00 Scholte, Trmp

Floodplain rehabilitation and the future of conservation & development : adaptive management of success in Waza- Logone, Cameroon Scholte, P.T.

Citation Scholte, P. T. (2005, November 23). Floodplain rehabilitation and the future of conservation & development : adaptive management of success in Waza- Logone, Cameroon. Institute of Environmental Sciences (CML), Leiden. Retrieved from https://hdl.handle.net/1887/4290

Version: Not Applicable (or Unknown) Licence agreement concerning inclusion of doctoral License: thesis in the Institutional Repository of the University of Leiden Downloaded from: https://hdl.handle.net/1887/4290

Note: To cite this publication please use the final published version (if applicable). FLOODPLAIN REHABILITATION AND THE FUTURE OF CONSERVATION & DEVELOPMENT

2005 Tropical Resource Management Papers, No. 67 (2005); ISBN 90-6754-953-3

Also published as thesis (2005), Leiden University ISBN-10: 90-9019780-X ISBN-13: 978-90-9019780-7 Floodplain Rehabilitation and the Future of Conservation & Development

Adaptive management of succes in Waza-Logone, Cameroon

Paul Scholte

Preface

I first learnt of the Waza-Logone floodplain rehabilitation plans in 1992 and was immediately thrilled by its perspectives. I expected that reflooding could trigger a cascade of developments and provide a new élan in a rather paralysed environment. At that time I was working in nearby Chad in a Sahelian environment where improved management without substantial inputs was out of scope under the reigning economic conditions. In Waza-Logone, an area still endowed with spectacular wildlife, we could start this unique experiment, improving the livelihoods of fishermen and pastoralists, rehabilitating wildlife habitat and, ‘on top’, triggering an improved management system. This study follows the quest of preparing, implementing, observing, discussing and analysing floodplain rehabilitation and, as proved necessary, developing the human capacity to assure that both wildlife and humans would benefit from the rehabilitation interventions.

This study reflects the development of my activities in the Lake Chad Basin in the 1990s. My assignment from 1990 till 1993 at the Programme Ecologie Pastorale at the Laboratoire de Farcha (Chad) was an excellent preparation for the research I carried out in neighbouring Cameroon at the Waza-Logone Projet (Maroua, 1993- 1997) and at the Ecole pour la Formation des Spécialistes de la Faune (Garoua, 1998-2000). Temporarily back in the Netherlands, various organisations requested my presence in the region, through short-term assignments, in 2000-2003, allowing me to keep track of some of the latest developments.

The following chapters also reflect my personal development. I started in the Lake Chad basin as an ecologist with an interest for social and development activities, developing into an ‘environmental scientist’ in the CML-Leiden tradition. Yet, I did not neglect my field biological roots and was stimulated by the request to contribute to the ‘Important Bird Areas of Africa’ and ‘Mammals of Africa’. I discovered another challenge when teaching an MSc course on rangeland management at the Uni- versity of Ndjamena. I was captured by the exchange of experiences in professional education. In 1998 I was asked to initiate Garoua’s community conservation curriculum. Supervising students during their field research has been amongst my most rewarding experiences.

Presenting our Waza-Logone reflooding studies at congresses and in scientific journals, I realised how special our experiences actually were. The size and the impact of reflooding were at a different scale than floodplain rehabilitation studies in Europe or the USA. Reflooding was an excellent tool to test concepts of ecological change that have seldom been assessed at such a scale in practice. The landscape scale also widened the scope of this study from the ecological impact of floodplain rehabilitation to encompass also management issues such as planning and capacity building. The process of writing the papers and this thesis In the mid-1990s, I started reporting preliminary results of our work, discussing the set-up of our monitoring and presenting a series of base-line studies and inventories, which laid a foundation for the later reported studies. This thesis has slowly grown through the elaboration of individual papers that, although based on the Waza- Logone and Garoua work, became small projects with their own objectives and conclusions. This allowed me to take some distance from the project and concentrate on linkages with upcoming scientific discussions. In this process, a number of papers were ultimately discarded from this thesis, especially the older ones that were largely descriptive and the ones that did not sufficiently contribute to the connecting (scientific) thread of this study, see annexed list of background publications.

The first half of this study ‘Impact of reflooding’ is predominantly based on ecological science. Chapters vary in scope, set-up and pretensions. Presented work on vegetation (Chapters 3,4) and pastoralist responses (Chapter 7) was developed and carried out as part of systematic research. Chapters on waterbirds- and antelope dynamics (5,6) on the other hand, compile a large number of surveys, carried out at various times for various purposes. The variety of survey methods and personnel as well as their large temporal and spatial scale, prevent a strong analysis on causal relations. These latter chapters aim, based on best available science, to draw broad conclusions on developments in bird and antelope populations, the area’s ‘conservation assets’. Moreover, they are of value for the total picture to be drawn in Chapters 8 and 12.

The second half of this study, ‘Enhancing Conservation – Development Integra- tion’, is only occasionally linked with theory. This second half is needed however to lay the foundation for my aim to address the issue of adaptive management that is required to understand the role of floodplain rehabilitation and how it may enhance conservation-development integration. This study obliged the use of different research styles: ecological science, qualitative participatory methods and occasional steps into social and education science. I do not pretend to master these latter, and limited myself to a straightforward methodology based on case-based reasoning, greatly assisted by exchanges with subject specialists and reviewers.

Hydrology and fisheries’ ecology also called for attention in this floodplain environment. At the Waza-Logone project, hydrologists and fishery researchers with whom I closely co-operated, have reported their findings in a series of field reports. A scientific synthesis would require more time and a different expertise than I possess.

The title I initially had chosen for this study was ‘Rehabilitating the floods, for people or wildlife?’ as this issue had remained at the forefront of my field research. This question was already answered in the late 1990s when finalising a first version of Chapter 8 (‘time-bomb’). At that time, my work was already aimed at the development of ways to redirect the conservation-development balance. The present title reflects this ongoing quest. Remarks on terminology The studies presented are the result of the work of a large number of people with whom I worked in the field, as authors or as reviewers. Throughout this book I use ‘we’ to stress common work and ideas of my colleagues and I, and ‘I’ when I refer to my own specific opinion.

Throughout this book, indicated years of ecological monitoring refer to the rainy and flooding season and not necessarily the calendar year when the consequences of the reflooding were monitored. Vegetation composition monitored in May 1994 for example, was referred to as 1993 vegetation because 1993 was the year of the relevant growing (= rainy + flooding) season. To allow comparison of flooding differences, also bird and antelope counts have been referred to the relevant rainy and flooding season. The January 1995 waterbird counts and late April 1995 wildlife counts were thus indicated as 1994 counts. In descriptions not specifically related to ecological monitoring (Chapters 2, 8-11), years are indicated as calendar years.

Nomenclature of woody plants follows Geerling (1982), grasses van der Zon (1992), and other plant species the second edition of the Flora of West Tropical Africa (Hepper, F.N. 1954-1972). Names of birds follow the checklist of the area (Scholte et al. 1999), names of mammals follow Kingdon (1997).

Despite the multitude of more or less similar expressions for Community Conservation, I have used the term (Integrated) Conservation-Development (Project) as most relevant term for the activities conducted by the Waza-Logone Project.

Table of Contents

PART I – Introduction

1 General Introduction and Study Outline 13

2 The Ecological History of Waza-Logone: Constructing a reference image for floodplain rehabilitation 65

PART II – The Impact of Reflooding in Waza-Logone

Impact on Vegetation 3 Floodplain Rehabilitation in North Cameroon: Impact on vegetation dynamics 89

4 Maximum Flood Depth Determines Above-ground Biomass in African Seasonally Shallowly Flooded Grasslands 107

Impact on Wildlife 5 Waterbird Recovery in Waza-Logone (Cameroon), resulting from increased rainfall, floodplain rehabilitation and colony protection 129

6 Antelope Populations in Waza National Park (Cameroon) from 1960 till 2001: Impact of changes in rainfall, hydrology and human pressure 145

Impact on Pastoralists 7 Pastoralist Responses to Floodplain Rehabilitation in North Cameroon 161

PART III – Enhancing Conservation – Development Integration by Management Planning and Training

Risks: the overshoot of success 8 Immigration: A potential time bomb under the integration of conservation and development 185 Grip on the Whole: management planning 9 At the Interface of Legislation and Wildlife Management: A decade of experience in consensual protected area management planning in Cameroon 203

Foundation: development of human capacities 10 Curriculum Development at the African Regional Wildlife Colleges, with special reference to the Ecole de Faune (Cameroon) 233

11 Wildlife Managers’ Perceptions of Community Conservation Training in West and Central Africa 251

PART IV – Synthesis

12 Floodplain Rehabilitation and the Future of Conservation & Development: Synthesis and concluding remarks 269

References 299

Summary 321

Samenvatting 329

About the Author 337

List of Background Publications 338

Acknowledgments 340

The Photographic Outline is found between Chapter 1 and Chapter 2 33 PART I

Introduction

1 General Introduction and Study Outline

2 The Ecological History of Waza-Logone: Constructing a reference image for floodplain rehabilitation

1 General Introduction and Study Outline

1.1 Introduction to African floodplains and their rehabilitation

‘From a European’s point of view Nuerland has no favourable qualities, unless its severity be counted as such, for its endless marshes and wide savannah plains have an austere, monotonous charm. It is throughout hard on man and beast, being for most of the year either parched or a swamp. But Nuer think that they live in the finest country on earth and, it must be admitted, for herdsmen their country has many admirable features’ Evans-Pritchard (1940) introducing the ecology of the Sudd (South Sudan), the largest stretch of seasonally flooded grassland in Africa.

African floodplains The fertility of floodplains is legendary and some of humankind’s main civilisa- tions, such as in Egypt and Mesopotamia have developed around them. Floodplains are fed by the regular spilling, once or several times a year, of river water, loaded with sediments over the levees into the surrounding plains. The great fluctuations in water level create a seasonal cycle of flood and drought, allowing a high primary production, abundant wildlife and often a high human population density.

Nearly all African rivers have fringing floodplains, but the vast areas of seasonally flooded grasslands, which are so important to the African economies and wildlife, are associated with rivers that have seasonal rainfall catchment areas (Denny 1993). The largest are the Inner Niger Delta in Mali, Lake Chad and floodplains associated with its tributary rivers (Cameroon, Chad, Nigeria), the Sudd in Sudan, the Zambezi valley floodplains with the Kafue Flats and Barotse plains in Zambia and the Okavango Delta (Botswana) (Fig. 1.1). These floodplain rivers have diverse and abundant fish faunas, which support some of the richest inland fisheries (Wellcomme 1979). After the floods recede, herbivores, wild as well as domestic, find abundant and high quality grazing in the floodplain grasslands and reach densities seldom encountered in other habitats (de Bie 1991). Other expressions of these floodplains’ richness are the large number of waterbirds, resident as well as migratory. All major sub-Saharan African seasonally flooded grasslands (Fig. 1.1) feature prominently in the list of Important Bird Areas (Fishpool & Evans 2001). This biological abundance notwithstanding, ‘African wetlands may not qualify other than second-order hotspots especially as concerns endemism’ (Myers 1997), indicating the frequent bias of the ‘biodiversity’ concept towards species diversity (Denny 1994; Scholes & Biggs 2005). Species-richness assessments of African habitats therefore tend to neglect the role of seasonally flooded grasslands (e.g. Fjeldsa et al. 2004) and relatively few of them have an effective protection status (WCMP 2004). 16 Part I – Introduction

No and name Country Area at peak flood (km2) Floodplain type

1 Senegal Delta Senegal 8 000 Coastal delta 2 Senegal Valley Senegal 5 000 Fringing floodplain 3 Inner Niger Delta Mali 30 000 Internal delta 4 Niger fringing plains Nigeria 5 000 Fringing floodplain 5 Niger Delta Nigeria 36 000 Coastal delta 6 Volta River Ghana 8 500 Fringing floodplain 7 Benoué River Nigeria 3 100 Fringing floodplain 8 Benoué River Cameroon 1 000 Fringing floodplain 9 Logone Floodplain Cameroon 5 000 Fringing floodplain 10 Chari and Logone Chad 63 000 Fringing floodplain 11 Congo River Congo ? Fringing floodplain 12 Barotse plain Zambia 10 750 Fringing floodplain 13 Kafue Flats Zambia 4 300 Fringing floodplain 14 Okavango Botswana 17 000 Internal delta 15 Shire River Malawi 1 000 Fringing floodplain 16 Kifakula Depression Congo 1 500 Fringing floodplain 17 Kamulondo Congo 12 000 Fringing floodplain 18 Sudd Sudan 92 000 Fringing floodplain 19 Tana Delta Kenya 1000 Coastal delta

Adapted from Drijver and Marchand 1986 and Welcomme 1979.

Fig. 1.1 – African Seasonally Flooded Grasslands 1 – General introduction and study outline 17

Since about 5000 BC, when the earliest systematic colonisation of the Nile, Meso- potaminan and Indus rivers occurred, there has been concentrated human efforts to tame the floods for agriculture and other purposes. Not all these developments have been sustainable. Overexploitation of renewable resources led to the collapse of some ancient societies, such as in Mesopotamia with its irrigated agriculture (Janssen & Scheffer 2004). Land uses that did not interfere with the hydrological regime such as grazing and associated burning and extensive agriculture, have also had a profound influence on the natural vegetation cover (Drijver & Marchand 1986).

The hydrology of the wetlands of sub-Saharan Africa remained, until about 50 years ago, relatively unchanged explaining their general natural appearance (Denny 1993). More recently, however, demand for water and electricity has resulted in an increasing number of water management projects (Drijver & Marchand 1986). Off-shoots off these developments were varied and include expansion of inland fisheries and irrigation schemes. Large man-made lakes not only flood valleys but also regulate the water flow in catchment and drainage areas, thus depriving the seasonally flooded grasslands of their main inputs of water and sediments (Denny 1993). The loss of seasonally flooded grasslands often meant the loss of the resource base on which local communities and wildlife depend, with devastating impacts (McCully 2001).

Few of the large African seasonally flooded grasslands are effectively protected, although recently efforts have been undertaken to designate large parts of the Inner Niger Delta, Lake Chad and others as Ramsar sites (WCPA 2004). But even if these floodplains have an internationally recognised protected area status, such as Waza National Park and the Kafue Flats in Zambia, this has not necessarily stopped the upstream construction of dams. Lack of impact assessments was blamed for many of these interventions (Drijver & Marchand 1986). Even though such assessments are increasingly undertaken, as the one of the Jonglei canal in Sudan (Howell et al. 1988), large scale hydrological interventions have continued on the African continent (McCully 2001).

Floodplain rehabilitation A number of studies has shown the economic irrationality of large-scale hydrological interventions in African floodplains, showing that their costs, including environmental ones, largely exceed benefits (Drijver & Marchand 1986; Marchand 1987). These studies and rising popular protests (McCully 2001) have led to an emerging tendency to discuss the rehabilitation of these floodplains to their original functions. 18 Part I – Introduction

Unfortunately the knowledge about African wetlands remains poor (Denny 2004) and the few studies on the impact of processes, such as water releases from reser- voirs, do not surpass a ‘black box’ stage (Hollis et al. 1993). One is therefore inclined to look at experiences with floodplain rehabilitation experiments in Europe and the USA that have lately taken flight (INTECOL 2004). These nature restoration programmes are often carried out on a sub-landscape level only, in areas measured in hectares and not in square kilometres. With such applied scales it is not surprising that most restoration programmes are based on a sectoral nature conservation approach, studied by predominantly natural sciences. Larger-scale rehabilitation efforts almost always associate people, and require interactions of natural and human sciences. The subsequent increasing complexity makes it particularly difficult to transpose experiences from one place to another.

These observations were earlier voiced by Schouten (2001) and Nienhuis (1998) in their comments on rehabilitation experiments in the Netherlands and Europe respectively:

‘In many cases, nature rehabilitation cannot be isolated from the rehabilitation and management of the landscape...nor can the landscape be considered in isolation from various socio-economic functions. The integration of these layers of interest forms one of the biggest and most urgent challenges of spatial planning in the Netherlands’ (Schouten 2001).

and

‘...experiences in Europe with the rehabilitation of large rivers are rare, relative to smaller streams, due to the large numbers of interwoven societal demands on the river, the connected economic costs and the complexity of the physical and biological systems involved. Proposals and concepts for large river restoration are much more abundant than demonstrations in the field. The empirical large-scale testing of the models, connected to the political willingness of catchment-scale restoration, are forming the real bottlenecks’ (Nienhuis et al. 1998).

The Everglades (USA) hosts the best-known large-scale (sub) tropical wetland rehabilitation program, aiming at restoring parts of the 5000 km2 degraded area (Kiker et al. 2001). Compared to African seasonally flooded grasslands, the Ever- glades’ oligotrophy, the dominance of Cyperaceae and the limited role of wild and domestic large herbivores is striking. Most striking differences are related with the area’s socio-economic context as can be expected from its location in the USA. This is illustrated by the Everglades restoration programme’s budget of approximately 10 billion USD. In contrast, the budget of the Waza-Logone floodplain rehabilitation totals 20 million SFr, i.e. 1% of the recovery cost per km2 of the Ever- glades (Kiker et al. 2001; IUCN 1999). Despite the multitude of technical studies, 1 – General introduction and study outline 19

many questions remain on the impact of restoration on the Everglades ecosystem. This has led to a still rather basic restoration vision ‘to mimic as closely as possible the appearance and behaviour of the system as if drainage and development had not occurred’ (Davis & Ogden 1994). The lack of a common vision is further illustrated by the repeated delays in the implementation of the Everglades restoration program (Grunwald 2002).

One can conclude that for the large-scale rehabilitation of tropical floodplains in developing countries, not only demonstrations hardly exist, but also the concepts and proposals (‘models’) have yet to be formulated. The present study analyses a floodplain rehabilitation test at a landscape level in a tropical floodplain ecosystem in North Cameroon and aims to contribute to the development of these concepts and proposals for future rehabilitation trials in other African floodplains.

1.2 Introducton to ‘Conservation & Development’ and capacity building needs

The community conservation panacea The image with which conservation organisations in Africa present themselves has changed over the last 20 years from wildlife protection to people-oriented conservation (e.g. Adams & Hulme 2001). The protection focus of the colonial and early independence period was reflected in autobiographies of the last European directors in charge of protected areas in Sub-Saharan Africa with titles such as ‘Mourir pour les éléphants’ (Verschuren1 1970), ‘Les guardiens de la vie sauvage’ (Dupuy2 & Dupont 1984 and ‘An impossible dream’ (Parker & Bleazard3 2001).

In the early 1990s, publications such as ‘The Myth of Wild Africa’ (Adams & McShane 1992) and ‘Out of Eden’ (IIED 1994) reflected the focus of an emerging conservation brand that refuted the concept of a ‘wild Africa that had to be protected against Africans’. The new ‘leitmotif’ was ‘conservation with, instead of against people’ and ‘win-win scenarios’ (Inamdar et al. 1999). This new conservation paradigm, so-called Community Conservation, was considered a means of re- conciling Conservation and Development by ensuring that the interests and knowledge of local communities were taken into account (Adams & Hulme 2001). Local communities were to be provided with trade-offs of the protected area through a share in tourism and hunting revenues or the (legal) use of specific natural resources. Alternatively, compensations should be provided for the opportunity (non-

1 Former director of wildlife of DRC (former Zaire). 2 Former director of wildlife of Senegal. 3 Game wardens in Kenya at the end of the colonial period who dedicate their book to ‘a Pleistocene Africa which we so enjoyed and sought to preserve, but which is gone’. 20 Part I – Introduction

exploitation) costs of protected area resources by the initiation of small-scale development projects. For some, community conservation was also considered a pragmatic way of dealing with increasingly poor and often corrupt governments (Adams & Hulme 2001). Community Conservation (CC) embraces a range of initiatives such as Integrated Conservation-Development Projects (ICDPs), Community-Based Conservation (CBC), Community-Based Natural Resource Management (CBNRM) (Adams & Hulme 2001). These expressions are often used as synonyms, despite their different (regional) origins.

The marriage of Conservation with Development has become an attractive target for international funding that, at the end of the 1990s, totalled annually hundreds of millions of dollars (El Ashri 2001). The development of the Community Con- servation concept in Africa was based on experiences of which the CAMPFIRE initiative on wildlife utilization in Southern Africa is one of the best known (Child4 1995). The community-based mountain gorilla tourism in Rwanda and Uganda with its high revenues per tourist funnelled into local communities was another experience that stimulated the early development of Community Conservation in Africa (Weber & Vedder 2001).

Community conservation, the debate In the late 1990s, however, critical reports on Conservation-Development projects began to emerge (Berret & Arcese 1995; Kramer et al. 1997; Terborgh et al. 2002). Some biologists went further and questioned the underlying principles of Com- munity Conservation and pleaded for a return to the ‘fortress Conservation’ approach (Oates 1999; Spinage 1998). This so-called ‘pro-park lobby’ is in an unlikely alliance with critics who detect in community conservation a shallow façade to hide old-style preservation (Adams & Hulme 2001). Although few protected area managers seem to follow this criticism, it has led to an emerging consensus that Community Conservation is not the panacea that it initially had appeared. Com- munity Conservation has its limits when wildlife resources or socio-economic circumstances do not allow their sustainable use in sufficient quantities compared to the opportunity costs of the protected areas (Adams & Hulme 2001). The potential of tourism and safari hunting has, most notably, shown to be more limited than previously assumed (Wilkie & Carpenter 1999ab). The practical organisation of community conservation has also proven to require long-term investments and stable and decentralised institutional environments (Adams & Hulme 2001).

Sometimes, development near protected areas may do more harm than good. Ex- periences from West Africa, the Central African Republic and the one presented in this study have shown that development in the vicinity of protected areas may appear to be a Trojan Horse jeopardising conservation (Noss 1997; Oates 1998;

4 Former director of wildlife of Zimbabwe. 1 – General introduction and study outline 21

Scholte 1998). Outside Africa, the Galapagos islands are an example of the risks of developing ecotourism in a prime conservation area thus triggering massive immigration and over-exploitation (Grenier 2000). Some, like Oates (1999), have therefore pleaded for a total ban on Community Conservation.

Developing tools One may argue that not enough has been done to develop the tools for Conser- vation-Development approaches that should also take into account the risks of development. At present Conservation-Development approaches through Inte- grated Conservation Development Projects remind us of the expression ‘If your only tool is a hammer everything looks like a nail’. Or, as Adams & Hulme (2001) concluded pragmatically ‘The real issue is not whether conservation should be done with people, but how’.

Developing the capacity of park managers The changes mentioned above in conservation image notwithstanding, one has the impression that daily African wildlife management practice has, since the days of the last European wardens, little changed in its pre-occupation with poaching control as described by Parker & Bleazard (2001). The title ‘Wildlife Wars’ of the autobiography of the former director of the Kenya Wildlife Service (Leakey & Morell 2001), covering the period in the 1990s, is meaningful. I attribute this ‘business as usual’ at least partly to the limited efforts to involve protected area personnel in the design and implementation of conservation-development (see also Western 2003). ‘Approaches to natural resource management... tend to assume that the manager is outside the system being managed. However, where the objectives include long-term sustainability, ... the managers [are] integral components of the system’ (Walker et al. 2002). With the increasing international attention for Com- munity Conservation, the role of protected area managers has been further erod- ed into a caretaker with an agenda (i.e. management plan) filled in at the nation’s capital. This situation contrasts the early days of independence when African Park wardens were amongst the best-educated personnel in their country.5 Early graduates of the regional African wildlife colleges in Mweka and Garoua became later in their careers minister or occupied other high-ranking posts in their countries, a situation increasingly difficult to imagine. Their high-ranking positions allowed them to influence developments beyond the borders of the protected area.

In this study I develop a pragmatic approach on Conservation-Development. The study further aims at understanding the role of management planning and the potential of raising the capacity of protected area managers, to offer a perspective of a ‘future for Conservation & Development’.

5 Parker and Bleazard (2001) show that although colonial park wardens in Africa were materially not well off, they enjoyed high status. After independence this status has dropped gradually. 22 Part I – Introduction

1.3 Reflooding in Waza-Logone6

This study discusses the quest for adaptive management that was started in 1993 with reflooding parts of the Waza-Logone floodplain in North Cameroon. The reflooding aimed to rehabilitate vegetation, birds, wildlife, pastoral grazing and fisheries that had come under pressure as result of the construction of an upstream dam and embankment in 1979. Initially, our attention was aimed at the monitoring of the impact of the reflooding and the design and discussion of large scale reflooding options. During this process, our attention shifted to the mitigation of existing and rising conflicts in the floodplain, in particular the one between Waza National Park and the surrounding communities. Learning by doing not only has given insight on the impact and consequences of reflooding but also highlighted the need to know more about the condition of the floodplain prior to the construction of the dam in 1979.

Following the widespread famine during the 1970s in the Sahel, the Government of Cameroon, assisted by several donor countries, initiated a number of large-scale irrigated rice schemes. One of them was a gravitation irrigated rice scheme, for which a dam and an embankment along the Logone river were constructed to form the Lake Maga reservoir (Fig. 1.2). In combination with lower than average rainfall,7 the depth, duration and extent of downstream inundations were reduced, which led to serious ecological degradation of the Logone floodplain, both in and outside Waza National Park (see Fig. 1.2). Annual grasses invaded the productive perennial grasslands, limiting regrowth in the dry season and reducing carrying capacity for wildlife and livestock (Scholte et al. 1996a). Fishing resources also dropped dramatically, provoking an emigration of a significant part of the human population from the floodplain (Drijver et al. 1995). Leiden University, in collaboration with the Garoua Wildlife College, initiated several impact studies that showed not only the devastating impact of the rice scheme but also its failure in food production. In the 1980s, protests of local populations were still considered taboo; when desperate floodplain inhabitants tried to reopen a watercourse they were deterred by the army. Despite all investments, rice cultivation did not become successful and in the early 1990s half of the irrigated rice scheme area was left un- used. A climate was thus slowly opened in which measures could be discussed to mitigate and even counteract the impact of the dam and the embankment.

This allowed the start of the Waza-Logone project in 1992, a collaboration between the Cameroonian Ministry of Environment and Forestry and the World Conserva- tion Union (IUCN), the Institute of Environmental Sciences (Leiden University), the Netherlands Development Organisation (SNV) and WWF-The Netherlands.

6 The Waza-Logone area will be introduced in Chapter 2 and the photographic outline. 7 As measured since the 1930s (Beauvilain 1995). 1 – General introduction and study outline 23

Fig. 1.2 – The Waza-Logone Area 24 Part I – Introduction

The overall goal of the project was to assist the government of Cameroon in pursuing the integrated management of natural resources in the Waza Logone region, so as to provide a sustainable livelihood for the local people and to maintain the diversity of the biological resources of the floodplain.

Amongst the specific objectives8 were: • To investigate, design and implement an operational plan for the rehabilitation of the hydrological regime of the Logone floodplain • To formulate and test an operational action programme for the management and development of the area around Waza National Park (IUCN 1999; Loth & de Iongh 2004).

Studies of the Waza-Logone project in 1993 identified a number of reflooding options, which were expected to have a major impact on the floodplain, including its human population. When in early 1994, the hydrologists Emmanuel Naah and Hans Wesseling, topographer Paul Kouamou and I visited the floodplain after a year of reviewing studies on hydrology, ecology and land use, we realised how little we knew on the possible impact of reflooding. The extremely flat area, with a slope of no more than a few centimetres per km, made that even basic questions such as those on the direction of flooding had to remain unanswered, let alone the impact of reflooding on vegetation or land use. The plan described in the project document, to prepare and implement a 1000 km2 reflooding in one step, looked increasingly grotesque.

Although not planned and budgeted for, subsequent discussions held at the Waza- Logone project lead to a strong backing of a pilot release as a learning by doing experiment. An ideal candidate, without the risk of flooding habitations and crop- land, was the opening of the Petit Goroma, a watercourse that was cut off by the embankment along the Logone river (Fig. 1.2). After discussions with communities and authorities the pilot floodplain reflooding was initiated by breaching the embankment, reflooding from September 1994 onwards 180 km2 of desiccated floodplain and raising water levels in an additional area of approximately 600 km2.

The impact of the reflooding, already in the first year, on especially water levels, fish production and grazing exceeded our expectations. But should we attribute this to

8 Although formulated as specific objectives, these are in fact activities or outputs of the project. Also mentioned: • To contribute to ongoing government efforts towards regional development planning by means of field evaluations of the environmental impacts of existing development projects. • To develop a methodology for the design and implementation of conservation and development activities in other situations in the Sudano-Sahelian region. • To strenghten institutional capacity to manage the natural resources of the Waza-Logone region, in particular through training of national staff participating in the project. 1 – General introduction and study outline 25

the opening of the embankment or were there other factors, such as the favourable rainfall, that interfered? The large impact of the pilot reflooding stressed the need to strengthen the management of Waza National Park and its direct surroundings to cope with the rapidly changing situation. Especially the attraction of pastoralists and fishermen to exploit the newly reflooded areas drew our attention. At the same time a hitherto relatively strong governmental presence, including a relatively well equipped and staffed Waza National Park, was disappearing. Solutions had to be developed to mitigate the negative impacts of these changes. Through the formulation of a management plan, such measures were formulated and programmed with local communities, sedentary as well as nomadic, protected area authorities, traditional and administrative authorities and the Ministry of Environ- ment and Forestry. Experiences with the implementation of this management plan highlighted, however, the low capacity of park personnel to deal with new forms of collaboration with local communities. This motivated the training of not only park guards of the national parks in north Cameroon, but also at the Garoua Wildlife College, thus increasing the impact beyond the Waza-Logone area.

The Waza-Logone pilot reflooding was a unique large-scale experiment in tropical ecosystem and land use recovery. The induced changes in floristic composition and vegetation productivity challenged vegetation dynamics theory and generally assumed flooding – productivity relationships. Monitored changes in waterbird and antelope numbers questioned a simple recovery mechanism and highlighted interactions of wildlife with competing land uses, in particular pastoralism that showed a steady increase in grazing intensity. With an impact zone of several hundred square kilometers, the reflooding covered site-specific and landscape issues, linking local impacts, e.g. on vegetation composition, with region-wide changes, such as on transhumant pastoralism patterns.

1.4 Research questions and study outline

The present study is organised in four parts. The first introduces the area and available information sources. The second part discusses the impact of the reflooding induced in 1994 on the Waza-Logone floodplain. The third part questions the risks of this floodplain rehabilitation that were mitigated by management planning and training. The fourth part syntheses these experiences.

In more detail, and mentioning the respective chapters where the information can be found, the research questions and the outline of this study is as follows. 26 Part I – Introduction

PART I – Introduction

How has our reference image of Waza-Logone developed in time?

The present knowledge of the ecology of Waza-Logone has developed in a period of widely fluctuating climatic conditions and a rapidly changing land use. I start this study with a presentation of the main sources of information on Waza-Logone in a time sequence over the last 150 years (Chapter 2). The presented historical dia- gram shows the link between available information and the reference image for floodplain rehabilitation. Presented information also stresses the ecological significance of Waza-Logone, even in its degraded post-dam period, justifying the rehabilitation efforts. The papers cited in the list of background publications provide additional information on these issues.

PART II – The Impact of Reflooding in Waza-Logone

The overall research question is: Does reflooding lead to the restoration of the Waza- Logone floodplain to its pre-dam structure and conservation and development functions?

In other words, How resilient is this floodplain environment after the dam construction? Has it switched into another state without possibility to return to its pre-dam state?

Scientists active in the Waza-Logone area in the 1980s and early 1990s attributed much of the observed degradation of the Waza-Logone area to the desiccation following the Maga dam construction, a process that has taken over 15 years and was still under way at the start of the 1994 reflooding. Moreover, it occurred in a period of drought stress and with a rapidly changing land use. We did not take for granted that reflooding would automatically reverse this process. In the meantime, population pressure in and around the floodplain had increased considerably and it was not excluded that rehabilitated resources could rapidly become overexploit- ed again. In this study the question of ecosystem restoration is answered along three lines, (1). vegetation composition and productivity, (2) numbers of waterbirds and antelopes and (3) pastoral recovery. The question of part II is answered under these latter three subheadings and their respective sub-questions, in five chapters:

Does reflooding lead to a 100% perennial grass cover with the same floristic composition and production that existed prior to the Maga dam construction? What are the mechanisms with which these changes take place? 1 – General introduction and study outline 27

Vegetation composition In Chapter 3 we synthesise the main findings of unpublished studies in the 1960s and 1970s, prior to the Maga dam construction and after the dam in the 1980s on the changes in vegetation composition, that were our de facto reference image. My colleagues and I systematised this monitoring to study the vegetation changes triggered by the 1994 pilot reflooding. We expected that reflooding would induce the restoration of perennial vegetation in five years to its pre-dam situation, characterised by the perennial grasses Echinochloa pyramidalis, Vetiveria nigritana, Hyparrhenia rufa and Oryza longistaminata (Wesseling et al. 1994). This optimism was based on observations in a 40 km2 area in the central, relatively well flooded floodplain where the water level was raised by the construction of a small dam (Drij- ver & Kouahou 1995). Studies on hydroseries, the sequence of plant species along a flooding gradient, explained the impact of this water raising (van der Zon 1992). This use of hydroseries is, however, not uncontested (e.g. John et al. 1993; Leen- dertse et al. 1997). The area downstream of the Maga dam had become entirely desiccated and a slower or only partial vegetation recovery could alternatively be expected.

Above questions are linked to the discussion on ‘ (Gleasonian) Succession versus State and Transition’ alternatively called ‘gradual changes’ versus ‘catastrophic shifts in ecosystems’ (Scheffers 2001). This discussion has taken flight in rangeland management in the 1980s (Westoby 1980; Westoby et al. 1989, see also Rietkerk et al. 1996), followed by aquatic and other ecosystems (e.g. Scheffer et al. 1993). Floodplains have been considered environments where cyclic (Gleasonian) succession takes place; the ‘lineair’ Clementsian succession model (evolving into a single state) is obviously of little use (Van der Valk 1992; Middleton 1999). The alternative State and Transition model predicts a different scenario, in which sudden switches into different vegetation communities will take place only once a threshold has been passed. This unpredictably threshold may be difficult to pass hampering necessary interventions to reach the desired state. Following this model, reinstalling the pre-dam flooding regime would not automatically lead to the restoration to the pre-dam vegetation.

Vegetation productivity Many floodplain plant species are said to have higher rates of production in less flooded conditions (Middleton 1999), but this reasoning may be biased towards temperate zones. We expected that perennial grasses in Waza-Logone would show an increased production with reflooding, but no reference information existed on the impact of the Maga dam on vegetation production. We further hypothesised that in case of increased production there would be a time lag between the moment of water raising and the full expression of vegetation production increase, because of the predominance of perennial rhizomatous grasses of which a large part of the biomass is in their rhizomes. In Chapter 4 I therefore linked maximum flood depth 28 Part I – Introduction

with above-ground biomass at the end of the wet season allowing an assessment of the impact of the 1994 pilot reflooding.

Does reflooding lead to the recovery of birds and wildlife to their pre-dam numbers?

Waterbirds Waterbirds constitute one of the principal conservation assets of the Waza-Logone area (Fotso et al. 2001), but have hitherto received little attention other than occasional, qualitative observations. The only quantitative information on the occurrence of waterbirds prior to the Maga dam was on Black-crowned Crane, one of the area’s flagship species that numbered at least 10 000 individuals in the 1970s, almost as much as the present entire subspecies population. The resident Black- crowned Crane depends on moist grassland habitat during the crucial nesting and dry season periods. Its population drop to about 2000-2500 individuals in the early 1990s was therefore attributed to the Maga dam construction (Scholte et al. 1999; Sinibaldi et al. 2004). A comparison of the dominant habitats showed the importance of perennial grasslands and 20-40 cm deep water for waterbirds (Scholte et al. 2000a). Against this background, Chapter 5 investigates whether reflooding, that increased the extent of shallow water and allowed the recovery of perennial grasslands, has resulted in increased numbers of waterbirds. The relatively favourable rainfall that characterised the period since the reflooding, may also have influenced the observed waterbird numbers. This also holds for protection, especially of vulnerable colonies.

Antelopes Waza National Park is one of the last remaining areas in dryland West-Central Africa with abundant wildlife. Antelopes are amongst the wildlife most prominently present in the area, but the floodplain antelope Kob and Korrigum (‘Topi’) in particular have undergone a dramatic drop in population numbers after the Maga dam construction. Chapter 6 assesses the development of these antelope populations from the early 1960s, allowing appreciating their possible (partial) recovery to the spectacular population size in the early 1970s. As discussed for waterbirds, the positive influence of the relatively favourable rainfall since the mid-1980s on the one hand and the increasing human pressure on the national park on the other hand was expected to influence possible reflooding impacts.

Does reflooding lead to full recovery of pastoral use of the floodplain?

Pastoralists Transhumant pastoralists are a wary population group in Waza-Logone, which has been largely neglected by developers and authorities. However, grazing of their estimated 200 000 cattle constitutes one of the main land uses in the Waza-Logone area. Our activities with pastoralists were based on confidence building through 1 – General introduction and study outline 29 the organisation of discussion sessions with authorities in the pursuit of solutions to the reigning insecurity that troubled their daily life. In Chapter 7, we present the impact of the 1994 pilot reflooding on the pastoral migration pattern and the resulting increasing cattle grazing intensities. Three scenarios were formulated to capture potential reactions to reflooding: an overshoot scenario with a higher grazing intensity than resources available, a territorial scenario with a lower grazing intensity and an ideal distribution scenario with a grazing intensity in line with the availability of resources. An important justification of the pilot reflooding was that the rehabilitated floodplain would become an alternative for grazing inside Waza NP. In the late 1980s and early 1990s incursions of cattle, searching for dry season forage and water, constituted a major source of conflict with the Waza National Park authorities.

PART III – Enhancing Conservation-Development Integration by Management Planning and Training

The overall research question is: If a rehabilitation is ecologically successful, how may then the ecosystem’s functions that underlie the balance between Conservation and Development be assured?

Risks: The overshoot of success Chapter 8 shows that pastoralists and fishermen, attracted by the newly available resources, compete with waterbirds and floodplain antelopes, thus threatening the balance of the pursued conservation-development integration. I subsequently analyse if comparable risks exist in other Integrated Conservation-Development Projects (ICDPs) and how these are dealt with. The immigration risks motivated the development of a policy, based on a categorisation of local communities, of which some first promising results are presented. The experiences from Waza-Logone are of great interest for other Conservation-Development Projects whose impacts are often difficult to distinguish from other development impacts.

Grip on the whole: Management planning Participatory studies showed that exploitations from inside Waza National Park, despite their illegality, have continued under the presence of, till recently, fair numbers of park guards. These observations refute the ‘victim image’ that sometimes is held of local communities living around protected areas (IIED 1994). Moreover, the continuing exploitation of park resources, reinforced by the results of the reflooding monitoring, question if the discussed ecosystem rehabilitation also leads to the recovery of wildlife and improved land use, the reference image sketched in Chapter 2. A new management policy on the status of the park villages was one of the outcomes of the Waza NP management plan, the first of its kind in Cameroon and indeed Central Africa. The plan further offered a consensual interpretation of the 1994 environmental law to control the continuing exploitation of resources 30 Part I – Introduction

from inside Waza National Park. The plan also allowed the creation of a management committee in which, for the first time, park authorities and local communities discussed problems related with the park management. I discuss the formulation process of this management plan and review its outcome based on an independent evaluation in 2002. These lessons are put in a perspective of some of the key issues of ongoing management planning in African protected areas (Chapter 9).

Foundation: Development of human capacities The policies analysed were facilitated and driven by third parties, i.e. international NGOs and expatriates mainly. Protected area managers played a rather limited role only, due to lack of training and experience, undermining their involvement in the development of conservation-development integration. I address necessary training of protected area personnel on two levels. In Chapter 10 I address training requirements on an institutional level and discuss curriculum reforms at the Garoua regional wildlife college for West-Central Africa, through a comparison with the other regional African wildlife colleges in Eastern and Southern Africa. In Chapter 11, I evaluate courses addressing community conservation and park planning. These pilot courses were developed at the Garoua Wildlife College during the Waza management plan formulation and implementation. Presented experiences are also relevant for other protected areas in West and Central Africa.

PART IV – Floodplain Rehabilitation and the Future of Conservation- Development: Synthesis and Concluding Remarks

In the concluding Chapter 12, I review the outcome of the reflooding. I compare the reflooding responses that, because of their different response times, caused a domination of human supported resources. I continue analysing the contributions of management planning and capacity building to correct these undesired side effects of the otherwise successful reflooding. Based on these experiences I comment empirical concepts of ecosystem change most notably resilience and hysteresis. Especially the time dimension, that played such an overruling role in outcome of the floodplain rehabilitation, receives special attention. The normative ecosystem approach, is subsequently used to present the key findings of this study in a wider context.

Monitoring observations and insights allowed the reformulation of the floodplain rehabilitation expectations and hypotheses postulated above and discussed in the respective chapters. The consequences for the overall floodplain rehabilitation targets as programmed in Waza-Logone (IUCN 1999) and possibly in other (African) floodplains, are discussed in section 12.6. I conclude with reviewing the overarching adaptive management approach that is introduced below. 1 – General introduction and study outline 31

1.5 The quest for adaptive management

Adaptive management, can be loosely defined as the learning by doing, relies on an accumulation of credible evidences to support a decision that demands action (Walters & Holling 1990).

In addition to the discipline specific discussions introduced in the preceding para- graphs, I would like to present this book as test of an integrated approach of tropical floodplain rehabilitation.

The 1980s and 1990s witnessed a blossoming of concepts and theories which tried to capture the ecological as well as social aspects of the management of ecosystems. The Ecosystem Approach (Mitchell 2002), Ecosystem Based Management (Pirot et al. 2000), Resilience management (Walker et al. 2002) are some of the more recent expressions, inspired by Adaptive Management, introduced by Hol- ling (1978). Adaptive management was justified by the difficulties to assess the control of all aspects of ecosystem management motivating ‘a shift to adaptive management, which relies on the flexible, diverse and redundant regulation, monitoring for responsiveness and experimental probing’ (Holling 1995). Key desirable attributes highlighted are the capacity for learning and to be flexible and adaptive, hence the link that is often made with ‘learning organizations’ (Mitchell 2002; Salafsky et al. 2001).

Walters & Holling (1990) classified adaptive management into three categories, ranging from trial and error, to passive and subsequently active adaptive management. For some floodplain resources, i.e. waterbirds, lack of any quantitative reference information allowed us initially only to pursue a trial and error approach as outlined in the previous paragraph. For other resources, most notably perennial vegetation and floodplain antelopes, we started with a single hypothesis ‘requiring a natural habitat’, i.e. with as target ‘back to the pre-dam situation’. Over the years this passive adaptive management evolved into a more active adaptive management and alternative hypotheses of floodplain rehabilitation were formulated, amongst others related with favourable rainfall conditions and unfavourable conditions related to human pressure.

This thesis follows a thread of adaptive management: From [I] the conduct of inventories and overview of studies (Chapter 2 and annexes), allowing the formulation of an initial reference and target image and rehabilitation hypotheses (Chapter 1) to [II] probing, pilot reflooding & monitoring, impact studies on vegetation, waterbirds, floodplain antelopes and pastoralists (Chapters 3-7), to [III] problem analysis on the risks of conservation-development integration (Chapter 8). Subsequently solutions are planned through management planning and reviewing (Chapter 9) and capacity building is undertaken for future Conservation & Development (Chapters 32 Part I – Introduction

10, 11). [V] In the synthesis the contributions of the undertaken approach and the initially postulated hypotheses are evaluated and reformulated. Floodplain Rehabilitation and the Future of ‘Conservation & Development’

A PHOTOGRAPHIC OUTLINE

This story of Waza-Logone is long and has loops and many details. The following photographs and their captions illustrate the essential aspects of the story. They touch upon each of the chapters in the same sequence as the text. PART I – Introduction

The ecological background of the floodplain

The Adamaoua mountains, situated 500 km south of Waza- Logone, are the main catchment area of the Logone River.

The Mandara mountains, 100 km south-west of Waza-Logone, no longer drain directly into the Waza-Logone floodplain.

The dam and embankments of Lake Maga, built in 1979 for irrigated rice cultivation, block flooding of the Logone river and from streams of the Mandara Mountains into the southern Logone floodplain.

34 Part I – Introduction Within weeks of the floods receding, the area changes from a carpet of green lush grass into scorched earth.

The annual cycle of flood and drought characterises African floodplains. This thesis examines the ecological effects of this annual cycle, and what happens when is no longer occurs and when it is reinstated again. The reinstatement of the annual flood and drought cycle is what is meant by ‘reflooding’.

The ecological background of the floodplain 35 Acacia seyal shrubland borders the floodplain. It is not only an important source of food for wildlife, but also provides Arabic gum and firewood to local communities

The rocks near Waza town tower above Sclerocarya birrea woodland that dominates the sandier, slightly elevated parts of the western Waza-Logone area. Here wildlife concentrates during the rainy and flooding seasons.

36 Part I – Introduction Three species of vultures with Giraffes near a waterhole in the western part of Waza NP. The importance of Waza NP was recognised in its designation as an UNESCO-Man and Biosphere Reserve.

Concentrations of waterbirds and Kob in the eastern floodplain part of Waza NP. The national park and floodplain were selected as Important Bird Area and presented as the first Ramsar Site of Cameroon.

Lion escaping the heat. Spectacular wildlife such as Elephant and Lion attracts more tourists to Waza NP than to than any other protected area in Central Africa.

The ecological background of the floodplain 37 The floodplain people and the resources they use

Ramparts of Zina, a Kotoko town in the centre of the floodplain. For centuries, the Kotoko used to be the rulers of the floodplain, which has a long history of conflicts over floodplain resources.

The Kotoko used to be the exclusive owners of fish canals. This unique fishing technique drains depressions towards the main watercourses at the end of the flooding season. At the end of each canal the fisherman places his nets and traps almost the entire fish population that lived the previous season on the floodplain.

Fish canals from the air, taking receding waters towards the Logomatya, a branch of the Logo- ne River that can be seen in the background.

38 Part I – Introduction Dieguéré, a Musgum village on the edge of Waza NP. The Musgum are generalists who cultivate, fish and keep livestock. Most of them settled in the floodplain in the early 1900s.

Fishing on the plains near Zwang at the onset of the floods.

Sorting fish. In Waza- Logone, fisheries are second only to grazing in its economic importance. Andirni, an agropastoral village created in 1917 because of the proximity of several pools. Many FulBe and Arab agropastoralists settled in villages like Andirni south of Waza NP after the droughts of the 1970s.

Inside the house of Fannie, wife of Moussa Barka, a park guide living in Andirni. On the left the decorat- ed bed, showing the family’s transhumant roots. For years, Moussa and Fannie were our wonderful hosts.

40 Part I – Introduction Young men herd the cattle of the agropastoralists.

FulBe and Arab transhumants move with their families. This harsh life is increasingly being abandoned for life in villages and towns with health, marketing and schooling facilities.

The floodplain people and the resources they use 41 The ecological impact of the Maga dam

Deserted quarter of Mahé, in the north- east corner of Waza NP. After the Maga dam construction approximately one third of the sedentary and mobile population of the floodplain left.

Dead Giraffe in Talabal waterhole.

Dead Kob antelope. Following the Maga dam construction the Kob population crashed from 20 000 to 5 000. Its population further declined to 2000 by the end of the 1985 drought. Rinderpest hit the Kob hard during those years as well.

42 Part I – Introduction Since the completion of the Maga dam, waterholes are filled by tankers towards the end of the long dry seasons to help keep wildlife alive.

The only Vetiveria nigritana left in the western floodplain, where prior to the Maga dam this tussock grass used to be dominant. It is not the lack of water but especially the drop in perennial grass production that has caused the ecological catastrophe following the dam construction.

The ecological impact of the Maga dam 43 The lead-up to the reflooding and the first water coming through

The Maga rice scheme with large parts of its area left fallow. In the early 1990s, it had become increasingly clear that the Maga dam was an ecological as well as economic failure. The continuing import of Asian rice, available in the whole of northern Came- roon, testifies the failure of the Maga irrigated rice cultivation scheme. With the start of the Waza-Logone Project in 1992, the issue of reflooding was no longer a taboo.

Consultation of local people showed unanimous support for the reflooding. Local communities had never been consulted before and were hostile to the Maga dam.

44 Part I – Introduction Breaching of the embankment in 1994.

View from the Logone at the reopened ‘Petit Goroma’ in flooding, with on the left the remaining embankment. From September 1994 onwards, 20 000 litres water per second flowed additionally onto the floodplain for three months of the year.

The lead-up to the reflooding and the first water coming through 45 Car stuck near Andirni

The Waza-Logone project has put much effort in monitoring the impact of the (pilot) reflooding to prepare large-scale reflooding. Transport conditions were often difficult, it took one day to travel by car and boat to Zina, the centre of the floodplain. From there it was another two days by canoe or foot to the fur- thest monitoring sites.

Tracks become naviga- ble for canoes during the flooding season.

Following the monitoring transect from Zina to Waza NP on foot.

46 Part I – Introduction PART II – The Impact of Reflooding in Waza-Logone

Impact of the reflooding on vegetation dynamics

Installation of a monitoring grid, in the heart of the reflooded area where flooding and vegetation dynamics were intensively monitored. In 1994, prior to the reflooding, the grid was largely covered by Sorghum arundinaceum, a stout annual grass that invaded the floodplain during the absence of flooding in the mid-1980s.

Sorghum arundinaceum survived several years of flooding but it ultimately disappeared from the reflooded area. With its hard-to-digest reed-like stems containing few nutrients, it is of limited grazing value.

Aerial view of the eastern part of Waza NP. In 1996 Sorghum arundinaceum still covered large stretches of the floodplain, especially around waterholes, such as at Tchikam. (the woody vegetation is dark green, the shorter mostly biannual grasses, are clear green, whereas the stout Sorghum can be recognised by its beige-greenish colour). The transect running from Zina into the heart of Waza National Park was installed in 1984 by students of Leiden University. Covering always flooded, reflooded since 1994, and desiccated floodplain since the Maga dam, the transect was an important base for (qualitative) vegetation monitoring.

The typical perennial floodplain vegetation, with the rhizomatous perennial grasses wild rice (Oryza longistaminata) and Echinochloa pyramidalis. The cover of perennial grasses increased between 1993 and 1999 from 41 to 75% of the area affected by the reflooding. In 2003, almost the entire reflooded area was covered with perennial grasses again.

In the foreground (bi-) annual vegetation with its large open spaces, which allowed a recolonisation by the perennial floodplain grasses through lateral rhizomatous growth (from the background).

Vetiveria nigritana, a tussock grass, has not been able to recover in the reflooded area despite its good condition elsewhere in the floodplain. We attribute this set-back to the changed direction of flooding, with a lower quantity of sediments it carries.

48 Part II – The impact of reflooding in Waza-Logone Impact of the reflooding on vegetation production

This study clarified the link between maximum Measuring dry season regrowth following fire. flood depth and vegetation production. With a Regrowth, important because of its high quality 20cm rise in maximum waterlevel through reflood- and availability in the dry season, is expected to ing, above-ground biomass increased with approxi- increase only after several years when sufficient mately one third. rhizome biomass has been accumulated.

Kob antelopes gathering in area with fresh regrowth after burning.

Impact of the reflooding on vegetation production 49 Impact of the reflooding on waterbirds

Foraging crane family on perennial floodplain grassland. Estimated numbers of Black-crowned Crane dropped from over 10.000 in the early 1970s to 2000-2500 during the 1990s. The latter number still represents one sixth of the world population of the western subspecies of the threatened Black-crowned Crane. Black-crowned Cranes are present in Waza-Logone during the entire year and strongly depend on flooding for both breeding and foraging.

Black-crowned Crane nest built on an inundated Vetiveria nigritana tussock, the grass species that showed a dramatic decline after the Maga dam construction and has not yet shown any recovery.

50 Part II – The impact of reflooding in Waza-Logone Waterbird counting team of the Garoua Wildlife College and Waza-Logone project. Total numbers of observed waterbirds increased between 1992 and 2000 from 60 000 to 105 000. A combination of factors was found to be responsible for this increase, including improved rainfall (especially ducks and geese), floodplain rehabilitation (omnivorous storks and herons) and protection measures (some locally breeding storks and herons).

White-faced Whistling Duck. The increase in ducks and The increase of Black-headed Heron (an omni- geese in Waza-Logone corresponds to their recovery vorous heron) was attributed to a combina- over most of West Africa following the droughts in the tion of floodplain rehabilitation and colony 1980s. protection.

Impact of the reflooding on waterbirds 51 Black-headed Heron colony in Andirni. Because of protection by the village and floodplain rehabilitation, this colony has increased from 750 in 1993 to the exceptional size of 2500 nests from 1999 onwards.

Large fish-eating birds such as Yellow-billed Storks did not benefit from the reflooding because of repeated destruction of their colonies. Young birds are increasingly taken for consumption, not hindered by fishermen who believe that fish-eating birds reduce their own catch.

Destroyed Marabou Stork colony at Halé in 1997. Afterwards Marabou Storks moved to Kalya in the centre of Waza NP where the colony increased to more than 500 nests but was destroyed again in 2002.

52 Part II – The impact of reflooding in Waza-Logone Impact of the reflooding on antelopes

Park guard counting animals at Gobe waterhole in Waza NP. At the end of the dry season, only a hand- ful of water points remain, allowing fairly accurate counts of animals that drink daily, such as most antelopes.

Kob antelopes. The initial increase in number of Kob after the reflooding did not continue beyond 1997, possibly because of competition from increasing numbers of livestock.

Red-fronted Gazelle, a Sahelian species, the only species that has benefited from the drying out of the floodplain.

Impact of the reflooding on antelopes 53 The responses of pastoralists to the reflooding

Discussions with agropastoralists from Fadaré. Responses of mobile pastoralists to the floodplain rehabilitation programme were assessed through interviews with leaders of over 100 pastoral camps, which were held at the end of each grazing season from 1993 to 1999.

Park warden questioning a herder inside Waza NP. One of the goals of the reflooding was to stop the incursions of pastoralists into the park to reduce the competition between wildlife and livestock for dry season grazing.

54 Part II – The impact of reflooding in Waza-Logone Signing of the contract between the Fadaré pastoralists, Waza NP authorities and Waza-Logone project (December 1997). This formalised the verbal commitment of Yerima, a respected pastoral leader, that he and his followers would not longer send their herds into Waza NP.

Grazing intensity tripled from 1993 to 1999 due to a sharp increase in number of pastoralists migrating onto the floodplain each year. Because of this increase in livestock numbers, the condition of individual cattle has hardly improved. This leaves few incentives for pastoralists to refrain from grazing the good pastures inside Waza NP.

The responses of pastoralists to the reflooding 55 PART III – Enhancing Conservation – Development Integration by Management Planning and Training

The risks of a Conservation & Development project that is too successful

Lougouma, village on the border of Waza NP. If the park was really completely closed for the exploitation of natural resources such as wood, thatch grasses and pastures, the 15 villages surrounding it would not be able to stay where they are.

Newly built quarter at Lougouma village hosting people arriving after the improved flooding.

The Sultan of Logone Birni in front of his palace. Traditional authorities have lost much of their power and no longer control settlements in ‘their’ area. It is in the interest of village chiefs to allow an increase in population, as it gives them more power and financial means. As also the government has little control over these settlements, there is a need for developing alternative local management structures.

56 PART III – Enhancing conservation – development integration Fishermen returning from fishing in Waza NP. After the reflooding, the number of fishermen increased by a third, whereas the number of seasonal fishermen increased even more.

Kalamaloué National Park, where the disappearance of wildlife due to human encroachment is a bleak perspective for nearby Waza NP.

The risks of a Conservation & Development project that is too successful 57 The difference that management planning can make

Practising a new fire policy in Waza NP. Traditionally, management plans focus on technical and ecological issues, which were also included in the Waza Management Plan.

Removing mud out of the waterhole at Kalya, in the centre of Waza NP. As long as the flooding remains inferior to the pre-dam period, the maintenance of waterholes will be an essential management activity in Waza NP.

Elephant poached inside Waza NP. Also with increased local community involvement in conservation, anti- poaching remains an important activity because of the continuing presence of well-armed commercial poachers.

58 PART III – Enhancing conservation – development integration Triggered by Participatory Rural Appraisals in each of the villages around Waza National Park, a dialogue was started between local communities and national park authorities. This was an integral part of the Waza management planning, building consensus on sensitive issues such as demarcation of the national park and the exploitation of resources such as straw and Arabic gum in the park.

Meeting of the Waza NP committee, with representatives of park villages and transhumant pastoralists. One of the most significant results of the Waza management plan was this platform where local communities are consulted on park management issues.

The difference that management planning can make 59 Building the new village of Halé, outside the national park. Whereas decades of oppression towards Baram, the village in Waza NP, only caused frustration, the new more balanced policy of the ‘carrot and stick’ resulted in the voluntary movement of more than half of its population to Halé.

Water pump that was installed at Halé with the support of the Waza- Logone project.

A new future.

View of Halé village

60 PART III – Enhancing conservation – development integration The importance of developing capacities of protected area managers

Mweka wildlife college. The three African regional wildlife colleges together have trained more than 4000 protected area managers many of who are presently in charge of protected areas. Protected area managers generally have little input in protected area planning and community conservation. Their lack of capacity in these disciplines was not unique to Waza-Logone but was also identified as a major constraint for Integrated Conservation & Development Programmes elsewhere in Africa.

Examination at the Garoua Wildlife College. This francophone college has trained protected area personnel from over 20 West and Central African countries since 1970. Its curriculum was outdated however. A new curriculum was therefore developed and implemented in the late 1990s. Our study analyses the factors that influenced this curriculum reform in the college’s endeavour to respond to new developments in African wildlife management.

The importance of developing capacities of protected area managers 61 Trainees expressed their interest in field trips and exercises as the most effective type of training, addressing skills and complementary to the predominantly knowledge-based classroom training.

Field training, counting wildlife in a protected Botanical field training. area. Planning and financial problems are the main reasons why field trips rarely cover the quarter of the training time they are programmed to occupy.

Computer exercises at the Southern African Wildife College. After some initial hesitation, computer science has become an integral part of the wildlife curricula at the three African wildlife colleges.

62 PART III – Enhancing conservation – development integration Anti-poaching training for Waza National Park Guards. This was their first training for over a decade in these essential skills.

Garoua students practising Participatory Rural Appraisal, a field exercise in the newly developed Com- munity Conservation course. This study reviews the trainees’ evaluations of this course to appreciate its relevance and support its further development. Their reactions suggest that protected area personnel are not ‘attitude limited’ as often suggested. Their constraints to develop a more people-oriented work style lie largely in the areas of knowledge and skills. These findings motivate increased efforts to implement training for protected area personnel in community conservation.

The importance of developing capacities of protected area managers 63 Exhausted after monitoring the transect from Zina to Tchikam and continuing to Waza town (with Moussa Barka, October 1997). This three-day trip by foot during the flooding season was my last with the Waza-Logone project. 2 The Ecological History of Waza-Logone: Constructing a reference image for floodplain rehabilitation

‘...the country, however, became exceedingly interesting and pleasant when we reached one of the numerous water-courses of these African Netherlands...., an open and clear river about seventy yards broad, which being fringed on each bank with a border of slender deléb palms or kamelutu, in the clear, magnificent morning sky afforded a most pitoresque view’ Barth (1857), the first European visitor to Waza-Logone on 13 January 1852

2.1 Historical periods and main ecological characteristics

Rainfall and flooding are the dominant factors, which trigger changes in vegetation composition and production (Chapter 3, 4), waterbird and large antelope population sizes (Chapter 5, 6) and pastoral migration and livestock densities (Chapter 7). They should be known in order to interpret the changes induced in 1994, subject of this thesis, in a wider, spatial and chronological/temporal context. Below, I will show that our reference image of the ‘pre-Maga dam’ period when the ‘floodplain was still intact’ was biased by the period from which information was consulted. Changes in hydrology, land use, vegetation and especially wildlife have often occurred already years earlier, but in several cases, the system’s resilience has led to their recovery. The intensity and speed of changes in the post Maga dam period were, however, unprecedented in recent times, triggered by lower than average rainfall and man-induced drought.

I distinguished eight historical periods, based on criteria such as rainfall, hydrology and the impact of human interventions and, above all, the quantity and quality of available sources of information. Information sources prior to the 1930s encompassed indirect information from excavations, travel descriptions and oral history, whereas scientific descriptions, contacts with scientists and local communities constitute our more recent sources of information.

Rainfall data was collected at nearby N’djamena (Chad) from 1905 till 1914 and from 1931 onwards (Figs. 2.1, 2.3). We therefore also have to rely on less detailed information on climatic periods in the region, derived from Lake Chad water levels during the last millennium (Fig. 2.2), and other geological, palynological and historical sources (Maley 1981; Nicolson 1986).

1 Prior to ±1820

Information sources Sources of information on Waza-Logone prior to 1820 are essentially archaeological studies and travel descriptions, hand-written manuscripts in Italian or Arabic, which were translated and reinterpreted before being accessible (Hopkins & Levt- zion 1981; Rauchenberger 1999). One often has the impression that these secondary sources are illustrative of the interpreter’s own perception and interest, making it difficult to draw conclusions on the state of the environment.

Rainfall and hydrology Wet and drought spells have occurred in Waza-Logone throughout the last millennium, their reported frequency is often a function of the quality of information (Fig. 2.2). The 17th century was particularly wet and floods probably inundated the entire floodplain (Zeltner 1997). From approximately 1700 onwards, information 68 Part I – Introduction change of park warden (Saleh) Committee created Waza project end of FAO Ecole de Faune created NP created Waza Flizot park warden (or earlier?) Change of park warden (Habibou) Start WL-project Designation as Biosphere reserve (1979 or 1982?) New park warden (Badjoda) Independence of Cameroon 2 2 ‘first’ Elephants appearing frequent crop raiding by Elephants ) Management plan approved -1 ) enlarged with 180 km s -1 3 s 3 first reflooding study (Wit & Grijs) Park warden killed 1996 1995 1994 reflooding 1 (20 m 19991998 1997 reflooding 2 (7-10 m 19741973 construction of upstream embankments (1971-1977) Move of Baram (park village) 1972 limiting drainage into the Logone plain 1970 road construction none 1969 1968 1965 19641963 19621960 1956 1955 1954 1953 1952 1951 Elephants destroying crops1950 assistance project FAO park village Zeila burned crop protection unit created 2000 1978 19771975 extinction Cheetah 1959 1958 1957 19931992 1991 1990 198719861985198419831982 1981 1980 1979 start of vegetation monitoring Maga dam none none reduced with 1500 km none low Kob numbers start of dam impact studies extinction Waterbuck Civil war in Chad (1982-1987) emigration out of floodplain HIGH 0 500 0 = long-term average LOW 2000 1990 1980 1970 1960 1950 -500 Rainfall Interventions Flood & wildlife Vegetation Communities/pastoralists NP management Waza 2 – Reference images 69 s, Zina-Waza hunting reserve created Zina-Waza Dieguere village created (Musgum) creation of Andirni and Badadaye villages (Arab Fulbe) rhino observed rise in ivory commerce creation of Mbili-2 by Musgum from Pouss ‘first Elephants since memory in Waza’ no elephants observed buffalo present buffalo Anané large village 1920 19131910 major drought no elephants observed 1900 Rabah defeated / colonisation 1930 19501940 1935 Dragesco 1952 HIGH ??? 1880? 1870 1860 1850 elephant dead, Niwadji Zina brisk market for ivory elephant prolific ? 1890 rinderpest? (park village) – Historical chart, events in Waza-Logone from 1850 onwards, with periods of major upheaval indicated (cf. collapsing panarchie – Historical chart, events in Waza-Logone 0 500 LOW 1950 1947 1944 1941 1938 1936 1932 1929 1926 1923 1920 1917 1914 1911 1908 1985 0= average conquest of Rabah, massacre in Mahé Figure 2.1 Figure Chapter 12.4) -500 70 Part I – Introduction

t1 t2 285 m

t3 283 m

281 m

279 m z1 900 1000 1100 1200 1300 1400 1500 1600 1700 18001900 1950 Year t1, t2, t3: spilling into Bahr el Ghazal z1: complete drying up of lake Chad Note: siltation caused the rise of the average level of the lake bottom (280 → 281) during the last millenium.

Figure 2.2 – Levels of Lake Chad (adapted from Maley 1981, 1993).

is sufficiently detailed to suggest the occurrence of drought-wet cycles with a length of approximately three-quarter of a century (Fig. 2.2).

Archaeological sources Dwelling mounds, presumably man-made, are widely dispersed on the Waza-Lo- gone floodplain. Some of them are at present still inhabited, such as the district centre Zina and the villages bordering Waza NP in the East and North. It is assumed that all dwelling mounds used to be inhabited during wet spells, as they were the last spots of non-inundated land in a sea of inundated land. Inside present-day Waza National Park pottery remains can be found on virtually all dwelling mounds that witness the existence of (temporary) fishing camps. Several settlement mounds have been the subject of archaeological surveys. Daima in nearby Nigeria gives insight into the subsistence economy of its inhabitants during the last two and a half thousand years (Connah 1981). The large majority of bones found were from fish, suggesting a predominance of fish protein in daily consumption, as is presently also the case. Remaining bones found were from birds and domestic cattle, a few from small stock, presumably goats, and a small number of wild animals, including Bohor Reedbuck, Kob, Warthog, Gazelle and Elephant. It is likely that sorghum was already grown from the earliest occupation, but remained for some time below archaeological threshold visibility (Connah 1981).

Travel accounts Amongst the first Arabian visitors to West Africa, Ibn Abi Zayd al-Qayrawani (922-996) and Al-Maliki (around 1057), pronounced themselves against the trade in elephants tusks: ‘It came from trading in ivory and I was loath to be involved in something which had received adverse comments from men of learning’ (Hopkins 2 – Reference images 71

& Levtzon 1981). Yet 800 years later, increasing demand for ivory in Europe would ravage the elephant herds in Waza-Logone. Leo Africanus passed the northern part of the Waza-Logone area in the dry season of 1513-1514 (Rauchenberger 1999). From the Lake Chad area he mentioned slaves as main export product but only once noticed the export of ivory, in contrast with Dangala (Sudan) he visited and described as home to renowned elephant hunters.

Human populations The Sao, ancestors of present Kotoko (Lebeuf 1969), were predominantly sedentary and inhabited the various dwelling mounds in the floodplain such as present day Mbili, Baram and Mahe in the north and eastern part of Waza National Park. Fish was their predominant source of protein, probably supplemented through hunting. After the wet 17th century (Fig. 2.2), the first nomadic Fulbe and Arab Choa entered the area, diversifying the human exploitation by their pastoral way of life (see also Chapter 7). It seems likely that with their arrival local breeds of Bos taurus cattle, possibly related to the Kuri cattle still surviving in the Lake Chad (CIRAD-EMVT 1997), were replaced with the now dominant Bos zebu breeds well adapted to transhumance.

2 ±1820 – ±1900

Information sources The first European travelers visiting the Waza-Logone area in the early 19th century found an area that had just undergone major changes with the arrival of Fulbe and Arab pastoralists bringing, amongst other things, the islamic culture into the area (Chapter 7). Denham & Clapperton (1826) travelled in the Lake Chad area and visited the sultan of Logone Birni in February 1824, who, though with far less authority, still reigns over the floodplain (Lebeuf 1969; van Est 1999). Unfortu- nately, Denham’s account provides hardly any environmental information of the floodplain. Thanks to later travellers, especially Heinrich Barth (1857), detailed first-hand descriptions of the floodplain environment became available. In March 1872, Nachtigal visited like Denham and Barth, Logone Birni and collected detailed, yet mostly secondary information on the floodplain on the eve of the turmoil caused by rinderpest and the conquest by Rabah that would change the area dramatically.

Rainfall and hydrology The 19th century had both droughts and wet spells, ending the 19th century as it started with a relatively wet period (Fig. 2.2).

Vegetation Descriptions of Barth (1857) suggest that the levees along watercourses were much more densely wooded than nowadays. The same plant species seem to have been 72 Part I – Introduction

dominant (Celtis africana, Ficus spp, Tamarindus indica), although some Sudanian savanna tree species, rather scarce in the area nowadays, were mentioned as well (Vitellaria paradoxa). The vegetation of the lower parts of the floodplain was largely comparable to the present one, although explorers paid little attention to its floristic composition.

Birds and large mammals Birds did not attract the attention of most European explorers and their observations are therefore limited to either conspicuous species as Ostrich or to abundant and frequently eaten species such as Guineafowl (Barth 1857).

Nachtigal, who traveled several years in the Lake Chad Basin, remarked on his passing through Sudan in 1874: ‘When one recalled the large number of antelopes which the traveller encounters on all sides in the regions of Bornu, even in the neighborhood of inhabited places, the difference was astonishing’ (Nachtigal 1987). Even today, densities of larger mammals seem to be higher than in most other parts of the Sahel (Scholte & Hashim 2005). Of these large mammals, Giraffe attracted the attention of Barth (1857) who mapped these observations, showing the early importance of present Waza NP). Barth (1857) frequently mentioned Hippopo- tamus from the floodplain, where it is now reduced to Lake Maga and its direct environs.

Elephant is amongst the most conspicuous wildlife species of Waza-Logone, easily traceable by the deep prints they leave in moist soil, remaining well visible till the next rainy season. Elephant prints are still the fear of any driver and one of the horses with which Barth’s company travelled fell and his rider broke his arm. Without doubt the interest of the explorers for Elephants can also be explained by the increasing demand for ivory during the 19th century. UK ivory imports from Tripoli, through which most ivory from Central-West Africa was transported by camel, shot up in the early 1850s to some 70 tons a year remaining fairly constant and only reducing at the end of the 1890s (Johnson 1978). Barth (1857) described several encounters with Elephant herds in Waza-Logone and mapped some observations and prints. Barth and his company frequently ate Elephant flesh, said to be the most commonly available. During the passage of Nachtigal, Elephant hunting had become so important that he described Zina as a ‘brisk market for ivory’ (Nach- tigal 1987). The slaughter of Elephants was also reflected by the naming of the village of Niwadji that, created in the late 19th century south of Waza NP, signifies the ‘place of a dead Elephant’.

Human populations: turmoil Whereas the first part of the 19th century passed under relatively stable conditions, the second part of the 19th century was marked by continuing slaving raids, rinderpest, the devastating passage of the army of Rabah and ended with the colonisation 2 – Reference images 73

of the entire sub region (Zeltner 1997, 1988). Apart from the resulting human tragedy, these developments have had a profound impact on the ecology of the area, still visible today.

From 1890s till deep into the 20th century, northern Cameroon was hit by rinderpest that wiped out entire herds of livestock, not only upsetting the lives of pastoralists, but the entire agro-ecology of the area (Beauvilain 1989). Losses of wildlife due to rinderpest have only been reported in the second part of the 20th century, but it is assumed that it was responsible for the start of the decline of African Buf- falo. ln Sudan, a link was laid between rinderpest and increased Elephants hunting to reconstitute the depleted cattle herds (Cunnison 1960).

The spreading of rinderpest was triggered by the movements of people prior to and with the arrival of Rabah’s army that swept from East to West through the Lake Chad Basin and reached Waza-Logone in 1893. Rabah’s ‘Sudanese’ or ‘Arab’ soldiers targeted especially Kotoko villages of which 33 out of the 43 in the Logone sultanate were completely destroyed. In Mahé, a small village on the northeastern corner of Waza NP, 60 adult men were killed, nearby Garlé lost 20 adults and its remaining population was driven in slavery (Beauvilain 1989). Anané was mentioned as an important village by Nachtigal (1987), whereas there remains at present only a waterhole surrounded with some Balanites aegyptiaca trees in the Acacia seyal zone of Waza National Park. The army of Rabah lived mostly off looting (Zelt- ner 1988), and combined with the increasing availability of firearms may have contributed to the decline of several large mammals, most easily detectable by the Elephant population dynamics (see below).

3 ±1900 – 1950

Information sources Scattered reports by European travelers appeared, some of them with considerable zoological details (Herzog zum Mecklenburg 1912; Jeannin 1936; Zwilling 1940), yet none with the genuine observation quality of the earlier travelers Barth and Nachtigal. Our PRA-studies and other oral history surveys (Chapter 9; van Est 1999) clarified much of the local history.

Rainfall The first part of the 20th century has known both droughts, around 1913 and 1934, and a relatively wet period from ± 1936 till 1945 (Figs. 2.1, 2.3).

Birds and large mammals Bates (1927) was the first ornithologist to publish his observations of his passage through the Waza-Logone area on his trip to Lake Chad. None of his observations has been included in the checklist of the area due to the lack of accuracy on obser- 74 Part I – Introduction

vation locations (Scholte et al. 1999). Gide (1927-1928) and Allegret (1993) who travelled in 1927 downstream the Logone river, described in passing a conspicuous avifauna with abundant Black-crowned Cranes and ducks. No species were mentioned which do not exist today anymore and their description gives an ‘impression’ rather similar to the one that is presently available.

Jeanin (1936), a French North-Cameroon-based veterinarian, wrote the first account on the mammals of Waza reserve, and commented on their decline since 1900. For some larger animals his descriptions are sufficiently detailed to allow a semi- qualitative comparison with later estimates. Giraffe for example, had declined from 1900 onwards when groups of 30-40 Giraffes could still be observed (like at present), to a maximum groups size of 15 in 1935. The entire reserve was estimated to number 100 Giraffes only, compared with 1200-1500 in the 1970s and 1990s (Esser & van Lavieren 1979; Tchamba & Elkan 1994). Early 1900, African Buffalo Syncerus caffer used to be common in the floodplain: an inhabitant of Zina was said to have hunted over 100 specimens during flooding when animals concentrated on the non-inundated mounds. In 1935, African Buffalo was already rare in the area (Jeannin 1936). Jeannin (1936) observed Korrigum (‘Topi’) herds of 100-150, occasionally up to 250, like the present situation (see Chapter 6). In 1935, however, Korrigum’s distribution spanned the entire area between the Mandara Mountains and the Logone river, whereas nowadays Korrigum is largely restricted to Waza NP and the bordering floodplain. Kob was observed in ‘groups of 5-50 individuals’ and ‘near the villages of Mahé and Baram hundreds can be observed in a morning’. Its population was under heavy hunting pressure, however. Once the area was gazetted as hunting reserve, numbers of Kob increased dramatically to the tens of thousands that could be observed the 1950s-1970s (see Chapter 6).

Herzog zum Mecklenburg (1912) did not observe Elephants on his journey through present northern Cameroon in 1910-1911 and remarked that Elephants’ preferential habitat was ‘with sufficiently large trees in steppe land with drinking places’. He added that former Elephant hunters around Lake Chad had changed their way of living because of the lack of Elephants. A decade later, Zwilling (1940) did not observe Elephants in north Cameroon either. Twenty years later Jeannin (1936) described the presence of a few Elephants in Northern Cameroon, but not in the Waza-Zina reserve. Flizot (1948), the first Waza National Park warden, described that in 1947 one single Elephant roomed in the Mandara Mountain area, with Ele- phants increasing in the Benoue valley further south due to supposed immigration from Nigeria (Flizot 1948). The number of Elephants in Waza National Park increased from nil in the 1940s to 50-80 in 1954 (Robin 1954) to 600 in 1968 (Fli- zot 1968) and over 1000 in the 1990s, mostly as the result of immigration from Chad (Tchamba 1996). 2 – Reference images 75

2500 wetdwet average d adry w

2000 Rainfall with: increasing line > average

horizontal = average 1500 decline < average

1000 no flooding

500

0 cumulative deviation of mean annual rainfall (mm) 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 -500

Figure 2.3 – Cumulative rainfall 1930-2001

Human populations: settling The relative security that arrived with the colonial rule triggered from the early 1900s onwards the migration of Musgum groups into the Logone floodplain. Rich fishing grounds and fertile soils for agriculture were the principle attractions. Temporary fishing camps developed into settlements on the dwelling mounds deserted after the turnoil of the previous decades. In several cases the Musgum settled in Kotoko villages, but created slightly separate quarters. The arriving Mus- gum conformed themselves to the authority of the Kotoko and their leaders till deep into the 20th century (van Est 1999).

Fulbe and Arab transhumant pastoralists settled increasingly at the sandier southern edge of the present Waza National Park. Most of these villages, such as present-day Badadaye, Andirni and AmKodje were already used as transitory camps because of the presence of water in their immediate surroundings.

Management The ‘Waza-Zina Hunting Reserve’ was created in 1935 (CCTA 1953). Discussions in villages, with the sultan of Logone Birni and with Badjoda (former park warden) stressed the tensions in the negotiations of colonial authorities with the sultan of Logone Birni on the creation of the reserve. This resulted in a status-quo, lasting till present, of the presence of villages right on the border on the north and east side of the (present-day) national park (Chapter 9). In contrast, the lamido of Pette, reigning over villages on the southern part of Waza NP, agreed on the des- 76 Part I – Introduction

ignation of the hunting reserve and assisted in the retreat of villages from the borders, creating the only de-facto buffer zone (see Chapter 9).

4 The 1950 and 1960s

Information sources After the Second World War, there was an increasing attention of French agricultural and veterinary officers for the Waza-Logone area, studying vegetation (Vaillant 1956; Letouzey 1968), wildlife (Robin 1954; Flizot 1962) and birds (Dragesco 1960, 1961).

Rainfall and hydrology The 1950s and 1960s were characterised by abundant rainfall and floods (Figure 2.3). The hydrological situation of Waza NP described by Flizot (1962) illustrates the recent changes in environmental conditions (the post 1979 situation): Water has completely retreated at the end of February or March (December-January), afterwards the dry herbs are being burned and the subsequent regrowth forms an excellent pasture for the herbivorous animals. Approximately twenty water holes used to contain water till the first rains in May (March). The Yaérés give room to the entire fauna from mid March till June (from December onwards), with notable exception from Elephant and Giraffe, who remain dependent on the forage sources in the Acacia seyal zone (no changes).

Vegetation Letouzey (1968) in his general account on the vegetation of Cameroon, described a dominance of Echinochloa pyramidalis, with Vetiveria nigritana as second and Oryza barthii (= O. longistaminata) and Hyparrhenia rufa as co-dominant. He further indicated an inundation depth of 1 to 2 metres (<1 metre) and gave a list of a number of species, which can still be found in association with before mentioned species, namely Sorghum arundinaceum, Dinebra retroflexa, Pennisetum ramosum and Ischaemum afrum.

Birds and Large Mammals Dragesco (1960, 1961) reported his, mostly qualitative, ornithological observations, some of species (eg. African Skimmer Rynchops flavirostris) which have not been observed in recent times. Behind Waza camp, Dragesco observed Black-crowned Cranes Balearica pavonina pavonina, whose numbers he estimated in December 1958 at over 5000. On the same spot I observed them regularly during different periods in 1995-2000, yet never counted more than 1000.

In 1954, the first estimates on large mammals in Waza NP were published, probably based on ground total counts by its warden Flizot (Robin 1954, see also Chap- ter 6). The 1962 total ground count by Flizot (1962) showed an increased number 2 – Reference images 77 of large mammals compared to the earlier reports by Jeannin (1936) probably because of the above-average rainfall and impact of protection measures.

Human populations When we discussed the former floodplain conditions with people, they often referred, with nostalgia, to this period. In 1960 Cameroon became independent and in subsequent years the state was further consolidated. It was not only peace and prosperity however. Zeila, Garle and other Kotoko villages that had rallied against the ruling party, were burned down and moved to neighbouring villages and towns. Zeila, inside Waza NP, has remained a deserted dwelling mound with its occasional lion pride. Badadaye, a Fulbe village, moved further from the Waza National Park boundary.

Management Tourist numbers to Waza NP rose steadily up to 4000 persons per year, and made Waza amongst the best-known West-Central African protected areas. The number of park guards rose during the 1950s and 1960s steadily till fifteen, reflecting the increasing presence of the state in the Waza-Logone area (Fig. 2.4). Apart from this governmental personnel, Flizot appointed young men as so-called park guides in Waza and Andirni village near the main entrances of Waza NP. Their main task was to guide tourists inside the park for which they received a modest remunera- tion. In return for their exclusive position, park guides assisted park guards in incidental tasks such as anti-poaching campaigns and were held to inform the park authorities on any infraction in their area.

50 bar: park guards (armed) + village guards (unarmed) 40 line: park guides (unarmed)

20 number of park personnel 10

0 1953 1962 1965 1970 1974 1978 1980 1982 1985 1988 1990 1994 1996 1998 2000 2002

Figure 2.4 – Number of Waza National Park guards 78 Part I – Introduction

5 1970 – 1979

Information sources In the 1960s-1970s, wildlife conservation had captured the personal interest of the first president of Cameroon who guided several of his international visitors to Waza National Park where he possessed a lodge. The president’s interest in wildlife conservation was also expressed by the choice of the northern capital of Garoua for the regional francophone Wildlife College that was created in 1970s (Chapter 10). Staff and students of the Garoua Wildlife College would play a crucial role in the initiation of the first substantial ecological studies of Waza National Park. In the 1970s, a Garoua based FAO project supported the development of national parks, in particular Waza and Benoue NP for which several base-line studies would be carried out (Vanpraet 1977).

Rainfall and hydrology The 1970s were characterised by the great Sahelian droughts of 1973-75. These droughts coincided with the construction of the Waza-Kousseri road, that cut off some areas of flooding of the Logone river as well as the ephemeral streams that drain the northern part of the Mandara mountains into the Waza area. At the same time, parts of the Logone upstream of the later Lake Maga were embanked, on both the Cameroonian and Chadian sides, further reducing flooding.

Vegetation Staff of the Garoua Wildlife College conducted several studies on the vegetation of Waza NP and produced Floras on woody species and grasses (Geerling 1982; van der Zon 1989). Throughout this thesis, I refer to the discussions with Garoua ecologists whose experiences have been used to reconstruct a qualitative image of the former vegetation (photographs and personal communications of Geerling, van Lavieren and Wit). Vegetation maps of respectively the entire area and Waza NP were produced based on 1:50.000 aerial photographs from the mid-1960s (Gaston & Dulieu 1976; Wit 1975). In the description of Gaston & Dulieu (1976), Hypar- rhenia rufa was indicated as dominant in the western parts of the floodplain with Vetiveria nigritana as second only. Amongst the communities of the wetter parts of the floodplain, the lack of Echinochloa pyramidalis was striking. Their description differed from Wit (1975), who distinguished five plant communities for the floodplain of Waza NP: 1 Vetiveria nigritana community which was the most common type of grassland (observations confirmed by our discussions with pastoralists); 2 Sesbania pachycarpa community found on the slightly sandier parts of the floodplain (area of Zeila and Baram, NE Waza NP); 3 Hyparrhenia rufa that was restricted to ridges along drains near Tchede and Lougouma, contrary to Gaston & Dulieu (1976). Since their field work was car- 2 – Reference images 79

ried out in subsequent years, this difference is probably due to different locations the authors have visited; 4 Echinochloa pyramidalis vegetation community was found to be locally dominant, especially on the Acacia seyal side of the plain, subject to less inundation, with large numbers of dead Vetiveria nigritana tussocks; 5 Forb communities: Just east of the Acacia seyal zone, on the relatively higher parts of the floodplain, a vegetation community without grasses was found. Vetiveria nigritana used to be common as well, given the fact that numerous dead tussocks were found.

Birds and large mammals Waza-Logone was frequently visited in the late 1960s and early 1970s by mostly Nigerian-based ornithologists, coinciding with the construction of a tourist lodge (Fry 1970; Broadbent 1971; Greling 1972ab; Mundy 1972; Holmes 1972, 1974; Pettet 1976; Kavanagh 1977). Their visits were limited to the more accessible western part and lasted a few days only. Only Greling (1972a) carried out a longer study of bird densities in the Acacia seyal shrubland and of the birds of the wooded savanna of the southern part of Waza NP. Thiollay (1978) was the first ornithologist who reported the ornithological significance of Waza, based on his systematic raptor observations. Louette (1981) compiled most of these early observations from Waza-Logone in the first checklist of birds of Cameroon.

The staff of Garoua Wildlife College (see below) initiated wildife studies in Waza NP including the first published systematic large mammals census of Waza NP (Esser & van Lavieren 1979, see Chapter 6), as well as an ecological study on Giraffe (Ngog Nje 1983, 1984).

Human populations The first Sahelian drought after a long period of good rainfall struck hard not only in Waza-Logone, but especially further North in Northern Nigeria and Niger. In- creasing land pressure caused a steadily influx of pastoralists from these areas into Waza-Logone.

Management The designation of the Waza-Zinah Faunal reserve into Waza National Park in 1968, a few years later followed by the construction of a tourist camp near Waza village, triggered a rapid development of the park’s management infrastructure. Additio- nal park guards were recruited and their number rose steadily till the mid-1980s (Fig 2.4). In the early 1970s, the number of tourists rose to 5000-6000, remaining constant till the early 1990s (Fig. 2.5). The number of park guides increased from 4 to over 20 (Fig. 2.4). 80 Part I – Introduction

8000

6000

4000 number of visitors

2000 no information available

966 968 970 972 974 976 978 980 982 984 986 988 990 992 994 996 998 000 002 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 Year Figure 2.5 – Number of visitors to Waza National Park

6 Post Maga dam period, 1979 – 1993

Information sources Garoua Wildlife College continued to play an important role in the initiation of ecological studies in Waza-Logone, joined in the mid-1980s by Leiden University through its field station with the University of Dschang, CEDC.

Rainfall and hydrology The Maga dam construction in 1979 proceeded a relatively long period of below- average rainfall, that some called the worst drought in the area’s history. From 1987 till 1989 rainfall improved and subsequently deteriorated somewhat (Fig. 2.3).

1 Vegetation In the early 1980s, staff of the Garoua Wildlife College reported that Vetiveria nigritana was dying massively in the western part of the floodplain, especially inside Waza NP. This motivated the start of vegetation studies in 1984-1985 by Came- roonian and Dutch students, under the supervision of Carel Drijver (Leiden Uni- versity) and Ton van der Zon (Ecole de Faune-DGIS). They installed a transect from the centre of Waza National Park to Zina in the heart of the floodplain and included a grid near Tchikam in the centre of the no longer flooded floodplain.

1 For more specific information on vegetation, see Chapter 3. 2 – Reference images 81

Birds and large mammals Contrasting the interest expressed during the 1970s, the only ornithological studies undertaken in the 1980s were aerial waterfowl counts on the newly created Lake Maga and other wetlands of Far North Cameroon (Roux & Jarry 1984, 1986, 1987) as well as a terrestrial count of Ruff Philomachus pugnax on the rice fields of Maga by OAG Münster (1991). In the early 1990s, Njifortii (1997) conducted a Ph.D. study on the potential of harvesting Guineafowl that has not had an implementation so far.

Crop damage by the expanding Elephant populations attracted more and more the attention of authorities and became a political issue in the 1990s, making the provincial head of wildlife say: ‘when the Elephants are out it is the minister of environment who calls me, the moment Elephants are back into the park, I have to call him for any discussion.’ The CEDC initiated a Ph.D. study on the migration of Elephants during the rainy season out of Waza NP towards the south, and during the dry season towards Kalamaloué NP (Tchamba 1996). Forage scarcity inside Waza NP was assumed to be the main reason for this migration, possibly enhanced by the lack of flooding, although it remains speculating on the exact reason (Thouless et al. 1995).

Human populations To counter some of the negative impacts of the Maga dam construction the Came- roonian government mobilised a series of development interventions in the area. Village waterpumps and large pastoral drinking ponds were installed during this period. Nevertheless, the Maga dam construction, followed by the severe 1980s drought triggered a considerable out migration of the floodplain, of sedentary as well as of mobile population groups. The newly created Lake Maga and Lake Lagdo further south attracted fishermen from Waza-Logone. Some pastoralists moved into neighbouring Chad, from where even larger numbers of people left because of the long lasting civil war.

Management Following the rapidly degrading situation in Waza NP due to the construction of the Maga dam, the Government of Cameroon launched the ‘Sauvetage Waza’ campaign, digging a large number of waterholes. Although not addressing the real cause of the problem, the reduced flooding, nor its major consequence, the reduction in forage resources, these water holes have played an indispensable role in the management of Waza’s wildlife during the 1980s-1990s droughts. Outside Waza NP, comparable water holes were dug for livestock.

In the mid-1980s, the Cameroonian government recruited additional park guards, of who most were on duty when we provided their first training course in 1996. Despite the reduced flooding and the virtual collapse of several of its wildlife pop- 82 Part I – Introduction

ulations, Waza NP continued to attract large numbers of tourists (Fig. 2.5). In 1979 or 1982 Waza NP was designated as Biosphere reserve (Batisse 1992), although this has remained a designation on paper only with no consequences for its management (Chapter 9).

In 1985, a small dam was constructed by the livestock department to raise the flooding level in an area downstream of Zilim, north of Zina (Fig. 1.2), which was still relatively well flooded, of approximately 400 km2. Studies were undertaken to assess the state of its vegetation, livestock keeping and fisheries after the intervention. Although no reference existed from prior to the intervention, positive effects were observed, as well as increased conflicts between fishermen and pastoralists. This experience would inspire the formulation of the Waza-Logone project (Drijver & Kouahou 1995).

7 Pilot reflooding and consensual park management planning, 1993 – 2003

The subject of this thesis; the reader is referred to Chapter 1 for an introduction to the research questions and discussions of this study. Below I emphasise the links with the topics discussed in this overview.

Information sources, the rationale of pilot reflooding In 1994, an initial study into the potential of reflooding concluded that only by pilot testing the impact of reflooding could be assessed (Wesseling et al. 1994). This pilot testing should include hydrological monitoring (the extremely flat area made predicting rather difficult), ecological-monitoring and the monitoring of land use recovery, integrated into a cost – benefit analysis of the various reflooding options.

The pilot reflooding was initiated in 1994 by breaching an embankment that had blocked a former watercourse, the ‘Petit Goroma’ (Fig. 1.2). Through this newly opened watercourse an additional flow of approximately 20 m3/s streamed into the floodplain. In 1997 another watercourse, the Areitekele, located just south of the Petit Goroma was opened, triggering a further 8-10 m3/s inflow into the floodplain.

Rainfall and hydrology In 1994, the first year of reflooding, rainfall was favourable, followed by three years of below average rainfall. From 1998 onwards rainfall has been above-average (Fig. 2.3). The 1994 flooding was abundant, more than could have been expected from the rainfall, strongly setting everybody’s mind on the impact of the breaching of the embankment. The reflooding was strongly and unanimously appreciated by people from the Waza-Logone area, even though its strength led to the destruction of crops and houses. 2 – Reference images 83

Vegetation From 1993 onwards, vegetation monitoring was re-activated by the Waza-Logone project in the Waza-Zina and Tchikam transect which were earlier used in 1984- 1985. Starting prior to the pilot reflooding, floristic composition was also monitored in over 100 plots in a grid that was situated in the centre of the pilot release impact zone (Fig. 1.2). Vegetation production was assessed as a function of flooding depth (Chapter 4).

Birds and large mammals Despite the previous, mostly isolated ornithological publications, Waza-Logone was poorly known for its rich birdlife. Based on the 1970s studies and our own fieldwork, we elaborated a checklist with 379 bird species, more than most other areas in West-Central Africa, attributed to the area’s varied habitat and study intensity (Scholte et al. 1999). Systematic waterbird counts were initiated in 1993 (Wetten & Spierenburg 1998) and followed up by the Waza-Logone team in subsequent years (Chapter 5). The resulting, quantitative information motivated the inclusion of Kalamaloué and Waza National Park, the Logone floodplain and Lake Maga each as Important Bird Area (Fotso et al. 2001). The Cameroonian Govern- ment also presented Waza National Park as the country’s first Ramsar site.

Human populations The economic crisis that started in the late 1980s led to the devaluation of the local currency in 1994 with an impact on the functioning of central and local governments. Although Cameroon has remained a rather stable country in a troubling region, lack of security became a major problem in the Waza-Logone area (Chap- ter 7). Interventions of the Waza-Logone project and its offshoots encompassed besides the reflooding, small scale development interventions (Chapter 9).

Management The recruitment stop into the Cameroonian civil service from the early 1990s onwards, said to be dictated by international financing bodies, had serious consequences for the management of Waza and other national parks in Cameroon. Most of their park guards were recruited in the late 1970s and mid 1980s, and either approached the age of retirement in the late 1990s or became eligible for muta- tion to their region of origin. Although Waza NP had a reasonable number of park guards in 1990, in the following years their numbers would drop dramatically and at the end of the 1990s, Waza NP was back to its early 1960s protection level (Fig. 2.4).

Park guides, appointed to guide tourists and researchers, are not armed nor formally trained for anti-poaching and park outreach activities. Yet their commitment and continuous presence in villages on various corners of the park has made them over the years indispensable for the protection of the park, first complementary, 84 Part I – Introduction

but due to shrinking number of park guards, more and more as the only representatives of the park authorities. This experience motivated the initiative of the appointment of so-called village guides who play a more formal role in patrolling the park. It should be stressed however, that true anti-poaching control can only be carried by armed and properly trained park guards.

8 Envisaged large scale reflooding

In Chapter 12, the consequences of the present study for envisaged large scale reflooding are discussed. Below I address generalities of the proposed reflooding options.

Rationale Large scale reflooding aiming to rehabilitate the quasi-totality of the Logone floodplain, was initially programmed to take place in 1998, postponed till 2000 and has at the moment of writing not started. The Waza-Logone project distributed a project document, summarising the pilot reflooding results, including some of those presented in this thesis, to attract the necessary 12 million euro (IUCN 1999). Despite the economic efficiency, as suggested by the various versions of the Cost- Benefit analysis (IUCN 1999; Emerton 2004), no follow-up has been given by any of the contacted donors.

Information sources The impact study of the reflooding was mainly based on systematic monitoring. Unfortunately some of this monitoring infrastructure is no longer in place due to lack of maintenance, allowing in future, without considerable efforts, only qualitative observations.

Rainfall and hydrology Three reflooding options were described, amongst which the maximum one totals 640 km2 that includes not only flooding from the East (direction Logone river) but also through the Mayo Vrick (spillway Lake Maga) with an impact also in the western floodplain (IUCN 1999).

Management Originally the 180 km2 reflooding, subject of this thesis, was aimed as a pilot to develop insights on the impact of large scale reflooding on the hydrology, ecology as well socio-economics of the floodplain. Results of this study will show that the integration of these subjects into management recommendation to mitigate potential conflicts has proven to be an even more important challenge. 2 – Reference images 85

2.3 Reference and target image of the Waza-Logone reflooding

Throughout this book I discuss floodplain rehabilitation against the background of a reference image of the area that we constructed in 1993 at the start of the Waza-Logone project (Wesseling et al. 1994). This, in our initial perception, ideal situation served not only as source of inspiration (Lenders et al. 1998), but also seemed, in retrospect with some naiveté, to be an achievable goal. This reference image was formed by impressions from above-mentioned information sources from the 1960s and 1970s. Simplified, our target image for a completely restored floodplain was characterised by a dominant cover of perennial grassland vegetation with Echinochloa pyramidalis and Vetiveria nigritana on the plains (Wit 1975), abundant wildlife like in the 1960s and early 1970s e.g. with 10 000 Black-crowned Cranes (Dragesco 1960, 1961, Chapter 5), 25 000 Kobs and 20 000 Korrigum (Fizot 1962; Chapter 6), both in and outside the national park, and prosperous pastoralists and fishermen with functioning (conflict resolution) management institutions (Chapter 7; van Est 1999).

The ecosystem as described in the 1960 and 1970s was, however, the result of a long period of above-average rainfall (Fig. 2.1). When considering the time-span of the last millennium, the 1960s reference image has been just part of the returning wet-drought spell cycles (Fig. 2.2). The severe rainfall deficiencies that the area experienced in the 1970s and especially in the 1980s, after the Maga dam construction, was by no means exceptional, neither in recent history (1913 drought) nor during earlier periods (Fig. 2.2).

Spells of wet and drought periods are characteristic for the Waza-Logone ecosystem, to which its plant and animal species are thought to be well adapted. This ecosystem resilience became apparent with the recovery of antelopes and Giraffe between 1935 and 1960 after lifting much of the hunting pressure (Jeannin 1936) and assisted by above-average rainfall. The resilience of the Elephant populations was even more spectacular. Hunted to local extinction in the late 19th century and absent from the Waza-Logone area till the 1940s, Elephant has recovered to such an extent that it is now locally considered almost a problem animal (Tchamba 1996).2 Yet the extinction of animal species during the early 20th century, most notably Western Black Rhino (Diceros bicornis longipes) and Painted Dog (Lycaon pic- tus) followed by African Buffalo, Cheetah (Acinonyx jubatus), Leopard (Panthera par- dus) and Red-flanked Duiker (Cephalophus rufilatus) in the 1970 and 1980s (Chap- ter 6) may be more difficult to be reversed. The spectacular return of Waterbuck, observed in 1998 and 2003, shows the difficulty to make such predictions though.

2 One should realise that the concentration of Elephants in Waza-Logone, is the result of pressure in almost the entire Lake Chad Basin and a rather artificial situation. 86 Part I – Introduction

Changes in wildife and livestock densities have had an impact on vegetation as well. The late 19th century extinction of the Waza-Logone Elephant population, concurring with the drop in numbers in cattle and several Antelope species due to the ravaging rinderpest, may have created a window of opportunity for the regeneration of trees, in analogy to the situation in East Africa (Prins & van der Jeugd 1993). The Sclerocarya birrea woodlands of southern Waza and the Balanites aegyptiaca and Tamarindus indica woodlands on the dwelling mounds in the floodplain can possibly be traced to this period. Although still speculative, this would explain the lack of regeneration of these woodlands that has preoccupied ecologists working in the area since the 1960s (Flizot 1968; Vanpraet 1977; MINEF/IUCN 1997). The disappearance of Hippopotamus (Hippopotamus amphibius) from the Logone floodplain, still present till the end of the 19th century, may have had an important impact on the floodplain environment. In the somewhat comparable Okavango delta, hippos create a network of small channels between their day-time retreat and non-inundated grazing grounds causing diversion of water and sediment to the backswamp areas (McCarthy et al. 1998).

The floodplain rehabilitation hypotheses, postulated in Chapter 1, were based on above presented reference and target images. During the 1993-2000 study period these images were gradually developed as discussed in more or less detail in each of the following chapters. In the syntheses, Chapter 12, I reformulate these floodplain rehabilitation hypotheses based on the overall study results, leading to new floodplain rehabilitation objectives. PART II

The Impact of Reflooding in Waza-Logone

Impact on Vegetation 3 Floodplain Rehabilitation in North Cameroon: Impact on vegetation dynamics

4 Maximum Flood Depth Determines Above-ground Biomass in African Seasonally Shallowly Flooded Grasslands

Impact on Wildlife 5 Waterbird Recovery in Waza-Logone (Cameroon), resulting from increased rainfall, floodplain rehabilitation and colony protection

6 Antelope Populations in Waza National Park (Cameroon) from 1960 till 2001: Impact of changes in rainfall, hydrology and human pressure

Impact on Pastoralists 7 Pastoralist Responses to Floodplain Rehabilitation in North Cameroon

3 Floodplain Rehabilitation in North Cameroon: Impact on vegetation dynamics

Paul Scholte, Philippe Kirda, Saleh Adam and Bobo Kadiri

Updated version, including 1998, 1999 and 2002 data, from Applied Vegetation Science 3: 33-42, 2000.

Abstract Since the construction in 1979 of a dam in the Logone floodplain in the Sahelo- Sudanian zone of Cameroon, annual inundations have decreased, reducing perennial vegetation as important grazing for nomadic herds and wildlife during the dry season. Presently, possibilities exist to release excess water for floodplain 2 rehabilitation. In 1994, a pilot release was executed reflooding 180 km to verify predicted advantages. Vegetation has been studied from 1984 onwards along a transect covering flooded, recently reflooded as well as desiccated parts of the floodplain. Since 1993, the floristic composition has been monitored in a grid in the centre of the impact zone. Cover of perennial grasses, most notably Echino- chloa pyramidalis and Oryza longistaminata increased again from 41 to 75% in the reflooded zone. Vetiveria nigritana, a tussock grass that used to be dominant, disappeared slowly after the dam construction and has not shown a comeback. The cover of annual species, most notably Sorghum arundinaceum, a dominant annual grass only since the mid-1980s, decreased in the reflooded zone from 58% to 23%. If the observed conversion rate of annual into perennial grassland is extrapolated, a recovery towards a ‘100%’ perennial state was most likely reached after the 2003 flooding season, as indicated by qualitative observations in 2002. Apart from favourable climatic conditions, recovery to the pre- dam situation might be dependent on the restoration of soil fertility, limiting an approach focusing on flooding depth only.

3.1 Introduction

Seasonally flooded grasslands are amongst the most productive ecosystems known in sub-Saharan Africa (Denny 1985). They are essential habitats for water birds and can harbour large concentrations of antelopes. Many people have adapted to life on the floodplains and are exploiting them with increasing intensity, competing with wildlife. Extreme cases of competition occur when the water supply is cut off for irrigation or hydroelectric purposes, causing abrupt environmental changes due to desiccation and lack of silt input (Breen et al. 1988; Middleton 1999). After widespread food shortages in the early 1970s, such water control projects were undertaken in almost all major river systems in sub-Saharan Africa (Drijver & Mar- chand 1986). Following a change in general attitudes in which the value of downstream, small-scale human activities as well as of wildlife is recognised, more diver- sified approaches of river basin development are currently being developed (Adams 1996). In several river systems possibilities exist for the controlled release of water to deliberately flood downstream areas, potentially rehabilitating the degraded environment and restoring its functions. Such controlled flood releases have been tried on the Phogolo river in South Africa (Bruwer et al. 1996), in the Senegal valley (Hol- lis 1996) and in the Hadejia-Nguru wetlands in Nigeria (Polet & Thompson 1996). They are reported as being ‘successful’, but remain ecological ‘black boxes’ with available information restricted to water flows (in) and local production increases (out).

In 1979, the Waza-Logone area has followed the fate of many African floodplains, with the construction of a dam, forming Lake Maga, and an embankment along the Logone river, as part of a parastatal rice scheme (Fig. 3.1). In combination with lower than average rainfall for the last two decades (Beauvilain 1995; Fig. 2.3), this dam has led to a reduction in depth and extent of flooding over an area of 1500 km2. Previously productive perennial grasslands were replaced by stands dominated by annual-grass dominated stands, limiting regrowth in the dry season and reducing the carrying capacity for wildlife and livestock (Scholte et al. 1996a,b). A pilot reflooding was started in 1994 by the opening of a watercourse, formerly closed by the embankment along the Logone River, reflooding an area of ca. 180 km2. Based on initial positive impressions, options were formulated for a further 1000 km2 of floodplain rehabilitation.

No ecological information following reflooding has been reported from elsewhere in (West) Africa (John et al. 1993). Expectations of the Waza-Logone rehabilitation were, however, optimistic based on the impressive dynamics of floodplains (Breen et al. 1988). In the reflooding appraisal study (Wesseling et al. 1994), we assumed that by reflooding, disturbed floodplain vegetation in the impact zone would fully rehabilitate into its previous ‘perennial’ state within five years. One of the objec- 92 Part II – Impact of reflooding

Fig. 3.1 – The Waza-Logone area 3 – Vegetation dynamics 93

tives of the pilot reflooding experiment was to test these assumptions, summarised in the following questions: 1 How did the floodplain vegetation change during the desiccation process following construction of the Maga dam? 2 What is the rehabilitation capacity of the disturbed vegetation?

The answers to these questions will be pivotal in assessing the viability of the planned large-scale reflooding in Waza-Logone and for similar releases elsewhere. Moreover it may indicate possible bottlenecks in floodplain rehabilitation and inspire actions to solve them.

3.2 Description of study area

Hydrology The Waza-Logone area covers ca. 8000 km2 in the Far North province of Came- roon (Fig.3.1) and has a mean annual rainfall of 650 mm (Beauvilain 1985). The rainy season is from June to September, which is followed, in parts of the area, by flooding from the Logone River and its branches until December. The floodplain is dominated by heavy cracking clay soils (vertisols), with loamy soils on the levees along watercourses (Brabant & Gavaud 1985). The area is extremely flat, intersected by only a few developed watercourses. Most of the flooding occurs over the land surface, also called creeping flow (Howell et al. 1988).

No systematic information is available on water levels on the floodplain prior to the reflooding. Information on former flooding extent was obtained by Landsat MSS satellite image interpretation and discussions with local people employed for hydrological monitoring. The year 1994 was considered as a standard for rainfall and river discharges for reasons of comparison as only slight differences were detected in the period 1994-1997 (internal reports Waza-Logone project). In 1997, a second watercourse was opened, increasing flooding beyond the 1994 floodline. Although the reflooding aimed to (partly) restore the pre-dam hydrological regime, major differences remained. Flood water could now only arrive from the Logone River in the east, whereas in the past it also arrived from the west, through water courses draining the nearby Mandara Mountains. These waters are now trapped by Lake Maga. This change of water direction has the following consequences: • A shorter flooding season, due to the relatively late arrival of flood waters from the Logone River (mid-August onwards), which also takes a long time before arriving in the downstream, western part of the floodplain. This is particularly apparent in the Tchikam plots and the western part of the transect, where flood water does not arrive until late October, well after the area has dried up following the rainy season. 94 Part II – Impact of reflooding

• Flooding water now has to travel over 40 km before arriving in the western area. It passes through dense vegetation that reduces its speed and probably leads to a loss of sediment load, resulting in less fertile soil conditions downstream (Breen et al. 1988).

Vegetation history Barth (1857) and Nachtigal (1889) gave the impression that in the middle of the 19th century the levees along water courses were far more densely wooded than nowadays, most probably with the same dominant species (Celtis africana, Ficus spp, Tamarindus indica), although some Sudanian tree species, scarce nowadays, were mentioned as well (Vitellaria paradoxa). Floodplain vegetation composition seems to be comparable to the present.

Flizot’s (1962) description indicated a drying of the floodplain 1-2 months later than presently observed. This is due to the relatively abundant rainfall and floods of the 1950 and 1960s (Beauvilain 1995). Letouzey (1968) mentioned the dominance of Echinochloa pyramidalis, with Vetiveria nigritana as second and Oryza barthii (= O. longistaminata) and Hyparrhenia rufa as co-dominants. He indicated an inundation depth of 1 to 2 metres (at present <1 metre) and mentioned Sor- ghum arundinaceum, Dinebra retroflexa, Pennisetum ramosum and Ischaemum afrum, which can still be found in the desiccated floodplain.

The period of 1970-1979 was characterised by the great Sahelian droughts of 1973- 1975, coinciding with limited changes in the hydrological infrastructure due to the construction of the Waza-Kousseri road (Fig. 3.1). Vegetation maps of the entire area and Waza National Park, based on aerial photographs from the mid 1960s (Gaston & Dulieu 1976 and Wit 1975, respectively), showed that presently inundated floodplain communities used to extend 10-15 km further west. According to Gaston & Dulieu (1976), Hyparrhenia rufa was dominant in the western parts of the floodplain with Vetiveria nigritana as second. The lack of Echinochloa pyramidalis in the floodplain is striking. This description differed from Wit (1975) who found Vetiveria nigritana the most common grass species in the western floodplain and Hyparrhenia rufa restricted to ridges along drains. Wit (1975) found Echinochloa pyramidalis locally dominant, especially on the western side of the plain which is subject to less inundation, with the presence of large numbers of dead Vetiveria nigritana tussocks. Sorghum arundinaceum was present in Acacia seyal thickets above the flood line in the western part of Waza NP.

The presently inundated floodplain vegetation is typical of African seasonally flooded grasslands with its species-poor perennial grasslands (Denny 1993). Ory- za longistaminata and Echinochloa pyramidalis constitute single (or two) species- stands on the floodplain, but Vetiveria nigritana is abundant on the levees of the drainage lines and on the higher parts of the intact plain. Hardly any perennial 3 – Vegetation dynamics 95

grasses can be found on the clay soils above the present flood line, where annual grasses (in particular Sorghum arundinaceum) and herbs occur as well as, locally, thickets of Acacia seyal and Piliostigma reticulatum.

3.3 Methods

Vegetation monitoring Maga dam period, 1984-1985: A 35 km long transect, with six permanent plots (P) each 20x50m, was established. It started at the western edge of the floodplain, which is no longer inundated, with exclusively annual vegetation where Vetiveria nigritana used to be dominant (Wit 1975). The transect ended in the East in the well-inundated floodplain, dominated by perennial grasses, little changed since Gaston & Dulieu (1976) and Wit (1975) (Table 3.1; Fig 3.1). In the desiccated floodplain, 23 permanent 10 x 10m plots were installed in a 1 x 0.5km area around Tchikam (Table 3.1; Fig. 3.1). Data was collected along the transect in 1984 and 1985 and from the Tchikam plots in 1985, forming a reference for the present study.

Table 3.1 – Observation methods

Observation Location Flooding1 No of Length Plot size Observation Observation platform (Fig. 3.1) permanent (km) (m) years2 period plots

Transect Zina → Tchikam F (D) 16 (Z) 8 6 x 6 1993-97, 2002 flood (east) R (D) 18 (Z) 9 6 x 6 1993-97, 2002 flood F R (D) 4 (P) 13 50 x 20 1984-853, flood 1993-97, 2002

Transect Tchikam → west D 2 (P) 5 50 x 20 1984-853, mid-dry (west) 1993-97, 2002 season

Tchikam Tchikam D 23 0.96 x 0.54 10 x 10 19854, 1993-96, mid-dry plots 1999 season 22/9/94 Grid F R D floodline 20 x 16 6 x 6 1993-97-99 early dry above under season 28 108 late dry season 1993

1 Flooding types F: always well flooded (also after 1979); R: reflooded from 1994 onwards; D: desiccated since 1979, 2 excluding late flooding ; ( ) marginally present; Throughout this paper we refer to the flooding years which were stud- 3 ied, and not to the observation period; Based on Oijen & Kemdo (unpubl. MSc thesis 1986) and ten Cate (unpubl. MSc 4 thesis 1988); Based on Brummelen & Specken (unpubl.MSc thesis 1986). 96 Part II – Impact of reflooding

Pilot reflooding, 1993-1999(-2002): From 1993 onwards, prior to the pilot reflooding, vegetation monitoring was re-activated in the same Tchikam plots as in 1985 (Table 3.1). The Tchikam plots were only marginally affected by the pilot release because of the late arrival of flooding water. Their position above the present flood line makes them complementary to the ‘above flood line’ part of the grid.

In 1993, the number of permanent plots in the transect was increased from six to 40, the new 6 x 6 m plots (Z) were located at each 0.5 km (Table 3.1). Floristic composition was monitored in a newly installed 20 x 16 km grid, situated in the centre of the pilot release impact zone, upstream from the transect and Tchikam plots (Fig. 3.1). The grid had 136 (6m x 6m) plots. Vegetation in the grid was initially composed of 40% perennial grasses (mainly in the eastern part, bordered by the water course which brought in the release water) and 60% annual species in the centre and west, influenced by the Maga dam. Here the dominance of the 3-4 m tall Sorghum arundinaceum was striking. Initially aimed at hydrological monitoring, these 136 permanent plots were located at a distance of 0.5 or 1 km (Table 3.1).

Data collection and analysis Based on the 1984-1985 methodology (Tchikam plots and transect), the following parameters were determined in the three observation areas (see Table 3.1 for additional details):

Tchikam plots • Number of individuals of Echinochloa pyramidalis, Ischaemum afrum, Vetiveria nigritana, and Pennisetum ramosum. From 1993 onwards, Panicum anabaptistum and Sorghum arundinaceum were also counted. • From 1993 onwards, Oryza longistaminata cover was estimated

Transect (east and west) • Dominance ranking of Echinochloa pyramidalis, Vetiveria nigritana and Oryza longistaminata (the latter only since 1993) • To detect possible trends of Vetiveria nigritana, its tussock condition was classified in 1984, 1985, 1996, 1999 and 2002 (adapted from Riney 1963, see Fig. 3.2).

Grid • Estimation of total ground cover (except 1993) • Estimation of cover for each species and life form (except 1993) • Ranking of species according to dominance (1993, 1994).

Observations in the grid in 1993 were only possible at the end of the dry season, when fires had passed, preventing reliable cover estimations. General dominance assessments were still possible, however, and were compared with the 1994 early 3 – Vegetation dynamics 97

100

% belonging to condition class 20

0 Year 84 85 96 99 02 84 85 Dominance 2 ? 0 96 99 0 0 1 02 84 85 4 Plot no. P1 1 12 5 96 99 02 0 1 2 84 85 96 P3 0 0 0 99 02 84 P4 1 1 3 85 96 99 1 4 3 1 02 84 85 degrading P5 2 3 3 96 99 02 P6 3 2 2 tussock 3 3 3 P7 condition 1 2 3A 3B 4

Figure 3.2 – Transect: Vetiveria nigritana vigour class development (1984-2002) (adapted from Riney 1963) 1 base of tussock firm when kicked gently 2 central part of base dead, often charred stumps remain; death is commonly from the centre outwards 3A living exterior is not intact anymore 3B few minor living parts remain on the exterior 4 only dead tussock base still remaining

dry season records. Cover per species or life form is expressed as percentage of total vegetation cover (relative cover) to limit bias related with the observation period. In the grid, plots were separated according to their position to the 22 Sep- tember 1994 flood line, the maximum height of the floods in the first year of reflooding (Table 3.1). As explained above, from 1997 onwards the actual floodline expanded into the area above the 1994 floodline, preventing quantitative use of these data beyond 1997.

Monitoring results of all plots have been treated as paired samples, eliminating plots where data for one or more years were missing. Given the general non-normal distribution of species and plant form cover, 1993-1997 results were analysed with Friedman two-way ANOVA test (for multi, paired, non-parametric variables) or, in a few cases, Wilcoxon signed ranks test (for two, paired, non-parametric variables) (SPSS 1993). 98 Part II – Impact of reflooding

3.4 Results

Changes in vegetation attributed to the impact of the Maga dam In 1984 Vetiveria nigritana was the dominant species in four of the six studied plots of the transect, and was second in dominance in the two extreme plots in the west and east (Fig.3.2). Between 1984 and 1985, during a severe drought, deterioration of its condition in the transect was observed and it lost its dominant position to Panicum anabaptistum and Echinochloa pyramidalis. In the 1990s, only one individual remained in the western part of the transect, but was found dead in 2002. In the eastern part of the transect, which was still well flooded, Vetiveria nigritana remained strongly present although it had lost its dominant position to Echino- chloa pyramidalis and Oryza longistaminata (Fig. 3.2; Table 3.2). Its condition seemed, however, to be improved in 1996 and 2002 compared to 1984 and 1985. In the Tchi- kam plots, Vetiveria nigritana also disappeared between 1985 and 1993 (Fig. 3.3).

Information on Echinochloa pyramidalis and Oryza longistaminata grasses indicated a reduction of their cover following the Maga dam construction. In the Tchi- kam plots, Echinochloa pyramidalis disappeared between 1985 and 1993 (Fig. 3.3). On the transect it disappeared from two plots between 1985 and 1993 (P3, P4), although further to the east its presence continued in 1993 in the lower laying parts of the desiccated plain (Z24, P5, Z26-29) (Table 3.2). In the still well flooded part of the transect (P6, P7) both species remained dominant (Table 3.2).

200 1985 1993 1994 1995 1996 1999 ) 2

150 (per 100 m

100

50 number of individuals

0 Pennisetum Sorghum Panicum Ischaemum Vetiveria Echinochloa Oryza ramosum* arundinaceum anabaptistum* afrum* nigritana* pyramidalis* longistaminata (relative cover)

Figure 3.3 – Tchikam plots: Variation in number of individuals per species (1985-1996, 1999) *: significant difference 1993-1996 (p < 0.05), see Table 3.6 3 – Vegetation dynamics 99

Table 3.2 – Abundance of Echinochloa pyramidalis and Oryza longistaminata along the transect

Part no. Z2 Z3 Z4 Z5 P7 Z6 Z7 Z8 Z9 Z10 Z11 Z12 P6 Z13 Z14 Z15 Z16 Z17 Z18 Z19 Z20 Flood1 EFFFFFFFFFFFFFFRDEEED

Echinochloa pyramidalis 1984 12 4 1985 1 1 1993 023111111222124000000 1994 02222221102?12?000000 1995 022222211024221000000 1996 011212221012221200000 1997 022222221023222400000 2002 012212122023221000300

Oryza longistaminata 1993 0122?222?111111000000 1994 011111122111210000000 1995 011111122111113000000 1996 022121112121112000000 1997 011111112111111000000 2002 021121211111112200000

Part no. Z21 Z22 Z23 Z24 P5 Z25 Z26 Z27 Z28 Z29 Z30 P4 Z31 Z32 Z33 Z34 Z35 P3 P1 Flood1 RRRRRRRRRRDRRRRRRDD

Echinochloa pyramidalis 1984 2 4 3 0 1985 2 3 2 0 1993 0002203124000050000 1994 0003000??5000000000 1995 0002400232000000000 1996 0002500242000000000 1997 0002500232000000000 2002 1112132222000300000

Oryza longistaminata 1993 0000300001000030000 1994 0000535003000000000 1995 0000000101000054000 1996 0000005121055023000 1997 0000005111050432000 2002 2001401111031111100

1 F: Flooded, R: Reflooded, D: Desiccated, E: Exondated (situation 1994 onwards). 2 Dominance: 0 not present, 1 dominant, 2 second, 3 third, etc. (see Table 3.3 for P-values) 100 Part II – Impact of reflooding

Table 3.3 – P-values of shift in abundance along transect, see Table 3.2 (Wilcoxon Matched-pairs Signed ranks test, Anon. 1993).

Echinochloa pyramidalis Oryza longistaminata

1993 1994 1995 1996 1997 1993 1994 1995 1996 1997

1994 0.92 0.30 1995 0.66 0.79 0.44 1.00 1996 0.55 0.50 0.80 0.39 0.15 0.02* 1997 0.87 0.78 0.16 0.035* 0.87* 0.31 0.14 0.41 2002 <0.0001* <0.0001* <0.0001* <0.001* 0.001* 0.013* <0.001* <0.001* <0.001* <0.001*

*: significant at 0.05 level

Sorghum arundinaceum, was again found in the Acacia seyal thickets above the flood line in the western part of Waza NP (Wit 1975). In 1984 and 1985 it was not mentioned in the plots on the transect although it dominated in several of them from 1993 onwards (including two of the 1984 plots). It was also not mentioned in the Tchikam grid either, but it remains unclear whether or not it did occur there.

Changes in vegetation after reflooding

Tchikam plots Results showed the high inter-annual fluctuations in the number of plant species as well as in the cover of Oryza longistaminata (Fig. 3.3). The appearance of Pani- cum anabaptistum in 1995, also observed in the grid above the flood line, is remarkable. As observed in the grid above the flood line, the number of Sorghum arundinaceum individuals did not change significantly between 1993 and 1997, contrasting a remarkable drop between 1996 and 1999 (Fig. 3.3).

Transect On the transect, Echinochloa pyramidalis lost its dominant position to Oryza longistaminata in the parts with the highest water levels from 1994 onwards (Table 3.2). In the period 1994-1997, Echinochloa pyramidalis colonised only one plot, situated at the edge of the flooded plain. In the same period Oryza longistaminata, appeared in eight plots, but did not persist in two of them, and disappeared from one plot. The annual fluctuation in dominance of both rhizomatous grasses, apparently un- related to inundation dept, is striking. This is also observed for Echinochloa pyramidalis between 1996 and 1997 and Oryza longistaminata between 1995 and 1996 (Table 3.2). Changes in the transect accelerated between 1997 and 2002 when Echi- nochloa pyramidalis and Oryza longistaminata appeared in seven respectively six of the 16 reflooded plots, each of which became the dominant species in three plots. 3 – Vegetation dynamics 101

In addition, Oryza longistaminata became the dominant species in four reflooded plots where it was already present. At the end of 2002, 13 out of 16 reflooded plots were dominated once again by rhizomatous grasses, in the remaining three, these grasses had appeared and were expected to dominate soon.

Grid From 1993 onwards, a total of 67 annual grasses, eight perennial grasses, five perennial herbs and eleven woody species were found. Presented data on woody, annual, perennial and total cover are based on all species. We also present the results in detail of all perennial grasses (with exception of three rare ones), Acacia seyal, which was reported to be regenerating after the Maga dam construction, and the two dominant annual grasses (Table 3.4). This selection is based on the previous pilot research (Table 3.1), Gaston & Dulieu (1976) and Wit (1975) and forms the base of the floodplain grazing exploitation (Scholte et al. 1996b). The contribution of this selection to the total cover in the flooded grid from 1994 to 1999 was 67, 77, 72,82, 72 and 76 % respectively.

Table 3.4 – Selected species contributing 67-82 % of total cover in the flooded grid from 1994 to 1999 (see Text)

Species Plant type Hydrosere position2

Acacia seyal woody perennial 8 Pennisetum ramosum annual grass 7 Sorghum arundinaceum tall annual grass1 6 Panicum anabaptistum (short living) tussock grass 5 Ischaemum afrum (short living) tussock grass 4 Vetiveria nigritana tussock grass 3? Echinochloa pyramidalis rhizomatous grass 2 Oryza longistaminata rhizomatous grass 1

1 In the study area bi-annual forms were observed in very low densities 2 Series, with increasing inundation from 8 to 1, grasses adapted from van der Zon (1992)

In 1994, after the first reflooding season, Echinochloa pyramidalis, Ischaemum afrum, and Vetiveria nigritana were found to be less dominant in the reflooded zone than in 1993 (Table 3.5). In the reflooded part of the grid, relative cover of perennial grasses increased from 41% in 1994 to 62% in 1997 and 75% in 1999, a 7% annual average (Fig. 3.4). The change in perennial grass cover can largely be attributed to the relative cover increase of Oryza longistaminata and Echinochloa pyramidalis (Fig. 3.5). The limited number of observations in the grid above the 1994 flood line did not allow a clear distinction in the perennial species’ increase, which seems to be initially merely Panicum anabaptistum and Ischaemum afrum, Fig.3.5 (but see under ‘Tchikam plots’). In the reflooded part of the grid, annual grasses and herbs 102 Part II – Impact of reflooding

cover % 40

Total cover (real) 0 Annuals* Perennials* 1994 1995 Woody 1996 1997 1998 1999

Figure 3.4 – Development in life form cover (1994-1999) in grid plots *: significant difference 1993-1997 (p < 0.05), see Table 3.6

Table 3.5 – Changes in dominance from 1993 to 1994 in the grid area (Wilcoxon signed ranks test, SPSS 1993)

Dominance in 1994 compared with 1993 P-values

Acacia seyal > 0.80 Pennisetum ramosum > 0.51 Sorghum arundinaceum < 0.18 Panicum anabaptistum < 0.34 Ischaemum afrum < < 0.001* Vetiveria nigritana < 0.004* Echinochloa pyramidalis < <0.001* Oryza longistaminata > 0.16

*significant at 0.05 level

decreased in relative cover from 58% to 23% from 1994 to 1999 (Fig. 3.4). Despite high inter annual fluctuations, Sorghum arundinaceum’s relative cover decreased in the flooded grid from 26 to 16% between 1994 and 1997 but above the 1994 flood line it fluctuated around 35% (Fig. 3.5). From 1998 onwards, coinciding with 3 – Vegetation dynamics 103

50 left: below '94 floodline right: above '94 floodline, from 1997 onwards including several flooded plots 1994 1995 1996 1997 1998 1999

cover % 40

0 Acacia seyal Pennisetum * Sorghum Panicum Ischaemum Vetiveria *Echinochloa *Oryza ramosum arundinaceum anabaptistum afrum nigritana pyramidalis longistaminata

Figure 3.5 – Changes in percentage cover of individual species (1994-1999) in grid plots *: significant difference 1993-1997 (p < 0.05), see Table 3.6

Table 3.6 – P-values Figs. 3.3-3.5 at Tchikam and Grid plots (1993-1997) (Friedman two-way anova test, SPSS 1993)

Tested variables Fig. 3.3 Fig. 3.4 Fig. 3.5 under ’94 floodline above ’94 floodline

Acacia seyal 0.94 0.82 Pennisetum ramosum <0.0001* 0.99 0.94 Sorghum arundinaceum <0.0001* <0.0001* 0.56 Panicum anabaptistum <0.0001* 0.90 0.72 Ischaemum afrum <0.0001* 0.89 0.88 Vetiveria nigritana <0.0001* 0.60 1.0 Echinochloa pyramidalis <0.0001* 0.040* 1.0 Oryza longistaminata 0.038* 0.0053* 0.81 Total cover 0.24 Annual cover <0.0001* Perennial cover <0.0001* Woody cover 0.8

*: significant at 0.05 level 104 Part II – Impact of reflooding

the 1997 extension of the floodline, the cover of Sorghum arundinaceum above the 1994 floodline dropped dramatically as well. The increase in cover of Oryza longistaminata above the 1994 floodline was due to its dominance in four plots, situated below the 1997 floodline.

3.5 Discussion

Pre-Maga dam vegetation Available information from the pre-dam situation is highly fragmented and sometimes contradictory (Gaston & Dulieu 1976). However, none of the old landscape descriptions indicated the presence of plant species or plant communities that are no longer encountered in the floodplain. A reduction in some of the floodplain species (e.g. Vetiveria nigritana), has already started well before the Maga dam construction.

Impact of the dam The limited number of observations did not allow a comprehensive assessment of the disturbance process following the 1979 Maga dam construction. Data on Veti- veria nigritana showed, however, that changes took place over a period of six years and were strongly influenced by climatic and hydrological fluctuations, most notably the 1984-1985 drought. Floodplain rehabilitation is, obviously, subject to the same variable environmental fluctuations and fits in the ‘succession and fluctuation’ model of Van der Valk (1985). In 1988, a year with rainfall well above the decade average (Beauvilain 1995), local people considered the flooding conditions as comparable to ‘pre-dam averages’. This suggests that even before the reflooding flooding levels were still comparable to pre-dam levels. The Maga dam apparently caused ‘only’ a change in relative cover of existing plant communities, resulting in a different vegetation pattern. The slow changes following the disturbance in 1979 might be indicative of the future rehabilitation process.

The rehabilitation The hydrosere observed by van der Zon (1992), and from other African seasonally flooded grasslands (Denny 1993; Howell et al. 1988) matched with our field observations (Table 3.4). The validity of such a hydrosere as a way of possible succession in West Africa was, however, strongly contested by John et al. (1993) and is also contradicted by the presented rehabilitation succession. The lack of Vetiveria nigritana in the rehabilitation succession is especially illustrative of the different species’ regeneration capacity, irrespective of their position in the hydrosere. The decrease in dominance of Echinochloa pyramidalis on the transect in the first reflooding year and its replacement by Oryza longistaminata showed, however, that the hydrosere sequence may be of some help in predicting changes if dealing with species with comparable regeneration strategies. Oryza longistaminata and Echi- 3 – Vegetation dynamics 105 nochloa pyramidalis, both rhizomatous grasses, re-established relatively quickly by lateral vegetative propagation (Van der Valk 1992). This ‘peripheric propagation’ was probably influenced by the proximity of depressions with remaining perennial vegetation, the ‘border’ effect (Barrat-Segretain & Amoros 1996). The fluctuating dominance of both rhizomatous grasses was surprising, given their perennial nature and the period of relatively comparable rainfall and flooding conditions. This fluctuation can only be explained by contraction and spreading speeds of at least the plot length (6 m/year), which is less than the estimated lateral growth -1 speed of 50 m year of Phragmites australis (van der Valk 1992). Some of the more sudden appearances of Oryza longistaminata and Echinochloa pyramidalis can, however, only be explained by propagule dispersal (Noble & Slatyer 1980; Cellot et al. 1998), almost certainly during the flooding.

The tussock grasses Ischaemum afrum and Panicum anabaptistum, both depend on generative propagation and appeared as quickly as the dominant rhizomatous grasses. Vetiveria nigritana, also dependent on generative propagation, did not regenerate at all. Given the dominance of Vetiveria nigritana in the study area prior to the Maga dam construction and its importance as fodder for livestock and wildlife, its lack of regeneration is a major set-back of the otherwise positive pilot reflooding. The 1996, 1999 and 2002 monitoring showed the general good condition of Vetiveria nigritana tussocks, producing seeds in abundance in an area where vegetation cover is very low, potentially permitting gap phase replacement (Van der Valk 1992). Further west in the transect, where Vetiveria nigritana used to be dominant and is still present in depressions, the lack of regeneration might be attributed to the earlier described late arrival of flood water and/or the low sediment level of flood water. It remains, however, difficult to establish the characteristics of these safe sites (Van der Valk 1992). In the eastern part of the transect where, between 1985 and 1993 and despite the continuing flooding, Vetiveria nigritana lost its dominant position to Oryza longistaminata and Echinochloa pyramidalis, it is assumed that the strong competitive nature of these species prevented regeneration of Vetiveria nigritana.

We attribute the 1998 and 1999 changes in species composition in the area situated above the 1994 floodline (Figs. 3.3, 3.5), to the impact of the additional flooding from 1997 onwards. Above average rainfall in 1997-2000 (Fig. 2.3) may further have reduced the differences between flooded and non-flooded areas.

Speed of recovery The decrease in dominance of Echinochloa pyramidalis, Ischaemum afrum and Veti- veria nigritana in the first flooding season contrasted with their later overall increase. We attribute this to the bias of the late 1993 observation period when after bush fires, these species were relatively well visible. We speculated on the speed of a 100% perennial grassland conversion and assumed that the increase would 106 Part II – Impact of reflooding

remain on the same level as in the period 1994-1999. Perennial grass cover in the grid was initially 40% and with an annual 7% conversion, it would take nine years, i.e. till after the flooding season of 2003, before the area would again be entirely dominated by perennial grassland. This prediction is in line with presented qualitative observations on the transect (Table 3.2). These reported results are satis- factory compared to temperate wetland rehabilitation projects such as mire rehabilitation in Central Europe where return to ‘original stage’ is said to be ‘impossible within reasonable time spans’ (Pfadenhauer & Klotzli 1996). Three years after reflooding wetlands in the prairie pothole region in the USA, its vegetation composition still differed significantly from natural wetlands (Galatowitsch & Van der Valk 1996). Thoraug (1980) mentioned recovery rates of two to over 30 years in restored conditions and at least double that time under natural conditions in restoration efforts in various ecosystems in the USA. The relative speed of wetland rehabilitation in the Waza-Logone area may be explained by the dominance of grasses (Thoraug 1980) and the productive tropical floodplain environment. A further factor may be the short, 15 year time span between disturbance and rehabilitation (or 25 years when the climate-induced disturbances are included). This resulted in relatively few changes in the hydrological infrastructure and the continued presence of all major floodplain species.

3.6 Conclusions

The changes in the period 1993 till 1999 showed partial recovery towards a perennial dominated vegetation due to the 1994 pilot release. The ‘100%’ perennial state, probably reached in 2003, differs, however, from the pre-Maga dam vegetation, which was the result of the relatively humid 1940-1970 period. Most remarkable is the lack of Vetiveria nigritana recovery in the reflooded downstream areas. Its comeback may depend on brief periods of entirely different environmental conditions, possibly not occurring in the space of a decade. Another unknown factor in the rehabilitation is the possible lack of sediments brought in with floodwater in the downstream areas. 4 Maximum Flood Depth Determines Above-ground Biomass in African Seasonally Shallowly Flooded Grasslands

Paul Scholte

Abstract African seasonally flooded grasslands are an important source of forage for wildlife and livestock. Flood depth has been frequently used to explain the distribution of plant species, but its relation with vegetation production has remained ambiguous. I studied the relation between flooding and above-ground biomass at the end the flooding season, as well as the dry season to assess the impact of reflooding on the shallowly inundated Logone floodplain, Cameroon. Above-ground biomass of a combination of all species and of the individual perennial grasses Oryza longistaminata and Echinochloa pyramidalis showed a positive linear relationship with maximum flood depth up to at least one meter. These relations became stronger during the two years following the water raising, showing the lag in reaction time to floodplain rehabilitation. Above-ground biomass data from other major floodplains in the three main African geographic regions showed a similar relation with maximum flood depth up to one meter. Dry season regrowth, important because of its high nutrient quality in a period of forage scarcity, was not directly related with maximum flood depth, possibly because of its dependency on the period of burning and residual soil moisture. Although the mechanism of the maximum flood depth - above-ground biomass relation is not yet fully understood, presented data allow an assessment of the impact of (re-) flooding on vegetation production.

109

4.1 Introduction

African seasonally flooded grasslands have drawn widespread attention as source of forage for wildlife (Rees 1978a; Howell et al. 1988) and livestock (Evans-Pritchard 1940; Hiernaux & Diarra 1983; Breman & De Wit 1983). Flooding intensity characterises floodplain vegetation composition (Ellery et al. 1991) as has become apparent by the construction of upstream dams (Marchand 1987; McCully 2001; Scholte et al. 2000b). The raising of waterlevels by flood releases, as practised in the Logone floodplain in North Cameroon, showed the reversibility of these changes on floodplain vegetation composition (Scholte et al. 2000b). The impact of the flooding regime on the productivity of seasonally flooded grasslands remained unknown, as only average productivity characteristics, not linked to flood depth, are reported. In relatively well-studied temporate zones, where flooding usually occurs outside the growing season, it is assumed that the lower the inundation the higher the production becomes (Middleton 1999). Although this may also hold for forested tropical floodplains (Hughes 1990), scattered data of seasonally flooded grasslands in Africa suggest a higher production with increasing flood depth (Francois et al. 1989; Hiernaux & Diarra 1983; Ellenbroek 1987). Only under rapid and extremely deep flooding, more than several meters, tropical seasonally flooded grasslands have shown a relatively low production (Ellenbroek 1987; Junk et al. 1989; Morton & Obot 1984). Flood pulsing, i.e. periods of flood followed by a pronounced dry season (Junk et al. 1989; Middleton 1999), allows fire and grazing, reducing above-ground biomass and triggering regrowth of perennial grasses. This regrowth is marked for its nutritional quality and is the base of wildlife and livestock production (Howell et al. 1988; Hiernaux & Diarra 1983). Regrowth production has been linked with above-ground biomass (Breman & de Ridder 1991), but not with flood depth so far.

To assess the impact of increased flooding on vegetation production, I studied the relationship between above-ground biomass and regrowth of grasses with maximum flood depth. Fieldwork was conducted in the Logone floodplain that is part of the Lake Chad seasonally flooded grasslands (Fig. 4.1). I compared these findings with the main seasonally flooded grasslands in West, North-East and South- ern Africa (Denny 1993). Above-ground biomass of depression and stream communities (Hiernaux & Diarra 1984a; Ellenbroek 1987; Francois et al. 1989), with a maximum flood depth of over one meter will not be discussed because of their limited occurrence in Logone. Because of the floodplain’s importance for grazing, Logone assessments included key forage quality parameters allowing a link with monthly nutritional quality assessments of comparable grassland communities in the Sudd (Howell et al. 1988). Changes in local variations in soil properties such as cracking depth were assessed as they were expected to lead to differences in soil moisture and influence (regrowth) production. Tobias & Vanpraet (1981) and Ellen- broek (1987) who concluded that pedological differences in the Logone and Kafue floodplain were insignificant compared to the impact of flooding on vegetation, limited their observations to general soil characteristics. 110 Part II – Impact of reflooding

Figure 4.1 – Waza-Logone area and position of the Transect. 4 – Vegetation production 111

4.2 Materials and methods

Study area The Logone floodplain has a Sahelo-Sudanian climate, with an average annual rainfall of 650 mm that falls from June to September (Beauvilain 1995). In the study period 1994-1996, rainfall in Zina (Fig. 4.1) was 691, 656 and 761 mm respectively.

The Logone floodplain is situated in the alluvial plains of the former Mega Lake Chad (Schneider 1994; Brabant & Gavaud 1985). The lacustrine and more recent fluvial deposits of the Logone river attain a thickness of hundreds of meters and largely consist of fine textured smectitic clays. Soils have been classified as undif- ferentiated vertisols (Brabant & Gavaud 1985). The eastern and central parts of the area are subject to flooding, from September till December, by the Logone River that has its origin in the volcanic Adamaoua highlands, 500 km further south. The nearby predominantly gneissic Precambrian Mandara Mountains feed seasonal rivers, draining, from July till September, into the south-western part of the study area. The area is extremely flat, intersected by a few little developed watercourses. Most of the flooding therefore occurs over the land surface, still saturated by the rains, also called creeping flow (Howell et al. 1988).

The Logone floodplain vegetation with its species-poor perennial grasslands is typical of African seasonally flooded grasslands (Denny 1993; Ellenbroek 1987; Howell et al. 1987). Oryza longistaminata and Echinochloa pyramidalis constitute one- or two species-stands on the floodplain, whereas Echinochloa stagnina and Vossia cuspidata dominate depressions and stream borders. Hardly any perennial grass is found on the clay soils above the present flood line, where annual grasses (especially Sorghum arundinaceum) and herbs occur, as well as scattered thickets of Acacia seyal and Piliostigma reticulatum. Vetiveria nigritana used to be abundant on the higher parts of the floodplain, but was replaced, due to desiccation (see below), by Ischaemum afrum and Panicum anabaptistum.

The Logone floodplain severely degraded when it was cut off from its main water supply through the construction of a dam upstream (Fig. 4.1). Built in 1979, this dam has caused the disappearance downstream of the perennial floodplain vegetation in an area of 1500 km2. This had a negative impact on livestock and fisheries, as well as on large mammals and birds (Drijver et al. 1995; Scholte et al. 1996a). To counter the ongoing degradation, an embankment was breached in 1994, opening a watercourse that had been closed off since 1979 (Fig. 4.1). This brought back the annual floods in an area of 180 km2 and triggered recovery of perennial floodplain vegetation that was expected to have fully covered the reflooded area in 2003 (Scholte et al. 2000b). 112 Part II – Impact of reflooding

Description of sample sites In 1985, a 16 km transect was established starting at Tchikam in Waza National Park (Fig. 4.1). Prior to the Maga dam its vegetation was dominated by Vetiveria nigritana (Wit 1975) subsequently replaced by Sorghum arundinaceum, Ischaemum afrum and Panicum anabaptistum. This area has been reflooded since 1994 (Fig. 4.2), and is gradually being occupied by Oryza longistaminata and Echinochloa pyrami- alis (Scholte et al. 2000b). Zwang village, situated on a rise on the border of Waza National Park (Fig. 4.1), separates the reflooded western from the eastern part of the transect that has always been flooded. The transect ended near Zina, in the well- inundated floodplain, dominated by Oryza longistaminata and Echinochloa pyramidalis (Annex 4.1; Fig. 4.1). In 1994, the transect’s 4 permanent (P-) observation sites were increased with 28 (Z-) sites, equidistant at 0.5 km, followed by four (S-) plots in 1995 (Annex 4.1), (Scholte et al. 2000b).

70 1994 1995 60 1996 flood duration 1994 (weeks) 50

30 Max. flood depth (cm) 20

0 sites: Z34 Z32 P4 Z29 Z26 S4 P5 Z22 Z20 S3 P6 Z11 S2 Z8 P7 S1 flooding: r r r f r r r r d f f f f f f f West East Tchikam 1 km Zwang Zina

Figure 4.2 – Maximum Flood Depth in the transect Zina-Zwang-Tchikam (see also Fig. 4.1) Sites, see Annex 4.1 for more information; Flood situation: f: flooded (before and after 1994), r: reflooded from 1994 onwards, d: desiccated (before and after 1994)

In 1994, 1995 and 1996, water depth was measured at each south-central site border, twice a month during the flooding period from mid-August till mid-Novem- ber. In 1994, every second week flood duration (Le Houérou 1989; Van der Valk 1981) was assessed, but was discontinued in subsequent years because of the inaccessibility of the western part of the transect. 4 – Vegetation production 113 5 3 Oryza longistaminata, Echinochloa pyramidalis, Vetiveria nigritana Ischaemum afrum, anabaptistum Sorghum arundinaceum, Cichorus olitorius Panicum Soil moisture Dominant plant 3 70 no 3.01 well common, developed – 122 60 60-120 1.67 4 (cm) soil layer (cm) (cm) In layer (cm) season ratio Root OM % pH Depth to sandy Surface Cracking CaCO Layer Top-sub (KCl) clay (loam) cracks depth nodules mid/end dry species 2 Gilgai 2000b et al. 1 – Differentiating soil characteristics – Differentiating

S2 limited 0-10S3 – pronounced nopronounced 6.4–0.3 –S4 mod. no 5.0–6.8 115 142 – rare – 72 65 150 no no 1.57 5.21 S1 limited 0-10 9.2-0.5 3.9-5.5 115 60 10-60 1.36 P6 pronounced 0-15 6.1–0.6 4.5–6.2 125 100 no 2.54 P5 mod. pronounced no – – >180 65 no 3.34 P4 no no 6.4 – 0.5 4.7 – 6.3 110 40 no 2.01 P7 limited 0-10 –

Site

aAlways flooded aAlways reflooded see Scholte see Figure 4.2 limited: 1:50 (wavelength: amplitudo); mod. pronounced: 1:25; 1:15 not determined a non-forage forb of which no nutritional quality was assessed. Table 4.1 Table 2 3 4 5 1 114 Part II – Impact of reflooding

In 1995, at the end of the dry season, eight soil pits were dug along the transect (Annex 4.1), described following FAO guidelines (1977) with additional observations on gilgai (wavelength and amplitudo, Blokhuis 1993), Table 4.1. Soils were

classified as Eutric Vertisols (FAO 1988) as even CaCO3 concentrations in S1 and P7 did not meet the 15% level for a diagnostic calcic horizon (Driessen & Dudal 1989). Of four selected pits, granular composition, pH-KCl, organic matter and

CaCO3 were analysed for each distinguished horizon (Table 4.1) at BLGG labora- tories (The Netherlands). The upper soil layers had high organic matter contents and low associated pH (KCl) values (Table 4.1), respectively higher and lower than usually observed in vertisols (Blokhuis 1993; Bunting & Lea 1962), but comparable to seasonally flooded grassland vertisols in the Kafue Flats (Rees 1978b; Ellen- broek 1987). With their common and well developed cracks, sites P6 and S3 situated in the eastern, continuously flooded part of the transect, marked a transition to soils characteristic of more limited flood depths (Table 4.1; Fig. 4.2). Soil moisture (‘residual’ moisture) was determined during the excavation at 20-40 cm, 60- 80 and around 120 cm. In 1996, soil moisture content was determined in February, mid-dry season, in 16 unprotected plots at depths of 15, 20, 25, 30 and 35 cm (An- nex 4.1). The 110-ml samples were oven-dried for 24 hours at 110°C and subsequently assessed soil moisture was expressed as percentage of dry weight.

Biomass

Above-ground biomass In 1994, 32 sites were sampled at equal distances in the transect, reduced to eight and 16 equidistant sites in 1995 and 1996 respectively (Fig. 4.2, Annex 4.1). Im- mediately after the floods receded (see below), in each selected site 1 m2 random- ly taken samples of above-ground biomass were hand clipped at ground level.

The difficulty to reach the western part of the transect in 1994 allowed only four samples to be taken and another four additional replicates in each site to be visu- ally estimated after cutting, which were corrected following double-sampling weight- estimate (’t Mannetje 1987). In 1995 and 1996, ten samples in each selected site were cut. In the laboratory, clipped biomass was separated according to species. Samples were oven-dried at 65°C for 48 hours and expressed as dry weight percentage. The average of the above-ground biomass per site, expressed as DM (dry matter) gm-2, was used for comparison with maximum flood depth, below-ground biomass, regrowth and soil moisture. Comparison of biomass between years was based on the individual samples in each site.

In view of the low percentage of dead above-ground biomass (<5%), we assumed that biomass sampled when sites dried up was also the maximum above-ground biomass of the growing season. Data from the Kafue Flats (Annex 4.2) suggest that this maximum might in some cases be reached somewhat earlier. The Logo- 4 – Vegetation production 115 ne floodplain as well as the Inner Niger Delta, Sudd and Kafue Flats, has only one growing season however, and with a relatively low regrowth production, standing crop at the end of the main flooding season approaches total annual production.

Below-ground biomass In 1996, three months after the recession of the floods, eight below-ground biomass samples of 0.5 litres were taken horizontally at depths of 0-15 cm and 15-50 cm respectively in the same 16 pits as used for the assessment of soil moisture (Annex 4.1). Roots and rhizomes with a diameter larger than approximately 0.2 mm were separated by hand from the extremely firm soil, dry cleaned and oven- dried for 48 hours at 65°C. No distinction was made between living and dead below-ground biomass, as the latter was estimated to be less than 10%. The average per site was expressed as grams of root l-1 soil, and converted into gm-2.

Regrowth after burning In 1995 and 1996, eight and 16 sites respectively adjacent to the soil moisture and below-ground biomass assessment sites (Annex 4.1) were burnt in late and mid February respectively, as commonly applied in the area. In the 1995 and 1996 plots, 12 respectively eight, 1 m3 wire-netted exclosures had been placed to prevent grazing on each of the sites. In April regrowth turned brown and was hand cut at ground level, irrespective of species composition. For comparison we took ten random samples outside the exclosures at each site. Samples were treated as described under above-ground biomass.

Forage Quality In 1994 and 1995, 30 oven-dried samples were selected, targeting sites dominated by the six dominant grass species (Table 4.1). The entire plants were sampled (cf. Mefit-Babtie 1983), allowing a comparison with the monthly measurements in the Sudd (Howell et al. 1988) and an assessment of total N-yield (N% x Biomass). N-content was analysed at the BLGG laboratory (The Netherlands) and converted to digestible crude protein content, following DCP (‰) = 58.06 (%N) – 35.2 (Bou- det 1983).

Data analysis Normality has been assessed with Kolmogorov-Smirnov test (SPSS 1999), reject- ing non-normality for water depth (p = 0.999, 0.59, 0.77 for 1994, 1995 and 1996 respectively), as well as for total standing crop (p = 0.10, 0.30, 0.20 respectively). Above-ground Oryza longistaminata biomass, excluding 0-sites, non-normality was rejected (p = 0.84, 0.34, 0.85), no longer when including all sites (p < 0.0001). The situation with above-ground Echinochloa pyramidalis biomass was mixed: excluding 0-sites (p = 0.27, 0.001, 0.006, all sites (p < 0.0001). For regrowth in 1995 and 1996, non-normality was not rejected (p < 0.0001), nor for under-ground bio- 116 Part II – Impact of reflooding

mass (p < 0.0001). For soil moisture in 1995 at 20-40, 60-80 and 120 cm, p was respectively 0.32, 0.003, 0.82, in 1996 non-normality was not rejected (p = 0.026).

Above-ground biomass observations from the same plots were compared amongst years, depending on normality with t-tests or Mann-Whitney (SPSS 1999). The calculations of correlation coefficients, Pearson’s or Spearmann’s and linear regressions followed SPSS (1999).

4.3 Results

Maximum flood depth and flood duration In 1994, flooding lasted 5 to 15 weeks, prolonging the rainy season that started with the first rains in June that ponded on the surface from July onwards. The rains ended in early October, when floodwater had arrived in all sites. Some sites dried up in the first week of November, most of the others in the last week of November; the remaining three sites west in the transect during the second week of December.

Maximum flood depth followed the same pattern in the three observation years and was in 1996 5.2 and 4.4 cm lower than in 1994 and 1995 respectively (1994*1995, p = 0.68; 1994*1996, p = 0.007; 1995*1996, p < 0.001) (Fig. 4.2). Flood duration was correlated with maximum flood depth (R = 0.55, p < 0.001), but with limited amplitudes, partly due to the two-week interval of flood duration measurements (Fig. 4.2).

Above-ground biomass Total above-ground biomass (Fig. 4.3), as well as that of Oryza longistaminata (Fig. 4.4) and Echinochloa pyramidalis (Fig. 4.5), increased with increasing maximum flood depth. In 1996, in the third year of the reflooding, this relation had become more pronounced than in the first year 1994. In 1995 the number of observations was too limited for such calculation. Total above-ground biomass in both 1996 and 1995 was higher than in 1994 (p < 0.0001) and in 1995 equal to 1996 (p = 0.82).

Looking at the individual species, the following can be observed. Sorghum arundinaceum a stout annual reed-like grass that had invaded the desiccated floodplain in the mid-1980s, produced in 1994 substantial above-ground biomass till a maximum flood depth of 38 cm, in eight out of 32 reflooded sites. In 1996, its occurrence was reduced to two out of 16 reflooded sites with less than 15 cm maximum flood depth. The limited number of observations due to sorghum’s absence in the continuously flooded eastern part of the transect did not allow a functional relation to be drawn. 4 – Vegetation production 117

1600 ) -2 1994 1995 1996

1200 y = 12.43 x + 263.7 R2 = 0.68 p <0.0001 1996

800

1994 400 y = 3.88x + 174.5 R2 = 0.25 p =0.004

total above-ground biomass (g DM m 0 0 20406080 max. flood depth (cm)

Figure 4.3 – Relation between maximum flood depth and total standing crop (0-65 cm max. flooddepth)

1000

1994 1995 1996 y = 11.45x - 48.4 800 R2 = 0.54 p =0.001

) 1996 -2 600 1995 y = 9.85x - 100.6 R2 = 0.7014 p =0.009 400 Oryza longistaminata 1994 y = 5.45x - 64.9 biomass (g DM m 200 R2 = 0.42 p <0.0001

above-ground 0 0 1020304050607080 max.flood depth (cm)

Figure 4.4 – Relation between maximum flood depth and above-ground biomass of Oryza longistaminata (0-65 cm max. flooddepth) 118 Part II – Impact of reflooding

600 1994 1995 1996

) 500 -2

y = 6.86x - 52.9 400 1996 R2 = 0.76 p <0.0001

300 Echinochloa pyramidalis 200 biomass (g DM m 1994 y = 3.25x - 47.4 R2 = 0.40 p <0.0001 100

above-ground 0 0 20406080 max.flood depth (cm)

Figure 4.5 – Relation between maximum flood depth and above-ground biomass of Echinochloa pyramidalis (0-65 max. flooddepth)

For wild rice Oryza longistaminata, the dominant perennial grass, above-ground biomass was a function of maximum flood depth in 1994, a relation that improved in subsequent years (Fig. 4.4). In 1994, above-ground biomass of Oryza longistaminata was lower than in 1995 and 1996 (p < 0.0001) and in 1996 slightly higher than in 1995 (p = 0.042).

Echinochloa pyramidalis, co-dominant in inundated sites, showed an increasingly pronounced relation with maximum flood depth (Fig. 4.5). In 1995, the limited number of observations did not allow such calculation. In 1994, above-ground biomass was lower than in 1995 and 1996 (respectively p = 0.002 and p < 0.0001) and in 1995 equal to 1996 (p = 0.50).

Below-ground biomass In the topsoil (0-15 cm), below-ground biomass varied from 8 g DM m-2 (Z32, with annual vegetation) to 2750 g DM m-2, double the above-ground biomass (Z29, with a dense rhizome layer of O.longistaminata). In the 15-50 cm layer, below-ground biomass was reduced to 3-165 g DM m-2. Below-ground biomass at both 0-15 and 15-50 cm was correlated with above-ground biomass and with the percentage of perennials (Table 4.2). Maximum flood depth was correlated with especially below- ground biomass at 15-50 cm (Table 4.2). The shoot/root ratio was negatively correlated with the percentage of perennials (R = -0.60, p = 0.014). 4 – Vegetation production 119

Regrowth after burning Regrowth was higher inside than outside the exclosures (p < 0.05) in one site (S1) in 1995 and in 1996 in 5 sites (S1, Z8, Z11, P6, Z22). With one exception (Z22), these sites were situated in the well-flooded eastern part of the transect, outside Waza National Park (Fig. 4.2). The following presentation will be limited to the exclosure data.

In 1995 regrowth was correlated with residual soil moisture still present at the end of the dry season at a depth of 20-40 cm but not with residual soil moisture at 60- 80 cm, and at 120 cm (Table 4.2). The 1996 data indicated that regrowth was hardly influenced by mid-dry season soil moisture. Both mid-dry season soil moisture and residual moisture was influenced by flood depth (Table 4.2). A positive correlation existed between maximum flood depth, below-ground biomass and % perennials, but these parameters had little influence on regrowth that was only correlated with soil moisture at 20-40 cm (Table 4.2).

Forage quality and N-yield Nitrogen content of aboveground biomass was between 0.38-0.74%, with minor differences between species and not correlated with biomass (R = -0.051, n = 23, p = 0.82), or with maximum flood depth (R = -0.27, n = 23, p = 0.21). Nitrogen yield was not correlated with maximum flood depth either (R = 0.26, n = 9, p = 0.50). Regrowth had a N-content between 0.9 and 2.2 %; the limited number of observations tend to have a negative correlation with above-ground biomass (R = -0.55, n = 7, p = 0.21).

4.4 Discussion

Comparison with other African seasonally-flooded grasslands Howell et al. (1988) considered the plains around Lake Chad too dry to be compared with the Sudd (Sudan), neglecting the Logone floodplain further south. In the Nyanya study area in the Sudd, long-term rainfall was 100-150 mm higher than in Logone (1981: 739 mm; 1982: 777 mm) (Mefit-Babtie 1983). Its species composition was comparable with Logone with Oryza longistaminata and Echinochloa pyramidalis as dominant species and Hyparrhenia rufa and Vetiveria nigritana occurring on the edges of the floodplain. In the Kafue flats (Zambia) that have a -1 somewhat higher rainfall of 880 mm year and comparable soils, Echinochloa stagnina and Vossia cuspidata dominated the deeper flooded parts (> 1 m), as in Logone (Ellenbroek 1987). The area with shallow floods (< 1 m) had however a different vegetation composition from Logone with only occasionally Oryza longistaminata as dominant species (Ellenbroek 1987). In the Inner Niger Delta (Mali), with a low -1 rainfall of 380 mm year , Oryza longistaminata dominated the parts with a maximum flood depth of about 50 cm, with Echinochloa stagnina and Vossia cuspidata 120 Part II – Impact of reflooding 1.00 – –.47 – –.11 – 1.00 – 1.00 1.00 3 .59 – – 1.00 2 .86 .66 Soil moisture half-way the dry season (February) 3 – – 1.00 – 3 –––––.52– ––––– .63 .75 .53 .036 .006 .016 2 – – .24 1.00 1.00 3 .63 .56 .59 .42 .009 .024 .009 .016 .001 <.0001 2 Biomass 0-15 cm 15-50cm 20-40 cm 60-80 cm 120cm .49 3 .33 .791 0.072 – depth ground moisture moisture moisture 2 .41.36 .158 1.00 .56 .49 1.00 .44 .46 .27 .086 .25 .88 Soil moisture at the end of dry season (May); 0.009 .22 .22 .10 .57 2 3 < 0.05) p .83 .69 .69 .53 .035 .003 0.03 <.0001 2 Max flood Regrowth Above- % Perennials Below-ground Biomass Soil Soil Soil 1.00 1.00 1 Sig .16 .47 Sig .054 cm Sig 0-15 R .49 –.011 – – Correlations between maximum flooddepth, regrowth, and intermediate parameters (1995, n = 7; 1996 16) Pearson Correlation Coefficient (SPSS 1999); Pearson Correlation Coefficient depthRegrowthAbove-ground Sig Biomass R R .49 % Perennials 0.60 .20 Sig 1.0 R .22 1.0 .70 Moisture Sig .17 Table 4.2 Table 1 –: not determined; in bold: significant ( 20-40 cm Soil R .16 – .22 – .31 – –.34 Flooding season:Max. flood 1995 1996 1995 R 1996 1995 1996 1995 1996 1995 1996 1995 1996 1995 1996 1995 1996 1995 1996 Below-ground Biomass cm 15-50 Sig R .052 .967 120 cm Sig .061 .12 – .96 .25 .24 .81 Moisture60-80 cm Soil Moisture Sig .71 R –.68 – .63 –.64 – – .45 –.024 – –.47 .42 .19 Soil R .54 4 – Vegetation production 121

occurring at a maximum flood depth of 1-2.8 m, occasionally up to 4-5 m (Hier- naux & Diarra 1983, 1984a). Biomass measurements in these three areas generally followed the methodology described for Logone, for differences see Annex 4.2. Information on maximum flood depth was not very accurate and limited the use of data (Annex 4.2).

In this sample of African seasonally flooded grasslands, above-ground biomass measured at the retreat of the floods was function of maximum flood depth as well (Fig. 4.6). Together with the Logone 1996 data, a consistent relation for African seasonally flooded grasslands, with a maximum flood depth of less than 1 m was found: Above-ground Biomass (g DM m-2) = 14.8 Max. Flood depth (cm) + 242.6 2 (R = 0.69, p < 0.0001).

Maximum flood depth – flood duration In addition to flood depth, flood duration has often been considered a main determinant of vegetation composition in wetlands (Ellenbroek 1987; Thompson 1985). In the Inner Niger Delta (Hiernaux & Diarra 1984a), maximum flood depth was more strongly correlated with flood duration than in Logone (R = 0.94, p < 0.0001 versus R = 0.545, p < 0.001). I attribute this difference on one hand to the large number of deeply inundated sites in the Inner Niger Delta and on the other hand to the large number of fish canals in Logone that accelerate draining (Drijver et al. 1995). The difficulties to measure flood duration and its low amplitude (see above),

Sudd y=30.1X- 8.94 R2 =0.97 ) 3000 -2 Kafue y=22.8X 504.9 R2 = 0.66

Inner Delta y=19.2X + 386.5 R2 = 0.75 (g DM m 2000 combined: Y=15X+273.5 (R2 = 0.68, p < 0.0001)

1000 Logone:Y=12.4X+263.7

above-ground biomass 0 0 20 40 60 80 100 120 140 max.flood depth (cm)

Figure 4.6 – Relation between maximum flood depth and total above-ground biomass in the Inner Niger Delta (Mali), Kafue flats (Zambia) and Sudd (Sudan) (0-120 cm max. flooddepth) (see also Annex 4.2) 122 Part II – Impact of reflooding

made us concentrate on the maximum flood depth-above ground biomass relation, discussed below.

Floodplain rehabilitation In 1995 and 1996, above-ground biomass of all species together as well as of the individual species Oryza longistaminata and Echinochloa pyramidalis were higher than in 1994, the first reflooding year. The correlation coefficient between maximum flood depth and biomass also increased after 1994, showing the two-year time lag in reaction of vegetation production to water level increase. We reported earlier the initial vegetation stress, due to the sudden rise of the water level, leading to the lack of flowering of Oryza longistaminata and the limited increase in cover of this species and Echinochloa pyramidalis compared to the following years (Scholte et al. 2000b). In 1994, Sorghum arundinaceum was still widespread and attained a high production in the re-flooded sites, but gradually disappeared in following years (Scholte et al. 2000b). This contrasted with the sudden die-off of Sorghum sp. in standing water in Sudan, reported by Bunting & Lea (1962). Con- trary to these changes in species composition and production was the lack of any noticeable impact on the morphological characteristics of the (top) soils within this three-year study period (Table 4.1). The impact of soil properties on (regrowth) production remained therefore unclear.

Maximum flood depth – Above-ground biomass Flood depth is a well-known environmental determinant of species distribution in wetlands (Ellery et al. 1991). My results show that maximum flood depth was also the main environmental factor explaining above-ground biomass production in the shallowly inundated Logone floodplain.

The difference in intercept between 1994 and 1996 of Oryza longistaminata and Echinochloa pyramidalis (Figs. 4.4, 4.5) can be explained partly by the 5-cm difference in maximum flood depth (Fig. 4.2). The Rain-Use Efficiency, based on the 0- intercept of maximum flood depth (Fig. 4.3), was with 2.5 and 3.5 kg DM ha-1yr-1 mm-1 in 1994 and 1996 respectively, well in line with the 1.5-6 kg DM ha-1yr-1 mm-1 range noted for the Sudano-Sahelian transition subzone (Le Houerou 1989).

At a maximum flood depth of 60 cm, virtually all biomass consisted of Oryza longistaminata and Echinochloa pyramidalis (Fig. 4.3 versus Figs. 4.4, 4.5). Under deeper flooding, Echinochloa stagnina and Vossia cuspidata were dominant, but restricted to depressions that were not considered in this study.

Micro-topical heterogeneity, caused by gilgai, is expected to have influenced the maximum flood depth – above-ground biomass relation that was assessed in 1 m2 plots. An experimental set-up that takes this heterogeneity into account may lead 4 – Vegetation production 123 to the establishment of a stronger maximum flood depth – above-ground biomass crop relationship.

Regrowth Mid-late dry season regrowth was very limited, but not neglected by grazing animals as shown by the difference in biomass in and outside the exclosures. The equation of Breman & de Ridder (1991) that relates above-ground biomass and regrowth would imply a regrowth of 42-147 g DM m-2 with a maximum flood depth of 20- 60 cm. This is considerably more than the maximum regrowth measurements of 10 g DM m-2 (1995) and 27 g DM m-2 (1996), following mid-dry season burning as commonly applied in the area. Regrowth in the Sudd area, under similar flood depths, was also low (Annex 4.2). More regrowth is only produced with a maximum flooding depth exceeding 50-100 cm and early dry season burning (Table 4.4; Hiernaux and Diarra 1984b), contradicting the linear biomass-regrowth relation of Breman & de Ridder (1991). The role of above and under-ground biomass and key soil properties in the production of regrowth remains to be clarified.

Biomass forage quality Digestible crude protein (DCP) of regrowth (1.7-9.3%), of above-ground biomass of Oryza longistaminata (0-0.75%) and of Echinochloa pyramidalis (0%) correspond to the extremes of the biomass-nutritional quality relationship of Howell et al. (1988) to which I refer for the development of forage quality throughout the year. Herbivores select the more nutritious parts of the grasses and target nutritious regrowth, maintaining a DCP content of 2.9% in their diet in the Inner Niger Delta (Hiernaux & Diarra 1984b).

Wet season growth was N-limited, as virtually none of the standing crop samples surpassed the physiological threshold of 0.5% N (Breman & de Ridder 1991; Boudet 1984). Also the lack of correlation between N-content and above-ground biomass points to this. In contrast, dry season regrowth has been largely water limited, but irrigation experiments in the Inner Niger Delta showed that with higher regrowth (>110 g DM m-2), production was greatly enhanced by NP fertili- sation (Hiernaux & Diarra 1984a).

Functional explanation of the maximum flood depth – above-ground biomass relation Junk et al. (1989) emphasised the local character of a possible flood depth - production relation, arguing that differences in inundation rates regularly disturb this relationship. The above-discussed time lag of vegetation production (2 year) and species recovery (1-10 year, Scholte et al. 2000b) to the start of reflooding are illustrative for such disturbances. Catling (2001), based on a large number of field measurements, concluded that most floating rices are adapted to an optimum flooding regime at which their full grain yield potential is expressed. 124 Part II – Impact of reflooding

The growing season is a function of soil moisture that was influenced beyond a depth of 120 cm, even with shallow flooding. An assessment of the development of soil moisture throughout the year was, however, beyond the scope of this field study. The N-growth limitation notwithstanding, results of this study did not support a link between N-yield and maximum flood depth or with the vegetation’s position in the transect, a measure for the distance water travelled and potentially an indication of sediment load (Breen et al. 1988; Scholte et al. 2000b).

Flooding drives adapted plants to grow up and keep pace with the rising water level (Ellenbroek 1987; Breen et al. 1988; Blom & Voesenek 1996; Cronk & Fennessey 2001). This shoot elongation may be practised by the deeply flooded Vossia cuspidata – Echinochloa stagnina communities with relatively weak stems that, supported by the buoyancy of the water, may reach 250 cm or more with a height of the aerial parts of about 80-100 cm at all seasons (Ellenbroek 1987). The maximum flood depth-above-ground biomass relation and the N-content that reached physiological threshold values suggest a comparable shoot elongation in the more shallowly inundated studied communities. The discussed Echinochloa pyramidalis has relatively strong fibrous shoots, however, that tower, already at the start of the flooding season above the approaching floodwater, refuting this explanation. In addition, Oryza longistaminata, the only important C3 grass on the floodplain, is also known to photosynthesize under reduced light conditions and relatively low ambient temperatures under water (Ellenbroek 1987; Cronk & Fennessey 2001).

4.5 Conclusion

It can be concluded that changing maximum flood depth, through water extraction or reflooding, not only has an impact on vegetation composition but also on its production. Based on data from Central as well as West, North-East and South- ern Africa, I showed that a rise in maximum flood depth with 1 cm, to a maximum of 1 m, corresponds to an increase in above-ground biomass with approximately 140 kg DM ha1. Regrowth after burning possibly increases as well, but is indirect- ly related with maximum flood depth, and only becomes substantial once maximum flood depth surpasses 50-100 cm. The factors, influenced by maximum flood depth, that determine plant production such as the length of growing season, soil moisture and induction of shoot elongation, have yet to be clarified. An experimental follow-up of this study should also take into account the relation of maximum flood depth with sediment load and other parameters that might influence productivity. 4 – Vegetation production 125

Annex 4.1 Overview of observation sites and measurements along the transect (see Figure 4.1)

Flooding season

1994 1995 1996

Study parameter Above-ground biomass, Above-ground Biomass, Above-ground Biomass, Fodder quality Regrowth, Fodder quality, Below-ground Biomass, Soil characteristics and Regrowth Plot size -moisture Soil moisture

50 x 20 m P7, P6, P5, P4 P7, P6, P5, P4 P7, P6, P5, P4

6 x 6 m – SI, S2, S3, S4 SI, S2, S3, S4

6 x 6 m Z2, Z4, Z6, Z8, Z9, Z10, Z11, – Z8, Z11, Z20, Z22, Z12, Z13, Z14, Z15, Z17, Z18, Z26, Z29, Z32, Z34 Z19, Z20, Z21, Z22, Z23, Z24, Z25, Z26, Z28, Z29, Z30, Z31, Z32, Z33, Z34

Total number of sites 32 8 16 126 Part II – Impact of reflooding

Annex 4.2 Production characteristics of seasonally flooded grasslands in the Sudd, Kafue Flats and Inner Niger Delta (see also Fig. 4.6) 2 2 1m 0.5 m ? times 10 times -2 8 8 8 8 8 9 10 – – 167 9±5 7±4 5±3 7±4 17±8 Vossia cuspidata Vossia Vossia cuspidata Vossia Echinochloa stagnina Echinochloa stagnina Echinochloa stagnina Oryza longistaminata Oryza longistaminata Oryza longistaminata Oryza longistaminata Oryza longistaminata Echinochloa pyramidalis Echinochloa pyramidalis 2 2 1 m ? times Replicates after burning replicates 5 ) -2 (234±63)(210±70) 10 times (102±49) (231±45) (821±337) (idem dominant species) Max gm (cm) -ground Biomass (gm 4 36 1100 ±220 120 2486 depth 2 1 Sudd 2 9 277±57 Site no. Max. Flood Sudd 1 Max. total Above Sample size Dominant species Regrowth Sample size Kafue Flats 2 100Kafue Flats 5 85 1198 1547 Sudd 4Sudd 5 6Kafue Flats 1 13 254 ±71 283± 35 Sudd 3 10 270 ±62 0.5 m 4 – Vegetation production 127 2 meas- 12 4 m 12 -2 11 11 12 11 – – – – 3 times 9.2±0.4 10.9±3.2 92.4±6.6 11.9±1.4 137.3±9.8 Flood depth has generally only 4 Measured monthly, maximum value presented, Measured monthly, 8 Measured at the end of main flooding sea- 5 measured in June , 6 months after burning; 11 Panicum sp. Panicum repens Leersia denudata Setaria sphacelata Vetiveria nigritana Vetiveria nigritana Vetiveria Oryza longistaminata Oryza longistaminata Oryza longistaminata Oryza longistaminata Hiernaux and Diarra (1984a); Paspalidium obtusifolium 3 not determined; 2 10 Replicates after burning replicates Ellenbroek (1987); 2 Measured throughout the dry season; 7 5 ) -2 (1988); 7 7 et al. 6 (idem dominant species) Max gm ± 2 months after drying up, not burned; 9 1491 at moment of drying up; → (cm) -ground Biomass (gm 4 49 1220.5±124 1747 (top) 3 6 depth Interpreted from Mefit-Babtie (1983), see also Howell normally 2-3 months after burning; ured in June, 4 months after burning Kafue Flats 7 60 350 Kafue Flats 8 40Inner Delta 79-80 479 Inner Delta 2 774±89 Site no. Max. Flood Kafue Flats 6 Max. total Above 70 Sample size Dominant species Regrowth 1747 Sample size Inner Delta 80-81 9Inner Delta 81-82 24Inner Delta 82-83 0 471.3±121 541.3±37Inner Delta 279.1±18 49 1m 24 times 1582.8±255 1 son (as in Logone); been assessed twice, resulting in an underestimation of maximum flood depth compared to our study;

5 Waterbird Recovery in Waza-Logone (Cameroon), resulting from increased rainfall, floodplain rehabilitation and colony protection

Paul Scholte

Abstract The impact of reflooding of a desiccated floodplain on waterbirds in North Cameroon was monitored by January total counts from 1992-2000. Bird numbers in the dry season increased from 60 000 to 105 000, whereas the number of species surpassing international 1% criteria doubled from 6 to 12. The increase in Anatidae corresponds to their recovery in West Africa following the droughts in the 1980s. The increase in especially Ciconiiformes in Waza-Logone was not paralleled by similar trends in other main West African floodplains, suggesting that floodplain rehabilitation has played an important role. The limited increase of the large piscivorous Ciconiiformes Leptoptilos crumeniferus, Mycteria ibis and Pelecanus rufescens is associated with repeated destruction of their breeding colonies, possibly leading to their decline as in the rest of West Africa. In contrast, an Ardea melanocephala colony increased from 750 to 2500 nests between 1993-2003. The exceptional colony size, a multiple of the next largest known colonies, implies that besides improved habitat due to reflooding, protection also played a vital role. It can be concluded that the increase in waterbirds in Waza-Logone is due to a combination of factors: improved rainfall (especially Anatidae); floodplain rehabilitation (especially omnivorous Ciconiiformes); and protection measures (selected Afrotropical Ciconiiformes).

131

5.1 Introduction

Studies on the population dynamics of waterbirds have centered on the impact of climatic and hydrological variability. Their scale varied from continent-wide responses to erratic mass flooding in the arid zones of Australia (Roshier et al. 2001), to reactions to local changes in water regime in temperate France (Paillisson et al. 2002). In line with these population dynamics are the responses of waterbirds to man-made changes in the hydrology of wetlands, the subject of this chapter.

After the severe drought in Sahelian Africa in 1972-1974, its impact on numbers of West European migratory waterbirds such as purple heron Ardea purpurea became clear (Held 1981). In reaction to the 1980s Sahelian droughts, a range of other studies appeared on the impact of the variation in West African rainfall on Palearctic birds (Szep 1995; Foppen et al. 1999; Fasola et al. 2000; Schricke et al. 2001; Barbraud & Hafner 2001). The impact of these droughts has been more obvious than have been the long-lasting effects of water control projects undertaken in many major African wetland systems. Triplet & Yesou (2000) attributed negative trends in numbers of Anatidae in the Senegal Delta to reduced flooding following the construction of embankments and dams in the Senegal River. The impact of drainage works in the Everglades (USA) on breeding waterbirds is amongst the best documented. Since the 1930s, the number of pairs decreased from 50% in case of Great Egrets (Casmerodius albus) to 90% for Woodstork (Myc- teria americana) (Bancroft 1989).

In response to these negative impacts, wetland rehabilitation efforts have received increasing attention, especially in Europe and North America, but only on relatively small areas (Middleton 1999). Their impact on a large scale is still poorly known, the potential importance of rehabilitation for waterbird numbers notwithstanding (Roshier et al. 2001). In Hadjia-Nguru, Nigeria, numbers of Anatidae were correlated to the size of the area impacted by flood releases (Polet 2000). Anatidae numbers also correlated with time, suggesting an alternative explanation in the recovery of waterbird populations, as observed elsewhere in West Africa after the mid-1980s droughts (Schricke et al. 2001). In addition, high water levels are not necessarily an indication of suitable habitat for shorebirds and egrets (Bancroft 1989; Butler et al. 2000), and may even have a negative impact (Crivelli et al. 1995). An analysis of counts of waterbirds using temporarily flooded rainfed lakes in Niger showed that bird densities were inversely correlated to the size of the inundated area, with smaller lakes having higher nutrient loads (Mullié et al. 1999). In addition to these trophic related factors, nest-site availability may also be a limiting factor for colonial waterbirds (Hafner 2000). These experiences show the need to study the development of waterbird populations not only in relation to flood levels, but also to other environmental and anthropological factors. 132 Part II – Impact of reflooding

Figure 5.1 – The Waza-Logone floodplain and colony locations (see section 5.3). Year of colony destruction is indicated behind the location name. 5 – Waterbirds 133

The Waza-Logone floodplain in North Cameroon (Fig. 5.1) was severely degraded due to restricted water supply resulting from the construction of a dam and embankment upstream. Built in 1979, these structures caused the disappearance downstream of the perennial floodplain vegetation in an area covering 1500 km2. This habitat loss had a strong negative impact on livestock and fisheries, as well as on antelopes (Scholte et al. 1996a). Lack of sound data prevented a quantitative assessment of the effects of the embankment and dam on waterbirds except for the western Black-crowned Crane Balearica pavonina pavonina. For this endangered species Waza-Logone is a stronghold with about 2500 individuals or 16% of the estimated population (IUCN 2003; Scholte 1996; Beilfuss et al. 2003). Prior to the construction of the dam, however, its population numbered over 10 000 individuals (Holmes 1972). Balearica pavonina is resident and the decreased population size can be attributed to changes in the Waza-Logone area proper. Floodplain degradation following the construction of the dam is the most likely cause thereof.

To counter the ongoing degradation, an embankment was breached in 1994, raising flooding levels in an area of 600 km2 and rehabilitating 180 km2 of perennial floodplain vegetation (Scholte et al. 2000b). Annual total waterbird counts, organized in the framework of the African Waterbird Census (Dodman & Diagana 2003), formed the main tool to monitor the impact of the reflooding on waterbirds (Schol- te et al. 1999a; Scholte et al. 2000a; Fotso et al. 2001).

In this chapter, I analyse the impact of long-term rainfall, floodplain rehabilitation and protection measures to the observed increases in numbers of waterbirds. To discern the impact of floodplain rehabilitation from long-term rainfall changes, a comparison is made with counts from West Africa sites with a comparable rainfall history. An expanding colony of the Black-headed Heron Ardea melanocephala allowed a complementary assessment of the impact of floodplain rehabilitation on this terrestrial feeder preferring damp grassland (Brown et al. 1982; Kushlan 2000), greatly extended with the reflooding (Scholte et al. 2000b). The developments of Ardea melanocephala numbers are compared with those of other, but unprotected, colonial breeders in the area. This allowed an appreciation of the importance of colony protection for the recovery of waterbirds as well.

5.2 Materials and methods

Study area The Waza-Logone area covers approximately 8000 km2 in the Far North province of Cameroon (Fig. 5.1). It receives an average annual rainfall of about 650 mm. The rainy season lasts from June to September, followed by flooding of the Logo- ne River, in parts of the area, from August to November. Rainfall and flooding are complementary in their impact on flood depth. 134 Part II – Impact of reflooding

After the floods ceased in 1979, annual grasses invaded the desiccated floodplain. The reflooding, started in 1994, initiated an annual conversion of 7% of the annu- 2 ally reflooded 180 km back into productive perennial grasslands (Scholte et al. 2000b). Each year at the end of December, immediately after the drying up, large stretches of the floodplain are burnt by pastoralists, leaving a bare landscape that soon turns green again with perennial grasses’ regrowth, attracting livestock, antelopes and birds.

Environmental factors The Ndjamena station, although located at a distance of c. 100 km, is representative of rainfall in the Waza-Logone area. Discharge of the Logone river, at the entry of the floodplain (Fig. 5.1), was measured during the flooding season. Discharge is considered to be a function of rainfall in the upstream catchment area and determines the extent of flooding in Waza-Logone (Sighomnou 2000). Vegetation composition and cover were monitored in a 20 x 16 km grid (Scholte et al. 2000b).

Dry season waterbird counts in Waza-Logone Waterbird counts correspond to the 1992 and 1994-2000 rainy and flooding seasons. Counts took place in January, halfway through the dry season, except for the 1994 and 2000 flooding seasons when counts were conducted in February. Two teams, of 2-3 observers each, covered, by car and foot, the entire Logone floodplain in two weeks for a total waterbird count (Van Wetten & Spierenburg 1998; Dijk- stra et al. 2002; Bredenbeek 2004). Each team consisted of at least one observer who participated in all counts. Special attention was paid to the remaining wet parts of the floodplain, such as depressions and streams. Observation intensity, expressed as number of observation days, was not correlated with the number of observed 2 birds (R = 0.05, p = 0.69).

Waterbird species were categorised following the main taxonomic groups, Anse- riformes; Charadriiformes, with Laridae and Sternidae kept separate; Ciconiiformes, plus Pelecaniformes (i.e. Phalocrocorax africana and Pelecanus spp.) and Balearica pavonina. Ciconiiformes were subdivided according to status and annual lifestyle, i.e. Palearctic migrants versus Afrotropical residents or inter-Africa migrants (Schol- te et al. 1999a), to examine the possible dependency on breeding possibilities in the area. Trophic characteristics correspond largely to habitat use during the count period, i.e. depressions for piscivorous birds and damp grasslands for omnivorous birds.

Ardea melanocephala colony description and counting methods The Ardea melanocephala colony is situated in Sclerocarya birrea woodland in the southeastern corner of Waza National Park, 10 km west of the reflooded area (Fig. 5.1). The colony lies next to Andirni (Fig. 5.1), a village home to several Waza Na- tional Park guides, who regularly survey the colony and protect it from heron rob- 5 – Waterbirds 135

bers. The area has witnessed an influx of seasonal labour from Chad, bringing in cultures with other eating habits that pose a threat to large birds. In 1998, for instance, some of them were caught at night trying to trap young herons.

In 1993, the colony had existed over a decade and was restricted to the woodland west of the village. In 1994 it expanded into the area south of the village, exposing it to heron robbers. In 1998 the colony moved again into the western woodland and in 2002 to the north of the village, where the most effective protection could be expected. Till 1995, three small Ardea melanocephala colonies (<50 nests) were found in the direct vicinity of Waza National Park.

Nest counts were carried out during 1993-2000 and again in 2003 in June-August, after the first heavy rains (Table 5.1). Organisational problems forced us to bring forward our visits of 1995 and 1998 when the main rains had not yet started. In 1997 the colony was counted both prior to and during the main rains. The heronry was subdivided in five parts, based on position relative to the village. From 7.30 till 12 am, we counted the number of nests and adult herons in each nest tree. During the August counts in 1993 and 1997, the number of juveniles per nest tree was counted to evaluate the colony’s success-rate. This number should be considered a minimum, as only the larger juveniles were detectable.

Table 5.1 – Breeding parameters of Black-headed Heron Ardea melanocephala colony

date of presence of Number Adult/nest Young/nest nest trees nests/tree count Bubulcus ibis of nests1 (median)2 (median)4 (number) (median)2

1993 7 Aug breeding 742 nd3 0.60a 76 9.0bc 1994 23 June 533 present 1160 nd3 139 nd 1995 5 June no 718 2.33a 102 5.8e 1996 2 June >100 present 1604 1.71c 201 7.0cd 1997 17 May no 1229 2.00ab 157 7.0de 1997 16 Aug breeding 2012 0.94f 0.55a 180 10.0b 1998 12 June no 1372 1.80b 150 8.0c 1999 1 July 1052 present 2418 1.60c 167 14.0a 2000 15 July breeding 2016 1.43d 197 9.0bc 2003 12 July breeding 2479 1.00e 180 13.0a total 1.50 0.56 9.0

1see Fig. 5.4 for corrected numbers (explained in Results) 2different letter (e.g. a-b), p < 0.05, if at least one letter is the same p > 0.05 (e.g. a-ab) 3nd: not determined, only total known 4p = 0.91 136 Part II – Impact of reflooding

Other Ciconiiformes colonies Our year-round presence in the study area in 1992-1997 and frequent contacts with local communities allowed the survey of colonies of other Ciconiiformes. The species that were involved were egrets, in particular Cattle Egret Bubulcus ibis and the large piscivorous species, Marabou Stork Leptoptilos crumeniferus, Yellow-billed Stork Mycteria ibis and Pink-backed Pelican Pelecanus rufescens.

Data analysis Categories of the dry season counts were mutually compared with GLM analysis. Developments in waterbird numbers, rainfall, river discharge and perennial vegetation cover were assessed by linear regression. Developments of the waterbird counts and colony were described with regression analysis with best-fitted curve. Median values were compared with Mann-Whitney tests. Correlation between number of nests and adults per nest-trees was investigated with Spearman rank correlation (SPSS 1999).

Laridae & Sternidae R2 120000 Ruff Shorebirds (-ruff) Total: 0.67 Anatidae Afrotropical omnivorous 0.24 Palearctic omnivorous 100000 Ciconiiformes, Pelecanus, Balearica Afrotropical piscivorous } Palearctic piscivorous 0.01 80000

0.41 60000 0.63

number of waterbirds 40000 0.61

0.01 20000 0.35 0.29

0 1992 1994 1995 1996 1997 1998 1999 2000

reflooding flooding season

Figure 5.2 – Dry season waterbird numbers in Waza-Logone Legend: R2 of linear regressions, in bold if significant (p < 0.05) 5 – Waterbirds 137

5.3 Results

Waterbird dry season numbers Total waterbird numbers increased from 60 000 to 105 000 during 1992-2000 (Fig. 5.2). Palearctic migrants made up 53% of the observed birds. Their numbers 2 fluctuated strongly from year to year and did not show an increase (R = 0.12, p = 0.40). The Afrotropical waterbirds on the other hand averaged 47% of the observations and showed a marked increase during the observation period (R2 =0.60, p = 0.025).

In the period 1992-2000, 19 water bird species were present in numbers surpassing 1% of their known population size (Wetlands International 2002). The number of species passing this criterion found in a single year doubled between 1992 and 2000 (Fig. 5.3). Species that surpassed the 1% criterion in all years were White- faced Whistling Duck Dendrocygna viduata, with numbers up to 6% of known population size, Balearica pavonina pavonina (up to 16%) and Collared Pratincole Glareola pratincola (up to 39%).

The total number of waterbirds per category increased significantly in the years between 1992 and 2000 (F = 19.3, p < 0.0001, df = 1,55) for all categories. The relative increase in bird numbers over time did, however, not differ significantly amongst the categories (F = 1.4, p = 0.23, df = 7,48). The absolute increase in num-

14 12 y = 3.56Ln(year) + 4.28 R2 = 0.63 p = 0.019 10 8 6

of population size 4 2 number of species present with > 1% 0 1992 1994 1995 1996 1997 1998 1999 2000

reflooding flooding season

Figure 5.3 – Number of waterbird species in Waza-Logone surpassing 1% of known population size, 1992-2000. 138 Part II – Impact of reflooding

bers of waterbirds was mainly due to the Afrotropical Anatidae such as Sarkidior- nis melanota and to Afrotropical omnivorous Ciconiiformes, in particular Ardea melanocephala and Bubulcus ibis (Fig. 5.2). Other species that showed at least a five- fold increase were Squacco Heron Ardeola ralloides (Palearctic piscivorous Ciconii- formes) and the shore birds Wood Sandpiper Tringa glareola and Black-winged Stilt Himantopus himantopus (Palearctic), Jacana Actophilornis africana and Spur-winged Plover Vanellus spinosus (Afrotropical). No increase was observed for Ruff Philo- machus pugnax nor for migratory omnivorous Ciconiiformes, whereas numbers of the large resident piscivorous Ciconiiformes Leptopillos crumeniferus, Mycteria ibis 2 and Pelecanus rufescens fluctuated over time (R = 0.11).

The preceding season’s rainfall was not correlated with numbers observed for any of the waterbird species or categories counted. Logone discharge, correlated with the flooding extent, was only correlated with total number of waterbirds observed 2 (R = 0.61, p = 0.038).

Ardea melanocephala Dry season numbers of Ardea melanocephala increased only from 1997 onwards reaching 5800 in 2000 (Fig. 5.4), largely surpassing 1% of the known population.

In the early 1995, 1997 and 1998 colony counts, Ardea melanocephala was present in large numbers but not yet breeding, as shown by the higher number of adults per nest compared to later counts (p < 0.0001) (Table 5.1). The number of adult herons per nest in the early 1997 count was intermediate to the early 1995 and 1998 counts (Table 5.1), which have been corrected with the 1997 late/ early season ratio of 1.64.

The size of the colony increased steadily from 742 nests in 1993 to 2418 nests in 1999 and subsequently leveled off (Fig. 5.4). There was no correlation between the 2 number of Ardea melanocephala nests and rainfall (R = 0.27, p = 0.19), nor with 2 Logone discharge (R = 0.14, p = 0.40). Perennial vegetation cover in the reflood- 2 ing impact zone was correlated to the number of nests (R = 0.92, p = 0.001). The relation with preceding year’s perennial vegetation, expected to have supported the population during the harsh dry season, was however less pronounced (R2 = 0.63, p = 0.032), suggesting non-causality.

Other Ciconiiformes colonies Bubulcus ibis nested scatteredly in villages in the area, with no indications of disturbance. Traces of other egret colonies were found in the northern part of the area that showed signs of destruction. Marabou stork Leptoptilos crumeniferus colonies have been subject to frequent disturbances, often targeting the young for consumption. After the destruction of the Zina town colony in 1994 (±250 nests), colonies with about 50 nests were found in the towns of Pouss and Guirvidig (Fig. 5.1). 5 – Waterbirds 139

6000 nests A.melanocephala dry season numbers A. melanocephala 5000 dry season numbers large piscivorous Ciconiiformes.

number 4000

3000 y = 1073Ln(x) - 50.6 R2 = 0.91

2000

1000

0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2003 reflooding flooding season

Figure 5.4 – Breeding and dry season Ardea melanocephala and large piscivorous waterbirds, Waza-Logone 1992-2003.

A Leptoptilos crumeniferus colony was found in 1996 near Halé on the border of Waza National Park, where at least 210 nests were destroyed in March 1997 (Fig. 5.1). No colonies were found in 1998. In 1999, Leptoptilos crumeniferus was found breeding in the center of Waza NP (80 nests), where the colony grew steadily to over 500 nests. Young were robbed again in 2002 after which the site was abandoned. Till 1997, a small colony of Mycteria ibis (±60 nests) and pink-backed pelican Pelecanus rufescens (±30 nests) found relative protection inside Waza National Park (Fig. 5.1). Villagers explained that its subsequent disappearance was caused by destruction by fishermen. Only once we found a larger Mycteria ibis colony with 550 nest-building individuals near Mahé village in October 1997 (Fig. 5.1).

5.4 Discussion

General developments in waterbird numbers The standardised itinerary and stable personnel assured a constant counting intensity, not achieved in most African waterbird counts (Dodman & Diagana 2003). We found, however, large inter-annual fluctuations, especially of Palearctic birds, also observed in Niger (Brouwer & Mullié 2001). Initially the reflooding resulted in an increase in large piscivorous species, i.e. Great White Pelican Pelecanus ono- crotalus, subsequently followed by an increase in smaller piscivorous egrets. I attrib- 140 Part II – Impact of reflooding

ute these developments to increasing fishing intensity that targets large fish (Scholte et al. 1998; Scholte 2003a).

Increased ornithological importance since the reflooding Published waterbird population estimates seldom cover Central Africa (Wetlands International 2002). A choice for (North-) East African instead of the applied West- Africa population estimates would have given 17 species surpassing the 1% criterion. For species such as White Stork Ciconia ciconia, a choice between West and East African populations ignores the position of Central Africa at the crossroad of these populations (Brouwer et al. 2003).

Not only the number of species surpassing the 1% criterion has increased but also the observed numbers of each of these species important for conservation. The endangered Balearica pavonina pavonina, for example, showed an upward trend 2 (R = 0.48, p = 0.057).

Comparison with other West African waterbird counts: impact of generally improved rainfall conditions The Sahel region from Senegal to Chad, including the Waza-Logone area, was distinguished as a region with a strongly correlated rainfall pattern (Gommes & Pe- trassi 1996). In the 1950s and early 1960s, rainfall was well above long-term-average. After this, a gradual decline took pace which included the Sahelian droughts of 1969 and 1971-74. This ‘first Sahelian drought’, was followed by a relatively humid period that ended with the severe 1983-85 drought, after which a gradual, though highly fluctuating, increase in rainfall has taken place (Beauvilain 1995; Gommes & Petrassi 1996).

To discern the impact of generally increased rainfall from the impact of floodplain rehabilitation, developments in waterfowl numbers in Waza-Logone were compared with January counts from other West African countries (Dodman & Diaga- na 2003). We applied the same taxonomic and functional categories as distinguished above. The only data for which observation effort was more or less constant is from the Senegalese part of the Senegal Delta (Schricke et al. 2001; Triplet pers.comm. 2004) and, looking at Anatidae only, from Hadjia-Nguru in neighboring Nigeria (Polet 2000; Dodman & Diagana 2003). Although only covering the period 1992-1997, the Niger country-wide data was used because of the prominent presence of Ciconiiformes, as in Waza-Logone (Brouwer & Mullié 2001).

The total numbers of birds per distinguished category (Fig. 5.2) in the Senegal Delta did not show an increase (F = 1.95, p = 0.17, df = 1,62), but the categories differed significantly in their relative increase (F = 7.05, p < 0.001, df = 7,55). This was due to the remarkable increase in Afrotropical and Palearctic Anatidae in Sene- 2 gal (R = 0.73, p = 0.004). This was paralleled in 1993-1994 in Niger and in 1994- 5 – Waterbirds 141

1997 in Nigeria. Anatidae numbers in Niger and Nigeria dropped since 1994 and 1998 respectively, possibly related to their concurrent increase in Waza-Logone.

Shorebirds that increased in Waza-Logone, did not show any trend in Niger nor in Senegal. The lack of trend in numbers of Philomachus pugnax in Waza-Logone (Fig. 5.2) was paralleled in other West African wetlands (Trolliet & Girard 2001).

Omnivorous Ciconiiformes, including herons, showed an increase from 1992-2000 in Waza-Logone but not elsewhere in West Africa. This was partly due to the sudden increase of Bubulcus ibis in 1994 in Niger, the Senegal Delta, albeit in low numbers, and up to 260 000 in the Inner Niger Delta in Mali, attributed to the exceptional 1994 flooding (van der Kamp et al. 2002a). A comparable increase of Bubulcus ibis in Waza-Logone took place in 1998 with high numbers still present in 1999. A comparison of the population trend of the Ciconiiformes Ardea melanocephala, Mycteria ibis and Leptoptilus crumeniformes between Waza-Logone and other West African sites is hampered by the rarity of these species in the latter sites. The smaller piscivorous Ciconiiformes, i.e. egrets, showed widely fluctuating numbers in all areas, with exception of Ardeola ralloides with its upward trend in Waza-Logone. Of the omnivorous Ciconiiformes, the abundant Bubulcus ibis and the rare Ardea melanocephala showed a 15% decrease (Van der Kamp 2002b). Amongst the few increasing resident piscivorous Ciconiiformes in the Inner Niger Delta one should note Phalocrocorax africanus (van der Kamp 2002b).

Improved rainfall conditions may largely explain the increase of Anatidae in Waza- Logone. The increase of especially Ciconiiformes relied on other factors, of which floodplain rehabilitation and colony protection will be discussed below.

Impact of floodplain rehabilitation on waterbird numbers The impact of floodplain rehabilitation, induced by the 1994 reflooding, was two- fold: a change of annual into perennial vegetation in an area of approximately 180 2 km (Scholte et al. 2000b) and an increase in vegetation production and water availability in an area of 600 km2 (Chapter 4).

In 1994, the first reflooding year, we compared waterbird diversity in always-flooded perennial grassland, reflooded grassland with still a predominantly annual grassland vegetation that would convert in later years into perennial grassland and degraded annual grassland vegetation, not subject to reflooding (Scholte et al. 2000a). Recently reflooded and especially undisturbed flooded habitats with a water depth of 20-35 cm showed the highest density of passerines and waterbirds. These observations correspond to the observed range for small herons and large herons in the Everglades (USA) with a water depth of 10 to 40 cm respectively (Bancroft, 1989) and in flooded grasslands in the Inner Niger Delta, Mali (van der Kamp et al. 2005). With the progressing floodplain rehabilitation, differences be- 142 Part II – Impact of reflooding

tween reflooded and undisturbed habitats in Waza-Logone have gradually disappeared (Scholte et al. 2000b).

The impact of the floodplain rehabilitation is limited to about a third of the yearly-inundated floodplain. The floodplain rehabilitation targeted the area adjacent to Waza National Park, where wildlife, including (resident) waterbirds receive protection. This may explain the relatively large impact of reflooding, further enhanced by colony protection as discussed below.

Ardea melanocephala versus other Ciconiiformes: linking colony protection with floodplain rehabilitation The number of Ardea melanocephala counted in the Waza-Logone floodplain from 1992-2000 equaled 25-80% of the numbers counted in 23-36 African countries (excluding Chad and Sudan). Only in the 10 000 km2 Sudd area in Sudan, surveys in the early 1980s revealed Ardea melanocephala numbers comparable to those in Waza-Logone (Howell et al. 1988).

From 1993 to 1999, total dry season Ardea melanocephala numbers were lower than numbers counted at the colony (Fig. 5.4). A. melanocephala also showed a time lag of three – four years between the reflooding and the increase in dry season numbers, longer than other waterbirds (Fig. 5.2). These observations may be attributed to dispersal of juveniles and young adults that was reported in Southern Africa to cover distances of up to 800 km (Tarboton 1977). Presented results showed that the direct impact of rainfall during the study period on the increase of the Ardea melanocephala population was limited. Ardea melanocephala dry season increase was not related with the Logone river discharges that influence the flooding in the entire Waza-Logone floodplain. I suspect that the large changes due to reflooding have masked the influence of rainfall and flooding.

With 742 nests (Fig. 5.3), the Ardea melanocephala heronry was in 1992 already three times the size of the largest known colonies in Kenya (North 1963; Parsons 1976; Taylor 1972) and Somalia (Ash & Miskell 1998) and a multitude of the maximum size of 100-150 reported from South Africa (Skead 1964), Tanzania (Beesley 1972) and West-Sudan (Wilson 1982). In West Africa, reported colony sizes are smaller than 100 nests (Elgood et al. 1994; Cheke & Walsh 1996; van der Kamp et al. 2002b).

The number of breeding pairs in the Ardea melanocephala colony increased from the first reflooding year onwards. The constant size of the colony since 1999, suggests that availability of feeding in the colony surroundings limits further increases (Fasola & Barbieri 1978; Gibbs et al. 1987). Low breeding success, less than one young per nest (Table 5.1), was also reported from E.Africa (Brown et al. 1982). The increase in colony size in especially the first years can not be explained by the 5 – Waterbirds 143 output of the colony only and must have been supplemented by immigration. This reasoning is supported by the above-mentioned reduction in number of smaller colonies in the area, probably due to human pressure.

Hafner (2000) argued that exceptionally large nesting assemblages are, on a regional scale, caused by a shortage of suitable nesting places. The increased Logone colony occupied an increasing number of nesting trees that did not lead to an increasing number of nests per tree (Table 5.1). This suggests ample tree availability. The colony’s location, in the eastern woodland area on the edge of the floodplain, predicts ample tree availability in the near future as well. The availability of nest trees close to the village, source of protection, may be of crucial importance however.

The development of the large Ardea melanocephala colony in Waza-Logone is rem- iniscent of the grey heron Ardea cinerea population in France. The latter recovered, after severe prosecution, from almost nil in 1900 to 19 000 pairs in 1990. In 1958, more than half of the French population was found in one single well-protected colony of 1 300 pairs, subsequently splitting into several new colonies (Ma- rion et al. 2000). In the Netherlands, the Ardea cinerea population increased from about 6 500 pairs in 1926 to 10 000 pairs at present, while the number of colonies increased from 130 to 500 due to increased protection. Large colonies with over 500-1000 nests, in well-protected privately owned woodlands as occurred during and just after periods of prosecution, are no longer found (Braaksma et al. 1950; Van der Weide 2002). The Ardea melanocephala colony in Waza-Logone reminds of the earlier European situation, although the tendency, towards increased prosecution, is the reverse.

The lack of increase of large piscivorous Leptoptilos crumeniferus, Mycteria ibis and Pelecanus rufescens in 1992-2000 in Waza-Logone is attributed to the repeated destruction of their colonies. These species have become increasingly rare in the Inner Niger Delta and the Senegal Delta, despite their rich fish resources (Van der Kamp et al. 2002a). Consumption of chicks by immigrants has often been the direct cause of colony disturbance in Waza-Logone. I expect that increasing competition for fish resources has made local communities indifferent to the destruction of the colonies in their territories. In the Inner Niger Delta in Mali, where larger piscivorous birds have already become rare, reproduction sites of the remaining smaller piscivorous Egretta species have now come under increasing pressure (Van der Kamp et al. 2002b, 2005). 144 Part II – Impact of reflooding

5.5 Conclusions

Developments in the waterbird populations of Waza-Logone show the link between long-term rainfall conditions and floodplain rehabilitation. To quantify the relative contributions of these factors a more thorough knowledge of regional waterbird population dynamics is necessary to reveal possible exchanges and compare developments between West-central African wetland areas. Long-term monitoring, with constant or at least quantified effort, is vital in this context.

The observations in Waza-Logone show that in addition to improved rainfall conditions and floodplain rehabilitation, colony protection has played a crucial role for at least some of the Afrotropical waterbirds. The role of colony- and habitat protection is likely to become more prominent in future because of the increasing human pressure on floodplain resources (Scholte 2003a). Initiatives such as the colony protection at Andirni merit further support to ensure their continuity, possibly through a modest touristic exploitation that elsewhere has proven its success (Bouton & Frederick 2003). 6 Antelope Populations in Waza National Park (Cameroon) from 1960 till 2001: Impact of changes in rainfall, hydrology and human pressure

Paul Scholte, Saleh Adam and Bobo Kadiri

Abstract Antelopes are amongst the most prominent wildlife in Waza National Park which, situated in the Sahelo-Sudanian savanna of Cameroon, has witnessed changes in rainfall, hydrology and human encroachment. To assess their impacts, we reviewed 26 surveys, comprising total, transect and localised counts, both aerial and terrestrial. Estimated Kob numbers dropped from 20 000 in the 1960s and 1970s to 2000 in the mid-1980s, recovering to 5000 in the 1990s. Estimated Korrigum (‘Topi’) and Roan numbers dropped already in the early 1970s, for reasons largely unknown, and slightly recovered to respectively 2000 and 1000 in the 1990s. The diversity of counting methods notwithstanding, the drop in Kob numbers and the extinction of Waterbuck can be attributed to the construction of the upstream Maga dam in 1979 and subsequent low rainfall. Yet population structure data showed that Kob’s reproduction capacity was not severely hit and Kob recovered in the late 1980 and early-mid 1990s, but did not persist, despite increased flooding. Although few signs of poaching have been reported, it might have been important in periods of drought, when Kob and Korrigum left the protected area.

147

6.1 Introduction

Waza National Park (NP), situated in the Sahelo-Sudanian zone in North Cameroon, has been the subject of ecological studies such as those on Giraffe Giraffa camelopardalis L. (Ngog Nje 1983, 1984), Elephant Loxodonta africana Blumenbach (Tchamba 1995, 1996) and Lion (Panthera leo L.) (Bauer 2003). Waza NP also harbours large populations of Kob (Kobus kob Erxleben), Korrigum (Damaliscus lunatus korrigum Burchell), Roan (Hippotragus equinus Desmarest) and Red-fronted Gazelle (Gazelle rufifrons Gray), as well as minor populations of Bohor Reedbuck (Redunca redunca Pallas) and Common Duiker (Sylvicapra grimmia L.). These antelopes are important resources, from a conservation as well as tourist perspective (Flizot 1962; East 1999). Previously published antelope population assessments in Waza NP (Esser & Lavieren 1979; Tchamba & Elkan 1995) described the distribution of antelopes based on a single survey in December, at the end of the flooding season, underestimating the role of the floodplain during the subsequent dry season. Their limited time-span also hampered an assessment of the changes triggered by droughts, natural as well as man-made.

The Waza-Logone floodplain was severely degraded when it was cut off from its main water supply through the construction of a dam upstream. Built in 1979, this dam has caused the disappearance downstream of the perennial floodplain vegetation in an area of 1500 km2, including approximately a third of Waza NP, with a dramatic impact on livestock and fisheries. To counter this ongoing degradation, the IUCN Waza-Logone project was started. In 1994 an embankment that had closed off a watercourse since 1979 was breached, bringing back the annual floods in an area of 180 km2. This triggered an annual 7% restoration of perennial floodplain vegetation in the reflooding impact zone, including a part of Waza NP and a tripling of livestock densities (Scholte et al. 2000b; Scholte 2003a).

It was argued that antelope populations also suffered from the decreased flooding and should recover after the 1994 reflooding. To assess the antelope populations during the floodplain rehabilitation, annual counts were initiated at the waterholes in Waza NP still holding water at the end of the dry season. However, changes in flooding cannot be interpreted in isolation from the inter-annual fluctuating rainfall. We therefore reviewed all 26, mostly unpublished wildlife counts, conducted from 1960 till 2001 with varying target species and methodology. The number of counts and their repartition in time allowed an assessment of the impact on antelope populations of the dramatic changes in hydrology and rainfall that Waza NP witnessed in the second part of the 20th century. 148 Part II – Impact of reflooding

6.2 Materials and methods

Study area Waza NP covers 160,000 ha and lies in the transition zone between the Sudan and Sahel savannas. The sandy soils in the slightly elevated south-western and southern part of Waza NP support a wooded Sclerocarya birrea and Anogeissus leio- carpus savanna with Hyperthelia dissoluta and other Andropogon-like perennial grasses. On the transition towards the lower parts of the area, this wooded savanna is replaced by Lannea humilis open shrubland on planosols with a sparse cover of annual grasses and herbs (eg. Schoenfeldia gracilis, Blepharis linariifolia). Tree cover has been steadily decreasing since the early-1970s (Esser & Lavieren 1979; Piet Wit pers. comm.), probably due to lower rainfall and the late fire regime applied. Most of Waza NP is dominated by vertisols, which are or used to be subject to flooding. Vegetation types differ in relation to the depth of present and past flooding. In general, the clayey soils above the 1979 flood line, mostly in the central part of Waza NP, are dominated by Acacia seyal shrublands. The areas formerly flooded, which used to be productive perennial grasslands, are now covered by annual herbs and grasses especially Sorghum arundinaceum. Flooding arrives as a slowly progressing sheetflow from the east and south-east, the over-flow of the Logone and its branches, from over 10 km of the Waza NP border. In the flooded (north)-eastern part of Waza NP, covering ca. 30.000 ha, perennial grasses such as Echinochloa pyramidalis, Hyparrhenia rufa and Oryza longistaminata and Vetiveria nigritana dominate and on the occasionally flooded parts further west the biannu- als Ischaemum afrum and Panicum anabaptistum. These rather monotonous grasslands are broken by (manmade) dwelling mounds, previously inhabited by fishing communities, wooded with Balanites aegyptiaca and Tamarindus indica and often with a waterhole in their vicinity. Each year, immediately after the area dries up at the end of December, many parts are burned, leaving a bare landscape. Only some local humid spots remain with green regrowth and concentrations of wildlife.

Rainfall and flooding The rainy season is from June to September, followed in the eastern parts by flooding from the Logone river and its branches from August to November. During the subsequent dry season (December-May) temperatures rise as high as 450C, and the only remaining water sources are a number of waterholes. With a long-term average annual rainfall of about 650 mm, large inter-annual differences have been recorded, as measured at nearby Ndjamena. Four rainfall periods have been identified throughout the West-Central Sahel, amongst others based on the N’djamena data-set (Gommes & Petrassi 1996). In the 1950s and early 1960s, rainfall was well above long-term-average after which a gradual decline has taken place resulting in the Sahelian droughts of (1969-) 1971-74. This ‘first Sahelian drought’ since memory of the present generation, was followed by a relatively humid period that 6 – Antelopes 149 ended with the severe 1983-85 drought, after which a gradual, yet highly fluctuating, increase in rainfall has taken place (Fig. 2.3).

Antelope population numbers were compared with developments in annual rainfall that has been averaged for the preceding three rainy seasons. This choice was based on the rate of conception of most adult female antelopes with a time-lag of at least 10-12 months, i.e. two preceding rainy seasons, standardised into the mobile average over three years (Beauvilain 1995).

Flooding of the Logone River ceased during the 1971-74 and the 1983-85 droughts. The first cession of flooding coincided with the construction of the Waza-Kousseri road that has cut several watercourses feeding the northern and northwestern part of Waza NP. It was, however, the 1979 Maga dam construction that has had a profound impact on the floodplain, both in and outside Waza NP, by reducing flooding from the Logone River and secondary water courses.

Review of 1960-2000 count methodology To facilitate comparison, we categorised counts into total, transect and local counts conducted by either terrestrial or aerial surveys (Scholte 2000a), Table 6.1. For reasons of standardisation, indicated years correspond to the preceding rainy- and flooding season and not necessarily the survey year.

• Total terrestrial counts - Waterhole counts. Two observers, a park guard and an observer, were placed from 6 am till 6 pm, during two consecutive days, at each of the 14 waterholes still holding water at the end of the dry season (April, May), The extreme heat (450C), low air humidity (<20%) and limited shade, made all but the wariest species (Giraffe and Ostrich Struthio camelus L.) drink in the observers’ presence. Numbers, time of drinking and species have been recorded. In the open floodplain mainly, the number of animals that did not drink was counted as well. Animals outside Waza NP, mainly Kob and Korrigum, were counted by car during the following days. - Counts by car and from strategic observation points focussed on the areas around waterholes where antelopes concentrated at the end of the dry season. Flizot, park director during the 1960s and early 1970s, probably also followed this method.

• Total aerial counts. For Kob, concentrated around water holes, this method has been used as reinterpretation of the 1991 transect aerial count that targeted elephants and was biased towards water holes (Tchamba & Elkan 1995), Table 6.1. The 1974 counts of Vanpraet (1977), by helicopter, lacking a methodology description, were presented as total counts as well. 150 Part II – Impact of reflooding Ecole de Faune (internal report ‘75) Ecole de Faune (internal report ‘76) Ecole de Faune (internal report ‘81) Ecole de Faune (internal report ‘89) Ecole de Faune (internal report ‘91) Floodplain outside Author 4 4 4 4 4 3 (sampling intensity) NP included? A (14%) Yes & Elkan (1995) Tchamba TO: Total count; TR: Transect count, TOP: Total count of part the area; count, TOP: Total count; TR: Transect TO: Total 2 2 1 in the NE corner of Waza NP, where Kob concentrates in the dry season. NP, in the NE corner of Waza 2 targeted Transect In an area of 16 km 4 – Background information on antelope counts in Waza NP – Background information on antelope counts in Waza Kob data have been recalculated as of total count (see text); Terrestrial; A: Aerial; T: Year Month/Aerial/Terrestrial Species Total Table 6.1 Table 197419751976 Sept. May, 1977 April1980 January all1982 Dec.1982 April1986 Dry season Kob all1987 April1988 Kob March - May TO/TR all1990 Idem Kob Kob1991 April TOPTR A (7%) April Kob TR Dec. Elephant TO Kob TR TOP NE corner only T, Kob TO A (13%) Kob TOP T No (?) NE corner only T, A (13%) TO T TOP T TOP No T NE corner only T, (1977) No Vanpraet 1 NE corner only T, 3 ? Yes ? Lavieren (1979) Esser & Van Yes Lavieren (1979) Esser & Van Thal (unpubl) Korthof (unpubl) Ecole de Faune (internal report ‘83) Njiforti (unpubl) 196019671974 General General April all all Kob TO TO TOP T T1991 NE corner only T, 19931994 Feb. - May1995 End April1996 End April all1997 End April all1997 End April No all1999 March, April No all1999 April Kob + Korrigum all TO 2000 April TR2001 End April TO April TO all End April road T, TO all all road T, Flizot (1962) TO Flizot (annual report 1968) waterhole T, all all waterhole T, waterhole T, TR waterhole T, TO TR Yes Yes No Yes TO TR Yes foot (31%) T, Yes waterhole T, foot (?) T, waterhole T, foot (31%) T, Yes Hebou (unpubl) Scholte (unpubl) Yes Bos & Bus (unpubl) Scholte (unpubl) Yes Scholte (unpubl) Yes Yes Bobo (unpubl) Bobo (unpubl) Saleh (unpubl) Saleh & Njiforti (unpubl) Saleh (unpubl) Saleh (unpubl) 6 – Antelopes 151

• Transect counts, terrestrial - Road count. While driving every accessible track in the park, observed animals were counted and their perpendicular distance to the road estimated. Observations were stratified into the four main landscapes (see above) and corrected with a detection curve (Burnham et al. 1980). Roads have, however, been designed for tourist purposes and pass the areas with water holes, where wildlife concentrates. Tracks also tend to follow the Lannea humilis shrubland, with its smooth soil surface, an area preferred by Red-fronted Gazelle, Korrigum and Roan (Esser & van Lavieren 1979; Vanpraet 1977). To avoid the waterhole bias, an alternative calculation was made, excluding observed animals at a distance of less than 1 km of the waterholes. - Counts on foot. The park and neighbouring floodplain were subdivided in 78 5-15 km long transects. Garoua Wildlife College students, assisted by park tourist guides, counted in a fixed band of 200 m on each side of the observer. Difficulty existed with the assessment of the bandwidth, especially in the floodplain grasslands with little reference points.

• Aerial Transect counts. Counts were conducted with a four-seated Cessna plane at an altitude of 100 m in fixed bands of 100 or 200 m (Esser & van Lavieren 1979; Tchamba & Elkan 1995) and with an Alouette helicopter (van Lavieren pers. comm.; Esser & van Lavieren 1979).

2 • Local counts. A 16 km triangle in the NE floodplain part of Waza NP, where kob antelope concentrates, was counted by 40-50 observers, ‘sweeping’ the area in bands of 200-300 m (van Lavieren 1976).

Kob population structure During total counts in 1997, Kob’s sex and age class was assessed to detect possible trends in population, allowing a comparison with assessments in 1972 and 1986 during total counts as well. Methodology followed Wanzie (1988).

6.3 Results

Importance of the floodplain Throughout the dry-season, almost the entire Kob population was found in the floodplain (Fig. 6.1). Korrigum and Roan moved from the Sclerocarya woodland and A.seyal shrublands into the floodplain during the second part of the dry season. A major part of the Korrigum population left the floodplain at the onset of the rains, when rainwater stagnated in small pools in the Sclerocarya woodland and neighbouring Lannea humilis shrubland and were not observed in the waterhole counts. Korrigum that remained in the floodplain continued drinking at the waterholes (Fig. 6.1), but their numbers were limited (Fig. 6.2: 1994, 1996 and 2001). 152 Part II – Impact of reflooding

100

percentage of all observations 20 Kob 0 Korrigum Roan 7 7 1 9 Red-fronted Gazelle .19 91 c .19 9 93 c 9 4 De . 1 .9 1 De l.1 y 0 eb ri a y. F M a Ap M Progressing dry season First rains

Figure 6.1 – Percentage of antelope observations in the floodplain. Details of count, see Table 6.1

At the onset of the rains, Roan also left the floodplain (Fig. 6.1) but contrary to korrigum, Roan were included in the counts when drinking at the waterholes in the A.seyal shrubland (Fig. 6.3). Only a limited, stable part of the Red-fronted Gazelle population used the floodplain (Fig. 6.1). Whereas most antelopes came from some distance drinking at the waterholes during day-time, an important part of the Kob population was already present at dawn. After drinking in the morning, Kob generally did not leave the area around the waterhole and sometimes also drank in the afternoon.

Population numbers Korrigum population numbers were strongly correlated with numbers of Roan and to a lesser extent with Bohor Reedbuck and Kob. The population developments of Bohor Reedbuck and Kob, both floodplain species, were not correlated (Figs. 6.2, 6.3; Table 6.2). Only Kob population numbers showed a positive correlation with rainfall, whereas numbers of Red-fronted Gazelle were negatively correlated with rainfall (Table 6.2).

A drop in Kob population numbers occurred in the early 1970s, coinciding with the 1969-74 drought and the lack of flooding (Fig. 6.2; Table 6.2). In 1971, 1972 and 1982, Kob losses, extrapolated from respectively 198, an unknown and 400 6 – Antelopes 153

30000

KOB KORRIGUM

Number 25000 1

20000 overestimation

underestimation 2 15000

10000

67 5000 5 343 4 4 0 '60 '67 '74 '76 '77 '82 '86 '87 '91 '93 '94 '95 '96 '97 '99 '01

drought dam drought reflooding Year

1 Probably biased by experiences of the early 1960s, see text 2 A concentration of Kob outside the survey area (i.e. NE of the park) was observed (van Lavieren, pers. comm.) 3 Targeting Kob around waterholes only (Table 6.1) 4 Early rains caused Korrigum leaving the floodplain, no longer drinking at waterholes (see text) 5 Estimations by road counts were 10730 Kob and 2800 Korrigum, corrected for waterhole bias (see methods) 2300 and 898 respectively, 6 Estimations of terrestrial transects were 7665 kob and 2138 Korrigum 7 idem, 9548 Kob and 4284 Korrigum; in 2000, 8929 Kob and 1512 Korrigum.

Figure 6.2 – Estimated numbers of Kob and Korrigum, 1960-2001

carcasses found, were estimated at about 1000 (Table 6.3; Vanpraet 1977; Thal unpubl.). In the same period, Korrigum and Bohor Reedbuck numbers seemed to have dropped more dramatically (Fig. 6.2, 6.3). Yet in 1974, Flizot only remarked ‘Korrigum seems to be less abundant and seems to have migrated to the north’ and ‘few Bohor Reedbuck have been observed, contrasting the numbers before the droughts’.

The Waterbuck (Kobus elypsiprymmus Ogilby) population was estimated at more than 200 throughout the 1960s, but not detected by the aerial surveys of 1976 and 1977 (Esser & Van Lavieren 1979). Waterbuck was no longer observed from the mid-1980s onwards till one and two individuals were once seen again in the floodplain part of Waza National Park in 1998 and 2003 respectively. Bushbuck and Red-flanked Duiker (Cephalophus rufilatus Gray) still present in the 1970s (Wit 154 Part II – Impact of reflooding

4000 4000 2000 ROAN RED-FRONTED GAZELLE BOHOR REEDBUCK 1 2 Number 1500 overestimation

underestimation 4 4 1000

9 500 7 3 3 3 6 5 0 '60 '67 '74 '76 '77 '91 '93 '94 '95 '96 '97 '99 '01 drought dam / drought reflooding Year

1 Presumed overestimation, explanation not know. 2 Probably based on experiences of the early 1960s, see text 3 Aerial counts, underestimated smaller antelope species, see text 4 Waterhole count, overestimating Roan, probably drinking twice 5 Waterhole count, underestimation for unknown reasons 6 Waterhole count, underestimation because of early rains (see also Fig.6.3 and Text) 7 Estimations by road counts were 1446 Roan and 1574 Gazelles, corrected for waterhole bias (see methods) 935 and 1036 respectively . 8 Estimations of terrestrial transects were 1331 Roan and 525 Gazelles 9 idem, 3868 Roan and 593 Gazelles; in 2000, 496 Roan and 228 Gazelles.

Figure 6.3 – Estimated numbers of Roan, Red-fronted Gazelle and Bohor Reedbuck, 1960-2001.

pers. comm.) have no longer been observed in Waza NP. Common Duiker still occurs in limited numbers.

The 1977-1982 drop in Kob population coincided with the construction of the Maga dam and subsequent lack of flooding (Fig. 6.2; Table 6.2). The further drop from 1982 to 1986 in population numbers of Kob and possibly Korrigum populations were accentuated by the 1985-86 drought (Figs. 6.2, 6.3; Table 6.2). Observed Kob densities in the area with the largest concentrations dropped from 356 (1974) and 210 km-2 (1975) to 131 (1980) and 91 km-2 (1982) to 45 (1988) and 23 kob km-2 (1990). Population numbers of Kob and Korrigum recovered between 1986-1987 and 1991-1993 (Fig. 6.2). This recovery continued from 1994 till 1997, possibly enhanced by the increased (re-)flooding, but numbers stabilised subsequently. 6 – Antelopes 155

Table 6.2 – Correlation coefficients amongst antelope populations and key environmental parameters

Kob Korrigum Roan Bohor Red-fronted Rainfall3 Reedbuck Gazelle

Kob r1 1 p Korrigum r 0.732 1 p 0.003 Roan r 0.66 0.99 1 p 0.013 < 0.001 Bohor Reedbuck r 0.47 0.81 0.79 1 p 0.15 0.004 0.004 Red-fronted Gazelle r -0.21 0.14 0.26 0.26 1 p 0.54 0.70 0.44 0.47 Rainfall3 r 0.54 0.48 0.39 0.49 -0.64 1 p 0.032 0.084 0.19 0.13 0.032 Presence of Maga dam S4 -0.80 -0.32 -0.17 -0.24 0.32 -0.45 p < 0.001 0.27 0.58 0.48 0.33 0.078

1Pearson Correlation Coefficient (SPSS 1999); 2Significant (p < 0.05) in bold; 3Cumulative rainfall of the preceding three rainy seasons; 4Spearmann Correlation Coefficient (SPSS 1999).

Kob population structure In 1986, half of the Kob population was observed outside Waza NP and had a sex ratio (male/female) of 0.47, that was 0.60, 0.70 in the East and North-East of Waza NP respectively; overall sex ratio was 0.53 (Table 6.3). The number of young per female varied from 0.31 (E.Waza NP) to 0.39 (outside) to 0.50 (NE Waza NP). Amongst carcasses found, juvenile Kobs were over-represented (Table 6.3), due to their vulnerability to muddy water and warble flies (van Lavieren pers. comm.). In February and April 1997, sex ratio was 0.51 and 0.53 and young/female ratio 1.04 and 1.18 respectively (Table 6.3).

6.4 Discussion

Count biases To some extent differences in counting methods masked changing environmental conditions. This held especially for the 150-200% higher Kob numbers estimated in transect than in total counts (Fig. 6.2), due to their congregation around water holes in all but the rainy and flooding season (Esser & van Lavieren 1979), see also Fischer & Lindenmair (2001a). The terrestrial transect counts resulted in deviating results for other species as well (Fig. 6.3), attributed to the difficulty of 156 Part II – Impact of reflooding 6 6 6 6 a 508 b 5 5 177 (35%)a 176 (35%) 4 carcasses carcasses carcasses 198 Adult male Subadult male (Sub) Adult Juvenile Number 3 = 0.045 (chi-square test). 1 p < 0.0001, except g-h: p Waza NPWaza female Part of Month Author – Population structure of Kob 2 Classification following Wanzie 1988. Classification following Wanzie corresponds to the preceeding rainy- and flooding season not necessarily survey year. Year 6.1. See also Table letter: Different Includes some subadult males, the 1972 counts can only be mutually compared. Some double-counting (50%) has taken place (compare with Fig. 6.2). 1972 North East April EFG 1973 16 (8%)b 77 (39%)b 105 (53%) 1972 North East April 155 (31%)a Table 6.3 Table Year 1986 East1986 NP East of Waza 1986 Febr. All 1997 Febr. Park and 1997 Korthof (unpubl.)1 surroundings2 24 (15%)d Febr. Febr.3 April4 Hebou (unpubl.) 28 (17%)d5 6 96 (7%)e 551 (12%)g 176 (53%)d 234 (18%)e 334 (7%)g 54 (16%)d 704 (54%)e 1741 (39%)g 230 (8%)f 335 563 (12%)h 1817 (41%)g 273 (21%)e 524 (19%)f 366 (8%)h 4472 1307 1414 (50%)f 1761 (36%)h 592 (21%)f 2082 (43%)h 4832 2813 1986 North East Febr. 110 (9%)c 262 (22%)c 534 (46%)c 265 (23%)c 1171 6 – Antelopes 157 estimating the bandwidth by inexperienced observers, of which their large number (>40) and frequent changes did not allow any correction. The aerial surveys were biased towards the larger species neglecting Red-fronted Gazelle and Bohor Reedbuck (Van Esser & Lavieren 1979; Tchamba & Elkan 1995; Dejace et al. 2000). During the waterhole total counts, frequent, i.e. more than once a day drinking, was observed of Kob, leading to a potential overestimation of its population numbers. None of the other antelope species showed comparable drinking behaviour. Other imaginable biases of the waterhole counts could only have led to an underestimation of population numbers. Red-fronted Gazelle is less dependant on daily drinking than the other antelope species in Waza NP (Estes 1992). It is further likely that some antelopes drank before or after the observers had been installed.

In addition to these methodological biases, it cannot be excluded that ‘counter psy- chology’ has influenced survey results. Flizot probably overestimated the numbers of Korrigum, Roan and Bohor Reedbuck in 1967, neglecting a decline that is expected to have taken place after the above-average rainfall and flooding in the 1950s and early 1960s when he was frequently in the field. Flizot, as well as his succes- sors, further could have had an interest in showing a healthy animal population in the park they managed.

Kob Despite the above-mentioned count biases, major impacts of (man-made) droughts have been detected, especially on Kob. The Kob population dynamics show the intrinsic link between stress from periods of low rainfall and man-made droughts, induced by the Maga dam construction (Fig. 6.2; Table 6.2). This also holds for the rehabilitation process, as the 1994 reflooding started when the Kob population had already partly recovered from the mid-1980s drought.

Kob’s sex ratio (male/female) in Comoé NP (Ivory Coast) was only 0.34, which Fischer & Lindsenmair (2002) attributed to high poaching pressure. Here, Geer- ling & Bokdam (1971) observed a sex ratio of 0.74 in a time when poaching pressure was much lower. The somewhat lower Kob sex ratio outside Waza NP in 1986 (0.47), compared to inside Waza NP (0.60), may possibly be attributed to poaching. Sex ratio of the entire Kob population in 1986 and 1997 was comparable (0.51-0.53), suggesting no trend in poaching. Predation pressure has remained relatively constant during the last decades (Bauer 2003) and does not explain observed differences in Kob nor any of the other antelope species.

Kob year classes of Fischer & Lindsenmair (2002) did not entirely match Wanzie (1988), yet with the broad class ‘young’ (calves and juveniles) this source of differences was limited. Fischer & Lindsenmair (2002) observed a low 0.49 ± 0.05 young per adult female, which they attributed to the high poaching pressure, as argued for the above-mentioned low sex ratio. The comparably low young per 158 Part II – Impact of reflooding

adult female ratio observed in Waza in 1986 (0.42) can be attributed to drought induced food stress as the juvenile percentage in 1997, preceded by three years of almost average rainfall, was twice as high (Table 6.3). Although some differences in interpretation cannot be excluded, drought seems to have affected especially the youngest and eldest age classes. Although the 1985-1986 drought had limited the reproduction output, it does not seem to have hit the (future) reproduction capacity, as confirmed in the Kob population increase in subsequent years (Fig. 6.2).

Korrigum, Roan and Bohor Reedbuck The poorly documented late 1960s-early 1970s drop of Korrigum, Roan and Bohor Reedbuck populations (Figs. 6.2, 6.3) remains difficult to explain. Contrary to common belief, presented results show that lack of flooding due to the Maga dam construction, that took place only after 1979, was not the reason for their decline (Figs. 6.2, 6.3; Table 6.2). The lack of flooding since 1979 may, however, have been a constraint for their full recovery. Elsewhere in West-Central Africa these antelopes species have also been under increasing pressure, due to poaching, competition and disease transmission by livestock (Esser & van Lavieren 1979; Sayer 1982; East 1999; Fischer & Lindsenmair 2001b; Sinsin et al. 2002). With a continuing drop in the Arly-Singou-Pendjari area (Sinsin et al. 2002), Waza NP holds the only remaining viable Korrigum population, probably due to the variation in protected habitat, including both floodplain and upland savanna (Sayer 1982). Although the floodplain habitat has been improved with the recent reflooding (Scholte et al. 2000b), livestock competition and potential disease transmission in the immediate surroundings of the park, have also increased (Scholte 2003a).

Extinction and human impact The disappearance of Waterbuck that used to occur in the floodplain, yet in small numbers since the early 1970s, appears to be linked to the hydrological changes. Its timid come-back may possibly be associated with the reflooding from 1994 onwards. In addition to these changes in the floodplain, Vanpraet (1976) stressed the negative impact of habitat changes in other parts of Waza NP such as increased fire in the preferred Lannea humilis habitat, a practice that, though addressed in its management plan remains difficult to change (Scholte 2000b). The extinct Bushbuck and Red-flanked Duiker used to occur in the wooded parts of Waza NP that experienced a loss of tree cover due to lack of flooding from upland water courses, the low rainfall in the 1970s and early 1980s (Fig. 6.2) as well as an increasing impact of Elephant (Tchamba 1996).

Direct human impact on these extinctions can not be excluded as reported for neighbouring Kalamaloué NP where the once common Waterbuck disappeared in the same period, due to human encroachment (Scholte 2003a). Red-fronted Ga- zelle, with a predominant Sahelian distribution (East 1999; Scholte & Hashim 2005), seems to have increased in both national parks. Few direct signs of poach- 6 – Antelopes 159 ing have, however, been reported in and around Waza NP (Tchamba & Elkan 1995; Scholte et al. 1999b). This may be due to the difficult relations that existed for most of the time between park authorities and local communities and the inaccessibility of the area during the rainy and flooding season (Scholte et al. 1999b). Kob and Korrigum populations that leave Waza NP for considerable part of the year (Flizot 1974; Esser & van Lavieren 1979) are especially vulnerable to poaching.

It has been difficult to assess with one single method the population size of all antelope species, each with a particular distribution and habits. A combination of standardised counts, such as annual waterhole counts and occasional aerial counts, as proposed in the Waza management plan, seems appropriate and cost effective. This will allow monitoring the antelope populations after the recently increased confidence between park authorities and local communities and allow an assessment of the impact of more extensive reflooding (Scholte 2003a).

7 Pastoralist Responses to Floodplain Rehabilitation in North Cameroon

Paul Scholte, Saïdou Kari, Mark Moritz and Herbert Prins

Human Ecology, in press

Abstract This chapter examines the responses of mobile pastoralists to a floodplain rehabilitation program in North Cameroon. From 1993 to 1999, we measured 2 changes in number of camps and herds, and the time they spent in the 600 km of the Logone floodplain that was reflooded in 1994. The first year, few pastoralists anticipated the reflooding or its impact, and the increase in grazing intensity was caused by a prolonged stay of pastoralists who already used the area for transit. The following three years showed a sharp increase in the number of camps and herds, which stabilised from 1997 onwards. Overall, grazing intensity increased three-fold, following the gradually recovering perennial grasslands, with no signs of over-exploitation of the area. These developments closely match the Ideal Preemptive Distribution model. We also examined how reflooding affected pastoral incursions in Waza National Park located in the floodplain.

163

7.1 Introduction

Mobile pastoralists are remarkably well adapted to the drylands of Africa, much better than sedentary pastoralists or ranchers (Breman & Wit 1983; Niamir-Fuller & Turner 1999). Floodplains play a key role in the yearly migration of many pastoralists, due to the availability of nutritious grass regrowth and surface water far into the dry season, when surrounding grazing lands have dried up (Hiernaux & Diarra 1983). However, human pressure on these floodplains is increasing and re- stricting pastoralists’ access to them. One of the more devastating impacts on floodplains is the construction of dams for electricity and irrigation purposes (Drijver & Marchand 1986). Increasing awareness of the importance of wetlands for rural economies in semi-arid Africa has led to a series of initiatives to reduce the impacts of existing dams (Acreman & Hollis 1996). Most initiatives aim at improving water management, so that ‘excess’ water can be released and restore former floodplain functions, including that of dry season pastures for transhumant pastoralists. In several places in Africa experiments with such releases have started. But despite the importance of African floodplains for mobile pastoralists, there had not yet been an assessment of their responses to floodplain rehabilitation. This paper examines the responses of mobile pastoralists to the reflooding of the Logone floodplain, focusing on changes in transhumance patterns and grazing intensity from 1993 until 1999. It offered a unique chance to study pastoralists’ reactions to improving grazing conditions in a Sahelian environment.

In 1979, the Waza-Logone area in the Far North Province of Cameroon met the fate of other African floodplains, when a dam and embankment along the Logone River were constructed forming a reservoir for a parastatal irrigation scheme (Fig. 7.1). In combination with lower than average rainfall during the last two decades (Beauvilain 1995), the dams reduced flooding in an area of about 1,500 km2, which included Waza National Park, an important refuge for wildlife such as elephant, lion, antelopes and waterbirds (Scholte et al. 1996a). Annual grasses invaded previously productive perennial grasslands, limiting regrowth in the dry season and reducing the floodplain’s carrying capacity for wildlife and livestock (Scholte et al. 1996a). Studies that documented the detrimental effects of the dam on ecology and local economies (Drijver & Marchand 1986) were the impetus for a reflooding program, which took shape in the Waza-Logone Project. An appraisal study showed the potential for reflooding (Wesseling et al. 1996), which was supported by local people, authorities and wildlife managers. Reflooding was started in May 1994 with the re-opening of a watercourse that had been closed off by the embankment along the Logone River, starting the recovery of perennial grassland vegetation (Scholte et al. 2000b). 164 Part II – Impact of reflooding

Fig. 7.1 – The Far North Province with the Waza-Logone area 7 – Pastoralists responses 165

One of the main goals of the reflooding was to stop the incursions of pastoralists into Waza National Park in order to reduce the competition between wildlife and livestock for dry season grazing (Scholte 2003a). Livestock is not allowed in national parks and this has been the subject of numerous confrontations between park authorities and pastoralists (Scholte et al. 1996b, 1999b). The project anticipated that the floodplain rehabilitation would motivate pastoralists to find grazing lands adjacent to the park, where the reflooding was expected to have its main impact.

The reflooding feasibility study predicted that the productivity of the floodplain grasslands would increase two- or three fold (Wesseling et al. 1996). One of the questions was how mobile pastoralists would respond to this increase in rangeland productivity and whether a potential influx of ‘new’ pastoralists would jeop- ardize the projected rehabilitation of the Logone floodplain. Another question was whether pastoralists would indeed refrain from entering Waza National Park with increasing productivity of the floodplain.

Predicting pastoralists’ responses Visions from pastoralists, researchers, and authorities diverged on the reactions of mobile pastoralists to the floodplain rehabilitation and its subsequent impact on Waza National Park. The visions can be grouped in three different scenarios:

1 Overshoot scenario The first scenario assumed a ‘free-for-all’ open access situation in which pastoralists would flock to the improved grazing opportunities. This would result in grazing intensities largely exceeding the improved carrying capacity, i.e., an ‘overshoot’ of cattle. The fear was that this would lead to overgrazing and conflicts with other pastoralists, thereby undoing the rehabilitation of the floodplain. In this scenario, pastoralists would most likely continue to take their cattle into Waza National Park because there would be no alternative grazing opportunities.

2 Territorial scenario The second scenario assumed that pastoralists that used the area before reflooding would somehow defend their grazing lands, i.e., display territorial behavior (Casimir 1992). Territoriality among mobile pastoralists has been observed in neighboring Chad (pers. obs.) and in the Niger floodplain in Mali (Legrosse 1999; Turner 1999). Although there seemed to be a system in the floodplain in which pastoralists had customary rights to specific campsites and surrounding pastures, it was unclear whether and how pastoralists would defend these rights if there were an influx of ‘newcomers’. In this scenario, there would be most likely a slow increase in grazing intensity and a low exploitation of perennial grass cover in the rehabilitated floodplain. Potentially this would allow 166 Part II – Impact of reflooding

pastoralists with claims in the floodplain to return to their old campsites and refrain from taking their cattle inside Waza National Park.

3 Ideal free distribution scenario The third scenario predicted that any improvement in grazing land condition, i.e. perennial grass cover, would lead to a corresponding increase in the number of cattle. The assumption underlying this scenario was that pastoralists would adjust the number of herds in the reflooded area and/or the time they spent there to make optimal use of the available resources. The result would be a gradual increase in grazing intensity following a similar increase in perennial grass cover. In this scenario, grazing resources outside the park would gradually come available to pastoralists that were taking their cattle into the park.

The data presented in this article allows for an evaluation of which scenario best matches pastoralist responses to reflooding. We will also consider how other factors, such as insecurity, cattle theft and predation, and conflicts with fishers and farmers affected transhumance patterns.

7.2 Background

The Waza-Logone floodplain The Waza-Logone floodplain receives a mean annual rainfall of c. 650 mm with between-year fluctuations of up to 50%, a typical Sahelo-Sudanian climate (Beau- vilain 1995). The rainy season is from June to September. In parts of the area, the Logone River and its branches flood during August to November. In April and May, temperatures rise as high as 45°C. In areas subject to annual flooding, perennial grasses such as Echinochloa pyramidalis, Oryza longistaminata and, on the slightly elevated parts, Vetiveria nigritana and Hyparrhenia rufa dominate. When the floods recede at the end of December, most parts are burned by pastoralists and hunters, leaving a bare landscape and only some humid spots remain with vivid green vegetation and concentrations of fishermen, livestock and birdlife.

Until the Maga dam construction in 1979, the study area, the ‘reflooded area’ located in the southern floodplain, used to be flooded and was composed of productive perennial grasslands (Fig. 7.1). When flooding ceased, perennial grasses were replaced by annual grasses, especially the reed-like unpalatable Sorghum arundinaceum. The reflooding in 1994 brought back the annual floods in an area of about 180 km2 and raised the water level in 600 km2. Following the reflooding an annual 7% increase in perennial grass cover was measured (Scholte et al. 2000b). The improvements were dramatic: Sorghum arundinaceum has disappeared and one can again see for miles and miles fertile perennial grasslands. 7 – Pastoralists responses 167

This period coincided with average rainfall and flooding conditions, and without droughts like those that plagued the area in the 1970s and 1980s (Beauvilain 1995).

Pastoralists in the Waza-Logone floodplain The Logone floodplain historically has been a Key Resource Area for pastoralists in the Lake Chad Basin (Scholte & Brouwer 2005). When rangelands in the surrounding plains and mountains are drying up and sources of surface water are disappearing, pastoralists trek to the Logone floodplain to find forage and easily accessible surface water for their animals. Each year, thousands of pastoralists with about 200,000 cattle from Cameroon, Chad, Nigeria, and Niger enter the floodplain for a period of six to eight months (Beauvilain 1981).1

The mobile pastoralists that use the floodplain pastures belong to two ethnic groups: (Shuwa) Arabs (Braukamper 1996) and FulBe (e.g. Azarya et al. 1999), (Table 7.1). The Arabs do not sell milk, which they use for household consumption or leave for the calves. This is one of the reasons why their herds grow at a higher rate than those of FulBe pastoralists. Other reasons are their involvement in livestock trade, the profits of which are invested in cattle, and the fact that they care for entrusted animals from relatives in Nigeria. The Arab pastoralists that use the reflooded area are nomadic and practice transhumance from the Diamaré plains, south of the floodplain, where also most FulBe pastoralists come from (Fig. 7.1).

The FulBe can also be divided in nomadic and agro-pastoral groups. This distinction refers to the fact that the former do not practice agriculture and are without permanent settlements. Agro-pastoral FulBe, on the other hand, have families that live in villages where they cultivate. Agro-pastoral herds may be permanently on transhumance and come back to the village only once or twice a year for a few weeks. There are two distinct groups of agro-pastoral FulBe that practice transhumance to the reflooded area. One group comes from the area south of Waza National Park and is referred to as the FulBe Ngara’en, which is the name of the ruling clan of the towns of Pétté and Fadaré (Fig. 7.1). Their transhumance takes them along the borders of Waza National Park, and not surprisingly, the FulBe Ngara’en are the ones responsible for incursions into the park. The other agro- pastoral FulBe group comes from further south and is often referred to as the FulBe Yillaga’en, which is the ruling clan in the towns of Mindif and Moulvou- daye (Fig. 7.1). Most agro-pastoralists from both groups are descendents of FulBe that conquered the area two centuries ago (Mohammadou 1988).

1 Use of floodplains in the Chad basin is facilitated by international agreements (CIRAD-CTA, 1996). In principle, pastoralists have to pay a small fee per head of cattle and show certificates of herd vacci- nations when they cross the border, but in practice border controls are not very strict. 168 Part II – Impact of reflooding

Table 7.1 – Characteristics of Pastoral Groups at the end of the Study Period, 1999

Nomadic Pastoralists Agro-Pastoralists

FulBe FulBe FulBe Arabs FulBe FulBe Adanko’en Alijam’en Mare’en Ngara’en Yillaga’en

1 Number of households 25 89 114 135 89 (or herding groups) Number of households per camp 2.5 9.8 3.8 9.6 NA NA Number of camps 10 9 30 14 26 16 Number of herds per camp 7.2 9.4 8.4 13.9 10.9 12.0 Number of herds per household 2.9 1 2.2 1.5 NA NA Total number of herds 72 85 252 195 283 192 Estimated number of cattle 4,593 5,338 21,344 23,653 23,829 16,320 Presence in the floodplain 20 30 60 30 >100 >100 (in years)

1 Unlike the nomadic groups, the FulBe Ngara’en and FulBe Yillaga’en are mostly herders on transhumance without their families. The term household cannot be used for this group. It is better to think of these units as herding groups in which a number of herders of one village cooperate. Large herding groups often constitute one camp; smaller herding groups camp together with herding groups from neighboring villages. The herds in these herding groups, in turn, are often composed of animals from multiple households in the village.

The nomadic FulBe group can be divided into three distinct sub-ethnic groups Adanko’en, Alijam’en and Mare’en.2 The majority of nomadic FulBe originally came from Nigeria, but have been in Cameroon for the last twenty to sixty years. They have Cameroonian identity cards and identify themselves as Cameroonians. Their kin network extends across all borders, which reflects their readiness to move, responding to ecological as well as socio-political conditions.

Although cattle account for more than 90% of the total Tropical Livestock Units in the floodplain, there are also FulBe Uuda’en shepherds who transhumance from Nigeria and Niger to the Logone floodplain. But since they do not exploit the reflooded area we have excluded them from this study. We also excluded the sedentary, village-based Musgum, who possess less than 10% of the total number of cattle in the floodplain.

2 A sub-ethnic group refers to a group consisting of different clan fractions that have in common a migratory history and certain cultural traits (Burnham 1996). Other, numerically smaller nomadic FulBe groups in the Logone floodplain have been grouped with one of the three main nomadic FulBe groups with whom they associate. 7 – Pastoralists responses 169

Transhumance patterns One can distinguish two phases of pastoral exploitation of the floodplain. During the first phase from November through January, after the recession of flooding, cattle feed on young grass shoots. At the beginning of this phase, when pastoralists enter the floodplain, they often split their herds. Older and lactating animals stay with the families, while young herders take the stronger, healthy animals on a separate transhumance, called luci, following the retreat of the water in search of the best pastures. During this period camps and herds move frequently. In January-February, the herds are re-united again and pastoralists often settle in one or two campsites for the rest of the dry season. During this second phase, cattle feed on the regrowth of perennial grasses that result from fires. At the onset of the rainy season in May-June, pastoralists direct their animals to those places where the first rains have fallen and leave the floodplain till the following dry season.

One can also distinguish different ways in how the reflooded area fit into the transhumance of different pastoral groups before it was reflooded in 1994. A first group of pastoralists, mainly FulBe Alijam’en and Arabs, used the reflooded area as a transition zone from the rainy season pastures in the south to their dry season pastures further north and east (Fig. 7.1). A second group, FulBe Mare’en and FulBe Adanko’en, used the reflooded area only in the cold dry season (November- January), before they went to their dry season pastures at the borders of Lake Maga, further south. A third group of pastoralists spent the entire dry season in the reflooded area. This last group consisted predominantly of the agro-pastoral FulBe Ngara’en and FulBe Yillaga’en, and the FulBe Mare’en. We will see below that most pastoral groups continued these transhumance patterns after reflooding, while spending increasing time in the reflooded area.

7.3 Methods

Early 1994, before the reflooding, we visited all pastoral camps in the southern part of the floodplain (Fig. 7.1). The main objective was to start a dialogue with a wary group that had suffered from previous governmental and project interventions (Moritz et al. 2002; Scholte et al. 1996b). Most of the time was spent on discussing ongoing developments in the floodplain and pastoralists’ concerns, in particular insecurity. During the visits in following years, we obtained information on ‘new’ pastoralists in the study area, which were included in our annual monitoring of pastoralists’ responses to the reflooding. We also obtained information on pastoralists that we missed in our earlier campaigns. Generally, these ‘missing’ pastoralists used the reflooded area only as a transition zone from rainy season pastures in the south to dry season pastures further north outside our study area. At the end of each campaign, we organized a session in a centrally located market town to discuss results and upcoming issues, such as further reflooding initiatives. 170 Part II – Impact of reflooding

Assessing grazing intensity The visits in 1994 and subsequent years also served to establish a database of all pastoral groups that spent time in the reflooded area (Table 7.1). Groups were entered in the database by the name of the camp’s leader and for each group we listed the campsites visited in the floodplain and the number of weeks spent there. Early in our work, we compiled a list with all campsites and marked them on a topo- graphic map. We used this list to record transhumance patterns and determine the number of weeks that pastoralists spent in the reflooded area. The interview method was an effective and reliable way to assess changes in grazing intensities; pastoralists could recall without difficulty how many days they had camped in each site over the last year.3 In subsequent campaigns, we also collected data on the number of herds and herd size (i.e. the numbers of cattle per herd) in informal discussions with group leaders.4 The trust we had built up through the projects’ commitment to resolving pastoralists’ primary concern of insecurity (Scholte et al. 1996b), allowed the estimation of these otherwise very sensitive topics. We checked self-reported herd numbers and size with direct counts during overnight stays, crosschecking with pastoralists in other camps, and data from vaccination campaigns (Table 7.2).5 Because there were significant differences between self- reported herd size from FulBe Mare’en and Arabs and our own direct counts and vaccination campaigns, we have disregarded the former. In this study of changes in grazing intensity, we consider herd size to be a constant value, although grassland productivity increased after the reflooding and assumedly led to greater repro- ductive rates and thus larger herds. Moritz (2003), for example, found that nomadic herds in the Logone floodplain grew on average 4.5% per year, slightly higher than the 3.4% postulated for a ‘normal’ herd by Djahl & Hjort (1976). Increased repro- ductive rates and herd growth is reflected in an increase in the number of herds per pastoral group when pastoralists ‘split’ their herds.

We have operationalised grazing intensity as the number of herds multiplied by the number of weeks each camp spent in the reflooded zone and that multiplied by the number of cattle per herd of the concerning pastoral group. The resulting grazing intensity is expressed as cattle densities, i.e., the number of cattle per km2 aver-

3 Pastoralists used moon cycles and weekly market days as reference points to recall how long they had camped in particular sites. We cross-checked this information with other pastoralists. 4 We have considered herds as management units, i.e., the animals that graze together during day- time under the guidance of a herder, and not as property units (Dahl and Hjort 1976: 134). 5 Initial analysis using stem-and leaf plots in SPSS (1999) showed that herd size data followed a normal distribution. Herd size data from different sources have subsequently been analyzed with t-tests, variances of herd sizes with Levene’s Test for equality of variances (SPSS 1999) to enable possible lumping. Subsequently, herd size data from different pastoral groups were lumped if results of ANOVA and Tamhane Post Hoc tests (SPSS 1999) allowed for it. 7 – Pastoralists responses 171

Table 7.2 – Average Herd Size by Source of Information

Self-report Vaccination Direct count Average herd size used in records this study (vaccination records and direct counts lumped)

Nomadic Pastoralists FulBe Adanko’en 65.8 a1 63.8 a – 63.8i (25) (5) FulBe Alijam’en 64.4 b2 64.5 b 50.0 b 62.8i (36) (15) (2) FulBe Mare’en 75.3 e 84.0 f 92.3 f 84.7k (72) (38) (3) Arabs 98.7 c 120.8 d 122.3d 121.3j (15) (13) (6)

Agro-Pastoralists FulBe Ngara’en 80.8 g 84.2 g – 84.2k (62) (20) FulBe Yillaga’en 79.1 h 85.4 h 82.7 h 85.0k (51) (29) (6)

Number of reports or counts is given in parentheses. 1 2 same letter: p (means) > 0.05 and p (variance) > 0.05; same letter in italics: p (means) > 0.05 (t-test, SPSS 1999) and p (variance) < 0.05 (Levene’s Test for equality of variances, SPSS 1999).

aged over the 6 months of the dry season. This also allows us to compare our data with livestock surveys in other floodplains.6

7.4 Results

Changes in transhumance patterns In individual and group discussions, pastoralists invariably expressed their appreciation of the reflooding initiatives. At a meeting in 1997 with about twenty representatives of different pastoral groups, a proposed opening of a second watercourse, in addition to the 1994 reflooding, was applauded with ‘the more water the better’. How pastoralists responded in terms of changing transhumance patterns will be discussed below.

6 The 27 to 69 cattle km2 increase in the reflooded area in the Logone floodplain falls within the range of aerial surveys of other African floodplains, which assessed average cattle densities from 25 to 80 cattle per km2 (Scholte & Brouwer 2005). Obviously, care should be taken to compare surveys of different scales and methodologies. 172 Part II – Impact of reflooding

Changes in the number of pastoral camps In the first year of reflooding, few pastoralists had heard about the reflooding or had anticipated its impact, reflected by the limited increase in the number of pastoral camps (Fig. 7.2). When the positive impact of the reflooding became apparent, it was subject of much discussion on market places and other pastoralists’ fora. The number of pastoral camps increased dramatically in the second year, 1995, and continued until 1997 after which it stabilised (Fig. 7.2). Unlike the other pastoral groups, the number of Arabs camps did not increase, although some ‘new- comer’ Arabs pastoralists are said to have joined the existing camps, contributing to the highest ratio of herds per camp (Table 7.1).

120

Y(total) = 119 - 64.94/ t ( p = 0.003) 100 t = years after reflooding

80 Fulbe Mare'en Arabs 60 Fulbe Ngara'en Fulbe Yillaga'en Fulbe Alijam'en Fulbe Adanko'en 40 number of camps

0 1993 1994 1995 1996 1997 1998 1999 reflooding

Figure 7.2 – Development of number of camps in the reflooded area, 1993-1999.

Y R2 p Total 119.2-64.9/t 0.96 0.003 FulBe Adanko’en 12.7 - 11.4/t 0.94 0.007 FulBe Alijam’en LnY = 2.54 - 1.06/t 0.87 0.020 Arabs LnY = 2.64-0.14/t 0.52 0.169 FulBe Mare’en 35.1-15/t 0.99 0.000 FulBe Ngara’en LnY = 3.44-1.36/t 0.93 0.008 FulBe Yillaga’en 18.9-10.0/t 0.86 0.023

Bold: table-wide significant (α = 0.05). Regression equations have been determined between the number of camps and grazing intensity with increasing time, using best-fitted curve estimation (SPSS, 1999). Significance levels have been adjusted on a table-wide level, using the Bonferroni test for multiple tests at a 0.05 significance level (Rice 1989). 7 – Pastoralists responses 173

1200

Y(total)= 261ln t +601 (p =0.002) 1000 t = years after reflooding

800 Fulbe Mare'en Arabs Fulbe Ngara'en 600 Fulbe Yillaga'en Fulbe Alijam'en Fulbe Adanko'en number of herds 400

200

0 1994 1995 1996 1997 1998 1999 reflooding

Figure 7.3 – Development of number of herds in the reflooded area, 1993-1999.

Y R2 p Total Y=261.3 ln t + 601.4 0.97 0.002 FulBe Adanko’en LnY = 5.02- 3.12/t 0.96 0.004 FulBe Alijam’en 2.49X + 68.2 0.69 0.082 Arabs 6.73X + 152.9 0.74 0.061 FulBe Mare’en LnY = 5.56- 0.55/t 0.93 0.009 FulBe Ngara’en LnY = 5.97 - 1.71/t 0.93 0.008 FulBe Ngara’en 13.15X + 111.3 0.99 0.000

Bold: table-wide significant (α = 0.05), see also remark Fig. 7.2

Changes in the number of herds Initially, the number of herds followed a relatively low increase rate as newcomers had relatively few herds compared to the increase in the number of camps (Fig. 7.3). The number of herds continued to increase until the last monitoring year. Part of the increase is due to normal herd growth and subsequent splitting of herds that were already present. The 4.5% annual growth rate of nomadic herds (Moritz 2003) would translate in an increase of 30% in the number of cattle from 1993 to 1999. If we assume a similar increase in the number of herds, then approximately one-third of the increase in the number of herds in the study area during that period is due to growth of existing herds, while two-thirds is due to the arrival of new herds.7

7 This assumption seems reasonable given the large number of animals and herds in our study, respectively 95,077 and 1,079 in 1999. 174 Part II – Impact of reflooding

Changes in grazing duration We also measured how long pastoralists stayed in the reflooded zone, which we refer here to as grazing duration, i.e. the number of weeks pastoralists stayed in the study area. Before reflooding in 1994, FulBe Mare’en stayed in the reflooded zone almost twice as long as other groups. After reflooding the difference between the FulBe Mare’en and the other pastoral groups has become less distinct as contrary to most other pastoral groups, FulBe Mare’en stayed only slightly longer than before. From 1993 until 1999, FulBe Alijam’en and Arabs groups have stayed increasingly longer in the reflooded zone on their yearly migration to pastures further north and east into Chad.8

70 Y(total) = 7.1 t + 17.9 (R2=0.92 p =0.003) t(1993) = 1 )

-2 60 ng i 50 graz intensity (cattle km 40

Total 10 Fulbe Mare'en Arabs Fulbe Ngara'en 0 Fulbe Yillaga'en Fulbe Alijam'en 1993 1994 1995 Fulbe Adanko'en 1996 1997 reflooding 1998 1999

Figure 7.4 – Development of grazing intensities in the reflooded area, 1993-1999.

Y R2 p Total 7.1X + 17.9 0.91 0.003 Fulbe Adanko’en 0.38X -0.017 0.54 0.096 Fulbe Alijam’en 0.59X - 0.40 0.93 0.002 Arab 1.7X + 2.16 0.88 0.006 Fulbe Mare’en LnY = 2.9 - 0.63/t 0.67 0.046 Fulbe Ngara’en 2.33X + 3.4 0.71 0.035 Fulbe Yillaga’en 0.99X - 0.26 0.85 0.009

Bold: table-wide significant (α = 0.05), see also remark Fig. 7.2

2 2 8 Respectively Y = 0.00344(Year)-6.9, R = 0.91, p = 0.003 and Y = 0.0039(Year)-7.8, R = 0.87, p = 0.006. 7 – Pastoralists responses 175

Changes in grazing intensity With the data on the changes in the number of herds and the time that they spent in the reflooded area, we can estimate the changes in overall grazing intensity for the impact zone. Grazing intensity in the reflooded area increased from 27 to 69 cattle km2 for the six-months dry season between 1993 and 1999 (Fig. 7.4). Three pastoral groups, FulBe Mare’en, Arabs and agro-pastoral FulBe Ngara’en accounted for the largest increase in total grazing intensity. The FulBe Ngara’en and the Arabs pastoralists have been the main beneficiaries of the reflooding, accounting for respectively 33% and 24% of the increase in grazing intensity (Table 7.3).

Table 7.3 – Contributions of Pastoral Groups to Increase in Relative Grazing Intensity, 1993-1999

Nomadic Pastoralists Agro-Pastoralists

FulBe FulBe FulBe Arabs FulBe FulBe Adanko’en Alijam’en Mare’en Ngara’en Yillaga’en

Increase in grazing intensity (Fig. 7.4) 5.0% 8.4% 15.0% 24.3% 33.2% 14.1% Increase in grazing intensity per herd 0.2% 0.09% 0.13% 0.18% 0.53% Increase in grazing intensity -3 per head of cattle (10 ) 1.1 1.5 0.7 1.0 1.1

In terms of increase in grazing intensity per individual household, agro-pastoral Ngara’en and Yillaga’en FulBe have benefited most of the reflooding, although this was biased by the low number of households accompanying the herds (Table 7.3). FulBe Mare’en, who used to be numerically the most prominent pastoralists in the reflooded area before reflooding, were the only group for whom grazing intensity did not increase, meaning that their cattle did not increase and did not stay longer in the study area. Overall, FulBe Mare’en have benefited the least from the floodplain rehabilitation (Table 7.3).

In the first year the increase in grazing intensity was for a large part due to a prolonged stay of pastoralists in the reflooded zone, reducing their stay elsewhere in the floodplain. This was most evident in the case of the FulBe Alijam’en, who on their yearly migration to the northern floodplain, have increased their stay in the reflooded area. When grazing duration in the reflooded area increased, this did not lead to lower grazing intensities in other parts of the floodplain. We observed that other groups, some coming from Nigeria, have replaced groups in the northern and eastern parts of the floodplain that are now visiting the reflooding area. 176 Part II – Impact of reflooding

Other factors affecting transhumance patterns It is important to keep in mind that pastoralists’ transhumance patterns are not only determined by ecological factors; socio-political factors may be important as well (Niamir-Fuller 1999b; Stenning 1957). We expected that this might also have been here the case, because the topic of cattle losses due to theft, diseases or predation, was nearly always raised in our discussions with pastoralists. Moreover, pastoralists argued that the insecurity threatened their access to the rehabilitated floodplain (Scholte et al. 1996b). The start of the floodplain rehabilitation coincided with a period of insecurity, unparalleled in recent history (Issa 2004), in which several pastoralists lost their lives defending their herds. We anticipated that cattle losses might have influenced pastoralists’ transhumance patterns, thereby potentially reducing the impact of the reflooding.

In 1997 we assessed the effects of cattle losses on transhumance patterns and the exploitation of the rehabilitated floodplain. In our annual visits to camps we asked about the number of cattle lost and categorised them as diseases, accidents, predation and theft. We calculated cattle losses as a function of grazing duration in a particular part of the floodplain, following the equation below:

Reported losses Loss rate (fraction per week) = (Number of herds) x (herd size) x (grazing duration)

This survey shows the spatial distribution of risks in the floodplain (Fig. 7.5). We were unable to make a causal link between data on cattle losses with grazing intensities in particular areas of the floodplain and thus limit our discussion here to the following general observations. When we recorded cattle losses in 1997, pastoralists reported the theft of 600 cattle. This was considerably less than they had experienced in the early 1990s before the Waza-Logone Project had undertaken action (Scholte et al. 1996b). A year later, cattle losses were even lower, as special governmental forces had been stationed in the area to counter the insecurity. In subsequent years losses to cattle theft continued to decline. Pastoralists frequently argued that the reflooded area was the area with the highest risk of cattle raids as confirmed by the reported loss rate in 1997 (Fig. 7.5). This may have inhibited some pastoralists visiting this area in the early years of reflooding. The fact that there was no sharp increase in the number of camps and herds after 1998, the year that the special forces practically halted cattle thefts, is an indication that insecurity did not influence the transhumance into the reflooded area.

Later discussions with nomadic pastoralists focused on conflicts with fishers and agriculturalists (Kouokam et al. 2004). These conflicts concerned the blocking of transhumance routes into and within the Logone floodplain by agricultural fields and fish canals (Moritz et al. 2002). The Waza-Logone Project later assisted in the 7 – Pastoralists responses 177

in la a p re in d la o A p a d d re a lo e o g F d lo A a o a rn F M e flo in e rn Z ke rth R a o ste L N a E

Figure 7.5 – Number of cattle lost in different floodplain areas in 1997 For areas, see Fig. 7.1 resolution of these conflicts through the demarcation of livestock corridors into the floodplain and establishment of conflict resolution mechanisms (Kari & Scholte 2001). These conflicts did not seem to have had a lasting impact on transhumance patterns neither.

Another factor that might have affected transhumance patterns in the reflooded area was livestock predation. Agro-pastoral FulBe Ngara’en frequently raised this problem in meetings with the Waza-Logone Project (Scholte et al. 1996b). Although predation by lions was a great concern for individual FulBe Ngara’en, the 155 cattle lost to lions in 1997 represent only a fraction of the total losses in the floodplain. Lion predation only occurred in areas adjacent to Waza National Park, and even there accounted for a relatively small percentage of the total losses (Fig. 7.5). Hyena predation was reported in areas far from the park. Overall, predation accounted for only a small percentage of total livestock losses. It is unclear what exactly the effect of predation was on transhumance patterns, but the fact that FulBe Nga- ra’en continued to use the pastures in and adjacent to Waza National Park suggests that it affected their transhumance patterns only minimally.

Pastoralists suffered the largest cattle losses because of diseases, such as try- panosomiases, foot-and-mouth disease and a variety of parasites. These diseases are endemic, but outbreaks more often affect areas with concentrations of herds, like 178 Part II – Impact of reflooding

the Logone floodplain. The impact of diseases on transhumance patterns is unclear, save in the case of serious outbreaks. In 1997, for example, an epidemic of foot- and-mouth disease struck the borders of Lake Maga, where many pastoralists stayed towards the end of the dry season. The pastoralists there suffered high losses; 1,512 of their cattle died (Fig. 7.5). Other pastoralists, who used the pastures of Lake Maga as a transition zone to their rainy season pastures, waited as long as possible in the reflooded area before moving south, thereby increasing the grazing intensity in the reflooded area that year.

Pastoralists’ incursions into Waza National Park One of the goals of the floodplain rehabilitation was to reduce the competition between livestock and wildlife in Waza National Park by stopping pastoralists’ incursions into the park. The assumption was that reflooding would offer the pastoralists responsible for illegal grazing in the park alternative grasslands outside the park.

After the first year of reflooding, the Waza-Logone Project started a dialogue with informal leaders of the FulBe Ngara’en from the Pétté-Fadaré area, the pastoralists who most frequently were caught in the park. This resulted in a verbal commitment in October 1995 of one of the most influential and respected leaders that his herds and those of his followers would no longer enter the park. In December 1997, this commitment was formalized into a written contract between his group, the Waza National Park authorities and the Waza-Logone Project.

In the period 1995-1997, the FulBe Ngara’en group showed a striking reduction in grazing duration in the reflooded area, following the decision of the (absentee) owners to send their herds further north during the early dry season to avoid any risk of entering Waza National Park. Other FulBe Ngara’en pastoralists followed this example, in part because their leader no longer intervened on their behalf in conflicts with the national park’s authorities. But FulBe Ngara’en pastoralists from Pétté and Fadaré also argued that a major consideration for their commitment to refrain from grazing inside Waza National Park was the improved condition of grazing lands due to the reflooding. This left sufficient grasses in the reflooded zone for the hot dry season when herd mobility is limited. Nonetheless, pastoralists noted that they suffered major losses of newborn calves during the first year of changing transhumance routes and pastures, a situation that improved in subsequent years when herders acquainted themselves with the site-specific circumstances. The reduction in pastoralists’ incursions into the park during 1995-1997 was confirmed by park authorities who only occasionally reported herds within park boundaries.

However, from 1997 onwards, pastoralists complained about the increasing number of cattle in the reflooded area and elsewhere in the floodplain due to an influx 7 – Pastoralists responses 179

of ‘new’ pastoralists, which left little forage for them towards the end of the dry season. These developments undermined the agreement between the FulBe Ngara’en and the Waza-Logone Project, which was already under pressure because of illness of the FulBe leader and personnel changes at the Waza-Logone Project. In the end, despite the increase in perennial grassland productivity, FulBe Ngara’en pastoralists returned to graze their animals in the park.

7.5 Discussion

How did mobile pastoralists respond to the rehabilitation of the Logone floodplain? Did territorial behavior or the lack thereof lead to a lower or higher exploitation of the gradual increases in perennial grassland productivity? Did the rehabilitated floodplain limit the incursions of pastoralists into Waza National Park? The data presented above allows to answer these questions.

Our data shows that risks of cattle losses are spatially distributed (Fig. 7.5). Pas- toralists are aware of this and make decisions, if relevant, accordingly. Cattle losses due to predation were limited in both relative and absolute terms, and probably did not influence transhumance patterns. Cattle losses to theft were much higher, particularly in the reflooded area. But the effect on transhumance patterns seemed minimal; grazing intensities did not increase sharply when security improved. Outbreaks of contagious diseases, however, did affect transhumance patterns in 1997, but only temporarily.

Our data on pastoralist responses can be summarised as follows. During the first six years of reflooding, grazing intensity in the reflooded area has increased linearly, from 27 to 69 cattle km2, contrasting the density dependent increase in the number of camps that after a rapid increase in 1995 levelled off. The continuing increase in the number of herds, albeit slowing down to 7% in 1999, can be partly attributed to natural growth by splitting of the herds present, but the influx of new herds account for most of the increase. The increase in grazing intensity in 1994 and from 1997 onwards is for a large part due to the increasing grazing duration in the reflooded area, because pastoralists spent there more time before they moved further north. The question is what this means for the three scenarios of pastoralists’ responses to floodplain rehabilitation that we earlier postulated.

The overshoot scenario predicted social and ecological ‘chaos’, triggered by a rapidly increasing number of pastoralists and their cattle. In 1995 the area indeed witnessed a rapid increase in the number of camps (Fig. 7.2). This was followed by violent clashes between Arabs and Musgum fishers. However, these conflicts did not continue in the following years, in part due to mediation by the Waza-Logone Pro- ject. The number of camps continued to increase after 1995, but at a much slower 180 Part II – Impact of reflooding

rate. Between 1993 and 1999 overall grazing intensity increased steadily, in part because of the influx of newcomers, in part because of natural growth of existing herds, and in part because some pastoralists spent more time in the reflooded area before they moved further north to their dry season pastures. There are no indications that the increase in grazing intensity has led to overgrazing as from 1995 onwards perennial vegetation has increased with an annual 7% of the reflooded area (Scholte et al. 2000b). We conclude therefore that the overshoot scenario has not taken place.

The territorial scenario predicted that pastoralists who used to exploit the area before reflooding would prevent newcomers from using the reflooded area. In the Logone floodplain, there is not a system in which pastoralists ask formal authorisation to access the floodplain as in the Inner Niger Delta in Mali (Legrosse 1999). There used to be, however, a system in which pastoralists pay taxes and tributes to traditional chiefs in exchange for protection of access to pastures and personal safety (Moritz 2002). This ‘nomadic contract’ has come under considerable pressure in the last decades. Contrary to what could be expected in a territorial situation, most pastoral groups have received proportional benefits from the reflooding, calculated as grazing intensity increase per household (Table 7.3). FulBe Mare’en, who could have expected to behave territorially given their dominant presence prior to the reflooding, even have had lower benefits, calculated per head of livestock, than the other pastoral groups (Table 7.3). The limited increase in number of camps in the first reflooding year could be attributed to territoriality, but in view of the sharp increase in number of camps in subsequent years, it is more likely that lack of information was the main cause. The stabilisation in number of camps in 1996, the low increase in number of herds and the prolonged stay in the reflooded area might suggest a territory effect. In 1997, however, the number of camps increased again, while the number of herds stabilised and grazing duration declined. In 1999, the 1996 pattern repeated. This fluctuating pattern is difficult to explain with a territorial scenario.

The fluctuating pattern seems to indicate the existence of a feedback mechanism between the number of camps, herds and grazing duration, the product of which, grazing intensity, fluctuates around the saturation level of a linearly increasing grazing capacity. We imagine the following mechanism. In a given year with a high number of camps and herds, saturation of grazing intensity is relatively quickly reached, leading to reduced grazing duration, i.e., early departure of pastoralists to other parts of the floodplain. This negative experience will dissuade potential newcomers from visiting the following year the reflooded area. In the following year, the reflooded area will have an increased carrying capacity because of its further developing perennial vegetation. With an equal number of pastoralists, the reflooded area offers more forage, meaning that they stay there longer. This leads to positive evaluations and an increasing number of herds in the subsequent year, etc. 7 – Pastoralists responses 181

This fluctuating pattern supports an Ideal Free Distribution Scenario, which assumes that any increase in perennial vegetation would be subject of consumption by cattle. Apart from the lag in vegetation development in the first year (Scholte et al. 2000b), perennial vegetation and grazing intensity both increased linearly during the study period, suggesting optimal exploitation of grassland resources. Models of the ideal distribution posit different ways in which ideal distribution and thus optimal exploitation is achieved. The Ideal Free Preemptive distribution model assumes that the first individual to use a territory can preempt it. Individuals first occupy the territories with highest quality and only the lower quality ones when the population density increases (Pulliam & Danielson 1991).9 The preemptive distribution gives distinct advantages to someone who arrives first with his herds, over his competitors in exploiting and depleting the vegetation around his campsite before moving on. The question is why our data on changes in grazing intensity fits the ideal free distribution scenario and whether there was a pre-emption effect, i.e., how did pastoralists adjust their transhumance patterns to achieve an optimal exploitation of the rehabilitating grassland resources.

Pastoralist responses The Logone floodplain is best described as an ‘annual grazing area’ for mobile pastoralists (Niamir-Fuller 1999a), i.e., an area used by one or more ethnic groups in which land is not held in common and no action is undertaken against intrud- ers (Casimir 1992). However, in the case of mobile pastoralists in the Logone floodplain this does not mean that access to grazing lands is regulated by the principle of first-come-first-served, as Casimir suggests (1992:160).10 Mobile pastoralists in the Far North of Cameroon have negotiated access to grazing lands through higher-level institutions, in what we have called the ‘nomadic contract’ (Moritz et al. 2002). Compared to the rainy season transhumance area of the Diamaré, there is limited contact between mobile pastoralists and higher-level institutions in the Logo- ne floodplain, although pastoralists pay taxes and tributes to the various authorities in the floodplain.

The overall result is that no pastoralists are denied access, as long as they pay dues to the authorities. In the Logone floodplain, negotiations and coordination occur therefore primarily amongst pastoralists. Niamir-Fuller (1990) calls this ‘passive coordination’ in which no formal agreements are made between pastoralists but where coordinated movements result from individual decision-making. When spa- tio-temporal distribution of cattle is determined by herders on the basis of social

9 The relatively uniform floodplain vegetation masks possible quality distinctions, an issue that has not been further studied. 10 Although access to grazing lands for main camps is not regulated by the principle of first-come- first-served; it is the principle by which the luci herds gain access to pastures. There is increasing competition among luciji herders to enter the floodplain as soon as possible, even though the partial flooding carries health risks. 182 Part II – Impact of reflooding

and ecological characteristics (Turner 1999: 104, 105), this does not automatically lead to competition over the same campsites or grazing lands, since individual herders weigh costs and benefits differently and consequently make different choices. Passive coordination in the floodplain is likely also achieved through formal and informal arrangements or mechanisms. Preemptive use, for example, might be reflected in the customary rights that pastoralists have to campsites in ‘sojourn pastures’ that they have occupied for a long time (i.e., a first-come first- serve basis at a different time-scale). However, pastoralists have only limited usu- fruct rights over the grazing lands radiating from their campsite, the boundaries of which ‘expand and contract in response to the local availability of fodder’ (Tur- ner 1999: 108). Because the radii of different campsites overlap, pastoralists from a number of different camps use the same grazing lands. When reflooding led to greater grassland productivity, this allowed newcomers to occupy new sites in proximity to existing customary campsites while avoiding conflicts over campsites and grazing lands. This explains how the number of camps in the study area could more than double without leading to significant conflicts among mobile pastoralists.11

The other major contributors to the increase in grazing intensity were pastoralists who use the reflooded area as a transition zone to pastures further north. These pastoralists use ‘transit pastures’ and occupy ‘travel campsites’ along the transhumance route that takes them further north. We did not find any timing conventions on these transit pastures, meaning that pastoralists can adjust the time they spend in the reflooded area depending on the resources available. The data presented above shows that pastoralists indeed adjusted their transhumance patterns primarily by shortening or prolonging the time they spent in the rehabilitated part of the Logone floodplain.

The development of pastoral exploitation of the perennial grasslands in the reflooded area confirms that mobility and flexibility are key adaptations to the African drylands (Behnke et al. 1993; Breman & de Wit 1983; Niamir-Fuller 1999b). More importantly, there is no evidence that the reflooding led to a ‘tragedy of the commons’ (Hardin 1968). On the contrary, pastoralists adjusted their numbers and the time spent in the reflooded area quickly and effectively in response to changing productivity of perennial grasslands through ‘passive coordination’ (Nia- mir 1990). Our studies thus not only show the resilience of the Logone floodplain ecosystem (Scholte et al. 2000b) but also the continued flexibility of mobile pastoralism.

11 In all the years of working with mobile pastoralists in the Logone floodplain, we have not come across conflicts between pastoralists over campsites or grazing ranges. PART III

Enhancing Conservation – Development Integration by Management Planning and Training

Risks: the overshoot of success 8 Immigration: A potential time bomb under the integration of conservation and development

Grip on the Whole: management planning 9 At the Interface of Legislation and Wildlife Management: A decade of experience in consensual protected area management planning in Cameroon

Foundation: development of human capacities 10 Curriculum Development at the African Regional Wildlife Colleges, with special reference to the Ecole de Faune (Cameroon)

11 Wildlife Managers’ Perceptions of Community Conservation Training in West and Central Africa

8 Immigration: A potential time-bomb under the integration of conservation and development

Paul Scholte

Published in Ambio 32: 58-64

Abstract Integrated Conservation and Development Projects (ICDPs) aim to stimulate conservation without the previous negative experiences for local people, but pay little attention to their long-term impact such as immigration. The rehabilitation of the Logone floodplain in North Cameroon, the core activity of the Waza- Logone ICDP, has led to a 34% increase of sedentary fishermen and a multiple number of temporary fishermen. Whereas livestock pressure tripled, Kob antelopes, a key floodplain species, have not increased, reducing their competitiveness. The virtual disappearance of wildlife in nearby Kalamaloué NP, due to advanced human encroachment forms, is therefore a bleak perspective for Waza NP. Examples from the Central African Republic, Galapagos, Nigeria and Zimbab- we also showed that in open-access systems, improvement in living standards (development) may stimulate immigration, jeopardising the stability necessary in protected areas (conservation). Most ICDPs lack demographic monitoring, masking its possible immigration risk. To counter the immigration risk in Waza, a policy was formulated based on local stakeholder categorisation and subsequent privileges, resulting in the voluntarily displacement of a village out of Waza NP. It is further recommended that ICDPs should be involved in regional land-use planning and discourage development activities that stimulate immigration.

187

8.1 Introduction

Decades of conservation efforts have left a global network of protected areas, rich in wildlife, forming a true heritage for mankind (Bruner et al. 2001). However, in many instances, the eviction of villagers and nomadic groups associated with the creation of these areas has left deep scars. Subsequent lack of participation in the profits of mass tourism or other park revenues has nourished an almost universal local hostility to protected areas (Anderson & Grove 1989; IIED 1994). In the mid- 1980s this led to a change in conservation paradigm, whereby development was incorporated to counter the negative aspects of conservation for local people (Hulme & Murphree 1999), in particular their loss of access to exploitable resources, generally called people-oriented conservation (Jeanrenaud 1999) or community conservation (Adams & Hulme 2001). Further, it was felt that the efficiency of conservation could be improved by linking conservation and development (Brown & Wyckoff-Baird 1992), hence the dominant project form of Integrated Conserva- tion and Development Projects (ICDPs) referred to in this paper. In an idealised form ICDPs cover three subjects: i) protected area management, ii) management of buffer zones, and iii) local social and economic development (Brown & Wyckoff- Baird 1992). Emphasis has been placed, at least on paper, on stimulating the participation of local communities in the formulation and implementation of these projects.

The decade after the start of the first ICDPs, a variety of reports on their activities have appeared (IIED 1994; Brown & Wyckoff-Baird 1992; Wells et al. 1992; Bran- don 1997; Brandon & Wells 1992; Caldecott 1996). Initially the discussion centred on the justification of people-oriented conservation (IIED, 1994; Wells et al. 1992) or, alternatively, its reject (e.g. Spinage 1998). Later reports discussed preliminary results of ICDPs and the need for adjustments. Brandon (1997), for example, identified questionable assumptions on which many ICDPs have been based which led to disappointing results, especially with regard to their conservation objectives. She argued that many ICDPs have been set up as aid projects where typical governmental responsibilities, such as law reinforcement, have been neglected. Lately, Adams & Hulme (2001) argued to move beyond rhetoric and proposed more diver- sified community conservation, tuned towards the specific situation.

The long-term role of ICDPs in the future of protected areas, most notably the danger of immigration encouraged by its own activities, has thus far received only scant attention. In a review of ICDP case studies Salafsky (1994: 455) noted that ‘none of the case studies mentions developing strategies to deal with the influx of poor migrants that can be expected if the project does indeed succeed in raising standards of living relative to surrounding areas’. Brandon & Wells (1992) and to some extent Caldecott (1996) also discussed the potential risk of people being attracted by the activities of an ICDP. First-hand experiences with the phenomenon 188 Part III – Enhancing conservation – development integration

have been reported, in narrative terms only, from the forests of West Africa (Oates 1999) and the Central African Republic (Noss 1997).

One can well imagine the following, highly simplified, immigration-risk scenario for open-access systems (Ostrom 1990). A protected area is surrounded by a few poor communities, who depend on the (illegal) exploitation of the protected area’s natural resources. However, this low-intensity exploitation has hardly any impact on the area’s ecology. For a variety of reasons, an ICDP is started in the area, improving local living conditions, thereby attracting (poor) newcomers, who want to share in the increased resources. In the absence of barriers, immigration will continue until everybody has once more reached the same (low) level of development. The result is that the protected area is now surrounded by many, instead of a few, poor communities, who still have to make use of the same resources, thereby jeopardising its ecology due to their increased numbers. Is this development or conservation or neither one?

This article provides case studies of an ICDP in the Waza-Logone area of Cameroon with which to examine the validity of this scenario for open-access and possibly other property systems as well. These experiences are compared with those from other ICDPs, and strategies to deal with immigration problems associated with such projects are illustrated with a recent example from Waza-Logone.

8.2 Waza-Logone case studies

The Waza-Logone area The Waza-Logone project area covers about 8000 km2 in the Sahelo-Sudanian zone of Cameroon (Fig. 8.1). The area is bordered in the west by Nigeria and in the east and north by the Logone River, which forms the border with Chad. The region is characterised by floodplains, which are intensively used for fishing (Drijver et al. 1995) and dry-season grazing (Scholte et al. 1996b). The area includes two national parks, Waza and Kalamaloué, containing a diverse population of large mammals: Elephant (Loxodonta africana), Giraffe (Giraffa camelopardalis), Hippopotamus (Hippopotamus amphibius), seven antelope species, three primate species, War- thog (Phacochoerus africanus) and predators such as Lion (Panthera leo), Spotted and Striped Hyena (Hyaena hyaena, Crocuta crocuta). The Waza-Logone area is also considered as one of the principal waterbird areas in West-Central Africa (Scholte et al. 1999a).

Since 1979, the natural hydrological regime of the Logone floodplain has been affected by the construction upstream of a dam near Lake Maga and an embankment along the Logone river, as part of the parastatal irrigated rice scheme (see Fig. 8.1). These structures have reduced both the depth and extent of flooding in an 8 – Immigration, a potential time-bomb 189

Figure 8.1 – Map of the Waza-Logone area. 190 Part III – Enhancing conservation – development integration

area of 1500 km2. Subsequently, productive perennial grasslands have been invaded by annual grasses, limiting dry-season regrowth and reducing the carrying capacity of the area for both livestock and wildlife (Scholte et al. 1996b). Fishing resources have declined dramatically as well (Drijver et al. 1995). Since 1979, about 40% of the human population then settled on the floodplain has left the area, while a large part of the remaining population has shifted from fisheries and animal husbandry to agriculture (Scholte et al. 1996b). Lack of flooding is perceived as the major problem in villages around Waza National Park (NP), followed by the denial of access to fishing and grazing grounds in the park, and intimidation by park authorities (Scholte et al. 1999b; Chapter 9).

Impact of human encroachment on Kalamaloué NP In contrast to Waza NP, which has a size of 170 000 ha, Kalamaloué NP is a small national park of 4500 ha situated within a densely inhabited area (Fig. 8.1). As will be argued below, developments that have taken place around Kalamaloué NP and subsequently had their impact on the park, might well take place around Waza NP if the ICDP, which was only started in 1992 and therefore has only had a limited impact as yet, was to become too successful. Kavanagh (1978) concluded, based on his experiences in 1974 and 1975, that Kalamaloué NP had tremendous potential for tourism with its wide variety of mammals and birds. The ten park guards, since 1974, and their reasonable working conditions should have been sufficient to maintain its condition (Bruner et al. 2001). Nonetheless, 20 years later dramat- 1 ic changes in wildlife numbers have taken place: Waterbuck (Kobus ellipsiprym- nus) became extinct, and the reduction of Kob (Kobus kob) and Warthog overshad- ow variation in numbers due to counting bias (Table 8.1).

Recent studies have indicated an enormous pressure on the park from surrounding communities.2 Firewood extraction has been estimated at more than ten tons (wet mass) a day, considered to be unsustainable as indicated by the declining cover of woody species. Each day an average of five herds of 60-120 cattle and ten small- ruminant herds from neighbouring villages, complemented by several herds from further away, are found inside the park. Annually approximately 300 fishermen

1 Wildlife numbers in Kalamaloué NP (based on unpublished reports by the Ecole de Faune Garoua and Scholte and Saleh unpublished data): Line transect counts, covered by foot with three observers per line, have been carried out in 1979, 1981 and 1984. All animals observed in a 100 m band on both sides of the 15 transects, 1 km apart, were counted, corresponding to a 20% overall sample. In 1994 and 1996 these counts were repeated by park guards. The latter also confirmed the disappearance of species that were no longer found. 2 Natural resource exploitation in Kalamaloué NP, based on Saleh (1994): A general survey was carried out by a forestry student, originally from the area, in each of the 26 villages bordering Kalamaloué NP, obtaining information on village history and demography. This student assessed the number of fishermen within the park during a three-month study. Firewood use extraction from the park was estimated in 1996 during three sessions each of 4 days on 15 main transit points around the park in different seasons. Information on the presence of livestock in the park was obtained during these studies. 8 – Immigration, a potential time-bomb 191

Table 8.1 – Wildlife counts in Kalamaloué NP, 1979-19961

Species 1979 1981 1984 1994 1996

Warthog 563 618 141 175 50 Red-fronted Gazelle 38 132 0 410 195 Kob Antelope 2733 1837 112 920 95 Waterbuck 481 1159 extinct extinct extinct Bushbuck 68 73 9 60 present Common Duiker 88 0 0 5 10

1Based on line transect counts, see footnote 1. make use of the fish resources of the park. Despite the official ban on exploiting park resources, many people have deliberately chosen to live on its borders. Of the 26 villages in or on the border of the park (within a distance of less than 1 km), eight were created after its delimitation as a forest reserve in 1946 followed by another six after its designation as a national park in 1973. This substantial immigration was caused by the availability of fish as well as pastures and wood in the park, which are easy to market as a result of the proximity of the ever-growing markets of Kousseri and N’djamena.

The Waza-Logone ICDP’s reflooding’s impact on Waza National Park The Waza-Logone project started in 1992 with the goals of planning and implementing interventions to restore the area’s biological diversity, and to sustain the development of its inhabitants: a genuine ICDP. After consultation with local communities and (national park) authorities, a water course blocked by the embankment along the Logone river, was re-opened in May 1994, reflooding an area of 180 km2 of desiccated floodplain.

Monitoring (Scholte et al. 2000b) showed that in 1994, the first year of reflooding, vegetation production was high and perennial grasses set seed abundantly, but no changes in vegetation composition were observed. In subsequent years an annual 7% increase in perennial grasses, most notably Echinochloa pyramidalis and Oryza longistaminata, has been observed, and these are expected to fully cover the area once again in 2003 (Chapter 3). Vetiveria nigritana, dominant before the dam construction, has not shown any increase as yet. Waza NP harbours an important large-mammal population, with fluctuating numbers during the last 34 years of counts based on various methods.3 Numbers of Kob antelope have reduced almost ten-fold between 1977 and 1993, a variation that can not be explained

3 Wildlife numbers in Waza NP (based on Esser & van Lavieren 1979, Scholte et al. 1996a, Scholte unpublished data, see also Chapter 6). 192 Part III – Enhancing conservation – development integration

by count bias (Fig. 8.2). The decline of this typical floodplain species (Kingdon 1997) may be primarily attributed to the reduction in dry-season forage availability due to the Maga dam construction in 1979. The impact of poaching on antelope’ numbers was probably of secondary importance. With the rehabilitation of the floodplain, the Kob population has initially increased, but numbers stabilised from 1995 onwards (Fig. 8.2).

40000 Kob antelope Cattle in Waza NP and number in reflooded zone reflooded zone (dry season)

30000 number

20000

10000

0 1962 1976 1977 1993 1994 1995 1996 1997 1999 dam reflooding Legend: Methodology Kobs (Chapter 6), cattle (Chapter 7). No cattle surveys have been conducted prior to 1994. Indicated years reflect the relevant flooding season (= actual survey year -1).

Figure 8.2 – Number of Kob antelopes and cattle in Waza National Park and the reflooded area.

In spite of 40 years of protection, with 20-30 reasonably well equipped park guards, almost all surrounding communities continued to make intensive use of Waza NP. Or, as they themselves put it: ‘we have always fished in the water holes our parents dug; the park only came afterwards’. When the park is flooded (from June to December), people are not hindered by the presence of game guards, who cannot easily access the flooded area. ‘It is with the grader opening the roads that the park boundaries return’.

In 1994-5, a 25% increase was recorded in the numbers of sedentary fishermen fishing within and near Waza NP, increasing to 34% in 1995-6, the second year of reflooding (Table 8.2).4 In 1994-5 the harvest was estimated at over 10-kg fish ha-1

4 Fishermen population in Waza NP, based on Saleh (1995): The number of sedentary fishermen exploiting Waza national park and its direct surroundings (<5 km of boundary) in the previous and 8 – Immigration, a potential time-bomb 193

Table 8.2 – Changes in number of sedentary fishermen in Waza NP villages, 1993-19951 villages 1993-4 1994-5 1995-6 % increase location of origin2

Tchédé 22 26 28 27 Lagdo Zouang 21 21 21 0 – Lougouma 14 26 30 114 Lagdo, Maga, Garlé Mahé 15 19 21 40 Zina, Tila, Guidiba Baram 16 18 18 13 Lagdo Total 88 110 118 34

1Methodology, see footnote 4; 2Lagdo, situated 300 km to the south; Maga, situated 50 km to the south; others situated nearby in floodplain. with a wet-season income per fisherman of US$ 400, much more than a farmer can make. The number of people fishing in and near Waza NP has been estimated to have tripled with the arrival of seasonal fishermen from neighbouring villages. Their highly variable presence and the inaccessibility of the floodplain during flooding prevented monitoring their numbers in any detail.

Livestock intrusion from nearby grazing lands is considered to be one of the most urgent problems by the park authorities (Scholte et al. 1999b). In the first three years after the reflooding, the number of pastoral camp groups in the area near Waza NP increased from 47 to 106. Livestock pressure in the dry season increased in the period 1993-1999 from 14 000 to 39 000 cattle5 (Fig. 8.2). The cattle: Kob ratio increased from 5 till 8, illustrating the marginalisation of the antelope populations that, towards the end of the critical harsh dry season, compete with the abundant cattle for the same, scarce grasses (Noupa 1995).

In the first years after the reflooding, the sedentary population of the villages on the Waza NP edges was increasing at an annual overall rate of 5%, composed of 3% immigration and 2% natural growth, representing a doubling within a period of 14 years.6 Only one new village is planned to be established, at Doudou Ndiyam near Goulou (Table 8.3; Fig. 8.1). Five households who left in 1985 have written to the administrative authorities asking to be allowed to return to their natal site. current year was assessed in 1994-5. Fishing revenues were estimated based on direct observations and interviews. Repeated visits complemented data for 1995-6. 5 Grazing intensity of nomadic and transhumance pastoral population in the Waza area (Scholte et al. 2005; Chapter 7). 6 Waza park village demography, based on Tambekou et al. (1997): During PRA studies (Scholte et al. 1999b), a baseline demography study was executed, nine villages were covered in 1994-5, the remaining five in 1995-6. In 1996-7, the number of households was re-examined and information was obtained on motivation of settlement and village of origin. Households were used as sampling unit, being more reliable than individuals. 194 Part III – Enhancing conservation – development integration

Waza NP complements the fishing and grazing grounds elsewhere in the plain, due to its position downstream in the floodplain (Fig. 8.1). Depressions outside the park dry up several weeks earlier than those inside and fishermen shift their fishing activities into the park when catches have fallen elsewhere. With the reflooding, water holes deep in the interior of the park have become well stocked with fish and nutritious vegetation until the end of the dry season, and therefore attractive to fishermen and pastoralists. The reflooding has had major benefits for the park’s natural resources as well as for the surrounding human communities, at first sight a ‘win-win situation’. As suggested by the deteriorating cattle-antelope ratio, it seems unlikely that the increased benefits for the park can keep pace with the demands of the increasing human population and, without interventions to restrict access to park resources and immigration, the situation at Waza may ultimately resemble that at Kalamaloué NP.

Table 8.3 – Changes in number of households in Waza NP villages, 1994-19961

Number of households migration balance household/year 1996-7 1995-6 1994-5

Amkodje 14 11 3 Andirni 68 70 -1 Badadaye 29 27 1 Baram 17 17 0 Diéguéré 23 23 0 Goulou 52 31 Lougouma 22 20 1 Ndiguina 91 90 1 Niwadji 69 64 2.5 Mahé 22 19 1.5 Mbili 79 78 1 Tagawa 1,2 29 26 3 Tchede 24 22 1 Zwang 242 23 0 TOTAL 515 208 285 15 (3.0%)

1Methodology, see footnote 6; 2Includes one household ‘naturally’ constituted (not included in migration calculations). 8 – Immigration, a potential time-bomb 195

8.3 Other immigration examples

The Campfire program in Zimbabwe is more a rural development program than an ICDP (Metcalfe 1994). Nevertheless, owing to its success it is internationally seen as a prime example of a community conservation project. In and around the various districts where it is active, ‘land hunger’ is the major (push) factor for immigrant installation (Dziginrai 1996). Immigration in wildlife areas causes fragmentation of wildlife habitat and competes with Campfire revenues and projects (Bird & Metcalfe 1995). In Hurungwe district (West Mashonaland), the achieve- ments of the Campfire project itself, visible in the form of revenue distribution, rural clinics and schools, also attract people from other, non-Campfire, villages in the district as well as from outside the district. This is considered ‘a very serious problem’ by the Hurungwe district officials (Mr. Banda and Hurungwe district councillor, pers. com. 1996) and Campfire officials alike (Anon. 1996). In a neighbouring district (Omay), mechanisms have been developed to control immigration by an adapted form of land-use planning and its, till recently, (rigorous) application (Taylor, pers. comm.).

Lately, the impact of immigration on the forests of West and Central Africa has received particular attention (Noss 1997; Oates 1999). Low initial population densities, a lack of land tenure and sudden large-scale (conservation) interventions in an environment with few employment opportunities, are some of its principal reasons. Passionately, Oates (1999) described how various prime wildlife areas have been destroyed through the bias of development projects and ICDPs alike. He concluded that in the West-African rainforest context any mutual benefit between conservation and development is an illusion.

The Galapagos Islands are a prime example to show that conservation and development are intrinsically linked. Wildlife tourism, with an estimated annual value of US$ 100 million, and fisheries, has made the islands the richest province of the country and triggered an influx of 12 000 people or three-quarter of the Galapagos’ present population (Abbott 1996; Int. Am. Dev. Bank 2000). Despite the restricted area open to human habitation, 3%, this influx threatens the islands’ long-term conservation, an issue that has received a fair amount of attention in the popular press (e.g. Benchly & Franklin 2001). For years, conservationists have pleaded for restrictions on immigration, arguing that any loss in biodiversity would have major consequences for the booming tourist industry as well (Abbott 1996).

Immigration has also been described as a major threat for protected areas by Sherbenin and Freudenbergen (1998), who analysed various push and pull factors that lie at its origin. Surprisingly, they rejected the link between the activities of ICDPs, influencing several of these push and pull factors, and immigration based 196 Part III – Enhancing conservation – development integration

on ‘incomplete evidence’ (Sherbenin & Freudenberger 1998: 51). Indeed almost no ICDP monitors demographic trends in its project area, making it particularly difficult to assess the impact on immigration. This is at odds with ICDPs’ development objectives and complements the concern of Kremen et al. (1994), who earlier noted a comparable lack of wildlife monitoring in ICDPs.

8.4 Discussion

Justification for ICDPs The doling-out of development activities such as wells and roads is widely used to motivate people to refrain from exploiting protected area resources, although there is little experimental evidence for the success of this approach. It is also often seen as a means of compensating for lost resources (Sayer 1991) and destined to all people living at the edge of the protected area. It becomes somewhat theoretical to compensate already evicted people for resources lost in the past. Instead of con- ceptualising solutions in terms of compensation, it seems more appropriate to approach the problem by ‘learning by doing’, with initial activities such as the Waza reflooding constituting a first step in a process of stimulating an atmosphere of trust and dialogue in which possible strategies to control the immigration risks can be negotiated.

But why should people in Waza receive such a ‘development compensation’ if they still use park resources? The studies referred to above (Saleh 1994, 1995; Scholte et al. 1999b), showed that hundreds of people exploit fish, wood and pastures from inside Kalamaloué NP and Waza NP on a very regular basis. Their habitation on the park boundary would no longer be possible if the parks were really closed to them. In a number of cases, the exclusion of local people has even privileged others, like the inhabitants of a remaining village in Waza NP who fish intensively in water holes on the territory of one of the evicted villages.

Risks of ICDPs Compensation for lost resources to all people in and around protected areas may provide incentives leading to an intensification of exploitation of park resources, based on increased migration into the area around the park. Even if exploitation can be prevented in the short term with an improved park control system, the sharing of improved resources by a larger number of people around the park, to the extent that everybody is as poor as before, will only result in increased long-term risks to the park’s survival. The occupation of forest reserves in Nigeria by immigrants attracted by rural development opportunities (Oates 1999), showed the necessity of establishing a link between conservation and development objectives (Brown & Wyckoff-Baird 1992). The underlying problems are even more far-reaching: how can an ICDP work in such frequently occurring open-access systems, where an 8 – Immigration, a potential time-bomb 197

increase in the standard of living attracts people and thereby increases pressure on the protected area?

It is proposed to emphasise the long-term objectives of the ICDP that must be realistic, even under changing conditions. Conservation risks should be explicitly distinguished, if possible, from the start of project implementation, and a system of adaptive management able to deal with them should be developed during the execution of short-term activities. In cases of collapse of the state, or more common its advancing erosion, dependency on governmental agencies has shown to be far less successful compared to a system based on people with traditional ties to the area supported by international non-governmental organisations (Hart & Hart 1997). Experiences from the Waza-Logone area have shown the unpredictabil- ity of changes, such as the insecurity in the area from 1990 onwards and the presence of key persons brooking deals between authorities and local communities.

Sinclair & Arcese (1995), Hart & Hart (1997) and especially Oates (1999) have previously questioned the efficiency of ICDPs. They argued that ICDPs fail to provide success vis-à-vis conservation and enhance dependency on funding agencies, risking over-exploitation if the support system collapses. Without doubt the general complexity of ICDPs and conflicting interests will lead to these problems, which are well known in the development-project world (Brandon & Wells 1992). One can question whether it is realistic to expect ICDPs to function properly in terms of all their objectives, even though this is essential for their output (Brown & Wyckoff-Baird 1992).

8.5 A way forward

Who is local? Not everybody found in or around a protected area can be considered ‘local’. It is therefore surprising that little effort has been made to identify local user/stakeholder groups in and around protected areas (but see Jeanrenaud 1999). In a publication on community wildlife management (IIED 1994), local resource users were not further categorised, masking diversity in interests and, consequently, management potential. Around Waza NP, communities have been categorised based on i) territory/space, ii) activities undertaken and iii) kinship (Fig. 8.3). Other, predominantly non-local, stakeholders can be categorised as park authorities, (governmental) agencies, (inter) national visitors, donors, researchers, business people etc. (IIED 1994). 198 Part III – Enhancing conservation – development integration

WAZA COMMUNITIES

Territory (i) Kinship (iii) protected area • people returning to their home village 8 • people of direct descent, village on village ter- but not born in or in the boundary ritory on vicinity of the protected boundary area 9 2 village inside 3 1

• agropastoralists 4 Activities undertaken (II) • seasonal fishermen with activities outside their own village territory 4 villages/towns further away

• village or town based pastoralists 5 transhumant pastoralists 7 • seasonal fishermen 6

Legend: Categories based on i) territory ii) activity undertaken iii) kinship, see also text.

Figure 8.3 – Categorisation of Waza communities

An appropriate categorisation of stakeholders is a crucial start for immigration control. A solution formulated by Caldecott (1996: 52) for the open-access system around the national parks in Cross River State, Nigeria: ‘an indigene was therefore anyone whose claim to be one was endorsed by an appropriate village council’ remains too open and may still lead to an unmanageable situation. National park and/or ICDP authorities should also have a say in excluding some categories. ‘Ad- vising villagers that registration of non-indigenes would be against their own interests because it would cause benefits to be diluted’ (Caldecott 1996: 52) will work only if the committees who decide on the criteria of who is local are small enough, without the danger of being dominated by leaders.

Barriers to immigration: a (remaining) role for buffer zones? Buffer zones have long been regarded as a tool to integrate protected areas in their regional setting (Batisse 1982). They are assumed to extend the habitat for wildlife and offer products for surrounding communities (MacKinnon et al. 1986). They have also been proposed to halt further encroachment arising from the increasing 8 – Immigration, a potential time-bomb 199 demand for agricultural land by subsistence farmers (Nepal & Weber 1994). Im- plementation of the buffer zone concept is difficult in Waza-Logone and elsewhere, as its legal status, as a type of protected area, may result in the perception of an encroachment of the national park (Neumann 1997).

But even if such constraints are lifted, administrative regulations are by no means sufficient for influencing people’s behaviour. Traditional leaders play a major role in the decision-making process regarding the creation of new villages (Dzingirai 1996). In many areas the authority of traditional leaders is eroding, making these leaders no longer the single deciding authority. Worse, this ‘power vacuum’ is one of the main causes of the development of open-access situations, the single-most important condition for immigration (Ostrom 1990). The traditional leader of Kousseri, who controls part of the area covering Kalamaloué NP, explained that he decides on the creation of new villages. Asked about the villages recently established, he was not aware of them: ‘Apparently my say in this is not always necessary’. It is, further, in the interest of local leaders to increase population numbers in their territory, an economic incentive that cannot easily be reversed (Sultan of Logone-Birni, pers. comm. 1996).

Analysing the human encroachment around Kalamaloué NP, one can conclude that parameters that are regularly the focus of improvement by ICDPs, such as accessibility, year-round water availability, availability of land for cultivation, grazing and wood collection and a nearby market, were all present. To avoid a similar encroachment near Waza NP, the access to these resources should be regulated, for example by strengthening the control of the park (Bruner et al. 2001). Expe- riences from Kalamaloué NP, with its relatively high park guard-size ratio (1: 500 ha), showed the limits of such an approach, however.

A more powerful tool would be a policy of active discouragement of development activities in the surroundings of protected areas. In a buffer zone adjacent to the protected area, no rural development activities should be practised, or only such activities that have no appeal to human communities. In the transition or support zones (Wells et al. 1992), sustained rural development as part of a general strategy of regional development may be promoted, in a manner that limits the risks of human attraction. These support zones should provide alternatives to those communities that exploit the buffer zones and protected areas and will not receive development assistance in this framework. General guidelines are difficult to provide and depend also on pragmatic choices that often must be made. However, the utmost caution should be adopted vis-à-vis road construction (Wilkie et al. 2000) and increasing pastoral and domestic water availability. This contradicts present ICDP practices and buffer zone development as described by Sayer (1991), whereby various extractive uses originally proposed for the support zones only, are promoted right in the buffer zones. 200 Part III – Enhancing conservation – development integration

Some preliminary results of a new immigration policy in Waza-Logone In September 1997, a management plan for Waza NP and its periphery zone was approved by the government of Cameroon to improve relations between park authorities and its neighbors (Scholte 2000b; Chapter 9). Another objective was to prepare the park for the immigration of people to its borders with the floodplain rehabilitation efforts by the Waza-Logone Project. In this management plan, a committee was created, with representatives of selected communities and park authorities (both with voting rights) and administrative and traditional authorities (without voting rights). The Waza management committee, formalised by a ministerial decision, has representatives of all villages either situated on the park boundary, 2, or with their territories bordering it, 3 (Fig. 8.3). Agropastoralists as well as transhumant pastoralists 5, 7, are represented as long as they exploit at present the periphery zone of Waza NP. Seasonal fishermen, 6, were excluded. Baram, the only village inside the park boundary, 1, was also excluded to participate in the committee with the idea that this could ultimately persuade them to settle outside the park boundaries, in accordance with the official conservation policy. In 1997 and 1998, village water pumps, small-scale irrigated rice cultivation and local tourist camps, were promoted by the Waza-Logone Project in the park villages, 2, 3, but excluded Baram, 1. People returning to their home or parents’ village, 8 and 9, have been considered for the time being as belonging to one of the categories 1-7.

Characteristically, all Waza villages originate from well before the establishment of the national park (Scholte et al. 1999b). In other situations it might be useful to distinguish those villages and inhabitants present prior to the establishment of the protected area from those which arrived more recently with, obviously, less rights. The list of distinguished activities inside the park boundary may be extended to include other natural resources than grazing lands and fishing waters as well as ceremonial or religious attractions, as is the case of Niokolo Koba NP in Senegal (pers.comm. students Ecole de Faune).

Early 1998, one of the families in Baram persuaded the other 14 families to move out of the village (and out of the park). A letter was sent to various authorities and to the Waza-Logone Project asking for assistance. In a subsequent correspondence, the village was promised that, once moved beyond the park boundary, they would benefit from the same advantages as the villages outside the park. In October 1998, the initiator settled with his family in Halé on the edge of the park (Fig. 8.1) followed by eight Baram families between March and June 1999. Whereas in 1996 and 1997 the initiator received fines of about USD 150 for fishing inside the park, in 1998 his relationship with the park warden improved considerably. In June 1999, the Waza-Logone Project installed a highly appreciated water pump in the new village. Whereas decades of oppression towards Baram by virtually all park wardens did not change their continued presence, the new more balanced policy of ‘the carrot and stick’ has yielded a tangible result. As the initiator of the 8 – Immigration, a potential time-bomb 201

move told me (June 1999), ‘as long as they pressed us to leave we were determined to stay, but as soon as there was a perspective of a more prosperous stay outside the park and people were no longer daily harassed, we changed our opinion’. This experience should guide the further development of the policy with regard to the newcomers in a regional context.

8.6 Conclusions and recommendations

The Waza-Logone area is unusual in having experienced a degradation process with subsequent emigration and having (partly) been successfully rehabilitated, provoking a population build-up on the park’s borders. The risk of ICDPs becoming poles of attraction for people has also been noticed in the Central African Republic (Noss 1997), Nigeria (Oates 1999) and Zimbabwe (Anon. 1996). These areas have in common that, for a variety of reasons, access is not well regulated and surrounding areas exercise a high (emigration) pressure. In other open-access areas it is expected that the lack of real impact of ICDPs, often running for a limited period only, has masked the threat of immigration. Anyhow without a proper demographic monitoring system only massive immigration can be detected. Such demographic information is prerequisite to further analyse the risk of ICDPs, in other than open-access property systems possibly as well.

These negative experiences do not detract from the necessity of integrating protected areas in their regional setting, for which the ICDP concept, if more realis- tically designed and executed, remains the best alternative. Taking into account the following suggestions can already make some major improvements: • More emphasis should be put on conservation objectives as experience shows that they are often marginalised in daily ICDP routine (Brandon 1997). Improved park control can reduce some of the impacts of immigration. • ICDPs will have to devote more effort to targeting local resource users. Se- lected groups can play a key role in the management of the areas surrounding the protected areas and may form a social fence for further influxes. Special privileges will have to be given to these people and in some cases exclusive land rights (Neumann 1997), without risking a loss of control over the protected area. This could reduce the ‘openness’ of the area. • To influence the development of activities around protected areas, ICDPs should devote more attention to regional planning, an activity demanding the presence of project personnel in the lobbying worlds in district, provincial and national capitals. This will have to be accompanied by further commitments of the various authorities involved, most importantly an appropriate legal context and its subsequent implementation. 202 Part III – Enhancing conservation – development integration

• Generally, smaller budgets spread over longer periods will allow ICDPs to concentrate on their role as facilitator in conservation-development planning instead of becoming a financing party with all its ‘spending obligations’ that often are easier funnelled into development than conservation.

If these recommendations result in an ICDP becoming so complex that its conservation and development objectives are not realistic, while immigration remains a threat for the protected area, its continuation is no longer justified. If continuing, the ICDP may otherwise appear to be a Trojan Horse. 9 At the Interface of Legislation and Wildlife Management: A decade of experience with consensual protected area management planning in Cameroon

Paul Scholte

In: The Future of Conservation in Africa: Law, Biodiversity, Livelihoods and Development (2005).

Abstract Management plans are considered to be indispensable for the daily management of protected areas. Historically, their focus was on relatively straightforward technical or ecological issues. Since the 1990s management plans have increasingly been based on the consensual interpretation of pressing management issues, often related to the legal context of protected areas. The most important product of the modern management plan is a consensus building process based on negotiation among a large number of stakeholders. This chapter asks whether the high expectations for management planning as an interface between protected area management and wildlife law and policy are realistic. The analysis first sketches protected area management policy in Cameroon, particularly the limitations of the legal context in which it occurs. This is followed by an account of the formulation process for the Waza NP management plan, a process that has led to a consensus on sensitive key issues among local communities, local and national authorities, and international environmental NGOs. This particular exercise in management planning was a learning process for all concerned. The analysis questioned, however, whether the resulting management plan was sufficiently realistic to lead to its successful implementation. The Waza case is then compared with management planning experiences in other protected areas in Cameroon and in other African countries.

205

9.1 Introduction

Recent developments in African wildlife conservation In the late 1980s and 1990s, African wildlife conservation underwent profound changes. It became increasingly evident, for example, that wildlife conservation could no longer be considered the sole responsibility of government. Local communities and, in many regions, the private sector would have to play a vital role. Conservation has now come to be seen as part of an overall sustainable development strategy. Although the precise meaning and consequences of this new conservation (Hulme & Murphree 1999) will continue to be debated by some (Spinage 1998) and needs to be conceptualised more fully (Adams & Hulme 2001) protected area managers can no longer ignore the pressure to come to grips with the new expectations imposed on them by the outside world.

In an increasing number of African countries, elements of the paradigm shift are taken into account in newly drafted national wildlife legislation. But in most countries, including Cameroon, the situation on the ground in local communities seems not to have changed very much (Bomba & Bigombo Logo 2001). This com- plicates things for protected area managers who are trapped between the ‘unrealistic’ aspirations of local people, ‘rigid’ legislation, ‘distant’ superiors and ‘impossible’ demands of donors and pressure groups.’ (Scholte 2000b). African protected area managers also have to face conflict situations (Hart & Hart 1997), the intrinsic limits of conservation and development integration (Scholte 2003a) and a lack of essential equipment and personnel. In the face of these constraints it is hard to be sure what is actually happening in the field. Evidence from various parts of Africa suggests, however, that there is a ‘silent’ wildlife crisis. Some explain this by pointing to the continued dominance of fortress conservation, which limits the viability of wildlife resources on common land (Brockington 2002; Gibson 1999). Others blame the new conservation for the lack of success (Oates 1999).

Here, I focus on the long implementation trajectory for changing wildlife law and policy. Now that the rules are shifting, is wildlife conservation still being pursued as business as usual? Or are some of the offshoots of the new conservation paradigm really having an impact on the ground? Management plans are increasingly seen as valuable tools for redirecting protected area management, especially when they are part of donor assisted integrated community development projects (ICDPs) (Thomas & Middleton 2003). Given the controversies surrounding many African protected areas, it is no surprise that external support is often conditioned by consensus building on the need for management planning. But does it work? How does it work? 206 Part III – Enhancing conservation – development integration

Management plans

The ‘fashion’ for protected area management planning Management plans are said to be indispensable tools for the daily management of protected areas around the world (MacKinnon et al. 1986; Ortega-Rubio & Ar- guelles-Méndez 1999; Parr 1998; Rettie 1995). Management plans have multiple functions, the most frequently cited of which are the identification of management needs for a protected area, the setting of its management priorities, and or- ganising an approach to its future (MacKinnon et al. 1986), see also Fig. 9.1. His- torically, management plans were focused on relatively straightforward technical or ecological issues, for which ecologists, together with protected area managers themselves provided the information base. Now, the ‘fashion’ for management plans brings a broader mix of scientists, decision makers, and financing agencies into the picture, each with their own motivations and expectations.

For scientists, management plans seem to play a role as baskets for research results. A literature search of peer-reviewed science journals returned 172 conservation related hits on the key words management plan.1 Sixteen per cent of these papers analysed (parts of) existing management plans, 28 per cent described research that, according to the author, potentially supports the design of elements of (future) management plans, and 50 per cent described more fundamental research and superficially concluded that it could be useful for management planning.

In a number of countries, the preparation and implementation of management plans is now legally prescribed for national parks. They include Australia (ANZECC 2000), Mexico (Ortega-Rubio & Arguelles-Méndez 1999), England and Wales (Tho- mas & Middleton 2003) and Cameroon, where

2 A management plan shall be drawn up for the State forests defining, in accordance with the conditions laid down by decree, the management objectives and rules for each forest, the means needed to achieve the said objectives, as well as the conditions under which the local population may exercise their logging 3 4 rights in accordance with the provisions of the classification instrument.

1 Based on a search through the abstracts of Institute of Scientific Information (ISI) journals from 1995-2003, 26% of the hits concerned papers dealing with conservation, the others had a medical (46%), hydrological or geological (11%), technical (10%), fisheries/agronomic (7%) or management (1%) orientation. 2 State forests include wildlife protected areas, such as national parks, game and hunting reserves, state game ranches, wildlife sanctuaries, buffer zones and state zoological gardens as well as several forest reserves categories (MINEF 1994: 9-10). 3 In the French law text the broader term ‘droits d’usage’ is used (traditional user rights) (MINEF 1994: 11). 4 MINEF 1994: 11. 9 – Management planning 207

Management plans are also required for natural sites being considered for inclusion on the World Heritage List (Thomas & Middleton 2003) and ‘...provisions should be made for a management policy or plan’ for Biosphere Reserves (UNESCO- MAB 2002). NGOs and donor agencies have arguably become the strongest sup- porters of management planning. They see them as indispensable instruments for the integration of conservation and development (Margolius & Salafsky 1998).

Legislation Agency, policies, strategies, business plan

Regional plan, broadscale land management planning

MANAGEMENT PLANS FOR INDIVIDUAL OR GROUPS OF RESERVES

Subsidiary, e.g. fire plans, site plans etc.

Operational/action plans, works programs

Implementations

Figure 9.1 – Planning hierarchy (adapted from ANZECC 2000) In bold: relevant strata for Cameroon

Recent development of management plans Parallel to the change in the wildlife conservation paradigm in the 1980s and 1990s, the aim of management plans also changed. They developed from being essentially technical documents into more formal presentations of the outcomes of negotiation. Prior to the 1980s, a management plan was often a technical document, informally prepared by a dynamic protected area warden. It had no formal significance and was only for internal distribution, often in photocopy form or even as a carbon copy. Consultants were often asked to co-ordinate the management plan formulation process and describe in detail the biological environment and pro- 208 Part III – Enhancing conservation – development integration Interpretation? gives more flexibility exploitation (as Waza) communities No Developments No of park villages (village hunting) co-management local new internal organisation No Submitted for approval of the minister. 4 Last version submitted for approval of the minister, two earlier versions rejected Last version submitted for approval of the minister, WWF Consultants Park exploitations,YesWWF Consultants Park 2 Tropenbos MINEF, Poaching, tourism EU-ECOFAC Consultants + User rights in Comparable to 2 1 3 4 World Heritage Site World +warden but status as game reserve (198?) Status Status (1950) Reserve project team relation to zoning but status Waza, (1968) Reserve to immigration, (1986) status pursued mainly re-settlement (2001) (since) Period Agency Coordination An earlier ‘development plan’, serving as project document, had no formal status; 3 (1934-1968) Reserve Warden – Management Planning in Protected Areas Cameroon This list is limited to protected areas where management planning ongoing; Korup National Park Heritage World 1996-2002 Benoué National Park Biosphere 1998-2002 WWF Project team Zoning in relation Table 9.1 Table Name National SAVANNA Waza International National Park Biosphere Faro Management Plan formulation 1995-1997LOWLAND National Park FOREST Dja IUCN Game Reserve – Main Issues Project team + Biosphere New policy/law 2001 - ongoing User rights 1994 – 2004 WWF Project team Yes Park infrastructure 1 Lobéké National Park – Ongoing WWF Consultants? Zoning – park Yes Ma’an (2000)team Int. Project development, (1999 and 2000); Campo- National Park – 2001-2004 9 – Management planning 209 posed management measures for its conservation. The internally discussed plan sometimes achieved a more formal status when, in addition to instructing the park warden and his/her colleagues, it was used to attract financing (as was the case in Korup NP, see Table 9.1). More recently, management plans have become bulky documents clearly aimed at external constituencies.

Lengthy descriptions in a management plan can mask the nature of its formulation process. Are the proposed management options and subsequent actions the choices of the authors and their superiors, or do they reflect a widespread consensus? Consensus could bring wider support for sensitive management issues. And recognising this, management plans have increasingly been based since the 1990s on the consensual interpretation of pressing management issues, where there is often a conflict of interest among stakeholders. Indeed, the most important product of the modern management plan is a consensus building process based on negotiation among a large number of stakeholders.5 In the case of Waza National Park (NP) (Fig. 9.2), elaborated below, the key conflicts were between local people and park authorities over the illegal exploitation of natural resources inside protected areas.

Structure of the analysis This chapter asks whether the high expectations for management planning as an interface between protected area management and wildlife law and policy are realistic. The analysis first sketches protected area management policy in Cameroon, particularly the limitations of the legal context in which it occurs. This is followed by an account of the formulation process for the Waza NP management plan, a process that has led to a consensus on sensitive key issues among local communities, local and national authorities, and international environmental NGOs. Everyone agrees that this particular exercise in management planning was a learning process for all concerned (MacKinnon et al. 1986.) The analysis goes on to question, however, whether the resulting management plan was sufficiently realistic to lead to its successful implementation. The Waza case is then compared with management planning experiences in other protected areas in Cameroon and in other African countries.

5 In the case of Kruger N.P., sensitive and highly controversial international issues involving elephant culling were involved (Braack 1997). In the case of Hwange NP in Zimbabwe, the key frictions were between tour operators and park authorities (UNDP 1998). 210 Part III – Enhancing conservation – development integration

Figure 9.2 – Map of the Waza-Logone area, Cameroon. 9 – Management planning 211

9.2 Background of protected area management in Cameroon

Introduction Cameroon has lowland and mountainous rainforest in the south and mountains and lowland savannas in the north. Its wide-variety of ethnic groups and cultures lead people to call it a ‘miniature Africa.’ Biogeographic diversity makes Came- roon one of the African countries with the highest number of mammal, bird and plant species (Alpert 1993; Fotso et al. 2001). In the 1960s and 1970s, a strong conservation policy led to the creation of some of the first national parks in West Africa and to the setting up of a widely respected regional wildlife training centre (Scholte 2003b). National parks were well staffed and had reasonable working conditions. Although in the mid-1980s governmental support for conservation diminished in a context of stagnating national economic growth, a Ministry of Environment and Forestry was created in 1993, partly in response to the energy generated by the Rio Summit. This was followed, in 1994, by a new law on the environment designed to facilitate a more participatory style of conservation and, in 1995, by its application decree.

The need for local participation in the development process has been underscored by traditional development organisations, which have spent about ten million US dollars a year on Integrated Conservation and Development Projects (ICDPs) in Cameroon since the early 1990s. The Government of Cameroon has increased the number of protected areas – a favored focus for ICDPs – especially in the rain- forests and montane forests of southern Cameroon (Alpert 1993; Fotso et al. 2001). These changes have substantially complicated protected area management in Cameroon, with a multitude of actors involved in an increasing number of lo- calities.6 The results of all this activity have been mixed. The recently created rainforest protected areas, for example, have had only limited attraction for the tourism and safari hunting businesses, even though their biodiversity is remarkable (Blom 2001; Wilkie & Carpenter 1999ab).

Ambiguous wildlife law The 1978 law on the protection of national parks, though still valid, has been largely forgotten with the introduction of a new series of legal texts in the mid-1990s (Kamto 1996). In the 1994 forestry and wildlife law broad, inclusive purposes were suggested in the first section:

This law and the implementing instruments thereof lay down forestry, wildlife and fisheries regulations with a view to attaining the general objectives of the

6 Cameroon strives to bring 20 per cent of its territory under some form of protection. There are currently ten national parks and seven faunal reserves, including three international Biosphere reserves and one World Heritage Site (see Table 9.1). National parks situated in the savanna zones originated from colonial times and have played an important role as catalysers for tourism (Fotso et al. 2001). 212 Part III – Enhancing conservation – development integration

forestry, wildlife and fisheries policy, within the framework of an integrated management ensuring sustainable conservation and use of the said resources 7 and of the various ecosystems.

In subsequent sections, however, the 1994 law also appeared to lay the basis for a more rigid and centralised approach. In this and other respects the law contained significant ambiguities. ‘The state shall ensure,’ it said, ‘the protection of the forestry, wildlife and fishery heritage’8 and ‘The genetic resources of the national heritage shall belong to the State of Cameroon.’9 One analysis of the process by which the 1994 law was drafted, however, stressed the dominant role of the World Bank, the lack of leadership from the Government of Cameroon, especially from the newly created Ministry of Environment and Forestry (MINEF), and the lack of involvement of other stakeholders (Ekoko 2000).

The 1995 implementation decree, in describing categories of protected areas to be designated and managed under the 1994 law, did not depart substantially from a strict protection regime,10 and was in some respects stricter than the IUCN protected area designations.11 The decree article on resource user rights in protected areas lacked any reference to the grazing of domestic animals, a major land use in Northern Cameroon, and further highlighted a legal bias towards the economi-

7 MINEF 1994: 5. 8 MINEF 1994: 7. 9 MINEF 1994; 7. 10 The decree recognized the following types of areas: Integral ecological reserves (no area has been designed as such); Game Reserves (see Table 9.1, infra); Areas set aside for the conservation, management and propagation of wildlife and for the protection and management of its habitat, where hunting is forbidden, except by authorisation of the minister in charge of wildlife, as part of duly approved management operations, and where dwelling and other human activities are regulated or forbidden; National Parks (see Table 9.1, infra), defined as uninterrupted areas whose fauna, flora, soil, subsoil, atmosphere, waters and natural environment as a whole are of special interest and should be preserved from any natural deterioration and protected against any human interference likely to alter their outlook, composition and evolution [in which respects there shall be taken into consideration the preservation of endangered animals and plant species, as well as habitats, on all or parts of the national territory; the preservation or gradation of major migration routes of wildlife; the scientific or technical studies indispensable for the development of human knowledge; and where there are prohibi- tions on hunting and fishing, except as part of a park management operation; industrial activities; extraction of materials; pollution of any nature; farming, grazing and forestry activities; stray domestic animals; aircraft flying at an altitude of less than 200m; the introduction of local or imported animal and plant species, except for scientific purposes or as part of management operation authorised by the minister in charge of wildlife]; Sanctuary, zoological garden, game-ranch, and game-farming areas; and Buffer zones (infra, section 3.2.4) (MINEF 1995: 120-121). 11 In comparison with the IUCN (1994), categories of protected areas, for example, only the strictly protected areas are represented in the Cameroon decree. And in the case of national parks the prohibition on subsistence resource use is stricter than suggested by IUCN criteria. On the other hand, human habitation is not explicitly excluded in Cameroon national parks. 9 – Management planning 213 cally important forest zone of Southern Cameroon.12 Wildlife in Cameroon is not confined, however, to protected areas and, because of increasing pressures on the remaining habitat, serious resource conflicts can arise in and around human settlements and pastoral camps. The 1990s legal texts are based on the assumption that local wildlife authorities will have a strong presence and adequate intervention capacity, but this is unrealistic given the low number of personnel and limited transport and other equipment. This has made it particularly difficult for farmers and pastoralists to react to frequent crop and livestock damage, provoking bit- ter reactions (Scholte et al. 1999b). The 1995 decree says that ‘Any battue [culling] shall be undertaken by officials of the services in charge of wildlife.’13 But it also says that ‘...no person may be charged with hunting a protected animal if his act was dictated by the urgent need to defend himself, his livestock or his crops. Proof of lawful defence must be given within 72 hours to the nearest official in charge of wildlife.’14

Some of the criticism of the 1994 law and 1995 decree has been directed at their fragmentation, loopholes and inconsistencies (Mayaka 2002). More problematic, however, are the implementation difficulties they pose, because they lack clear guidance on the meaning of crucial concepts, such as community, and because they do not reflect consensus amongst the main stakeholders (Egbe 2001). The more general problem is that these legal texts were drafted in isolation, too far removed from the field and lacking the feedback from field experiences.

Management planning: a way forward? The implementation of a specific ICDP project, like Waza-Logone, cannot correct ambiguities and other weaknesses in national legislation. Carrying out such a project does, however, provide a window of opportunity for making decentralised interpretations and policy decisions under the law. The making and adoption of management plans are especially important sites for this activity, because the law states that ‘Any activity in a state forest shall, in all cases, be carried out in accordance with the management plan.’15

Although two types of plans, for management and development, are recognised in the 1995 decree, the application portion of the decree focuses only on the management plan, which it defines as ‘A technical document drawn up by the service

12 The decree defined user rights as ‘the exploitation of forestry, wildlife and fishery produce by the local population for personal use. However except for game reserves, sanctuaries and buffer zones where they may be authorised, such right shall apply neither to integral ecological reserves, national parks, zoological gardens nor to game-ranches’ (MINEF 1995: 123). 13 MINEF 1995: 127. 14 MINEF 1995: 128. Not surprisingly the difficulty of presenting this proof has frequently created conflict between wildlife authorities and farmers and pastoralists. 15 MINEF 1994: 11. 214 Part III – Enhancing conservation – development integration

in charge of wildlife, or by any other natural person or corporate body commissioned by the service, for the purpose of planning in space and time all strategies to be implemented for the sustainable use of one or several given wildlife resources.’16

The room for manoeuvre in the process of making and adopting a management plan is not entirely clear, however, especially with regard to user rights inside national parks. In what follows I try to answer this question. The first step is to describe the process whereby the first formally approved management plan in Cameroon, for Waza NP, was drafted and adopted. This is then compared with management planning experiences in other protected areas in Cameroon. A further brief comparison with management planning in other African protected areas provides some insight into the extent to which experience in Cameroon offers more general lessons for improving the management of protected areas.

9.3 The Waza-Logone case

Introduction Waza NP is probably the best-known protected area in Cameroon. It receives as many tourists (6000 yr-1) as all other Cameroonian protected areas combined and more than any other protected area in Central Africa. It is situated in the Sahelo- Sudanian Far North Province and covers large stretches of wooded grassland and bushland (Fig. 9.2). The eastern part is characterised by tree-less grasslands and is highly productive, thanks to the annual flooding of the Logone River. Waza’s main attractions are the large Elephant, Giraffe, and Lion populations, as well as high concentrations of antelopes and waterbirds during the dry season. Waza NP was created in 1934 as a game reserve. It became a national park in 1968 and an international biosphere reserve in 1979. Its history is closely linked with the various villages that surround the park, all dating from before its creation. They are composed of ethnic Kotoko fisheries communities, Musgum and Bornouas agri- culturists, and Fulani and Arab pastoralists (Drijver et al. 1995; van Est 1997; Scholte et al. 1996b). At present only one village, Baram, remains inside the park.

The process of developing a management plan for Waza NP was based on previous work, which beginning in 1984 had identified four major problems. The first was that the construction of a dam upstream, on the Logone River, had degraded the environment in the eastern part of the park, causing dramatic reductions in

16 By contrast, a development plan is defined as ‘a technical document drawn up by the service in charge of wildlife, or any natural person of corporate body commissioned by it, which determines in space and time the nature and programme of works and studies to be carried out in a protected area and which shall apply to the said area.’ However, development plans for protected areas managed by individuals may be drawn up by the individuals themselves and approved by the service in charge of wildlife (MINEF 1995: 120). 9 – Management planning 215 wildlife populations and serious declines in fisheries and livestock production (Drij- ver et al. 1995; Scholte et al. 1996b). There were, secondly, severe tensions between national park authorities and neighbouring populations leading to an almost complete lack of collaboration facilitating access to poachers (Scholte et al. 1999b). A third problem stemmed from a reduction in government commitment to park management, which led to deteriorations in vehicle condition, a lack of maintenance for park infrastructure, and low salaries and bleak career prospects for park personnel. The park was also plagued by a legacy of poor decisions involving uncon- trolled bush fires and poor road building and maintenance, which also resulted in unnecessary habitat degradation. This interrelated set of problems was gradually reducing wildlife numbers and yielding low economic benefits from the park to neighbouring populations.

Management plan formulation (1995-1997)

Introduction In 1995, the government of Cameroon delegated the formulation of a management plan for Waza NP to the Waza-Logone Project.17 A core team to co-ordinate the process was formed, comprising the warden of Waza NP, a short-term con- sultant who had worked in the area 25 years ago, the director of the Waza-Logone Project, who had extensive experience in government service, and the author. Additional project teams played important roles in carrying out various surveys, as well as in stimulating discussions with local populations and between the Project and its partners. Local consultancies were arranged to assist with financial, institutional, and legal issues.

To address the four problems noted earlier four main intervention strategies were developed. One was the creation of a park protection and anti-poaching policy. A second sought to improve the technical and ecological basis for park management through better infrastructure and resource monitoring. A third strategy aimed at better integration of development and protection goals, through improved community relations and tourism initiatives. And fourthly, there was a strategy to improve the internal environment of park management by ensuring that the necessary human and material resources would be available. Some of these strategies are relatively straightforward and require no major changes in policy. They essentially depend for their success on the availability and commitment of qualified

17 The decree requires that ‘Development plans,... , shall be rendered enforceable by order of the minister in charge of wildlife... All development plans shall be prepared in accordance with directives issued by the ministry in charge of wildlife, and shall stipulate in particular: a general description of the protected area; the basic objectives to be achieved, taking into consideration the interests of the local populations and the necessity for the conservation of biodiversity; the works to be carried out and the schedule thereof; the cost of the project; and indicators for its follow-up and assessment.’ (MINEF 1995: 127). 216 Part III – Enhancing conservation – development integration

people. The one strategy that clearly challenged existing policy for the park was the third one, analysed more closely below.

A new start As result of the dam and embankment built along the Logone River in 1979 to enable irrigated rice cultivation in the Maga area, upstream of Waza NP (Fig. 9.2), the lack of flooding has had major impacts on the Waza-Logone floodplain ecosystem. Fisheries production has dropped and livestock and wildlife can no longer eat the perennial grasses which used to produce palatable regrowth during the dry season. Floodplain rehabilitation by reflooding has, therefore, been the central activity of the Waza-Logone Project since its inception. The first pilot release of water in 1994 reflooded about 180 km2 and coincided with the start of the management planning process. When the release rehabilitated perennial grasslands and produced more fish (Scholte et al. 1996a), it also created an atmosphere of renewed confidence among local people in their collaboration with outsiders. Increased resource availability, especially on the rangelands, enabled initiatives to exploit alternative resources outside the national park.

Coincident with the floodplain rehabilitation there was also a change in park management. The former warden, who had been in place for over twelve years and had tense relations especially with the Kotoko fishermen communities, had tried several times to remove a village only to be stopped by the traditional authorities. His ethnic background and marriages made him much more aware of the situation in Fulbé and Arab villages than elsewhere. The new warden, by contrast, had an academic background, was young, and was a Kotoko. Local people took this change in manager as a sign of increased tolerance towards their presence, and they immediately asked the new warden for permission to exploit natural resources, such as fish and firewood, within park boundaries.

Intelligence and information (Scholte et al. 1999b) To supplement available information on village-park relations (Njiforti et al. 1991) a systematic survey was undertaken in 1995 in each village, following the methods of Participatory Rural Appraisal (Gueye & Schoonmaker Freudenberger 1991). Complementary studies analysed fishery and pastoral exploitation inside the park, as well as transhumant and nomadic groups originating outside the park villages. The results showed that among a wide variety of problems several were directly related to the existence of the park (Table 9.2). In addition to the lack of flooding, previously mentioned and substantially alleviated by the pilot water release, concern was raised amongst the management plan co-ordination team about the high degree of the dependence of local communities on park resources, especially for fishing and grazing. Were the park really closed for exploitation, villages would not be able to stay where they were. The new information strongly suggested that 9 – Management planning 217

Table 9.2 – Ranking of the ten main problems perceived by Waza NP village communities1

Problem Times the problem has been ranked as

1st 2nd 3rd 4th 5th 6th 7th 8th

Lack of flooding 6 1 1 Lack of formal authorisation to fish in the park 1 1 1 Animal theft 1 1 Lack of grinding mill 1 1 Crop damage 2 1 1 1 2 Lack of pastoral water in dry season 2 1 1 Intimidation by the Waza NP authorities 1 3 Well without water 1 Predation on domestic animals 2 1 1 3 Lack of grazing 2

1Not all figures add up because of the omission of less important problems.

resource exploitation within the park, although formally prohibited, was being tolerated informally by local park personnel in exchange for an illegal ‘tax.’

When the management plan was being put together, Waza NP had about 20 guards. Most of them were recruited in the 1970s and based in the park for years (Fig. 9.3). The warden, who was the only staff member on the core team, felt that the guards were to some extent unaware of the feelings villagers had about them. They did understand, however, the ingenious methods local people used to circumvent the rules, and by not addressing these they were missing an opportunity to create better relationships with local communities.

There are also in the park about 35 tourist guides, who originate from the park villages and are paid by their clients. They are registered at the park office and, apart from their tourist work, are supposed to support actively the surveillance performed by the game guards. In some ways, they have been the true guardians of Waza NP over the years. They have excellent local knowledge and have shown their commitment to the park during confrontations with poachers. They are pursuing, in effect, a participatory conservation approach to managing the park. The survey work showed that their main problem was lack of support from their fellow villagers. The guides, in other words, were the ones most directly confronted by tensions between park and people. 218 Part III – Enhancing conservation – development integration

50 bar: park guards (armed) + village guards 40 line: park guides (unarmed) (unarmed)

20 number of park personnel 10

0 1953 1962 1965 1970 1974 1978 1980 1982 1985 1988 1990 1994 1996 1998 2000 2002

Figuur 9.3 – Development of personnel numbers of Waza NP

Preparing for consensus The process of trying to reach a new consensus on how to manage Waza NP began in February 1996, with a three-day workshop in which selected invitees met to discuss the 1995 survey work. The outcome was a set of preliminary proposals to improve both resource exploitation inside the park and the management of surrounding areas. They formed the basis for further discussions in subsequent months with authorities and local populations.

The workshop agreed that existing park borders should be regarded as inviolable and definitive. They were needed to guarantee conservation values and to reassure local villages that they did not have to fear any further extension of the park.18 The idea of authorising specific resource exploitations inside the park also found agreement, on the basis that it would counter an increase in park exploitations. This would require

18 Park boundaries cannot be changed through a management plan, but widespread suspicion on this subject reigned during the planning process. People’s attitudes towards Waza NP are rooted in the past. The removal of villages, through frequently changing park boundaries, has left deep scars and is still feared. In 1966, one village inside the park, Zeila, was burned down for political as well as park management reasons. Its inhabitants, dispersed over neighbouring villages often on the park borders, continued exploiting its former fishing grounds in the park. During the late 1940s, two other villages, Andirni and Badadaye, shifted a few kilometres to escape being trapped in the park. All villages have seen the park approaching or as someone put it ‘as soon as there are animals you make sure that you leave.’ Fear of the park has been especially acute in the eastern villages, which found themselves right on the park border in the 1950s. In the late 1980s, the park warden caused havoc when he used a road grader to redraw the border of the park, putting many villages inside the park that had not been there before. Although the management plan corrected this situation, local uncertainty can only be alleviated by giving park boundaries a more obvious material presence on the ground. 9 – Management planning 219 finding a way around the prohibition in the 1995 decree on resource exploitation within national parks, and legal advice was sought for this purpose. The third notable outcome of the workshop involved buffer zones.19 Although a buffer zone is itself a protected area and falls under the jurisdiction of the Minister of Environment and Forestry, there is a risk that proposals to create buffer zones will be viewed by local people as a reduction in their ability to control what they think of as their area. The workshop, therefore, proposed that all existing and legal resource exploitation be guaranteed for those living in buffer zones, and for their descendants.

Negotiations In January 1997, after weeks of preparation to help select preliminary representatives from the local population, a meeting was held to set up a management committee. Additional meetings were organised in March 1997 in the provincial capital to decide on the main components of the management plan. After a day on which ministry officials dealt with a number of procedural planning issues, the second day of these meetings constituted in effect the first session of the management committee. The workshop proposals on park resource exploitation proved uncontroversial. The proposals on buffer zones, however, provoked heated discussion and clearly made ministry officials uncomfortable. They did not want to con- firm that present users of resources in potential buffer zones would have rights of access to those resources. It was eventually agreed that the management plan would refer to a periphery zone, a concept not defined in law and, therefore, open to future elaboration. The core team of the Waza-Logone Project, together with a senior staff member of the Ministry of Environment and Forestry, then wrote a draft of a five-year management plan for Waza NP. After some additional discussions at the ministry, the draft was finalised without fundamental changes and was approved by the Minister of Environment and Forestry in September 1997.20

Implementation (1998-2003) and the 2002 evaluation Although the implementation of the management plan, for various reasons, did not get off to a fast start, within a year of its approval a new warden was appointed for Waza NP. He warmly endorsed the plan and took particular steps to see that an outside expert would evaluate it (Amougou 2002).

In line with the intervention strategy for improving internal management, the management plan called for the creation of a sub-unit to pull together and enhance the

19 According to the 1995 decree a buffer zone is ‘A protected belt around each national park, natural reserve or game reserve designed to mark a transition between the above-mentioned areas and the zones where cynegetic, agricultural and other activities are freely carried out. Nevertheless, certain human activities may be regulated therein following a management plan duly approved by the minister in charge of wildlife. The instrument setting up a protected area shall lay down the limits of its buffer zone’ (MINEF 1995: 121). 20 Arrêté No. 732/A/MINEF/DFAP/DAJ du 16 Septembre 1997, rendant exécutoire le Plan Direc- teur d’Aménagement du Parc National de Waza. 220 Part III – Enhancing conservation – development integration

protection, ecological, development integration, and administrative and financial aspects of park management. But circumstances worked against this goal. The recruitment of park guards was frozen following the imposition of International Monetary Fund restrictions on the government.21 The number of staff functions at the central ministry increased during the same period, however. Clearly, field positions did not have high priority. Initiatives to put middle-level technicians at the head of various park units were not followed up, despite financial support for this promised by the Waza-Logone project. So, the park warden remains the only higher-level person on the park staff. And the fact that he is surrounded by an ageing group of park guards seriously compromises the intervention capacity of park management. Since 2001, the recruitment of 600 park guards has been approved in principle, but there has been no implementation.22 In fact, the number of park guards has declined by 50 per cent (Fig. 9.3).23

The management plan also tries to improve the technical and ecological basis for park management decision making. Although elementary ecological and wildlife research and monitoring, set up in the early 1990s, has continued during the implementation phase (Table 9.3), this strategy has also not made much progress. While the various institutions that can and could play a role in research have been identified, their several responsibilities and budget allocations have not been settled. Long term monitoring of park conditions has been neglected and limited follow up means that research and monitoring have had almost no impact on the management of the park (Table 9.3). Zoning, a subject the plan left to be developed during the implementation phase, has had no priority, probably because of a lack of ideas and directives (Table 9.3). A scientific committee, composed of nine Came- roon-based experts,24 advises the plan’s management committee on ecological issues. Its recommendations on the controlled exploitation of park resources attracted some attention. But the scientific committee has met only five times in the management plan implementation period, and this is simply not sufficient for either effective supervision or advice.

The management committee25 has a consultative role on subjects dealing with activities inside the park and a management role in the park’s periphery. It approves the annual plan for the park and brings together representatives of local populations and park authorities, who have, respectively, 15 and 6 voting members. Ad-

21 In 1992, only 5 per cent of MINEF’s total budget was allocated to non-staff expenditures (Brunner & Ekoko 2000). 22 Personal communication from the Director of Wildlife, March 2001 and January 2002. 23 The lack of park guards has led to speeding up the recruitment and training of village guards. But they were never meant to replace park guards. Village guards are not allowed to carry arms and, therefore, can play only a secondary role in anti-poaching campaigns. 24 Including the author. 25 Officially approved by a ministerial decision in April 1998. Decision No. 0728/D/MINEF/DFAP/ DAJ portant organisation et fonctionnement du Comité de concertation/gestion du Parc National de Waza et de sa zone péripherique. 9 – Management planning 221

Table 9.3 – Evaluation of Waza Management Plan’s implementation (Amougou 2002)

Actions prescribed in Realisations % Management Plan (own impression)1 Comments (adapted from Amougou 2002)

1 Protection and anti-poaching control 1.1 Recruitment of park guards 0% (-50%) During the implementation period a 50% drop, see Fig. 9.3 and text 1.2 Recruitment of village guards 100% Initiative of the park warden, yet village guards cannot replace park guards 1.3 Training of guards 100% (30%) Limited, as no new park guards had to be trained 1.4 NP border demarcation 10% A proposal has been drafted 2 (Ecological) management 2.1 Research2 60% Generally ecological research 2.2 Monitoring2 20% (40%) Ecological monitoring mainly, responsibilities of partners involved not clear 2.3 Follow-up Research & Monitoring 10 % Only one study has had an impact on management 2.4 Zoning 0% Not considered a priority 2.5 Fire policy developed 0 % (10%) Some field knowledge being applied 2.6 Infrastructures developed 40% (30%) Especially maintenance of water holes 3 Integration into local development 3.1 Tourism Tourist promotion 70% Brochures, local and international press cover Tourist management 60% (30%) Some ecotourism activities started but failed 3.2 Relations with local communities Committee 90% Functioning, several conventions signed, as yet external investments only Access to park resources 50% (25%) Only partly executed as strongly dependant on presence of park guards (1.1) Local development 70% See Table 9.4 The Baram case 50% A part of the village has voluntarily moved out of (village inside Waza) Waza NP (Chapter 8) 3.3 Awareness creation amongst: Neighboring communities 80% (50%) See Table 9.4 Women 100% (50%) In all structures there is a representation of women Youth 50% Waza NP visited by pupils Private sector 0% (10%) Decision makers 0% 3.4 Other forms of Integration 0% (10%) Considerable publicity in scientific fora

1Note that a score of 50% should be considered as reasonable 2Scores and comments ‘averaged’ over six categories, including vegetation, fish, birds and large mammals. 222 Part III – Enhancing conservation – development integration

ministrative and traditional authorities are present as consultative members. Clusters of 2-3 villages, as well as the transhumant pastoralists, also select one man and one woman as committee representatives. Agropastoralists, whose families are outside the area, select one representative, who was subsequently designated as the first president of the committee, with the park warden as vice-president. Baram, the only village inside the park boundary, was excluded from participation in the committee, on the theory that could ultimately persuade the people there to settle outside the boundaries, in accordance with the official conservation policy.

The creation of this committee has yielded some of the most remarkable impacts that are attributable to the management plan. The committee meets at least once a year and has shown itself on several occasions to be concerned not only with development issues important to the communities it represents but also with the conservation of Waza NP. In 1999, for example, the committee decided to send a delegation to the ministry in Yaoundé, the national capital, to plea for additional park guards. On the other hand, park communities are so diverse that their commitment to the park planning and management has been difficult to secure through their representation by only a limited number of committee members, whose feedback to local communities is necessarily limited (Table 9.4). The most visible impacts of the Waza-Logone Project have clearly come from the pilot reflooding, beginning in 1994. Although these efforts were not extended for financial and technical reasons, the positive correlation between reflooding and the presence and intensity of development activities is clear (Table 9.4). In other places, even small changes in material conditions have yielded results.26

9.4 Management planning in other protected areas in Cameroon

Why management planning? In all the major state protected areas in Cameroon where an ICDP project is being carried out, management planning has started (Table 9.1). In contrast, there are no management planning initiatives in the mountain forest areas in western Cameroon (at Mt. Cameroon, Mt. Kupé, and Mt. Kilum-Ijum), where protection is a matter not for the state but for the community, with ICDP support. At first glance, this suggests that the planning under way in the state protected areas is just part of a management fashion, a nominal response to the legal requirement that each protected area have a management plan. Or do these management planning initiatives, on closer scrutiny, turn out to address real and pressing management needs, as is arguably the case in Waza NP ? The principal issues addressed

26 In the village of Baram, where no development interventions were initiated by design (Table 9.4), one of the families persuaded the others to move out of the village, and out of the park, in 1998 (see Chapter 8). 9 – Management planning 223 2 , A , A S S resources Village created in Village 4 As prescribed in the manage- 3 All None (illegally) access to park leave the national park; 1 + Fish S, F → 3 Agro-pastoralists have a village base, transhumant pas- – Fish S, F 5 – none none – in Waza presently exploited conditional et al. well 1996b) culture Tourist camp Tourist S: straw, A: Arabic Gum, F: Fish; in bold already conditionally active; S: straw, 2 culture non-grazing Insecurity issue non-grazing Insecurity 3 1 + pump Water – 5 5 – Impact of Waza-Logone ICDP activities on park communities (1995-2002) – Impact of Waza-Logone 4 – –: none, slight, +/–: moderate, +: reasonably high, ++: very high; Agro ++pastoralists on +/– Convention Grazing None (Fig. 9.2) none failed continuing committee 1999 with descendants from Baram, subsequently integrated in development activities as well the committee; Halé Table 9.4 Table Village/Community reflooding Ndiguina Impact ofGoulou MbiliBaram – – Development interventions – + X Representation Park resources Apiculture X X Proposed 1 – – Arabic Gum +/– S, A Arabic Gum, fish S, F S toralists not. MahéLougoumaZwangTchede +Diégueré + AndirniAmkodje ++BadadayeNiwadji +Tagawa – townWaza +/– +/– +/– Improved rice pump Water – – – X X X camp Tourist – pump Water X Improved rice camp Tourist +/- well + pump Fish +/– Pastoral +/– Reception hall + Fish S, F Fish + +/- +/– + Grazing +/– S,F S, F +/– Grazing Various Grazing S Grazing Grazing Grazing S, A S S S Transhumant +Transhumant pastoralists ment plan, no development activities nor any representation in the committee were pursued order to dissuade this village (x) (Scholte 224 Part III – Enhancing conservation – development integration

by management planning in Cameroon are zoning and resource exploitation within protected areas (Table 9.1). These are also subjects that generally require a consensual interpretation of legislation and policy. Zoning has received little attention in the defined categories of protected areas in Cameroon, because even a buffer zone is considered a protected area offering few, if any, rights for local communities.

How does management planning in Cameroon take place? International conservation NGOs, such as the World Wildlife Fund (WWF) and the World Conservation Union (IUCN) play an important role in the organisation and facilitation of management planning in Cameroon. They help provide and find project staffs, who are often supported by short-term consultants and who invariably constitute the planning co-ordination teams. The warden of the respective protected area is generally closely associated with these teams but seldom plays a more active co-ordination role, as he did at Waza. In Cameroon, the co-ordination teams are usually constituted informally, as with Waza and Benoué NPs, but in some cases (Korup NP, for example) ministerial approval of the team is pursued. Most planning activities started in the mid-1990s, with three management plans that have been formally approved, Waza, Korup and Benoué NPs. In the Dja game reserve, management planning has dragged on for years, in part due to the multitude of actors involved and also because of the complexity of the area and issues addressed. The Dja case shows that planning can be overly ambitious in scope in relation to the quantity and quality of the human and other resources available.

Participatory studies with local communities (Gueye & Schoonmaker Freudenber- ger 1991) are generally being used as preparation for management planning (Wild & Mutebi 1997; Vabie 2000). Unfortunately, these studies are seldom carried out or supervised by the management planning co-ordination team. The opportunity to establish a dialogue with local communities is, therefore, missed. Furthermore, these studies are not undertaken, overall, on a systematic basis and they neglect evasive communities, such as the nomadic pastoralists in the Benoué and Faro areas.

Review of the management planning outcomes Apart from Waza NP, other protected areas in Cameroon have insufficient experience with management planning to be the basis of detailed analysis. Some useful impressions can be gleaned, however, from Benoué NP, where the second management plan in Cameroon was approved in 2002 and the Dja World Heritage and Biosphere Reserve, which plan was submitted for the minister’s approval in 200427 (Table 9.1). The Benoué plan generally resembles that for Waza, especially in its layout and internal organisation, no doubt because several ministry officials had ear-

27 This section is inspired by a review of the Benoué and Dja management plans by participants (park wardens of Central Africa) of the bushmeat training courses by the Garoua Wildlife College in July 2003 and September 2004 respectively. 9 – Management planning 225

lier participated in the Waza planning. More emphasis is put on zoning, however, which is the main theme in the Benoué plan, and a detailed work plan and budget are included. Proposals for the internal organisation of the park include a doubling of staff, which is even more ambitious than for Waza. But in the monitoring and research section no collaboration mechanism is proposed, even though at Waza this has proven to be a main limitation on implementation. There is to be a committee representing the main stakeholders at Benoué, and a scientific committee is announced. So, overall the main features of management planning at Benoué suggest that there has been a fairly direct transfer of experience at Waza. The Dja management plan is part of a generation of plans in which the central ministry, after the failures of previous versions, has taken the lead during its formulation (Table 9.1). Although previous versions were inspired by the Waza experience, the role of local communities in the consultation fora seems to have been limited to a representative role only. Much of the sensitive zoning issues have, as at Waza, been postponed for further investigations during the implementation of the first plan.

9.5 Management planning in other African countries

Why management planning? On the basis of an unsystematic sample of seven management planning exercises from a range of other African countries some admittedly limited comparisons with the Cameroon experience can be made. Based on an analysis only of the management plan text, none of these other seven management plans is an explicit response to detailed legal requirements, although several interpret the law in specific local contexts on issues such as the exploitation of resources from within the protected area and zoning (Table 9.5). Perhaps the most interesting comparison is with the Réserve Spéciale de Forêt Dense Dzanga Sangha in the Central African Republic, which was created by a specific law, signed by the head of state in 1990,28 and which describes local user rights in a site-specific 1992 ministerial arrêté,29 in much the same way as in management plans in Cameroon.

How did the management planning take place? All seven management plans in one way or another have involved local communities, a distinct difference with earlier practices. This holds for the long established protected areas of Kruger, Hwange, and Serengeti where the involvement of local communities is limited to a consultative role, as well as to some newly created protected areas in which a stronger involvement is sometimes pursued (Table 9.5).

28 Loi No. 90.017 du 29 Décembre 1990. Portant création d’une Réserve Spéciale de Forêt Dense Dzanga-Sangha. 29 Arrêté No. 007 du 25 Mars 1992. Portant règlement Intérieur de la Réserve Spéciale de Forêt Dense Dzanga-Sangha. 226 Part III – Enhancing conservation – development integration uled any impact uctures interpretations consultation experiences park borders Principal themes New policy Proposed Implementation park authorities overr Involvement local Communities / Restitution of mgt plan str a few with others / more or less. reception Tourist not treated attention, justifying planning undertaken never approved) communities / no Coordination Duration plan) and scientists and programmes warden team from the park (1995, none) pygmy communities local communities as positive WH – Management Planning in selected African protected areas BR Table 9.5 Table by Protected Formulationarea Approval previous (year, BR: Biosphere Reserve; WH: World Heritage Site BR: Biosphere Reserve; WH: World Banc d’Arguin WWF with(Mauritania) park teamWH ? Minister (1995, none) ? Framework for more detailed plans No No Claimed as positive Serengeti (Tanzania) IUCN-project team 5 year Minister (4 2 seminars with Livestock / agri- earlier versions representatives of Likely culture in the area NoBoucle du Baoulé NP Consultants(Mali) ? NA, as not Minister approved (1998, none) Zoning and newly No, follow-up No drawn borders ? of new law on Bwindi NP(Uganda) Project team,WH 2,5 year consultants Board / yes Yes Hwange NP (1995, none)(Zimbabwe) Consultants + 1 year park team MinisterKruger NP Zoning, Exploitation(S.Africa) Few Park scientists 3 year several Yes, (1999, 1992) BoardDjoudj NP seminars / yes of NTFP-resources Yes(Senegal) IUCN-project Relations between 5 55 workshop days, tourist agencies and No Negative publicity Elephant culling Minister (1998-1986) No, park No internal mainly, / yes Yes No Fire management Deteriorating poli- around eviction of exploitation Elephant culling Collaboration with tical situation has ? under increased Yes Widely claimed 9 – Management planning 227

Review of the outcome As in Cameroon, all of the management plans reviewed cover a period of five years, except for Banc d’Arguin NP where the period is ten years. Since the making of a plan can take a year or more (Table 9.5) and took more than three years in Came- roon (Table 9.1), a longer time span for the plan seems realistic. This is underscored by the experience at Kruger NP, where the management plan was formed as an annex to a plan prepared twelve years earlier but still took three years to complete. As in Cameroon, most management plans emphasise zoning, which can be a sensitive issue; so sensitive at Waza that zoning was omitted from the final plan document.

None of the management plans reviewed anticipates independent review and only two, Bwindi and the Banc d’Arguin, provide for internal review. This will make it unusually difficult to assess and evaluate plan implementation and is probably a hangover from the time when plans were regarded as essentially technical documents. Anecdotal information suggests that actual experience with implementation of the management plans ranges along the entire spectrum from good to bad.

9.6 Discussion and conclusions

If meaningful and effective, management planning is assumed for purposes of analysis to require success at every phase of the process, the lessons of the experiences reviewed, here, are usefully considered phase by phase. It might be useful to consider management planning as a process in three phases: (1) reaching a consensus on main management issues; (2) ‘writing’ the management plan, based on this consensus and an analysis on less controversial management issues and (3) ensuring the continued commitment of stakeholders during the management plan’s implementation. The success of management planning depends on all three stages.

The process of management plan formulation At Waza, which was already the best known protected area in Cameroon, the additional studies and surveys deemed necessary for the management plan, and the negotiation surrounding these, still took two years. This was a shorter time than was needed elsewhere in Cameroon but longer than in most other African countries (Table 9.5). While additional information and intelligence were being gath- ered, at least one staff member was occupied full-time in co-ordinating project teams and consultants. Regular meetings brought together up to ten governmental staff members and more than fifty local representatives. The costs of making the Waza plan, including studies, meetings, and tours but not salaries and con- sultancy fees were nearly US$ 80,000. This is at the upper end of cost estimates for similar management plans in Australia and the US, which generally did not 228 Part III – Enhancing conservation – development integration

include any base-line studies (ANZECC 2000) but lower than the US$ 250,000 estimated cost of management plans in Africa as a whole (Clarke 1999).

The planning at Waza occurred in quite exceptional circumstances, created by the successful floodplain rehabilitation and the arrival of a new park warden. No changes in associated governmental staff and key contact persons occurred. This was unusual but it did facilitate the learning process and made it productive for plan staff to spend time on study tours for central government officials, who were then encouraged to begin other planning exercises elsewhere in Cameroon. Under less favourable circumstances, a plan preparation period of one to two years would be much too optimistic (Table 9.1). And when planning is under pressure from time, budget, and donor-imposed constraints, third parties may come to dominate the process to the extent that local people and park authorities feel no attachment to either the process or its outcomes.

The first five-year management plan for Waza has, to some extent, been dominated by the Waza-Logone project staff. One important reason for this is imbal- ance in power relations between authorities and local populations. It is not clear how participatory management should be approached in conditions where tension is prevalent and people are reluctant to mention park exploitation. The stronger parties, namely the park authorities and those who have good relations with them, may fear a loss of power. Weaker parties may be reluctant to begin an open process that will change the status quo when the outcome is uncertain. This dilemma played itself out during the buffer zone discussions where, in principle, the Waza-Logone project’s proposal might have facilitated land use control and could have been of mutual benefit. The central authorities, however, would not provide guarantees against misuse, thus making it too risky to rely on project promises. Local people have been able to play a somewhat more active role in park decisions through the management committee, but a co-ordinating third-party agency, like the project, with a combined role of financing and facilitation (Groot 2002) is essential in making this work.

Generally speaking, a park warden is too much of an interested party to be an effective facilitator of the management planning process (Groot 2002). However, park wardens could be more involved than is now the case in Cameroon (Table 9.1), especially if they were taught the requisite skills.30 More could also be done, as suggested by Clarke (1999) to review the successes and failures of park management planning and build these experiences into the wildlife-training curriculum.

30 I attribute their present limited role to a lack of necessary skills and not, as often assumed, to a lack of interest (Chapter 11). The regional wildlife colleges have taught elements of park management planning since the 1980s and recently updated their courses, although facilitation skills may not have received sufficient attention as yet. The Garoua wildlife college recently developed a park management planning refresher course for former students who occupy park warden posts (Scholte 2003b). 9 – Management planning 229

Overall, the Waza management planning process led to a break-through in protected area management in Cameroon. For the first time, park authorities committed themselves to explaining a plan in a committee setting dominated by local populations. Despite legal constraints, procedures for access to park resources were formulated, bringing transparency to an old practice of individual negotiations with park authorities. This has had major implications for other protected areas in Came- roon as well.

Management plan approval authority Although the management plan approval authority in Cameroon is the minister of Environment and Forestry, this has been subject of discussion. In protected areas where the planning process was controversial, such as Dja and Benoué, interventions made by ministries for Livestock, Agriculture, and Mines, for example, contradicted positions taken by the Ministry of Environment and Forestry, prompting the Dja management plan co-ordinator to suggest that approvals should come from the Office of the Prime Minister. This seems certain, however, to further compli- cate what is already a difficult political process. If the view is that management plans are primarily technical documents, approval can more readily occur at lower levels in the ministerial hierarchy, and this may be an option for Cameroon in the future, when approvals deal with second versions of management plans.

Form, presentation and distribution of management plans Management plans still consist mostly of technical and ecological descriptions (Cla- ridge 1999). Opinion is divided about why this is and whether it is a good thing. Description is favoured by external consultants, who would rather not deal with complex policy issues. It also gives plans a scientific aura and, thus, some claim to neutrality as between conflicting interests. But the value of lengthy descriptions in plans has been questioned on the grounds that that protected area managers are more interested in conservation values and the management choices that might help achieve them than in compendia of descriptive information (Claridge 1999). One simple way to accommodate both would be to split planning documents into an annex and a main volume, with the former containing the descriptions and the latter addressing the management and policy issues.

Only the Hwange and Kruger management plans include either a detailed account of how the planning process actually unfolded or a list of who participated in various workshops, stakeholder meetings, and other relevant deliberations. Given the current tendency to emphasise that the quality of management planning is better gauged by looking at the process rather than at the final document, in order to ‘create ownership’, this is a strange omission. Certainly, a wide variety of methods exist to contact and reach out to various publics, all the way from the global con- stituency that cares about Kruger NP, for example, to local communities (Pretty et al. 1995). Participants in management planning exercises ought to take better ad- 230 Part III – Enhancing conservation – development integration

vantage of these methods, whether they involve the internet or more convention- al printed materials.

Management plan implementation Before starting the Waza management plan implementation, the prediction was that ‘Essential for the success of the implementation of the Waza management plan will be that at least some of the results will be put quickly into practice, justifying a continued commitment of local people’ (Scholte 2000b). In fact, there was genuine commitment to the plan by local communities and protected area authorities, and this contrasted with the disappointing roles played by central government and by the international donor community (Table 9.3). But for the governmental partners and other officials involved, management planning is a job, a part of their normal work, for which they are paid. For local people the situation is very different. To participate, they have to sacrifice time and energy that could be spent doing other more immediately productive things, and for them the expectation that planning will lead to improvement in their daily lives is acute. One way to hold people accountable for their adherence to a plan would be to make the plan contractual (Kirkpatrick 2001; Pollack & Horty 1999). But even this could not guarantee success.

The tendency to predict and to try to guarantee the success or failure of management planning by focusing on the form and content of the plans is understand- able (Clarke 1999). But there are some larger contextual variables at work, too. Management planning is one of the weakest links in the conservation chain and the conservation community puts too little emphasis on it. None of the seven management planning formulation processes in Cameroon, for example, has costed more than five per cent of the total budget of conservation projects. This limits the financial impact of their failure, but it also causes major frustration for the local people and private companies who invest their time and energy in the expectation that the conservation community will get behind the implementation of plans in a serious and immediate way.

Clearly, management planning has proven in many cases to be too ambitious, given the availability of human resources, the complexity of problems, and the multiple and to some extent conflicting expectations of local communities, authorities and donors. It might make more sense to focus on removing the most salient obstacles to the day-to-day management of protected areas and only then begin a more comprehensive management planning process (Thomas & Middleton 2003). Cer- tainly, if no obstacles for management exist, or no consensus can be envisaged, management planning should be limited to a simple (internally) formulated working document. 9 – Management planning 231

The mixed results of management planning shows that the implementation trajectory for changed wildlife law and policy is often longer than many anticipated. Offshoots of the new conservation paradigm do have an impact on the ground, but they are often too limited to counter the increasing conflicting interests of conservation and development. This should not deter one from the need to continue with management planning, as this may provide much needed feed-back to national policy and legislation that will be discussed below.

Implications for national policy and legislation

‘Empowering poor people to conduct their own appraisal and analysis and to present their realities is one thing. Whether their voices are heard, understood and acted on is another. There are two weak links: from voice to policy change (policy-in-principle); and from policy change to practice (policy-in-practice)’ (Chambers 1998).

Although most of this chapter deals with policy-in-practice, the question of how local experience with management planning can lead to changes in policy and legislation also deserves some comment.

Policy It is important to remember, despite all the previous criticism of ministry personnel and their poor use of human resources, that a radical change in ideas and perceptions has taken place at the Ministry of Environment and Forestry in Cameroon since the appearance of the 1995 decree (Egbe 2001) Although there was specula- tion that the 1995 decree would bring about a turn towards a new community wildlife policy in Cameroon, even going beyond the obvious implications of the 1994 law, there was at the time very little practical understanding among ministry officials of what this might mean. The idea that policy should involve and encour- age the participation of local communities has now become mainstream thinking, in part because of the exposure ministry officials had to the day-to-day problems of park communities as evidenced during the Waza management planning process. There has even been discussion of creating a distinct unit for the management of protected areas, a sort of parastatal wildlife service that could give local people and officials more independence in day-to-day management decision making than they have, now.

Legislation Experience with the Waza-Lagone Project showed that new national legislation is needed in Cameroon to diversify protected area categories and, thus, to reflect different resource exploitation regimes, especially where there is substantial human presence, as with IUCN categories IV and V (IUCN 1994). This would be helpful in national parks, which now lump together areas that serve various objectives, 232 Part III – Enhancing conservation – development integration

some of them incompatible with those laid down in the 1995 decree. This diver- sification would not yield quick results, but without this legal change only limited further progress with management planning can be made. The legal definition of buffer zones, now seen as extensions of protected areas, also needs to be amend- ed (Neumann 1997; Scholte 2003a)31

The Cameroon experience also has lessons for other African countries. In Gabon, for example, where a network of 13 national parks was created in 2002 (Quammen 2003), problems identified for Cameroon’s protected areas, such as above-mentioned lack of differentiation in status of national parks, are also present. A wildlife decree for Gabon is expected to give management plans a prominent role in decen- tralising national park decision making and include terms of reference for plans that call for consultation with local communities.32

The final point to be made about legislation concerns the incompatibility between the protected area management regime now legally in place in Cameroon and the management scheme envisioned for international biosphere reserves. As previously noted, management planning for Waza, Benoue and Dja biosphere reserves in Cameroon began more than ten years after their designation. Yet, even though biosphere reserves are meant to serve multiple functions33 and are, therefore, unsuitable for management under a single authority striving to meet the requirements of a single type of protected area,34 the 1994 law and the 1995 decree in Cameroon proceeded as if this were not the case. The ambiguous legal status of Cameroon’s national parks vis-à-vis biosphere reserves needs to be resolved. The decennial evaluation, proposed by UNESCO-MAB for all biosphere reserves, but never conducted in Cameroon, could play a role in stimulating this reflection. The experiences with the World Heritage Sites’ periodic review (World Heritage Centre 2002), that encompass the Dja Reserve in Cameroon, shows that feed-back from periodic reporting into national policy and legislation cannot be taken for granted and could be more actively pursued by UNESCO.

31 These changes in national policy need to be accompanied by change at the grass roots level, too, if they are to have tangible impacts. Grass roots initiatives are presently limited, however, by the unwill- ingness of the Ministry of Environment and Forestry to assign more field personnel, one of the principal indicators of the central government’s lack of commitment to management plan implementation. 32 Personal Communication, René Hilaire Adiaheno, secrétaire permanent Conseil National des Parcs Nationaux, Gabon. December 2003. 33 ‘A biosphere reserve should perform three complementary functions: a biodiversity conservation function (with a focus on conserving a representative sample of major ecosystems); a development function (with a focus on humans in the biosphere, emphasising an integrative role for local communities) and a logistical function (combining conservation research, education, training and monitoring). Biosphere Reserves are a special kind of conservation area-traditionally a nested series of zones with differing management intensities (core area, buffer zone and transition area)’ (Bridge- water et al. 1996). 34 ‘It should be noted that a Biosphere Reserve which is equivalent to a single IUCN [protected area] category does not conform fully to the Biosphere Reserve concept. Such cases should promote a re- examination of the Biosphere Reserve to test its effectiveness’ (Bridgewater et al. 1996). 10 Curriculum development at the African Regional Wildlife Colleges, with special reference to the Ecole de Faune (Cameroon)

Paul Scholte

Environmental Conservation 30: 249-258

Abstract Regional colleges in Tanzania, Cameroon and, recently, South Africa have trained some 4000 wildlife managers. Training need assessments called for major curriculum reforms, which were developed and implemented in the late 1990s. This is an analysis of the factors that influenced this curriculum reform in the colleges’ endeavour to respond to new developments in African wildlife management. Since 1979, the curriculum of Garoua Wildlife College, Cameroon, has changed only gradually, whereas work placement subjects, selected by students and their employers, have quickly responded to developments in wildlife management, with an increase in the number of people-oriented subjects amongst other things. In the new curriculum, Garoua’s mid-career students appreciated biology and inventory disciplines for their relevance, as well as courses in other disciplines tailored to conservation practice. The curriculum reforms implemented at Garoua depended on the presence of interested lecturers with an appropriate background, often obtained by additional training. The curricula of the regional wildlife colleges at Garoua (Cameroon), Mweka (Tanzania) and SAWC (South Africa) showed important differences, as a result of regional differences and differing visions of the wildlife management profession. All three colleges have struggled to establish a feed-back system for continuous curriculum review. Increasing the exchanges between the colleges could further develop the curricula. While pursuing necessary changes in curriculum and institutions, care should be taken not to reduce the colleges’ sustainability.

235

10.1 Introduction

European managers ran African protected areas long after independence and the international conservation community was generally reluctant to see the custodi- anship of wildlife transferred into African hands (Robins 1970). The subsequent lack of qualified national wildlife managers in independent Africa drove the establishment of the College of African Wildlife Management in Mweka (Tanzania) in 1963 and the Ecole pour la Formation des Spécialistes de la Faune in Garoua (Came- roon) in 1970. Both colleges had a regional orientation because of the limited size of national markets. In total, the colleges of Mweka and Garoua have trained 2500 and 800 students, respectively, from some 20 African anglophone and francophone countries. They focus on mid-career, medium-level wildlife professionals, and many of their graduates have become protected area or wildlife service managers (Gamassa 1995; Ngog Nje 1995a; Njoya 2001). The Southern African Wildlife Col- lege (SAWC) was founded in 1997, focussing on Southern Africa, including luso- phone Mozambique and Angola (Venter 1995).

National training centres in Botswana, the Central African Republic, Kenya, Nigeria and Zimbabwe (Child 1976) aimed at the low and, occasionally, medium professional levels (i.e. park guards and rangers), but some collapsed due to their dependence on foreign assistance (Ngog Nje 1995b). Although universities also provided courses in ecology and wildlife biology, Mweka, Garoua and SAWC have remained virtually without competition for mid-career professional wildlife management training (Snelson & Lanjouw 1997).

Since the creation of the regional wildlife colleges, major changes have taken place in wildlife management that, apart from ecological and technical management, now also includes communication and community development (Cannon et al. 1996; Saberwal & Kothari 1998; Hulme & Murphree 1999). This new orientation has resulted in increased awareness of the role of women in environmental management (de Bruijn et al. 1997). The private sector has taken over governmental responsibilities especially in Southern Africa (Hulme & Murphree 1999). Armed conflicts in Central and West Africa, the student catchment for Garoua, have resulted in loss of state power, without a private sector or other institutions to take over. New models for conservation and relevant training need to be implemented during such critical periods (Hart & Hart 1997; Shambaugh et al. 2001). As in other professions (Van den Bor 1989), the information technology revolution has helped generate powerful tools such as geographical information technology (GIS) and remote sensing, which few wildlife managers have mastered.

In 1993, Mweka and Garoua were criticised, among others, for being slow in adapt- ing their curricula to meet evolving needs, providing insufficient field practice, having curricula insufficiently responsive to identified needs, failing to evaluate 236 Part III – Enhancing conservation – development integration

training effectiveness and not including a greater representation of ecosystems in training programs (Snelson & Lanjouw 1997). In response to such criticism, clients and experts became involved in curricula reforms that hitherto had been conducted internally. None of the colleges had a feedback system to evaluate and review the curriculum. Participatory curriculum development (Rogers & Taylor 1998) can only be effective with appropriate facilities and qualified and motivated staff to implement reform (Tuntivanich 1989); this required institutional changes when external support for the colleges stagnated.

In this chapter I review developments in African wildlife management and the resulting training requirements and organisation of the colleges. I assess how Ga- roua has adapted its curriculum over the last 20 years and compare its 1996 reform with those of the other wildlife colleges. An analysis of the factors that have influenced curriculum reform and review will show the importance of the organisational and institutional settings of the colleges.

10.2 Methods

I began this study during a two-year assignment at Garoua, holding informal discussions with students and staff, which were presented at feedback sessions. In- formation on SAWC and Mweka was obtained from internal documents and discussed during the SADAC Wildlife Management Training Workshop in Harare, Zimbabwe and at SAWC in November 1999 and Mweka in May 2002.

Training needs and developments in African wildlife management I used the two wildlife training needs assessments conducted in the mid-1990s that covered central, eastern and southern Africa (Pitkin 1995; ULG 1998). Because of the lack of earlier systematic training needs assessments, I analysed 492 work placement subjects. Since 1979, Garoua students have spent six weeks in the field between their first and second year, often at their previous position of employment. Generally, the students’ employers or financing agencies and, to a lesser extent, the students themselves selected these work placements or research projects (Fig. 10.1). The role of the college at Garoua has usually been limited to academic approval. Work placement subjects were categorised into four disciplines, namely biology and ecology, protected area management, tourism and other forms of exploitation and people-oriented subjects, and the Pearson Correlation (SPSS 1999) coefficients were calculated. It was not possible to perform a comparable analysis at Mweka and the recently-created SAWC, because of the limited role of work placements or research projects at these colleges (see later). 10 – Curriculum development 237

20 Percentage of students 10

0 EmployerEmployer + Student Student + College All student college

Figure 10.1 – Who chose students’ work placement subjects in 1998 (as disclosed by students, n=48)

Background to students and colleges Garoua students indicated their prior work and training experience as part of the evaluation. No systematic information was available from Mweka or SAWC. Tests of normality were carried out with Kolmogorov-Smirnov tests (with Lillefors significance correlation).

Garoua’s curriculum development and students’ perception To study Garoua’s curriculum development, its >40 courses were categorised into the ten disciplines, including work placements, commonly used at Garoua.

An anonymous evaluation of students’ perceptions was introduced to provide feedback on Garoua’s 1997 curriculum. The 31 diploma and 18 certificate students judged the relevance of the courses’ contents, their presentation and the balance between theory and practice on a scale of 1-5. Scores were averaged and ranked for each course. I also allocated the scores to the disciplines and subsequently ranked them. I was unable to obtain data on the perceptions of Mweka or SAWC students for comparison (see later).

Curriculum comparison between the regional colleges Information on the curricula of Mweka and SAWC was obtained from the 2000- 2002 course prospectuses and syllabi and categorised according to the disciplines applied at Garoua. The number of hours each subject occupied in the curriculum, the main criterion in weighing student results at the three colleges, was used as comparison. 238 Part III – Enhancing conservation – development integration

10.3 Results

African wildlife managers’ job requirements and training needs Graduation from the wildlife colleges has become an important criterion for promotion to medium level staff functions. Whereas in the past certificate students accounted for most enrolments, their numbers at Garoua and Mweka, and recently at SAWC, have dropped (Table 10.1), suggesting an increase in job requirements. In Cameroon, Niger, Tanzania, Uganda and elsewhere, some of the more presti- gious national parks have recruited M.Sc. graduates as park wardens, thus bypass- ing the wildlife colleges. Only Mweka delivers advanced diploma and post-graduate courses, but struggles with their accreditation.

Relatively few wildlife college graduates continued advanced training. Out of a sample of 221 who received training at Mweka from 1989 to 1999, 29 continued their training after graduation. However, more than 50% of the 221 expressed a desire to continue to study for a B.Sc. or M.Sc. in order to improve career advancement options (Lipya 2000).

Table 10.1 – Student numbers at the African regional wildlife colleges.

Year Garoua Mweka SAWC

Certificate Diploma Certificate Diploma Certificate Diploma Male Female Male Female Male Female Male Female Male Female Male Female

1998 14 4 27 4 432 492 25 2 0 0 19991 00 00 222 522 22 2 12 1 2000 11 4 21 5 432 552 15 1 18 1 20011 00 00 542 27 6 nd nd nd nd 2002 8 4 26 6 37 4 34 2 18 8 17 5 % 732783179010881286148713

1Garoua had a biannual intake: 2Male and female numbers not specified. nd not determined.

Developments in wildlife management reflected by work placement subjects Amongst the work placements, there has been a steady increase in people-oriented conservation subjects between 1980 to 2000, their frequency being negatively correlated (p = 0.006) with the number of biology and ecology subjects (Fig. 10.2). The number of protected area management subjects received a largely constant level of attention (Fig. 10.2). 10 – Curriculum development 239

Biology & ecology [1,2]

People-oriented [8] 80 Protected area management [3,4,6,9]

Tourism & exploitation [5,7]

20 Percentage of work placements

0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000

Year

Figure 10.2 – Subject areas of Garoua work placements, 1980-2000. Numbers in parentheses correspond to the disciplines on the x-axes of Figures 3 and 4.

Garoua’s work placement subjects have been responsive to topical developments in wildlife management. An example is the rinderpest outbreak in 1983 that received considerable attention amongst the 1984 subjects, contributing to the peak of interest in biology and ecology (Fig. 10.2). People-oriented subjects peaked in 1994 (after the United Nations Conference on Environment and Development) and subsequently decreased. On average 47%, and never less than 30%, of the work placement subjects dealt with tourism and exploitation, and people-oriented conservation.

Background of colleges

Students With on average approximately 10 years prior work experience, Garoua students have been professionals, working in protected areas or forestry or agricultural services (Table 10.2). Mweka and SAWC have also attracted an increasing number of students from community organisations, and hunting and tour operators, while some students have been self-financed.

Since the late 1980s, there has been a steady increase in the number of female students. The preferential scholarship policy at Garoua has raised the number of female students to 25%, double that of Mweka and SAWC (Table 10.1). 240 Part III – Enhancing conservation – development integration

Table 10.2 – Prior education and work experience (years ± 95% confidence interval) of Garoua students between 1997 and 1999 (BAC roughly equivalent to English A-levels, BEPC is roughly equivalent to English O-levels/GCSE).

Experience Average Diploma Certificate

Prior education Median BAC + 2.0a BEPC+ 2.0b Prior field work Mean 6.4e ± 1.7 6.7e ± 2.4 Prior office work Median 1.0c 5.0d Prior total work Mean 9.2f ± 1.7 12.3g ± 2.4

Comparisons between diploma and certificate students are given by a different superscript: p < 0.05 by Mann-Whitney (a-d) or independent t-test (e-g) (SPPS 1999).

Trainers Five of Garoua’s nine permanent training staff had university training in biology or agronomy; the others were Garoua graduates. Increasingly, temporary staff are hired, while expatriate staff have declined from an initial majority to none. Most of the staff used to have a management or scientific background, whereas the current staff members have spent their entire career at the college. At its establishment, most Mweka staff had a B.Sc. and there were few diploma holders. At present, all but one of the 14 staff at Mweka hold M.Sc. degrees. SAWC has four permanent staff to ensure the functioning of the college, drawing on a variety of subject specialists for training.

Facilities and financing All students have access to computers, but it has been difficult to keep track of the latest developments, such as in the Internet, at Garoua. Garoua and Mweka spend periods of up to two weeks in the field, requiring important investments in cars, camping gear and other field equipment. Day trips dominate at the SAWC, which is surrounded by Kruger NP. Garoua and Mweka manage nearby savannah and mountain rainforest reserves for fieldwork.

Garoua has received significant Cameroonian government financing, albeit on a lower level than in the 1980s, ensuring the salaries of its personnel. Tuition fees cover most of the running costs, including the fees for temporary trainers, whereas donors have provided investments and staff training. The annual costs of a Garoua student is estimated to be € 8800, including scholarships (€ 4600) and annual tuition fees (€ 2100) (Njoya 2001).

Although a government structure, Mweka does not receive government financing for recurrent activities and depends on tuition fees (US$ 3125 yr-1 and US$ 6000 yr-1 for Tanzanian and foreign students, respectively) and project support. The 10 – Curriculum development 241

SAWC is a private institution and tuition fees (US$ 4500 yr-1) and company and donor grants assure its continued functioning and cover investments.

Curriculum reform and review Three internal curriculum reforms have taken place since Garoua’s establishment, namely in 1977, when certificate and diploma courses increased from one to two years in duration (Allo 1978), 1981 and 1989. The 1996-1997 reform, which led to a modular programme, was based on a five-day workshop with a variety of participants, including three non-Cameroonian graduates, followed by a workshop with donors and ministerial personnel (Table 10.3 and 10.4). Garoua staff produced syllabi and began curriculum implementation in September 1997.

Table 10.3 – Curriculum reform and ongoing review at the African wildlife colleges in the 1990s

Mweka Garoua SAWC

PROCESS • Starting base White blanket Existing curriculum White blanket for reform • Reform methods Workshops, Technical workshop Consultancies consultancies PR follow-up Workshop workshop • Period of reform 1993–1997 1996-1997 1994-1997 • Frequency of review Annually Bi-annually ?

FEEDBACK • Course evaluation: Trainee Ineffective (lack Yes Initially by each of confidence) trainer, since recently centrally Faculty Yes Yes Limited Outsiders Yes No Yes • Tracer studies 2 limited studies Ongoing (n>>100) Results not yet (n=55, 18) available

Mweka has gone through six curriculum reforms (Snelson 1993), driven mainly by consultants. The 1990s reform took five years as it deliberately ignored the existing curriculum (Table 10.3). Although facilitated by Mweka staff, substantial technical assistance was delivered through workshops and the production of syllabi in 2001. The production of course manuals is, as at Garoua, still underway. 242 Part III – Enhancing conservation – development integration

Table 10.4 – Participants of curriculum reform and review at the African wildlife colleges.

Mweka Garoua SAWC

Reform Review Reform Review Envisaged1 Initiation2 Review

(Potential) trainees + – – – – + + – – – – Protected Area managers ++ + + – + ? + Local communities / – – – – – – – – – – – – Development NGOs Clients: Governmental ++ + + – – + ? + Private sector + + – – – – + + Conservation NGOs + – + – – + + ++ Trainers, inside college ++ ++ ++ ++ + – – Trainers, outside college + + + – – + + + Scientists + + + – – + + +

Presence of participants: ++ = majority, + = one-several, – = occasionally one, – – = none. ?= unknown. 1Pending a presidential decree, (re-) institutionalising the college; 2due to the recent start of the college

The SAWC leaves much initiative to its visiting expert staff that, not bound to syllabi, have had the ability to develop their own courses for which they generally produce manuals.

Recent curriculum review at Garoua has been driven mainly by student feedback (see below), reported to a commission of training staff (Table 10.3). A similar commission deals with course development at Mweka, which has recently installed an external curriculum review commission like that at SAWC (Table 10.3).

Developments in Garoua’s curriculum Changes from 1979 to 1997 in Garoua’s curriculum have been gradual (Fig. 10.3). One of the few changes was the 1981 introduction of ornithology, a biology subject that has remained a major course, possibly because of the enthusiasm of the trainer. In 1997, veterinary techniques (biology), a prominent course due to the presence of French veterinary assistance until 1996, was reduced. In 1997, computer science (administration) and GIS (inventories) were introduced, as well as participatory approaches and environmental education, thus tripling the importance of socio-economics. 10 – Curriculum development 243

40 1979 1981 1989 1997 30

10 Percentage of curriculum 0 t s g s g s n nt gy gy e ic in ic o e o o ri m n t m ti iol o a la nt tat h un is e tourism c h g . B Ecol v & e lacem 1 2 n m le p . I s s & & k 3 . Habi rk & io-econo r 4 a g rm c in o m Wo . P . A d 5 7 . S 0. eerin 8 . A 1 n 9 gi n . E 6 Figure 10.3 – Curriculum subject composition of the diploma course at Garoua Wildlife College, 1979-1997.

Feedback on the new Garoua curriculum The perceived relevance of courses at Garoua has been remarkably consistent but varies between disciplines (Table 10.5). Courses on biology, inventories, parks and tourism, and arms and hunting were considered to have above-average relevance (Table 10.5). Surprisingly, courses in the field of ecology and habitat management were ranked lower. New courses ranked from very high (for example participatory approaches: socio-economics), to very low (for example remote sensing and GIS: inventories). The relevance of the extensive ornithology course (biology), a topic that received far less attention at Mweka and SAWC, was considered above average.

Students and training staff alike expressed their interest in field trips and exercises as the most effective type of training, addressing skills and complementary to the predominantly knowledge-based classroom training. Moreover, fieldwork (several weeks of bush experience) was considered unsurpassed in developing professional attitudes by students at all three colleges.

Curriculum comparison between the three African wildlife colleges The curricula of the two-year certificate and diploma courses were similar at Garoua (Fig. 10.4). Mweka and the SAWC focussed their one-year certificate course on traditional wildlife management (inventories, mechanics) and their diploma courses of two years and one year, respectively, on socio-economics and administrative subjects. This contrasted with the substantial prior office experience of certificate stu- 244 Part III – Enhancing conservation – development integration

Table 10.5 – Relevance ranking of disciplines by Garoua students (1997-1999).

Rank Diploma students Mean1,2 Certificate students Mean2,3

1 Parks & tourism 4.49a Inventories & wildlife 4.37a management 2 Biology 4.41a Arms & hunting 4.28a 3 Inventories & wildlife 4.39a Biology 4.26a management 4 Arms & hunting 4.38a Parks & tourism 4.25a 5 Habitat management 4.20ab Administration & legislation 4.18a 6 Socio-economics 4.19ab Socio-economics 4.07ab 7 Ecology 4.14bc Ecology 4.09b 8 Administration & legislation 4.08c Habitat management 3.93b 9 Engineering & mechanics 3.91cd Engineering & mechanics 3.54c

1Median of rank 1 and 2 is 5; rank 3 is 4.25, rank 4-9 is 4; 2Comparison of medians amongst disciplines, different 3 letters in superscript indicate p < 0.05 (Mann-Whitney); For all ranks median is 4.

dents at Garoua (Table 10.2). Differentiation between certificate and diploma courses enabled students to start at certificate level and continue the following year with the diploma course. Although possible, this has created considerable overlap at Garoua.

Biology and ecology received ample attention at Garoua and at Mweka (Fig. 10.4). The lack of these disciplines at diploma level at SAWC was striking, given the preference for biology exhibited by Garoua students (Table 10.5). Inventories and wildlife management and habitat management received attention in all curricula. Parks and tourism were dealt with at all colleges and contributed to a third of the diploma curriculum at the SAWC. Garoua and Mweka scarcely covered engineering and mechanics, nor did the SAWC diploma course, a development consistent with the low appreciation of these subjects by Garoua students (Table 10.5). Socio-economics received attention in all curricula, with courses at SAWC and the diploma course at Mweka spending up to 20% of the time on this area. The SAWC further spent an important part of its curriculum on administration.

10.4 Discussion

The wildlife colleges have made considerable internal efforts to keep pace with the new developments in wildlife management. Do these efforts result in a curriculum that meets the demands of wildlife practice? 10 – Curriculum development 245

Garoua/C Garoua/D Mweka/C Mweka/D SAWC/C SAWC/D 40

10 Percentage of curriculum 0

y t t s g s n gy g g c c o nt o og m i tin i ti e l m n m a iol o at a un l c ife t tourism h h is . B dl c g 1 . E l & e lacem 2 s m s & le p wi . Habi o-econo & k & rk & rm i r s 4 a g c in e o m Wo ri . P . A d o 5 7 . S 0. nt eerin 8 . A 1 e n 9 v gi n n . I . E 3 6

Garoua/C=Garoua, certificate course. Garoua/D = Garoua, diploma course. Mweka/C= Mweka, certificate course. Mweka/D = Mweka, diploma course. SAWC/C= SAWC, certificate course. SAWC/D= SAWC, diploma course.

Figure 10.4 – Curriculum subject comparison of the three African wildlife colleges.

Continuing discrepancy between curricula and professional requirements There are different requirements for a two-year curriculum and a six-week work placement and these cannot be translated automatically into developments of job requirements. For example, protected area management subjects have received less attention in work placements than in Garoua’s curriculum, yet are expected to occupy a major share of a wildlife manager’s time. Tourism and exploitation and people-oriented conservation have received more than twice as much attention in work placements as in the curriculum. More important than this absolute difference was the observed time lag in attention. Whereas in 1994 people-oriented subjects peaked amongst work placements, their subsequent decline coincided with an increase in attention in the curriculum, a lag of five years.

Three hundred interviews with protected area managers across eastern, central and southern Africa, conducted in the early 1990s, identified training needs in three job responsibilities, namely (1) implementing intervention programs, (2) ensuring visitor satisfaction and (3) promoting conservation in local communities. Managers further expressed the need for (4) knowledge-oriented training in policies and procedures, planning and administration, as well as (5) skills-oriented training in cre- ativity, problem analysis and evaluation techniques (Pitkin 1995). A training needs 246 Part III – Enhancing conservation – development integration

assessment in Southern Africa concluded that the wildlife colleges remain ‘by their nature’ weak in providing skills and attitudes especially in administration, project and business management (need 4) and community wildlife management (need 3) (ULG 1998).

These training needs were already addressed in the curricula, needs (1) and (2) in inventories, habitat management, and parks and tourism (Fig. 10.3), introduced or reinforced during the recent reform, such as needs (3) and (4) in socio-economics, and administration and legislation (Figs. 10.3, 10.4), or have remained largely un- covered (need 5), with the possible exception of the SAWC. In the early and mid 1990s, a reported general lack of skills amongst graduates of Garoua and Mweka (Pitkin 1995; Snelson & Lanjouw 1997; ULG 1998), explained the continuing need for training in responsibilities that were already addressed by the curricula (needs 1 and 2), but probably too theoretically. This was in line with the evaluation by Garoua students in 1998-2000, showing that training after the curriculum reform was still too knowledge-oriented.

Curriculum reform and review The differences between the curricula (Fig. 10.4) can be partly explained by the reform processes. The short intensive reform at Garoua with a limited number of participants (Table 10.4) who were mostly familiar with the old curriculum, led to gradual changes imprinted by the training staff, as reflected by the reported course introduction. The elaborate reform process at Mweka (Table 10.3) led to a break with the previous curriculum, as illustrated by the new one–year only certificate course. The newly-developed focus on the wildlife profession probably played a critical role.

None of the colleges had a feedback system in place prior to the curriculum reform. At Garoua, feedback from students with an average of 9-12 years’ work experience (Table 10.2) was introduced in 1998 and considered an important indication of quality. Despite sometimes-harsh judgements that could have provoked resistance (Tuntivanich 1989), trainers used the subsequent evaluations to monitor their functioning and impact. This has also led to a reduction in the number of temporary trainers and adjustments in overlapping courses, but not to changes in the course syllabi. The external review commissions at Mweka and the SAWC have not yet had an impact on the curriculum.

All three colleges recognised the potential of communicating with their alumni and started tracer studies (Table 10.3); however, communication and methodological problems prevented their feedback into the curriculum. 10 – Curriculum development 247

Training and learning The three colleges have the distinct advantage of attracting experienced students, who are motivated to learn and, with the exception of Cameroonian students, will be rewarded with promotion once back home. Trainers with relatively little field exposure and little experience with participatory training (Moss 1991) may have difficulties using students’ prior knowledge (Dochy et al. 1997).

Socio-economic and administrative courses should especially be tailored to wildlife practice. These courses are not automatically in the interest of students, who are after all wildlife managers and not human-resource or business managers, and therefore demand specialised trainers. The student feedback highlighted the discrepancy in socio-economics where the course participatory approaches was ranked fourth in relevance, while the more general economics course ranked only forty- third.

Curriculum reform has been hampered by the difficulty of introducing new ideas in an environment where students and trainers will not recognise them and hardly ever receive incentives for somewhat-controversial knowledge. An example is the rejection of the transfer of technology paradigm (Chambers & Jiggins 1987) in the Garoua participatory approaches course, whereas extension services, considered to be subject specialists, were still working with it. The SAWC has the distinct advantage of being located in South Africa, where it is in the forefront of developments in wildlife management.

Field work and exercises Assignments and field exercises, appreciated for their effectiveness (see also Galin- do-Leal 2001) were programmed to occupy a quarter of the time at Garoua and slightly more at Mweka (Table 10.6). In practice they hardly ever attained this as Garoua students noted, mainly due to planning problems, and not for financial reasons as Garoua staff sometimes claimed. An increasing problem was the dis- integrating protected area system, especially in Cameroon, depriving Garoua of prime wildlife conservation field examples (Alpert 1993; Scholte 2003a).

Modularisation and short courses Modularisation has been presented as an instrument to facilitate a transition from a classical towards a more flexible education system (Meel 1997) and the introduction of a modular system brought advantages such as more focused learning and training at the three colleges (Banham 1999). Some expected results, such as increased exchanges between lecturers were, however, a function of changes in the colleges’ management. At all three colleges, little use was made of the ability to combine long and short courses. Modularization did, however, allow trainers to be involved in additional research, consultancies and short-course training. 248 Part III – Enhancing conservation – development integration

Table 10.6 – Type of education at the three wildlife colleges.

Garoua Mweka SAWC

Certificate Diploma Certificate Diploma Certificate Diploma

Course length (hrs) (years) 2200 (2) 2700 (2) 1152 (1) 2352 (2) 1272 (1) 1314 (1) Long field trips (%) 18 15 27 20 0 3 Exercises/ Assignments (%) 25 23 291 291 ?3 ?3 Lectures (%) 39 47 442 442 ?3 ?3 Work placement / Research period (%) 18 15 0 6 0 19

1Includes student presentations and directed study; 2Includes seminars and self-directed study; 3Not centrally fixed, dependant on trainer.

Since the mid-1980s, Garoua has organised short courses on wetland management that were initially externally financed and attended by both outsiders and students. These stimulated the 1996 modularization based on ecosystems and were subsequently integrated as modules in the curriculum. Mweka and the SAWC organised short courses more regularly than Garoua, providing the well-situated SAWC in particular with regular income. Moreover, short courses allowed the colleges to keep in touch with senior wildlife managers (Pitkin 1995; Stone 1997).

Science bias The comparison between work-placement subjects and curricula indicated a bias at Garoua towards biological sciences, whereas employers were of more diverse backgrounds. Science bias is widespread in conservation training (Cannon et al. 1996) and an argument for including people from various disciplines and interest groups in curriculum reform (Rogers & Taylor 1998). Stakeholders absent from the Garoua 1996 reform were development organisations and local people (Taylor 1997), whereas Garoua trainers dominated in numbers (Table 10.4). With a comparable background Mweka staff probably also had a science bias, but they involved more outsiders in curriculum reform and review (Tables 10.3 and 10.4). The SAWC with its temporary staff has had the flexibility to draw trainers from other disciplines.

Personnel aspects Given the small size of the colleges, the staff have inevitably placed a personal imprint on the curriculum and this is to be applauded when the result of initiatives of dedicated staff. At Garoua, trainers with what at first sight appeared to be unsuitable professional backgrounds developed several new courses. Examples were a veterinary trainer who attended a six-week participatory rural appraisal (PRA) 10 – Curriculum development 249

training course and a wildlife utilisation trainer who attended a 9-month remote sensing/GIS course; both subsequently introduced these topics at Garoua. Trai- ners at Garoua attended courses and workshops organised by international NGOs and there has also been specific funding to Garoua for such training. Expatriate staff at Garoua have sometimes set important directions, such as on rainforest and wetlands management, which have been continued by counterpart staff after their departure. Subject-specialist trainers have had the freedom to develop their own courses with their own personal imprint at the SAWC.

Some pronounced personnel impacts on the curricula, however, were due to institutional deficiencies. In Garoua’s curriculum, training personnel could be held responsible for several deplored developments, and the English language course was abandoned with the departure of the anglophone director. That choice of specific trainer may have an impact beyond presentation was illustrated by the differences in perceived relevance between diploma and certificate courses that could sometimes be related to the trainer concerned (for example, habitat management, Table 10.5). Garoua and Mweka have been linked to their wildlife ministries and these should have allowed frequent rotation of personnel, maintaining contacts with wildlife management practice. Unfortunately, two-thirds of the training staff have been at Garoua and Mweka for more than 15 years, probably because of the personal preferences of trainers, low status of the colleges and the lack of appropriate human resource policies at the ministries and colleges. Given their regional orientation, it is regrettable that none of the colleges has employed trainers from within their regions.

10.5 Conclusions

Curriculum reform, especially at Garoua, has been related with personnel management, in particular the limited recruitment of trainers with a practical wildlife management background and interest, and subsequent training and follow-up. The 1996 Garoua curriculum reform workshop did not institute an external follow-up, nor has Mweka’s reform been effective in this regard. Curriculum reform should, however, be a regular phenomenon and include representatives of development agencies and local people (Taylor 1997). Students’ feedback on the curriculum has already played an important role at Garoua.

Obviously there is no ‘best’ curriculum and opinions will always differ on the ideal composition of any curriculum. Nevertheless, the differences between the curricula with similar target groups are remarkable. To some extent this may reflect regional differences as indicated by different views of the wildlife profession, such as the ‘wildlife MBA’ diploma course at the SAWC. It would therefore be interesting to compare the reactions of students to their respective curricula among the 250 Part III – Enhancing conservation – development integration

three colleges. Interest has been shown in stimulating such exchanges (Kanya- mibwa 1999), although as yet not on a structural base. This may also result in student exchanges for a specific module such as private wildlife management at SAWC and lowland rainforest wildlife management at Garoua.

The regional wildlife colleges have been pressured to orient training towards pres- tigious university level, as in their short courses. Collaboration with universities could avoid the accreditation difficulties experienced by Mweka and enable the small colleges to concentrate on their practical training niche.

Despite the discussed drawbacks, Garoua and Mweka have shown by their continued existence to be sustainable, an achievement given the harsh institutional environment (Ebohon et al. 1997) that has led to the virtual collapse of the university system in Cameroon and other countries. While pursuing changes in curriculum and the colleges’ structures, care should be taken not to jeopardise their sustainability. Donor fatigue in supporting the colleges is not entirely justified given their regional role and, in the case of Garoua, continuous support from the host government. Outside support has been crucial, not only as a source of investment, but also through staff training as an impetus for the implementation of curriculum reform. This has led to the education of several generations of African wildlife managers, including women, at a third of the costs in Europe (Van den Bor 1989). Future conservation success in Africa will depend to a large extent on the capacity and dedication of wildlife college students, as it has in the past. 11 Wildlife Managers’ Perceptions of Community Conservation Training in West and Central Africa

Paul Scholte, Wouter T. de Groot, Zacharie Mayna and Talla

Environmental Conservation, in press

Abstract Training need assessments for African protected area managers have revealed the need for people-oriented training to increase their job performance. The Garoua Wildlife College, Cameroon, developed long (diploma and certificate) and refresher courses in community conservation for mid-career protected area managers and park guards from West and Central Africa in the late 1990s. Through lectures, case studies and Participatory Rural Appraisal exercises, the courses emphasised the development of skills for tuning principles of people participation to the conservation objectives of protected areas. The present study reviews the trainees’ evaluations of these courses to appreciate their relevance and support their further development. Diploma students judged the courses as highly relevant because of the acquired analytical skills, whereas certificate students considered them only of medium relevance. The reaction on short refresher courses varied as a function of the use of cases from either the trainees’ professional experience or from the fieldwork location. The reactions of trainees to this learning opportunity suggest that protected area personnel are not ‘attitude-limited’ as often suggested. Their constraints to develop more people-oriented work style lie largely in the areas of knowledge and skills. These findings motivate increased efforts to implement training of protected area personnel in community conservation, preferably early in their careers.

253

11.1 Introduction

Community conservation, i.e. the notion that conservation cannot and should not be pursued against the interests and wishes of local people, has received increasing attention from researchers, authorities, development and conservation agencies and local community organisations (Adams & Hulme 2001; Hulme & Murphree 1999). Bruner et al. (2001) and Vanclay (2001) showed the importance of the number of park guards and their awareness building activities in local communities for the effectiveness of protected areas. Protected area personnel, i.e. wardens and park guards, have seldom been involved in the development of community conservation, however, in spite of its large international funding (El-Ashry 2001).

Training needs assessments for Eastern and Southern Africa (ULG 1998) and the African continent (Pitkin 1995) have highlighted the need to train protected area managers in human-related subjects, see also Western (2003). Also in India and the USA the need to include the human dimension in conservation education has been stressed (Saberwal & Kothari 1996; Jacobson & McDuff 1998). The African training needs assessments have remained, unfortunately, without further speci- fication, e.g. stressing a ‘need for all aspects of Community Based Natural Resource Management training’ (ULG 1998). In some publications, emphasis is put on required changes in attitudes of protected area personnel (IIED 1994).

Existing training courses in participatory natural resource management (Borrini- Feyerabend et al. 2000; Nguinguiri 2001) are based on insights from participatory rural development (Pretty et al. 1995) that do not take potential frictions between conservation and short-term development objectives into account (Oates 1999; Scholte 2003a). For example, protected area personnel also have a policing role, rendering their position towards local communities sometimes ambiguous, which requires special skills and attitudes. Another difference with participatory rural development is that protected area personnel are evaluated on overall conservation results. From their perspective therefore, community conservation tends to be seen as a means rather than a goal.

This chapter reviews the evaluation by the trainees of the first long and refresher courses in community conservation for West and Central African protected area personnel given by the basic training college at Garoua (Cameroon). A comparison will be made with courses at the College of African Wildlife Management at Mweka (Tanzania) and the Southern African Wildlife College, SAWC (South- Africa), that train mid-career medium level wildlife professionals from especially Eastern and Southern Africa respectively. This study should enhance an appreciation of the relevance of training in community conservation and support its further development. 254 Part III – Enhancing conservation – development integration

11.2 Garoua college and the community conversation courses

The Garoua Wildlife College and its students The Garoua Wildlife College or simply ‘Garoua’ is one of the three regional Afri- can wildlife colleges that train mid-career protected area managers (Scholte 2003b). Founded in 1970, Garoua draws its students from twenty francophone, mainly West and Central African countries. Garoua provides courses of two years on Cer- tificate and Diploma levels, with entry requirements BEPC (‘O-levels’) and BAC (‘A-levels’) respectively, and at least three years of working experience (Table 11.1). Many park wardens in West-Central Africa are Garoua diploma graduates, whereas certificate graduates occupy medium-management positions such as assistant warden.

Garoua’s curriculum has been dominated by courses on species identification and the functioning of ecosystems, while some attention has been paid to tourism and sport hunting. Prior to 1996, hardly any attention was paid to the role of local people in wildlife exploitation and protected area management, apart from incidental theoretical courses on sociology and extension science (Scholte 2003b). Garoua is not an exception; regional colleges in India (Saberwal & Kothari 1996) and Tan- zania (Snelson 1993) shared, till recently, this almost complete focus on bio-ecological wildlife management. An external review carried out in 1996, stressed the introduction of a community conservation course and an update of courses on park management planning and environmental education. A total of about 300 hours, or 10 percent of Garoua’s curriculum, has now become people-oriented. In addition, 40% of the students presently choose a people-oriented subject for their practical period (Scholte 2003b). With the curriculum reform also the need was identified to upgrade skills of former Garoua students who are active as protected area managers.

Community conservation courses Protected area personnel are often considered insensitive to the needs of local communities (IIED 1994), resulting in a lack of involvement and difficult implementation of Integrated Conservation and Development Projects (ICDPs; Scholte et al. 1999b). The goal for the short and long courses was therefore to raise awareness and increase capacities for an atmosphere of transparency and mutual trust between protected area managers, local communities and conservation-development project personnel (Table 11.2).

Long courses for mid-career protected area managers The average professional experience of Garoua students is over ten years (Table 11.1). Such prior professional experience has been argued to be the most important variable influencing learning results (Dochy et al. 1997), and calls for the active participation of students that often have more field experience than the trainers. 11 – Community Conservation Training 255 2 in community conservation In other profession: 20% In other profession: 18% 1 Prior work Prior experience From 17 francophone West-Central African countries; From 17 francophone West-Central 3 1 primary education + about 3 months initial little change of posts professional training years (average) fieldwork + 5 years (average) Passive: 30% fieldwork + 1 year Passive: 45% Information on long courses based 1999-2001; 2 Burkina Faso, Guinée C., Mali, Mauritania, Senegal 5 Protected area Length of participants experience village exercises and Benoué, Waza Cameroon NP, Cameroon additional training responsibilities 1997-991997-99 18 O-level + 2 additional 4.3 to 9.1 years 33 None: 50% A-level + 2 years 4.7 to 8.1 years None: 27% 1999-01 20 days /5 1999-01 14 25 (average) office Active: 0% (average) office Active: 9% 3 3 ) Senegal Niokolo NP, – Background of trainees and community conservation courses at the Garoua Wildlife College (1996 –2000) ) Dja Reserve, course at Garoua with variety of posts and no experience 5 4 Information on long courses based 1997-1999; CAR, Cameroon, Chad, Congo-B. and Gabon; Cameroon) Park guards(North Cameroon)Park wardens (C. Africa 2000 2000 3 days / 0.5 day role play 10 days / 2 2 x 30 15 Predominantly Diploma 5-20 years, often a Majority with little to Table 11.1 Table LONG COURSES Certificate course SHORT COURSES Park guards (Far North 1996 2 days /none 2 x 13 < 5 years1 4 10–30 years, with None Type of coursesType Sessions Course length / Number of Prior Education Diploma course Park wardens Africa (W. 2000 10 days / 1 day 16 256 Part III – Enhancing conservation – development integration

In view of the present predominant lecturing style Garoua faces the challenge to include insights from participatory adult education (van der Bor et al. 1995; Crow- der et al. 1998).

Both long and short courses started out with discussing the objectives of community conservation, addressing knowledge and attitudes. These objectives were specified by the tuning of objectives of a protected area (IUCN 1994) with possible participation level such as described by Pretty et al. (1995) and Barrow & Murphree (2001), essentially addressing analytical skills. For the remainder of especially the long courses, a ‘community conservation cycle’ was put at central stage, in analogy to the project cycle (Cracknell 2000), to stress the interrelationship between undertaken activities (Table 11.2). The diagnostic stages of Participatory Rural Appraisal (PRA) and problem- and stakeholder analysis (de Groot 1998) were followed by the facilitation of stakeholder representation. Confidence building, addressing attitude and communication skills, was part of the course as well because of its relevance for protected areas with tense relations between protected area authorities and local communities. Problem analysis and categorisation allowed trainees to get a grip on development problems with conservation requirements and their inclusion into the community conservation cycle.

Long courses were designed with a focus on analytical and communication skills, addressed by lectures, fieldwork and examples from protected areas either visited or from students’ home areas (Table 11.2). A one-day village PRA exercise was taken up in the college’s regular 1-2 week field trips to one of the national parks in northern Cameroon. Unfortunately, fieldwork planning did not allow for overnight stays in a village, which could have stimulated more informal contacts.

Short courses for park guards Both short courses (Table 11.1) aimed at initiating park guards to collaborate with local communities (Table 11.2) and the development of a dialogue between ICDP projects and guards. The courses were structured around the main phases of the community conservation cycle, with examples from the local protected area and its ICDP project, and based on the concept that involvement of local communities should not imply loss of authority over the protected area.

Short courses for park wardens The refresher course was limited to 10 days because of financial and absence- from-post constraints, motivating a specific course theme: ‘The role of management planning in the involvement of local people in protected area management’. This choice was based on the interest that Cameroonian park wardens had shown in the park management planning of Waza National Park and the frequent discussions on community conservation (Chapter 9). These refresher courses focused participants from several West and Central Africa countries on a selected protect- 11 – Community Conservation Training 257 2 – SAWC 2 – mostly limited to participa- 4 1 2 –– + ++++ – – – + ++ Garoua Mweka 3 3 Long Short Long Long 3 Unspec. Guard Warden Certificate Diploma unspec Level College Course length K: Knowledge – A: Attitude Sa: Skills, analytical; Sc: Skills in Communication Sf: Facilitation; 3 also subject of follow-up long courses Protected Area Planning and Extension 5 1995); et al. 1998 course manuel; 2 – Contents of community conservation courses at the three African Wildlife Colleges strategies 2001 syllabi; Table 11.2 Table Community conservation: goal or means? Synonyms of community conservation participation level to objectives Tuning K,A + K Sa + + + – – + ++ + –++: dominant; +: much attention; –: fair – no attention: 1 – + +/– + – + Description of training Type • Communication S + The community conservation cycle, including: • Confidence building• Diagnosis: PRA • Diagnosis: Problem / stakeholder analysis • Representation of stakeholders• Negotiation of contracts, • Implementation of contracts• and evaluation Monitoring ++ SaSkills addressed in: • Analysing park – people interactions + ++• Formulating conservation & development K, S, A A,Sc ++• + Facilitation +• K, Sf Conflict management ++ Sa, K Sc Sc Sa K – + – – – – ++ – + – – – – – – – – – – – – – – – – – S – – S + – – – – – – – + – – – – + – – – – + – – – + + – – + – + + + – tion typology (Pretty 258 Part III – Enhancing conservation – development integration

ed area where management planning was ongoing. In preparation of the short courses, park wardens prepared 10 minutes presentations on management planning in relation to local community involvement in their own protected area, allowing a focus of the introductory lectures on management planning and community conservation, for both trainers and trainees. In the session with Central African park wardens the hosting park warden presented the ongoing park planning process, subject of subsequent fieldwork. In the training session with West African park wardens, the host park warden was not able to carry out his presentations and emphasis of the course was shifted to an exchange of experiences amongst participants, with fieldwork limited to a one-day visit of development realisations. Both sessions included lectures and discussions on PRA and communication skills, with PRA field exercises in the Central African session only.

Comparison with other African Regional Wildlife Colleges With the Garoua courses as starting point, we compared the 2001 syllabi of the one-year certificate and two-year diploma courses at the College of African Wild- life Management (Mweka, Tanzania) and the 1998 course manual of the one-year course at the Southern African Wildlife College (SAWC, S.Africa).

11.3 Evaluation methods

Training is a systematic process, which helps individuals develop knowledge (facts and information), skills (the ability to carry out an action) and attitudes (feelings expressed to others through behaviour or language) to increase their job performance (Rothwell & Sredl 1992; Stone 1997). These three training dimensions are subject of the review in this paper. We followed Kirkpatrick’s hierarchy (Kirk- patrick 1994; Rothwell & Sredl 1992), based on trainees’ (1) reaction, (2) learning, (3) behaviour change during job performance. No course evaluations from Mweka and SAWC were available for comparison.

Trainees’ reactions At the start of the 1999-2001 long course, students were invited to write a state- ment on people’s involvement in conservation, its justification and their earlier experiences. These statements were categorised in four classes, ranging from development to conservation focus. Long course students judged the ‘relevance’ of each course of the curriculum for their (future) profession, as part of the curriculum evaluation. Students also judged the ‘presentation’ of each course, in which they addressed whether they ‘liked’ the course and its training methods. Both long and short course participants were asked to judge the relevance of course themes and training methods on a scale of 1 (non-relevant) to 5 (highly relevant) and explain their choice in few words. 11 – Community Conservation Training 259

Trainees’ learning At the end of all but the 1996 short course, trainees’ reactions on their perceived learning experiences were anonymously assessed with two multiple-choice questions. Half way through and at the end of each long course, Garoua organised a written assessment, addressing acquired knowledge mainly, constituting half the students’ mark. An oral assessment at the end of the semester constituted the other part, together with, in 1999-2001, an appreciation of the PRA field exercise’s report written by groups of four diploma students. Oral assessments generally emphasised students’ analytical skills and attitude and often referred to work placement experiences. These assessments fulfilled regular educational requirements, and were a check on learning progress.

Behaviour change during job performance Experience with tracer studies of Garoua has shown that post-course assessments based on sent-out questionnaires only do not yield valid results and need more cost intensive interviews conducted in the respective countries of former students. For the present courses, post-training observations could be made for park guards based in North Cameroon only.

Statistical analysis Correlations of formal assessments were analysed with Pearson correlation, comparison of their means with paired t-test (SPSS 1999). The five point rating scale used in the curriculum and course contents and -methods evaluation was analysed with a Mann-Whitney test (SPSS 1999).

11.4 Results

General reactions to relevance and presentation of the courses Three-quarters of the diploma students and half of the certificate students indicated they had some experience in community approaches prior to the course, especially where they had a background in agriculture or forestry (Table 11.1). Asked to justify community conservation at the start of the course, certificate students emphasised conservation objectives, whereas diploma students tended to emphasise development objectives.

In the curriculum evaluation of the diploma long course, the relevance of both sessions of the community conservation course was evaluated as above average amongst the well-established courses in wildlife and vegetation inventories (p < 0.01). Diploma students judged the presentation of the 1997-99 session as average amongst these, improving in 99-2001 (p < 0.0001), without correlation with perceived course relevance. Certificate course students perceived the relevance as 260 Part III – Enhancing conservation – development integration

well as the presentation of both sessions of the community conservation course as average, outperformed by courses on mammals and legislation.

The theory-practice balance of the community conservation course was judged equal to the other courses that, with the exception of courses on mechanics, computer science and vegetation inventories, were all considered too theoretical.

Park guards judged the course as useful (4.3 and 3.7 on a 1 to 5 scale for the 1996 and 2000 session respectively). The park warden refresher course was considered very useful by half and two-third of the trainees of the West and Central African sessions respectively. The remaining wardens considered the course merely useful.

Specific reactions to course contents and training methods Courses were held in a stimulating atmosphere, with trainees eager to know what this community conservation approach that so many people refer to really is about. Discussions on fieldwork sessions were, especially at certificate level, sometimes considered overwhelming. Frequent structuring and summarising course sessions allowed students to take notes that they could consult and ask questions about in a next session.

The introduction of both long and short courses emphasised the responsibility of park personnel towards conservation objectives, stressing that community conservation is a means and not a goal as such. This was a response to some trainees who felt threatened by the mere idea of community conservation, which they initially perceived as loss of authority.

Certificate students considered problem analysis, problem categorisation and communities’ representation as very relevant compared to the introductory subjects (p < 0.05). The former had received much attention in the course, stimulated by the large number of questions, exercises and discussions with examples from the visited protected areas, whereas the latter had remained on a somewhat more abstract level. Diploma students singled out the topic of ‘Who is local?’ as more relevant than the introduction (p = 0.04).

Diploma and certificate students expressed a strong preference for cases that were familiar through their own fieldwork. The discussion of the Waza National Park committee and especially the visit of a local community member who had signed a contract with the Waza National Park authorities were thus considered above average relevant (p < 0.05). Certificate students considered cases presented by fellow students on protected area committees very relevant (p < 0.05), contrasting with a merely average appreciation by diploma students. This difference may be attributed to a highly skilled certificate student who presented an innovative case 11 – Community Conservation Training 261 from Dzangha-Sangha (CAR). Contrasting with the high relevance of these direct experiences were the discussions on an article on Campfire in Zimbabwe (Feron 1997) and a video on buffer zone management in Uganda (Brown & Singer 1991) (p < 0.05).

In the park guards’ short course, stimulating question and answer sessions developed following each introductory lecture. These discussions were often centred on the ambiguous legislation with regard to community involvement. Attention was paid to development problems of the region and the role park guards could play to gain confidence of local communities. The participation of wardens, the guards’ supervisors, in the course limited the frankness of the discussions, but assured a consistent approach towards local communities. The ICDPs that financed the short courses were happily surprised by the discussions as they had expected a somewhat hostile attitude of the park guards towards them.

In the refresher course for Central African park wardens, park management planning received a lot of attention with a high appreciation for the course theme (4.8 on a scale of 1 to 5) and perceived relevance (4.4). In the West African park wardens course, with the poorly developed practical case, exchanges of experiences (4.4) were considered above average (p = 0.014), whereas community conservation approach ranked second (4.2). Contrary to the long-course students, park wardens expressed a lot of interest in the somewhat abstract subject of tuning participation level to protected area objectives.

Trainees’ learning Student assessments were (partly) based on fieldwork and associated classroom discussions, and presentations of fellow students. Assessments showed generally large differences amongst students that Garoua trainers generally attribute to the students’ educational background that vary between their countries of origin. Repeated assessments yielded little individual variation. With exception of the 1999 certificate long course, oral assessments targeting skills and attitudes, had higher results than written tests (p < 0.05). Results of the diploma students’ exercise on PRA fieldwork, that stimulated working till late at night were marked even higher.

Half to three-quarters of the diploma and certificate students declared that the course had changed their attitude on the need to involve local people (Table 11.3). All diploma students and a majority of the certificate students declared that the course had changed their (analytical) skills on how to involve local people in conservation (Table 11.4). 262 Part III – Enhancing conservation – development integration

Table 11.3 – . Community conservation course evaluation, attitude dimension, assessment of change of opinion by course participants on the need to involve, in one way or another, local people in nature conservation (% of respondents).

Long courses Short Course

1997-1999 1999-2001 2000 2nd year 1st year

Diploma Certificate Diploma Certificate Guards n=31 n=16 n=24 n=14 n=55

Yes, a lot 3944756462 Yes, a little bit 23 38 17 14 33 No, I was already aware 39 13 8 7 5 No, the course was not of 0 6 0 14 0 any use

Table 11.4 – Community conservation course evaluation, skills dimension, assessment by course participants on the change of their capacity to involve local people in nature conservation (% of respondents).

Long courses Short Course

1997-1999 1999-2001 2000 2nd year 1st year

Diploma Certificate Diploma Certificate Guards n=30 n=15 n=24 n=12 n=51

Yes, a lot 7747756729 Yes, a little bit 23 40 21 33 65 No, I already mastered 0 13 0 0 6 the subject No, the course was not of 00400 any news

Virtually all park guards declared having changed their attitude on the need to involve local people (Table 11.3), which was the primary objective of the park guard course. This contrasted with the limited perceived changes in skills (Table 11.4), not specifically addressed by this introductory course (Table 11.2).

Post-course behaviour change during job performance In the months following the 1996 training, at least five (out of 26) guards under- took, stimulated by follow-up instructions of their wardens, initiatives such as the organisation of village discussion sessions, the assistance of villagers against problem animals and a bushfire campaign. In addition, the warden of one of the northern province national parks initiated with his guards a public awareness campaign 11 – Community Conservation Training 263

in all park villages, the first time during their presence in the area. These initiatives much improved, at least temporarily, their hitherto negative relation with local communities. The fact that half of the Far North Province guards went on retirement between 1996 and 2004 limited the long-term impact of the course.

Community conservation courses at the other African Regional Wildlife Colleges In the 2001 curricula, Mweka has developed a two-week certificate course ‘People and Conservation’ and, at diploma level, a two-week communication skills- and a three-week Community Conservation course. A three-week course on participatory planning and community conservation is given at post-graduate level for senior wildlife managers without prior training in participatory approaches, like the Garoua warden refresher course. Long course training methods at Mweka and Garoua were comparable, with predominantly lectures (50%), case studies by students (15%) and only 15% of the course time consisting of field exercises. Although course objectives differed amongst the wildlife colleges, the contents were largely comparable (Table 11.2). SAWC paid more attention to Community Based Natural Resource Management (conservation outside protected areas), whereas Mweka and Garoua centred their courses on protected areas.

11.5 Discussion and conclusions

Relevance of courses Reaction to training is considered as one of several parameters to study the impact of training (Kirkpatrick 1994; Rothwell & Sredl 1992; Stone 1997). Long and short course participants had an average of over ten years of professional experience and we are therefore inclined to consider their reaction on the relevance of the courses and their perceived learning experience as an indication of the impact of training for their future job performance.

Are park personnel ‘attitude-limited’ towards community conservation as generally assumed? (IIED 1994). In other words, do they reject learning and changing? Rather our hypotheses is that, as with all people, it is difficult to have a positive attitude towards something that you are insecure about and that you cannot put into action. The trainees’ reactions show, however, that once roles and needs are clarified and basic skills acquired, attitudes change quickly (Tables 11.3 and 11.4).

The evaluation suggests that even the three-day introductory courses were already successful in changing attitudes on the importance of involving, in one way or another, local people in conservation. The high appreciation of the relevance of the course by diploma students is consistent with the number that declared that the course had increased their analytical skills (Table 11.4). Prior to the course, these 264 Part III – Enhancing conservation – development integration

diploma students associated community conservation with economic development objectives mainly. The eye-opener was that communities can also be involved in conservation. The merely average perception of certificate students and lower numbers who had acquired necessary skills, suggest that the long course was more tuned to diploma students’ level or interest, probably because of its largely analytical character.

The course evaluations indicated the need for continuous reference to practice. Training on the spot, as practised with the park guards, thus almost becomes mentoring, requiring trainers who are familiar with the working situation. For park wardens, experiences from colleagues played an important role, which need to be streamlined with the course theme, putting high demands on the trainers as well. Experiences with long course students showed that sharing the experiences strongly depends on the availability of interesting cases and showed the need to assist students in presenting them. For long course students, experiences from fieldwork locations have played an essential role. Cases presented on paper or video should be limited to only the truly necessary, e.g. for broader comparison.

A more participatory training should be pursued to allow effective adult learning. Didactic courses and follow-up mentoring may assist trainers, who generally use the lecturing training style they themselves have been trained in, to develop the necessary skills. Long course students with their hitherto passive training experiences should also be introduced to this participatory training, e.g. at the start of the first year as at the other regional African wildlife colleges.

Perspectives Experiences from Gabon suggest a low perseverance of training experiences of governmental protected area personnel, often based in the capital cities (Thibault & Blaney 2001). Participants of the Garoua courses have field-based posts however and may therefore be expected either to use the training results themselves or to stimulate their application by others.

Garoua proposed to train a selection of park guards in communication skills, once the policy described in the management plan would be clarified and clear instructions could be given. This would prepare them as community liaison agents, reducing some of the difficulties in combining policing and extension tasks (Van den Ban & Hawkins 1988). Protected area personnel should receive training in community conservation attuned to their education level and working experience, thus developing not only necessary knowledge and attitudes but skills as well. Apart from the necessity to organise refresher courses for existing protected area personnel, a need exists to develop community conservation courses as part of pre-recruitment training for future personnel. 11 – Community Conservation Training 265

Training in community conservation allows protected area managers to participate in public outreach (Barrow & Murphree 2001; Gilbert 1971) and play an active role in community involvement. This, in turn, may stimulate protected area personnel to influence authorities and projects towards a balanced integration of conservation and development objectives (Scholte 2003a) and the institutionalisation of community conservation (Bergin 2001).

PART IV

Synthesis

12 Floodplain Rehabilitation and the Future of Conservation & Development: Synthesis and concluding remarks

In this concluding chapter, I review the outcome of the floodplain rehabilitation by comparing the reflooding responses that, because of their different response times, caused human supported resources to become dominant. I continue analysing the contributions of management planning and capacity building to correct these undesired effects of the otherwise successful reflooding.

I subsequently review different concepts of ecosystem change and management that may explain the observed changes. I first apply panarchy (Gunderson & Holling 2002), a holistic empirical concept, that allows observed ecological as well socio-economic developments and interventions to be linked in a temporal and spatial hierarchy. It provides limited causal explanations however. Secondly, I use the empirical concepts hysteresis and resilience that show that the rehabilitation of Waza-Logone, despite its success in the field, can be considered as a classic case of hysteresis in which both end results and the rehabilitation path are different from the pre-dam situation and the degradation process. I introduce the concept of ‘evolving states’ to explain that not all observed states are static, but may undergo cyclic succession, such as has taken place with the reflooding. I subsequently applied the holistic-normative ecosystem approach to present the key findings of this study in a wider context by suggesting some further development of these guidelines for analysis.

At the end of this chapter I analyse the adaptive management approach pursued, that I use as overarching concept including elements of the above discussed empirical and normative concepts. I revisit the monitoring observations and insights to reformulate the floodplain rehabilitation expectations and hypotheses postulated in the beginning of this chapter and discussed in the respective chapters. The consequences for the overall floodplain rehabilitation targets as programmed in Waza-Logone (IUCN 1999) and possibly in other (African) floodplains are discussed. I call especially attention to the role of a ‘vision’ , including not only targets for rehabilitation but also understanding the required enabling environment. Empirical concepts of change and exchanges with other experiences played an important role in the development of this vision. One may conclude that adaptive management, based on monitoring of separate ecosystem components and a regular review of targets within an overall vision, forms a useful framework for the rehabilitation and management of poorly known tropical (floodplain) environments.

271

12.1 Synthesis of the monitoring results on reflooding responses

In this section I synthesise the changes in vegetation, wildlife and pastoralism in Waza-Logone observed during the last two decades. I will start, insofar available information allows, with a description of the degradation processes induced by the construction of the upstream Maga dam in 1979.1 This is followed by a summary of the changes observed from 1993-2000(-2003) and their interpretation. I conclude with the reformulated hypotheses for future reflooding. A particular point of attention is whether observed changes can be attributed to the reflooding or to alternative explanations such as the relatively favourable rainfall during the study period (Chapter 2), (see Table 12.1 for an overview of their relative impacts).

Table 12.1 – Impact of reflooding and the importance of other factors that explain observed developments (for temporal and spatial scale see Fig. 12.2)

Resource Reflooding Increased Extraction Protection Diseases Develop- rainfall + (grazing, (by humans) ments related poaching elsewhere flooding or theft)

Vegetation Composition ++1 +/– 1994-1998 – – – – – – + 1998-2000 Production +1 +–2 –– –– –

Waterbirds Anatidae +/– ++ +/– +/– – ? (dry season Omnivorous + +/– +/– + – + numbers) Ciconiiformes Piscivorous large + +/– ++ – – – Ciconiiformes

Waterbird Black–headed + +/– +/–3 ++ – + (?) colonies heron (nests)

Floodplain Kob (number) + +/– + + +4 – antelopes Korrigum (number) +/– (?) + – ? + ? + ? –

Pastoralism Camps (number) + – +/– +/– + Grazing intensity + + – +/– + (no. cattle km-2)

Impact : ++: major; +: important; +/–: existing; –: marginal; – –: none; ?: unknown; nothing mentioned: not relevant. 1Includes the in the long–term impact of increased physical protection; 2no grazing during flood season; 3potentially important; 4rinderpest

1 For earlier changes, see Chapter 2. 272 Part IV – Synthesis

Vegetation composition

Impact of the Maga dam • Fragmentary information suggests that, in the early 1980s, in extended parts of the area perennial grasses died. This can be largely attributed to the dam construction, although there were indications that the condition of especially Veti- veria nigritana was already poor due to the mid-1970s drought, the first since the 1930s (Fig. 3.2). Till 1985 and in some occasions 1993, remaining individuals of the rhizomatous Echinochloa pyramidalis were found in the desiccated parts of the floodplain (Table 3.2). In the central part of Waza NP, out of reach of the reflooding but with local rainfall ponding, a few individuals of the tussock grass Vetiveria nigritana survived for over twenty years.

Recent changes • The 1993-2000 monitoring, supported by field observations in 2002, indicates that in 2002 the perennial rhizomatous grasses Oryza longistaminata and Echi- nochoa pyramidalis occupied almost the totality of the reflooded area once again. • In terms of species composition, this rehabilitation was only partial however: of the previously dominant perennial floodplain grasses, Vetiveria nigritana did not show any sign of re-establishment despite its good condition elsewhere in the floodplain.

Impact of the reflooding • The initial hypotheses of a ‘full recovery, in five years, to the pre-Maga state from the 1970s’ was therefore rejected. • The rehabilitation has taken place under (long-term) average rainfall and flooding conditions, but above average if considering rainfall of the last 30 years only. The lay-out of our study, with plots both in and outside the area impacted by the reflooding, showed the large impact of reflooding compared to a low impact of merely average rainfall.

Interpretations • The gradual but steady recovery of Oryza longistaminata and Echinochloa pyramidalis grasses can be explained by lateral growth through their rhizomes. • The tussock grass Vetiveria nigritana depends for its multiplication on seed dispersal, a less secure strategy that may depend on (temporarily) entirely different environmental conditions. This holds in particular for the direction of flooding water, and therefore the sediments it carries, that has remained entirely different from the pre-dam situation. 12 – Synthesis 273

Hypotheses for future reflooding • With future reflooding one should take into account a period of at least ten years for the recovery of rhizomatous grasses. If rainfall is lower than average, this period may be much longer. • Full recovery that includes again a predominant presence of Vetiveria nigritana, may last much longer than 10 years and may depend on siltation processes.

Vegetation production

Impact of the reflooding • Above-ground biomass production showed a 2-year time lag in response to the start of the reflooding. • Assessments in Waza-Logone, supported by data from three other seasonally flooded grasslands from the three main African geographic regions, showed an increasing above-ground biomass production with maximum flood depth in a 0-1 m range. • Data from Waza-Logone did not support an established linear relation between dry season regrowth production and wet season above-ground biomass from the Inner Niger Delta (Mali), and thus with maximum flood depth. A comparison with other African floodplains hints at an exponential relation of regrowth production if maximum flood depth exceeds ca. 100 cm. Below this maximum flood depth, the timing of burning has an overwhelming influence on the (low) regrowth production.

Interpretations • We applied the relationship of maximum flood depth and above-ground biomass to the 20 cm water level rise, due to the reflooding, in an area of 60 000 ha. This way it may be estimated that wet season above-ground biomass has increased by approximately 800 kg DM ha-1 in the first reflooding year (≈ 10% production increase) increasing to an additional 2800 kg DM ha1 (≈ 37% increase) in later years.

Hypotheses for future reflooding • I postulate a two year time-lag and the maximum flood depth - above-ground biomass equation, for future reflooding in the plain grasslands (with a maximum flood depth < 1m). • I hypothesise that the time lag of the response of dry season regrowth production to reflooding is a multitude of the two years time-lag of wet season above- ground biomass. Regrowth production is expected to start only after several years, when sufficient rhizome biomass has been accumulated. 274 Part IV – Synthesis

Waterbirds

Impact of the Maga dam • The list of bird species no longer observed since 1980 suggests changes amongst especially the woody savanna species, more influenced by long-term rainfall change than by the flooding (Scholte et al. 1999). • I attribute the drop in numbers of the area’s flagship species, the resident Black-crowned Crane, from 10 000 in 1973 to 2000- 2500 individuals in 1992 to the Maga dam construction. No quantitative information was available about other waterbird species.

Recent changes • The numbers of waterbirds, observed during the dry season, increased from 60 000 to 105 000 between 1992 and 2000. • The numbers of waterbirds surpassing the international 1% criterion of their population size, doubled from 6 to 12 per year. Also the population numbers of each of these species important for conservation increased, as was shown for Black-headed Heron and Black-crowned Crane. • The already exceptional size of the Black-headed Heron colony at the start of the reflooding tripled during the study period to a total of 2500 nests.

Impact of the reflooding and interpretations • The increase in numbers of Anatidae (ducks and geese) corresponds to the generally observed trend in their recovery from the mid-1980s droughts as observed elsewhere in West Africa. • The change in numbers of Ciconiiformes (storks, herons and egrets) exceeded their developments elsewhere in West Africa. • The spectacular development of the Black-headed Heron colony was attributed to the reflooding, and the incorporation of smaller colonies previously subject of destruction. • Amongst the water birds that showed a trend less than the general two-fold increase, were the large predominantly piscivorous resident Marabou and Yel- low-billed Storks. This is attributed to the repeated destruction of their colonies by (fisher) men, a recent development in Waza-Logone that has elsewhere in West Africa already been under way for years.

Hypotheses for future reflooding • Hypotheses for the recovery of waterbird by future reflooding should take into account their long-term dynamics based on regional changes (e.g. rainfall) and local changes especially human protection and prosecution. The categorisation in functional groups, with a comparable reaction to reflooding, should therefore not only be based on status (Palearctic migrant – Afrotropical resident) 12 – Synthesis 275

and feeding ecology (piscivorous – omnivorous), but also on vulnerability to persecution.

Floodplain antelopes

Impact of Maga dam • Numbers of Kob antelopes crashed from 20 000 to 5 000 in the period 1979- 1983 due to the man-induced drought caused by the Maga dam construction. The subsequent drying up of the area coincided with a period of rinder pest. Numbers further declined to 2 000 at the end of the 1985 drought, showing the antelope’s vulnerability to natural droughts as well. • The disappearance of Waterbuck, soon after the construction of the Maga dam, has become a symbol of the area’s degradation.

Recent changes • The Kob population numbers increased between the late 1980s and the early- mid 1990s from 2 000 till 5 000 individuals, after which the population size stabilised. • Korrigum (‘Topi’) antelope numbers have increased steadily since the early- mid 1980s as well. • In 1998 and again in 2003, two Waterbuck individuals were observed once again.

Impact of the reflooding and interpretations • Kob antelope is the only large mammal whose population dynamics in the period 1960-2000 could be linked to water management, i.e. the 1979 Maga dam construction and the reflooding from 1994 onwards. • The initial recovery in Kob numbers has not continued beyond 1997. The main explanation is the increased competition with livestock due to the rapidly increasing cattle grazing intensity, also inside Waza NP (see below). Direct persecution has almost certainly played an important, yet poorly known, role as well. The dramatic drop in number of park guards (see below) is an indication of the reduced conservation effectiveness.

Hypotheses for future reflooding • Floodplain rehabilitation hypotheses should not only be based on water management but include feedback mechanisms of conservation effectiveness as well.

Pastoralist Responses

Impact of Maga dam • Only scarce qualitative information is available on the reduction in grazing intensity following the Maga dam construction. One may assume a rapid drop in 276 Part IV – Synthesis

cattle grazing intensity by a reduction in length of stay in the area impacted by the Maga dam. If production of the perennial vegetation did indeed cease abrupt- ly, many pastoralists may have had to decide already at the end of the rainy season to direct their herds South (as during the 1985 drought, Chapter 7). In any of these cases, the grazing intensity decline will have taken place faster than the recovery process.

Recent changes and impact of reflooding • From 1992 to 2000, we observed a doubling of the number of cattle herds, two- thirds of which was due to newcomers and one-third to ‘natural’ increases based on increased floodplain productivity due to reflooding and favourable rainfall. • The average time cattle herds were present in the reflooded area increased by 60%. • The resulting grazing intensity, expressed as cattle density, almost tripled between 1993 and 1999. • The increase in the number of herds and camps slowed down after 1996. Gra- zing intensity continued to increase linearly however, showing density independence, following the gradually recovering perennial grasslands. It is expected that this increase has levelled off at a maximum of about 100 cattle km-2 (≈ 70 TLU) around 2003. • I attribute the quasi-totality of the observed changes to the reflooding, the only factor that differed between the studied floodplain and surrounding areas from where pastoralists moved away.

Interpretations • No ‘territorial blocking’ nor ‘chaotic overshoot’ scenario has taken place with the reflooding. The Ideal Free Preemptive Distribution scenario, assuming that any increase in perennial vegetation is consumed by cattle with the first individuals using the territories pre-empting them, best explained pastoralist responses to floodplain rehabilitation. Pre-emptive use might be reflected in the customary rights that pastoralists have to campsites, which they have occupied for a long time. • One of implications of the Ideal Free Pre-emptive distribution is that with reflooding, apart from a short lag in pastoralist reaction time, all additionally produced forage is consumed, leaving few incentives for pastoralists to refrain from grazing inside Waza National Park.

Hyptheses for future reflooding • I postulate the Ideal Free Preemptive Distribution scenario for future reflooding of the area. Unfortunately, this scenario predicts continued grazing of livestock inside Waza National Park, unless strict measures are taken. This scenario is conditional on reflooding being carried out in steps with comparable size of impact area. If the impact size is much larger, as proposed by IUCN 12 – Synthesis 277

(1999), the time-lag of the arrival of new groups may be somewhat longer than observed during the 1994 reflooding.

12.2 Comparison of reflooding respronses

Floodplain rehabilitation may lead to unexpected changes in land use. Quickly recovering species will have a distinct advantage over slowly recovering competitors, driving the latter into a secondary role or even towards their local extinction. In the Waza-Logone floodplain with competing land uses, speed of response has therefore a profound influence on their relations. One may argue that area’s continuing resilience (see 12.4) will lead to corrections in the long run. With the virtually irreversible changes in hydrology and landuse, exchanges between wildlife populations in the Lake Chad Basin have become very limited, however, hampering any recolonisation.

To allow a comparison of the 1993-2000 (-2003) developments of perennial grass cover, numbers of Black-headed Heron nests, dry season numbers of waterbirds, Kob antelope numbers and cattle grazing intensity, I indexed these parameters to their 1993 pre-reflooding situation (Fig. 12.1). I consider the development of perennial grass cover and vegetation production as reference lines for the floodplain ecosystem recovery.

This comparison shows that the recovery of human-supported resources such as cattle grazing intensity and villagers’ protected heron nests, exceed the increase in resources with which humans compete, i.e. Kob antelopes and large piscivorous waterbirds. Human populations have been quick to respond to the improved conditions, out-competing wildlife (Fig. 12.1). The isolated Waza-Logone population of 4000 Kob antelopes had a limited capacity to increase, at most at an annual rate of 50% i.e. 2000 antelopes. The ‘reservoir’ of 4.7 million cattle in the Convention- 2 al Lake Chad Basin (Martin et al. 1996) of which 200 000 cattle visit the floodplain has, however, a potential of increasing cattle grazing intensity in the floodplain many times over, by redirecting migration routes. A number of species that compete with people for the same resources (e.g. piscivorous birds) have also come under increased pressure, not only because of the increased competition (as indicated by the number of fishermen, Chapter 8) but also because of direct persecution (Chapter 5).

2 Comprising Nigeria, NE of Damaturu, Cameroon north of Guider, Chad in a line of (Figuil) Pala- Lai-Melfi-Moussoro-Mao, and extreme SE Niger. 278 Part IV – Synthesis

400

heron nests

300 cattle density perennial grass cover

200 waterbirds

Vegetation production

Increase (100=1993) Kob 100 large piscivorous birds

0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 start reflooding increase reflooding (0--> 20 m3/s) (20 --> 30 m3/s) flooding season Notes: 1. From 1999 onwards, development of perennial grass cover is based on transect data 2. The total number of these parameters increased significantly over time (GLM, F = 31.1 p < 0.0001, Df=1,49), their relative increase over time differed as well (groups*time F = 6.083, p > 0.001, Df = 6,43). One should take into account that the impact areas cover with 180 km2 and 600 km2 respectively, only parts of the area where waterbirds have been counted. A comparison of the ecosystem components that were monitored on approximately the same scale gives, however, comparable results. Cattle grazing intensity (cattle km-2) with Kob antelope numbers: significant increase in time (F = 6.8 p = 0.026, df = 1,10), and significant difference amongst components: (F = 12.7, p = 0.006, df = 1,9). Numbers of waterbirds with heron nests and numbers of piscivorous birds: significant increase in time (F = 8.7, p = 0.008, df = 1,21) and significant differences between components (F = 4.4, p = 0.027, df = 2,19).

Figure 12.1 – Development of Waza-Logone Resources, 1993-2003.

12.3 Enhancing the integration of conservation and development

Human population and land use dynamics The Waza-Logone floodplain rehabilitation has led to the partial, yet substantial recovery of its main land use types (fisheries,3 pastoralism and conservation), truly the success story of the Waza-Logone project and worth the investments (Chapter 3-7; IUCN 1999; Emerton 2004). However, already strained relations between pastoralists, fishermen and farmers have come under further pressure, leading to violent clashes during the first reflooding years (Kari & Scholte 2001). For changing relations between farmers and fishermen one is referred to Kouakam et al.

3 As explained in the preface, not treated in this thesis. See, however, Loth 2004. 12 – Synthesis 279

(2004). Also within the same category of land use, different groups have benefited differently. For example Mare’en pastoralists, most prominent prior to the reflooding, benefited less than Arabs and agro-pastoralists (Chapter 7). This has, as far as we are aware of, not led to tensions between these groups, probably because of their long existing and as yet little disputed power relations.

More problematically, the rise of human population numbers at the borders of Waza National Park has put the conservation – development balance under pressure (Chapter 8) leading for instance to increasing domestic grazing inside the national park. One might argue that Waza-Logone is an exception because it has experienced a degradation process with subsequent emigration. Experiences with development activities around conservation areas becoming poles of attraction for people have, however, been reported in an increasing number of cases (Chapter 8; ABCG 2004).

In the long term, people may become convinced of the need to refrain from exploiting resources from inside the protected areas. A population build-up may potentially even have advantages in awareness building. There does not seem to be an analogy with cases of positive influences of population density on land management (‘more people, less erosion’ Tiffen et al. 1994) unless one would accept to follow the path of the developed countries with the loss of species which may be reintroduced subsequently.

Wildlife policies: developing instruments for Conservation & Development At present, no politically or ethically viable alternatives for conservation & development approaches exist (Adams & Hulme 2001; Chapter 8). Most of the interest in conservation by authorities and politicians in Cameroon and elsewhere, even though still limited, is due to the activities of conservation & development projects. Continuing on the path of integrating conservation with development is further justified because present failures can be largely attributed to the poorly developed instruments in the field. It is as one says ‘If your only tool is a hammer everything looks like a nail’.

One of the most promising tools explored in Waza-Logone was social fencing, where local communities exclude ‘outsiders’ from exploiting the surroundings of a protected area (Chapter 8), an approach that is no longer taboo (Anonymous 2004). This was part of management agreements between protected area authorities and local communities, where a controlling and information duty of local communities was rewarded by exclusive exploitation rights as well as other privileges such as a small-scale development assistance (Chapter 9). The pursued ‘carrot and stick’ approach towards local communities was another promising instrument applied for conservation with development. New opportunities (‘carrots’) were created for park border villages that were targeted for small-scale development activi- 280 Part IV – Synthesis

ties, whereas Baram, the village inside Waza National Park, was actively deprived of such opportunities (‘stick’). A local initiative resulted in Baram being moved thus breaking a deadlock of tense relations with the park authorities and assuring a new future for the village (Chapter 8).

The convention with the pastoralists from Fadaré, at 40 km from Waza National Park, showed the conservation potential of development activities at a certain distance from the protected area (Chapter 7). For two years this convention reduced livestock pressure on Waza NP, but the rapidly increasing grazing intensity in the floodplain and difficulties to monitor the convention with a limited number of guards (see below) made this tool less successful in the longer term.

The implementation of the management plan has been successful on a local level as the greatly improved relationship between park authorities and local communities testified. Lack of commitment of especially the Cameroonian government has led, however, to a decline in the number of park guards that compromised the momentum of gained confidence between protected area authorities and local communities (see next section). These experiences should be taken into account in other protected areas in Cameroon and Central Africa, where consensual management planning has just started. More emphasis should be put on the design and implementation of locally adapted arrangements in which not only local people but also local and central governments should keep their side of the bargain.

Capacity building of wildlife managers Protected area personnel such as wardens and guards generally have little input in protected area planning and community conservation. Their lack of capacity in these disciplines was not unique to Waza Logone, but also identified as a significant constraint in a WWF-portfolio of seven ICDPs in Gabon, Ethiopia, Ecuador, Honduras, Pakistan and the Philippines (McShane & Newby 2004). The need for capacity building in these disciplines motivated the development of a training programme at the Garoua Wildlife College, the first of its kind in West and Central Africa. De- veloped courses aimed at raising knowledge and skills of protected area personnel to support the dialogue with local communities enhancing the conservation of Waza-Logone and other protected areas in Cameroon and West-Central Africa. Experience with these courses showed the potential to upgrade the knowledge and skills of present personnel that are not attitude-limited as often supposed.

The study on the curriculum development at the Garoua wildlife college in Came- roon showed only gradual changes contrasting with the rapidly changing market demand that calls for more attention to community conservation and other subjects that reflect present park management objectives. Garoua’s curriculum differs from the other African regional wildlife colleges at Mweka (E.Africa) and SAWC (S.Africa) in its limited attention to community conservation and business skills. 12 – Synthesis 281

Erosion of the Cameroonian wildlife sector Poor human resource management in the Cameroonian wildlife sector has overshadowed many of the reported results of management planning and training. Within four years following the first refresher courses the number of Waza park guards had dropped from 20 to 10, due to retirement, illness or change of posts (Fig. 2.4). The number of guards in the northern province dropped less dramatically but also experienced ageing and lack of post rotation. Personnel management has captured our attention since the first discussions with the officials of the Human Resources department of the Ministry of Environment and Forestry (Chap- ter 9), but can be largely attributed to a lack of government commitment (van Est & Scholte 2001). The proposed institutional changes of the wildlife department into a parastatal (Director of Wildlife, pers. comm. 2002), inspired by the Kenya Wildlife Service model (Leakey & Morell 2001), has the potential to change the precarious personnel situation. The forthcoming recruitment of 600 park guards (Director of Wildlife, pers. comm. 2001 repeated in subsequent years) may be an outstanding opportunity for preparatory training in new management techniques.

In contrast with this lack of government commitment towards biodiversity conservation was the protection of the area against armed robbery and cattle theft. At the start of the reflooding in 1994, pastoralists expressed their concern about the problematic security situation that threatened to block access to the newly reflooded area (Chapter 7; Scholte et al. 1996b). This situation improved dramatically after 1998 when the government sent special troops to control the situation (Chapter 8). This shows that the difficulties with conservation did not only result from government weakness, but ultimately from a lack of government commitment.

12.4 Empirical concepts of ecosystem change

In the sections 12.1 and 12.2, I synthesised and compared the changes of the individual Waza-Logone ecosystem components, followed by an overall comparison. In this section I will discuss approaches that aim to explain ecosystem changes from a whole-system perspective. Experiences from the Waza-Logone rehabilitation suggest that the time-dimension should play a prominent role in these approaches. In addition I argue that some of these ‘holistic’ approaches do not help identify causal relations between change agents and the subsequent ecosystem responses. They should therefore not be considered as models of change. Their role, as metaphor, lies largely in facilitating communication on theories and insights.

Exploring the holistic panarchy concept for Waza-Logone Panarchy is a holistic concept that aims to help explain developments of ecosystems as well as societies (Gunderson & Holling 2002). Two features distinguish panarchy from traditional hierarchical models (Holling et al. 2002). The first is its 282 Part IV – Synthesis

elementary unit, the adaptive cycle. The second feature of panarchy concerns the manner in which elements of complex adaptive cycles nest to one another in a hierarchy (Holling et al. 2002).

I applied the hierarchical concept to illustrate the different recovery speeds and scales at which the floodplain rehabilitation has taken place, see Fig. 12.2. In addition, Panarchy allows putting the importance of the main historic events, introduced in Chapter 2, in a wider perspective, thus highlighting the evolving reference image (Gunderson et al. 2002). Periods of collapsing ‘panarchies’ encompassed the late 1800s that were hit by a cascade of events (rinderpest, elephant slaughter, war and colonisation) that still determine the present state of the area (Chapter 2). A second downward spiral occurred after the construction in 1979 of the Maga dam, immediately followed by the 1980s droughts, a drop of investments in public services and, as some would argue, a collapsing state. These developments took place against a background of an increasing population pressure, not only in the area itself but with surrounding regions exercising a push of human migration as well.

= adaptive cycle 3 1000 = floodplain rehabilitation monitoring Elephant/Woodland cycle = intervention wet dry spells

institutional cycle

political 2 100 Waza mgt.

veg.

Time (year) Time impact of 1 10 comp. pastoral Maga-dam recovery recovery curr. mgt. develop- veg. pro- planning long waterbird ment duction courses recovery recovery

0 1

pilot refresh- veg.- release ment flood (log) course

1 100 1000 10.000 100.000 0 2 345(log)

Area (km2) Figure 12.2 – The Waza-Logone Panarchy 12 – Synthesis 283

One may expand this conceptual framework by adding to the factors time (T) [years] and spatial (S)- scale [km-2], a ‘system level’ (L) that represents the delay in response to reflooding, or its impact on ecosystem components and activities, based on levels of production (e.g. 1 = primary, 2 = secondary), communication, etc. in a hierarchy (see Table 12.2).

The recovery of perennial vegetation, waterbirds, wildlife and pastoral use, characterised by its direct link with the reflooding (L ≤ 2) has taken place in a period of 3 to 10 years (Fig. 12.2; Table 12.2). Although not without surprises (Chapter 3), it has been possible to forecast most of its impact. The short training courses were of limited impact in a relatively short period only (< 1 year), due to the rapidly degrading personnel situation. The effects of the long-course training, let alone of curriculum development, took more time before being felt (Fig. 12.2). The many steps in its implementation (L ≥ 4, Table 12.2), has made it rather complicated to forecast their outcome.

Table 12.2 – Hierarchical levels of ecosystem component’s response to reflooding

System level

Abiotic 1 2 3 4 5 6

Hydrology Change agent

Grassland primary production production

Grassland composition composition change

Waterbirds waterbird responses

Antelopes antelope responses

Pastoral use pastoralist responses

Management diagnostic, negotiationscommitment- planning dialogue implementation

Training course ↓ design trainer implementation

Curriculum ↓ development wildlife design department college community 284 Part IV – Synthesis

One of the lessons learnt is that training people the moment one is confronted with problems is too late. This may seem ‘an open door’, yet it is common practice with conservation & development interventions (Chapter 11). Refresher courses should build on proper pre-recruitment training as earlier argued by Lindley (1999) for extension professionals. This is an additional argument for the need for sound pre-recruitment training and the presence of institutions that can take care of this (Chapter 10).

Box 12.1 – Vegetation states and (cyclic) succession in Waza-Logone

Three states (sensu Westoby et al. 1989) can be distinguished for the present Waza- Logone floodplain vegetation. State A, the ‘pre-dam’ floodplain state, has a diverse perennial grass vegetation, including rhizomatous grasses (Echinochloa pyramidalis and Oryza longistaminata) and tussock grasses (Vetiveria nigritana). This state can still be found in the area that has continued receiving floodwater after the Maga dam construction (Chapter 3). State A has a somewhat perpetuating nature (Rietkerk et al. 2002), the numbers and the condition of Vetiveria nigritana have for example remained remarkably constant throughout a period of relative drought (mid-1980) and more abundant rainfall (study period), Fig. 3.2. This does not exclude long-term changes such as those that occurred in the density of Vetiveria nigritana tussocks in the late 1970s when decades of relatively humid conditions were halted by a drought (12.1). State B, derived from State A after the dam construction, has different appearances, due to the cyclic (Gleasonian) succession that takes place (Fig. 12.3). With the reflooding, Sorghum arundinaceum annual grasses were replaced by short living perennial grasses and subsequently by rhizomatous perennial grasses. Such cyclic succession also occurs in Dinder National Park (Central Sudan) where the same range of annual to perennial grass communities were found in neighbouring backswamps (pers.obs. Feb. 2005). The flooding conditions of these backswamps vary every few years, depending on the shape and therefore discharge capacity of the feeders from the river. Van der Valk (1992) analysed this Gleasonian succession for environments with more predictable cyclic circumstances. State C, on the edge of the floodplain, is dominated by annuals herbs. It can be considered a further ‘degeneration’ of State B, where even in the presence of (late) flood water no further changes in vegetation were observed.

The variation in plant species within the three states, most notably B, is larger than differences between the three states. This is particularly important for the interpretation of ecosystem functions, such as providing forage for grazers. 12 – Synthesis 285

Predam State A Rhizo.grasses + Vetiveria State B. Cyclic succession amongst: nigitana

Echinochloa / no changes Oryza rhizo. observed grasses Dam catastrophe Panicum / Ischaemum per. grasses

Sorghum annual grasses

State C annual Above-ground Biomass – (Nutrient Status?) herb vegetation

Max. flood depth

Figure 12.3 – Vegetation States and cyclic Succession in Waza-Logone (1993-2003)

Hysteresis and resilience Ecological restoration was defined as the complete and functional return to a pre- disturbance state, whereas rehabilitation has often been defined as the only partial structural and functional return to a pre-disturbance state (Tapsell 1995; Mid- dleton 1999). The 2000-2003 reflooding ‘end’ result was different from the 1979 pre-dam situation with regard to the area’s hydrology, vegetation composition, the ratio of numbers of livestock and kob antelopes and, probably, the numbers of large piscivorous birds compared to the number of fishermen. As shown in section 12.2, land use has shifted into a direction in which conservation has lost some of its (relative) pre-dam importance. I therefore conclude that the 1994 reflooding induced the rehabilitation and not the restoration of the floodplain. The Waza- Logone results are not less profound, however, than the results of many ‘restoration’ projects, such as those envisaged by the Everglades Restoration Plan (Sklar et al. 2001).

The end results of the reflooding, discussed in section 12.1, are obviously of importance, but we should not neglect the rehabilitation process itself. A central concept in restoration ecology is hysteresis. Hysteresis, etymologically ‘coming behind’, has its background in physics, where it is the ‘retardation of the effect when the forces acting upon a body are changed. Hysteresis represent the history dependence of (physical) systems: If you push on something, it will yield: when you release does 286 Part IV – Synthesis

it spring back completely? If it does not it is hysteresis in some broad sense’ (Setna 2004). In ecology hysteresis is used in a more holistic sense, emphasising the word ‘completely’ in above description. This results in the concept that ‘that forward and backward switches [of the changes into a state] occur at different critical conditions’ (Scheffers 2001).

With the reinstatement of flooding, tussock grasses and floodplain antelopes have experienced a slower, if any, recovery than the dramatic drop in their population numbers that took place following the Maga dam construction (12.1). One may speculate on the entirely different conditions that should reign to cause, if possible, a switch of the vegetation into its genuine pre-dam state (Box 12.1), such as increased sediment deposits, triggered by water flows feeding the western part of the floodplain (Chapter 3). For the antelopes this may hold for a collapse of livestock grazing intensities or protection measures not only more effective than at present but also more than when the population drop occurred. The hysteresis of these vital resources in the floodplain system recovery has led to a different land use state outcome of which the consequences were discussed in section 12.2.

Resilience, in its broadest sense the ability of a system to persist in an uncertain world, has intrigued ecologists and increasingly social scientists (Perrings 1998). Resilience was initially defined as ‘the ability to experience change and disturbance without catastrophic qualitative change in the basic functional organisation’ (Holling 1973), and more recently as ‘the long-term adaptive property of the system to withstand external perturbations without changing its basic structure’ (Holling 1995). Resilience also refers to ‘...the maximum perturbation that can be taken without causing a shift to another ecosystem state’ (Scheffer 2001). Resilience is not necessarily constant during the adaptive cycle (Holling & Gunderson 2002). Populations are more vulnerable when under stress, as became clear with the cascade of events (drought, rinderpest) following the Maga dam construction (12.1).

Pimm (1984) follows an alternative approach in defining resilience as ‘the time taken for a system to return to its state prior to disturbance’. The concept of resilience appears thus to have different meanings (Perrings 1998; Van den Berth et al. 2004), of which Pimm resilience may be easier to measure (Batabyal 1998) and is closer to the original meaning of hysteresis as simple time lag between cause and effect (see above). Holling resilience is the most frequently being used, reflecting more lasting changes (e.g. Levin et al. 1998). Below I discuss both concepts of resilience for the Waza-Logone floodplain in the light of the changes induced by the construction in 1979 of the Maga dam and, since 1994, the reflooding.

Local people argued that in 1988, a year of abundant rainfall, the flooding in Waza- Logone resembled the floods in the years before the construction of the Maga dam 12 – Synthesis 287 in 1979 (Chapter 3). Also in 1994, the first reflooding year, the flooding appeared once more as its pre-dam stage. The dam has ‘only’ changed the frequency of flooding in the study area, well within the area’s (hydrological) resilience. Yet following the Maga dam construction ‘catastrophic changes’ (Holling 1973; Rietkerk 1998) of the ecosystem have taken place indicating that the area’s resilience has been exceeded, switching from a productive floodplain into a desiccated unpro- ductive state (Chapter 12.1).

In the ten years following the reflooding, the perennial grassland vegetation rehabilitation in the reflooding impact zone was virtually complete, albeit with a different species composition than before the dam construction (Box 12.1, Chapter 3). This rehabilitation took place in a period with relatively favourable rainfall, when perennial grasslands generally expand into the annually grass dominated drier parts of the floodplain (r-phase in the adaptive cycle). In the areas further west, only partly reached by the reflooding, a vegetation occurred that showed only an increasing dominance of annual herbs (Figure 12.3). Because the pre-dam state (A in Fig. 12.3) may not have been reached, I conclude that Pimm resilience is (much) more than ten years. If neglecting the species composition and only considering perennial grass cover as target for floodplain rehabilitation as we initially did (Wesseling et al. 1994), one may conclude that Pimm resilience is about ten year (Fig. 12.4: target image range). Pimm resilience allows a comparison with other areas where rehabilitation trials were conducted, indicating the high resilience of the Waza-Logone floodplain compared to other areas (Chapter 3).

Characteristics that explain the resilience of the seasonally flooded grasslands include the high below-ground biomass (± 70% of total biomass) that is inacces- sible to grazing animals because of the firm vertisols and the six months of seasonal protection due to rainfall ponding and subsequent flooding. Drent & Prins (1987) also argued that, because the herbivore is prisoner of its food supply, vegetation under ‘natural conditions’ is free from disturbance due to the herbivore. Ideal Free Distribution is one the mechanisms that makes it rather unlikely that herbivores exercise long-term disturbance (Scholte & Brouwer 2005).

The Lake Chad Basin showed earlier in the 1900s a remarkable resilience with the recovery of its elephant populations (Chapter 2). A return to the ‘basic pre-disturbance state’ was obviously impossible with the changes in environment that have taken place in the 70-year time-span. The initial increase in numbers of Kob antelopes in response to the reflooding, but also the recent return of some individual waterbuck antelopes, showed the resilience of the Waza-Logone floodplain area in maintaining an appropriate habitat. Pimm resilience, certainly much more than ten years, shows that this ‘antelope resilience’ is limited however. The best example of the area’s resilience became visible with the development of pastoralism in the reflooded area. Grazing intensities recovered in ten years to a pre-dam 288 Part IV – Synthesis

Before After

Range of reference image

Range of target image

Holling resilience with vs-without dam

Recovery time (=Pimm resilience)

Dam Reflooding

Note: The range of reference and target image represent their development during the study period (see text).

Figure 12.4 – Evolving States and with-vs-without comparison

cattle density as suggested by a comparison with cattle densities in other African floodplains (Scholte & Brouwer 2005).

In analogy to the discussion on restoration versus rehabilitation (12.4.2), we may conclude that the concept of a static or basic pre-disturbance state as used in both the Holling (1973, 1995) and Pimm (1983) resilience concepts is difficult to apply. A more realistic concept for resilience takes into account that changes in the pre- disturbance state have taken place, in analogy to what economists call ‘with-versus-without comparison’ in Cost Benefit Analysis (Gittinger 1982). Van den Berg et al. (2004) developed a ‘with-without scenario’ for a simple linear succession in marshland in Western Netherlands in case of a single instantaneous disturbance. ‘Evolving states’ may be a more appropriate concept that one may conceive as the pre-disturbance state that has continued developing as could have been expected in the absence of the disturbance. Imagine, for example, that the Maga dam was not built and the area had developed with the continuing impact of factors such as drought and poaching (Table 12.1). Almost certainly the cover of Vetiveria nigritana would also have dropped from 1979 onwards (Fig. 12.4). This also holds for the size of the floodplain antelope populations. Both Vetiveria nigritana and antelopes would of course not have dropped to the same extent as with the presence of the dam. Reference areas play an important role to assess these background changes, in analogy to the monitoring of the reflooding (Chapters 3 and 6). 12 – Synthesis 289

Empirical concepts of ecosystem change, concluding remarks Holling & Gunderson (2002) stated that ‘the adaptive cycle should be considered as metaphor to help interpreting events and their gross causes, and not as a [causal] theory’. This did not, however, stop them as well as others using panarchy as a theoretical framework for ecosystem and societal change (e.g. Gallopin 2002). Laplante (2004) rightly questioned such straightforward interpretations towards a unique ecological theory. The straightforward discussion on results of the reflooding in section 12.1 has a relatively high causal content, contrasting the discussions on panarchy. Not by coincidence I started this chapter by presenting conclusions based on field data and brought them subsequently together in an overall comparison. Only afterwards a comparison was made with empirical concepts of ecosystem change that would have been of little value if not fed with field experiences. The vegetation rehabilitation experiences, for example, showed that the concept of target communities (‘pre-dam perennial grasslands’) has limits because of their dynamics (Fig. 12.4). Ultimately we deal with individual species and their traits and not vegetation communities that are important for rehabilitation (Bakker et al. 2000; Chapter 3). The concept of ‘State & Transition’ plays an important role in the understanding of the dynamics of the system. At least in Waza-Logone, it has less relevance to understand the services its provides.

To understand functional relations one has to study individual ecosystem components often targeting individual species. The somewhat ‘dull’ description and interpretation of the floodplain rehabilitation case becomes however much more attractive when put in a hierarchical time and space dimension such as panarchy. These holistic metaphors play a complementary role in communication and exchange of experiences, contributing to vision forming and ultimately to interpretations of ecological philosophy (Laplante 2004)

12.5 The key findings of this study and the ecosystem approach

A discussion on concepts of ecosystem change and management, should include the normative ecosystem approach, formally endorsed by the fifth Conference of Parties (COP-5) of the Convention of Biological Diversity in 2000 (www.biodiv.org/ decisions; Shepherd 2004). The 12 principles of the Ecosystem Approach reflect the present state of ‘practical wisdom’ on ecosystem management, obviously an improvement from the inflexible approaches that dominated till recently. Korn et al. (2003) developed several criteria to create an enabling environment for the application of the ecosystem approach. Shepherd (2004) facilitated the implementation of the ecosystem approach by categorising its 12 principles into five implementation steps. 290 Part IV – Synthesis

The potential use of the ecosystem approach was recently voiced by the Director General of IUCN:

‘The ecosystem approach has emerged as a core element that can bring the widespread and diverse activities of this movement [IUCN] together, not by sub- suming any individual activity, but by providing an orientation point. Anything from the management of a local wetland to issues such as global climate change can be aligned under the ecosystem approach. This allows us to speak in terms and concepts that can influence an entire global debate... and indeed other sectors’ (Steiner 2004).

In restoration practice the principle ‘management must recognise that change is inevitable’ is especially relevant. For conservation practice this holds for the principle ‘management should be decentralised to the lowest appropriate level’. As such the ecosystem approach is an excellent framework for analysis, especially if reflec- tions such as of de Groot (2001), Korn et al. (2003) and Shepherd (2004) are kept in mind.

The Waza-Logone project was implemented well before the ecosystem approach was conceived and operationalised. It remains speculative if its knowledge beforehand would have influenced the Waza-Logone approach. It is, however, the present lack of relevant examples that makes it difficult to interpret the present state of the Ecosystem approach into practical recommendations. In 2002, the sixth Conference of Parties therefore requested the submission of case studies and lessons-learnt on the development and implementation of the ecosystem approach on national and regional levels (www.biodiv.org/decisions). Below I will present the main lessons of this study, allowing the further development of the ecosystem approach, by first addressing the five key findings of the present study, namely:

1 The impact of different response times of ecosystem components to the rehabilitation results (resulting in ‘rehabilitation instead of restoration’ and ‘win-loose situations’). 2 The Ideal Free Pre-emptive distribution, probably reigning in other (tropical floodplains) systems as well, makes it particularly hard to set resources aside for conservation purposes (failure of convention with pastoralists on refraining from grazing inside Waza NP to be compensated by reflooding benefits). 3 The risk of human population build-up at the border of protected areas, accelerated by the integration of conservation and evelopment (immigration time- bomb). 4 Need to assure the commitment of all parties in consensual protected area management (lack of governmental commitment). 12 – Synthesis 291

5 Need to have adequate intervention capacities of all parties (training of protected area personnel in planning and community conservation)

Finding 1 has a clear link with principle 8 ‘Recognizing the varying temporal scales and time-lags that characterise ecosystem processes, objectives for ecosystem management should be set for the long-term’ (Table 12.2). Finding 2 shows the difficulty to seek an appropriate balance between use and conservation of resources (principle 10, Table 12.2). Finding 3 may be linked, if some adaptations are made, to principle 3, ‘ecosystem managers should consider the short and long term effects (actual or potential) of their activities including [on] adjacent and other ecosystems’. Findings 4 and 5 are somewhat related to principle 12 ‘The ecosystem approach should involve all relevant sectors of society and scientific disciplines’. Findings 4 and 5 further provide input in the categories of measures proposed by Korn et al. (2003), namely ‘institutional strengthening and co-operation’ and ‘Information – knowledge – capacity’.

In addition to above key findings, other comments linked to the 12 ecosystem principles are made based on insights from this study (Table 12.2). Of these comments, I will in the last part of this chapter specifically address the ones on adaptive management, applied by Shepherd (2004) as implementation steps of the ecosystem approach. Adaptive management is particularly relevant as overarching framework for our experiences. It needs further development to become fully operational, however.

12.6 Towards adaptive management of success

Introduction As Lee (1993) puts it

‘Adaptive management is an approach to natural resource policy that embodies a simple imperative: policies are experiments; learn from them... Adaptive management takes uncertainty seriously, treating human interventions in natural ecosystems as experimental probes. Its practitioners take special care with information. First they are explicit about what they expect, so that they can design methods and apparatus to make measurements. Second they collect and analyse information so that expectations can be compared with actual... Finally, they transform comparison into learning – they correct errors, improve their imper- fect understanding, and change action and plans. Linking science and human purpose, adaptive management serves as a compass for us to use in searching for a sustainable future’. 292 Part IV – Synthesis

Table 12.2 – Findings of this study as elaboration upon the principles of the Ecosystem Approach

No Ecosystem Approach Principle Findings of this study

1 The objectives of management of land, water True, although the term ‘societal’ leads to differ- and living resources are a matter of societal ent interpretations (e.g. de Groot 2001) choice

2 Management should be decentralised to the ‘Lowest appropriate’ may also mean higher-level lowest appropriate level government (e.g. in land use planning and security issues)

3 Ecosystem managers should consider the Human population build-up should be included in effects (actual or potential) of their activities this principle on adjacent and other ecosystems

4 Recognising potential gains from management Economic context should be broadly interpreted there is usually a need to understand and and does, for example, not necessarily imply a manage the ecosystem in an economic context monetary context

5 Conservation of ecosystem structure and ‘Conservation’ should be broadly interpreted and functioning, in order to maintain ecosystem include rehabilitation services, should be a priority target of the ecosystem approach

6 Ecosystems must be managed within the limits ‘Ecosystem’ should be defined broadly such as to of their functioning enable removal of undesirables states (eg. annual herbs State C Box.12.1)

7 The ecosystem approach should be undertaken True, varying from village level to Lake Basin at the appropriate spatial and temporal scales levels. Care should be taken to be flexible to take into account new insights on what appropriate scales are (see 12.6)

8 Recognising the varying temporal scales and Very true, different response times will make short lag-effects that characterise ecosystem term ‘restoration’ objectives out of reach processes, objectives for ecosystem management should be set for the long term

9 Management must recognise that change is Very true, hence the need for adaptive manage- inevitable ment, guided by a long term vision (12.6).

10 The ecosystem approach should seek the True, but Ideal Free Distribution disputes the ‘bal- appropriate balance between and integration ance’ concept. of conservation and use of biological diversity

11 The ecosystem approach should consider all Note the importance of historic data that comple- forms of relevant information, including ments and sometimes even contradicts ‘scientific’ scientific and indigenous and local knowledge, and local knowledge (case of elephants, Chapter innovation and practices 2).

12 The ecosystem approach should involve all – Sometimes, relevant sectors should be excluded relevant sectors of society and scientific (see case of park village) disciplines – Need to assure the commitment of ‘all’ parties in consensual protected area management – Need to have adequate intervention capacities of ‘all’ parties 12 – Synthesis 293

Following Kessler (2003), the above definition of adaptive management runs the risk of leading to a sum of small adaptations that may end, finally, in an undesired situation. Adaptive management therefore needs to be guided by a long-term vision. Adaptive management then entails the use of monitoring for continuous adaptation plus inputs for reflection for the identification, re-identification and work towards this long-term vision.

In this section, I review the management approach that was pursued in the reflooding experiment in the light of adaptive management principles. In particular I want to discuss which insights obtained from monitoring have led to changes in the field. For this purpose I distinguish between three levels of abstraction: concrete actions, planning, and targets. I will first address the difficulty to monitor in a project context beyond a time-span of five years. I will subsequently illustrate the impact of monitoring on concrete actions and policies and on planning. A revisit of the earlier discussed hypotheses for future reflooding allows the development of an overall vision of floodplain rehabilitation. Lessons from this approach allow a further development of a planning framework for managed flood releases in African river systems.

Monitoring, studies and concrete actions and policies In 1993 we hypothesised that restoration of the floodplain environment and its land use, to the pre-Maga dam situation, would take five years (Chapter 1; Wesseling et al. 1994). A considerable part of the (financial) resources of the Waza-Logone project during the first two project phases (1992-1997) were allocated to the monitoring of this scenario4 and to discuss and formulate larger-scale reflooding options. The general appreciation of the reflooding in these years was overwhelmingly positive (Acreman & Pirot 2004). Difficulties to assure the necessary financial sources for large scale reflooding caused the shift of attention of the Waza-Logone project from 1997 onwards to more mainstream development interests such as small- scale projects for sanitation and agriculture (Table 9.4). Local communities as well as project personnel lost interest in monitoring beyond a time span of 3-5 years in which tangible results had to be presented and continuing financing had to be assured.

Many obtained insights of monitoring were put into action, often even before being formally reported. Decisions and subsequent actions taken on the management of the Waza-Logone floodplain can be summarised as follows:

4 Not restricted to hypotheses of resources analysed in this thesis. Fisheries and agriculture were amongst the extensively monitored resources as well. Project monitoring also encompassed social relations, including gender, see also Loth (2004). 294 Part IV – Synthesis

Wildlife policies • Based on results from the developing grazing pressure, efforts were undertaken that resulted in the convention with pastoralists on refraining from grazing inside Waza NP (Chapter 7) • Based on the monitoring of impact of reflooding, the ‘social buffering policy’ was explored (Chapter 9). • Based on results from village PLAs (Scholte et al. 1999b), the ‘carrot & stick policy’ with regard to Baram village inside Waza National Park, was implemented (Chapters 8,9)

Capacity building • Based on the observations on the poorly developed skills of park personnel, a course in People-oriented Conservation and Planning was developed and implemented, initially for Waza personnel (1996), later for personnel of other protected areas as well. • Experiences with these courses highlighted the need for curriculum and institutional reform of the Garoua wildlife training college.

Above-presented monitoring results of the reflooding years 1993-1997 were integrated in the management plan of Waza National Park (Chapter 9) and the progressing Cost-Benefit Analysis of large-scale reflooding options (Wesseling et al. 1994; IUCN 1999; Emerton 2004). Based on continued monitoring from 1998 till 2003, studies were undertaken outside the project framework. Reported in the preceding chapters, the resulting progressing insights can be summarised as follows: • In addition to the changed vegetation composition in an area of 180 km2 (Chapter 3), the importance of water raising in an area of 600 km2 was clarified (Chapter 4). The increase in vegetation production explained the larger than initially calculated impact of the reflooding on livestock grazing and fisheries. • The importance of the generally increasing trend of waterbirds recovery from the 1980s drought was highlighted by the comparison with developments in waterbird populations elsewhere in West Africa (Chapter 5). Most probably, the same holds for other wildlife, including floodplain antelopes, but this has been difficult to show due to the lack of reference areas. • The divergence of developments of floodplain resources from 1997 onwards (Figure 12.1, see above; Chapter 8) reinforced the necessity of proposed actions in the management plan.

These decisions and actions were undertaken to meet the overall goal of the floodplain rehabilitation in achieving both sustainable development and biodiversity conservation. Progressing insight showed the need to pursue corrections to plan- 12 – Synthesis 295 ning, to the rehabilitation targets and to the overall vision on floodplain rehabilitation that will be discussed below.

Monitoring and planning Probably the most far-reaching decision on floodplain rehabilitation, motivated by the difficulty to translate initial survey results into actions in the field, was to implement the 1994 reflooding (Chapter 1). This ‘pilot release’ became the identity of the Waza-Logone project and it is difficult to imagine that it had initially not even been programmed. Based on the preliminary monitoring results of this first reflooding, subsequent negotiations and the resulting consensual decision, a follow-up of the pilot reflooding was implemented by an opening of a second watercourse, the Areitekele, in 1997.

Important, yet expensive, adaptations were made of the larger-scale reflooding proposals, the so-called ‘main release’ (IUCN 1999). Monitoring showed that although flooding water reached the more central parts of Waza National Park, its arrival was delayed till well into the dry season. To allow the recovery of the more central parts of Waza National Park, possibly including Vetiveria nigritana stands, it was deemed necessary to imitate the original hydrographic pattern where floodwater reaches already during the rainy season the (western) floodplain.

An extrapolation of the preliminary reported benefits of reflooding for vegetation production (Chapter 3) and especially livestock grazing (Chapter 7) played an important role as justification of the proposed large scale reflooding options. Live- stock grazing for example was estimated to represent about 50% of the incremen- tal benefits of the various reflooding options (Emerton 2004).

Monitoring and floodplain rehabilitation targets

A target image for future (large-scale) reflooding can be regarded as the ‘em- bodiment of insights of feasible policy and management goals along temporal and spatial scales’ (Lenders et al. 1998).

The goal of the floodplain rehabilitation, voiced at the start of the reflooding, was: ‘The hydrological and ecological rehabilitation of the floodplain that involves the restoration, as far as possible, of the physical conditions on the floodplain that existed before the construction of the SEMRY infrastructure in 1979’ (IUCN 1999). Although not explicitly formulated prior to the start of the 1994 reflooding, this goal has, rather surprisingly, not evolved over the years.5 The reflooding goal was based on a single hypothesis: ‘back to the natural hydrology’, resembling the target for the Everglades restoration (Davis & Ogden 1994). The reformulated flood-

5 It should be mentioned that I was part of the project team preparing these formulations! 296 Part IV – Synthesis

plain rehabilitation hypotheses allow the adaptation of the overall floodplain rehabilitation goal, the subject of this section.

Below, I present some thoughts of how this floodplain rehabilitation target should look like taking into account above-presented hypotheses (12.1) categorised into specific temporal and spatial scales and reigning distribution models.

Different temporal scales: • The recovery of tussock grasses lags several years behind the recovery of rhizomatous perennial grasses, which in their turn lag several years behind changes in annual grasses’ composition and cover • Dry season regrowth production lags several years behind the increase in wet season above-ground biomass that in its turn lags two years behind the installation of reflooding • The recovery of ‘human-competing’ resources, if taking place at all, lags behind ‘human-supported’ resources • The impact of management planning lags several years behind immigration and other developments described above • The impacts of training lag behind management planning • Institutional adaptations lag behind curriculum requirements

Different spatial scales: • Waterbird (sub-) continent changes related to environmental conditions versus site-specific situations such as protection. • The recovery of livestock whose recovery is regionally determined versus the recovery of antelope populations that is generally based on locally determined factors only.

Different distribution models: • For pastoralism, but possibly for other land uses such as fisheries as well, the Ideal Free Pre-emptive Distribution model was postulated. This model predicts that animals and plants will distribute evenly according to availability of resources. This is a strong counter force against the reservation of resources for conservation purposes.

From reference to target image and back: developing a vision for the rehabilitation and management of African floodplains The reference image(s) described in Chapter 2 was obviously not within the target of a rehabilitation program. Progressing knowledge, obtained by the conduct of historical and other surveys and the monitoring of the reflooding, has led to the reformulation of the reference and target images (respectively Chapter 2 and 12.1). Only after this first stage of reflooding probing (1993-1999), have we been able to formulate more precise floodplain rehabilitation hypotheses and targets. These 12 – Synthesis 297

hypotheses and targets should be adapted based on progressing insights in the future. In situations with poorly known reference situations and insufficient insight to formulate a target for rehabilitation, one should consider floodplain rehabilitation as an iterative process. One should pass not only from trial & error into passive and subsequently active adaptive management (Walter & Holling 1990), but also continue along in this reiterative spiral, passing on to future trial & error, passive adaptive management stages. As hysteresis is almost always present, therefore prediction, even based on events in the past, will often be poor.

Integrating these insights in a planning framework for managed flood releases, instead of the (one) circle framework proposed by Acreman & Pirot (2004), I pro- pose the spiral as shown in Figure 12.5.

CHANGING THEORIES (reflected in empirical concepts)

adaptive vision

ACTIONS

continuous monitoring

CHANGING OUTSIDE WORLD

Figure 12.5 – Planning framework for flood releases

A vision encompasses not only the aspiration of (developing) targets but also envisages the enabling environment in which rehabilitation and management targets can move. I express my vision for (further) floodplain rehabilitation in Waza- Logone as follows:

Floodplain rehabilitation should be conducted in a probing approach with simple and long-term (‘continuous’) monitoring that allows interventions and regular redefinitions of targets. Because of expected imbalances in Conservation versus Development, appropriate mitigating measures should be prepared beforehand, e.g. by the designation of distinct areas for conservation and other competing land uses. In addition one may consider zoning with exclusive rights for present local communities that form a social buffer against the installation of newcomers. For- mal agreements on the exclusive use of resources may facilitate such measures. To enhance the efficiency of conservation, it will be important to increase the intervention capacity of protected area personnel well before major changes such as those induced by floodplain rehabilitation, are undertaken. This often implies that institutional arrangements have to be taken up early. In the long term one aspires towards governmental commitment for conservation and local development in the way they have been mobilised in Cameroon for security in the late 1990s. In addition one should strive to more coherence between local, national and international conservation policies and interventions. References

ABCG (2004). Meeting on Migration and Conservation: Issues and Interventions. www.ABCG.org (accessed Feb. 2005). Abbott, A. (1996). Conservationists seek compromise over new threat to Galapagos. Nature 383: 3. Acreman, M. and J.Y. Pirot (2004). Tools and practices for Floodplain Restoration. Pp. 125- 144 in Loth, P. (ed.) The return of the water. Restoring the Waza Logone floodplain in Came- roon. IUCN, Gland, Switserland. Acreman, M.C. and G.E. Hollis (1996). Water management and wetlands in sub-Saharan Africa. IUCN, Gland, Switzerland. Adams, W.M. (1996). Economics and hydrological management of floodplains. Pp. 21-33 in Acreman, M.C. and G.E. Hollis (eds.) Water Management and Wetlands in Sub-saharan Africa. IUCN, Gland, Switserland. Adams, W.M. and T. McShane (1992). The Myth of Wild Africa. Conservation without Illusion. Norton & Company, New York, USA. Adams, W.M. and D.Hulme (2001). If community conservation is the answer in Africa, what is the question? Oryx 35: 193-200. Allegret, M. (1993). Carnets du Congo. Voyage avec André Gide. CNRS Litterature, Paris, France. Allo, A. (1978). The Garoua College of Wildlife Management. Parks 3: 8-9. Alpert, P. (1993). Conserving biodiversity in Cameroon. Ambio 22: 44-49. Amougou (2002). Rapport de la mission d’évaluation du plan d’aménagement du PN Waza. In- ternal Report, Waza, Cameroon. Anderson, D. and R. Grove (eds.) (1989). Conservation in Africa. People, policies and practice. Cambridge University Press, Cambridge, UK. Anon. (1996). Campfire News, December 1996, Harare, Zimbabwe. Anonymus (2004). Peering at the Future. Conservationists are, reluctantly, starting to get real. The Economist June, 19th 2004. ANZECC (2000). Working Group on National Parks and Protected Areas. Best practice in Protected Area Management Planning. (available at www.pandora.nla.gov.au/tep/24147). Ash, J.S. and J.E. Miskell (1998). Birds of Somalia. Pica press, Sussex, UK. Azarya, V., A. Breedveld, M. de Bruin and H. van Dijk (1999). Pastoralists under pressure? Fulbe societies confronting change in West Africa. Brill, Leiden, The Netherlands. Bakker, J.P., A.P. Grootjans, M. Hermy and P. Poschlod (2000). How to define targets for eco- logial restoration? Introduction. Applied Vegetation Science 3: 3-6. Bancroft, G.T. (1989). Status and Conservation of wading birds in the Everglades. American Birds 43: 1258-1265. Banham, W. (1999). Modularisation: will this affect the way training is done: the experience of Mweka. Unpublished report, SADC Wildlife Training Programme Training workshop, Harare, Zimbabwe. Barbraud, C. and H. Hafner (2001). Variation des effectifs nicheurs de hérons pourprés Ardea purpurea sur le littoral Méditerranéen Français en relation avec la pluviométrie sur les quartiers d’hivernage. Alauda 69: 373-380. Barrat-Segretain, M.H. and C. Amoros (1996). Recolonisation of cleared riverine macrophyte patches: importance of the border effect. J. Veg. Sci. 7: 769-776. Barrett, C.B. and P. Arcese (1995). Are Integrated Conservation-Development Projects (ICDPs) Sustainable? On the Conservation of Large Mammals in Sub-Saharan Africa. World Development 23: 1073-1084. Barrow, E. and M. Murphree (2001). Community conservation, from Concept to Practice. Pp. 24-37 in Hulme, D. and M. Murphree (eds.) African Wildlife & Livelihoods. James Currey, Oxford, UK. Barth, H. (1857). Reisen und Entdeckungen in Nord-und Central-Afrika in den Jahren 1849 bis 1855. Tagebuch seiner im Auftrag der Brittschen Regierung unternommenen Reise. Volume 1-5. Gotha Justus Perthes. Translations in English taken from the first American edition: Barth, H 1857. Travels and discoveries in North and Central Africa,being a journal of an expedition undertaken in the years 1849-1855. New York: Harper and brothers, Volume 1-3. Batabyal, A.A. (1998). The Concept of Resilience: retrospect and prospect. Environment and Development Economics 3: 234-239. Bates G.L. (1927). Notes on some birds of Cameroon and the Lake Chad Region: their status and breeding-times. Ibis 3: 1-65. Batisse M. (1982). The biosphere reserve: A tool for environmental conservation and management. Environmental Conservation 9: 101-111. Bauer, H. (2003). Lion Conservation in West and Central Africa: the case of Waza National Park, Cameroon. Thesis Leiden University, The Netherlands. Beauvilain, A. (1981). Élevage et éleveurs du grand yaéré (Nord Cameroun). Revue de geogra- phie du Cameroun 2: 163-176. Beauvilain, A. (1989). Nord-Cameroun. Crises et Peuplements. Volumes 1 and 2. Thesis Univer- sity of Rouen, France. Beauvilain, A. (1995). Tableau de la pluviometrie dans les bassins du Tchad et de la Benoué. Tableaux et documents scientifiques du Tchad. Documents pour la recherche III. CNAR, N’djamena, Chad. Beesley J.S.S. (1972). A nesting colony of Black-headed Herons Ardea melanocephala in Arusha, Tanzania. Bull. E. Afr. Nat. Hist. Soc. 1972: 47-48. Behnke Jr., R.H., I. Scoones and C. Kerven (1993). Range ecology at disequilibrium: new models of natural variability and pastoral adaptation in African Savannas. Overseas Development Institute, London, UK. Beilfuss, R., E. Williams, T. Dodman (2003). Status survey and Conservation Action Plan for the Black Crowned crane Balearica pavonina. International Crane Foundation, Wetlands International, Wageningen, The Netherlands. Benchley, P. and S. Franklin (2001). Galapagos. Paradies in Gefahr. National Geographic Deutschland, February 2001. Bergin, P. (2001). Accommodating New Narratives in a Conservation Bureaucracy. TANAPA & Community Conservation. Pp. 89-105 in Hulme, D. and M. Murphree (eds.) African Wildlife & Livelihoods. James Currey, Oxford, UK. Bird, C. and S. Metcalfe (1995). Two views from Campfire in Zimbabwe’s Hurungwe District. Wildlife and Development Series No. 5. IIED, London, UK. Blokhuis, W.A. (1993). Vertisols in the Central Clay Plain of the Sudan. PhD Thesis Wagenin- gen University, The Netherlands. Blom, A. (2001). The Impact of Tourism on Protected Area Management and the Local Eco- nomy in Dzanga-Sangha (Central African Republic). J. Sust. Tourism 8: 175-189. Blom, C.W.P.M. and L.A.C.J. Voesenek. (1996). Flooding: the survival strategies of plants. TREE 11: 290-295. Bomba, C.M. and P. Bigombo Logo (2001). Les Droits des Populations dans la Legislation Fo- restiere Camerounaise: Quels Acquis? Quelles Insuffisances? Bulletin Arbres, Forests et Com- munautres Rurales 22: 47-62. References 301

Borrini-Feyerabend, G., M., Taghi Farvar, J-C. Nguinguiri and V. Ndangang (2000). La gestion participative des ressources naturelles: Organisation, Négotiation et Apprentisssage par l’Action. Kasparek Verlag, Heidelberg, Germany. Boudet, G. (1983). Manuel sur les pâturages tropicaux et les cultures fourragères. Manuels et précis d’élevage. IEMVT, Ministère de la Coopération. La documentation française, Paris, France. Bouton, S.N. and P.C. Frederick (2003). Stakeholders’ Perceptions of a Wading Bird Colony as a Community Resource in the Brazilian Pantanal. Conservation Biology 17: 297-306. Braack, L. (1997). A revision of parts of the management plan for Kruger N.P. Vols. VII. & VIII (www.parks-sa.co.za). Braaksma, S., M.F. Morzer Bruyns and G.A. Brouwer (1950). Overzicht van de broedkolonies van de Blauwe Reiger Ardea cinerea L. in The Netherlands. Ardea 38: 135-162. Brabant, P. and M. Gavaud (1985). Les sols et les ressources en terre du Nord-Cameroun. MES- RES/IRA-ORSTOM, Yaoundé/Paris, France. Brandon, K.E. and M. Wells (1992). Planning for people and parks: Design Dilemmas. World Development 20: 557-70. Brandon, K. (1997). Policy and Practical Considerations in Land-use Strategies for Biodiversity Conservation. Pp. 90-114 in Kramer, R., C.van Schaik and J. Johnson (eds.) Last Stand. Oxford University Press, New York, USA. Braukamper, U. (1996). Strategies of environmental adaptation and patterns of transhumance of the Shuwa Arabs in the Nigerian Chad Basin. Nomadic Peoples 39: 53-68. Bredenbeek, J. (ed.) (2004). Wader and Waterbirds in the floodplain of the Logone Cameroon and Chad, February 2000. WIWO report, Zeist, The Netherlands. Breen, C.M., K.H. Rogers and P.J. Ashton (1988). Vegetation processes in swamps and flooded plains. Pp. 223-247 in Symoens, J.J. (ed.) Vegetation of inland waters. Handbook of Vegetation Science 15/1. Kluwer, Dordrecht, The Netherlands. Breman, H. and C.T. de Wit. (1983). Rangeland productivity and exploitation in the Sahel. Science 221: 1341-1347. Breman, H. and N. de Ridder (1991). Manuel sur les pâturages des pays sahéliens. Karthala, Paris, France. Bridgewater, P., A. Philips, M. Green and B. Amos (1996). Biosphere Reserves and the IUCN System of Protected Area Management Categories. IUCN, Gland, Switserland. Broadbent, J.A. (1971). Additions to the avifaunas of Waza (Cameroun) and Lake Natu (Soko- to). Bull. Niger. Orn. Soc. 8: 58-61. Brockington, D. (2002). Fortress Conservation: The Preservation of the Mkomazi Game Reserve, Tanzania. James Currey, Oxford, UK. Brouwer, J. and W.C. Mullié (2001). A method for making whole country waterbird population estimates, applied to annual waterbird census data from Niger. Ostrich supplement 15: 73-82. Brouwer, J., W.C. Mullié and P. Scholte (2003). White Storks wintering in Chad, northern Cameroon and Niger: a comment on Berthold et al. (2001). Ibis 145: 499-501. Brown, M and A. Singer (1991). La gestion des zones tampons en Afrique. PVO NGO/NRMS. Unpublished report + video Brown, L.H., E.K. Urban and K. Newman (1982). The Birds of Africa, Vol I. Academic Press, London, UK. Brown, M. and B. Wyckoff-Baird (1992). Projets Integrés de Conservation de la Nature et de Développement. World Wildlife Fund, The Nature Conservancy and World Resources In- stitute. WWF publications, Baltimore, USA. Bruner, A.G., R.E Gullison, R.E., Rice and G.A.B da Fonseca (2001). Effectiveness of Parks in Protecting Tropical Biodiversity. Science 291: 125-128. 302 Floodplain rehabilitation and the future of Conservation & Development

Brunner J. and F. Ekoko (2000). Cameroon in the Right conditions Pp. 59-80 in Seymour, F.J. and N.K. Dubash (eds.) The World Bank, Structural Adjustment and Forest Policy Reform. WRI, Washington DC, USA. Bruijn, M. de, I. van Halsema and H. van den Hombergh (1997). Gender and Land Use. Di- versity in Environmental Practices. Thesis publishers, Amsterdam, The Netherlands. Bruwer, C., C. Poultney and Z. Nyathi (1996). Community based hydrological management of the Phongolo floodplain. Pp. 199-211 in Acreman, M.C. and G.E. Hollis (eds.) Water Management and Wetlands in Sub-saharan Africa. IUCN, Gland, Switserland. Bunting, A.H. and J.D. Lea (1962). The Soils and Vegetation of the Fung. J. of Ecol. 50: 529- 558. Burnham, K.P., D.R. Anderson and J.L. Laake (1980). Estimating of density from line transect sampling of biological populations. Wildl. Mon. 72: 1-202. Burnham, P. (1996). The politics of cultural difference in northern Cameroon. Smithsonian In- stitution Press, Washington DC, USA. Butler, R.W., J.A. Kushlan and I.J. Davidson (2000). Herons in North America, Central Ame- rica and the West Indies. Pp. 151-176 in Kushlan, J.A. and H. Hafner (eds.) Heron Conser- vation. Academic Press, London, UK. Caldecott, J. (1996). Designing Conservation Projects. Cambridge University Press, Cambridge, UK. Cannon, J.R., J.M. Dietz and L.A. Dietz (1996). Training conservation biologists in human interaction skills. Conservation Biology 10: 1277-1282. Casimir, M. J. (1992). The determinants of rights to pasture: territorial organisation and ecological constraints. Pp. 153-204 in Casimir, M.J and A. Rao (eds.) Mobility and territoriality : social and spatial boundaries among foragers, fishers, pastoralists and peripatetics. Berg, New York, USA. Catling, D. (1992). Rice in deep water. IRRI. The Macmillan Press Ltd, London, UK. CCTA. (1953). Commission for Technical Co-operation in Africa south of the Sahara. Proceed- ings of the Third International Conference for the protection of the Fauna and Flora of Africa. Bukavu, October 1953. Bruxelles, Belgium. Cellot, B., F. Mouillot and C.P. Henry (1998). Flood drift and propagule bank of aquatic ma- crophytes in a riverine wetland. J. Veg. Sci. 9: 631-640. Chambers, R. (1998). Afterword. Pp. 197 in Holland, J. and J. Blackburn (eds.) Whose voice? Participatory research and policy change. IT Publications, London, UK. Chambers, R. and J. Jiggins (1987). Agricultural research for resource poor farmers part I. Trans- fer of Technology and farming system research. Agricultural Administration and Extension 27: 35-52. Cheke, R.E. and J.F. Walsh (1996). The Birds of Togo. B.O.U. Checklist No. 14. Tring, UK. Child, G.S. (1976). Training Personnel for National Parks. Parks 1: 14-16. Child, G. (1995). Wildlife and People: the Zimbabwean Success. How the conflict between animals and people became progress for both. Wisdom Foundation, Harare, New York. CIRAD-EMVT (1996). Atlas d’élevage du Bassin du Lac Tchad/Livestock Atlas of the Lake Chad Basin 1997. CTA, Wageningen, The Netherlands. Claridge, G. (1999). Protected Areas Management Planning. Tigerpaper 26: 15-17. Clarke, P. (1999). Park Management Planning in Africa. Oryx 33: 281-282. Connah, G. (1981). Three thousand years in Africa. Man and his environment in the Lake Chad region of Nigeria. New studies in archeology. Cambridge University Press, Cambridge, UK. Cracknell, B.E. (2000). Evaluating Development Aid. Sage Publications, New Delhi, India. References 303

Crivelli, A., P. Grillas, H. Jerrentrup and T. Nazirides (1995). Effects on fisheries and waterbirds of raising water levels at Kerkini reservoir, a Ramsar site in N. Greece. Environmental Management 19: 431-443. Cronk, J.K. and M.S. Fennessy (2001). Wetlands Plants. Biology and Ecology. CRC Press/Lewis Publishers, Boca Raton, Florida, USA. Crowder, L.K., W.I. Lindley, T.H. Bruening and N. Doron (1998). Agricultural Education for Sustainable Rural Development: Challenges for Developing Countries in the 21st Century. The Journal of Agricultural Education and Extension 5: 71-84. Cunnison, I. (1960). The Social Role of Cattle. The Sudan Journal of Veterinary Science and Ani- mal Husbandry 1: 8-25. Dahl, G., and A. Hjort (1976). Having herds: pastoral herd growth and household economy. Stock- holm (Sweden): Department of Social Anthropology, University of Stockholm, Norway. Davis, S.M. and J.C. Ogden (1994). Everglades. The Ecosystem and Its Restoration. St. Lucie Press, Delray Beach, Florida, USA. De Bie, S. (1991). Wildlife Resources of the West African Savanna. Wageningen Agric. Papers. 91-2. Wageningen, The Netherlands. Dejace, P., L. Gauthier and P. Bouché (2000). Les populations de grands mammiferes et d’autruches du Parc National de Zakouma au Tchad: Statuts et Tendances Evolutives. Rev. Ecol (Terre Vie) 55: 305-320. Denny, P. (1985). The ecology and management of African wetland vegetation. Junk, Dordrecht, The Netherlands. Denham and Clapperton. 1826 (1985 reprint). Narrative of Travels and Discoveries in Northern and Central Africa in the years 1822, 23 and 24. Darf publishers limited, London, UK. Denny, P. (1993). Wetlands of Africa: Introduction. Pp. 1-31 in Whigham, D.F., D. Dykyjova and S. Hejny (eds.) Wetlands of the World: Inventory, ecology and management I. Kluwer Aca- demic Publishers, The Netherlands. Denny, P. (1994). Biodiversity and wetlands. Wetlands Ecology and Management 3: 55-61. Denny, P. (2004). Wetlands are our life-blood. Valedictory Address UNESCO-IHE. Dochy, F.J.R.C., M.S.R. Segers and G. Moerkerke (1997). The importance of prior knowledge and assessment for increasing efficiency of the learning processes, especially in ‘problem- based’ powerful learning environments. Pp. 73-96 in Charrier, A. and R. van Haarlem (eds.) Biodiversity, an issue in Higher Education. Proceedings of the third European Conference on Higher Education for Agriculture. Wageningen University, Wageningen, The Nether- lands. Dodman, T. and C.H. Diagana (2003). African Waterbird Census/Les dénombrements des oiseaux d’eau en Afrique 1999, 2000, 2001. Wetlands Global Series 16, Wageningen, The Nether- lands. Dragesco J. (1960). Notes biologiques sur quelques oiseaux d’Afrique Equatoriale. Alauda 28: 81-92, 262-273. Dragesco J. (1961). Oiseaux des savanes d’Afrique équatoriale. Oiseau Rev. fr. Orn. 31: 179-192, 261-271. Drent, R.H. and H.H.T. Prins (1987). The herbivore as prisoner of its food supply. Pp. 133-149 in van Andel, J., J. Bakker and R.W. Snaydon (eds.) Disturbance in grasslands; species and population responses. Dr. W. Junk Publishing Company, Dordrecht, The Netherlands. Driessen, P.M. and R.Dudal (1989). Lecture Notes on the major soils of the world. Agricultural University Wageningen and Katholieke Universiteit Leuven, Wageningen, The Netherlands. Drijver, C.A. and M. Marchand. (1986). Taming the floods: environmental aspects of floodplain development in Africa. Nature and Resources 22: 13-22. 304 Floodplain rehabilitation and the future of Conservation & Development

Drijver, C.A., J.C.T. van Wetten and W.T. de Groot (1995). Working with Nature: Local Fishery Management on the Logone Floodplain in Chad and Cameroon. Pp 29-45 in van der Bree- mer, J.P.M., C.A. Drijver and L.B. Venema (eds.) Local Resource Management in Africa. Wiley, Chichester, UK. Drijver, C.A. and E. Kouahou (1995). What a small dam can do: local initiative in the Logone Floodplain, Cameroon. Pp. 179-184 in Roggeri, H. (ed.) Tropical Freshwater Wetlands. A guide to current knowledge and Sustainable Management. Kluwer Academic Publishers, Dor- drecht, The Netherlands. Dupuy, A.R. and P. Dupont (1984). Les guardiens de la vie sauvage. WWF-UICN, Gland, Swit- serland. Dijkstra, A.J., W. Ganzevles, G.J. Gerritsen and S. de Kort (2002). Waders and waterbirds in the floodplains of the Logone, Cameroon and Chad, January-February 1999. WIWO report 75, Zeist, The Netherlands. Dzingirai, V. (1996). Every man must resettle where he wants: The Politics of settlement in the context of community wildlife management programme in Binga, Zimbabwe. Zambe- zia 23: 19-30. East, R. (ed.) (1999). African Antelope database 1998. IUCN/SSC Antelope Specialist Group. IUCN, Gland, Switserland and Cambridge, UK. Ebohon, O.J., B.G. Field and R. Ford (1997). Institutional deficiencies and capacity building constraints: the dilemma for environmentally sustainable development in Africa. Int. J. of Sust. Devel. and World Ecol. 4: 204-213. Egbe, S.E. (2001). The law, communities and wildlife management in Cameroon. Rural Devel- opment Forestry Network Paper 25e 1-11 Overseas Development Institute, UK. Available at www.odi.org.uk/fpeg/publications/rdfn/25/rdfn-25e-i.pdf Ekoko, F. (2000). Balancing Politics, Economics and Conservation: The Case of the Cameroon Forestry Law Reform. Pp. 129-152 in Doornbos, M., A. Saith and B. White (eds.) Forests: Nature, People, Power. Blackwell, Oxford, UK. El-Ashry, M.T. (2001). Foreword. Pp x in Fishpool, L.D.C. and M.I. Evans (eds.) Important Bird Areas in Africa and Associated Islands: Priority Sites for Conservation. Pisces publications and BirdLife International (Birdlife Conservation Series No. 11), Newbury and Cambridge, UK. Elgood, J.H., J.B. Heigham, A.M. Moore, A.M. Nason, R.E. Sharland and N.J. Skinner (1994). The Birds of Nigeria. British Ornithologists’ Union, UK. Ellenbroek, G.A. (1987). Ecology and productivity of an African wetland vegetation. Junk publishers, Dordrecht, The Netherlands. Ellery, K., W.N. Ellery, K.H. Rogers and B.H. Walker (1991). Water depth and biotic insula- tion: Major determinants of black-swamp plant community composition. Wetlands Ecology and Management 1: 149-162. Emerton, L. (2004). Economic Justification of Additional Floodplain Re-inundation. Pp. 105- 121 in Loth, P. (ed.) The return of the water. Restoring the Waza Logone floodplain in Cameroon. IUCN, Gland, Switserland. Esser, J.D. and L.P. van Lavieren (1979). Importance, repartition et tendance evolutive des populations des grands herbivores et de l’autruche dans le parc national de Waza, Came- roun. Rev. Ecol (Terre Vie) 33: 3-26. Est, D. van (1997). The changing use and management of the floodplain environment by Mousgoum women in North Cameroon. Pp. 9-26 in de Bruyn, M., I. van Halsema and H. van den Hombergh (eds.) Gender and Land Use; Diversity in Environmental Practises. Thela Publishers, Amsterdam, The Netherlands. Est, D.M.E. van (1999). Vissen in Andermans Vijver. PhD thesis Leiden University, Leiden, The Netherlands. References 305

Est, van D. and P. Scholte (2001). Linking social and ecological diversity: organisations for adaptive natural resource management in the Waza-Logone floodplain (North Cameroon). Pp. 55-65 in Ali, M., P. Loth, H. Bauer and H. de Iongh (eds.) Management of fragile ecosystems in the North of Cameroon: the need for an adaptive approach. CEDC/CML, Leiden, The Netherlands. Estes, R.D. (1992). The behavior guide to African Mammals. The University of California Press, Berkeley, USA. Evans-Pritchard (1940). The Nuer. A description of the modes of livelihood and political institutions of a Nilotic people. Oxford University Press, New York. FAO (1977). Guidelines for soil profile description. FAO, Rome, Italy. FAO (1988). FAO/Unesco Soil Map of the World, revised legend. World Resources Report 60. FAO, Rome, Italy. Fasola, M. and F. Barbieri (1978). Factors affecting the distribution of heronries in Northern Italy. Ibis 120: 337-340. Fasola, M., H. Hafner, J. Prosper, H. van der Kooij and I.V. Schogolev (2000). Population changes in European herons in relation to African climate. Ostrich 71: 52-55. Feron, E. (1997). Les processus participatifs de gestion et de conservation de la faune en Afri- que Australe, ou la recherche d’un nouvel Eden. Le Flamboyant 42: 20-25. Fischer, F. and K.E. Linsenmair (2001a). Spatial and temporal habitat use of kob antelopes (Kobus kob kob, Erxleben 1777) in the Comoé National Park, Ivory Coast as revealed by radio tracking. Afr.J. Ecol. 39: 249-256. Fischer, F. and K.E. Lindsenmair (2001b). Decreases in ungulate population densitites. Exam- ples from the Comoe National Park, Ivory Coast. Biol. Conserv. 101: 131-135. Fischer, F. and K.E. Linsenmair (2002). Demography of a West African kob (Kobus kob kob) population. Afr. J. Ecol. 40: 130-137. Fishpool, L.D.C. and M.I. Evans (eds.) (2001). Important Bird Areas in Africa and Associated Islands: Priority Sites for Conservation. Pisces publications and BirdLife International (Bird- life Conservation Series No. 11), Newbury and Cambridge, UK. Fjeldsa, J., N.B. Burgess, S. Blyth and H.M. de Klerk (2004). Where are the major gaps in the reserve network for Africa’s mammals? Oryx 38: 17-25. Flizot, P. (1948). Les élephants des régions du Nord Cameroun et de la Bénoué. Mammalia 4: 148-151. Flizot (1962). The Waza National Park in Northern Cameroun. African Wildlife 16: 293-297. Flizot, P. (1962, 1968, 1974). Annual Reports. Provincial delegation, Garoua, Cameroon. Foppen, R.P.B., C.J.F. ter Braak, J. Verboom and R. Reijnen (1999). Dutch Sedge Warblers Acrocephalus schoenobaenus and West-African rainfall: empirical data and simulation modelling show low population resilience in fragmented marshlands. Ardea 87: 113-127. Fotso, R., F. Dowsett-Lemaire, R.J. Dowsett, COC, P.Scholte, M. Languy and C. Bowden (2001). Cameroon. Pp. 133-159 in Fishpool, L.D.C. and M.I. Evans (eds.) Important Bird Areas in Africa and Associated Islands: Priority Sites for Conservation. Pisces publications and BirdLife International, Newbury and Cambridge, UK. Francois, J., A. Rivas and R. Compère (1989). Le pâturage semi-aquatique à Echinochloa stagnina (Retz) P. Beauv. Etude approfondie de la plante ‘bourgou’ et des bourgoutières situées en zone lacustre du Mali. Bull. Rech. Agron. Gembloux 24: 145-189. Fry, C.H. (1970). Birds in Waza National Park, Cameroon. Bull. Niger. Orn. Soc. 7: 1-5. Galatowitsch, S.M. and A.G. van der Valk (1996). Vegetation and environmental conditions in recently restored wetlands in the prairie pothole region of the USA. Vegetatio 126: 89-99. 306 Floodplain rehabilitation and the future of Conservation & Development

Galindo-Leal, C. (2001). Design and analysis of conservation projects in Latin-America: an integrative approach to training. Conservation Ecology 5 (2): 16. online: www.consecol.org/ vol5/iss2/art16. Gallopin, G.C. (2002). Planning for Resilience: Scenario’s, Surprises, and Branch Point. Pp. 361-392 in Gunderson and C.S. Holling (eds.) Panarchy. Understanding transformations in human and natural systems. Island Press, Washington DC, USA. Gamassa, D.G. (1995). Mweka College of African Wildlife Management: Its role as a Regional Training Institution. Pp. 126-127 in Robinson, R. (ed.) Africa Heritage 2000: The future of Protected Areas in Africa. Proceedings of the IUCN Commission on National Parks and Protected Areas African Regional Working Session. National Parks Board, Pretoria, South Africa. Gaston, A. and D. Dulieu (1976). Etude agrostologique pour l’amènagement hydraulique pastorale des Yaérés. Etude agrostologique no. 46. IEMVT Maison Alfort. Geerling, C. (1982). Guide de terrain des ligneux sahéliens et soudano-guinéens. Mededelingen Landbouwhogeschool 82-3, Wageningen, The Netherlands. Geerling, C. and J. Bokdam (1971). The Senegal Kob (Adenota kob kob Erxleben) in the Co- moé National Park, Ivory Coast. Mammalia 35: 17-24. Gibbs, J.P. (1991). Spatial relationships between nesting colonies and foraging areas of Great Blue Herons. The Auk 108: 764-770. Gibson, C.C. (1999). Politicians & Poachers: The Political Economy of Wildlife in Africa. Cam- bridge University Press, Cambridge, UK. Gide, A. (1927, 1928). Voyage au Congo/Retour du Tchad. Editions Gallimard, reprinted in Folio 1995, France. Gilbert, D. (1971). Natural Resources and Public Relations. The Wildlife Society, Washington, USA. Gittinger, J.P. (1982). Economic Analysis of Agricultural Projects. John Hopkins University Press, Baltimore, USA. Gommes, R. and F. Petrassi (1996). Rainfall variability and drought in Sub-saharan Africa. FAO agrometeorology series working paper no.9. (www.fao.org/sd/EIdirect/Eian0004.htm). Greling, C. de (1972a). Sur les migrations et mouvements migratoires de l’avifaune Ethio- pienne, d’apres les fluctuations saisonnieres des densités de peuplements en savane Soudanienne au Nord Cameroun. Oiseau Rev. fr. Orn. 42: 1-27. Greling, C. de (1972b). New records from northern Cameroon. Bull. Brit. Orn. Club 92: 24-27. Grenier, C. (2000). Conservation contre nature. Les Iles Galapagos. IRD editions, Paris, France. Groot, A.E. (2002). Demystifying Facilitation of Multi-Actor Learning Processes. Ph.D. thesis, Wageningen University, Wageningen, The Netherlands. Groot, W.T. de (1998). Problem-In- Context. A framework for the analysis, explanation and solution of environmental problems. Pp. 22-43 in Nath, B., L. Hens, P. Compton and D. Devuyst (eds.) Environmental Management in Practice. Volume I. Instruments of Environ- mental Management. Routledge, London, UK. Groot, W.T. de (2001). Two paradigms for drylands research: a personal overview of the CEDC 2000 conference results. Pp. 11-22 in Ali, M., P. Loth, H.Bauer and H. de Iongh (eds.) Management of fragile ecosystems in the North of Cameroon: the need for an adaptive approach. CEDC and CML, Maroua, Leiden, The Netherlands. Grunwald, M. (2002). Everglades Revival Work Halted. Washington Post August 15, 2002. Gueye, B. and K. Schoonmaker Freudenberger (1991). Introduction à la methode accelerée de recherche participative (MARP). IIED, London, UK. Gunderson, L.H. and C.S. Holling (2002). Panarchy. Understanding transformations in human and natural systems. Island Press, Washington DC, USA. References 307

Gunderson L.H., C.S. Holling and G.D. Peterson (2002). Surprises and Sustainability: Cycles of Renewal in the Everglades. Pp. 315-332 in Gunderson, L.H. and C.S. Holling (eds.) Pa- narchy. Understanding transformations in human and natural systems. Island Press, Wash- ington DC, USA. Hafner, H. (2000). Heron Nest Site Conservation. Pp. 201-217 in Kushlan, J.A. and H. Hafner (eds.) Heron Conservation. Academic Press, London, UK. Hardin, G. (1968). The tragedy of the commons. Science 162: 1243-1248. Hart, T. and J. Hart (1997). Zaire: new models for an emerging state. Conservation Biology 11: 308-314. Held, J.J. den, (1981). Population changes of the Purple Heron in relation to drought in the wintering area. Ardea 69: 185-191. Hepper, F.N. (1954-1972). Flora of West Tropical Africa, 2nd edition. Crown Agents, London, UK. Herzog zu Mecklenburg, A.F. (1912). Vom Kongo zum Niger und Nil. Berichte der Deutschen Zentral Afrika-Expedition 1910-1911. Leipzig, Germany. Hiernaux, P. and L. Diarra. (1983). Pâturages de la zone d’inondation du Niger. Pp. 42-48 in Wilson, R.T., P.N. de Leeuw and C. de Haan (eds.) Recherches sur les systèmes des zones arides du Mali: résultats préliminaires. Rapport de recherche 5. ILCA, Addis Abeba, Ethiopia. Hiernaux, P. and L. Diarra (1984a). Is it possible to improve the traditional grazing in the flood plain of the Niger river in Central Mali? Proceedings 2nd International Rangeland Con- gress, Adelaide, Australia. Hiernaux, P. and L. Diarra. (1984b). Savanna burning, a controversial technique for rangeland management in the Niger floodplains of Central Mali. Pp. 238-243 in Tothill, J.C. and J.C. Mott (eds.) Ecology and Management of the World’s Savannas. International Savanna Symposium. Australian Academy of Science, Canberra, Australia. Holling (1973). Resilience and stability of ecological systems. Annual Review of Ecological Sys- tems 4: 1-24. Holling, C.S. (ed.) (1978). Adaptive environmental assessment and management. Wiley, Chich- ester, UK. Holling, C.S. (1995). What barriers? What bridges? Pp. 3-34 in Gunderson, L.H., C.S. Holling and S.S. Light (eds.) Barriers and Bridges to the Renewal of Ecosystems and Institutions. Columbia University Press, New York, USA. Holling, C.S. and L.H. Gunderson (2002). Resilience and Adaptive Cycles. Pp. 25-62 in Gun- derson, L.H. and C.S. Holling (eds.) Paranarchy: Understanding transformations in Human and Natural Systems. Island Press, Washington DC, USA. Holling, C.S., L.H. Gunderson and G.D. Peterson (2002). Sustainability and Panarchies. Pp. 63-102 in Gunderson, L.H. and C.S. Holling (eds.) Paranarchy: Understanding transformations in Human and Natural Systems. Island Press, Washington DC, USA. Hollis, G.E., W.M. Adams and M.Aminu-Kano (eds.) (1993). The Hadejia-Nguru Wetlands. En- vironment, Economy and Sustainable development of a Sahelian Floodplain Wetland. IUCN, Gland, Switserland. Hollis, G.E. (1996). Hydrological inputs to management policy for the Senegal River and its floodplain. Pp. 155-184 in Acreman, M.C. and Hollis, G.E. (eds.) Water Management and Wetlands in Sub-saharan Africa. IUCN, Gland, Switserland. Holmes, D.A. (1972). Bird notes from the plains south of Lake Chad, winter 1971-1972. Bull. Niger. Orn. Soc. 9: 47-55, 76-84. Holmes, D.A. (1974). Bird notes from the plains south of Lake Chad, winter 1971-1972. Part 3. Bull. Niger. Orn. Soc. 10: 28-37. Hopkins and N. Levtzion (1981). Corpus of early Arabic sources for West African History 2001. Cambridge University Press, Cambridge, UK. 308 Floodplain rehabilitation and the future of Conservation & Development

Howell, P., M. Lock and S.Cobb (1988). The Jonglei Canal. Impact and Opportunity. Cambridge University Press, Cambridge, UK. Hughes, F.M.R. (1990). The influence of flooding regimes on forest distribution and composition in the Tana river floodplain, Kenya. J. Appl. Ecol. 27: 475-491. Hulme, D. and M. Murphree (1999). Communities, Wildlife and the ‘New Conservation’ in Africa. Journal of International Development 11: 277-285. IIED (1994). Whose Eden? An Overview of Community Approaches to Wildlife Management. IIED, London, UK. Inamdar, A., H. de Jode, K. Lindsay and S. Cob (1999). Capitalizing on Nature: Protected Area Management. Science 283: 1856. INTECOL (2004). Conference Abstracts of the 7th Intecol Wetlands conference, Utrecht, The Netherlands. Inter-American Development Bank (2000). Ecuadorian Plan to protect Galapagos Islands’ ecosystems receives IDB loan for $10.4 million. Press release, November 1. (www.iadb.org/exr/ prensa/2000/cp19300e.htm). Issa, Saibou (2004). L’embuscade sur les routes des abords sud du lac Tchad. Politique Afri- caine 94: 82-104. IUCN (1994). Guidelines for Protected Area Management Categories. Gland, Switserland and Cam- bridge, UK: CNPPA-WCMC. www.unep-wcmc.org/protected_areas/categories/index.html. IUCN (1999). Rehabilitation of the Waza-Logone Floodplain, Republic of Cameroon: Proposals for the Reinundation Programme. IUCN, Gland, Switserland. IUCN (2003). Red List of Threatened Species. www.redlist.org. Jacobson, S.K. and M.D. McDuff (1998). Training idiot savants: the lack of human dimensions in conservation biology. Conservation Biology 12: 263-267. Janssen, M.A. and M. Scheffer (2004). Overexploitation of Renewable resources by ancient societes and the role of sunk-cost effects. Ecology and Society 9 (1): 6. www.ecologyandso- ciety.org/vol9/iss1/art6 Jeanin, A. (1936). Les Mammiferes sauvages du Cameroun. Paul Lechevalier, Paris, France. Jeanrenaud, S. (1999). People-oriented conservation: progress to date. Pp. 126-134 in Stolton, S. and N. Dudley (eds.) Partnerships for Protection. Earthscan, London, UK. John, D.M., C. Leveque and L.E. Newton (1993). Western Africa Pp. 47-78 in Whigham, D.F. D. Dykyjova and S. Hejny (eds.) Wetlands of the World: Inventory, ecology and management I. Kluwer Academic Publishers, The Netherlands. Johnson, M. (1978). By Ship or by Camel ? the Struggle for the Cameroons. Ivory trade in the 19th century. Journal of African history 17: 95-117. Junk, W.J., P.B. Bayley and R.E. Sparks (1989). The Flood Pulse concept in River-Floodplain Systems. Pp. 110-127 in Dodge, D.O. (ed.) Proceedings of the International Large River Symposium (LARS). Honey Harbour, Canadian Special publication of Fisheries and Aqua- tic Sciences 106, Department of Fisheries and Oceans, Ottowa, Ontario, Canada. Kamto, M. (1996). Rapport de Mission. Aspects juridiques du Plan d’Aménagement des Parcs Na- tionaux de Waza et de Kalamaloué. Report to the Waza-Logone Project, Maroua, Cameroon. Kanyamibwa, S. (1999). Project Development Workshop: Integrating Biodiversity Information Management into the Curricula of Wildlife Training Institutions. WCMC, Cambridge, UK. Kari, S. and P. Scholte (2001). La réhabilitation pastorale de la plaine d’inondation Waza- Logone (Cameroun): comment consolider sa réussite écologique? Pp. 315-316 in Tielkes, E. Schlecht and P. Hiernaux (eds.) Atelier regional: les approches de la gestion des pâturages et les projets de développement: quelles perspectives? Verlag Ulrich E. Grauer, Stuttgart, Ger- many. Kavanagh, M. (1977). Some birds in Cameroon. Bull. Niger. Orn. Soc. 13: 133-138. References 309

Kavanagh, M. (1978). National Park in the Sahel. Oryx 14: 241-245. Kessler, J.J. (2003). Working towards SEAN-ERA. A framework and principles for integrating environmental sustainability into planning. Tropical Resource Management Papers 43. Wage- ningen University, Wageningen, The Netherlands. Kiker, C.F., J.W. Miloton and A.W. Hodges (2001). Adaptive Learning for science-based policy: the Everglades restoration. Ecological Economics 37: 403-416. Kingdon, J. (1997). The Kingdon Field Guide to African Mammals. Academic Press. San Diego, USA. Kirkpatrick, D. (1994). Evaluating Training Programs, the Four Levels. Barrett-Koehler Publish- ers San Francisco, USA. Kirkpatrick, J.B. (2001). Ecotourism, Local and Indigenous People and the Conservation of the Tasmanian Wilderness World Heritage Area. Journal of the Royal Society of New Zealand 31: 819-828. Korn, H., R. Schliep and J. Stadler (eds.) (2003). Report of the International Workshop on the ‘Further Development of the Ecosystem Approach’. Bundesambt fur Naturschutz Skripten 78, available on www.bfn.de. Kouokam, R., M. Moritz, D.Ngantou and P. Loth (2004). Conflicts and conflict management in the Waza Logone Floodplain. Pp. 69-76 in Loth, P. (ed.) The return of the water. IUCN, Gland, Switzerland. Kramer, R., C. van Schaik and J. Johnson (eds.) (1997). Last Stand. Protected Areas & the defense of tropical biodiversity. Oxford University Press, New York, USA. Kremen, C., A.M. Merenlender and D.D. Murphy (1994). Ecological monitoring: a vital need for Integrated Conservation and Development Programs in the Tropics. Conservation Bio- logy 8: 388-397. Kushlan, J.A (2000). Heron Feeding Habitat Conservation. Pp. 219-236 in Kushlan, J.A. and H. Hafner (eds.) Heron Conservation. Academic Press, London, UK. Laplante, K. (2004). Environmental Alchemy: How to turn Ecological Science into Ecological Philosophy. Environmental Ethics 26: 361-380. Lavieren, L.P. van (1976). Methodes d’inventaire des populations des grands mammiferes d’Afrique. FAO document de travail RAP/74/056, Rome, Italy. Lebeuf, A.M.D. (1969). Les Principautés Kotokos: essai sur le caractère sacré de l’autorité. Paris. CNRS, Paris, France. Leakey, R. and V. Morell (2001). Wildlife Wars. My Fight to save Africa’s Natural Treasures. MacMillan, London, UK. Lee, K.N. (1993). Compass and Gyroscope: Integrating Science and Policy for the Environment. Island Press, Washington DC, USA. Leendertse, P.C., A.J.M. Roozen, and J. Rozema (1997). Long-term changes (1953-1990) in the salt marsh vegetation at the Boschplaat on Terschelling in relation to sedimentation and flooding. Plant Ecology 132: 49-58. Legrosse, P. (1999). Les règles d’accès des troupeaux peuls aux pâturages du Delta central du Niger (Mali). Pp. 76-96 in M. Niamir-Fuller (ed.) Managing mobility in African rangelands. Intermediate Technology Publications, London, UK. Le Houerou, H.N. (1989). The Grazing Land Ecosystems of the African Sahel. Ecological Stu- dies 75. Springer-Verlag, Berlin, Germany. Lenders, H.J.R., B.G.W. Arts, H. Strijbosch and G. Van der Velde (1998). The role of reference and target images in ecological recovery of river systems: lines of thought in the Netherlands. Pp. 35-52 in Nienhuis, P.H., R.S.E.W. Leeuwen and A.M.J. Ragas (eds.) New Concepts for sustainable Management of River Basins. Backhuys Publishers, Leiden, The Netherlands. 310 Floodplain rehabilitation and the future of Conservation & Development

Letouzey, R. (1968). Etude phytogeographique du Cameroun. Lechevalier, Paris, France. Levin, S.A., S. Barrett, S. Aniyat et al. (1998). Resilience in natural and socio-economic systems. Environment and Development Economics 3: 222-235. Lindley, W.I. (1999). Constraints and Potentials of Training Mid-Career Extension Professionals in Africa. International Workshop on Innovative Training Programmes for Mid-Career Agricultural Extension Field Staff in Sub-Saharan Africa, 6-8 July 1999. Addis Ababa, Ethiopia. Lipya, A.K. (2000). Evaluation of the progress of students trained at CAWM-Mweka. Unpublished report, SADC Wildlife Management Training Programme, Harare, Zimbabwe. Loth, P., and H. de Iongh (2004). The Waza-Logone Project: History, Objectives and Main Re- sults. Pp. 77-87 in Loth, P. (ed.) The return of the water. Restoring the Waza Logone floodplain in Cameroon. IUCN, Gland, Switserland. Loth, P. (ed.) (2004). The return of the water. Restoring the Waza Logone floodplain in Cameroon. IUCN, Gland, Switserland. Louette, M. (1981). The birds of Cameroon, an annotated check-list. Verhandel. Kon. Acad. Wetensch. Lett. Schone Kunst. Belg., Kl. Wetensch. 43 (163): 1-295. MacKinnon, J., K. MacKinnon, G. Child and J. Thorsell (1986). Managing Protected Areas in the Tropics. IUCN, Gland, Switzerland. Maley, J. (1981). Etudes palynologiques dans le bassin du Tchad et Paleoclimatologie de l’Afri- que nord tropicale de 30 000 ans à l’epoque actuelle. These Sc, Travaux et Documents de l’ORSTOM, Paris 129: 586. Consulted from Pp. 45-52 in Coetzee, J.A. and E.M. van Zin- deren Bakker (eds.) Palaeoecology of Africa and the surrounding Islands. Vol. 13. Balkema, Rotterdam, The Netherlands. Also consulted through Maley, J. 1993. Chronologie cal- endaire des principales fluctuations du lac Tchad au cours du dernier millenaire: le role des données historiques et de la tradition orale. Pp. 161-163 in Barretau, D. and C. von Graffenrie (eds.) Seminaire MegaTchad, ORSTOM, Paris, France. ’t Mannetje, L. (1987). Measuring quantity of grassland vegetation. Pp. 63-95 in ’t Mannetje, L. (ed.). Measurement of Grassland Vegetation and Animal Production. CAB International, Wallingford, UK. Marchand, M. (1987). The productivity of African floodplains. Int. J. of Env. Studies 29: 201-211. Margolius, R. and N. Salafsky (1998). Measures of Success: Designing, Managing and Monitoring Conservation and Development Projects. Island Press, Washington DC, USA. Marion, L., P. Ulenaers, P. and J. van Vessem (2000). Herons in Europe. Pp. 1-32 in Kushlan, J.A. and H. Hafner (eds.) Heron Conservation. Academic Press, London, UK. Martin, A., P. Bonnet, D. Bourzat and R. Lancelot (1996). Importance of livestock production and its economic contribution to the countries of the Lake Chad Basin Commission. Pp. 79- 96 in CIRAD/CTA Livestock Atlas of the Lake Chad Basin. CTA, Ede, The Netherlands. Mayaka, T.B. (2002). Wildlife Co-Management in the Bénoué National Park Complex, Came- roon: A Bumpy Road to Institutional Development. World Development 30: 2001-2016. McCarthy, T.S., W.N. Ellery and A. Bloem (1998). Some observations on the geomorphologi- cal impact of hippopotamus (Hippotamus amphibius L.) in the Okavango Delta, Botswana. Afr. J. Ecol. 36: 44-56. McCully (2001). Silenced rivers. The Ecology and Politics of Large dams. Zed Books, London, UK and New York, USA. McShane, T.O. and S.A. Newby. (2004). Expecting the Unattainable: The assumptions behind ICDPs. Pp. 49-74 in McShane, T.O. and M.P. Wells, Getting Biodiversity Project to Work. Columbia University Press, New York, USA. References 311

Mefit Babtie (1983). Development Studies in the Jonglei Canal Area. Technical Assistance Contract for Range Ecology Survey, Livestock Investigations and Water Supply. Final Report. Vol. 2, Background and Vol. 3. Vegetation Studies. Glasgow, Rome and Khartoum, Sudan. Metcalfe, S. (1994). The Zimbabwe Communal Area Management Programme for Indige- nous Resources (Campfire). Pp. 161-192 in Western, D., R.M. Wright and S. Strum (eds.) Natural Connections. Island Press, Washington D.C., USA. Middleton, B. (1999). Wetland Restoration. Flood pulsing and Disturbance Dynamics. Wiley, New York, USA. MINEF (1994). Cameroon Ministry of the Environment and Forestry. Law No. 94-01 of 20 January 1994 to lay down forestry, wildlife and fisheries regulations. Pp. 5-49 in A compendium of official instruments on forest and wildlife management in Cameroon. Ministry of Environment and Forestry, Yaoundé, Cameroon. MINEF (1995). Cameroon Ministry of Environment and Forestry, Decree No. 95-466-PM of 18 December 1995 to lay down the conditions for the implementation of wildlife regulations. Pp. 119-151 in A Compendium of Official Instruments on Forest and Wildlife Manage- ment in Cameroon. Ministry of Environment and Forestry, Yaoundé, Cameroon. MINEF-IUCN (1997). Plan Directeur d’Aménagement du Parc National de Waza et ses alentours. Unpublished report, Yaoundé, Maroua, Cameroon. Mitchell, B. (2002). Resource and Environmental Management. Prentice Hall, Harlow. Mohammadou, E. (1988). Les lamidats du Diamare et du Mayo-Louti au XIX siecle (Nord Came- roun) (Vol. XXII). Institute for the study of Languages and Cultures of Asia and Africa (ILCAA). Tokyo, Japan. Moritz, M. (2003). Commoditization and the pursuit of piety: the transformation of an African pastoral system. Dissertation, University of California at Los Angeles, Los Angeles, USA. Moritz, M., P. Scholte and S. Kari (2002). The demise of the nomadic contract: arrangements and rangelands under pressure in the Far North of Cameroon. Nomadic Peoples 6: 127-146. Morton, A.J. and E.A. Obot (1984). The Control of Echinochloa stagnina (Retz.) P. Beauv. by harvesting for dry season livestock fodder in Lake Kainji, Nigeria-A modelling approach. J. Appl. Ecol. 21: 687-694. Moss, G. (1991). The Trainers Handbook. Moss Associates, New Zealand. Mullié, W.C., J. Brouwer, S.F. Codjo and R. Decae (1999). Small isolated wetlands in the Central Sahel: a resource shared between people and waterbirds. Pp. 30-38 in Beintema, A. and van J. Vessem (eds.) Strategies for Conserving Migratory Waterbirds. Wetlands. Int. Publ. 55, Wageningen, The Netherlands. www.wetlands.org/pubs&/pub_online/StratMi- gratWBs.pdf. Mundy, P.J. (1972). Waza, Cameroun-access and avifauna. Bull. Niger. Orn. Soc. 9: 16-18. Myers, N. (1997). The rich diversity of biodiversity Issues. Pp. 125-138 in Reaka-Kudla, M.L., D.E. Wilson and E.O. Wilson (eds.) Biodiversity II. Understanding and Protecting Our Bio- logical Resources. Joseph Henri Press, Washington DC, USA. Nachtigal, G. (1987). Sahara and Sudan. Volume III. The Chad Basin and Baguirmi (translated by Fischer & Fischer). Hurst & Company, London, UK. Nepal, S.K. and K.E. Weber (1994). A Buffer Zone Concept for Biodiversity Conservation: Via- bility of the Concept in Nepal’s Royal Chitwan National Park. Environmental Conservation 21: 333-341. Neumann, R. (1997). Primitive Ideas: Protected Area Buffer Zones and the Politics of Land in Africa. Development and Change 28: 559-582. Nguinguiri, J.C. (2001). Guide pour la Formation en Gestion Participative des Ressources Naturelles. UICN-BRAC, Yaoundé, Cameroun. 312 Floodplain rehabilitation and the future of Conservation & Development

Niamir, M. (1990). Community forestry: herders’ decision-making in natural resource management in arid and semi-arid Africa. FAO, Rome, Italy. Niamir-Fuller, M. (1999a). Towards a synthesis of guidelines for legitimizing transhumance. Pp. 266-290 in Niamir-Fuller, M. (ed.) Managing mobility in African rangelands: the legitimization of transhumance. IT Publications, London, UK. Niamir-Fuller, M. (1999b). Managing mobility in African rangelands: the legitimization of transhumance. London. IT Publications, London, UK. Niamir-Fuller, M., and M. Turner (1999). A review of recent literature on pastoralism and transhumance in Africa. Pp. 18-46 in Niamir-Fuller, M. (ed.) Managing mobility in African rangelands: the legitimization of transhumance. IT Publications, London, UK. Nicolson, S.E. (1986). Climate, drought and famine in Africa. Pp. 107-129 in Hansen, A. and D.E. McMillan (eds.) 1986. Food in sub-Saharan Africa. Lynne Reiner Bouolder, Colo, USA. Nienhuis, P.H., R.S.E.W. Leeuwen and A.M.J. Ragas (1998). Discussion and State of the art: new concepts emerging. Pp. 353-367 in Nienhuis, P.H., R.S.E.W. Leeuwen and A.M.J. Ragas (eds.) New Concepts for sustainable Management of River Basins. Backhuys Publishers, Leiden, The Netherlands. Njifortii, H.L., T. Schrader and F. Toornstra (1991). LEARN: A Methodological Challenge for Rapid Environmental Assessment. Landscape & Urban Planning 20: 173-181. Ngog Nje, J. (1983). Structure et dynamique de la population des girafes au Parc National de Waza, Cameroun. Rev. Ecol (Terre Vie) 37: 3-20. Ngog Nje, J. (1984). Régime alimentaire de la girafe au Parc National de Waza, Cameroun. Mammalia 48: 173-183. Ngog Nje, J. (1995a). Central and West Africa. education and training requirements. Pp. 124- 126 in Robinson, R. (ed.) Africa Heritage 2000: The future of Protected Areas in Africa. Proceedings of the IUCN Commission on National Parks and Protected Areas African Re- gional Working Session. National Parks Board, Pretoria, South Africa. Ngog Nje, J. (1995b). La formation des spécialistes de la faune sauvage et de gestion des aires pro- tégées en Afrique. Unpublished report, Garoua Wildlife College, Garoua, Cameroon. Njiforti, H.L. (1997). The Biology and Management of Wild Helmeted Guineafowl (Numida meleagris galeata Pallas) in the Waza Region of North Cameroon. PhD thesis Wageningen University, Wageningen, The Netherlands. Njoya, I.S. (2001). L’Ecole de Faune de Garoua (hier-aujourd’hui-demain). Garoua Wildlife Col- lege, Garoua, Cameroon. Noble, I.R. and R.O. Slatyer (1980). The use of vital attributes to predict successional changes in plant communities subject to recurrent disturbance. Vegetatio 43: 5-21. North, M.E.W. (1963). Breeding of the Black-headed Heron at Nairobi, Kenya, 1958-62. J. E. Afr. Nat. Hist. Soc. 24: 33-63. Noss, A.J. (1997). Challenges to nature conservation with community development in central African forests. Oryx 31: 180-188. Noupa, P. (1995). Etude des interactions écologiques entre herbivores sauvages et domestiques dans l’écosystème de la plaine de Waza (Cameroun). Mémoire maitrise en sciences naturelles appliquées. Université de Liege, Liege, Belgium. OAG Münster (1991). Report of the Ornithological Expedition to Northern Cameroon. Biologische Station Münster, Germany. Oates, J.F. (1999). Myth and reality in the rainforest: how conservation strategies are failing in West Africa. University of California Press, Berkeley, USA. Ortega-Rubio, A. and C. Arguelles-Méndez (1999). Management Plans for Natural Protected Areas in Mexico: La Sierra de la Laguna Case Study. Int. J. Sustain. Dev. World Ecol. 6: 68- 75. References 313

Ostrom, E. (1990). Governing the Commons. The Evolution of Institutions for Collective Action. Cambridge University Press, Cambridge, UK. Paillisson, J.-M., S. Reeber and L. Marion (2002). Bird assemblages as bio-indicators of water regime management and hunting disturbance in natural wet grassland. Biological Conser- vation 106: 115-127. Parker, I. and S. Bleazard (2001). An impossible dream. Some of Kenya’s last colonial wardens recall the game department in the closing years of the British empire. Librario, Kinloss, Moray, UK. Parr, J. (1998). Discussion Paper for Protected Area Managers Writing a Management Plan. Tigerpaper 25: 9-13. Parsons, J. (1977). The effect of predation by Fish Eagles on the breeding success of various Ciconiiformes nesting near Kisumu, Kenya. J. Nat. Hist. 11: 337-353. Perrings, C. (1998). Resilience in the dynamics of economic-environment systems. Environ- mental and Resource Economics 11: 503-520 Pettet, A. (1976). The avifauna of Waza National Park, Cameroon, in December. Bull. Niger. Orn. Soc. 12: 18-24. Pfadenhauer, J. and F. Klotzli (1996). Restoration experiments in middle European wet terrestrial ecosystems: an overview. Vegetatio 126: 101-115. Pimm, S.L. (1984). The complexity and stability of ecosystems. Nature 307: 321-326. Pirot, J.Y., P.J. Meynell and D. Elder (2000). Ecosystem Management: Lessons from around the world. A guide for Development and Conservation Practitioners. IUCN, Gland, Switserland. Pitkin, B. (1995). Protected Area Conservation Strategy (PARCS): Training Needs and Oppor- tunities Among Protected Area Managers in Eastern, Central and Southern Africa. Biodiversity Support Program, Washington, USA. Poché, R.M. (1974). Notes on the roan antelope (Hippotragus equinus (Desmarest)) in West Africa. J. Appl. Ecol. 11: 963-968. Polet, G and J.R. Thompson (1996). Maintaining the floods – hydrological and institutional aspects of managing the Komadugu-Yobe River basin and its floodplain wetlands. Pp. 73- 100 in Acreman, M.C. and G.E. Hollis (eds.) Water Management and Wetlands in Sub-saharan Africa. IUCN, Gland, Switserland. Polet, G. (2000). Waterfowl and flood extent in the Hadejia-Nguru wetlands of north-east Ni- geria. Bird Cons. Int. 10: 203-209. Pollack, M.E. and J.F. Horty (1999). There is More to Life Than Making Plans: Plan Manage- ment In Dynamic, Multiagent Environments. Artificial Intelligence magazine 71-83. Pretty, J.N., I. Guijt, J. Thompson and I. Scoones (1995). Participatory Learning and Action, A Trainer’s Guide. IIED Participatory Methodology Series, IIED London, UK. Prins, H.H.T. and H.P.V. de Jeugd. (1993). Herbivore population crashes and woodland structure in East Africa. J. of Ecol. 81: 305-314. Pulliam, H. R. and B.J. Danielson (1991). Sources, Sinks, and Habitat Selection: A Landscape Perspective on Population Dynamics. American Naturalist 137 (suppl.): S50-S66. Quammen, D. (2003). Saving Africa’s Eden, 204 National Geographic, 50-72 (September 2003). Rauchenberger, D. (1999). Johannes Leo der Afrikaner. Seiner Beschreibung des Raumes zwischen Nil und Niger nach dem Urtext. Orientalia Biblica et Christiana 13. Harrasswitz Verlag. Wiesbaden, Germany. Rees, W.A. (1978a). The ecology of the Kafue lechwe: the food supply. J. Appl. Ecol. 15: 177-191. Rees, W.A. (1978b). The ecology of the Kafue lechwe: soils, water levels and vegetation. J. Appl. Ecol. 15: 163-176. Rettie, D.F. (1995). Our National Park System. University of Illinois, Urbana, USA. Rice, W. R. (1989). Analyzing tables of statistical tests. Evolution 43: 223-225. 314 Floodplain rehabilitation and the future of Conservation & Development

Rietkerk, M, P. Ketner, L. Stroosnijder and H.H.T. Prins (1996). Sahelian rangeland development: a catastrophe? Journal of Range Management 49: 512 - 519. Rietkerk, M. (1998). Catastrophic vegetation dynamics and soil degradation in semi-arid grazing systems. Ph.D. thesis, Wageningen Agricultural University, Wageningen, The Netherlands. Rietkerk, M., M. Boerlijst, F. van Langevelde, R. HilleRisLambers, J. van de Koppel, L. Kumar, H.H.T. Prins and A.M. de Roos (2002). Self-organization of vegetation in arid ecosystems. The American Naturalist 160: 524-530. Riney, T. (1963). A rapid field technique and its application in describing conservation status and trends in semi-pastoral areas. African Soils 8: 159-257. Robin (1954). Le livre des sanctuaires de la nature. Bibliothèque Géographique. Payot, Paris. Robins, E. (1970). The Ebony Ark. Black Africa’s Battle to save its Wildlife. Taplinger, New York, USA. Rogers, A. and P. Taylor (1998). Participatory Curriculum Development in Agricultural Edu- cation. A training guide. FAO, Rome, Italy. Roshier, D.A., A.I. Robertson, R.T. Kingsford and D.G. Green (2001). Continental-scale interactions with temporary resources may explain the paradox of large populations of desert waterbirds in Australia. Landscape Ecology 16: 547-556. Rothwell, W.J. and H.J. Sredl (1992). Professional Human Resources Development: roles and com- petencies. 2nd Edition. Vol II. HRD Press, Amherst, USA. Roux, F. and G. Jarry (1984). Importance, Composition et Distribution des Populations d’Anatidés Hivernant en Afrique de l’Ouest. Muséum National d’Histoire Naturelle, Paris, France. Roux, F. and G. Jarry (1986). Denombrements d’Anatidés Hivernant dans le Bassin Tchadien, Janvier 1986. Muséum National d’Histoire Naturelle, Paris, France. Roux, F. and G. Jarry (1987). Denombrements d’Anatidés Hivernant dans le Bassin Tchadien, Janvier 1987. Muséum National d’Histoire Naturelle, Paris, France. Saberwal, V.K. and A. Kothari (1996). The Human Dimension in Conservation Biology in Developing Countries. Conservation Biology 10: 1328-1331. Salafsky, N. (1994). Ecological Limits and Opportunities for Community-based Conservation. Pp. 448-471 in Western, D., R.M. Wright and S. Strum (eds.) Natural Connections. Island Press, Washington DC, USA. Salafsky, N., R. Margoluis and K. Redford (2001). Adaptive Management: A Tool for Conser- vation Practitioners. Biodiversity Support Program, Washington DC, USA. (www.foson- line.org/Resources.cfm). Saleh, A. (1994). La pêche dans le parc national de Kalamaloué: une activité à stimuler, à tolerer, ou à interdire? Rapport de stage optionnel. Université de Dschang, Dschang, Cameroon. Saleh, A. (1995). Impact de la reinondation de la plaine du Logone sur la peche: Cas particulier des villages riverains du Parc National de Waza. Rapport de stage d’insertion professionelle. Université de Dschang, Dschang, Cameroon. Sayer, J.A. (1982). The pattern of decline of the Korrigum Damaliscus lunatus in West Africa. Biol. Conserv. 23: 95-110. Sayer, J. (1991). Rainforest Buffer Zones. Guidelines for Protected Area Managers. IUCN, Gland, Switzerland. Scheffer, M. , S.H. Hosper, M.L. Meijer and B. Moss (1993). Alternative equilibria in shallow lakes. Trends Ecol. Evol. 8: 275-279. Scheffer, M., S. Carpenter, J.A. Foley, C.Folke and B. Walker (2001). Catastrophic shifts in ecosystems. Nature 413: 591-596. Schneider, J.L. (1994). Le Tchad depuis 25000 ans. Geologie, Archeologie Hydrogeologie. Mas- son, Paris. Scholes, R.J. and R. Biggs (2005). A biodiversity intactness index. Nature 434: 45-49. References 315

Scholte, P. (1996). Conservation status of cranes in north Cameroon and western Chad. Pp. 153-156 in Beilfuss, R.D., W.R. Tarboton, N.N. Gichuki (eds.) Proceedings of the 1993 Afri- can Crane and Wetland Training Workshop. www.savingcranes.org/data/program/pp153- 156.pdf Scholte, P. (1998). Waza National Park in Cameroon. Pp. 46 in Sherbenin, A. and M. Freu- denberger: Migration to Protected Areas and Buffer zones and Protected Areas: Can we stem the tide? Parks 8: 32-53. Scholte, P. (2000a). Inventaire de la biodiversité. Ressources en espèces sauvages. Pp. 50-53 in GEPIS. Vers une gestion durable des plaines d’inondation sahéliennes. IUCN, Gland, Swit- serland and Cambridge, UK. Scholte, P. (2000b). Towards consensual park management planning in Africa. Oryx 34: 87- 89. Scholte, P. (2003a). Immigration: a potential time bomb under the integration of Conserva- tion and Development. Ambio 32: 58-64. (www.iucn.org/themes/cem/cem/region/cameroon.htm) Scholte, P. (2003b). Curriculum development at the African Regional Wildlife colleges, with special reference to the Ecole de Faune (Cameroon). Environmental Conservation 30: 249- 258. (www.eldis.org/fulltext/CurriculumDev.EnvCons-Scholte.pdf). Scholte, P. (2005). At the interface of legislation and wildlife management: a decade of experiences in consensual protected area management planning in Cameroon. Pp. (in press) in Wandesforde-Smith (ed.) The Future of Conservation in Africa: Law, Biodiversity, Live- lihoods and Development. Cambridge University Press, Cambridge UK. Scholte, P. E. Pamo, S. Kari, S. Kersten and P. Kirda (1996a). Floodplain rehabilitation in N. Cameroon: expected impact on vegetation, pastoralists and wildlife. Pp. 492-493 in West, N. (ed.) Proceedings Fifth International Rangeland Congress. Society of Range Management, Colorado, USA Scholte, P., S. Kari and M. Moritz (1996b). The involvement of nomadic and transhumant pastoralists in the rehabilitation and management of the Logone floodplain, North Cameroon. Issues paper 66, Drylands programme. IIED, London, UK. Scholte, P., A. Saleh, K. Bobo and B. Boukar (1998). Who gets the fish? Assessing competition between piscivorous Birds and the fishery on the Logone floodplain of Waza National Park, North Cameroon. Pp. 378 in Farina, A., J. Kennedy and V. Bossu (eds.) Proceedings International Ecology Congress. INTECOL, Florence, Italy. Scholte, P., S. de Kort and M. van Weerd (1999a). The Birds of the Waza-Logone Area, Far North Province, Cameroon. Malimbus 21: 16-50. Scholte, S. Adam, S. Kari and J.H. Mbouche (1999b). Walking a tightrope: Using PRA in a conflict situation around Waza National Park, Cameroon. PLA Notes 35:7-12. (www.iied.org/ sarl/pla_notes/pla_backissues/documents/plan_03502.PDF) Scholte, P., S. de Kort and M. van Weerd (2000a). Floodplain rehabilitation in Far North Cameroon: expected impact on bird life. Ostrich 71: 112-117. Scholte, P., P. Kirda, S. Adam and B. Kadiri (2000b). Floodplain rehabilitation in North Came- roon: impact on vegetation dynamics. Applied Vegetation Science 3: 33-42. Scholte, P., S. Kari, M. Moritz and H.H.T. Prins (2005). Pastoralist Responses to Floodplain Rehabilitation in North Cameroon. Human Ecology (in press). Scholte, P. and J. Brouwer 2005. The Relevance of Key Resource Areas for Large-Scale Move- ments of Livestock: are Sahelian floodplains an example? Pp. (in press) in Prins, H.H.T. and F. van Langevelde, Large Herbivore Dynamics in Fragmented Landscapes. Scholte, P. and I. Hashim (2005). Red-fronted Gazelle. Pp. (in press) in Kingdon (ed.) Mam- mals of Africa. Academic Press, London, UK. 316 Floodplain rehabilitation and the future of Conservation & Development

Schouten, M.G.C. (2001). Nieuwe Heelmeesters: Wetenschap en praktijk van natuurherstel. In- augural address Wageningen University, Wageningen, The Netherlands. Schrader, T. (1986). Les yaérés au Nord du Cameroun: pâturages de saison sèche (?) Aspects socio- écologiques du développement pastoral dans la plaine inondable du Logone. Leiden University, Leiden, The Netherlands. Schricke, V., P. Triplet and P. Yésou (2001). Contributions Françaises à la connaissance des oiseaux d’eau Palearctiques hivernant dans le Delta du Senegal. Alauda 69: 135-148. Sethna (2004). What’s Hysteresis. www.lassp.cornell.edu/sethna/hysteresis/WhatIsHystere- sis.html. Accessed 27 March 2005. Shambaught, J., J. Oglethorpe and R. Ham (2001). The Trampled Grass. Mitigating the impacts of armed conflicts on the environment. Biodiversity Support Program, Washington DC, USA. Shepherd, J. (2004). The Ecosystem Approach. Five Steps to Implementation. IUCN, Gland, Swit- serland and Cambridge UK. Sherbenin, A. de and M. Freudenberger (1998). Migration to protected areas and buffer zones: can we stem the tide? Parks 8: 38-53. Sighomnou, D., (2000). La plaine du Yaere dans le Nord Cameroun: une experience de restaura- tion des inundations. Seminaire GIRN-ZIT, 20-23 Juin 2000, Bamako, Mali. Sinclair, A.R.E. and P. Arcese (1995). Serengeti in the Context of Worldwide Conservation Efforts. Pp. 31-46 in Sinclair, A.R.E. and P. Arcese (eds.) Serengeti II. Dynamics, Manage- ment and Conservation of an Ecosystem. University of Chicago Press, Chicago, USA. Sinibaldi,I., K. Schmidt, P. Scholte, I. van Duren, F. Corsi, J. Brouwer and H. Prins (2004). Dependence of large mammals and other flagship species on water and water management in Africa. A literature review. Scientific report commissioned by WNF, Zeist, The Netherlands. Sinsin, B., A.C. Tehou, I. Daouda, and A. Saidou (2002). Abundance and species richness of larger mammals in Pendjari National Park in Benin. Mammalia 66: 369-380. Skead, C.J. (1964). Black-headed Herons, Ardea melanocephala (Vigors and Children), nesting on a krans. Ostrich 35: 236. Sklar, F.H., H.C. Fritz, Y. Wu, R. van Zee and C. McVoy (2001). The design of ecological landscape models for Everglades restoration. Ecological economics 37: 379-401. Snelson, D. (1993). CEC Protected Areas and Biodiversity Strategy: Case study of Training and the College of African Wildlife Management. Unpublished report, IUCN/WWF Regional Office, Eastern Africa, Nairobi, Kenya. Snelson, D. and A. Lanjouw (1997). Institutionalizing in-service training in protected area authorities in Africa. Protected Areas Conservation Strategy (PARCS) Project. African Wildlife Foundation & Wildlife Conservation Society, Washington, USA. Spinage, C. (1998). Social change and conservation misrepresentation in Africa. Oryx 32: 265- 276. SPSS (1993). SPSS for Windows, release 6.01. SPSS Inc.. SPSS (1999). SPSS for Windows, release 9.0. SPSS Inc.. Steiner, A. (2004). News ReleaseWorld Congress Bangkok www.iucn.org/congress/documents/ press/2004-11-25-cem.pdf (accessed 22 March 205). Stenning, D.J. (1957). Transhumance, migratory drift, migration; patterns of pastoral Fulani nomadism. Journal of the Royal Anthropological Institute 87: 57-73. Stone, R. (1997). What’s Your Role? Training for Organizational Impact. A guide for training officers in protected area management. African Biodiversity Series, no. 5. Biodiversity Support Program, Washington DC, USA. Szép, T. (1995). Relationship between west African rainfall and the survival of central Euro- pean Sand Martins Riparia riparia. Ibis 137: 162-168. References 317

Tambekou, A., F. Djaoro and B. Boukar (1997). Rapport de synthèse des mouvements migratoires dans les villages périphériques du parc national de Waza. Internal report Waza-Logone project, Maroua, Cameroon. Tapsell, S.M. (1995). River restoration: what are we restoring to? A case study of the Raven- bourne River, Kondon. Landscape Research 20: 98-111. Tarboton (1977). The status of communal herons, ibis and cormorants on the Witwatersrand. S. Afr. J. Wildl. Res 7: 19-25. Taylor, J. (1972). The Black-headed Heron. Animals 13: 688-689. Taylor, P. (1997). Participatory curriculum development for agricultural education and training: experiences from Viet Nam and South Africa. Training for Agricultural and Rural De- velopment 1997-98: 4-15. Tchamba, M. (1996). Elephants and their Interactions with People and Vegetation in the Waza- Logone Region, Cameroon. PhD thesis, Utrecht University, The Netherlands. Tchamba, M.N. and P. Elkan (1995). Status and Trends of some large mammals and ostrich- es in Waza National Park, Cameroon. Afr. J. Ecol. 33: 366-376. Terborgh, J., C. van Schaik, L. Davenport and M. Rao (eds.) (2002). Making Parks work. Stra- tegies for Preserving Tropical Nature. Island Press, Washington DC, USA. Thibault, M. and S. Blaney (2001). Sustainable Human resources in a Protected Area in South- western Gabon. Conservation Biology 15: 591-595. Thiollay, J.M. (1978). Les plaines du Nord Cameroun. Centre d’hivernage de rapaces paléarc- tiques. Alauda 46: 314-326. Thomas, L. and J. Middleton (2003). Guidelines for Management Planning of Protected Areas. IUCN, Gland, Switserland. Thompson, K. (1985). Emergent plants of permanent and seasonally-flooded wetlands. Pp. 43- 107 in Denny, P. (ed.) The ecology and management of African wetland vegetation. Junk Publ- ishers, Dordrecht, The Netherlands. Thorhaug. A. (1980). Recovery patterns of restored major plant communities in the United States: High to low altitude, desert to marine. Pp. 113-124 in Cairns, J. (ed.) The recovery process in damaged ecosystems. Ann Arbor Science, Ann Arbor, USA. Thouless, C. R., M. Allen, C. Coetzee, H. Dublin, H. Mahamat, Mohamadou, A.D. Njoh, H. Peters, P. Scholte and M.N. Tchamba (1995). Management of conflict between humans and the migratory Waza elephants. Consultants’ report. Waza-Logone Project, Maroua, Cameroon. Tiffen, M., M. Mortimore and F. Gichuki. (1994). More people, Less erosion. Environmental Recovery in Kenya. John Wiley & Sons, Chichester, UK. Tobias, C., and C. Vanpraet (1981). Notes d’écologie Soudano-Sahelienne. Quelques relations sols-végétation dans le parc national de Waza (Nord Cameroun). Revue Science et Tech- nologie 1: 51-81. Triplet, P. and P. Yésou (2000). Controlling the flood in the Senegal Delta: do waterfowl populations adapt to their new environment. Ostrich 71: 106-111. Trolliet, B. and O. Girard (2001). Numbers of Ruff Philomachus pugnax wintering in West Africa. Wader Study Group Bull 96: 74-78. Tuntivanich, P. (1989). Curriculum development in veterinary science: a case-study from Thailand. Pp. 238-251 in van den Bor, W., J.C.M. Shute and G.A.B. Moore (eds.) South- North Partnership in Strengthening Higher Education in Agriculture. Pudoc, Wageningen, The Netherlands. Turner, M.D. (1999). The role of social networks, indefinite boundaries and politicalbargain- ing in maintaining the ecological and economic resiliency of the transhumance systems of Sudano-Sahelian West Africa. Pp. 97-123 in Niamir-Fuller, M. (ed.) Managing mobility in African rangelands. IT Publications, London, UK. 318 Floodplain rehabilitation and the future of Conservation & Development

ULG Consultants limited (1998). Training Needs Assessment. Unpublished report, SADC Wildlife Management Training Programme, Harare, Zimbabwe. UNDP (1998). Hwange National Park Management Plan 1999-2002, UNDP, Harare, Zim- babwe. Unesco-MAB (2002). Periodic Review for Biosphere Reserves. www.Unesco.org/mab. Vabie, M.B. (2000). Problèmes liés à l’utilisation des Méthodes Participatives: Enseignements tirés de l’application sur le terrain des PRA/RRA dans certains pays de la Sous-Région de l’Afrique Centrale. Arbres, Forets et Communautés rurales 15 & 16: 49-55. Vaillant, A. (1956). Les pâturages naturels du Nord-Cameroun. Unpublished report. Vanclay, J.K. (2001). The Effectiveness of Parks. Science 293: 1007. Van de Ban, A.W. and H.S. Hawkins (1988). Agricultural Extension. Longman, Essex, UK. Van den Bergh, J.C.J.M., A. Barendregt and A.J. Gilbert (2004). Spatial Ecology-Economic Ana- lysis for Wetland Management. Modelling and Scenario Evaluation of Land-Use. Cambridge University Press, Cambridge, UK. Van den Bor, W. (1989). Higher Agricultural education in development. Pp. 3-20 in van den Bor, W., J.C.M. Shute and G.A.B Moore (eds.) South-North Partnership in Strengthening Higher Education in Agriculture. Pudoc Wageningen, The Netherlands. Van den Bor, W., J.M. Bryden and C.T. Jongmans (1995). Rethinking higher agricultural education in a time of globalization and rural restructuring. European Journal of Agricultural Education and Extension 2: 29-40. Van der Kamp, J., M. Diallo and B. Fofana (2002a). Dynamique des populations d’oiseaux d’eau. Pp. 87-141 in Wymenga, E., B. Kone, J. van der Kamp, L. Zwarts (eds.) Delta Inté- rieur du Niger. Ecologie et gestion durable des resources naturelles. Wetlands International/ RIZA/Alterra/Altenburg & Wymenga, Veenwouden, The Netherlands. Van der Kamp, J., M. Diallo, B. Fofona, E. Wymenga, E. (2002b). Colonies nicheuses d’oiseaux d’eau. Pp. 163-186 in Wymenga, E., B. Kone, J. van der Kamp, L. Zwarts (eds.) Delta Inté- rieur du Niger. Ecologie et gestion durable des resources naturelles. Wetlands International/ RIZA/Alterra/Altenburg & Wymenga, Veenwouden, The Netherlands. Van der Kamp, J., B. Fofana and E. Wymenga (2005). Ecological values of the Inner Niger Delta. Pp. 155-177 in Zwarts, L., P. van Beukering, B. Kone and E. Wymenga (eds.) The Niger, a lifeline. Effective watermanagement in the Upper Niger Basin. Wetlands International/ RIZA/Altenburg & Wymenga, Veenwouden, The Netherlands. Van der Valk, A.G. (1981). Succession in wetlands: a Gleasonion approach. Ecology 62: 688- 696. Van der Valk, A.G. (1985). Vegetation dynamics of prairie glacial marshes. Pp. 293-312 in White, J. (ed.) The Population Structure of Vegetation. Junk, Dordrecht, The Netherlands. Van der Valk, A.G. (1992). Establishment, colonization and persistence. Pp. 60-102 in Glenn- Lewin, D.C., R.K. Peet and T.T. Veblen (eds.) Plant succession, theory and prediction. Chap- man & Hall, London, UK. Van der Weide, M. (2002). Blauwe Reiger. Ardea cinerea. Pp. 82-83 in SOVON. Atlas van de Nederlands broedvogels 1998-2000. Vogelonderzoek Nederland 2002, Nederlandse Fauna 5. Natopma; Natuurhistorisch Museum Naturalis, KNNV uitgeverij & European Invertebrate Survey-Nederland, Leiden, The Netherlands. Van Meel, van R.M. (1997). How to augment effectiveness and flexibility by curriculum development in agricultural higher education. A case for a study programme on biological diversity. Pp. 97-108 in Charrier, A. and R. van Haarlem (ed.) Biodiversity, an issue in High- er Education. Proceedings of the third European Conference on Higher Education for Agricul- ture. Wageningen University, Wageningen, The Netherlands. References 319

Vanpraet, C.L. (1977). L’écologie et l’amenagement du Parc National de Waza. PNUD/FAO, Rome, Italy. Venter, J. (1995). South African overview. Pp. 128-129 in Robinson, R. (ed.) Africa Heritage 2000: The Future of Protected Areas in Africa. Proceedings of the IUCN Commission on Na- tional Parks and Protected Areas African Regional Working Session. National Parks Board, Pretoria, South Africa. Verschuren, J. (1970). Sterven voor de olifanten (Mourir pour les Elephants). Het Wereldvenster, WNF, Zeist, The Netherlands. Walker, B., S. Carpenter, J. Anderies, N. Abel, G.S. Cumming, M. Janssen, L. Lebel, J. Nor- berg, G.D. Peterson, and R. Pritchard (2002). Resilience management in social-ecological systems: a working hypothesis for a participatory approach. Conservation Ecology 6(1): 14. [online] URL: http://www.consecol.org/vol6/iss1/art14/ Walters, C.J. and C.S. Holling (1990). Large-scale management experiments and learning by doing. Ecology 71: 2060-2068. Wanzie, C.S. (1988). Sociability of Buffon’s kob (Kobus kob kob Erxleben) in Waza National Park, Cameroon. Mammalia 52: 21-33. WCPA (2004). World Database on Protected Areas. www.sea.unep; wcmc.org/wdbpa/index.htm. Weber, B., and A. Vedder (2001). In the Kingdom of Gorillas. Simon and Schuster, New York, USA. Wellcomme, R.L. (1979). Fisheries Ecology of Floodplain Rivers. Longman, London, UK. Wells, M., K. Brandon and L. Hannah (1992). People and Parks. Linking Protected Area Man- agement with Local Communities. World Bank/WWF/ USAID, Washington DC, USA. Wesseling, J.W., C.A. Drijver, E. Naah, A.Namba, A. Zuiderwijk, D. Ngantou, P. Scholte and A. Tchappi (1994). Waza Logone flood restoration study. Delft Hydraulics, Delft, The Netherlands. Wesseling, H., E. Naah, C.A. Drijver and D. Ngantou (1996). Rehabilitation of the Logone floodplain, Cameroon, through hydrological management. Pp. 185-198 in Acreman, M. and G.E. Hollis (eds.) Water management and Wetlands in Sub-Saharan Africa. IUCN, Gland, Switserland. Western, D. (2003). Conservation science in Africa and the Role of International Collabor- ation. Conservation Biology 17: 11-19. Westoby, M. (1980). Elements of a theory of vegetation dynamics in arid rangelands. Israel Journal of Botany 28: 169-194. Westoby, M., B. Walker and I. Noy-Meir (1989). Opportunistic management of rangelands not at equilibrium. Journal of Rangeland Management 42: 266-74. Wetten, J.C.J. van and P. Spierenburg (eds.) (1998). Waders and waterfowl in the floodplains of the Logone, Cameroon, January, 1993. WIWO report nr. 67. Leiden, The Netherlands. Wetlands International (2002). Waterbird Population Estimates-Third Edition. Wetlands Inter- national Global Series No. 12, Wageningen, The Netherlands. Wild, R.G. and J. Mutebi (1997). Bwindi Impenetrable Forest, Uganda: Conservation through Collaborative Management. Nature and Resources 33: 33-51. Wilkie, D.S. and J.F. Carpenter (1999a). Can nature tourism help finance protected areas in the Congo Basin? Oryx 33: 332-338. Wilkie, D.S. and J.F. Carpenter (1999b). The potential role of safari hunting as a source of revenue for protected areas in the Congo Basin. Oryx 33: 339-345. Wilkie, D., E. Shaw, F. Rotberg, G. Morelli and P. Auzel (2000). Roads, Development and Con- servation in the Congo Basin. Conservation Biology 14: 1614-1622. Wilson, R.T. (1982). Environmental changes in western Darfur, Sudan, over half a century and their effects on selected bird species. Malimbus 4: 15-26. 320 Floodplain rehabilitation and the future of Conservation & Development

Wit, P. (1975). Preliminary notes on the vegetation of Waza National Park. FAO report, Rome, Italy. World Heritage Centre 2002. Periodic Report 3. Africa (2002). www.unesco.org/whc. Zeltner, J.C. (1988). Les pays du Tchad dans la tourmente 1880-1903. Racines du Présent. L’Har- mattan, Paris, France. Zeltner, J.C. (1997). Les pays du Tchad et la montée des périls 1795-1850. Racines du Présent. L’Harmattan, Paris, France. Zon, A.P.M. van der (1992). Graminées du Cameroun. Wageningen Agricultural University Papers 92-1. Wageningen, The Netherlands. Zwilling, G.U. (1940). Unvergessenes Kamerun. Zehn Jahre Wanderungen und Jagden 1928-1938. Verlag von Paul Baren, Berlin, Germany. 321 Summary

Introduction

The fertility of floodplains is legendary. Fluctuations in water level create a seasonal cycle of flood and drought, allowing a high primary production, abundant wildlife and high human population densities. Since the 1950s, demand for irrigation water and electricity has increasingly disturbed the natural flooding regimes of rivers in Africa. In 1979, the Waza-Logone floodplain in semi-arid Cameroon followed this fate by the construction upstream of the Maga dam and its embankment, greatly reducing the flooding intensity in an area of 1500 km2, including Waza National Park. As a consequence, annual grasses invaded productive perennial grasslands, reducing the carrying capacity of the area for livestock, fisheries and wildlife. The cultivation of irrigated rice, the main purpose of the Maga dam, was largely a failure, as testified by the use of less than half of the irrigation scheme’s capacity and continuing import of rice. In 1994, after intensive local consulta- tions, the Waza-Logone Integrated Conservation and Development Programme re-opened a watercourse previously blocked by the embankment along the Logone river. The reinstatement of the natural flooding regime in an area of 600 km2 aimed to restore both wildlife and human use of the desiccated floodplain. This study questions if, and how, reflooding may lead to the restoration of the Waza-Logone floodplain to its pre-dam structure and conservation and development functions. It does so by looking at the ecological history of the floodplain over the past two centuries (Ch. 2) and analysing the impact of the reflooding on the vegetation (Ch. 3 and 4), the waterbirds (Ch. 5), the antelopes (Ch. 6) and pastoral use (Ch. 7). Based on these analyses the integration of development and conservation and conflicts between the two are discussed (Ch. 8), as well as the usefulness and limitations of conflict mitigating management planning (Ch. 9). This is then related to present and possible future training of protected area managers in Africa (Ch. 10), and how these trainees view their own training in disciplines that have been lacking so far (Ch. 11). A synthesis and final conclusions are presented in Chapter 12.

The ecological history of the Waza-Logone floodplain In Chapter 2 I argue that the reference image of the pre-Maga dam period when the ‘floodplain was still intact’ was biased by the period from which information was consulted. Changes in hydrology, vegetation, land use, and especially wildlife have occurred regularly throughout the last two centuries, but in several cases the system’s resilience has led to their recovery. The intensity and speed of changes in 322 Floodplain rehabilitation and the future of Conservation & Development

the post Maga dam period, triggered by lower than average rainfall and man-induced drought, were, however, unprecedented in recent times. When studying the impact of reflooding, there is a need to distinguish between the different causes of these changes.

The impact of reflooding in Waza-Logone

In Chapters 3 and 4 the main question is: Does reflooding lead to a 100% perennial grass cover with the same floristic composition and production that existed prior to the Maga dam construction? And what are the mechanisms through which these floristic and production changes take place?

In Chapter 3, results are presented from vegetation studies from 1984 onwards along a transect covering always annually flooded, reflooded since 1994 as well as desiccated parts of the floodplain. Fragmentary information suggests that in the early 1980s in extended parts of the area perennial grasses died. This can be largely attributed to the dam construction, although there were indications that the condition of especially Vetiveria nigritana was already poor due to the mid-1970s drought. From 1993 to 1999, the floristic composition has been monitored in a grid in the centre of the reflooded zone. Following the reflooding, the cover of perennial grasses, most notably Echinochloa pyramidalis and Oryza longistaminata, increased again from 41 to 75% in the reflooded zone. The cover of annual species, most notably Sorghum arundinaceum, a dominant annual grass only since the mid- 1980s, decreased in the reflooded zone from 58% to 23%. If the observed conversion rate of annual into perennial grassland is extrapolated, a recovery towards a ‘100%’ perennial state was most likely reached after the 2003 flooding season. This was confirmed by qualitative observations in the transect in 2002. In terms of species composition, this rehabilitation was only partial as of the previously dominant perennial floodplain grasses, Vetiveria nigritana did not show any sign of re-establishment despite its good condition elsewhere in the floodplain. The initial hypotheses of a ‘full’ recovery, in five years, to the pre-Maga state from the 1970s was therefore rejected. The layout of our study, with plots both in and outside the reflooded area, showed the large impact of reflooding compared to a low impact of merely average rainfall. The gradual recovery of Oryza longistaminata and Echinochloa pyramidalis grasses can be explained by lateral rhizomatous growth. The tussock grass Vetiveria nigritana depends for its multiplication on seed dispersal, a less secure strategy that may depend on entirely different environmental conditions. This holds in particular for the direction of flooding water, and therefore the sediments it carries, which has remained different from the pre-dam situation. With future reflooding one should take into account a period of at least ten years for the recovery of rhizomatous grasses. If rainfall is lower than average, this period may be much longer. Full recovery that includes again a predominant pres- Summary 323 ence of Vetiveria nigritana may, apart from favourable climatic conditions, also be dependent on the restoration of soil fertility.

Flood depth has been frequently used to explain the distribution of plant species, but its relation with vegetation production has remained ambiguous. In Chapter 4, we studied the relation between flooding and above-ground biomass at the end of the flooding season, as well as the dry season, to assess the impact of reflooding on vegetation production. Above-ground biomass of a combination of all species and of the individual perennial grasses Oryza longistaminata and Echinochloa pyramidalis showed a positive linear relationship with maximum flood depth up to one meter. These relations became stronger during the two years following the water raising, showing the lag in reaction time to floodplain rehabilitation. Above- ground biomass data from other major floodplains in the three main African geographic regions showed a similar relation with maximum flood depth up to one meter. Dry season regrowth, important because of its high nutrient quality in a period of forage scarcity, was not directly related with maximum flood depth. A comparison with other African floodplains hints at an exponential relation of regrowth production if maximum flood depth exceeds one meter. Below this maximum flood depth, the timing of burning has an overwhelming influence on the (low) regrowth. Although the mechanism of the relationship between maximum flood depth and above-ground biomass is not yet fully understood, presented data allow an assessment of the impact of the 20 cm water level rise due to the reflooding. This way it may be estimated that wet season above-ground biomass has increased by approximately 10% in the first reflooding year increasing to approximately 37% in later years. For future reflooding I postulate the same two-year time-lag and a vegetation production increase following the presented maximum flood depth – above-ground biomass equation. In addition, dry season regrowth production is expected to increase only after several years, when sufficient rhizome biomass has been accumulated.

Does reflooding lead to a recovery of waterbirds? In Chapter 5, I attribute the drop in numbers of the area’s flagship species, the resident Black-crowned Crane, from 10 000 in 1973 to 2 000-2 500 individuals in 1992 to the Maga dam construction. The impact of the reflooding was monitored by January total counts from 1992-2000. Total waterbird numbers in the dry season increased from 60 000 to 105 000, whereas the number of species surpassing the international 1% criterion doubled from 6 to 12. The increase in Anatidae corresponds to their recovery over most of West Africa following the droughts in the 1980s. The increase in especially Ciconiiformes in Waza-Logone was not paralleled by similar trends in other main West African floodplains, suggesting that floodplain rehabilitation has played an important role. Amongst the water birds that showed a trend less than the general two-fold increase, were the large predominantly piscivorous Marabou and Yellow-billed Storks. This is attributed to the repeated destruction 324 Floodplain rehabilitation and the future

of their colonies by (fisher) men, a recent development in Waza-Logone that elsewhere in West Africa has already been under way for years. In contrast, a Black-headed Heron colony increased from 750 to 2500 nests between 1993-2003. The exceptional colony size, a multiple of the next largest known colony, suggests that besides improved habitat due to reflooding, protection also played a vital role. It can be concluded that the increase in waterbirds in Waza-Logone is due to a combination of factors: improved rainfall (especially Anatidae); floodplain rehabilitation (especially omnivorous Ciconiiformes); and protection measures (selected Afrotropical Cico- niiformes). Hypotheses for the recovery of waterbirds by future reflooding should take into account their long-term dynamics based on regional changes (e.g. rainfall) and site-specific changes especially human protection and prosecution.

Does reflooding lead to the recovery of antelopes to their pre-dam numbers? Antelopes, the subject of Chapter 6, are amongst the most prominent wildlife in Waza National Park that has witnessed changes in rainfall, hydrology and human encroachment during the last decades. To assess their impacts, we reviewed 26 surveys, comprising total, transect and localised counts, both aerial and terrestrial held from 1962 to 2001. Numbers of Kob antelopes, a key floodplain species, crashed from 20000 to 5000 in the period 1979-1983 following the Maga dam construction that also coincided with a period of rinderpest. Numbers further declined to 2000 at the end of the 1985 drought. The Kob population numbers increased between the late 1980s and the early-mid 1990s from 2000 to 5000 individuals. The initial recovery in Kob numbers following the reflooding did not continue beyond 1997. Estimated Korrigum (‘Topi’) and Roan numbers dropped already in the early 1970s and slightly recovered to respectively 2000 and 1000 in the 1990s. The disappearance of Waterbuck, soon after the construction of the Maga dam, has become a symbol of the area’s degradation. Two Waterbuck individuals were observed once again in 1998 and 2003. The diversity of counting methods notwithstanding, the drop in Kob numbers and the disappearance of Waterbuck can be attributed to the construction of the upstream Maga dam in 1979 and subsequent low rainfall. Yet Kob population structure data showed that its reproduction capacity was not severely hit. Kob recovered in the late 1980 and early-mid 1990s, but recovery did not continue beyond that, despite increased flooding. The postulated explanation is increased competition with livestock due to the rapidly increasing cattle grazing intensity (see below). Direct persecution, made possible by the dramatic drop in number of park guards, has probably played an important role as well. Floodplain rehabilitation hypotheses should therefore not only be based on water management but include feedback mechanisms of conservation effectiveness as well.

Does reflooding lead to full recovery of pastoral use of the floodplain? In Chapter 7, responses of mobile pastoralists to the floodplain rehabilitation program were assessed through interviews with leaders of over 100 pastoral camps, held at the end of each grazing season from 1993 to 1999. We registered changes in Summary 325 number of camps and herds, and the time spent in the 600 km2 of the Logone floodplain that was reflooded in 1994. The first year, few pastoralists anticipated the reflooding or its impact, and the increase in grazing intensity was caused by a prolonged stay of pastoralists who already used the area for transit. The following three years showed a sharp increase in the number of camps, which stabilised from 1997 onwards. The number of cattle herds doubled, two-thirds of which was due to newcomers and one-third to ‘natural’ increases. The average time cattle herds were present in the reflooded area increased by 60%. Overall, grazing intensity, expressed as cattle density, increased linearly following the gradually recovering perennial grasslands to three times its value in 1993. It is expected that this increase has levelled off at a maximum of about 100 cattle km-2 around 2003. Almost all observed changes were attributed to the reflooding, the only factor that differed between the studied floodplain and surrounding areas from where pastoralists moved away. No ‘territorial blocking’ nor ‘chaotic overshoot’ scenario has taken place with the reflooding. The Ideal Free Preemptive Distribution scenario, assuming that any increase in perennial vegetation is consumed by cattle with the first individuals using the territories pre-empting them, best explained pastoralist responses to floodplain rehabilitation. Pre-emptive use might be reflected in the customary rights that pastoralists have to campsites, which they have occupied for a long time. One of implications of the Ideal Free Pre-emptive Distribution is that with reflooding, apart from a short lag in pastoralist reaction time, all additionally produced forage is consumed, leaving few incentives for pastoralists to refrain from grazing inside Waza National Park. I postulate this Ideal Free Preemptive Distribution scenario for future reflooding. Un- fortunately, this scenario predicts continued grazing of livestock inside Waza Na- tional Park, unless strict measures are taken.

Enhancing Conservation & Development integration by management planning and training

If rehabilitation is ecologically successful, how may then the ecosystem’s functions that underlie the balance between Conservation and Development be assured?

In Chapter 8 I discuss the risks of Integrated Conservation and Development Pro- jects (ICDPs) in particular the overshoot of their successes. ICDPs aim to stimulate conservation without the previous negative experiences for local people, but pay little attention to their long-term impact such as immigration. The rehabilitation of the Logone floodplain, the core activity of the Waza-Logone ICDP, aimed at restoring the area’s services for both wildlife and human populations. The reflooding has led to a 34% increase of sedentary fishermen and a multiple number of temporary fishermen. Whereas livestock intensity tripled, Kob antelopes have not increased, indicating a reduction in their competitiveness. The virtual disappearance of wildlife in nearby Kalamaloué NP, due to advanced human encroach- 326 Floodplain rehabilitation and the future of Conservation & Development

ment, is therefore a bleak perspective for Waza NP. Examples from the Central African Republic, Galapagos, Nigeria and Zimbabwe also showed that in open- access systems, improvement in living standards (development) may stimulate immigration, jeopardising the stability necessary in protected areas (conservation). Most ICDPs lack demographic monitoring, masking its possible immigration risk. To counter the immigration risk in Waza, a management policy was formulated based on local stakeholder categorisation and subsequent privileges. The implementation of the management plan that offered development activities only in villages outside of the national park resulted in the voluntarily displacement of a village out of Waza NP. Whereas decades of oppression from park authorities towards this village only caused frustration, the new more balanced policy of the ‘carrot and stick’ yielded a tangible result. It is further recommended that ICDPs should be involved in regional land-use planning to discourage development activities that stimulate immigration.

In Chapter 9, management planning is discussed as means to get a ‘grip on the whole situation’. Management plans, such as the one formulated for Waza NP, are generally considered to be indispensable for the daily management of protected areas. Historically, their focus was on relatively straightforward technical or ecological issues. Since the 1990s, management plans have increasingly been based on the consensual interpretation of pressing management issues, often related to the legal context of protected areas. The most important product of the modern management plan is a consensus building process based on negotiation among stakeholders. This chapter asks whether the high expectations for management planning as an interface between protected area management and wildlife law and policy are realistic. The analysis first sketches protected area management policy in Cameroon, particularly the limitations of the legal context in which it occurs. This is followed by an account of the formulation process for the Waza NP management plan, a process that has led to a consensus on sensitive key issues among local communities, local and national authorities, and international environmental NGOs. This particular exercise in management planning was a learning process for all concerned. The analysis questioned, however, whether the resulting management plan was sufficiently realistic to lead to its successful implementation. The Waza case was not unusual, as shown by a comparison with management planning experiences in other protected areas in Cameroon and other African countries.

Protected area managers generally have little input in protected area planning and community conservation. Their lack of capacity in these disciplines was not unique to Waza-Logone, but also identified as a major constraint for Integrated Conserva- tion Development Programs elsewhere in Africa. In the chapters 10 and 11 I present the experiences at the Garoua Wildlife College, Cameroon, in its endeavours to reinforce the capacities of protected area personnel in Waza NP and more generally in West-Central Africa. Summary 327

Regional wildlife colleges in Cameroon, Tanzania and, recently, South Africa have been responsible for the training of 4000 protected area managers in Africa. Train- ing need assessments called for major curriculum reforms, which were developed and implemented in the late 1990s. Chapter 10 analyses the factors that influenced this curriculum reform in the colleges’ endeavour to respond to new developments in African wildlife management. Since 1979, the curriculum of Garoua Wildlife College has changed only gradually, whereas work placement subjects, selected by students and their employers, have quickly responded to developments in wildlife management, such as by an increase in the number of people-oriented subjects. In the new curriculum, Garoua’s mid-career students appreciated biology and inventory disciplines for their relevance, as well as courses in other disciplines tailored to conservation practice. The curriculum reforms implemented at Garoua depended on the presence of interested lecturers with an appropriate background, often obtained by additional training. The curricula of the regional wildlife colleges at Garoua, Mweka (Tanzania) and SAWC (South Africa) showed important differences as a result of regional characteristics and differing visions of the wildlife management profession. All three colleges have struggled to establish a feed-back system for continuous curriculum review. Increasing the exchanges between the colleges could further develop the curricula. While pursuing necessary changes in curriculum and institutions, care should be taken not to reduce the colleges’ sustainability.

Training need assessments for African protected area managers have revealed the need for people-oriented training. In the late 1990s, the Garoua Wildlife College, Cameroon developed long (diploma and certificate) and refresher courses in community conservation for mid-career managers and guards from Waza NP and other West and Central African protected areas. Through lectures, case studies and Participatory Rural Appraisal exercises, the courses emphasised the development of skills for tuning principles of people participation to the conservation objectives of protected areas. Chapter 11 reviews the trainees’ evaluations of these courses to appreciate their relevance and support their further development. Di- ploma students judged the course as highly relevant because of the acquired analytical skills, whereas certificate students considered them of medium relevance only. The reaction on short refresher courses varied as a function of the use of cases from either the trainees’ professional experience or from the fieldwork location. The reactions of trainees to this learning opportunity suggest that protected area personnel are not ‘attitude-limited’ as often suggested. Their constraints to develop a more people-oriented work style lie largely in the areas of knowledge and skills. These findings motivate increased efforts to implement training of protected area personnel in community conservation, preferably early in their careers. 328 Floodplain rehabilitation and the future of Conservation & Development

Synthesis

In the concluding Chapter 12, the outcome of the floodplain rehabilitation is reviewed by comparing the reflooding responses that, because of their different response times, caused human supported resources (cattle and Black-headed He- rons) to become dominant. I analyse the contributions of management planning and capacity building to correct these undesired effects of the otherwise successful reflooding. The implementation of the management plan and training program has been successful on a local level, as testified by the improved relationship between park authorities and local communities. Poor human resource management in the Cameroonian wildlife sector has overshadowed several results however.

I subsequently review different concepts of ecosystem change and management that help to understand and communicate the observed changes of floodplain resources. Panarchy, a holistic empirical concept, allows observed ecological and socio-economic developments to be linked in a temporal and spatial hierarchy. This hierarchy stresses the time lags between resources and the instalment of reflooding, as well as with capacity building, management and developments on which should be anticipated. Only by considering the separate ecosystem elements, causal explanations of the observed developments can be highlighted. Reflooding induced the rehabilitation of the floodplain, not its restoration, as the 2000-2003 ‘end result’ differed from the pre-dam situation with regard to the areas’ hydrology, vegetation composition, livestock-antelope ratio, etc.. In addition, the recovery process of several floodplain resources (tussock grasses, antelopes) has followed a different path and speed than the degradation following the dam construction. This hysteresis has led to the land use state outcome that differed from the pre-dam situation in its conservation value. I subsequently discuss the concept of resilience that illustrates the remarkable capacity of the Waza-Logone ecosystem to over- come disturbances. As this recovery not always leads to the identical reference situation, I introduce the concept of evolving states to explain that observed (vegetation) states are not necessarily static, but may undergo cyclic succession.

At the end of this thesis I analyse the pursued overarching adaptive management approach. I therefore revisit the monitoring observations and insights to reformulate the floodplain rehabilitation expectations. I call special attention to the role of a ‘vision’, including not only targets for rehabilitation but also understanding the required enabling environment. Empirical concepts of change and exchanges with other experiences played an important role in the development of this vision. One may conclude that adaptive management, based on long term monitoring of separate ecosystem components and a regular review of targets within an overall vision, forms a useful framework for the rehabilitation and management of tropical (floodplain) environments. 329 Samenvatting

Inleiding

De vruchtbaarheid van overstromingsvlakten is vermaard. Fluctuaties in water- niveaus creëren een cyclus van overstromingen en daaropvolgende droogten, met als gevolg een hoge primaire productie, talrijk wild en vaak een hoge bevolkings- dichtheid. De vraag naar irrigatiewater en elektriciteit heeft sinds de jaren vijftig echter in toenemende mate de natuurlijke loop van rivieren in sub-Sahara Afrika veranderd. Dat geldt ook voor de Waza-Logone vloedvlakte in het droge Noord- Kameroen. De Maga-dam en een dijk werden aangelegd, waardoor de jaarlijkse overstromingen in een gebied van 1500 km2, inclusief het Waza National Park, afnamen. Eenjarige grassen verdrongen de meerjarigen, waardoor de draagkracht van het gebied voor vee, visserij en wild sterk afnam. Het cultiveren van geïr- rigeerde rijst, de belangrijkste reden voor de bouw van de dam, werd een misluk- king. Minder dan de helft van de irrigatiecapaciteit wordt gebruikt en er wordt nog steeds rijst geïmporteerd. Na intensief lokaal overleg opende in 1994 het geïnte- greerde natuurbeschermings- en ontwikkelingsproject Waza-Logone een zijrivier die eerder door de dijk van de Logone rivier was afgesneden. Een gebied van zo’n 600 km2 werd hierdoor opnieuw bevloeid met als doel zowel de natuur als het menselijk gebruik ervan te herstellen. Deze studie onderzoekt of en hoe herbevloeiing kan leiden tot het herstel van de structuur en de natuurbeschermings- en ontwikkelings- functies van de Waza-Logone overstromingsvlakte van voor de dam.

De ecologische geschiedenis van de Waza-Logone overstromingsvlakte In hoofdstuk 2 beredeneer ik dat het referentiebeeld van de periode voor de dam, toen de vloedvlakte nog intact leek te zijn, beïnvloed was door de geraadpleegde informatie. Veranderingen in hydrologie, vegetatie, landgebruik en vooral wild, had- den vaak al veel eerder plaatsgevonden. De veerkracht van het gebied heeft meest- al geleid tot herstel. De intensiteit van veranderingen in de periode na de dam, nog verergerd door geringe regenval, waren echter ongekend. Het is daarom belang- rijk onderscheid te maken tussen de herkomst van de geobserveerde veranderingen.

De invloed van herbevloeiing in Waza-Logone

De leidende vraag in de hoofdstukken 3 en 4 is: Leidt herbevloeiing tot 100 procent meerjarig grasland met dezelfde floristische samenstelling en productie van voor de Maga- 330 Floodplain rehabilitation and the future of Conservation & Development

dam? En wat zijn de mechanismen die leiden tot deze veranderingen? In hoofdstuk 3 presenteren we de resultaten van vegetatiestudies vanaf 1984 langs een transect dat loopt van de jaarlijks overstroomde vloedvlakte, door de herbevloeide naar de sinds de Maga dam uitgedroogde delen van de vloedvlakte. Fragmentarische informatie suggereert dat begin jaren tachtig op veel plaatsen in het gebied meerjarige grassen stierven. Dit was grotendeels het gevolg van de komst van de dam, hoewel er indicaties zijn dat vooral Vetiveria nigritana al in slechte conditie was door de droogte van midden jaren zeventig.

Tussen 1993 en 1999 werd de floristische samenstelling gevolgd in een grid in het centrum van de invloedzone van de herbevloeiing. Na de herbevloeiing is de bedekking door meerjarige grassen, vooral Echinochloa pyramidalis en Oryza longistaminata, toegenomen van 41 naar 75 procent. De bedekking van eenjarige grassen, met name Sorghum arundinaceum die dominant was sinds midden jaren tachtig, nam af van 58 naar 23 procent. Als de geobserveerde conversiesnelheid van eenjarig naar meerjarig grasland wordt geëxtrapoleerd, is een 100 procent meerjarig graslandsituatie waarschijnlijk in 2003 bereikt. Observaties in het transect uit 2002 bevestigen dit beeld.

Wat soortensamenstelling betreft heeft de natuur zich slechts gedeeltelijk hersteld omdat Vetiveria nigritana nog ontbreekt, ondanks de goede conditie van deze soort elders in de overstromingsvlakte. De aanvankelijke hypothese dat binnen vijf jaar de situatie volledig hersteld zou zijn, moest daarmee worden verworpen. De opzet van de studie, met proefvlakken in en buiten het herbevloeiingsgebied, toont de sterke invloed van herbevloeiing aan en de relatief lage invloed van regenval in de betreffende periode. Het geleidelijke herstel van Oryza longistaminata en Echino- chloa pyramidalis kan worden verklaard uit laterale groei van rhizomen. Het pol- gras Vetiveria nigritana daarentegen is voor vermeerdering afhankelijk van zaadver- spreiding; een onzekere strategie die waarschijnlijk van volstrekt andere milieu- omstandigheden afhangt. Toekomstige herbevloeiing moet rekening houden met een periode van tenminste tien jaar voor het herstel van rhizoomgrassen. Een volledig herstel, inclusief Vetiveria nigritana is, naast regenval, mogelijk afhankelijk van het herstel van bodemvruchtbaarheid.

Vloeddiepte wordt regelmatig gebruikt om de verspreiding van plantensoorten te verklaren. De relatie met vegetatieproductie is echter onduidelijk. In hoofdstuk 4 bestudeer ik de relatie tussen overstroming en bovengrondse biomassa aan het einde van het vloedseizoen en tijdens het droge seizoen, om zo de invloed van herbevloeiing op de vegetatieproductie te bepalen. Bovengrondse biomassa van alle soorten samen en die van de individuele soorten Oryza longistaminata en Echino- chloa pyramidalis, laten een positieve lineaire correlatie zien met maximale vloeddiepte tot ten minste een meter. Deze correlaties werden sterker in de twee jaar na de herbevloeiing. Bovengrondse biomassagegevens van andere overstromings- Samenvatting 331 vlakten in de drie belangrijkste geografische zones in Afrika laten een soortgelijke relatie zien. Hergroei in het droge seizoen, dat vanwege het hoge nutriëntenge- halte essentieel is voor de voedselvoorziening van vee en wild, bleek niet gerela- teerd aan maximale vloeddiepte. Een vergelijking met andere overstromingsvlakten suggereert een exponentiële relatie van hergroei met vloeddiepte indien die meer dan een meter bedraagt. Hoewel het causale verband tussen bovengrondse biomassa en maximale vloeddiepte niet bekend is, kunnen we dankzij de gepre- senteerde gegevens de invloed van grotere waterdiepte door herbevloeiing op vegetatieproductie bepalen. De schatting is dat de hoeveelheid bovengrondse biomassa in het eerste overstromingsjaar met 10 procent en in latere jaren met 37 procent is toegenomen. Voor toekomstige herbevloeiing veronderstel ik eenzelfde ver- traging van twee jaar en een toename van de vegetatieproductie, waarvoor de bere- kende relatie tussen bovengrondse biomassa en maximale vloeddiepte geldt. Her- groei neemt waarschijnlijk pas jaren later toe wanneer voldoende rhizoom biomassa geaccumuleerd is.

Leidt herbevloeiing tot herstel van het aantal watervogels? In hoofdstuk 5 beredeneer ik dat de afname van het aantal Zwarte Kroonkraanvogels, Waza-Logone’s parade- paardje, van tienduizend in 1973 tot tweeduizend exemplaren in 1992, toe te schrij- ven is aan de komst van de Maga-dam. De invloed van de herbevloeiing hebben we gevolgd aan de hand van totaaltellingen in de maanden januari van 1992 tot 2000. Het aantal watervogels, geteld in het droge seizoen, nam toe van zestig- duizend tot 105 duizend. Het aantal soorten dat het 1-procentscriterium overstijgt, verdubbelde van zes naar twaalf. De toename van Anatidae (eend-achtigen) komt overeen met het populatieherstel na de droogten in de jaren tachtig elders in West- Afrika. De toename van Ciconiiformes (ooievaars en reigers) in Waza-Logone werd niet geëvenaard door de andere West-Afrikaanse overstromingsvlakten. Dit kan hoogstwaarschijnlijk worden toegeschreven aan de herbevloeiing. Watervogels die minder dan gemiddeld toenamen, waren de voornamelijk visetende Maraboe en Nimmerzat-ooievaars. Dit teleurstellende resultaat kan toegeschreven worden aan de regelmatige vernietiging van kolonies van deze soorten. Dit is een recent feno- meen in Waza-Logone dat elders in West Afrika al langer plaatsvindt. De bevindingen rond een Zwartkopreigerkolonie die tussen 1993 en 2003 toenam van 750 naar 2500 nesten, contrasteren hiermee. De uitzonderlijke grootte van de kolonie suggereert dat naast een verbeterd habitat door herbevloeiing ook bescherming een rol speelt.

Geconcludeerd kan worden dat de toename van watervogels in Waza-Logone het gevolg is van een combinatie van factoren: toenemende regenval (dat geldt vooral voor de Anatidae), vloedvlakteherstel (met name omnivore Ciconiiformes) en be- schermingsmaatregelen (specifieke Afrotropische Ciconiiformes). Hypotheses voor het verdere herstel van de watervogelpopulaties door toekomstige herbevloeiing dienen rekening te houden met lange-termijn dynamica, gebaseerd op regionale 332 Floodplain rehabilitation and the future of Conservation & Development

veranderingen (regenval) en gebiedsspecifieke veranderingen, met name menselijke vernietiging en bescherming.

Leidt herbevloeiing tot het herstel van antilopenpopulaties tot de pre-dam aantallen? Antilopen, onderwerp van hoofdstuk 6, behoren tot het meest opvallende wild in Waza National Park. Veranderingen in neerslag, hydrologie en menselijke druk hebben grote gevolgen gehad voor hun aantallen. Om deze gevolgen te bepalen, onderzochten we 26 totaal-, transect- en gelokaliseerde tellingen, uitgevoerd tussen 1962 en 2001, zowel vanuit de lucht als vanaf de grond. Het aantal Kob-antilopen, een belangrijke overstromingsvlaktesoort, daalde dramatisch van twintigduizend tot vijfduizend in de periode tussen 1979-1983: de periode na de dam en een bijkomende runderpestepidemie. Aantallen daalden verder naar tweeduizend aan het einde van de droogte in 1985. De Kob-populatie steeg tussen eind jaren tachtig en eind jaren negentig van tweeduizend naar vijfduizend. Het geschatte aantal Korrigum (‘Topi’) daalde begin jaren zeventig al en herstelde zich enigszins tot tweeduizend en duizend exemplaren in de jaren negentig. Het verdwijnen van de Waterbok na de komst van de Maga-dam, werd een symbool van het verval van het gebied. In 1998 en 2003 werden er echter weer twee gesignaleerd.

Ondanks de diversiteit in telmethodes kan de afname van Kob en Waterbok toegeschreven worden aan de Maga-dam en daarop volgende periode met weinig neerslag. De populatiestructuur gegevens van de Kob laten zien dat de reproductie- capaciteit niet verminderd was. Aantallen van de Kob herstelden zich tot de midden 90er jaren, maar stokte vervolgens, ondanks de toegenomen overstroming ten gevolge van de herbevloeiing. Een mogelijke verklaring hiervoor is de toenemende concurrentie met het aanwezige vee (zie hieronder). Directe vernietiging heeft waarschijnlijk ook een rol gespeeld. Het sterk afgenomen aantal parkwachters heeft hier ongetwijfeld invloed op gehad. Verwachtingen met betrekking tot resultaten van vloedvlakteherstel moeten daarom niet alleen gebaseerd zijn op water- beheer, maar ook rekening houden met de effectiviteit van de natuurbescherming.

Leidt herbevloeiing tot het volledige herstel van veehouderij in de overstromingsvlakte? In hoofdstuk 7 onderzoeken we de reacties van nomadische veehouders op het her- stelprogramma. Dat gebeurde door middel van interviews in meer dan honderd veehouderkampen aan het eind van de droge seizoenen van 1993 tot 1999. We registreerde veranderingen in de hoeveelheid kampen en kuddes en in de tijd die ze doorbrachten in de herbevloeide 600 km2 van de Logone overstromingsvlakte. In 1994, het eerste jaar van de herbevloeiing, anticipeerden maar een paar veehouders op de veranderingen. De begrazingsintensiteit nam toe doordat veehouders er langer bleven. In de drie daarop volgende jaren nam het aantal kampen sterk toe, maar dat stabiliseerde vanaf 1997. In totaal verdubbelde het aantal rundveekud- des, waarvan tweederde bestond uit nieuwkomers, en eenderde dankzij natuurlijke toename. De begrazingsintensiteit, uitgedrukt in runderdichtheid verdrie- Samenvatting 333 voudigde en nam lineair toe en volgde daarmee de geleidelijk herstellende meerjarige graslanden. De verwachting is dat de toename in 2003 is gestabiliseerd bij een maximum van 100 runderen per vierkante kilometer.

Praktisch alle geobserveerde veranderingen kunnen toegeschreven worden aan de herbevloeiing: de enige factor waarin de bestudeerde overstromingsvlakte en de omringende gebieden van elkaar verschilden. Met de herbevloeiing heeft geen ter- ritoriale afscherming noch chaotische overexploitatie plaatsgevonden. Het Ideal Free Pre-emptive Distribution-scenario, dat aanneemt dat iedere toename van meerjarige vegetatie wordt geconsumeerd, verklaarde als beste de reacties van veehouders op het herstel van de overstromingsvlakte. Preventief gebruik, waarbij de eerste individuen die de territoria gebruiken ze ook leeg maken, is mogelijk te verklaren uit de gebruiksrechten. Een van de gevolgen van Ideal Free Pre-emptive Dis- tribution is dat met herbevloeiing al het extra geproduceerde veevoer wordt geconsumeerd, waardoor veehouders niet gestimuleerd worden buiten Waza National Park te blijven. Ik postuleer ditzelfde Ideal Free Distribution-scenario ook voor toekomstige herbevloeiing. Helaas betekent dit ook dat er ook in de toekomst vee aan- wezigheid blijft in Waza National Park, tenzij hier strikte maatregelen tegen worden genomen.

Versterking van de integratie van natuurbescherming en lokale ontwikkeling door middel van beheersplanning en training

Als het herstel ecologisch succesvol is, hoe kunnen dan de functies van het ecosysteem die aan de basis liggen van de balans tussen natuurbescherming en lokale ontwikkeling worden gegarandeerd? In hoofdstuk 8 analyseer ik de risico’s van geïntegreerde natuurbeschermings- en ontwikkelingsprojecten (ICDP’s), met name het gevaar dat ze door hun succes hun doel voorbijschieten. ICDP’s beogen natuurbescherming te stimuleren zonder nadelige gevolgen voor een lokale bevolking. Ze besteden echter maar weinig aandacht aan de gevolgen voor de lange termijn, zoals immigratie.

Het herstel van de Logone overstromingsvlakte, de kernactiviteit van het Waza-Lo- gone ICDP, stond in het teken van een verbetering voor zowel natuur als mens. De herbevloeiing heeft geleid tot een toename van 34 procent van het aantal seden- taire vissers en een veelvoud daarvan aan seizoensvissers. Terwijl de veehouderij- intensiteit verdrievoudigde, nam het aantal Kob antilopen niet toe. Als gevolg van toenemende menselijke druk verdween in het naburig gelegen Kalamaloué Na- tional Park nagenoeg al het wild: een afschrikwekkend voorbeeld voor Waza Nation- al Park. Voorbeelden uit de Centraal Afrikaanse Republiek, Galapagos, Nigeria en Zimbabwe laten zien dat in dergelijke ‘open-acess’-systemen een verhoogde lev- ensstandaard immigratie kan stimuleren die de noodzakelijke stabiliteit ondermi- jnt. Het ontbreekt de meeste ICDP’s aan demografische gegevens, waardoor dit 334 Floodplain rehabilitation and the future of Conservation & Development

immigratierisico verborgen blijft. Om het immigratierisico in Waza-Logone te verlagen is een beheersbeleid opgezet dat is gebaseerd op een categorisatie van lokale belanghebbenden en hun privileges (zie hieronder). De uitvoering van het beheersplan stimuleerde ontwikkelingsactiviteiten in de dorpen buiten het park en leidde tot het vrijwillige verplaatsen van een dorp uit Waza National Park.

In hoofdstuk 9 bediscussieer ik beheersplanning als middel om greep op de situatie te krijgen. Beheersplannen worden gezien als onmisbaar voor het dagelijkse beheer van beschermde gebieden. Van oudsher ligt hun focus op technische of ecologische onderwerpen. Sinds de jaren negentig zijn ze vooral gericht op consensus, gebaseerd op onderhandelingen met een groot aantal belanghebbenden. In dit hoofdstuk vraag ik me af of de hoge verwachtingen van beheersplanning als schakel tussen het beheer van natuurgebieden enerzijds en natuurbeschermings- wetten en beleid anderzijds, realistisch zijn. Allereerst analyseer ik het natuurbe- schermingsbeleid in Kameroen, in het bijzonder de juridische beperkingen ervan. Daarna beschrijf ik de totstandkoming van het beheersplan voor het Waza Nation- al Park. Dat heeft geleid tot een consensus over onderwerpen die gevoelig lagen bij lokale gemeenschappen, lokale en nationale autoriteiten en internationale milieu-NGO’s. Deze beheersplanning was voor alle betrokkenen een leerproces. De analyse roept echter de vraag op of het managementplan realistisch genoeg was om uitgevoerd te worden. Hierin verschilt de Waza-case overigens niet van beheersplannen van andere natuurgebieden in Kameroen en elders in Afrika.

Natuurbeschermingspersoneel wordt over het algemeen nauwelijks betrokken bij planning en community conservation. Het ontbreken van vaardigheden in deze disciplines is niet uniek voor Waza-Logone, het is ook een probleem bij ICPD’s elders in Afrika. In de hoofdstukken 10 en 11 presenteer ik de ervaringen van het Garoua Wildlife College bij het verbeteren van de vaardigheden van het natuurbeschermingspersoneel in Waza National Park en elders in West-Centraal Afrika.

Regionale opleidingsinstituten in Kameroen, Tanzania en recentelijk Zuid-Afrika, hebben meer dan vierduizend Afrikaanse managers in de natuurbescherming op- geleid. Onderzoeken naar trainingsbehoeften toonden eerder aan dat er grote veranderingen nodig waren in de curricula. In de loop van de jaren negentig zijn die dan ook aangepast. Hoofdstuk 10 analyseert welke factoren van invloed zijn geweest op de pogingen van de opleidingsinstituten een antwoord te vinden op de nieuwe uitdagingen in natuurbescherming.

Sinds 1979 kende het curriculum van het Garoua College slechts geleidelijke veranderingen. Dit stond in schril contrast met de snelle ontwikkelingen die de studenten in hun stages tegenkwamen. Dat gold in het bijzonder voor mensgerelateerde onderwerpen. Midcareer-studenten waardeerden in het nieuwe curriculum de relevantie van vooral biologie en survey technieken. Dat gold ook voor andere Samenvatting 335 disciplines die sterk op de praktijk van de natuurbescherming waren gericht. Het doorvoeren van hervormingen in het curriculum was in Garoua afhankelijk van de interesse van trainers met de juiste achtergrond. De curricula van de regionale natuurbeschermingsopleidingsinstituten in Garoua (Kameroen), Mweka (Tanza- nia) en SAWC (Zuid Afrika) zijn onderling erg verschillend. Dit is het gevolg van regionale accenten en uiteenlopende visies op het natuurbeschermingsvak. Alle drie de instituten worstelen met het instellen van een feedbacksysteem voor een continue ontwikkeling van de curricula. Uitwisseling tussen de drie colleges kan een stimulans zijn voor verdere ontwikkeling ervan. Tegelijkertijd moet echter ook de continuïteit van de opleidingsinstituten niet uit het oog worden verloren.

Onderzoeken onder Afrikaans natuurbeschermingspersoneel brachten de behoef- te aan training in mensgerelateerde onderwerpen aan het licht. Eind jaren negentig heeft het Garoua College daarom cursussen ontwikkeld in community conservation voor haar mid-careerstudenten, alsmede voor beheerders en wachters van Waza National Park en andere beschermde gebieden in West-Centraal-Afrika. Door middel van hoorcolleges, casestudies en PRA-oefeningen, werden vaardigheden ontwikkeld om het betrekken van de lokale bevolking af te stemmen op de doelstellingen van beschermde gebieden.

In hoofdstuk 11 bespreken we de evaluaties van cursisten op de ontwikkelde cursussen. ‘Diploma-studenten’ (HBO-niveau) beoordeelden de cursussen als zeer relevant vanwege de opgedane analytische vaardigheden, terwijl ‘certificate-studenten’ (MBO-niveau) ze slechts als gemiddeld relevant beschouwden. De reacties op de korte opfriscursussen varieerden, afhankelijk van de mate waarin praktische cases werden gebruikt waarmee de cursisten vertrouwd waren. De reactie van de cursisten geeft aan dat de betrokkenheid van natuurbeschermingspersoneel grot- er is dan vaak wordt verondersteld en dat het ontbreken van een mensvriendelijk- er werkstijl veroorzaakt wordt door een gebrek aan vaardigheden en kennis. Deze bevindingen laten zien dat pogingen om natuurbeschermingspersoneel meer te betrekken bij community conservation, liefst ook vanaf het begin van hun car- rières, zinvol zijn.

Synthese

In hoofdstuk 12 evalueer ik het herstel van de overstromingsvlakte door de reacties van de afzonderlijke ecosysteemelementen te vergelijken. Door de verschillende reactietijden heeft dit geleid tot een dominantie van door mensen ondersteun- de hulpbronnen zoals vee en Zwartkopreigers. Ik analyseer vervolgens de bijdra- gen van beheersplanning en training om deze ongewenste effecten van de overigens succesvolle herbevloeiïng te corrigeren. De uitvoering van het beheersplan en het trainingsprogramma waren op lokale schaal succesvol, zoals de verbeterde 336 Floodplain rehabilitation and the future of Conservation & Development

verhoudingen lieten zien. Slecht personeelsbeleid in de Kameroense natuurbescher- mingssector heeft echter vele goede resultaten overschaduwd.

Ik bespreek vervolgens verschillende aspecten van veranderingen van ecosyste- men en management die de geobserveerde veranderingen kunnen verklaren. Pa- narchy, een holistisch empirisch concept, maakt het mogelijk de geobserveerde ecologische ontwikkelingen en interventies te verbinden in een hiërarchie van ruimte en tijd. Deze hiërarchie laat een vertraagde reactie zien, niet alleen tussen de hulpbronnen en de herbevloeiing, maar ook tussen beheer, capaciteitsopbouw en ontwikkelingen als immigratie. Alleen door de afzonderlijke ecosysteemelementen te bestuderen kunnen causale verbanden worden belicht. De herbevloeiing heeft een herstel van de overstromingsvlakte bewerkstelligd en geen restauratie. Het eindresultaat uit 2000-2003 wijkt af van de pre-dam situatie met betrekking tot hydrologie, de floristische samenstelling, de vee-antilope verhoudingen, enzo- voort. Ook het herstelproces van diverse hulpbronnen van de overstromingsvlakte is anders verlopen dan het degradatieproces na het aanleggen van de dam. Deze hysterese heeft geleid tot de huidige staat van landgebruik waarbij natuurbescherming een minder belangrijke rol speelt. Ik bediscussieer vervolgens het concept van veerkracht dat de opmerkelijke capaciteit van het Waza-Logone gebied weer- spiegelt om veranderingen te boven te komen. Omdat herstel niet altijd leidt tot de identieke referentiesituatie, introduceer ik het begrip ‘evolving states’ om uit te leggen dat de geobserveerde staat van de vegetatie niet noodzakelijkerwijs statisch is, maar ook cyclische successie kan ondergaan.

Ik besluit deze dissertatie met een analyse van de overkoepelende ‘adaptive management’ benadering. Hiervoor analyseer ik de observaties en verkregen inzichten om de herstelverwachtingen van de overstromingsvlakte nader te beschouwen. Ik vraag speciaal aandacht voor de rol van een visie, waarbij niet alleen oog moet zijn voor de doelstellingen van het herstel, maar ook de rol van ondersteunende dien- sten daarbij. Empirische begrippen van verandering en uitwisseling met andere ervaringen speelden een belangrijke rol in de ontwikkeling van deze visie. Ik con- cludeer dat adaptive management gebaseerd op lange-termijn monitoring van de afzonderlijke ecosysteemelementen en een regelmatige evaluatie van de doelstellingen binnen een algemene visie, een geschikt kader vormt voor het herstel en beheer van tropische (overstromingsvlakte) milieus. 337 About the Author

Born in 1962, Paul Scholte grew up in the area around Nijmegen. His passion for nature found an outlet through the NJN, the Dutch youth organisation for nature observation. Bird watching in the Ooijpolder and surrounding forests completed his VWO education at the Canisius College from 1974 to 1980.

In 1987, Paul graduated in Biology at Wageningen University. For a major in animal ecology he studied the link between land use, herpetofaune and birds of prey in Evros, Greece. During this first stay abroad, spontaneous contacts with local communities opened his eyes for the social dimension of nature conservation. He did majors in tropical soil science and vegetation science (mapping) during a one- year stay in Kenya. A minor was taken in extension science. During the first six years of his professional career Paul was active in rangeland ecology and management in Yemen (DGIS) and in Chad where he was project manager (DHV Consultants). With his colleagues, Paul prepared the first vegetation map of Yemen. In Chad he supervised studies on rangeland resources and their utilisation and translated results into policy advises to the Ministry of Livestock.

Employed by Leiden University, Paul headed the sections on national park assistance and ecological research at the Waza-Logone project (Cameroon) from 1993 to 1998. He co-ordinated the formulation of the Waza Management Plan and supervised impact studies of reflooding on vegetation, wildlife and pastoralists. Employed by DGIS and the Ministry of Environment and Forestry, Paul lectured at the Garoua Wildlife School (Cameroon) on participatory approaches in wildlife management and management planning from 1998 to 2000. He also co-ordinated curriculum reform at this regional college for protected area managers. The present thesis is a consolidation of the scientific publications he authored on these experiences.

Following his return to the Netherlands in 2000, Paul was independent consult- ant charged with project formulation, backstopping and evaluation for UNDP and Worldbank-GEF, IUCN, WWF, BCTF and others in Cameroon, CAR, Chad, Gabon, Mongolia, Morocco, Senegal and Sudan. Paul was appointed member of the Wad- den Sea Council and is member of the Dutch working group Ecology & Develop- ment, the Antelope Specialist Group (IUCN/SSC) and the Commission on Ecosys- tem Management (IUCN/CEM). From 2004 to 2005, Paul was senior biodiversity officer at NC-IUCN, focussing on corporate responsibilities with regard to biodiversity. Late 2005 Paul returned with Stephany, his partner, and children Thyne and Nika to Yemen as Chief Technical Advisor of the Socotra Conservation and Development Programme. 338 Background Publications

Research presented in this thesis is based on a series of field studies that elaborated upon methodologies and analysed preliminary experiences. Inventories complemented these background materials.

General Scholte, P., E. Pamo, S. Kari, S. Kersten and P. Kirda (1996). Floodplain rehabilitation in N. Came- roon: expected impact on vegetation, pastoralists and wildlife. Pp. 492-493 in West, N. (ed.) Pro- ceedings Fifth International Rangeland Congress. Society of Range Management, Colorado, USA. Scholte, P., W. Mullié, C.Batello, M. Marzot, A. H. Touré and D. Williamson (2004). Wildlife. Pp. 227-257 in Batello, C (ed.) The Future is an Ancient Lake. Traditional knowledge, biodiversity and genetic resources for food and agriculture in Lake Chad Basin ecosystems. FAO, Rome, Italy.

Birds Mullié‚ W.C., J. Brouwer and P.Scholte (1995). Numbers, distribution and habitat of wintering White Storks Ciconia ciconia in the East-Central Sahel in relation to rainfall, food and anthro- pogenic influences. Pp 219-240 in Biver, O., P. Enggist, C. Marti, T. Salath (eds.) Proceedings of the International Symposium on the White Stork (Western Population). Basel, Switserland. Scholte, P. (1996) Conservation status of Cranes in North Cameroon and Western Chad. Pp 153-156 in Beilfuss, R.D., W.R. Tarboton and N.N. Gichuki (eds.) Proceedings of the 1993 African Crane and Wetland Training Workshop. International Crane Foundation, Baraboo, Wisconsin, USA. www.savingcranes.org/data/program/pp153-156.pdf Scholte, P., S. de Kort and M. van Weerd (1999). The Birds of the Waza-Logone Area, Far North Province, Cameroon. Malimbus 21: 16-50. Scholte, P. (1999). Status of Vultures in the Lake Chad Basin, with special reference to North- ern Cameroon and Western Chad. Vulture News 39: 3-19. Scholte, P., S. de Kort and M. van Weerd (2000). Floodplain rehabilitation in Far North Cameroon: expected impact on bird life. Ostrich 71: 112-117. Scholte, P. and R.J. Dowsett (2000). Birds of Waza new to Cameroon: corrigenda and ad- denda. Malimbus 22: 29-31. Scholte, P. and P. Robertson (2001). Chad. Pp 177-184 in Fishpool, L.D.C. & Evans M.I. (eds.) Important Bird Areas in Africa and Associated Islands: Priority Sites for Conservation. Pisces publications and BirdLife International. Newbury and Cambridge, UK. Fotso, R., F. Dowsett-Lemaire, R.J. Dowsett, Cameroon Ornithological Club, P. Scholte, M. Languy and C. Bowden (2001). Cameroon. Pp 133-159 in Fishpool, L.D.C. & M.I. Evans (eds.) Important Bird Areas in Africa and Associated Islands: Priority Sites for Conservation. Pisces publications and BirdLife International. Newbury and Cambridge, UK.

Mammals Scholte, P (2000). Ressources en espèces sauvages. (Inventaire de la biodiversité). Pp 50-53 in GEPIS (2000) Vers une gestion durable des plaines d’inondation sahéliennes. IUCN, Gland. Switserland. Background publications 339

Scholte (2001). Notes on the status of antelopes in central and southern Chad. Pp 15-22 in East, R (ed.) Antelope Survey Update No 8. IUCN/SSC Antelope Specialist Group Report. Scholte, P. (2005). Dama gazelle. Pp (in press) in Kingdon et al. Mammals of Africa. Academic Press, London, UK. Scholte, P. and I. Hashim (2005). Red-fronted Gazelle. Pp (in press) in Kingdon et al. Mammals of Africa. Academic Press, London, UK.

Pastoralism Scholte, P, S. Kari and M. Moritz (1996). The involvement of nomadic and transhumant pastoralists in the rehabilitation and management of the Logone floodplain, North Cameroon. Issues paper 66, Drylands programme, IIED, London. (also in French). Kari, S. and P. Scholte (2001). La réhabilitation pastorale de la plaine d’inondation Waza-Logone (Cameroun): comment consolider sa réussite écologique? Pp 315-316 in Tielkes, E., E. Schlecht et P. Hiernaux (eds.) Elevage et gestion de parcours au Sahel, implications pour le développement. Comptes-rendus d’un atelier régional tenu à Niamey, Niger, du 2 au 6.10.2000. Verlag Ulrich E.Grauer, Beuren, Stuttgart Germany. Moritz, M., P. Scholte and S. Kari. (2002). The demise of the Nomadic Contract. Nomadic Peoples 6: 124-143 Scholte, P. and J. Brouwer (2005). The Relevance of Key Resource Areas for Large-Scale Movements of Livestock: are Sahelian floodplains an example? Pp (in press) in Prins, H.H.T. and F.van Langevelde. Spatial Ecology of Large Herbivores and Pastoralists.

On management Scholte, P., Saleh, A., Bobo, K and Boukar, B. (1998). Who gets the fish? Assessing competition between piscivorous Birds and the fishery on the Logone floodplain of Waza National Park, North Cameroon. Pp. 378 in Farina, A., J. Kennedy and V. Bossu. Proceedings International Eco- logy Congress, July 1998. Florence, Italy. Scholte, P., S.Adam, S.Kari and J.H. Mbouche. (1999) Walking a tightrope: Using PRA in a conflict situation around Waza National Park, Cameroon. PLA Notes 35: 7-12. (www.iied.org/sarl/pla_notes/pla_backissues/documents/plan_03502.PDF) Scholte, P. (2000) Migration Policy of the Waza-Logone Project in the Far North of Cameroon: from recognition to negotiation based on stakeholder identification. Pp. 13-14 in Vabi, M.B. Linking the conservation of biological diversity to sustainable development. WWF-Cameroon Programme Office, Yaoundé, Cameroon. Scholte, P. (2000). Towards consensual park management planning in Africa. Oryx 34: 87-89. Scholte, P. (2000) Towards collaborative management in Waza National Park: The Role of its Management Plan. Pp. 41-53 in Bauer, H. and A. Madi (eds.) People, parks and wildlife. Contri- butions from Cameroon. Proceedings of the Park-People Conference, Maroua February 1998. Centre for Environmental and Development Studies, Maroua, Cameroon. Est, van D. and P. Scholte (2001) Linking social and ecological diversity: organisations for adaptive natural resource management in the Waza-Logone floodplain (North Cameroon). Pp. 55-65 in Ali, M., P. Loth, H. Bauer and H. de Iongh (eds.) Management of fragile ecosystems in the North of Cameroon: the need for an adaptive approach. CEDC/CML. This publication was financially supported by WWF-Netherlands, Scholte Consulting and Leiden University.

The Waza-Logone project, under the auspices of which a major part of the fieldwork of this study was undertaken, was carried out by the Government of Cameroon and IUCN, in co-operation with the Institute of Environmental Sciences of Leiden University (CML), the Netherlands Development Organisation (SNV) and WWF-Cameroon, with the financial support of the Dutch Ministry of Foreign Affairs and WWF-Netherlands. My assignment at the Ecole de Faune was supported by the Dutch Ministry of Foreign Affairs (DGIS) and the Cameroonian Ministry of Environment and Forestry.

Acknowledgments

For the financial assistance of the production of this book, I am grateful to the efforts of Esther Blom (WWF-Netherlands). Matthew Parr (NC-IUCN) assisted with the language editing of the unpublished chapters, Linda Huijsmans with the Dutch translation of the summary, Mohamed Yesef with the French translation of the abstract. Sjoukje Rienks made a book out of all of this.

I am grateful to the staff of the Waza-Logone project, Michael Allen, Daniel Ngan- tou, Richard Braund, Maureen Roëll, Jaap Kok and Roger Kouakam as well as Patrick Dugan and Jean-Yves Pirot of the co-ordinating IUCN-wetland group for the support to the reported fieldwork. At the Waza-Logone Project I had a wonderful time with my colleagues who greatly shaped the outcome of the study: Saïdou Kari, Philippe Kirda, Saleh Adam, Bobo Serge Kadiri, Boukar Beladane, Paul Koua- mou, Etienne Pamo (Dschang University), the hydrologists Emmanuel Naah, and Daniel Sighomnou of IRGM and of the park village team Jean Hilaire Mbouche, Djebba and Madoum. In the floodplain Kaskala was our host and guide. His passing away is a great loss.

In the pastoral camps and villages we were always received with great hospitality and attention. Yerima Ouamarou, El Hadj Bouba, El Hadj Bello and El Hadj Eli learnt us the ins and outs of the pastoral life. The Waza National Park villages Andirni, Mahé, Halé and Zwang were our wet season floodplain homes. Saly of Halé was one of those immigrants, born in the floodplain, who made a difference. I would like to memorise the Blama (village chief) of Zwang who lost his life in the unscrupulous actions of the anti-gang movement in 1998. I am grateful to the Waza NP wardens Badjoda Daouda, Mahamat Habibou and Saleh Adam for the trustful collaboration, despite our sometimes different positions and responsibilities towards local communities. The Waza guides, Manga, Moussa Barka and his son Oumarou, and all the others, not only provided reliable guidance, but were a continuous source of information and feed-back on our observations. The friend- ship they offered has left deep marks (and the name of my son, Thyne Moussa). The Garoua Wildlife College was a stimulating environment for which I am grateful to the various generations of students. The directors Ibrahim Njoya, Jean Ngog Nje and Andrew Allo Allo, and recently Francis Tarla showed particular interest in my work. Mayna, Louis Tsague, Dong, Tchadel, Emmanuel Battokok, Rigobert Azombo, Etienne Hatungimana, Talla, Kwabong and Robert Ndim were inspiring colleagues. In addition, Battokok and Azombo were the driving force behind the waterbird counts. I was the last of a generation of Dutch trainers to the Ecole. I am grateful for the continuing interest of my predecessors Bart van Lavieren, Chris Geerling, Piet Wit, Ton van der Zon and Floris Deodatus.

Although the Ministry of Environment and Forestry as an institution was sometimes a source of frustration, the pleasant collaboration with the former directors of wildlife Denis Koulagna, Ndjoh à Ndiang and Mengang greatly improved our under- standings of the Yaoundé bureaucracy. At the CEDC (Centre for Environment and Development Studies), Madi Ali, Martin Tchamba, Franke Toornstra, Hanson Njiforti, Hannie Korthof and Waldo were our stimulating discussions partners.

My colleagues at CML, the Institute of Environmental Sciences of Leiden Univer- sity, tolerated my erratic presence for over a decade. Carel Drijver, who supervised the 1980s preparatory studies, was the driving force behind the initiation and early years of the Waza-Logone project. His engaged backstopping gave me a flying start. Jeroen van Wetten initiated several of the ornithological studies. Edith Roos, Annelies Oskam, Helias Udo de Haes, Hans de Iongh and Gerard Persoon have contributed in their various capacities to my work. The continuing feed-back of Wouter de Groot and that of Hans Bauer, Diny van Est, Ruth Noorduyn, Herwig Cleuren, Gerard Barendse and Paul Loth enriched my stay in Leiden. Evert Meelis advised on statistical analysis. I am grateful to Herbert Prins, Pieter Ketner, Frank Langeveld and their colleagues of the Resource Ecology Group, Wageningen Univer- sity, for their feed-back and stimulating discussions; Willemien Schouten kept the communication going. A rewarding part of my work was with students and later colleagues. Martine Graafland, Marlijn Hoogendoorn, Peter van der Jagt, Saskia van der Klundert, Bo Oosterhuis and Claudia van der Pot put in a vast effort to produce the data of the vegetation production studies. The ornithological studies by Selvino de Kort, Merlijn van Weerd, Ernée Raspe, Suzanne van der Giesen, Corinthe Zekveld and Hellen Elissen were a background for Chapter 5. Mark Mortiz vol- unteered in the beginning of the pastoral studies and has continued with us along the pastoral road ever since. 342 Floodplain rehabilitation and the future of Conservation & Development

I am further grateful to Joost Brouwer and Piet Wit for commenting on many versions of all chapters. Patrick Denny (IHE-UNESCO) reviewed the entire manu- script with formal and constructive informal comments. Approximately 70 reviewers of the various chapters overloaded me with hundreds of hours of work. Ulti- mately, many of their comments helped to sharpen my analyses and arguments, resulting in the presented chapters. Many reviewers have remained unknown, but in addition to above-mentioned people, the following persons have contributed to the final shape of the chapters. Jan Bakker and Ab Groothuis (Groningen Univer- sity) and Dick Visser (Nijmegen University) commented on the vegetation studies. Wim Mullié and Jan Wanink stimulated me to open my field notebooks and publish my ornithological observations. They also commented upon the subsequent manuscripts, as did Leo Bruinzeel and Eddy Wymenga (AW ecological consultants), Leo Zwarts (RIZA) and Marcel Klaassen (NIOO). Jeroen Bredenbeek and his WIWO colleagues provided the 1999 and 2000 waterbird counts data, as did Patrick Triplet (RN de la Baie de Somme) with the Senegal Delta data. Rod East, Chris Geerling, Bart van Lavieren and Peter van Bree commented on the antelope chapter. I am also grateful to Jean Ngog, Jean Thal and Louis Tsague (Ecole de Faune), Daan Bos, Luc Hebou and Hannie Korthof for the use of their unpublished antelope data. Han van Dijk (Africa Study Centre), Leslie Moore and Sean Overland commented the pastoralists chapter. Geoffrey Wandesforde-Smith (UC- Davis) edited the management planning chapter. Nancy Gellman, Deo-Gratias Gamassa and Freddy Manongi (Mweka), Fanie Greyling and Lynn Pullen (SAWC) and Rob van Haarlem (Wageningen University) contributed with their information and comments on the curriculum development and training chapters.

Passion for nature made this study fun. I am particular grateful to my old Ooijpol- der and Evros mates Jeroen en Wouter Helmer for accompanying me in this passion. With the stimulation and love of my parents I have been able to explore the directions of my life. Ever since the start of this study, Stephany Kersten has supported me with her love and help. Our children Thyne and Nika show there is even more in life. © 2005 Paul Scholte, Amsterdam

Lay out: Sjoukje Rienks, Amsterdam Photos: Paul Scholte, Amsterdam Maps: Yde Bouma, Leusden