Forest Decline in South Central Ethiopia Extent, History and Process

Forest Decline in South Central Ethiopia Extent, history and process

Gessesse Dessie

Department of Physical Geography and Quaternary Geology

Stockholm University

SE-106 91 Stockholm, Sweden © Gessesse Dessie ISSN: 1653-7211 ISBN: 978-91-7155-454-3 Layout: Clas Hättestrand Printed in Sweden: US-AB, 2007

Cover: Photograph showing Wondo Genet area in the eastern part of the study area. In the foreground smallholder agriculture is seen against the natural forest of the escarpments. Once a closed forest, the area is now agricultural land where annual crops such as maize are grown by smallholder farmers. Landholdings are less than one hectare in size. Whitish cluster in center left is the Wondo Genet College of Forestry. The light green tone at the foothills reveals predominantly khat farms pushing the forest frontier up the escarpment. Photograph by Gessesse Dessie, June 2004 Doctoral dissertation 2007 Degree of Doctor of Philosophy in Physical Geography Department of Physical Geography and Quaternary Geology Stockholm University Gessesse Dessie

Abstract

This study presents the extent, history and process of forest decline in Awassa watershed, south central Ethiopia. By combining different data sources such as satellite images, social surveys and historical documents, forest decline is described quantitatively and qualitatively and the main causes behind this process are identified. Forest decline in the study area is interpreted as the result of a combination of socio-political changes, economic activities, population growth, cultural patterns and agricultural developments while local conflicts over resources also play an important role. The findings of this study reveal forest decline to be a continuous process associated with spatial fragmentation and location specific losses. The recent increase in production of the cash crop khat has made a significant impact on the forest through several mechanisms: it relocates the agricultural/forest frontier; it causes intrusion and permanent settlement within forests; and fragments remaining forest. The analysis of human-spatial boundaries indicates unsystematic management of the natural forests by several administrative units. As a result, multiple claims have been made on the forests simultaneously as weak control and accountability conditions have negatively affected forest management. The main conclusions are as follows: Forest decline in the study area has a long history, spanning at least one century. The causes are identifiable as both temporally spaced individual events as well as chains of events. These interact with each other at different levels and scales as well as with the geographical properties of the study area. Land users’ rationale in weighing the advantages between keeping and replacing the forest is affected by economic gain, market conditions and transport facilities. Multiple claims to the forest land and weak accountability contribute to inefficient management, which accelerates forest decline.

Key words: Forest decline, people-forest interaction, boundaries, khat expansion, remote sensing, social survey, south central Ethiopia, Awassa watershed, Wondo Genet

Forest Decline in South Central Ethiopia Extent, history and process

Gessesse Dessie

This doctoral thesis consists of a summary and four appended papers.

List of Papers

Paper I: Gessesse Dessie and Kleman, J. 2007. Pattern and Magnitude of Deforestation in the South Central Rift Valley Region of Ethiopia. Mountain research and development 27(2):162-168. Paper II: Gessesse Dessie and Christiansson, C. 2007. Forest Decline and its Causes in the South Central Rift Valley of Ethiopia. Human Impact over a One Hundred Year Perspective. Ambio (accepted, resubmitted) Paper III: Gessesse Dessie and Kinlund, P. 2007. Khat Expansion and Forest Decline in Wondo Genet, Ethiopia. Geografiska Annaler (in review). Paper IV: Gessesse Dessie and Kleman, J. 2006. Ecological and Human Spatial Boundaries and their Impact on Forest Management and Forest Decline in Wondo Genet, Ethiopia. (Manuscript).

In this summary references to the above papers are made by their Roman numbers. The ideas and realization of all papers are entirely my own. All of the fieldwork, development of methods, data collection and analysis for all papers were undertaken by me. The writing part was done jointly with co-authors. They contributed to the discussion of findings and structures of the papers. In paper II and III the co-authors took part in the preliminary field trips. Paper I (uncorrected proof - see attached page for corrections) is presented in this thesis with permission from Journal of Mountain Research and Development.

Gessesse Dessie - Forest decline in South Central Ethiopia

Introduction decline rather than the causative processes. Therefore the need for a holistic approach is apparent. However, Natural forest is both an ecosystem and a resource: it is important to address the aforementioned chal- as an ecosystem it integrates diverse fauna, flora lenges, particularly the complexity of causative links and the physical environment; as a resource it has in time and space. various economic, ecological and social values. Fo- A more complete understanding of forest decline rest decline (Contreras-Hermosilla 2000) affects the requires 1) a means of determining magnitude and ecosystem as well as the usefulness of the forest as a characteristics with consistent estimation methods 2) resource. It involves multiple and interrelated factors an outline of processes: an intrinsic understanding of of people-forest interaction and results in significant functional chains and causative links 3) a pattern and environmental and social consequences (Lambin et trend analysis which delves beyond contemporary al 2001). snapshot observations 4) an investigation of whether People-forest interaction is determined by the current management/administrative conditions are interaction in space and over time of biophysical contributing to forest decline. and human factors (Moran and Ostrom 2005). A The magnitude of forest decline can be defined as study of such interaction needs to conceptualize the the difference between measurements taken at two relationship between the driving forces of human or more different points in time excluding possible induced changes, the processes and activities among errors (Contreras-Hermosilla 2000). The process them, and human behavior and organization (Turner analysis requires the identification of complex causa- et al 1994). tive links. Some causes imply complex links between Numerous studies have dealt with forest decline people and forest, such as in the case of cash crop issues. Various approaches exist, such as 1) regres- production, which has been shown to initiate links sion analysis of the causes and consequences with less at different scales between environment, people, emphasis on spatial links (e.g. Angelsen 1999); 2) the economy, culture and market (Grossman 1993; spatial representation of causes using a combination Nestel 1995; Govereh & Jayne 2002; Emich 2003; of remote sensing and socioeconomic techniques Keys 2004; Gebissa 2004). Landscape changes may (e.g. Fox et al 2003); 3) the analysis of causes in the occur over timescales longer than those that can be context of the larger political and economic environ- observed directly. Understanding the serial stage (time ment, at a number of temporal and spatial scales (e.g. line) of landscapes is necessary in order to identify Geist and Lambin 2002). While these approaches are more complex chains of causation (Antrop 2005; methodologically diverse, each provides essential per- Roberts 1996; Marcucci 2000: Batterbury and Beb- spectives necessary for the comprehension of tropical bington 1999). Boundaries are important elements forest decline. in the landscape because they institute claims and However, the following points express some of the are a preliminary form of management. Notably the practical, methodological and theoretical challenges human and ecological boundaries often do not match that still remain. and this mismatch and variation of boundaries may • Forest and biophysical variables are relatively over time lead to inefficient, inconsistent, wasteful, or fixed in space, while social variables are often destructive resource management (Meidinger 1998; fluid: they transcend space and time. These fun- Lovell 2002; Walker and Peters 2001; Cadenasso et damental differences in variable "stability" pose al 2003; Cumming 2006). problems of scale and cause identification, as well The problems in the study area are outlined as as problems in outlining functional links. follows: Forest decline occurs, but limited knowledge exists • Since processes involve the interaction of causes with regard to and consequences, activities such as filtering de- • The rate of decline, the spatial pattern and the terminant causes, evaluating the extent of impact, trend over an extended temporal perspective. as well as analyzing more complex causative chains are rendered problematical. • How long it has been in progress, its causes and consequences over time. • Interpretation is often context dependent, i.e. varying interpretations may result from the • Attributable causes in a short temporal per- application of differing human or environmental spective, specifically the impact of smallholder perspectives. This poses the problem of reducing agriculture. complex information to simplified knowledge. • Ecological and human spatial boundaries which In Ethiopia forest decline is identified as an important enclose and divide the forest resources, as well as problem (EFAP 1994). However, relevant data are their role in protecting and managing the forest scarce, the existing knowledge is incomplete and resources. interpretation is largely influenced by the prevailing This study addresses forest decline on two premises 1) environmental crises in the country (Hoben 1995; forest decline involves people-forest interaction and McCann 1999; Dessalegn 2001; Melaku 2003). The 2) a more complete understanding involves process existing few studies tend to emphasize the rate of identification, and an analysis of landscape history

Gessesse Dessie and management. Each is an entity where method and The floor of the depression is covered by lakes, wet- reliability of data can be evaluated and, furthermore, lands and alluvial plains, which together cover half when combined they can provide a more holistic the watershed. The remaining half consists of uplands perspective. and escarpments with slopes varying between 6 and Such an approach requires us to: 30 degrees. Awassa watershed is a closed hydrological • Integrate multiple underlying and proximate basin supplied by perennial streams emerging from causes at different scales and levels. the eastern escarpments. The rainfall is transitional between a bimodal • Combine qualitative and quantitative data, as pattern in the south and a single rainy season in the well as spatial and non-spatial data. north. Rain can be expected in March to April and • Spatially reference the fluid social variables to the June to August. The remaining part of the year is dry fixed forest and biophysical variables. with December and January being the driest months. • Investigate causative integration over time and Rainfall amounts differ between the drier western examine functional links at multiple layers and escarpments, which receive 900 mm a year, and the scales. wetter eastern escarpments, which receive 1200 mm. The mean annual temperature in the foothills varies • Consider different data formats and protocols, as from 17 to 19 degree centigrade. well as frequently scattered, scarce, and dispersed The soils are young and of volcanic origin. The data sources. valley floor is partly covered by lake deposits well The study area is a part of the south central Rift Val- supplied with plant nutrients. On higher ground the ley of Ethiopia and embraces a considerable diversity texture is sandy or silty while clay dominates around in topography, climate, vegetation types, and soils. the wetlands. The western section of the watershed Historically, the area has been influenced by the has acidic, coarse, shallow soils (Makin et al 1975). location of trade routes and in more recent time, the The vegetation zonation is influenced by altitude, emergence of highway nodes and population centers, rainfall and soil fertility. In the wet, eastern part of which have influenced land use patterns. The area the watershed, montane forest composed of Podo- has relatively recently (principally during the past carpus falcatus and Juniperus procera is dominant century) been settled by humans and there have been (Russ 1944). In the sub-humid western part shrubs several waves of immigration of different origin and and thickets dominate (Chernet 1982), while on the culture. As a result, significant changes have occurred lower foot-slopes and on the valley floor grassland, with respect to sociopolitical conditions, land tenure bush land and acacia woodland prevail. The forest regimes, agricultural expansion, population density is the partly fragmented remnant of a formerly more and road network expansion. extensive and coherent forest covering the eastern The central question of this thesis is: What has rift flank. It harbors important and rare fauna and happened to the forests of the south central Rift Valley flora, and is important for the watershed and the of Ethiopia during the past 100 years? The specific ecosystem, as well as for economic, research and questions that are addressed are: educational purposes. • What is the pattern and magnitude of recent forest The population of the study area is composed of decline in the area? six main ethnic groups together numbering about • What is the history of forest decline over a one half a million. About one third of the people live in hundred year perspective? Awassa and Shashemene towns. The recent rural and town populations are growing at rates of 2.1% and • What i s the significance of the recent increase in 4.1% respectively (CSA 1996). smallholders’ cash crop (khat) production? The current land use is predominantly smallholder • What are the roles of human and ecological agriculture with an average landholding size of less boundaries in forest decline? than one hectare per household. The major crops include enset, khat, sugarcane, maize and potatoes. Wondo Genet is agriculturally fertile, with irrigation farming dominating in the flat and undulating sites. The study area The Awassa watershed and its nearest surroundings in the south central Rift Valley was selected as the Methods study area (Fig. 1). It comprises 3060 km2 and lies at 6o 45’ N to 7o 15’ N latitude and 38o 15’E to 38o 45’ In this study a combination of different methods and E longitude. The area is located some 280 km south techniques were employed. Spatial and non-spatial of the capital Addis Ababa. Two sub-study areas, variables were measured. Satellite images from three Wondo Genet and Wosha, are located in the eastern points in time were interpreted and extensive field part of the Awassa watershed. checks were carried out. Social surveys (structured The large-scale physiographic setting is defined by interviews, PRA- Participatory Rural Appraisal tech- a tectonic depression bounded by steep escarpments. niques), historiography (travelers’ accounts, historical

Forest decline in South Central Ethiopia

Shashemene

Mt. Abaro Watershed boundary 2580 Lake Cheleleka Lake Wondo Genet Awassa Wosha PA College of Forestry Awassa Wetland 1675 Basha

� � � � ��

10� N Addis Ababa Ethiopia Study area

� � �� 40 � E

Figure 1. The study area, Awassa watershed (broken line) and its surroundings. Paper I and II cover the entire area. Paper III and IV focus on the eastern part, Wondo Genet area (the box), and the core area of the remaining forest (the oval).

records, documents) and an analysis of the physical Multi-Spectral-Scanner (MSS) in 1972 and Thema- variables (field observation) were the main methods tic-Mapper (TM) in 2000 and 2003, respectively. used. Universal Transverse Mercator (UTM) geographic The data sources and methods are described in the projection was used together with digitally converted separate papers while the timeline of the methods and and rectified topographical maps. A supervised maxi- their contribution to the different papers are shown mum likelihood classification method was employed in Fig 2. for the classification of the satellite images for both the forest and khat cases. Stereoscopic and digital interpretation of aerial photographs was used for Remote sensing identification of training areas in the MSS and TM Satellite images are capable of providing a synoptic data. Errors were reduced by extensive field checks, view of an area as it appears at a particular point repeated iterations of classification and the selection in time. In this way image interpretation shows the of appropriate resolution and seasons. conditions of the object at that time. Change detection techniques identify differences between two or more Social surveys points in time. Objects that may appear to remain constant for extended time periods and stand out in Social survey methods are a means of gathering infor- the landscape, such as forests, are relatively easy to in- mation using questionnaires or interviews. This thesis terpret, whereas more dynamic or transient objects are employed individual interviews, group discussions, more problematical. In view of this basic understan- public meetings and an archive survey. Quantitative ding, the remote sensing techniques were employed and qualitative data were co-analysed in order to al- to interpret and detect changes of the natural forest low identification and ranking of causative factors. cover and spatial distribution of khat farms. The purpose was to establish a historic time line and Images were obtained from two different sensors: identify important individual events. For the more

Gessesse Dessie

2ESEARCHCOMPONENTS -ETHODSTECHNIQUES 2ESEARCHPERIOD 0APER) 0APER)) 0APER))) 0APER)6 (ISTORIOGRAPHY

4RAVELLERSACCOUNTS 8

/RALSTORIES 8 8 4EXTSDOCUMENTS 8 8 3OCIALSURVEYS

3TRUCTUREDINTERVIEWS 8 )N DEPTHINTERVIEWS 8 8 02! 8 8 2EMOTESENSING

3ATELLITEIMAGES 8 8 8 !ERIALPHOTOGRAPHS 8 8

4OPOGRAPHICALMAPS 8 8 8

Figure 2. Methods and techniques employed in the research to analyze the long-term trend of forest decline. No single method has provided adequate information to span the entire temporal period or the multiple components of the research. Some of the methods are temporally specified (e.g. travelers accounts, satellite images), others are dependent on respondents’ memory and knowledge (e.g. social surveys) while others again can be verified when they cover recent periods (e.g. in depth interviews). However, although the individual methods have limitations they supplement each other to create an overriding understanding of the various components over the different time periods. recent decades, the interviews explored land users’ Field observations of the cultural landscape and decision making and resource sharing patterns, biophysical conditions were used to outline trends and particularly with reference to khat production and supplement the data acquired by the remote sensing illegal logging. and social survey methods. Randomly selected household interviews were employed to appraise the trend of khat expansion. PRA techniques including group discussions, key informant interviews, ranking, sequencing, listing and transect Presentation of the papers walks were used to secure relevant information from Paper I is an estimation of forest decline covering the peasant associations (PA) leaders, farmers (men and period between 1972 and 2000. Paper II explores women), development agents and elders. Attendance forest decline and human impact over a one hundred of public meetings assisted in gaining an understan- year perspective. Paper III discusses the process of ding of the forest decline issues revealed by local forest decline, specifically analyzing the recent people- people. Archival records from government offices, PA forest interaction resulting from the expansion of a offices, private people and the Wondo Genet College drug cash crop: khat. Paper IV presents a conceptual of Forestry were used to obtain information regarding discussion of boundaries and describes the effect of population, lists (and number) of households, land multiple boundaries on the conditions of forest decline holdings, market conditions and prices in the area.

Historiography

Narratives and topographic sketches in travelers’ Paper I: accounts were used to obtain information related Gessesse Dessie and Kleman, J. 2007. Pattern and to the state of the forests, human activities and land Magnitude of Deforestation in the South Central cover types of the past. Individual and group memory Rift Valley Region of Ethiopia. Mountain research surveys were employed in interviewing elders and and development 27(2):162-168. through group discussions. Historical texts, documents, research articles and academic theses were The main focus of this paper is the estimation of also utilized. the pattern and magnitude as well as possible consequences of deforestation occurring between the years 1972 and 2000, using remote sensing change Peasant association (PA) also called kebele is a grass-root level administration unit in Ethiopia with a defined spatial limit detection techniques. of approximately 800 hectare Three major issues are highlighted:

Forest decline in South Central Ethiopia

1) The areal decline and the resulting spatial pat- 2) The major causes and the mode of change tern The forest decline in the study area involves proximate The total natural forest loss between 1972 and 2000 causal factors, and causal factors that are more spa- amounts to over 40,000 ha, which is over 80 % of tially diffuse stemming from the long-term evolution the forest cover that was present in 1972. This cor- of a region much larger than the study area. Over the responds to an annual loss of over 1400 ha, equivalent course of the past half-century, South-Central Ethio- to 0.9% of the annual national loss. The decreasing pia has witnessed rapid growth of the population, trend of forest coverage in the study area coincides the number and size of settlements have increased, with the general forest decline pattern in Ethiopia. The the road network has been improved and economic forest decline during the period spanning from 1972 sectors have been developed. to 2000 (Fig 3) is not an isolated event, but rather a Expansion of agriculture, and in particular small- continuation of the past trend. holder farming contributes to over 80% of the forest In terms of spatial forest patterns, the decline area loss, characterized by two major modes of change: has substantially thinned the stands and distorted 1) internal: clearings created by the intrusion of small the linear spatial pattern naturally created by the farm plots, grazing lands and villages 2) external: ex- drainage network of fluvial valleys. Pockets of pansion of agriculture from the exterior into the forests. non-linear fragments of forest remain in the valleys and on steep slopes. These isolated patches are key habitats and harbor rare and threatened animal and 3) Selected consequences of forest decline plant species. Within the formerly closed natural forest, clearings Forest change results in habitat destruction and th- for farming and settlement create an initial pattern of reatens rare tree species such as Aningeria adolfi-frie- non-connected patches. In areas where forest clearing derici, Podocarpus falcatus and Prunus africanus with has been more complete, the result is a speckled pat- local extinction. Habitat destruction has furthermore tern of non-connected small forest patches remaining affected some endemic fauna e.g. the local extinction after the encroachment of clusters of farm plots, of the antelope species Mountain Nyala (Tragelaphus grazing units and settlements. buxtoni), one of the mammals endemic to Ethiopia.

Figure 3. Interpreted satellite images from 1972 (a) and 2000 (b) showing forest cover, black color, in the study area during those periods. The year 2000 image shows significant decrease of forest cover and the drying out (disappearance) of a small lake, grey color, in comparison to year 1972.

Gessesse Dessie

Another possible consequence is a decline in forests/ woodlands, covered approximately 40% of the availability of water. The natural location and the study area. distribution of the forest in the upstream regions Forest decline in the study area is interpreted as of catchments render these forests hydrologically a human–environment interaction in time and space important. During the study period, Lake Cheleleka (Fig 4). It is an integral part of the evolution of the has dried out. It appears probable that the drying out cultural (human made) landscape. Forest clearance is caused by a reduction in stream flow partly related was intensive during periods of large-scale immigra- to deforestation, and by the increased use of stream tion, increased commercial logging, and expansion water for irrigation purposes. of coffee farming activities as well as during the Forest loss has negatively affected practical fo- establishment of settlement schemes. restry training at Wondo Genet College of Forestry. The pattern of change displayed by forest decline The college is heavily reliant on the small area of is characterized by three types of phases: episodic, remaining natural forest for its outdoor exercises gradual and recurring. One of the most important and experiments. examples of an episodic phase is the catastrophic Major conclusions of the paper: famine in Ethiopia of the late 19th century, which • Significant and rapid forest decline has occurred resulted in abrupt environmental and social changes. in the area. Another significant example is the southern incorporation, which occurred shortly after the famine. • The remaining forest is highly fragmented. The southern incorporation resulted in a massive • Forest decline has resulted in habitat destruction social and environmental transformation in southern and threatens rare tree species with extinction. Ethiopia. Gradual change results from population • It also contributes to a decline in the availability growth, agricultural expansion and the increase in of water. number of settlements. Finally, an example of recurring impulses is the repeated sociopolitical changes, where each situation presents different socioeconomic conditions, administrative policies and principles of Paper II: land ownership. Gessesse Dessie and Christiansson, C. 2007. Forest Rainfall distribution, soil conditions, and wetlands Decline and its Causes in the South Central Rift and lakes shaped the distribution, size and continuity Valley of Ethiopia. Human Impact over a One of the forest/woodland in the study area. The earlier Hundred Year Perspective. Ambio (accepted). satellite image (from 1972) combined with the observable distribution of remnant trees over the landscape The main focus of this paper is to investigate, from a under study, and travelers’ accounts would seem to geographical perspective, linkages between the trend suggest the distribution of denser woody vegetation of forest decline and changes in the social, economic at the beginning of the 1900s. and political pattern in the Awassa watershed and its Geographical properties, for example topography, surroundings in south central Ethiopia over a 100- soil fertility and the relative location of the area stron- year perspective. Methods used are satellite images gly influence the process of change. In the actual case and map analysis, interviews and literature studies, the local land form, i.e. steep escarpments overlooking and travelers’ accounts and field observations. wide plains on the valley floor, in combination with Forest decline is analyzed by linking certain events the specific location between the South, rich in resour- temporally, spatially and functionally to the study ces and the political core of the country to the North, area as follows: constitute important reasons for the state, at an early • Changes in population, land tenure, agriculture stage, to strengthen its political control through esta- and transport capacity (hereafter called "causes"), blishing defense posts. This was an area suitable for of relevance for forest decline, were identified. settlement and economic development. Consequently, the strong interests of the state dictated the evolution • Each cause is subdivided into events important of the cultural landscape, which embraced the clearing for the study area, and the period of occurrence of forests for agricultural settlement. of each respective event is indicated on a timeline After the years of famine spanning the 1880s and spanning one century. 1890s, which had wiped out much of the human • The events are spatially represented on a sche- and animal life, human activities gradually started matic map, which gives an initial picture of the to escalate. Subsequently, following the southern space and time embraced by a specific event. incorporation migrants from the north settled in the • The process of change is outlined as a functional areas best suited for arable agriculture, which areas chain between events themselves, and between were savanna woodlands and the forest fringes. The events and forest decline. wetlands while providing important dry season grazing areas for the pastoralists constituted significant Analysis of early written accounts, interviews, distribution of remnant trees and other observations in the field show that, at the end of the19th century, natural The conquest and occupation of what is now southern Ethio- pia by northern Ethiopian armies

Forest decline in South Central Ethiopia

3HASHEMENE

7ONDO'ENET

2OAD

Figure 4. Graphical representation of important causes of landscape change during the period under study. breading ground for malaria. The wetland environ- were consequently cleared both for settlements and ment is furthermore unsuitable to the growth of to provide arable fields for inhabitants’ subsistence woody vegetation. farming. Improvement in the 1930s to motor vehicle stan- The main conclusions are as follows. dard of the old trade routes and subsequent upgra- • Decline of forests was already in progress in the ding in the 1960s to tarmac road with feeder roads early 20th century. from agricultural centers shortened travel time to • At the beginning of the last century forests were markets and increased load capacity. This motivated assigned low economic classification. an increase in cultivation of profitable, perishable or bulky crops. Clearing of remaining forests became a • Large areas of forest were cut down during poli- strong economic incentive. tical transition periods that resulted in a political Change in regime has implied far-reaching changes vacuum, in which the protection of resources was in the land tenure system. With new actors in power weak or lacking. differing political principles have been enforced, re- • Long-term planning efforts to manage forests sulting in changes in emphasis of land and resource were obstructed due to the uncertainty that resul- ownership. During privatization commercial farms, ted from changes in land tenure principle during logging enterprises, sawmills etc. were established each political period. in the absence of any long-term sustainability plans, • The little remaining forest land is becoming in- which inevitably led to forest decline. At the same creasingly attractive as potential land for arable time the establishment of these enterprises created agriculture as the successive upgrading of roads jobs and led to improvement in the infrastructure. improves accessibility between the study area and This in turn led to sporadic immigration into the study distant markets. area and increased population density. Forested areas

Gessesse Dessie

Paper III: distributing the need for labor more evenly. As a Gessesse Dessie and Kinlund, P. 2007. Khat Expan- result there is no bottleneck at a specific period sion and Forest Decline in Wondo Genet, Ethiopia. of time. Geografiska Annaler (submitted). • The plant grows under broad climatic conditions and can tolerate drought for long periods. The main focus of this paper is to investigate the forest decline process by analysing the people-environment Due to the aforementioned factors, khat is well interaction created by khat production. Khat, a stimu- adapted to the conditions experienced by the small- lant narcotic plant, is a high value agricultural crop holder farmers, who face land, labor and market in Ethiopia, which has recently become a cash crop limitations. Farmers are choosing to disregard the in the Wondo Genet area (Fig 5). social and financial costs posed by policy, cultural Remote sensing techniques are employed to survey and religious restrictions, in favor of khat production. the spatial distribution of khat farms while social sur- Consequently, successively more farmers are attracted vey methods are employed to reveal both the human to khat farming and the demand for more land for the activities driving the expansion of this crop, and also cultivation of khat has increased. Within individual the impact of its expansion on people’s livelihood and farm plots more land is allocated for khat than any forest cover. other crop while native trees with wide crowns are The findings reveal that: removed to avoid shading and to increase effective • The extent of khat production has increased since utilization of the space. All these factors contribute the late 1980s. The number of khat farmers has to a reduction in the number of individual trees and grown significantly, and more land is allocated to forest decline. within individual farms to khat. The findings also show that the economic benefits of khat production have attracted more people to the • Khat fetches larger incomes than any other crops area, with the result that the population density has per area unit. increased and the settlement pattern has changed. • The market demand for khat has increased con- While khat production contributes to a more disper- sistently. sed settlement pattern it also plays a part in shaping • Favourable road and transport facilities are avai- concentrated market settlements. This aspect, besi- lable to enable efficient transport of the perishable des influencing the direct conversion of forest land, leaves to distant markets. increases human activities proximate to forests e.g. grazing which can negatively affect the resilience of • The actors in the khat production process are the forest. strongly linked with one another and to the mar- Furthermore, Khat expansion partly contributes to ket by an informal but cohesive trade network. a decline in food production. The number of farmers • The leaves can be harvested throughout the year who mainly grow food crops is decreasing while the

Figure 5. Photograph showing khat for sale (wrapped in the bundles and on the weighing scale) in a kiosk in Mojo town 180 km from Wond Genet. This illustrates the marketability of Wondo Genet khat in distant markets and furthermore that khat is openly traded in a kiosk alongside other normal grocery items.

Forest decline in South Central Ethiopia number of khat farmers (those who allocate over 60% context, are exemplified by conditions in the remain- of their land to khat) is increasing. An interesting ex- ing natural forest of Wondo Genet Ethiopia. ception is in the case of enset production where the Conceptually, boundary refers to a wide range of number of growers is increasing. This increase may real and conceptual forms that divide larger hetero- however be attributed to the fact that farmers are geneous land units into smaller parts with a higher growing this particularly reliable staple food crop in degree of internal homogeneity. Boundaries are the their backyard as a strategy to cope with the general spatial manifestation of single or multiple natural or decrease in food production. human-defined properties. Ecological and human spa- Increased khat production has also led to smaller tial boundaries enclose and divide the forest resource. areas being allocated to fruit trees, garden crops and Human spatial boundaries control key elements in grazing for dairy production, activities which are forest management, such as accountability, claim and fundamental to family subsistence and traditionally control. There may exist a spatial mismatch between controlled by women. This development, together ecological boundaries and social spatial boundaries, with the fact that the substantial cash income gene- with potential negative consequences for natural rated by khat is controlled by men, and often spent resource management. outside the “family domain”, has led to a power shift The possible outcomes of boundary mismatch from women to men in the area. Food prices in the are area have increased, probably as a result of declining • Conflict of principles between ecosystem level production. These two conditions influence the wo- management and resource management decen- men and can lead to a tendency to rely on the forest tralization. and trees to extract additional income. • Multiple accountability and claims over an The expansion occurs both through encroachment ecosystem (the natural forest) by the various within the forests and through pushing back forest/ actors. agricultural frontiers (see front cover photograph). Main conclusions • Difficulties in integrating the various actors for • A significant amount of khat farms have replaced coherent management. forest in the study area and have promoted forest Mapping shows that ecological and social boundaries fragmentation. However, the crop occupies a rarely match in Wondo Genet (Fig 6). An example relatively small area and involves few farmers in of mismatch is presented using the Forestry College comparison to the total area and population of and the kebele boundaries (administrative bounda- the Wondo Genet area. • Khat is important to farmers whose land holdings are too small to yield sufficient income or to produce adequate amount of food. Khat precipitated forest decline may be regarded as a result of conscious decisions made out of necessity. • Khat expansion results in the establishment of permanent farm settlements in close proximity to the forests. • Khat has increased the value of land previously deemed unsuitable for cultivation, often covered by forests. . • Khat production causes the removal of trees growing on farms. KM

Paper IV: 7ATERSHEDBOUNDARY Gessesse Dessie and Kleman, J. 2006. Ecological and 5PSTREAMDOWNSTREAMBOUNDARY Human Spatial Boundaries and their Impact on 2EGIONALSTATEANDETHNICBOUNDARY Forest Management and Forest Decline in Wondo +EBELEPEASANTASSOCIATIONBOUNDARIES Genet, Ethiopia. (Manuscript). &ORESTRYPROJECTBOUNDARIES

The main focus of the paper is to investigate how the .ATURALFORESTCOREAREA complex processes driving forest decline are affected 7OSHAKEBELE by the structure and changes of human spatial and natural, ecologically defined, forest boundaries. The 7ONDO'ENET#OLLEGEOF&ORESTRY concepts of boundaries, and their importance in this Figure 6. Natural boundaries and human spatial boundaries A food crop (explained in paper III) in Wondo Genet.

Gessesse Dessie ries of peasant associations). The ecological system preferable means of ensuring local peoples’ rights and delineated by the college is divided between multi- protecting the forests, it can lead to a disintegration of ple kebeles and manifested by their administrative authority that can weaken accountability if not imple- boundaries. The two boundary systems mismatch mented properly. This indicates that decentralization in space, with the ecological boundaries defined by may not always bring about the desired solution in the college cutting through regional state and kebele natural forest management. boundaries. The mismatch causes problems for both Main conclusions: the kebeles and the college. In some cases the juris- • A complex structure of ecological and human diction of the kebeles is intruded, while in other cases spatial boundaries exist in Wondo Genet. the management interest and objectives of the college • Multiple human spatial (in particular administra- are violated. tive) boundaries overlay and crosscut an ecologi- The polycentric organization of administration cally defined forest structure in Wondo Genet. overlays and crosscuts an ecologically defined forest structure in Wondo Genet. This contributes to frag- • The multiple administrative boundaries contri- mented jurisdiction, ownership and management, bute to multiple claims over forest resources and which has led to weak accountability and adminis- weak accountability to protect the forests. trative inefficiency. There are no common forest ma- • The illegal activities are partly related to unresol- nagement objectives. The forest protection measures ved, overlapping and multiple boundary issues. are sporadic and inconsistent. Complications occur in enforcing legal and administrative measures. Il- legal loggers and illegal settlers exploit these weak links and are able to easily stifle the elected kebele leaders’ authority. Discussion The current human spatial boundaries are supe- This study presents, over an extended temporal per- rimposed on the landscape where different kinds of spective, the geography, history and process of forest boundaries have been recreated from past human decline in south central Ethiopia. The approach is activities. The creation and re-creation of bounda- holistic. By combining many different data sources ries have divided the society according to power, it is attempted to quantitatively and qualitatively de- between cultural/ethnic groups, between immigrants scribe the ongoing forest decline, and to identify the and indigenous people, between the rulers and the main drivers behind this evolution, thereby providing people and between individuals. In particular, periods a more complete picture of the forest decline process. of time when boundary changes have occurred are The approach differs from those studies which employ characterized by a “power vacuum” and a lack of quantitative analysis of limited causes over limited accountability. In those instances, forest decline is ac- time periods (e.g. Angelsen 1999) and those which celerated as a result of illegal actors exploiting a state employ concrete spatial representation of socio-eco- of uncertainty and lack of control. Thus, forest decline nomic variables (e.g. Walsh and Crews-Meyer 2002). is ultimately related to boundary issues and in parti- The holistic approach has the advantage that it is cular, to boundary changes. A functional chain can possible to capture and describe very diverse driving be defined by a number of links: boundary mismatch forces, in both the human and ecological domains, but and boundary changes  “power vacuum” and lack also has inherent limitations because of difficulties in of accountability  intrusion or illegal activities  combining data of different nature and data derived illegal forest cutting  forest decline. with different methods. With respect to the ongoing forest decline in the The following discussion focuses on the four key area, the illegal activities are at least partly related questions raised in the introduction. to unresolved, overlapping and multiple boundary issues. Possible indicators of this are; 1) continuing • What is the pattern and magnitude of recent forest violation despite strong government presence (poli- decline in the area? cing, military intervention, prosecution of violators, • What is the history of forest decline over a one mediation) and awareness of the general public 2) hundred year perspective? failure to solve causes of violation despite the decen- • What is the significance of the recent increase in tralized administration by kebeles intended to reach smallholders’ cash crop (khat) production? the grass root level. To control the illegal destruction various measures have been taken, including policing • What is the role of human and ecological boun- and guarding, consultation and mediation, creation daries in forest decline? of public awareness through public meetings and In terms of pattern and magnitude of forest decline, even military intervention. However, no significant the findings reveal forest decline to be a continuous progress has been registered. process, associated with spatial fragmentation and The existing decentralized administration in location specific losses. The remote sensing-based Wondo Genet, despite supposedly wielding power change detection result has shown substantial and and authority at a local level, seems unable to control rapid forest loss during the recent three decades, forest decline. Although decentralization is arguably a although tree plantations are increasing. A historic

10 Forest decline in South Central Ethiopia data analysis revealed that natural forest decline in the the large landowners. Such changes are ultimately area has been in progress for at least a century. This caused by socio-political changes. These observations aspect agrees with other works from Ethiopia (Russ confirm that observable contemporary events are the 1944; Mooney 1954; EFAP 1994; Reusing 1998). result of layers of historic conditions (Antrop 2005; Exceptions are McCann (1999) who has reported Marcucci 2000) and extraordinary events of wide forest gain in an area in southwestern Ethiopia and reaching social and natural impact can be oversha- Crummey (1998) who has reported an increase in dowed by recent events. tree cover in an otherwise highly degraded area in the This study shows that at an extended temporal north. However, in the study area the gain in forested scale, causes can have different impact and meaning, area through plantations is less than the magnitude individually and interactively. Major actors vary over of natural forest decline. time and the interaction between them increases in The spatial pattern of the remaining forest is sha- intensity and complexity. The fundamental social ped by the expansion of smallholders’ permanent cul- processes and local level immediate activities (Geist tivation and grazing. The small patch size, the lack of and Lambin 2002) are linked differently at different connectivity and the permanent agricultural activities time periods. probably undermine the resilience/regeneration poten- In terms of the driving processes behind forest tial of the forest. It appears that smallholders’ land decline, one particular recent trend, increased khat use contributes to forest decline. Location specific production has impacted the forest through several losses such as upstream forests of valuable watershed mechanisms; it relocates the agricultural/forest fron- services and key habitat areas are here concurrent tier, it causes intrusion and permanent settlement with other such losses in the country. within forests, and fragments the remaining forest. In terms of history, the analysis reveals that in the These changes in structure and use of the forest are one hundred years perspective, changes in population, however not always associated with large areal chan- land tenure, agriculture and road network have had ges. The highly fragmented pattern of khat fields make a strong impact on forest decline. The process can be accurate estimates of areal changes difficult. considered a human – environment interaction in time The contribution of khat to forest decline se- and space within which forest decline is an integral ems self-evident considering the magnitude of land part of the evolution of the cultural landscape. The transformation and the number of farmers involved. processes are related to the natural characteristics However, khat derives multiple causes due to the of the area and the socio-political and economic versatility of the crop, which grows under wide envi- processes which form the cultural landscape. Forest ronmental conditions and provides high comparative decline in the study area is interpreted as the combined advantages for smallholder farmers. This gives incen- result of socio-political changes, economic activities, tives to khat farmers to expand into forests where population growth, cultural patterns and agricultural other crops do not grow or where it is unfeasible to developments. grow them due to economic and protection reasons. The introduction of different land management Khat’s economic advantage attracts more smallholder practices, conflicts over resources and power struggle farmers and increases the value of farmland. The re- due to fluxes of immigration have occurred during liable market, favorable infrastructure and improved the last century. Changes in land tenure resulted road conditions further facilitate its expansion. in inconsistencies in ownership and in access to The khat related forest decline process was hence resources. Agricultural development has also been identified as the result of the combined effect of the influenced by land tenure changes and furthermore inherent properties of the particular crop, and socio- by road network development. The contribution of economic factors. Moreover, the khat case indicates population, land tenure, agriculture and improvement that the process of people-forest interaction is influ- of roads to environmental change in general, and enced by demographic factors such as population forest decline in particular, is documented by many size, economic conditions, the road network, and studies (e.g. Angelsen 1999; Contreras-Hermosilla market factors. This agrees with other studies that 2000; Deacon 1999) state that forest decline is not the result of a few Within an extended temporal perspective human causative factors related to population and poverty impact and forest decline vary in space and over alone (Angelsen and Kaimowitz 1999; Lambin et al time. The contributions of underlying causes, e.g. 2001; Moran and Ostrom 2005). socio-political developments, were important over a It is important to underline that while the ex- longer time perspective. Causes more specific in time pansion of cash crops such as khat is economically and space resulted in rapid transformations, e.g. the stimulated, reasons for protecting the natural forest emergence of cash crop production. The impact of are often an environmental issue. Unless the envi- some causes was more localized, e.g. commercial ronmental benefit is also economically important logging. Other causes interconnect several events in a time perspective that is realistic to farmers, it is and exert wide reaching impact e.g. the commercial unlikely that smallholder land users will perceive the farms. Importantly, the dominant actors/agents vary rationale in maintaining the natural forests. Where over time. The importance of smallholder farmers has farmers’ livelihood, and sometimes their very sur- increased in relation to a previously influential group, vival, is at stake, the objective of biodiversity and

11 Gessesse Dessie environment might not be incentive enough. Under gain, market conditions and transport facilities. such circumstances forest decline is imminent. Multiple claims to land and weak accountability In terms of boundary structure, in the study area contribute to inefficient management, which leads to as practically everywhere else, ecological and social forest decline. Boundaries need to establish a compro- spatial boundaries do not coincide, with negative con- mise between the sustenance of ecological elements sequences for accountability and management. The and the equitable share of resources. analysis of social-spatial boundaries and ecological boundaries indicate the unsystematic management of the natural forests by several administrative units. These aspects seem to influence the process of forest Research for the future decline due to non-integrated management, unresolved claims and weak accountability. Despite the fact Forest decline research often investigates observable that forest management was decentralized to localities recent conditions, while causes evolve over long time administered by elected officials, there appears to be periods. Some causes are sporadic, and in particular no concerted integration between them. Moreover, the causes that impact the power relations of actors tend changes of the social boundaries that occurred in the to be forgotten. Research that considers such issues is past seem to have resulted in unresolved claims crea- required to understand forest decline in a long-term ting situations where the forests are exposed to mul- perspective. tiple claims. Under such conditions where ownership Pragmatic and rigorous methods are needed to is diffuse, it is easy for violators to exploit the forest employ holistic forest decline studies linking the hu- for illegal activities. Such activities involve illegal log- man and biophysical aspects. Research is required ging, illegal settlement and other destructive activities. to develop such methods. The spatial integrity and the temporal continuity Boundaries play an important role in forest proof the forest that ensure its future are thus affected tection and management. Comprehensive research detrimentally. This condition supports conclusions is needed to link boundary issues with land tenure, that decentralized forest management by itself might access, claims on forests and accountability for ma- not guarantee efficient management (Larson 2003). nagement. This bears strong implications in Ethiopia where the Forest decline is both a quantifiable physical pro- few remaining natural forests are decentralized and cess and a contextual judgment at a given space and divided by boundaries of regions, ethnic groups and time. Consequently it can have different meaning for local communities (Melaku 2003). different stakeholders. Research is essential to identify This study reveals that consistent forest decline the common understanding of different stakeholders during the past century coincided with the steady regarding forest decline. ascendance of human activities. Processes involve Planting fast growing trees may solve some of the the physical removal and fragmentation of the forest, needs for forest products, but this can not replace na- weakening of its resilience and inconsistency in its tural forests’ intrinsic values: ecosystem, biodiversity, management. Forest decline, in general, seems to be and watershed services. These services are embedded the inevitable outcome of development activities and in the vegetation structure, variability of fauna and the destructive effect of various forms of resource flora and temporal continuity. Research is needed to conflict. find out to what extent plantations of fast growing trees can replace some of the services originally provided by the natural forests, and also relieve remaining natural forest of some of the pressure for timber. Conclusions The following are the main conclusions of the thesis: Acknowledgements Forest decline in the area has a long history, spanning at least one century. The recent rate of decline In doing this study I have received financial grants, is significant in comparison with the national rate of professional assistances, and friendly help from many forest loss. The causes of the forest decline in the area individuals and institutions. I sincerely thank you all. emerge from, on the one hand, layers of temporally You have contributed to an exiting period of learning. spaced separate events, and on the other, chains of The merits are based on your ideas and help. The rest events. These events interact with each other at dif- is fully my responsibility. ferent levels and scales as well as with the geograp- Financial support was provided primarily by Sida hical properties of the study area. Over an extended (Swedish International Development Cooperation temporal perspective forest decline has coincided with Agency), but also from IFS (International Foundation major sociopolitical regime transformations. for Science), the SSAG (The Swedish Society for An- Land users’ rationale in choosing the comparative thropology and Geography) and by Carl Mannerfelt advantage between keeping and replacing the forest is and Hans Ahlmann research foundations. Thank important. The choice made is affected by economic you all!

12 Forest decline in South Central Ethiopia

To my supervisors Professor Carl Christiansson, It was warm and shiny summer with you when the Professor Johan Kleman and Dr. Peter Kinlund: Thank Swedish winter itched my bones and darkened my you very much for your guidance, continuous con- spirit. Jenny Cadestedt thank you for your thought- structive and valuable support, particularly for your fulness and generosity. tireless engagement to structure the arduous manu- Shifferaw Dessie, Kelem Asfaw, Yiftusera Deribe, scripts. In addition: Carl thanks for the consistent Jambola Yilma and Birara Mengistu you are all great. encouragement at professional and personal level. I am eternally indebted to your affections. THANK Johan, thanks for those stimulating discussions, your YOU! generosity and troubleshooting when I stumble. Peter thanks for organizing literature courses that suited this study, helping to acquire relevant literature and References for being a good friend. There are very many people in Ethiopia, who in Angelsen, A. 1999. Agricultural expansion and deforesta- one way or another contributed to this study. My first tion. Modelling the impact of population, market forces appreciation shall go to the villagers particularly of and property rights. Journal of Development Economics Wosha Kebele who provided this study with essential 58: 185-218. information. Then to informants and research assis- Angelsen, A. and Kaimowitz, D. 1999. Rethinking the tants: Yonas Belachew, Abraham Hamza, Admassu causes of deforestation: Lessons from economic models. Mekonnen, Zerfu Sirmessa, Tadele Sebsibe, Waritu The World Bank Research Observer 14 (1): 73-98 Fareso, Esatu Enja, Asfaw Gidessa, Yirdachew Bekele, Antrop, M. 2005. Why landscapes of the past are important Jalo Bekure, Roba Worasa, Hasso Jula, Hirpeto for the future. Landscape and urban planning 70 (1- Wosena, Yosef Gulma, Girma Shumi. Also Zerihun 2): 21-34 Mohammed, Girma Negash, Teshome Sisay thank Batterbury, S.P.J. and Bebbington, A.J. 1999. Environmental you for sharing your wealth of knowledge of the Histories, access to resources and landscape change: study area. an introduction. Land degradation and development Deans, staff and friends at Wondo Genet Col- 10: 279-299 lege of Forestry and Natural Resources: Dr. Abdu Cadenasso, M.L., Pickett, S.T.A., Weathers, K.C., Bell, Abdulkadir, Dr. Melaku Bekele, Dr. Zebene Asfaw, S.S., Benning, T.L., Carreiro, M.M. and Dawson, T.E. Dr. Tsegaye Bekele, Menfese Tadesse, Mersha Ge- 2003. An Interdisciplinary and Synthetic Approach to brehiwot, Kefyalew Sahele, Yimiru Tesfaye, Genene Ecological Boundaries. BioSciences 53 (8): 717-722 Asefa, Fantaw Yimer, Megersa Debele, Endale Feyesa, Chernet, T. 1982. Hydrology of the Lakes Region, Ethiopia Ashenafi Talew, Afework Eguale. I thank you for all (Lakes Ziway, Langano, Abiyata, Shalla and Awassa). the support. Ethiopian Institute of Geological Survey. Addis Ababa, Members of the Department of Physical Geo- 97p graphy and Quaternary Geology, and Department of Human Geography of the Stockholm University: Contreras-Hermosilla, A. 2000. The underlying causes of forest decline. Occasional Paper No. 30. Centre for Thank you! Professor Karin Holmgren Head of the International Forestry Research (CIFOR). Bogor. Department of Physical Geography and Quaternary Geology: I am grateful for your consistent encourage- Crummey, D. 1998. Deforestation in Wällo: process or il- ment and also for allocating financial support when lusion? Journal of Ethiopian Studies xxi (1): 1-42 my scholarships run out. Professor Mats Widgren, CSA (Central Statistics Authority), 1996. The 1994 Po- Head of the Department of Human Geography, thank pulation and Housing Census of Ethiopia Results fro you for inviting me to join the PLATINA (People, Southern Nations, Nationalities and Peoples’ Region. Land and Time in Africa) research group. Joining Central Statistics Authority, Addis Ababa. this group of able and experienced researchers was an Cumming, G. S., Cumming, D. H., and Redman, C. L. 2006. immensely useful learning experience. Göran Alm for Scale Mismatches in Social-Ecological Systems: Causes, your expert support in Remote Sensing, Rolf Jacobson Consequences, and Solutions. Ecology and Society for your help with computer related issues, Carina 11(1): 14 [online] URL: http://www.ecologyandsociety. Henriksson for your support with administrative mat- org/vol11/iss/art14/ ters, Amanda Roberts for checking the language. . Deacon, R.T. 1999. Deforestation and ownership: evidence Professor Bo Dahlin University of Helsinki, thanks from historical accounts and contemporary data. Land for all your kindness and important advices at the economics 75(3):341-359 initial stage of this study. Mr. Johan Toborn, Mr. Sven- Dessalegn, R. 2001. The landscape of development: a Gunnar Larsson and Dr. Mats Sandewall thanks, your complex reality. In: Assefa, T. (editor). Food Security th- favorable decisions facilitated this study. . rough Sustainable Land Use. Population, Environment Professor Hans Hurni and the staff at Center for and Rural Development Issues for Sustainable Liveli- Development and Environment at the University of hood in Ethiopia. Proceedings of the Second National Berne, thanks for the hospitality and wholehearted Workshop of NOVIB partners Forum on Sustainable technical help while I was visiting your center. Land Use. 7-17 Friends in Sweden, Pia Barklund, Åke Barklund, EFAP (Ethiopian Forestry Action Program), 1994. Ethio- Junko Söderman and Torgny Söderman, thanks a lot. pian Forestry Action Program, Vol. 2. The Challenge for

13 Gessesse Dessie

Development. Ministry of Natural Resources Develop- Teferra, T. 1975. Development prospects in the southern ment and Environmental Protection, Addis Ababa. rift valley, Ethiopia (Vol.1&2). Land resource study Emich, M. 2003. The Human Ecology of Mayan Cacao 21. Land Resources Division, Ministry of Overseas Farming in Belize. Human Ecology 31(1): 111-131. Development: England Fox, J., Rindfuss, R. R., Walsh S. J. and Mishra, V. 2003. Marcucci, D. J. 2000. Landscape history as a planning tool. People and the Environment. Approaches for linking Landscape and urban planning 49:67-81 household and community surveys to remote sensing McCann, J.C. 1999. Green land, brown land, black land: and GIS. Kluwer Academic Publishers, Boston an environmental history of Africa, 1800-1990. James Gebissa, E. 2004. Leaf of Allah. James Currey, Oxford. Currey Ltd., Oxford. Geist, H.J. and Lambin, E.F., 2002. Proximate causes and Meidinger, E.E. 1998. Laws and Institutions in Cross- underlying driving forces of tropical deforestation. Boundary Stewardship. In: Knitigt, L.R. and Landres, Bioscience 52(2): 143-150. B.P. (editors). Stewardship across Boundaries, Island Press. 87-110 Govereh, J. and Jayne, T. S. 2003. Cash cropping and food crop productivity: synergies or trade-offs? Agricultural Melaku, B. 2003. Forest Property Rights, the Role of the Economics 28: 39-50. State, and Institutional Exigency: The Ethiopian Ex- perience. Doctoral Dissertation. Swedish University of Grossman, L. S. 1993. The Political Ecology of Banana Agricultural Sciences. Uppsala Sweden. Exports and Local Food-Production in St-Vincent, Eas- tern Caribbean. Annals of the Association of American Mooney, H.F. 1954. Report on forestry in Ethiopia with Geographers 83(2): 347-367. special references to the forests of Arussi and Sidamo. Imperial Forestry Institute. Oxford. Hoben, A.1995. Paradigms and politics: the cultural construction of environmental policy in Ethiopia. World Moran, E.F. and Ostrom, E. 2005. Seeing the forest and Development 23(6): 1007-1021 trees. Human-environment interactions in forest ecosystems. Cambridge: MIT Press. Keys, E. 2004. Commercial Agriculture as creative destruction or destructive creation a case study of Nestel, D. 1995. Coffee in Mexico: international market, Chili cultivation and plant-pest disease in the southern agricultural landscape and ecology. Ecological Econo- Yucatan Region. Land degradation and developmen mics 15: 165-178. 15:397-409. Reusing, M. 1998. Monitoring of natural high forest re- Lambin E.F., Turner B.L., Geist, H.G., Agbola, S.B., Angel- sources in Ethiopia. Addis Ababa, Ethiopia, Ministry sen, A., Bruce, J.W., Coomes, O.T., Dirzo, R., Fischer, of Agriculture. G., Folke, C., George, P.S., Homewood, K., Imbernon, Russ, G.W. 1944. Report on Ethiopian forests. Compiled J., Leemans, R., Li, X., Moran, E.F., Mortimore, M., in 1979 by Wolde Michael Kelecha. Addis Ababa, Ramakrishnan, P.S., Richards, J.F., Skånes, H., Steffen, Ethiopia: Ministry of Agriculture. W., Stone, G.D., Svedin, U., Veldkamp, T.A., Vogel, C. Roberts, N. 1996. The human transformation of the Earth’s and Xu, J. 2001. The causes of land-use and land-cover surface. ISSJ 50:493-510. change: moving beyond the myths. Global Environme- Turner, B.L., Meyer, W.B. and Skole, D.L. 1994. Global ntal Change 11(4): 261-269. Land-Use Land-Cover Change - Towards an Integrated Larson, A.M. 2003. Decentralization and forest mana- Study. Ambio 23(1): 91-95. gement in Latin America: towards a working model. Walker, P.A. and Peters, P. E. 2001. Maps, Metaphors, and Public Administration and Development 23:211-226 Meanings: Boundary Struggles and Village Forest Use on Lovell, C., Mandondo, A., Moriarty, P. 2002. The Question Private and State Land in Malawi. Society and Natural of Scale in Integral Natural Resource Management. Resources 14: 411-424 Conservation Ecology 5(2):25. [online] URL: http:// Walsh, S.J. and Cews-Meyer, K.A. 2002. Linking people, www.consecol.org/vol5/iss2/art/25/ place and policy. A GIScience approach: Kluwer Aca- Makin, MJ., Kingham, TJ., Waddams, AE., Birchall, CJ., demic Publishers, Boston

14 Paper I

Mountain Research and Development Vol 27 No 2 May 2007: 162–168

Gessesse Dessie and Johan Kleman Pattern and Magnitude of Deforestation in the

162 South Central Rift Valley Region of Ethiopia

The pattern and mag- 1962). The Wondo Genet forests support high biodiver- nitude of deforesta- sity and provide important watershed service (Mooney tion that occurred 1954; Makin et al 1975), although both aspects are now from 1972 to 2000 in threatened. About a century ago, early travelers (Maud the south central Rift 1904; Neumann 1902) reported seeing dense canopy Valley of Ethiopia were and continuous forest covering these sites. In the past analyzed using remote half century, the south central Rift Valley of Ethiopia sensing change detec- has evolved as one of the core regions in the country, in tion techniques. The terms of economy, population, urbanization, and road results show that nat- transport. This has caused rapid forest decline. ural forest cover In Ethiopia there is a shortage of accurate informa- declined from 16% in 1972 to 2.8% in 2000. The total tion regarding the rate of deforestation and the extent natural forest cleared between 1972 and 2000 amount- of earlier forest cover. The available local level land cov- ed to 40,324 ha, corresponding to an annual loss of er and land use change studies (Kebrom and Hedlund 1440 ha. The total loss was 82% of the 1972 forest 2000; Gete and Hurni 2001; Belay 2002; Woldeamlak cover and the annual loss was equivalent to 0.9% of 2002) are all from northern part of the country. The the national figure. The forest decline in the area objective of the present study is to detect, document, involved proximate causal factors as well as causal fac- and interpret the magnitude, trend, and spatial pat- tors that are more spatially diffuse and are part of the terns of forest cover change in the south central Rift long-term evolution of a region much larger than the Valley of Ethiopia between 1972 and 2000. The main study area. In order of importance, the major causes of research questions were: What was the rate of natural change were small-scale agriculture, commercial log- forest cover change between 1972 and 2000? What are ging, and commercial farms. Two major modes of the trends in fragmentation? What are the major causes change were observed; 1) internal, ie openings created and consequences of change? by small farm plots, grazing lands, and villages; and 2) external, ie expansion of agriculture from the exterior Material and methods into the forests. The main consequences of deforestation were habitat destruction and decline of water avail- The study area ability. The study area (Figure 1), which comprises 3060 km2, lies at 6°45' N to 7°15' N latitude and 38°15' E to 38°45' Keywords: Deforestation; pattern of change; change E longitude. It is located 280 km south of Addis Ababa, detection; remote sensing; biodiversity; Ethiopia. and forms part of the south central Rift Valley of Ethiopia. The area has a highly dissected topography Peer reviewed: March 2006 Accepted: June 2006 draining into 4 different watersheds. The altitude varies between 1675 and 2900 m. The climate is transitional, Introduction between a well-defined bimodal pattern to the south and a single rainy season to the north. The eastern half Tropical deforestation has become a global concern, gets annual rainfall reaching 1200 mm, whereas the with an annual total forest loss of 9.4 million ha (FAO rainfall on the western side subsides to around 900 mm 2000a). The total forested area of Ethiopia has at Lake Awassa. Within the area, the mean annual tem- decreased substantially during the past half-century or perature varies between 17°C and 19°C. so. Recent figures show that the country’s forest cover The forest is upland rain forest (Friis 1986), some- has shrunk to less than 3–3.6% (Reusing 1998; WBISPP times termed submontane seasonal rain forest (Chap- 2004). Forest loss in northern Ethiopia probably goes man and White 1970) or, more recently, Afromontane back many hundreds of years (Melaku 1992), with the rain forest (White 1983; Friis 1992). It contains a mix- main remaining forest now confined almost entirely to ture of Podocarpus falcatus and broad-leaved species at the southern part of the country (Logan 1946; Friis low elevations and in sheltered valleys, while species 1986; Friis 1992). that resemble dry single-dominant Afromontane forest In the 1970s, one of the major areas with remnant (White 1983; Friis 1992) with Juniperus procera as the high forests was the south central Rift Valley of most prominent species are found on ridges and at Ethiopia, including Shashemene, Wondo Genet, and higher elevations. A survey of the forest cover in the parts of Sidama (Chaffey 1979). Today only pockets of region was done by Chaffey (1979). Figures 2A to 2D high forest remain. The Shashemene forest was one of show representative land use types in the study area; the the most commercially valuable forests in the country different types also represent a sequence of deforesta- (Russ 1944; Logan 1946; Mooney 1954; Breitenbach tion and land use transformation. The population in Research

FIGURE 1 The study area. Digital terrain model showing the main physiographic features, including Lakes Awassa and Cheleka on the Rift Valley 163 floor, and the eastern escarpment (south of Mt Abaro) of the Rift Valley. (Map by authors) the study area is around half a million, about a third of which lives in the 2 major towns, Shashemene and Awassa. The area is comprised of parts of Oromya Regional State, and Southern Nations and Nationalities Peoples’ Regional State.

Registration of images The Landsat images were registered to the Universal Transverse Mercator (UTM) geographic projection, using modified Clarke 1880 spheroid, datum Adindan (Ethiopia) and zone 37 (see Table 1 for data source). Four topographic maps covering the area were scanned in Tag Image Format (TIFF) on a high-resolution scanner at 1200 dots per inch. The maps were then rectified (rectification error 0.3 pixels) and converted to image format. The image registration was done using distinc- tive and easily identifiable features such as road inter- sections and stream confluences.

Image classification We used a supervised maximum likelihood classification method to classify satellite images. Training areas were identified and defined as follows: for the entire study, the FAO definition of forest was adopted (FAO 2000b). readjusted until reasonable classification was attained. Forest included natural forest and established planta- To accommodate variability of forest cover types we tions >30% crown cover and >1 ha in area. Representa- used 80 training areas for the MSS and 50 for the TM tive forest cover types for MSS images were identified images. through stereoscopic and digital interpretation of aerial photographs. For the TM image, training areas were Classification accuracy identified based on knowledge of ground conditions Classification accuracy was tested by overlaying grids acquired through extensive multi-year fieldwork in the on the aerial photographs (Quirk and Scarpace 1982; area. The most homogenous portion of each sample Rembold et al 2000) and the IKONOS image. Three type was marked as a training area, after which the his- sets of aerial photographs representative of forest con- togram of each signature was checked for spectral sepa- ditions in 3 different parts of the study area were rability. The size and shape of the training areas were selected for verification of the 1972 classified image,

TABLE 1 Data sources and description.

Purpose Data/specifications Acquisition date/ Source Resolution/ scale

MSS Image 26 Dec 1972 (dry season) 57-m pixel Path/raw 181/55 USGS Classification ETM Image 5 Dec 2000 (dry season) 30-m pixel Path/raw 168/55 USGS

Aerial photo Nov-Dec1972 1:50,000 133ET1-133ET15 Ethiopian Mapping Agency Verification IKONOS 10 Dec 2000 4-m pixel Wondo Genet Wondo Genet College of Forestry

Topographic maps 1976, 1988 (prepared) Registration 0738 C4,0738 D3, 1:50,000 Ethiopian Mapping Agency 0638 A2, 0638 B3 Gessesse Dessie and Johan Kleman

FIGURES 2A–2D Pattern of deforestation. A) Unchanged forests located on the slopes protected by Wondo Genet College of Forestry. The upper canopy trees are Aningeria adolfi-friedericii, Podocarpus falcatus, and Prunus africanus. These species are rare and have a high market value. B) Land use transformation in what was originally the unbroken Shashemene forest. A plantation of mainly Cupressus lusitanica is seen in the background. The foreground shows small-scale farming, mainly maize mixed with grazing land and settlement. C) Pattern of change from forest to grazing land at high elevation >2500 m in Wondo Genet forest; the 164 dominant tree species is Juniperus procera, a durable construction wood with high market value. D) A landscape of farm plots that have completely replaced the forest, as a result of encroachment from the exterior. (Photos by Gessesse Dessie)

and one available IKONOS scene was used for verifica- Above 87% correct classification was attained for the 2 tion of the 2000 classified image. Digital orthophoto classifications. The change detection accuracy sets were created from the aerial photographs to elimi- (0.87*0.87) was 0.76. Percentage correct is a simple nate topographic and other distortion and to make measure of accuracy and well suited to a classification direct comparison with the classified 1972 image. The where only a few broadly defined classes are used orthophoto representing the true forest area and the (Campbell 1996). Various sources of error were recog- classified image representing forest area were then nized, including resolution and format difference compared (Figures 3A and 3B). The 2 units were print- between MSS and TM images, shadows in high-relief ed at a scale of 1:50,000; thereafter, the amount of for- terrain, localized haze, and the occurrence of restrict- est falling on the grid points was counted for both ed areas of other vegetation types (eg papyrus grass) prints using a square grid (4-mm2 unit). The accuracy that were occasionally classified as forest. We regard of classification was calculated by dividing the propor- the >87% correct as a classification accuracy which is tion of forest in the image by the proportion of forest fully satisfactory for the purpose of this study, where in the orthophoto. A similar procedure was followed the total magnitude of change between the 2 dates is for the year 2000 TM image using the IKONOS image. large (see below).

Mountain Research and Development Vol 27 No 2 May 2007 Research

165

FIGURES 3A AND 3B A) 1972 orthophoto compared to the 1972 classified image (B) for the estimation of classification accuracy. The site shows part of the Shashemene forest. Percentage correctness of the classification (100b/a) is >87%.

Results and discussion TABLE 2 Forest cover change between 1972 and 2000 estimated from satellite image analysis with 76 percentage correct accuracy. Quantitative trend The areas under forest in 1972 and 2000 are presented Forest area 1972 2000 in Table 2 (for spatial coverage, see Figure 4). A closer investigation of the changes revealed that of the 48,924 In hectares 48,924 8600 ha of forest in 1972, less than 1% remained unchanged. In percent 16.0 2.8 The forest loss of 1440 ha/yr is equivalent to 0.9% of the national annual forest cover loss (163,000 ha) reported by Reusing (1998). forests have glades of edaphic grasslands on seasonally The rate of deforestation in the study area can be waterlogged sites and marshes (Tewolde Berhan 1990). compared with studies elsewhere in Ethiopia. Northern The pattern is further shaped into linear form by a den- Ethiopia is generally regarded as heavily deforested, but dritic drainage pattern (see Figure 1). This linear struc- some data exist for areas with remaining pockets of for- ture is still apparent in Figure 4B, although substantial- est. In a study area of 11,000 ha in Northeast Ethiopia ly encroached and thinned. Location and distribution (1958–1986 study period), forest cover declined from in the upstream parts of catchments make these forests 7.8% to 5.4% (Kebrom and Hedlund 2000). A similar hydrologically important to the region. The non-linear study in Northwestern Ethiopia (covering 27,103 ha for fragmented patches are mainly located on slopes and in the period 1957–1995) documented a drop in natural inaccessible valleys. These isolated patches are key habi- forest cover from 27.1% to 0.3% (Gete and Hurni 2001). tats and harbor rare and threatened animal and plant The trend of forest decline in the study area corre- species. lates with the general pattern of forest decline in Figure 4B depicts a speckled spatial pattern formed Ethiopia over the last few hundred years (Kebrom and of non-connected small forest patches. Fragmentation Hedlund 2000; Belay 2002). However, there are a few generally resulted in a landscape that consisted of rem- examples of forest cover gain (Crummey 1998; McCann nant areas of native vegetation surrounded by a matrix 1999). There is an increasing trend of homestead exot- of agricultural or other developed lands (Saunders et al ic (non-indigenous trees) planting. A similar trend has 1991). A closer look (Figures 2B and 2C) shows a pat- been reported elsewhere in Ethiopia (Woldeamlak tern consisting of small farm plots and grazing land 2002). units. This pattern is in contrast to the large regularly Of the high forest lost (approximately 40,300 ha) in shaped openings in the Shashemene and Wondo Genet the study area, more than 82% was lost to mainly agri- forests (Figure 4A). These openings are created near 3 cultural land and about 17% was reforested and/or commercial sawmills. The spatial pattern is not only afforested. Studies elsewhere in Ethiopia likewise report- shaped by perforation and fragmentation inside the ed that forests were replaced mainly by cultivated land forests but also by changes on the exterior (Figure 5). (Gete and Hurni 2001; Kebrom and Hedlund 2000). Figure 2D shows complete replacement of border forests by agricultural land. Spatial patterns Examination of forest maps (Figures 4A and 4B) and Causes of change analysis of the topography revealed that the natural pat- Analysis of the satellite images and ground observation tern of forest distribution is governed by topographic revealed that small-scale permanent agriculture is the conditions and surface moisture states. The evergreen major cause of deforestation. This result is in line with Gessesse Dessie and Johan Kleman

166

FIGURES 4A AND 4B Deforestation 1972–2000 in the study area. A) Classified MSS 1972 image showing the forest cover during 1972. The forest is found at altitudes above 1800 m on upstream catchment area and covers 16% of the total study area. The regularly shaped openings (1, 2, 3) are created by commercial logging. B) Classified TM image showing the forest cover in 2000. The forest covers 2.8% of the total study area. Regularly shaped patches and strips are mainly plantations of Cupressus lusitanica, Pinus patula, and Eucalyptus spp. Lake Cheleleka has dried out.

previous studies (Mooney 1954; Russ 1944) and a diffuse and are part of the long-term evolution of a recent comprehensive study (WBISPP 2004). The other large surrounding region. According to Lambin et al causes of forest cover change are commercial logging (2001) various human–environment conditions react to and commercial agriculture. Commercial wood extrac- and reshape the impacts of drivers differently, leading tion was begun in the mid-1930s by 3 sawmills logging to specific pathways of land use change. in Shashemene and Wondo Genet forests (Mooney Forest decline in the period 1972–2000 is not an 1954). Cleared areas were immediately occupied by set- isolated event, but rather a continuation of past trends. tlers and converted to farmlands (Mooney 1954; Zeri- The forests of south Ethiopia have been subjected to hun 1988). The sawmill in Wondo Genet was closed in human influence for centuries. Pollen analysis (Bon- the early 1970s, while the other 2 remained in opera- nefille and Hamilton 1986) has revealed that the onset tion until the mid-1980s. Only a minor fraction of the of major disturbances, including mountain forest supply during the final period of operation came from destruction, occurred around 850 BP. In the study area, the study area. During the period 1950–1975, establish- soil carbon analysis (Zewdu and Högberg 2000) has ment of commercial farms (coffee and cereal crop revealed more recent major land cover change. Travel- farms) contributed to deforestation primarily through ers’ accounts at the beginning of the 20th century (Neu- clearance of forest and through the secondary effect of mann 1902; Maud 1904) indicate that at that time the in-forest settlement of evicted farmers (Ståhl 1974; study area was heavily forested. Our data show that by Makin et al 1975; Benti 1988). 1972 forested area was already down to 16%, suggesting The changes taking place are related to the gradual that a great deal was lost in the period 1900 to 1972. emergence of the region as one of the core centers in During the study period, the main causes of change Ethiopia. In south central Ethiopia, particularly around were the increasing expansion of small-scale farming, Awassa and Shashemene, the population grew fast, set- particularly cash crops, and the factors that drove it. tlements increased and expanded rapidly, road net- Another persistent cause was selective cutting and for- works improved, and economic sectors developed est fires. (Bjerén 1985). The forest decline in the area therefore Causes of abrupt change have also been identified. involves a combination of specific and proximate causal According to local residents in Wondo Genet, 2 periods factors, including causal factors that are more spatially known to be times of abrupt change are 1974–75 and

Mountain Research and Development Vol 27 No 2 May 2007 Research

167

FIGURE 5 Modes of change. The three primary modes of change (reduction) in forested areas: internal (1 and 2) and external (3). The internal modes appear as openings created by small farm plots, grazing, and settlements, while the external mode results from expansion of agriculture from the exterior into the forest.

1991–92. These periods correspond with the 2 most reduced streamflow related to deforestation and recent changes of government in Ethiopia. Farmers increased use of streamwater for irrigation purposes. point out that during periods of disorder and transition While deforestation contributes to the decline of between governments, more trees are cut and more water resources, it indirectly contributes to a decrease in forestland is settled. The other explanation is a lack of agricultural productivity and also affects peoples’ health. clear forest property rights. Melaku (2003) reported The people of Wondo Genet are heavily dependent on that state-owned forests experienced huge losses after irrigation agriculture, and Awassa and Shashemene the change of government in 1991. towns (with a combined population of about 200,000) rely on potable water drawn from the area. Additionally, Consequences forest loss has affected practical forestry training at Won- Habitat loss and habitat fragmentation create loss do Genet College of Forestry. The college, the only one and/or decline of species (Darren et al 1998). Accord- of its kind in the country, is heavily dependent on this ing to Saunders et al (1991) the influence of physical forest for outdoor exercises and experiments. and biogeographic change is modified by the size, shape, and position in the landscape of individual rem- Conclusion nants, with larger remnants being less adversely affected by the fragmentation process. The high biodiversity In addition to the severe loss in forested area, defor- of forests in the study area has been identified in many estation in the study area has also created a highly frag- studies (Lorna 1979; Cross 2003). Earlier, Makin et al mented pattern with a multitude of small patches. The (1975) proposed to establish a wildlife reserve in the patches are shaped by the expansion of permanent cul- Abaro and Wondo Genet area. Forest change leads to tivation and grazing by smallholders. Small patch size, habitat destruction and threatens to make rare tree lack of connectivity, and permanent agricultural activi- species such as Pouteria adolfi-friederici, Podocarpus falca- ties have seriously undermined the resilience/regenera- tus and Prunus africanus extinct. Already in the mid- tion potential of the forest. 1980s, the mountain nyala antelope, one of the mam- In the long run, plantations may ease the energy, mals endemic to Ethiopia, was extinct in the study area. feed, and construction needs of the people, but will not Another consequence may be a decline in availabili- contribute to any significant recovery of biodiversity. ty of water. For example, Lake Cheleleka dried up dur- Loss of natural forest has already caused substantial ing the study period (Figure 4B). Local sources reported damage to biodiversity and watershed services. To revive that during the past 30 years a number of streams previ- some of these services, it is important to conserve rem- ously feeding the lakes have dried up and/or the flow of nant patches before it is too late. water has substantially decreased. There are 3 possible The present study, by examining a nationally causes for the lake drying up: lowering of the outlet important site during a dynamic period, indicates the threshold; sedimentation in the lake basin; reduced magnitude, causes, and consequences of upland defor- streamflow. There are no indications of downcutting of estation in Ethiopia. The major proximate cause is the outlet and little erosional or depositional evidence expansion of smallholder agriculture. The conse- to suggest a major increase in sediment supply. It there- quences are far-reaching in terms of environmental, fore appears probable that drying up is caused by livelihood, and even academic impacts. Gessesse Dessie and Johan Kleman

168

ACKNOWLEDGMENTS AUTHORS

The authors wish to thank Göran Alm for his expertise during their work Gessesse Dessie with ERDAS orthophoto. Special thanks go to the Swedish International Stockholm University, Department of Physical Geography and Quaternary Development Agency (SIDA) and the International Foundation for Science Geology, Svante Arrhenius väg 8C, 106 91 Stockholm, Sweden. (IFS) for financing this study. Debub University, Wondo Genet College of Forestry, PO Box 128, Shashemene, Ethiopia. [email protected];

Johan Kleman Stockholm University, Department of Physical Geography and Quaternary Geology, Svante Arrhenius väg 8C, 106 91 Stockholm, Sweden. [email protected]

REFERENCES Logan WEM. 1946. An Introduction to the Forests of Central and Southern Ethiopia. Oxford, United Kingdom: Imperial Forestry Institute, University of Belay T. 2002. Land-cover/land-use changes in the Derekolli Catchment of Oxford. the south Welo Zone of Amhare Region, Ethiopia. Eastern African Social Lorna S. 1979. Birds of Wondo Genet. Stockholm, Sweden: Orgut-Swedfor- Science Research Review 18(1):1–20. est Consortium. Benti G. 1988. A History of Shashemene from its Foundation to 1974 [MA Makin MJ, Kingham TJ, Waddams AE, Birchall CJ, Teferra T. 1975. Develop- thesis]. Addis Ababa, Ethiopia: Addis Ababa University. ment Prospects in the Southern Rift Valley, Ethiopia. Land resource study Bjerén G. 1985. Migration to Shashemene. Ethnicity, Gender and Occupa- 21. 2 volumes. London, United Kingdom: Land Resources Division, Min- tion in urban Ethiopia [PhD dissertation]. Stockholm, Sweden: Stockholm istry of Overseas Development. University. Maud P. 1904. Exploration in the southern borderland of Abyssinia. Geo- Bonnefille R, Hamilton A. 1986. Quaternary and late tertiary history of graphical Journal 23(5):552–579. Ethiopian vegetation. In: Hedberg I, editor. Research on the Ethiopian Flo- McCann JC. 1999. Green Land, Brown Land, Black Land: An Environmental ra. Proceedings of the first Ethiopian Flora Symposium. Symbolae Botani- History of Africa, 1800–1990. Oxford, United Kingdom: James Currey. cae Upsalienses 26(2):48–63. Melaku B. 1992. Forest History of Ethiopia from Early Times to 1974 [MA Breitenback F. 1962. National forestry development planning. A feasibility thesis]. Bangor, UK: University College of North Wales. and priority study on the example of Ethiopia. Ethiopian Forestry Review Melaku B. 2003. Forest Property Rights, the Role of the State, and Institu- 3(4):41–68. tional Exigency: The Ethiopian Experience [PhD dissertation]. Uppsala, Swe- Campbell JB. 1996. Introduction to Remote Sensing. London, United King- den: Swedish University of Agricultural Sciences. dom: Taylor and Francis. Mooney HF. 1954. Report on Forestry in Ethiopia with Special References to Chaffey DR. 1979. South-west Ethiopia Forest Inventory Project. An Inventory the Forests of Arussi and Sidamo. Oxford, United Kingdom: Imperial Forestry of Forest at Munessa and Shashemene. Project Report 29. London, United Institute, University of Oxford. Kingdom: Land Resources Division, Ministry of Overseas Development. Neumann O. 1902. From the Somali coast through southern Ethiopia to the Chapman JD, White F. 1970. The Evergreen Forests of Malawi. Oxford, Sudan. Geographical Journal 20(4):373–398. United Kingdom: Commonwealth Forestry Institute, University of Oxford. Quirk BK, Scarpace FL. 1982. A comparison between aerial photography Cross P. 2003. The Butterflies of Wondo Genet. An Introduction to the But- and Landsat for computer land-cover mapping. Photogrammetric Engineer- terflies of Ethiopia. Wondo Genet, Ethiopia: Wondo Genet College of ing and Remote Sensing 48(2):235–240. Forestry. Rembold F, Carnicelli S, Nori M. 2000. Use of aerial photographs, Landsat Crummey D. 1998. Deforestation in Wällo: Process or illusion? Journal of TM imagery and multidisciplinary field survey for land-cover change analy- Ethiopian Studies 31(1):1–42. sis in the lakes region, Ethiopia. International Journal of Applied Earth Darren JB, Thomas AC, Lenore F. 1998. Habitat loss and population Observation and Geoinformation 2(3/4):181–188. decline: A meta-analysis of the patch size effect. Ecology 79(2):517–533. Reusing M. 1998. Monitoring of Natural High Forest Resources in Ethiopia. FAO [Food and Agriculture Organization]. 2000. The Challenges of Sustain- Addis Ababa, Ethiopia: Ministry of Agriculture. able Forestry Development in Africa. Yaoundé, Cameroon: FAO. Also avail- Russ GW. 1944. Reports on Ethiopian Forests. Compiled in 1979 by Wolde- able at: http://www.fao.org/docrep/meeting/x3976e.htm; accessed on 9 Michael Kelecha. Addis Ababa, Ethiopia: Ministry of Agriculture. January 2007. Saunders DA, Hobbs RJ, Margules CR. 1991. Biological consequences of FAO [Food and Agriculture Organization]. 2000a. Global Forest Resources ecosystem fragmentation—a review. Conservation Biology 5(1):18–32. Assessment. Rome, Italy: FAO. Ståhl M. 1974. Ethiopia: Political Contradictions in Agricultural Develop- Friis I. 1986. The forest vegetation of Ethiopia. In: Hedberg I, editor. ment. Uppsala, Sweden: Scandinavian Institute of African Studies. Research on the Ethiopian Flora. Proceedings of the first Ethiopian Flora Tewolde Berhan GE. 1990. Vegetation and environment of the mountains Symposium. Symbolae Botanicae Upsalienses 26(2):31-47. of Ethiopia: Implications for utilization and conservation. In: Messerli B, Friis I. 1992. Forests and Forest Trees of Northeast Tropical Africa. Their Hurni H, editors. African Mountains and Highlands: Problems and Perspec- Natural Habitats and Distribution Patterns in Ethiopia, Djibouti and Somalia. tives. Marceline, MO: African Mountains Association and Walsworth Press, Kew Bulletin Additional Series XV. Kew, United Kingdom: HMSO [Her pp 207–216. Majesty’s Stationery Office]. WBISPP [Woody Biomass Inventory and Strategic Planning Project]. 2004. Gete Z, Hurni H. 2001. Implications of land use and land cover dynamics Forest Resources of Ethiopia. Addis Ababa, Ethiopia: Ministry of Agriculture. for mountain resource degradation in the northwestern Ethiopian high- White F. 1983. The Vegetation of Africa. A Descriptive Memoir to Accompany lands. Mountain Research and Development 21(2):184–191. the UNESCO/AETFAT/UNSO Vegetation Map of Africa. Natural Resources Kebrom T, Hedlund L. 2000. Land cover changes between 1958 and 1986 Research 20. Paris, France: UNESCO [United Nations Educational, Scientif- in Kalu district, southern Wello, Ethiopia. Mountain Research and Develop- ic and Cultural Organization]. ment 20(1):42–51. Woldeamlak B. 2002. Land cover dynamics since the 1950s in Chemoga Lambin EF, Turner BL, Geist HG, Agbola SB, Angelsen A, Bruce JW, Coomes watershed, Blue Nile Basin, Ethiopia. Mountain Research and Development OT, Dirzo R, Fischer G, Folke C, George PS, Homewood K, Imbernon J, Lee- 22(3):263–269. mans R, Li X, Moran EF, Mortimore M, Ramakrishnan PS, Richards JF, Zerihun M. 1988. Jiggessa and Sole Sawmills and Joinery 1939–1985, Skånes H, Steffen W, Stone GD, Svedin U, Veldkamp TA, Vogel C, Xu J. Shashsamane [BA thesis]. Addis Ababa, Ethiopia: Addis Ababa University. 2001. The causes of land-use and land-cover change: Moving beyond the Zewdu E, Högberg P. 2000. Effect of land use on 15N natural abundance myths. Global Environmental Change 11(4):261–269. of soils in Ethiopian highlands. Plant and soil 222(1):109–117.

Mountain Research and Development Vol 27 No 2 May 2007

Paper II

Forest Decline and its Causes in the South Central Rift Valley of Ethiopia: Human Impact over a One Hundred Year Perspective

Gessesse Dessie1,2 and Carl Christiansson1

1 Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden 2 Wondo Genet College of Forestry and Natural Resources, Hawassa University, Ethiopia

Abstract

Forest decline in Ethiopia is highlighted by several authors but there is no consensus on its causes and consequences. The objective of this study is to investigate, from a geographical perspective, linkage between the trend of forest decline and changes in the social, economic and political pattern in the Awassa watershed over a one hundred year perspective. Field observations, satellite image and map analyses, interviews and literature studies were employed and natural indicators were analyzed. The findings indicate that the forest area declined from about 40% at the turn of the 19th century to less than 3% in the year 2000. Forest decline in the study area during the elected time period is the result of the combination of biophysical and social conditions. Important causes are geographic properties, sociopolitical changes, population growth, unstable land tenure principles, agricultural development and the improvement of transport capacity. The main conclusions are as follows. Already in the early 20th century forest decline was in progress and forests were attributed an insignificant economic classification. Large areas of forest were cut down during periods of political transition when as a result of the political vacuum interest in the protection of resources, including forests was lacking. Long term planning efforts to manage forests were obstructed by uncertainty resulting from land tenure principle change during each political period. The sparse area of forest land which remains is becoming increasingly attractive as potential land for arable agriculture due to improved road access between the study area and distant markets.

Key words: Forest decline, human impact, land use, land tenure, population increase, Ethiopia

Introduction 6, 7). However, a limited number of case studies from Ethiopia challenge the common explanations of forest Forest decline in Ethiopia is highlighted by several decline such as population growth (1, 2, 6, 9). Some of authors (1, 2, 3, 4). Disagreement, however, exists these emphasize that multiple social and biophysical about its causes and consequences (5, 6, 7). The factors act together to cause forest decline (10, 11, need for more space for cultivation is mentioned as 12). In order to reveal a pattern, studies of forest one reason for trees being increasingly removed (8). decline need to be based on analyses of spatially re- EFAP (7) claims that the population increase plays a presentative areas, while perceptions of general trends major role. McCann (1) points out broader sociopo- are revealed by considering an adequately extensive litical conditions as important factors in the process. temporal scale (12, 13, 14). Locally, systems of tree tenure influence the trend of Since the 1880s significant natural and political deforestation (9). events have deeply affected the social and natural In forest management at state and province level it environment of Ethiopia, not least in the southern re- is not uncommon that policy directives, development gions. Of specific impact can be named the combined objectives, research agendas and priorities of develop- effect of rinderpest, drought and locust invasion in the ment assistance to the forest sector are founded on late 19th century which led to a human catastrophe incomplete and biased data (1, 2, 5). Furthermore, and an extensive environmental crisis (15, 16) and forest development strategies are often outlined by furthermore facilitated the conquest and occupation external experts (6). of what is now southern Ethiopia by northern Ethio- National and local studies based on remote sensing pian armies, “the southern incorporation” (17). In the (3, 4) indicate the rapid decline of natural forests. 20th century sociopolitical, economic and cultural These instances point at two fundamental issues. conditions have changed gradually and several times Firstly, forest decline is in progress (3). Secondly, abruptly. As a result of these changes the natural the people-forest interaction has frequently been indigenous forest, which a century ago covered parts misinterpreted and simplistically explained (1, 2, 5, of southern Ethiopia, was greatly affected (1, 18).

Gessesse Dessie and Christiansson, C. 2007. Forest decline and its causes in the south central Rift Valley of Ethiopia. Human impact over a one hundred year perspective. Accepted for publication in Ambio. Gessesse Dessie and Carl Christiansson

N a

Shashemene

Mt. Abaro 2580 Lake Cheleleka Lake Wondo Genet Awassa Wosha PA College of Forestry Awassa Western escarpments Wetland 1675 Basha

Eastern escarpments

R e d S e a

100 N Addis Ababa Ethiopia Study area

1 0 k m 40 0 E

Figure 1. a) Digital terrain model showing the study area (the broken line indicates the Awassa watershed boundary). b) The northeastern escarpment of the study area during the wet season, and c) during the dry season. This area was covered by conifer forests as recently as the early 1970s. Prior to 1975 the major part of the land in the foreground was owned by one individual. It is currently cultivated by several farming households. d) A section of the wetland. There are few settlements in this area due to poor drainage and prevalence of malaria. Much of this land was, and still is, a common grazing area. Photos: Gessesse Dessie

However, a thorough understanding of the process of sed of mainly self supporting single stemmed woody forest decline during this period is incomplete. vegetation more than 5m in height and with a canopy Thus, the objective of this study is to investigate, cover of over 10% (19). In this article we combine from a geographical perspective, linkages between these two categories under the term “forest”. Where the trend of forest decline and changes in the social, relevant, however, the two categories are referred to economic and political pattern in Awassa watershed separately. in south central Ethiopia over a century. We here define forest decline as deforestation and forest degradation. Deforestation is the complete re- Study area moval of tree cover or substantial reduction of canopy cover (below 30%) over large areas while by forest The Awassa watershed and its nearest surroundings degradation we mean loss of the main attributes of in the south central Rift Valley were selected as the forests such as capacity to produce timber, fire wood, study area (Fig 1a). The study area comprises 3060 non-wood products and environmental services (10, km2 and lies at 6o45’ N to 7o15’ N latitude and 11). Natural forest in this context are forest stands 38o15’E to 38o45’ E longitude, located some 280 km composed predominantly of native tree species, es- south of Addis Ababa. Altitudes in the region vary tablished naturally, with a canopy cover above 30% markedly. Lake Awassa on the valley floor is situated while natural woodland is a tree community compo- at 1675 meters above see level while the crests of the

Forest decline and its causes in the south central Rift Valley of Ethiopia

Figure 1. Cont. rift escarpments reach 2900 and 2400 meters in the half of the Awassa watershed. The remainder of the east and west respectively. The large-scale landform is study area consists of uplands, escarpments with slo- a tectonic depression bounded by steep escarpments. pes varying between 6 and 30 degrees and foothills. The dissected plateau surfaces at the top of the es- The Awassa watershed is a closed hydrological basin carpments slope gently away from the rift. Thus, the supplied by perennial streams, which emerge from the highest points in the study area are found close to the eastern escarpments. escarpments. The floor of the depression is covered The pattern of rainfall is transitional between by Lake Awassa and the wetlands of the former Lake a bimodal pattern to the south and a single rainy Cheleleka and alluvial plains, which together cover season to the north. Rain can be expected in March

Gessesse Dessie and Carl Christiansson to April and June to August. Rainfall amounts differ in Awassa town and in Shashemene town. Rural and between the drier western escarpments and the valley town populations are presently growing at rates of floor, which receive 900 mm a year, and the wetter 2.1% and 4.1% per annum, respectively (23). Accor- eastern escarpments, which receive 1200 mm (20). ding to Yibeltal (24) the highest population density in The annual rainfall pattern in the study area for 1931 Ethiopia (the ratio of rural population to cropland), to 2000 is shown in Fig 2. The mean annual tempera- 1488/km2, is found in Wosha peasant association ture at Awassa town is 19 degrees centigrade and at (Fig 1a) within the study area. Wondo Genet 17 degrees centigrade. Over the whole study area mean temperatures vary greatly depending on altitude and aspect. Methods The soils are formed on recent volcanic and lake deposits and generally lack distinct profile develop- In this study we have focused on four variables which ment. The lowest part of the valley floor is covered by are of particular importance in explaining the process lake deposits well supplied with plant nutrients. In the of forest decline in Ethiopia (1, 6, 7, 9): changes in northern part of the study area the soils are coarse- population and settlement, new land tenure prin- textured, while the western section of the watershed ciples, agricultural innovations and improvements has acidic, coarse, shallow soils. The dominant soil in transport capacity. These variables are frequently types are Cambisols and Leptosols. The types of natu- referred to in studies of tropical forest decline (12, ral vegetation are determined by altitude, rainfall and 13, 14, 25, 26, 27, 28) and found highly important soil fertility. In the wet, eastern part of the watershed for the study area. Other variables that also influence montane forest composed of Podocarpus falcatus and forest decline are logging, forest fires (mainly human Juniperus procera, is dominant (22). In the sub hu- caused), and management principles. mid western part shrubs and thickets dominate (20), To define the long term perspective we employed while on the lower foot-slopes and on the valley floor methods described by Batterbury and Bebbington grassland and acacia woodland prevail. (29), Roberts (30) and Burgi and Russell (31) which The current land use is predominantly smallholder include field observations, interviews, travelers’ ac- agriculture (Fig 1b and Fig 1c) with an average land- counts, official history texts, dissertations, journal holding size of below one hectare per household in articles, maps and satellite images. To determine the the valley and foothills (3). Major crops grown in the initial forest area we employed a method (theoretical study area include maize, sugarcane, enset, khat and maximum extent of forest) described by Mather et potato. The population of the study area is composed al (27) and Matthews (32). Theoretical maximum of six main ethnic groups together numbering about extent of forest denotes the maximum forest cover half a million (23). About one third of the people live that existed in the past determinable by analyzing the

1600 1500

1400

1300 1200

1100

1000 Rainfall (mm) 900

800

700 600 1931 1935 1939 1943 1947 1951 1955 1959 1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 Year

Figure 2. Annual rainfall pattern in the study area for the interval 1931to 2000. The sources of these data are the Ethiopian Meteorological Agency and Kingham (21). The latter source provides rainfall records of Awassa, Shashemene and Wondo Genet from 1931 to 1970 while the Meteorological Agency source covers the period thereafter. The curve indicates mean values for the three stations. The average rainfall for the period of records is 1100 mm.

Forest decline and its causes in the south central Rift Valley of Ethiopia characteristics of biophysical factors (e.g. climate) that development project reports and archival records. are favorable or alternatively unfavorable for forest These documents were used to assemble time series growth, assuming minimal human impact. information on sociopolitical conditions, expansion of agriculture and settlements, and history of forests. 1. Natural indicators of relevance to forest distribu- Most of these sources have limitations as they em- tion are altitude (temperature), aspect, rainfall, slope phasize general themes and often lack specificity of gradients, rock outcrops, surface drainage, soil depth, time and area. soil moisture and soil pH. These indicators were iden- The precision of aggregated data was affected by tified using satellite images, air photos, topographical the difficulty of combining spatial and non-spatial maps (33) and field observations, including observa- data formats. The limitations were counteracted by 1) tion of remnant trees. These indicators were used in utilizing diverse and extensive data sources of spatial combination with the other methods, to determine and non-spatial data formats 2) careful evaluation the approximate extent of the natural forest at the of the consistency of the information obtained from beginning of the 20th century. various data sources 3) organizing the information to obtain a comprehensive and coherent picture, ac- 2. Travelers’ accounts consist of records by travelers hieved by presenting the information in a number of who visited the study area during the late 19th and ways: a graph is used to show the timeline of events; early 20th century. Their point of observation in the a sketch portrays the evolution of the cultural lands- terrain including the route they followed and locations cape; a flow diagram illustrates the interaction of the of their camping sites were considered. The routes causal factors and satellite images display the spatial selected pass across the eastern escarpments, across characteristics of the forest. the eastern foothills and across the valley by Lake Awassa and it may be confidently assumed that the 5. The estimation of the extent of the forest at the routes selected gave the travelers representative points beginning of the 20th century was carried out in two of observation of the study area. steps. First the boundaries of forest deterring indica- The travelers made first hand descriptions of the tors i.e. the lake, wetlands, steep slopes, non-favorable watershed and produced sketch maps, their journals soils and low rainfall were approximated using dif- thus presenting information regarding land cover ty- ferent maps (topographic, soil and hydrologic) and pes, people’s livelihoods and conditions of agricultural satellite images. Secondly the spatial distribution of landscapes and grazing areas. From these sources it forest was estimated directly from the distribution is possible to make an estimation of the extent of of remnant trees, and from information given in the forest cover in the past. The limitations of these interviews and travelers accounts. Thirdly the most sources arise from the fact that they are one-time feasible boundaries were determined according to the observations and confined to the areas close to the following relation: Forest = total study area – (lake travel routes. + wetland + steep slopes + unfavorable soils + low rainfall areas). 3. Interviews were conducted with 45 male residents with an average age of 60 years who have lived their entire lives within the study area. Of 15 interviews 9 were conducted in the form of group discussions, the remainder being conducted with key informants. Results Most of the interviews took place on high ground To illustrate the forest-people interrelations we crea- with a clear view of the surrounding landscape. The ted two graphical models: 1) a timeline of human interviews were structured around the framework activities in relation to the decline of the forest (Fig set by a checklist of topics including distribution of 3) and 2) a map showing, in layers, causes which over forests, past land use types, extent and patterns of time have contributed to forest decline (Fig 4). human immigration, cultivation patterns and the establishment of settlements. The respondents were asked to describe the extent of forest cover in the past Forests in the study area between 1900 and 2000 with reference to defined topographic fixtures. The th possible extent of forest was then determined after the We estimate that at the turn of the 19 century, natu- group had deliberated and come to agreement. The ral forests, including woodlands, covered about 40% result was further compared with descriptions made of the study area. The remaining 60% comprised of by travelers and an appraisal of natural indicators. savanna with scattered trees, bushland, grassland, Reference to the past was made using well-recognized wetland and lakes. The extent of the forested area events such as government changes, ethnic conflicts, in recent years was interpreted from satellite images. drought and famine years, initiation of development In 1972 the forest/woodlands cover was 16%. By projects, road construction and establishment of new the year 2000 it had decreased to less than 3% (3) settlements. (Fig 5). Two travelers who visited the study area between 4. The documents and records used for this study 1900 and 1904 have given indications of the extent of included scientific articles, unpublished theses, forest and woodland cover. Maud (34) who walked

Gessesse Dessie and Carl Christiansson

Socio-political Characteristic periods land uses 1 EPRDF Smallholder period agriculture Military Smallholder 1.3 2.3 period agriculture Italian Commercial 4.2 period logging Large-scale 4.3a 4.3b Hayla-Sellase commercial period agriculture 1.2 2.2 3.2 3.3 Menilek Garrison settlements/ 1.1 period Coffee farms 3.1 4.1 2.1 Pre-Menilek Pastoralism/ period Enset culture Forest decline factors 1. Land tenure change 1875 1900 1925 1950 1975 2000 1.1 Land measurement 1.2 Private ownership 50% 1.3 Public ownership Long-term trend in natural forest decline

2. Agricultural change 2.1 Coffee farming 2.2 Mechanized farming 2.3 Smallholder farming 25 3. Population change 3.1 Famine/disease 1972 3.2 Health/security

3.3 Immigration % forest cover

4. Transport capacity 2000 4.1 Caravan tracks 4.2 Direct road to south Ethiopia 0 4.3 Tarmac highway a) the first, b) upgraded

Figure 3. Human activities and the trend of forest decline over a 100-year timescale (1900-2000). The figure shows a time line divided into periods of political dominance and characteristic land uses of each period. The horizontal lines indicate the time each political period lasted and the circled numbers indicate periods of heavy impact on forests. The zigzag lines indicate violent transition periods between periods of strong political dominance. The broken line in the forest decline diagram shows the estimated loss in forest area based on circumstantial evidence while the solid line between 1972 and 2000 shows satellite image measured values. The three depressions in the line prior to 1972 indicate where it is likely that significant forest conversion occurred. 1) Ethiopian Peoples Revolutionary Democratic Front governing party from 1991 until present.

the escarpments and climbed Mount Abaro (Fig 1a) established, but it is likely that montane forests co- in 1901 wrote “Mount Abaro and the country to vered uplands and slopes in the east and woodlands the east is covered with dense forest as far as the covered parts of the valley floors. In places, though, eye could reach”. Neumann (35) who observed the shallow soils or waterlogged conditions created same forest from his camp near Lake Awassa (Fig glades, edaphic grasslands and marshes (34, 36, 37) 1a) in 1900 wrote, “primeval forest covered the that limited forest growth. escarpments east of Lake Awassa”. These accounts Furthermore, although human activities may suggest that forests covered the eastern escarpments have interrupted the continuity of the forest, this and adjacent land. Neither of the authors give an seems to have been insignificant in the area in the areal estimate or crown cover percentage of the early 20th century. Maud (34) reported that “the forest. In 1944 Russ (22) provided some area esti- country is sparsely inhabited, though it is very fertile mates: for example, Shashemene forest surrounding and well watered by many perennial streams”. The Mt. Abaro in the north east of the study area was same condition was observed by other travelers (35, reported as covering 35000 ha which constitutes 38). This is not surprising considering the generally roughly a third of the forest in the study area at that sparse population of Ethiopia in the past (37) and time. The northwestern part of the study area west settlement deterring factors in the study area, such of Shashemene town was covered with Euphorbia as malaria (39, 40) and ethnic conflicts (41). Accor- and Acacia woodland, described by Maud (34) and ding to Falcon et al (42) it was not until the 1960s Neumann (35). Whether the forest/woodlands were and 1970s that the population in the area began to continuous over large areas has not been entirely increase dramatically.

Forest decline and its causes in the south central Rift Valley of Ethiopia

The role of human activities in forest impact of the natural calamities contributed to the decline southern incorporation. Among the consequences were changes in land use, including introduction of The timeline (Fig 3) displays political and socio-eco- the ox-plow and cereal crop agriculture to the former nomic factors that explain the decline of forest cover. predominantly pastoral areas (34, 37, 43, 44, 45) as The 1880s and 1890s reveal a combination of natural well as to those where hoe cultivation was practiced and socio-political events that affected human life in (15, 47, 48). The immigrants from the north also southern Ethiopia in general and in the study area in brought with them Christianity with, as suggested particular. During the infamous Ethiopian famine of by McClellan (49), a different perception of land 1882-1892, contributory factors of which were an and forests compared to that of the inhabitants of outbreak of rinderpest, a locust invasion and crop the south. failure, substantial numbers of humans and animals perished throughout Ethiopia. According to Pank- The Menilek and early Haile Selassie period (late hurst (15) 90% of the animal and 30 to 60% of the 1800s to around 1930) (Fig 3, Fig 4) witnesses the human populations were wiped out. During the same occurrence of a combination of events relevant to period, the southern regions were incorporated into forest decline in the study area including land tenure the modern Ethiopian state by King Menilek II (46). change, the establishment of garrisons and the intro- The incorporation launched migration from northern duction of coffee farming. In this period land tenure in Ethiopia to the south. Rubenson (16) argues that the southern Ethiopia changed from common property to two events are linked in such a way that the combined state control by the introduction of the institution of

3HASHEMENE

7ONDO'ENET

2OAD

Figure 4. Temporal and spatial aspects of human activities and the evolution of the cultural landscape/human activities including the process of forest decline during the study period.

Gessesse Dessie and Carl Christiansson

1 2 3 a b

10 km

Figure 5. Classified satellite images showing forests (black color) in the study area in 1972 (a) and in 2000 (b). The map was created using standard satellite image classification technique after acceptable accuracy was reached (5). In the eastern part of the study area where rainfall and soil are favorable forest was prevalent. This was not the case in the low-lying areas partly covered by lakes and wetlands. Similarly the western part is devoid of forest due to a combination of low rainfall, steep slopes and shallow soils. The speckle distribution of the forest in (a) compared to (b), is due to the expansion of smallholder agriculture and settlements. Number 1, 2 and 3 in (a) show commercial logging areas started in the 1930s by the Italians. The human impact on the forest cover in the study area predominantly expands from the foothill areas uphill and eastwards.

to interviews, coffee berry disease forced the majority land measurement (qalad) (46). The qalad introduced of farms to close down. the registration and classification of land for taxation purposes, and enabled the state to expropriate land A review of the period from the 1930s to the 1950s adjudged vacant. For taxation purpose three value reveals political events that resulted in a significant categories were created: cultivated, semi-cultivated, impact on forests in the study area. The Italian oc- and uncultivated, where forests were classified in the cupation 1936-1942 was of particular importance. category of least value (46). During this period commercial logging schemes and The land measurement transferred the major part agro industry activities were started up (Fig 3, Fig of the land (in southern Ethiopia up to two thirds) 4). Three well-equipped sawmills were constructed to state ownership (50). In this way, the land became within the study area (54). A dense logging road net- a source of state revenue, a source of income for work was built (55). Logging was carried out without regional governors and a means of payment for indi- replacement planting or regeneration schemes. During vidual soldiers. According to Wood (51) the changes the same period a tomato farm and processing plant in land ownership and the introduction of tenancy were established (56). By the early 1950s a significant lowered the standard of living of the rural population portion of the forest had been cleared and peasant throughout the south and undermined the economic farmers began to occupy the logged sites (57). At security, which the former socio-political structures that time, royal engagement in the area guaranteed had afforded them. a high level of security. This in turn attracted the During this period coffee was promoted from its interest of buyers and leasers, including European status as a wild berry picked within the boundaries buyers interested in establishing a foothold in coffee of natural forests (1800-2200 meters above sea and dairy farming. level) to an arable crop of significant economic importance (47, 52, 53). This development contributed In the period spanning from the 1950s to 1974 land to forest decline in the study area, directly by conver- ownership was redefined once again by promoting ting forest to coffee farms and indirectly by attracting privatization through land grants to civil servants more people to the area, mainly in need of land for the and war veterans. During this period private owner- cultivation of subsistence crops. The land considered ship was the dominant form of land ownership (58). most suitable for coffee was the virgin forestland on Mechanized farming became increasingly attractive the lower slopes of the escarpments. Coffee cultivation (59). Between 100 and150 commercial farms of the continued to expand until the 1960s, when, according size 80 to 200 ha were established in the Shashemene,

Forest decline and its causes in the south central Rift Valley of Ethiopia

Awassa and Wondo Genet areas (60) (Fig 4). These of change is the introduction and prioritization of farms attracted large numbers of labor migrants and certain crops. McClellan (18), Benti (40) and Gessesse contributed to settlement expansion (40, 60, 61, 62,). Dessie and Kinlund (8) have shown how the emer- Establishment of the new farms in the study area gence of coffee, haricot beans and khat (a stimulating contributed to the eviction of small-scale farmers at a drug/narcotic plant) intensified changes by promoting rate of 100 to 200 families per year (60). Some of the economic activities, establishment of new markets, evictees, according to Benti (40), moved to forests. immigration and settlement, all of which factors have contributed ultimately to forest decline (Fig 6). The period from 1974/75 to 2000 discloses two important political transformations: Radical changes in land tenure and the formation of new local Patterns of immigration and their influence on the administration units (peasant associations). The process of forest decline land reform of March 1975 encouraged smallholder farming by abolishing commercial agriculture at the The transformation of the study area from sparsely same time as the peasant association system intro- populated forest/woodland to a densely populated duced a totally new version of land administration agricultural landscape is explained by different as- (63). The transition from Haile Selassie’s imperial rule pects of population movements. to Military rule in 1974 was manifested by significant Several ethnic groups have migrated to the area social disruption to which unplanned forest clearance during the study period. The migration was a conti- may be attributed. This was also the case during the nuous process with peak periods during the southern transition in 1991 from Military rule to the present incorporation and expansion of commercial agri- EPRDF rule. culture. According to Makin et al (62) during the 1960s around 500 families per year migrated from The trend of change and the interaction of causes various parts of the country to the Southern Rift Valley area. The timeline of change (Fig 3) reveals a trend shaped The present population composition displays four by causes that may be described as episodic, gradual areas of origin and six ethnic groups. For example, and recurring. During the study period two important a categorization of households for Wosha peasant episodic changes resulting in fundamental ecological association in the study area shows that in 2003 the and socio-economic change in the study area are population was composed of 17% inhabitants of identifiable: the famine of the late 19th century and northern origin, 27% of western origin and 56% of the southern incorporation. Examples of gradual people with eastern and southern origin. The dif- changes in the study area include population growth ferent groups represent different land management and the expansion of agricultural acreage and settle- principles and technology. People from the north, ments, characterized by the colonization of previously such as Amhara, brought with them ox-plow agri- uncultivated land. Shashemene town grew by 1500% culture and introduced a cropping system based on between 1935 and 1965, 300% between 1965 and teff and cereals. This facilitated plowing of large 1980 and 100% between 1980 and 1994 (61, 64, areas and the expansion of arable agriculture. The 65). Awassa town was established in the early 1960s system partly replaced a predominately pastoral eco- (66). Over 10 years, beginning in 1984, the popula- nomy on the plains and competed with an enset and tion grew from 36000 to 63000 (23). The recurring root crop economy on slopes and in forests (18, 67, causes appear intermittently and often alter the course 68). The northerners introduced Christianity to the of change set by previous trends. The political changes area which, according to McClellan (47), may have are good examples of recurring causes. contributed to forest decline as new perceptions on Girma (41) and Melaku (9) have shown how nature were introduced. People originating from the periods characterized by a combination of violent east, mainly Arsi Oromo including Gujji Oromo, are change and political vacuum result in socio-economic cattle herders (pastoralists). People from the west, insecurity and depletion of natural resources. This from Kembata, Hadiya, Wolayita and Gurage, are is exemplified by the Munesa-Shashemene forestry enset and root crop cultivators, and people coming project in the northeastern corner of the study area, from south of the study area are dominantly Sidama which lost 71% of its forest during the government enset cultivators and animal herders. The reasons change of 1991. The results of our interviews point for immigration in general are explained as the at two incidents of forest destruction related to the quest for land and resources (northern origin), labor recent transitional periods: The interviews indicated migration (western origin) and a gradual expansion that between 1974 and 1975, transition to Militery from proximate areas (eastern and southern origin) rule, many farmers from surrounding villages settled (39, 40, 41, 60, 61, 69). The ethnic differences among in the Wondo Genet “protected” forest and in 1991 these people may explain conflicts over resources (41, during the most recent change in political power a 69). To counteract these differences so-called conflict large area of government owned forest was cut down management schemes were introduced that included and converted to farm land. the establishment of buffer zones (which could be Among other events that may alter the trajectory forest areas), organization of settlement schemes,

Gessesse Dessie and Carl Christiansson

Conflict management – improved security

Malaria control – improved health status Population growth Settlement expansion Dynamic trade – high economic migration

Natural calamities Emplyment – high labor employment (1880s) Forest decline

Southern incorporation (1890s) Economic policy change

Political conditions Colonialism (1930s) Agricultural expansion Sociopolitical changes Land ownership change Resource exploitation Road network – expansion Commersial logging

Local National

Figure 6. Causes of forest decline in a temporal and spatial context. modification of administrative boundaries and puni- important in order to determine the distribution and tive measures (41). the extent of the forest at the end of the 19th century. With regard to changes in power and influence the Provided rainfall conditions have not changed consi- interview results show a defined pattern of political derably, the recorded rainfall of 1200mm would have control and thereby control over resources. Prior been sufficient to support montane humid forests on to1975 the majority of the land was controlled by the eastern upland plateau, escarpments and foothills. individuals who originated from the north. Between Similarly 900 mm rainfall on parts of valley floors 1975 and 1991 individuals of southern and western and western escarpments would have supported origin dominated the political and administrative woodlands, tree savanna, bushland and grassland. leadership, while after 1991 they were replaced by Extensive areas that lacked forest/woodland at the individuals of southern and eastern origin. Other end of the 19th century were characterized by soils categories of actors, which have exercised a significant that are too shallow or acidic (such as the soils on the impact, are big landowners, large scale mechanized majority of the western escarpment) and obviously farm enterprises, logging projects and sawmills, all of areas covered by lakes and wetlands. This agrees which have evolved and disappeared during different with early descriptions of the forest in the area (22, periods. An actor that remains important and active 34, 35, 37). to date is the peasant association (grass roots level Analysis of the forest/woodland pattern as it administrative unit). appears on the earliest satellite image (from 1972) combined with the observable distribution of remnant trees over the landscape under study would seem to suggest the distribution of denser woody vegetation Discussion several decades earlier. It appears likely that much the same areas as indicated on the satellite image were The process of forest decline can be considered a forested in the beginning of the last century, but that human–environment interaction in time and space. It coverage was more continuous with fewer and smaller is an integral part of the evolution of the cultural (hu- gaps and with the fringes stretching further towards man made) landscape. With this in mind we discuss areas with less favorable conditions for forest growth. the process in the light of the natural characteristics of Consequently it is possible to argue that forest/wo- the area and the sociopolitical and economic processes odland covered some 40% of the study area while which influence forest decline. about 60% displayed more sparse woody vegetation The biophysical conditions in the study area were or no tree growth.

10 Forest decline and its causes in the south central Rift Valley of Ethiopia

Human activities gradually escalated over time. It is probable that the extent of forest decline can The descriptions of early travelers indicate both sparse also be related to the size and frequency of caravans human and cattle populations in the first years of the plying the trade routes. While little is known about 20th century. This is logical as famine and rinderpest the extent of traffic along the trade routes in the study had wiped out much of the human and animal life in area, it is assumed that the need for wood for fuel the 1880s and 1890s and because of this, vegetation or building materials, etc. was of a similar magni- around 1900 may have been denser and somewhat tude as along the 19th century trade routes through more widespread compared to conditions 20 years Tanganyika where deforestation was severe (73). earlier. Improvement in the 1930s to motor vehicle standard Following the southern incorporation people from of the old trade routes through the study area and the north, the majority of whom were cultivators, subsequent upgrading, in the 1960s, to tarmac road started to move to the southern Rift Valley. It is likely with feeder roads from agricultural centers shortened that they settled in the areas best suited for arable travel time to markets and increased load capacity. agriculture. These areas were the savanna woodlands This motivated an increase in cultivation of profitable, and the forest fringes. Steep slopes were presumably perishable or bulky crops. Clearing of remaining fo- avoided for arable fields, and wetlands avoided both rests became a strong economic incentive, illustrated for settlements and cultivation. This definitely had by the fact that the forests in the study area declined an impact on the forest vegetation, although the from 16% to less than 3% during the period from settlements were still few (42) 1972 to 2000 (3). Wetland environments were certainly of importan- Organized and unorganized immigration in con- ce as dry season grazing areas for the pastoralists. nection with the southern incorporation seems to have However, the extent of the lakes and wetlands seem led to conflict over land between state, newcomers and to have fluctuated widely between years as indicated the limited number of original inhabitants. After the by both early travelers and later researchers. This may annexation the state claimed that the land was state be attributable to rainfall fluctuation or other environ- property and could therefore be distributed to the mental factors such as tectonic movements (20). newcomers. It is probable that the forced eviction, The earliest descriptions of the wetland appear in in some areas, of the original inhabitants from land the topographic sketch by Maud (34) and in accounts suited for agriculture including grazing, to land of by Erlanger and Neumann (70). The former show marginal potential added to deforestation (40, 60). lakes Awassa and Cheleleka to be joined, while the Another interesting aspect emerges from an ana- latter report that these two lakes merged during heavy lysis of land tenure and the qalad system. One of the rains. If these descriptions were correct a substantial intentions of the qalad system was to expropriate part of the valley floor was either permanently covered unoccupied land. As the study area was only sparsely by water or seasonally inundated preventing growth occupied it is likely that it was identified for expro- of woody vegetation and creating a significant bree- priation by the state. A number of indicators exist to ding ground for malaria. The current wetland area corroborate such expropriation, for example, many is much smaller. Hodson (38) and Nilsson (71) have places still bear the names of the royal family members discussed the gradual desiccation of this wetland. and dignitaries to whom large areas of land were al- Benti (40) and Hammer (39) identified malaria as one located (40). The results of the interviews imply that of the reasons certain areas in this part of Ethiopia from the 1930s onward about 20 individuals owned were left unoccupied. the majority of the best agricultural land and the The location of the study area on one of the ma- forests in the study area. It seems that in the expro- jor trade routes in the region is also of significance. priation process the primary occupants were evicted During the southern incorporation many garrison and they resettled by the wetland and in forests on posts were established in the south (17, 47, 72); four the slopes of the escarpments (40, 60). were located in or next to the study area (Fig 4). As has been shown, Ethiopia has witnessed several These posts were important for the state to effectively dramatic political changes during the course of the last organize control of the south central Rift Valley due century. These changes have each been accompanied to four factors: 1) The area was situated on one of by transitional periods characterized by uncertainty the main trade routes (40); 2) It constituted an area and insecurity. In the absence of firm political control, with competing ethnic groups (41); 3) It presented control of resources has also been lacking. Many a topographically convenient passage between the among the rural population have then taken the opp- resource rich Sidama in the south and the central core ortunity to usurp what has been available in terms of, of Ethiopia; and 4) A large proportion of the area was for instance, forest land and forest products, adding suitable for settlement and economic development. to the process of forest decline. Examples of this from Additionally the eastern escarpment overlooking the the study area have been described above. Rift Valley floor provided a strategic location from In addition, each change in regime has implied far- which trade routes could be protected and ethnic reaching changes in the land tenure system. With new conflicts could be contained. It is also important to actors in power differing political (alongside ethnic) note that most of the garrison posts were subsequently principals have been enforced, resulting in a change converted into urban settlements (17, 47). in emphasis from one focus on land tenure/resource

11 Gessesse Dessie and Carl Christiansson control to another (communal-state-private-large reach today’s very high population density, has added scale-small scale etc). Whenever this has happened to the pressure on the forests. it has obstructed earlier long term planning efforts, The qalad system introduced in the beginning again leaving room for individual initiatives to usurp of the last century resulted in forests being assigned vacant land in the remaining forests of the study area such a low economic classification that in effect the or resulting in small holders being forced off their land status of forest was equated with “marginal land”. leaving them with no alternative but to clear of plots Land expropriations at different times during the in the forest as their remaining means to survive. 20th century forced smallholders off the land they The privatization process, for instance, dislocated had been cultivating with the result that a number a large number of people from farmland to land that of them cleared new land for subsistence farming in was often forest. This process further resulted in a the forests. small number of individuals gaining influence over During the 20th century, as repeated violent chan- vast areas where exploitation of resources was sub- ges of government have been interspersed with peri- sequently initiated (Fig 1b and Fig 1c). Commercial ods of political vacuum, an environment conducive farms, logging enterprises, sawmills etc. were esta- to the protection of resources has been lacking. As blished in the absence of any long term sustainability a result, competition over resources has proceeded plans. This inevitably led to the degradation of soils uncontrolled and large areas of forest have been re- and vegetation. Examples of such widespread forest moved. Furthermore, with every political change new decline have been mentioned above. land tenure principles have been introduced, creating At the same time the establishment of these uncertainty among farmers and obstructing long term enterprises created jobs and led to improvements planning efforts. This has been highly detrimental for in the infrastructure. This in turn led to sporadic the remaining forests. immigration into the study area and increased the Finally the successive upgrading of the old trade population density creating yet another impulse route to modern vehicle standard has meant that the for uncontrolled resource use. Forested areas were study area has become more accessible to distant cleared both for settlements and to provide arable markets. It has consequently become economically fields for inhabitants’ subsistence farming. It is this profitable to grow cash crops such as khat. The little process that accounts for the speckled pattern of fo- remaining forest land is thus becoming increasingly rests as shown in Fig 5b. When the system of private attractive as potential land for arable farming. The land ownership was subsequently removed and the original speckled pattern of scattered cultivations in former private land was distributed to large numbers the forests has developed into a pattern displaying few of smallholders and landless people this pattern was and small patches of forest remnants, which today further emphasized. cover below 3% of the land.

References Conclusions 1. McCann, J.C. 1999. Green Land, Brown Land, This study touches on the geographical complexity Black Land: An Environmental History of Africa, behind forest decline in the study area. Spatial and 1800-1990. James Currey, Oxford, 201p. temporal methods such as satellite image analysis and interpretation of historical texts have been used 2. Crummey, D. 1998. Deforestation in Wällo: to estimate the change in forest cover during the last process or illusion? Journal of Ethiopian Studies. 100 years. It has been shown that in addition to lack xxi, (1), 1-42. of governance during repeated periods of violent 3. Gessesse Dessie, and Kleman, J. 2007. Pattern political, social and economic changes uncertainty in and Magnitude of Deforestation in the South land tenure system and introduction of new profitable Central Rift Valley Region of Ethiopia. Mountain crops have had major impacts on the forests. Research and Development, 27, (2), 162-168 A century ago forests/woodlands covered a con- 4. Reusing, M. 1998. Monitoring of natural siderable part of the study area. The exact extent high forest resources in Ethiopia. Ministry remains unknown as quantitative documentation is Agriculture, Addis Ababa. 61p. lacking. Through extrapolation of information from satellite images from the early 1970s, combined 5. Hoben, A.1995. Paradigms and politics: the with material from other sources it can be estimated cultural construction of environmental policy however that the forests covered about 40% of the in Ethiopia. World Development. 23, (6), 1007- study area. 1021 Decline of the forest was ongoing already in the 6. Dessalegn, R. 2001. The landscape of develop- early 20th century. The successive immigration of a ment: A complex reality. In Food Security through farming population, beginning in connection with Sustainable Land Use, Population, Environment the “southern incorporation” and intensified during and Rural Development Issues for Sustainable commercial agriculture expansion in the 1960s to Livelihood in Ethiopia. Taye, A. (ed.). NOVIB,

12 Forest decline and its causes in the south central Rift Valley of Ethiopia

Addis Ababa, pp. 7-17 yata, Shalla and Awassa). Ethiopian Institute of 7. EFAP (Ethiopian Forest Action Program). 1994. Geological Surveys. Addis Ababa, 97p Ethiopian Forest Action Program Vol 1 and 2. 21. Kingham, T.J. 1975. Rainfall Records for the Ministry of Natural Resources Development and Southern Rift Valley, Ethiopia. Report no. 18, Environmental Protection, Addis Ababa. Minstry of Overseas Development, Land Resour- 8. Gessesse Dessie, and Kinlund, P. 2006. Khat ces Division, London. expansion and forest decline in Wondo Genet, 22. Russ, G.W. 1944. Reports on Ethiopian forests Ethiopia. Department of Physical Geography (compiled by Wolde-Michael Kelecha 1979), Ad- and Quaternary Geology, Stockholm University, dis Ababa, Ethiopia. Stockholm, Manuscript. 23. CSA, 1996. The 1994 Population and Housing 9. Melaku, B. 2003. Forest Property Rights, the Census of Ethiopia Results for Southern Nations, Role of the State, and Institutional Exigency: The Nationalities and Peoples’ Region. Central Sta- Ethiopian Experience. PhD Thesis, Swedish Uni- tistics Authority, Addis Ababa versity of Agricultural Sciences, Uppsala, 227p. 24. Yibeltal, G. 1995. Population pressure, agricul- 10. Angelsen, A. and Kaimowitz, D. 1999. Rethinking tural land fragmentation and land use: a case the causes of forest decline: lessons from econo- study of Dale and Shashemene Weredas Southern mic models. The World Bank Research Observer. Ethiopia. In: Dejene, A and Mulat, D (ed.) Ethio- 14, (1), 73-98 pian agriculture problems of transformation. 11. Contreras-Hermosilla, A. 2000. The underlying Proceedings of the 4th annual conference on the causes of forest decline. Centre of International Ethiopian economy Addis Ababa. . pp. 43-64. Forest Research (CIFOR), Bogor. 25. Angelsen, A. 1999. Agricultural expansion and 12. Moran, E.F. and Ostrom, E. 2005. Seeing the Fo- deforestation: modeling the impact of population, rest and Trees Human-Environment Interactions market forces and property rights. Journal of in Forest Ecosystems. MIT Press, Cambridge, Development Economics. 58, (1), 185-218. 442p. 26. Kaimowitz, D. and Angelsen, A. 1998. Economic 13. Geist, H. J. and Lambin, E .F. 2002. Proximate models of tropical deforestation, A review. Centre causes and underlying driving forces of tropical for International Forestry Research (CIFOR). deforestation. Bioscience. 52, 143-150. Bogor 138 p. 14. Lambin, E.F., Turner, B.L., Geist, H.G., Agbola, 27. Mather, A.S., Needle, C.L. and Fairbairn, J. 1998. S.B., Angelsen, A., Bruce, J.W., Coomes, O.T., The human drivers of global land cover change: Dirzo, R., Fischer, G., Folke, C., George, P.S., the case of forests. Hydrological Processes. 12, Homewood. K., Imbernon, J., Leemans, R., Li, 1983-1994. X., Moran, E.F., Mortimore, M., Ramakrishnan, 28. Tole, L. 1998. Source of deforestation in tropical P.S., Richards, J.F., Skånes, H., Steffen, W., Stone, developing countries. Environmental Manage- G.D., Svedin, U., Veldkamp, T.A., Vogel, C., and ment. 22, (1), 19-33 Xu, J. 2001. The causes of land-use and land- 29. Batterbury, S.P.J. and Bebbington, A.J. 1999. cover change: moving beyond the myths. Global Environmental histories, access to resources and Environmental Change. 11,(4), 261-269. landscape change: An introduction. Land Degra- 15. Pankhurst, R. 1985. The History of Famine and dation & Development. 10,(4), 279-289. Epidemics in Ethiopia. Relief and Rehabilitation 30. Robert, N. 1996. The human transformation of Commission, Addis Ababa, 120p. the Earth’s surface. International Social Science 16. Rubenson, S. 1991. Environmental Stress and Journal. 48, 493-510 Conflict in Ethiopian History: Looking for Cor- 31. Burgi, M. and Russell E. 2001. Integrative relations. Ambio, XX, (5), 179-182. methods to study landscape changes. Land Use 17. Donham, D.L. and James, W. 2002. The South- Policy. 18, 9-16. ern Marches of Imperial Ethiopia. James Currey, 32. Matthews, E. 1983. Global vegetation and land Oxford, 320pp. use: new high-resolution data bases for climate 18. McClellan, W.C. 2002. Coffee in Centre-Periphe- studies. Journal of climate and applied meteoro- ry Relations: Gedeo in Early Twentieth Century. logy. 22, 474-487 In: The Southern Marches of Imperial Ethiopia. 33. EMA (Ethiopian Mapping Authority). 1988. Ohio University Press, pp 175-195. Topographic maps number 0738C4, 0738D3, 19. FAO. 2005. Third expert meeting on harmoni- 0638A2, 0638B3. Ethiopian Mapping Authority, zing forest-related definitions for use by various Addis Ababa. stakeholders. Proceedings, Rome. 34. Maud, P. 1904. Exploration in the southern 20. Chernet, T. 1982. Hydrogeology of the Lakes borderland of Abyssinia. Geographical Journal. Region, Ethiopia (Lakes Ziway, Langano, Abi- 23,(5), 552-579.

13 Gessesse Dessie and Carl Christiansson

35. Neumann, O. 1902. From the Somali coast th- 51. Wood, A.P. 1983. Rural development and natio- rough southern Ethiopia to the Sudan. Geograp- nal integration in Ethiopia. African Affairs 82, hical Journal. 20,(4), 373-398. (329):509-539 36. Tewolde Berhan, G.E. 1990. Vegetation and 52. Pankhurst, R. 1964. The trade of central Ethiopia Environment of the Mountains of Ethiopia: Im- in the nineteenth and early twentieth centuries. plications for Utilization and Conservation. In: Journal of Ethiopian Studies. 2, 41-91 Messerli, B. and Hurni, H.(ed.). African Moun- 53. Merid, W.A. 1988. The early history of Ethiopia’s tains and Highlands: Problems and Perspectives. coffee trade and the rise of shawa. Journal of pp. 207-216 African History. 29, 19-25. 37. Logan, W.E.M. 1946. An introduction to the 54. Zerihun, M. 1988. Jiggessa and Sole sawmills and forests of central and southern Ethiopia. Imperial joinery 1939-1985 Shashsamane. BA Thesis, Ad- Forest Institute, Oxford. dis Ababa University, Addis Ababa, Ethiopia. 38. Hodson, A. 1922. Southern Abyssinia. Geograp- 55. Chaffey, D.R. 1979. South-west Ethiopia forest hical Journal. 53,(2), 65-79. inventory project. An inventory of forest at 39. Hammer, J.H. 1987. Humane Development- Munessa and Shashemane Project report 29. Participation and Change among the Sadama of Land Resources Division, Ministry of Overseas Ethiopia. The University of Alabama Press, 292 Development, UK. pp. 56. MWMVF (Melge Wondo Meat and Vegetable 40. Benti, G. 1988. A History of Shashemene from Factory).1996. Ethiopian privatization agency its Foundation to 1974. MA Thesis, Addis Ababa Melge Wondo Meat and Vegetable Factory (A University, Addis Ababa. profile). Addis Ababa, 15p. 41. Girma, N. 2001. The Arsi and the Sidama: A 57. Mooney, H.F. 1954. Report on forestry in History of Inter-Ethnic Relations Along Their Ethiopia with special references to the forests of Borderlands, Ca 1900-1991. MA Thesis, Addis Arussi and Sidamo. Imperial Forestry Institute. Ababa University, Addis Ababa. Oxford 42. Falcon, N.L., Grove, A.T., Schove, D.J., Banister, 58. EEA/EEPRI (Ethiopian Economic Association/ K.E., Kingham, T.J., King, R.B., Lamb, H.H., Ethiopian Economic Policy Research Institute). Kent, P. and Peters, E.R. 1975. Former lake levels 2002. Land tenure and agricultural development and climatic change in the rift valley of southern in Ethiopia. A research report, Addis Ababa, pp. Ethiopia: Discussion. Geographical Journal. 157. 141,(2), 194-202. 59. Pausewang, S. 1983. Peasant, Land and Society. 43. Wellby, M.S.1900. King Menilek’s dominions A Social History of Land Reform in Ethiopia. and the country between Lake Gallop (Rudolf) Afrika-studien; Nr. 110. Munchen and the Nile valley. Geographical Journal.16, 60. Ståhl, M. 1974. Ethiopia: Political Contradictions 292-304 in Agricultural Development. Raben and Sjögren, 44. Harrison, J.J. 1901. A Journey from Zeila to Lake Uppsala. Rudolf. Geographical Journal. 18,(3), 258-275. 61. Bjeren, G. 1985. Migration to Shashemene. Ethni- 45. Gwynn, C.W. 1911. A Journey in Southern Abys- city, gender and occupation in urban Ethiopia. sinia. Geographical Journal. 38,(2), 113-139. PhD Thesis, University of Stockholm. 46. Bahru, Z. 2002. A History of Modern Ethiopia. 62. Makin, M..J., Kingham, T.J., Waddams, A.E., James Currey, London, 300p. Birchall, C.J. and Teferra, T. 1975. Development 47. McClellan, C.W. 1984. State transformation and prospects in the southern rift valley, Ethiopia. social reconstitution in Ethiopia: the allure of the Volume 1&2 No 21, Land Resources Division, south. International journal of African historical Ministry of Overseas Development, London. studies, 17(4): 655-675. 63. Teketel, A.K. 1998. ’Tenants of the state’ The 48. McCann, J.C. 1995. People of the Plow. An limitations of revolutionary agrarian transforma- Agricultural History of Ethiopia, 1800-1990. The tion in Ethiopia, 1974-1991. PhD Thesis, Lund University of Wisconsin Press, Madison, ,298p. University, Lund, Sweden. 49. McClellan, C.W. 1990. Articulating Economic 64. Mesfin, W.M. 1966. Problems of Urbanization. Modernization and National Integration at the Proceedings of the third international conference Periphery: Addis Ababa and Sidamo’s Provincial of Ethiopian Studies, Addis Ababa, pp. 20-38 Centers. African Studies Review. 33, (1), 29-54. 65. Kloos, H. and Adugna, A. 1989. The Ethiopian 50. Getahun, D. 1974. Emperor Menilek’s Ethiopia Population: Growth and Distribution. Geograp- 1865-1916: National unification or Amhara hical Journal. 155, (1), 33-51. communal domination. PhD Thesis, Howard 66. Ababu, A.A. 1995. A History of Yergaläm Town University, Washington DC. and Its Environs from 1933 to 1974. MA Thesis,

14 Forest decline and its causes in the south central Rift Valley of Ethiopia

Addis Ababa University, Addis Ababa. ancient lakes and glaciers. Geografiska Annaler. 67. Smeds, H. 1955. The ensete planting culture. Acta 22, 1-79. Geographica. 13, (4), 3-39. 72. Akalou, W.M. 1973. Urban Development in 68. Shack, W.A. 1963. Some Aspects of Ecology and Ethiopia 1889-1935: Early phase. Journal of Social Structure in the Ensete Complex in South- Ethiopian Studies. 11, (1), 1-16 West Ethiopia. Journal of the Royal Anthropolo- 73. Christiansson, C. 1981. Soil Erosion and Se- gical Institute. 93, 72-79. dimentation in Semi-arid Tanzania: Studies of 69. Zerihun, M. 1999. Natural Resource Competi- Environmental Change and Ecological Imba- tion and Interethnic Relations in Wondo Genet, lance. Uppsala: Scandinavian Institute of African South-Central Ethiopia. MA Thesis, Addis Ababa Studies, Uppsala: Dept. of Physical Geography, University, Addis Ababa. Stockholm University. 70. Erlanger, B. and Neumann O. 1901. Baron Er- langer and Herr Neumann in Southern Abyssinia. The financial support for the present work was Geographical Journal. 18, (2), 214-215. provided by the Swedish International Development 71. Nilsson, E. 1940. Ancient changes of climate Cooperation Agency (Sida) and International Foun- in British East Africa and Abyssinia. A study of dation for Science (IFS)

Paper III

Khat Expansion and Forest Decline in Wondo Genet, Ethiopia

Gessesse Dessie1,2 and Peter Kinlund3

1 Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden 2 Wondo Genet College of Forestry and Natural Resources, Hawassa University, Ethiopia 3 Department of Human Geography, Stockholm University, Sweden

Abstract

This study analyses the expansion of khat production in relation to forest decline, by assessing spatial variables and social factors, using remote sensing and social survey techniques. The study focuses on the Wondo Genet area in south-central Ethiopia. The results indicate that khat has expanded rapidly in terms of land area used for its production at forest frontiers, in isolated forest patches and within farmland since the mid 1980s. This is mainly due to high economic advantage, high market demand, and favorable means of transport as well as the existence of a cohesive trade network. Moreover, the properties of the crop also facilitate expansion. The increased khat production appears to be a result of conscious choice and rational decisions made by male farmers, regardless of religious, cultural and policy discouragement and despite khat’s possible negative impact on livelihood security. Khat expansion does not explain forest decline in the study area per se, but it plays an important role in enhancing multifaceted interaction between people and forest. The expansion influences forest decline directly by conversion, and indirectly through increased human activity in proximity to forests. The conversion has resulted in a reduction of forest area, resilience and regeneration. Khat production has changed human settlement patterns, suppressed production of other crops and influenced women’s income negatively. These aspects increase the demand for wood and it renders the forest an important source of supplementary incomes. Khat production may create tension, resulting from a conflict in interest between sustaining the native forests, with subsequent environmental benefits for the larger social group, and the economically driven choice land use made by khat farmers.

Key words: Khat, Ethiopia, Wondo Genet, deforestation, cash crop, human-environment interaction

Introduction important sources of hard currency for the country (Feyisa & Aune, 2003; Dechassa, 2001; US Depart- In the Wondo Genet area in south central Ethio- ment of Commerce, 2000). Khat also influences the pia, located some 260 km south of Addis Ababa, a economic status of local areas. Gebissa (2004), for dramatic agricultural change is taking place. In this example, reports that khat production and marketing intensively cultivated area a new cash crop, khat, has have made the Harerge province in Ethiopia an ex- rapidly become a major source of income. This has ample of prosperity and economic development in an profound effects on the livelihoods of the local people otherwise poor country. However, khat is a narcotic as well as on the local environment. crop, which incurs negative health and social conse- Natural forest is an important, but diminishing quences. Hence, khat is illegal in several countries. part of the Wondo Genet environment. This montane However, while in Ethiopia the government takes forest, which nurtures an extensive biodiversity, is no action against its cultivation, trade, or use in any today an isolated remnant of once extensive forests form, it does not encourage cultivation (Feyisa & of the south-central rift valley. Today the forest covers Aune 2003). upstream areas and is in proximity to agriculturally Khat-related studies, in Ethiopia and elsewhere, productive valleys with a rapidly growing farming have so far mainly emphasized the chemical proper- population. Extensive deforestation as a result of ties, the chewing culture and its social implications, human intervention is in progress. Khat production drug abuse, and marketing aspects (Brooke, 1960; may be an important aspect of this process. Dagne, 1984; Varisco, 1986; Selassie & Gebre, Khat (Catha edulis Forskal) is an evergreen tree 1996; UNODCCP, 1999, Al-Motarreb et al. 2002). cultivated for the production of fresh leaves that are Few studies deal with other issues. Anderson et al. chewed for their euphoric properties. (For detail de- (2003) are researching global khat consumption scription see Table 1) Khat is a high value cash crop in Ethiopia. Recently it has become one of the most More references on this subject are available at Project Cork, 2006.

Gessesse Dessie and Kinlund, P. 2007. Khat Expansion and Forest Decline in Wondo Genet, Ethiopia. Geografiska Annaler (submitted). Gessesse Dessie and Peter Kinlund

Table 1. General and specific characteristics of khat as a plant, agricultural crop and stimulant drug. The specific characteristics including ecological adaptability, methods of cultivation, harvesting, packaging and transport refer to the situation in Wondo Genet while the others refer to general characteristics of the plant.

Identification and geo- √ Botanical classification and scientific name: Family: Celastraceae, Genus: Catha, Species: graphical distribution edulis, Catha edulis Forskal √ Local names: Khat (qat, cat, kat, chat, tschat) (Yemen, Ethiopia), qaad or jaad (Somalia), miraa (Kenya), tohai (Ethiopia), mdimamadzi (South Africa).The leaves are also called Bush- man’s tea (South Africa), Abyssinian tea, African tea, Arabian tea as well as African salad. √ Geographic distribution: Ethiopia, Somalia, Kenya, Malawi, Uganda, Tanzania, Congo, Zam- bia, Zimbabwe, South Africa, Madagascar, Yemen, Afghanistan.

Description of the plant √ Khat is a flowering evergreen tree or large shrub that reaches heights over 30m. In cultivation and chemical content of it is usually pruned to less than 3 m. the leaves √ The leaves are coriaceous and up to 10 cm long and 2-5 cm wide. Fresh khat leaves are dark green or crimson-brown and glossy. √ The active ingredients of khat are cathinone and cathine, which produce psychotropic, euphoric, metabolic and cardiovascular effects similar to amphetamine. The stimulant effect of khat is most effective when the leaves are still fresh. Leaves less than 48 hours old are preferred to ensure a maximum potency of cathinone. However, Khat can be preserved by freezing.

Ecological adaptability √ Grow on 1500-2600 m.a.s.l. altitude, 18-29oC temperature, 900-1200 mm rainfall, drained soils. Susceptible to frost, less so to disease and insects than other highland crops.

Methods of cultivation √ Propagated by cutting, can be planted year round. Planting stock 2-3plants/square meters. Often intercropped but also cultivated as a single crop (monoculture). √ Fertilized (often) by manure, insect protection by DDT spray 1-2 times per crop/year. √ Tending operation are watering, weeding, manuring, topping, leaves pruning and stumping every 8-10 years. √ First harvest within 2 to 3 years. Will remain productive for over 75 years at the rate of 2 to 4 harvests per year.

Harvesting and packag- √ All harvesting is done by hand, with each branch picked individually. Harvesting time is often ing and transport early in the morning or late in the afternoon. √ The leaves usually wrapped in enset/banana leaves and also packaged in plastic bags to re- tain its moistness and freshness, often sprinkled with water during transport to keep the leaves moist. √ Due to perishability of the leaves, fast transport to markets and speedy delivery to consumers are crucial.

Use and effects √ The most favoured part of the leaves is the young shoots near the top of the plant. √ Khat is taken orally most commonly chewed and stored in the cheek: by filling the mouth with fresh leaves the user then chews intermittently to release the active components. Khat is usually chewed in company as a socializing drug and for recreation. Khat could also be prepared as a tea. √ It is consumed to enhance working capacity (farmers, labourers) and to improve attention (drivers, students). Khat chewing results in excitation, banishes sleep, dispels feelings of hun- ger and fatigue. Khat lifts spirits, sharpens thinking, and, when its effects wear off, generates mild lapses of depression. Khat can induce manic behaviors, psychosis and hallucinations. There appears to be no specific physical addiction to khat; however, there is a psychological dependence.

Drug level and legality √ According to the World Health Organization (WHO) khat is classified as an abusive drug that can produce mild to moderate psychic dependence. √ Khat is illegal in many, but not all, countries in western Europe and North America √ In Ethiopia it is not illegal, but its use is prohibited in civil servants rules of conduct as well as in university campus rules

Khat Expansion and Forest Decline in Wondo Genet, Ethiopia trends, prohibition debate and economical factors. tives of ecology and political economy to investigate Milich & Al-Sabbry (1995) discuss khat cultivation underlying contexts and processes of environmental in Yemen from the perspective of peasant rational and change in relation to the problems of conservation sustainable livelihoods for the peasant community. and the development of land based resources in a In Ethiopia, Gebissa (2004) has studied the contri- variety of biophysical and cultural settings. Of central bution made by khat to both household and regional importance in this study is the use of remote sensing economies, while Feyisa & Aune (2003) examine the and geographic information systems to link social effects of khat expansion on the farming system, as processes of change with spatial and environmental well as economic and social conditions. Both of these change. Useful in this context is the people and pixel Ethiopian studies, however, are from the eastern part approach (see e.g. Liverman et al. 1998; Walsh et al. of the country, Harerge, where khat has been a cash 2002 and Fox et al. 2003). crop for more than half a century. In Ethiopia where forest is a very limited resource, The environmental impact of khat expansion, es- forest decline is a serious problem. Agriculture, par- pecially dealing with forest cover change has not yet ticularly small scale, is the mainstay of the majority been studied. Thus, the focus of this paper is on the of the people, and smallholder agricultural expansion interaction between khat production and decline of and forest decline are strongly linked. Therefore, the the natural forest in the Wondo Genet environment. trends and patterns of agricultural expansion are Diminution of the natural forest has been in progress partly the result of change in farmers’ livelihood stra- in the study area for a long time. During the past 30 tegies, which in turn are a response to opportunities years this has mainly been due to agricultural expan- and or limitations arising during different periods. sion (Gessesse Dessie & Kleman 2007). The forest High value cash crops persuade farmers to change in the study area is one of the few patches with high their strategies. Such crop production initiates a environmental value remaining in the south central special farmer-environment interaction that is crop rift valley of Ethiopia. In the latter part of this 30 year specialised, economically motivated and market de- period, the economic importance of khat production pendent. At the local scale farmers tend to adapt their has grown substantially. Khat farms have been expan- land use systems in response to economic advantages ding spatially over time and the expansion is occur- while market and economic forces operating at larger ring both within farmlands and along the retreating scales and levels determine such economic advantage. frontier of the natural forest. Such interaction links changes in the local environ- Forest decline is a complicated process caused by ment to wider scale economic, policy and market multiple and interrelated factors and it is extremely forces. Cash crop production is strongly correlated to difficult to give a comprehensive picture of the proces- broader political, economic and market factors such ses involved (see e.g. Moran & Ostrom 2005). Too as demand, pricing and infrastructure (Grossman, often it is explained by oversimplified parameters 1993; Nestel, 1995; Govereh & Jayne, 2002; Emich, such as overpopulation, overconsumption or poverty. 2003; Keys, 2004). In this perspective, the factors that Differences in interpretations are also related to the encourage expansion of cash crops may cause forest diffuseness of the classification and measurement of decline (VanWey et al. 2005). forest decline. The contribution that khat production Khat, as a stimulant drug, presents an interesting makes to forest decline may be identified in both the case in the study of forest change, particularly in re- physical conversion of forests in relation to its ex- gard to farmers’ rationale to cultivating it, its input pansion, as well as its promotion of social processes to livelihood improvement and the social processes that are the underlying causes of forest decline. Thus, it steers. A study of khat production requires a focus in order to study the role of khat in the context of on both underlying and direct causes of forest decline forest change, it is useful to adopt an approach that (Geist & Lambin 2002; Contreras-Hermosilla 2000; combines “attention to local specificity with an ap- Angelsen & Kaimowitz 1999). Khat production is preciation of the wider political and economic context a profitable activity that may benefit a large number in shaping decision-makers’ behaviour” (Jones & of people, but it is also highly dependent on market Carswell, 2004:204). This study, therefore, adopts conditions. It can, on the one hand, improve the a perspective of political ecology (see e.g. Blaikie livelihood of the farmers, while on the other hand & Brookfield 1987; Bryant & Bailey 1997; Stott & expose the farmers to the risks imposed by market Sullivan 2000; Zimmerer & Bassett 2003; Robbins and economic forces operating at various scales 2004; Neumann 2005). Political ecology can be seen and levels. Moreover, as a profitable activity, it can as a multicolour approach that integrates the perspec- persuade farmers to alter their farming systems.

Various definitions of the term forest decline exist. A relevant Underlying causes are according to Geist & Lambin (2002) one is given by Contreras-Hermosilla (2000): Forest decline fundamental social processes, such as human population includes deforestation, forest degradation or a combination dynamics or agricultural policies that underpin proximate of both. Deforestation is a reduction of tree crown cover and causes. Proximate causes, in turn, are local scale human decline in area for long period of time. Forest degradation is a activities that have direct impact on forest cover. Intermediates qualitative change, which does not necessarily show a decrease between proximate and underlying causes, such as decision in the area rather a gradual reduction of biomass, and changes parameters and agents of forest decline are also important in species composition. (Angelsen & Kaimowitz 1999).

Gessesse Dessie and Peter Kinlund

3HASHEMENE

-T!BARO ,AKE ,AKE #HELELEKA !WASSA 7OSHA0! !WASSA

"ASHA

2ED 3EA

. !DDIS!BEBA %THIOPIA 3TUDY AREA

KM %

Figure 1. The study area: The Wondo Genet area (box) and Wosha PA in Awassa watershed in south-central Ethiopia.

Expansion along the forest frontier leads to land use particular interest is Wosha, a peasant association competition between khat farming and forestry where (PA) located in the northern part of Wondo Genet, the outcome is partly determined by the farmers’ about 12 km from Shashemene. The Wondo Genet evaluation of the comparative advantage of the two topography is characterized by rolling upland at land-uses. Therefore, an analysis of khat production 1700-2600 m.a.s.l where a third of the land is over can help to understand composite social, economic 2200 m.a.s.l and the major part of the area is steeply and market related causes of forest decline in progress sloped (>30%). The higher altitudes and steep slopes in the study area at different scales. This gives an support natural forests, while lower altitudes and opportunity to analyse forest decline as a process that gentle terrain consists mainly of farmland where a depends both on the consequences of the actual khat significant number of diverse natural on-farm trees production, as well as on the factors that promote grow. Most soils of the steep slopes are shallow khat production. (<40 cm) but at lower altitudes dark brown soils in The objective of this study is, thus, to understand deep alluvial sediments occur (Zewdu & Högberg the significance of khat production on people-forest 2000). The climate in Wondo Genet is characterized interaction in Wondo Genet, Ethiopia, using both re- by two rainy seasons: a long rainy season from July mote sensing and social survey methods, with a focus to September and a short rainy season from March on the extent of expansion, reasons for the expansion, to April. The mean annual rainfall is 1200 mm and farmers’ preferences and the consequences. mean annual temperature is 19ºC. Wondo Genet is one of the most densely populated areas in Ethiopia. Six major ethnic groups reside in the area and their The study area main source of livelihood is farming. The Wondo

The study area Wondo Genet is situated in the Awassa Peasant Association (PA) is a grass-root level administration watershed in the East African rift zone southeast of unit in Ethiopia with a defined spatial limit of approximately the town Shashemene (7º06’N; 38º37’E) (Fig 1). Of 800 hectares.

Khat Expansion and Forest Decline in Wondo Genet, Ethiopia

Genet environment is agriculturally fertile where ir- Methods rigation farming dominates in the flat and undulating areas. The dominant type of agriculture is smallholder Remote sensing perennial crop farming on holdings of less than half a hectare, on average. The major crops include enset, The spatial distribution of khat farms was surveyed khat, and sugarcane. using remote sensing techniques. This was achieved by analysis of a satellite image using supervised maximum likelihood classification. A Landsat Thematic Enset is a large, fibrous banana like plant containing an Mapper (TM) image from December (dry period) underground corm, a bundle of leaf sheaths that form the 2003 was used and bands 4,5,3 were combined to pseudostem, and large leaves. The corm and pseudostems discriminate khat farms from other vegetation types. provide its food value, when prepared as a starchy edible paste which is made into bread (qocho). Enset is only cultivated as The possible reflectance interference from other crops food in Ethiopia. Enset leaves are used in Wondo Genet to wrap was minimized by (1) selecting an image from the dry khat leaves to avoid desiccation. season when normally most other agricultural crops

Figure 2. Khat sites in Wondo Genet as of 2003: a) Khat landscape (circle 2, Fig 4. with clusters of iron roof houses of khat farms. Clusters of planted trees are seen around settlements. b) Khat farms in Wosha PA. In the foreground new mono-crop khat plots are visible (circle 1, Fig 4), natural forest is seen at the background. These areas shown were used as training samples for the remote sensing survey. (Photo by Gessesse Dessie)

Gessesse Dessie and Peter Kinlund

Table 2. Interpretation precision output based on 100 The coordinate points were saved in a dbf (data base independent GPS points collected during 2003 from khat format) attribute table. Subsequently, the table was plots categorized broadly. added (view-add event) to the images of 1972 and 2000 separately. The points falling on the forest were Farm properties Samples Correctly counted and proportions were estimated. interpreted Points % Social survey Crop stage Shrub/tree 20 18 90 Young 30 23 77 Structured survey. A structure survey was conducted in Wosha PA from December to April 2002 and Intercropping Mixed 10 7 70 January to May 2003. The PA was selected as a Mono-crop 40 35 87 unit of study for two reasons (1) PAs constitute relatively consistent spatial settings (though the exact boundaries are not clearly demarcated) that have existed since the mid 1970s, (2) they are recognized are harvested/dried, (2) the knowledge acquired of administrative units (grass root governments) and the ground conditions following extensive multi-year have an administrative organization and archives. field work in the area, (3) using a relatively high The spatial setting provides a framework within resolution Landsat TM image (30m pixel) where at which the pattern of khat expansion can be followed least two pixels constitute an average khat plot, and through time while the administrative setting helps to (4) repeating classification until the results concurred identify and select/approach farmers through official with known khat sites. channels. The Wosha PA in particular was selected At the canopy level, khat is composed of a relati- because khat production commenced recently in the vely smooth, evergreen canopy of dark green leaves. area and is expanding rapidly. Additionally Wosha The reflectance properties from different crop stages is conveniently located in a familiar and accessible and intercropping (secondary layer of other crops environment where fieldwork can be organized and grown in the khat fields) can vary due to variation in performed relatively easily. the density of khat leaves and the background provi- For the questionnaire survey, 140 households ded by other crops. At the shrub/tree stage khat plants were randomly selected from a list of households. are stocked densely and form a closed canopy. At an This number corresponds to approximately 10% of earlier stage, khat plants contain few branches and the total households in the PA. The respondents were leaves. As a result the space between individual plants is larger and the canopy closure is incomplete. In a khat producing landscape, the khat fields are One of the authors is based at Wondo Genet College of interspersed with small buildings, access roads, etc, Forestry in the study area. but the khat canopy is spatially dominant. The shape and the size of individual farms vary but overall they form an extensive and easily identified landscape type. Training samples were identified using GPS, 100 points were collected purposely from homogeneous khat areas. To encompass the signature variation caused by different growth stages, intercropping, etc, samples were collected from different age classes/crop stages (shrub/tree, young) and management/intercropping (mixed, mono-crop). See Fig 2 for types of khat farms. The precision of the classification was estimated using 100 GPS control points collected independently of the training points. Overall, 83 GPS control points (83%) agree with the result of image interpretation of khat farms (Table 2). The GPS points identified for precision estimation were also used to estimate the forest area replaced by khat farms (Fig 3). The points were compared with the area of forest decline recorded between 1972 and 2000. The images are Figure 3. Example of remote sensing classification output, independently created to estimate deforestation based ~3% of Wosha PA, The figure shows interpretation of on satellite image interpretation (Gessesse Dessie & forest converted to khat. Grey tone shows forest cover, Kleman 2007). The area of forest decline is estimated black indicates forest cover replaced by khat cultivation, by subtracting image 2000 from 1972 using IRDAS and white for other agricultural land uses including pasture. IMAGINE 8.6 modeller function. The GPS count The circled site is an example of conversion from forest to was performed in the ArcView GIS 3.2 environment. khat where khat has replaced forest.

Khat Expansion and Forest Decline in Wondo Genet, Ethiopia all household heads, and predominantly men. The ding the farmers’ reasoning, the value of khat, and respondents were asked to respond to a questionnaire trends of khat prices, and production organization. comprising questions within similar response cate- Discussions were held with both males and females in gories, which were evaluated and scaled to minimize order to ascertain an understanding of the difference variation in response. The questionnaire was designed in khat preference along a gender line. to reveal the trend of khat expansion quantitatively by Archival survey. Records relevant for the compre- employing questions in regard of farm management hension of khat expansion including files and reports decisions (farm size, land allocation) crops (food from the district government offices, Wosha PA office crops, cash crops) trees planted/removed (natural, and documents belonging to private traders were exotic), khat production (farm size, price) and pro- identified first. Information concerning population, gression of crop selection. A majority of the interviews list of households, land holdings, market supply and were conducted at farms. prices of khat were studied to establish temporal Qualitative survey. In addition to the quantitative trends. The data selected are consistent and can be survey, a number of qualitative techniques were em- counterchecked. However, in this paper data are used ployed in the study. These include group discussion, to achieve a generalized trend analysis, rather than key informant interviews, ranking, sequencing and an absolute quantity estimation. listings and transect walks. For group discussion and key informant interviews respondents were selected from among the farmers, PA leaders, development agents, khat traders and middlemen. The discussions The extent and expansion of khat and interviews were aimed to gain information regar- farms

Most agricultural activities are generally a male domain The satellite image interpretation (see Fig 4) shows an in Wondo Genet, while the women’s role is, to a large extent, extensive distribution of khat farms. The distribution restricted to domestic activities, petit-trade, wood collection etc. covers a wide altitudinal range spanning from 1700 The perspectives of women regarding khat were mainly obtained to 2600 m.a.s.l. (see contours).The distribution is to from the qualitative survey.

Mt. Abaro

Basha

1700 m

Figure 4. Satellite image (L71168055_05520030112) interpretation showing khat distribution in the 1900 m study area. The pattern of distribution reveals widely dispersed scattering in Wondo Genet and N 2400 m clusters in three major growing areas marked by 0 km 5 the circles. Circle 3 is not as continuous as 1 and 2 but includes noticeable clusters east and south 2600 m of Basha town.

Gessesse Dessie and Peter Kinlund

1000.0 4500.00 900.0 4000.00 Supply Ton 800.0 3500.00 700.0 Sell (x10000 birr) 3000.00 600.0 2500.00 500.0 ton 2000.00 Birr 400.0 1500.00 300.0 200.0 1000.00 100.0 500.00 0.0 0.00 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year

Figure 5. Khat trade and revenue in Wondo Genet 1991-2000. Source: Transaction records of khat traders. a large extent concentrated to three major clusters of 420 farmers (30% of 1400 total households in (see circles) in relatively sheltered valleys. Wosha PA) cultivating 105 hectares of khat (based The expansion rate and the increasing economic on the average figure of 0.25 hectares per farm) (Fig importance of khat has been analysed from various 6, dashed line). data sources (traders’ transaction records, Wosha PA In regard of the land allocation at the farm level, reports, and structured survey data) and for different the survey shows that farmers allocate more land to spatial scales that including Wondo Genet, Wosha PA khat production than to all other land uses combined and individual farms. (see Fig 7). Khat is grown on approximately 62% Over the ten year period 1991-2000, khat produc- (0.25 ha) of the farm area while living quarters, food tion in Wondo Genet increased by 180% in volume production and animal keeping occupy the remaining and by 800% in revenue (Fig 5). In 1991 khat was space. Within khat plots, farmers tend to remove trees cultivated in Wosha PA by only 10 farmers on barely 2 with wide crowns. On fields allocated to other land hectares of land (Fig 6). By the year 2000, 84 farmers uses, trees remain and are also sometimes planted. were cultivating khat on more than 27 hectares. In Boundaries separating the land of other farms usually 2003 our survey reveals that 30% of the interviewed consist of trees which may be planted or remaining farmers (n=140) were growing khat. An estimation forest trees. based on the results of the interviews reaches a figure

450 120 400 100 350 Farmers 300 Area (ha) 80 250 60 200 area farmers 150 40 100 20 50 0 0

1 2 3 4 5 6 7 8 9 0 1 2 3 9 9 9 9 9 9 9 9 9 0 0 0 0 9 9 9 9 9 9 9 9 9 0 0 0 0 1 1 1 1 1 1 1 1 1 2 2 2 2 year

Figure 6. Khat expansion in Wosha PA by area and number of farmers 1991 – 2003. Source: Wosha PA administration records (1991-2000) and survey result (2003).

Khat Expansion and Forest Decline in Wondo Genet, Ethiopia

Reasons for khat expansion in Table 3. Comparative advantage of khat cultivation in Wondo Genet. Wondo Genet Crops Frequency of harvest/year Birr/0.1ha/year The survey results show that the principle driving force behind khat farming expansion in Wondo Khat 2-3 harvest a year 5000-7000 Genet is the economic factor. Khat cultivation gene- Sugar- 1 harvest in 1-2 years 800-1200 rates an income 2-3 times a year, that is far superior cane to all other crops, e.g. 2500% to maize and 600% Enset 1 harvest in 2-3 years 500-800 to sugarcane from 0.1hectare of land (Table 3). This Maize 1 harvest a year 200-300 aspect is of particular importance to Wondo Genet farmers, whose land ownership is generally limited to less than half a hectare per household of 7-8 members. In Wosha PA land holdings per household decreased conditions, specifically demand, means of transport from 1.2 hectares in 1975 to 0.4 hectares in 2002, and trade networks. The qualitative survey reveals the while the household numbers increased from 400 to existence of strong demand indicators; the price of 1400, see Fig 8. khat in Wondo Genet increased 500% (9 to 45 birr per The results presented here concur with other bundle) within the 10-year period 1991-2000. More studies in Ethiopia and elsewhere, e.g. Gabissa then 20 market centres within 300 km buy Wondo (2004), and Feyisa & Aune (2003) in Harerge Genet khat consistently, to retail to the increasing Ethiopia; UNODCCP (1999) in Kenya and Varisco number of consumers. In general, both the number (1986) in Yemen cited the economic factor as the of khat consumers and consumption are increasing major propulsion behind khat expansion. However, in Ethiopia (Selassie & Gebre, 1996; UNODCCP, economic factors are not sufficient to explain the 1999; Dechassa, 2001; Gebissa, 2004). The market is timing of the expansion. also dependent on international stimuli. For example, Another important factor relates to market the ban of khat export from Kenya to Somalia in 1999 played a profound role in the development of Ethiopian production and export (Robinson, 1999). partition, mainly fruit trees Traders and consumers show particular interest in Wondo Genet khat, as indicated in Fig 9. A reliable road network and efficient means of Enset plot Living quarter transport are determinant factors for the efficient trade of the highly perishable leaves. In regard of roads, Wondo Genet is a favourable area because the farms and villages are well connected and the major southern highway is easily accessible. Means of transport suitable on feeder roads and highways exist, e.g. hundreds of horse carts operate to haul khat from farms to local markets and fast mini-trucks operates on highways to ensure speedy and reliable transport to distant markets, such as Addis Abeba. The significant role played by roads to khat expansion Khat plot in Wondo Genet concurs with the findings of other studies elsewhere. Verisco (1986) and Gabbisa (2004) reported khat expansion is strongly linked to the state of the road network and means of transport. The existence of a substantial link between road/transport and the change in crop production is 0 12 m well established. According to VanWey et al (2005) the types of crops planted are strongly determined by transportation costs while changes in external Live fence, mainly forest trees market values or the cost of transportation are reflected relatively rapidly in the spatial allocation of Figure 7. Sketch of a khat farm in Wosha PA showing av- agricultural activities. They have further stated that erage land allocation. The living quarters contains a house the level of deforestation as a result of agricultural and animal stalls, enset often mixed with vegetables, potato expansion will be directly related to the presence of and coffee bushes, while khat is grown unmixed with other roads and the transportation costs that farmers face crops including trees. In most farms partitions containing fruit trees exist. Around almost all farms live fences exist, It should be noted that very little khat is consumed locally. containing a mix of intermittent forest trees and hedges in Most farmers do not chew khat leaves. The road between Addis Abeba and Shashemene was rebuilt combination with trenches or stonewalls. in 2000

Gessesse Dessie and Peter Kinlund

1600 1.4

1400 1.2

1200 1 1000 0.8 800 0.6 600 Number of HH

400 0.4 Land holding (ha) 200 0.2 0 0

1974 1990 1996 2003 Number of HH Year Land holding

Figure 8. Changes in number of household and landholding in Wosha PA. Source: Office of WoshaPA administration.

for the crops that they may invest in (VanWey et al already mentioned, the plant grows under broad (2005). climatic conditions, and it can tolerate drought Another factor affecting the market is the system for long periods. It is furthermore less susceptible of the actors’ network. The qualitative survey reveals to disease and insects than coffee (Gebissa, 2004). the existence of informal but cohesive networks In all, khat suits the agricultural conditions of the that link actors involved in the khat production small-scale farmers in Wondo Genet, facing land and process, see Fig 10. The network functions with the labour shortage. purpose of matching the growing demand for khat However, factors that discourage khat production with cultivation. Consequently, farmers gain access also exist. Interviews with traders identify policy and to a reliable market that would otherwise prove religious restrictions. Government policy denies khat difficult. farmers participation in the agricultural extension Apart from the economic and market related package program to enhance crop productivity. The issues there are further factors which encourage government also requires khat traders to pay heavier khat production. The khat leaves can be picked tax (2-4 birr/kilo) than traders involved in other throughout the year. Consequently, the need for agricultural production. The Christian church also labour is distributed over the year and there are no discourages its members from khat cultivation, trade bottleneck situations at specific periods of time. As and consumption.

Figure 9. Marketing of Wondo Genet Khat in distant towns, a) Advertisement for Wondo Genet khat (arrow indicates an amharic ad “here you find the Wondo khat”) in Debre Zeit town 220 km from Wondo Genet, b) Khat retail in a kiosk in Modjo town 180 km from Wondo Genet: the upward arrow shows Khat leaves wrapped in bundles and the downward arrow show khat leaves on a weighing scale.

10 Khat Expansion and Forest Decline in Wondo Genet, Ethiopia

#OMMUNICATION LOGISTICANDSECURITY

-IDDLEMEN

+HATFARMERS +HATTRADERS

4RANSPORTERS +HATCONSUMERS

4RANSPORT SPEEDANDPROTECTION

Figure 10. Khat producers network. The middlemen facilitate the smooth flow of trade by ensuring communication between the farmers and traders, organizing logistics, including bulk purchase, storage and loading, insuring secure money transfer from traders to the farmers, and by giving farmers assurance and winning their trust. For these services the middlemen are paid by both traders and farmers. The transporters insure reliable and prompt transport to the consumers as well as protection to keep the leaves succulent and fresh during transportation. The prompt arrival of fresh khat at the market ensures the best prices. Drivers who reach their destination in time receive monetary bonuses from the traders.

Considerable arguments appear to exist for the soil erosion); its value/purpose (as food, fuel, fod- expansion of khat in the study area. However, it is der) and preference (i.e. the role of khat income in also important to take a closer look into the local the family economy). Farmers in general (men and people’s preferences in order to understand patterns women) give a high score to the general drawback of development. Explanations for the expansion of costs for khat (Table 4) in comparison to other food khat production may be found in the local setting as crops (enset, sweet potato) and cash crops (avocado). well as on the national and international scene. The reasons given by the farmers are that (1) khat incurs a high production cost (especially in respect of the cost of protection against possible looters), Local farmers’ preferences (2) khat prices fluctuate up to 50% seasonally, (3) disease may attack khat while the leaves are young The survey conducted in Wosha PA shows that far- (it is mandatory to chemically treat the new leaves), mers give a mixed assessment for khat in regard to and (4) new khat fields on steep slopes are prone to the general cost of drawbacks (production cost, price soil erosion (see Table 4 for khat properties). Farmers fluctuation, vulnerability to disease, contribution to in general give low rating for khat (no use as a food,

Table 4. Khat farmers’ views of crop preferences in Wosha PA.

Crops General drawback General value/purpose Men’s preference Women’s preference High Med. Low High Med. Low High Med. Low High Med. Low Enset √ √ √ √ Maize √ √ √ √ Sweet potato √ √ √ √ Coffee √ √ √ √ Avocado √ √ √ √ Sugarcane √ √ √ √ Khat √ √ √ √

11 Gessesse Dessie and Peter Kinlund

120

100 Enset 0 Sugercane Khat 0 Maize 0 Potato Number of farmers Teff 20

0 1 11 2002 Year

Figure 11. Change of crop preference by farmers. Source: Data obtained from structured interviews with farmers in Wosha PA. Crop history based on the number of farmers counted at three periods. Years 1985 (villagisation/resettlement) and 1991 (government change) are well recognized by all respondents while 2002 constitutes the interview period.

low value as fuel or fodder) when compared to other The study of the local preferences indicates that food and cash crops. Sugarcane rates highly because, economic factors are important in explaining the according to farmers, it is one of the main sources of expansion of khat in the area. However, it also animal fodder which is scarce due to the shortage of indicates that local decision-making is made by the pastureland. Enset is a highly regarded crop by both men who do not see or value the negative aspects. The men and women because (1) they consider it as one findings can be compared with the study in Harerge of the most important staple foods (2) they identify it (Gebissa 2004). There, the men’s views are similar to as disease and drought resistant, and (3) they regard the ones presented above, but many of the women it to be a good source of animal fodder. interviewed affirmed that khat indeed made them An investigation of the preference for khat in financially independent. comparison to other crops reveals differences in opinion between male khat farmers and their wives. The women give low preference scores for khat (Table 4) while men give high preference scores for khat and Consequences of khat production in other cash crops. The reasons given by the women Wondo Genet for their low appraisal are (1) that while men control the khat income, women who are responsible for Khat production influences the production of other feeding the family are not given adequate money to crops, the distribution of household incomes (men cover family expenditures (2) enset, dairy and some usually control the khat incomes) and patterns of fruit production, which constitute the domain of the settlement. This, in turn, influences the state of the women, is decreasing (3) men tend to marry additional forest in the area. wives and/or increase their alcohol consumption and One of the most significant impacts of khat, ac- get involved with prostitutes (thereby neglecting their cording to the farmers in Wosha PA, is a decline in families, and exposing themselves to HIV/AIDS). the production of food crops (Fig 11). The number The reasons given by men for their evaluation are of farmers growing food crops decreased significantly (1) the khat income makes it possible to buy food between 1985 and 2002. The number of maize gro- when their own production is not adequate, ( 2) the wers for example declined by 38%. For the period available land is not sufficient to produce food and from 1991 to 2002 alone the decrease was 58% (95 generate enough money to sustain the family, (3) to 40). The number of farmers growing crops such the khat income enables an improvement in living as potatoes and teff declined in number in a similar standards (better education, housing and clothing), manner. On the other hand, farmers who grew khat and (4) khat production requires less time and energy increased by 470% (10 to 57) between 1991 and (excepting protection against looters). 2002. This trend seems to substantiate the general

12 Khat Expansion and Forest Decline in Wondo Genet, Ethiopia picture that khat cultivation undermines food crop shaping concentrated market settlements. One typical production. This is, for example, Verisco’s view based example is Basha. This khat town grew from a small on studies in Yemen (Verisco 1986). Other studies village in the late 1980s to its present size with 7000 (e.g. Govereh & Jayne, 2003; Grossman, 1993), ho- residents when it became the major khat wholesale wever, argue that cash crop production can promote market in Wondo Genet. The growth of population food crop productivity. This may be the case for enset following cash crop production is not unique to khat. in the study area. Enset cultivation displays a different Other studies in Ethiopia reported that cash crop pattern and the number of farmers growing this crop booms have contributed to an increase in the number increased by 267% (30 to 110). Nevertheless, the qua- of the local population and concentrated settlements litative survey reveals a general perception of declining (Makin et al, 1975; Benti, 1988; McClellan, 1990). enset production in the study area. According to the interviewed farmers the amount of land allocated per family to enset cultivation has declined. One of possible explanations for this apparent anomaly could Khat production and the natural be that the increase in the number of growers should forest of Wondo Genet not be interpreted as an increase in total yield. Far- mers indicate that they produce enset for their own We have above discussed the extent of and driving consumption and that they can produce enough food forces behind khat production in Wondo Genet. We for annual consumption on a small area. Brandt et have also touched upon its consequences on popula- al. (1997) reports that enset that can be produced on tion and settlement. Let us now look into the role of about 500m2 area of land can feed an average family khat in the decline of the natural forest for a year. It is a general trend today in Wondo Genet The satellite image interpretation shows that the that enset is cultivated in residential quarters where natural forest in Wondo Genet declined from 14,270 planting enset does not interfere with other crops. ha to 2,270 ha between 1972 and 2000 (Fig 12). Khat Therefore, it may be argued that enset production has farms expand over upstream areas, appear along actually declined despite an increase in the number forest frontiers and are scattered widely within the of farmers growing the crop. One indicator is that forest. GPS point counts show that khat expansion the price of enset per unit has increased 1000% in contributes to about 30% of the natural forest 10 years time, according to information from local decline in the major khat areas (within the circles 1 markets. Thus, the increase in number of farmers who and 2 on Fig 4). The major khat areas are located in grow a reliable staple food crop in their backyard can sheltered valleys and on steep slopes. In addition to be taken as an indicator of farmers strategies to cope the permanent conversion of forest, khat contributes with the general decrease in food production. to the decline of forest resilience. This occurs as a Temporal analysis of crop preference (Fig 11) result of increased forest fragmentation due to wide shows a general trend of food crop decline while cash distribution (see Fig 4); biodiversity decline due to crops increase. clustering in sheltered valleys (see contours and the The economic benefits of khat production circles) where diverse and rare forest exist; and soil constitute a pull factor for the farmers outside Wondo erosion due to steep slope cultivation. It is also likely Genet. According to informants, people immigrate to that khat cultivation contributes to stream pollution khat areas from the neighbouring highland areas in due to the combination of upstream cultivation order to participate in khat cultivation, daily labour erosion and the use of pesticides and manure. Figure and trade. Analysis of archive records from Awassa 12 also shows a small amount of forest gain, in the finance office reveals that the population density in form of plantations by forestry projects (mainly major khat growing areas is 625 -1125 person/km2, west and south of Mt Abaro) and homestead trees which is much higher than in nearby non-growing established by farmers. The latter is significant in areas: 180-440 person/km2. khat growing areas particularly in the area of circle Standing on high ground one can see, on the slopes 2 (see also Fig 2a for clusters of planted trees near of Wondo Genet, widely scattered numerous clusters settlements). It is likely that khat expansion is one of of iron sheet roofs glittering in the sun. The majority a combination of factors contributing to this. of these are the recently constructed houses of khat Interview analyses reveal indications that further farmers, situated next to their plots. Khat expansion support the remote sensing outcome regarding khat in Wondo Genet is accompanied by a change from expansion within forests. The PA leaders claim that the nuclear to more dispersed settlement patterns illegal forest settlement has increased substantially (see Fig 2a). One of the main reasons for a dispersed during recent years, mainly by khat farmers. According settlement pattern is that khat is longstanding (up to to these sources the high influx of settlers is becoming 70 years) and farmers permanently reside near their a serious threat to the forest and hence it has become farms. According to farmers in Wosha, this is due a regional issue. As a result the regional government to the consistent need to protect khat leaves from has been involved in controlling settlement using potential looters. various means, including the military. While khat production contributes to a more The interviews also reveal that khat farms are dispersed settlement pattern it also plays a part in claiming areas previously considered unsuitable

13 Gessesse Dessie and Peter Kinlund

Figure 12. Forest decline between 1972 and 2000 in Wondo Genet based on satellite image interpretation. The circles indicate the major khat expansion areas (see also Fig 4) since the early 1980s. The forest gain is mainly of exotic tree species. The forest gain west and south of Mt Abaro is due to large-scale plantations planted by the Wondo Genet College of Forestry and a forestry project. The other scattered forest gain areas are farmers’ established homestead trees clustered around settlements. The classification ‘Other areas’ (grey) includes areas with no visible N Forest decline area 0 km 5 changes, including remnant forests and all other Forest gain areas land cover type changes/no changes. Other areas

for cultivation due to ground conditions, location a significant amount of khat farms have replaced near springs or stream banks, steep slope gradients forest areas. The results show that khat expansion is and/or prevalence of vermin. Such areas are often driven by land scarcity, as well as market, transport characterized by natural vegetation particularly and, above all, economic factors. Khat production clusters of old trees. It seems, thus, that khat also influences the trend of immigration and human production in the area increases land value. In general, settlement, which indirectly affects the state of the farmland prices in Wosha PA during the past decade forest. However, the role of khat production is a very have increased by 600% (500 birr to 3000 birr per complex issue. Khat is both a direct and an underlying 0.1ha). This increase can be related to the increasing factor in the variation in the extent of the forest in scarcity of farmlands, but the role of khat can also be Wondo Genet. seen in the price increase of poor lands (unsuitable for Khat is an undemanding crop displaying wide other crops, distant from access roads, steep slopes ecological adaptability. It enjoys favourable market or rough surfaces). According to the informants the conditions and, most importantly, it gives a high price of inferior land has increased on average by return per small area. These advantages are espe- 480% (250 birr to 1200birr per 0.1ha). cially important to farmers whose land holding are too small to yield sufficient income or to produce adequate amounts of food. It may be argued that Concluding discussion the fact that khat farming enables a large number of smallholder farmers to earn a good income on small The study has shown that khat production has plots may limit the tendency of agricultural expansion increased rapidly during the last two decades and into forests. However, the increasing population and

14 Khat Expansion and Forest Decline in Wondo Genet, Ethiopia the immigrant farmers still require new land. The them vulnerable to variations in market elements such comparative economic advantages have attracted as price fluctuation. Consequently, while the market farmers from other regions, resulting in increased prospers khat remains profitable and khat production migration to the study area. continues to expand. When the market stagnates far- Several studies (see e.g. Tiffen & Mortimore, mers may lose their principal means of income and 1994; Templeton & Sherr, 1999) show that popu- under such circumstances they may have to turn to lation growth results in an increase in the cost of the forest to obtain a supplementary income. land relative to labour, and as a result people tend Despite the general picture of khat as a source to improve crop and animal management, and to of high income, the study also indicates its negative initiate land improvements. One difference in the consequences for women’s income and food produc- case of khat is that the increase in value is not only tion, as has been shown in a large number of studies in the existing farmland, but also in lands previously from different parts of the world (see e.g. Bernstein deemed unsuitable for cultivation, such as slopes, et al. 1992). Women may not benefit from the khat stream banks, rough surfaces, and even vermin prone incomes and there is evidence of diminishing food areas. Consequently, khat has expanded in these production in the area. Increasing food prices in the areas indiscriminate of location and regardless of local markets also indicates this. The shrinking of their land holding size. It has also encouraged per- women’s incomes, decreasing food production and manent residence near khat fields. In this perspective growing in-migration together increase pressure on khat promotes both intensification within farms and forests by creating demand for more wood and iden- expansion. tifying forests as supplementary income source. The expansion on both farmlands and within Despite the rapid expansion of khat production forests has caused the removal of on-farm trees and in Wondo Genet, it still occupies a relatively small groups of trees, and the conversion of forestland. Mo- area and involves few farmers in comparison to the reover the expansion of khat farms has resulted in the total land size and population of the area. However, establishment of permanent farmers’ settlements, in its economic advantage is superior to all other crops close proximity to the forest. This contributes to forest grown in the area. Therefore, the issue of khat’s degradation through animal grazing, selective cutting, contribution to forest decline is more complex than charcoal burning, and fire. Additionally, khat farming simply the area it covers and the farmers it involves. has promoted forest fragmentation, which in turn has At present economic advantages make khat farming contributed to degradation of forest resilience. favourable for the individual farmer while maintain- Expansion is linked to the conscious choice and ing the native forests is not. The native forests offer rational decision by the (male) farmers. The study in- environmental benefits to the larger social group, dicates that farmers cultivating khat are well aware of which are not immediately advantageous to the in- the possible drawbacks of cultivating a controversial dividual khat farmers. Thus, tensions exist and may crop with limited other purposes, i.e. the disincentives increase between the local farmer making his econo- that are imposed by religion, high taxes and the denial mically driven decision and those concerned for the of extension services, as well as being at the cost of maintenance of the native forest for long-term benefit crop diversity by promoting khat as a monocrop. to the society as a whole. From this perspective, khat precipitated forest decline This study has shown that khat production does may be regarded as the result of a conscious decision not explain forest decline in Wondo Genet but it indi- made out of necessity. cates that khat plays an important role in this change. The manner of khat expansion in Wondo Genet This study also shows that a picture of deforestation is similar to the growth of the crop in Yemen. Milich cannot be based on a simple causal explanation. It & Al-Sabbry (1995) argue that farmers in Yemen is necessary to apply a broad-based approach that base their decisions on the fact that they are able to encompasses a variety of scales, methodologies and earn far more by growing khat than by growing any conclusions in the study of land use and land mana- other cash crop. Khat may also be compared with the gement. expansion of opium poppy production in Mexico. However, further research is required to position Bucardo et al. (2005) show that poppy farming in the impact of khat expansion in a wider socio-eco- Mexico has escalated during the past decade, in spite nomic perspective and with a view to revealing its of limited government assistance to farmers, due to its impact on landscape transformation, particularly high economic benefit which coincides with the fall of research that combines spatial information with socio- coffee and maize prices. Consequently, an increasing economic data on various levels. In regard to forestry, number of poor mountain farmers have changed their further studies are needed to understand the conflict crops from coffee to poppies. between the negative implications for the environment Economic factors promote khat expansion. Mar- resulting from forest decline regarded partly as a ket forces operating at various scales strongly dictate result of khat expansion, and the economics of khat the actions of the farmers. On the one hand, the expansion. Risk minimising and coping strategies dynamic commerce and strong economic advantages are of particular interest. The economic benefits of persuade farmers to continue producing khat. On the khat farming, as indicated above, need to be further other hand, increasing dependence on khat can make studied, especially with a gender perspective. Gross-

15 Gessesse Dessie and Peter Kinlund man (1993), for example, has shown that an export Ecology. Routledge, London crop can actually encourage local food production. Bucardo J, Brouwer K C, Magis-Rodrigues C M, Ramos Govereh & Jayne (2002) argue that the production R, Fraga M, Perez S G, Patterson T L & Strathdee of cash crops offer the farmers the potential to gain S A (2005). Historical trends in the production and access to credit markets and other resources that can consumption of illicit drugs in Mexico: Implica- contribute to increased food crop productivity. Ho- tions for the prevention of blood borne infections. wever, it is also argued that a result of the commercia- Drug and Alcohol Dependence 79: 281-293 lisation of agriculture is that “the rich and powerful Chambers, R (1983): Rural Development: Putting the get richer and more powerful; and the poor become last first. Longman, London relatively and often absolutely poorer and weaker” Contreras-Hermosilla, A (2000). The underlying causes (Chambers 1983:38). Thus, the role of khat in poverty of forest decline. Centre for International Forestry alleviation, which is often seen as a focal issue in the Research (CIFOR), Occasional paper no. 30. Bo- aim to improve environmental conditions, is still an gor. open question. Gebissa’s (2004) claim that khat gives Dagne, E (1984): Proceedings of the International Sym- prosperity and economic development to the area of posium on Khat (Catha Edulis) Chemical and Eth- cultivation does not, as yet, apply to the Wondo Genet nopharmacological aspects. Addis Ababa, 15th De- area, neither for the majority of the local people, nor cember 1981. Addis Ababa. for the environmental situation. Dechassa, L (2001): Khat (Catha edulis): Botany, Dis- tribution, Cultivation, Usage and Economics in Ethiopia, UN-Emergencies Unit for Ethiopia, Ad- Acknowledgements dis Ababa. Emich, M (2003): The Human Ecology of Mayan Cacao The authors wish to thank Professor Johan Kleman Farming in Belize. Human Ecology, 31(1): 111- Professor and Carl Christiansson for their construc- 131. tive comments and important inputs to this paper. Feyisa, H & Aune, J B (2003): Khat expansion in the Ethiopian Highlands, effects on the farming system in Habro district. Mountain Research and Develop- ment, 23(2): 185-189. References Fox, J, Rindfuss R R, Walsh S J, & Mishra V (2003): People and the Environment. Approaches for link- Al-Motarreb, A, Baker, K & Broadley, K J (2002): Khat: ing household and community surveys to remote Pharmacological and Medical Aspects and its So- sensing and GIS. Kluwer Academic Publishers, cial Use in Yemen. Phytotherapy Research, 16: Boston 403-413. Gebissa, E (2004): Leaf of Allah. James Currey, Ox- Anderson, D M, Hailu, D, Klein, A C & Beckerleg, S ford. (2004): The Khat Nexus: Transnational consump- Geist, H J & Lambin, E F (2002): Proximate causes and tion in a Global Economy. Available at: http:// underlying driving forces of tropical deforestation. www.consume.bbk.ac.uk/research/anderson_full. Bioscience 52, 143-150. html, accessed 12 March 2006. Gessesse Dessie & Kleman J (2007): Pattern and Magni- Angelsen, A., Kaimowitz, D (1999): Rethinking the tude of Deforestation in the south central rift valley causes of deforestation: Lessons from economic region of Ethiopia. Mountain Research and Devel- models. The World Bank Research Observer 14 opment, 27 (2):162-168. (1): 73-98 Govereh, J & Jayne, T S (2003): Cash cropping and food Benti, G (1988): A History of Shashemene from its crop productivity: synergies or trade-offs? Agricul- Foundation to 1974, Addis Ababa University, Ad- tural Economics, 28: 39-50. dis Ababa. Grossman, L S (1993): The Political Ecology of Banana Bernstein, H, Crow, B & Johnson H (eds.) (1992): Rural Exports and Local Food-Production in St-Vincent, Livelihoods. Crisis and Responses. Oxford Univer- Eastern Caribbean. Annals of the Association of sity Press, Oxford. American Geographers, 83(2): 347-367. Blaikie, P & Brookfield, H (1987): Land Degradation Jones, S & Carswell G (2004): Conceptual Frameworks and Society. Methuen, London. in Environment and Development. In: Jones, S & Brandt, S A, Spring, A, Hiebsch, C, McCabe, T Tabogie, Carswell G (eds.): Environment, Development & E, Diro, M, Wolde-Michael, G, Yntiso, G, Shigeta, Rural Livelihoods, Earthscan, London. M. & Tesfaye, S (1997): The “Tree Against Hun- Keys, E (2004). Commercial Agriculture as creative ger” Enset-Based Agricultural Systems in Ethio- destruction or destructive creation a case study pia. University of Florida. Also available at: http:// of Chili cultivation and plant-pest disease in the www.aaas.org/international/africa/enset/enset.pdf southern Yucatan Region. Land degradation and Brooke, C (1960): Khat (Catha Edulis): Its Production development, 15:397-409. and trade in the Middle East. The Geographical Liverman D, Moran E F, Rindfuss R R & Stem P C Journal, 126(1): 52-59. (1998): People and Pixels: Linking remote sens- Bryant, R & Bailey, S (1997): Third World Political ing and scoial sciences. Committee on the human

16 Khat Expansion and Forest Decline in Wondo Genet, Ethiopia

dimension of global change, commission on be- Stott, P & Sullivan, S (2000): Political Ecology: Sci- havioural and social science education. National ence, Myth and Power. Arnold, London Research Council. National Academy Press, Wash- Templeton S.R, & Sherr S.J. 1999. Effects of demo- ington DC. graphic and related microeconomic change on land Makin, M J, Kingham, T J , Waddams, A E , Birchall, C J quality in hills and mountains of developing coun- & Teferra, T (1975): Development prospects in the tries. World Development. 27(6):903-918 southern rift valley, Ethiopia. Volume 1. Land re- Tiffen, M, Mortimore, M & Gichuku, F (1994): More source number 21, Land Resources Division, Min- people, less erosion: Environment Recovery in istry of Overseas Development. London. Kenya. Wiley, London McClellan, C W (1990): Articulating Economic Mod- UNODCCP (1999): The drug nexus in Africa, ODCCP ernization and National Integration at the Periph- Studies on Drugs and Crime, Monographs # 1, ery: Addis Ababa and Sidamo’s Provincial Centers. United Nations Office for Drug Control and Crime African Studies Review, 33(1). Prevention, Vienna. Also available at: http://www. Milich, L & Al-Sabbry (1995): The “Rational Peas- unodc.org/pdf/report_1999-03-01_1.pdf ant” vs Sustainable Livelihoods: The Case of Qat US Department of Commerce (2000): Ethiopia. Overall in Yemen. http://www.togdheer.com/khat/rational. Export Down, “Khat” Sales Up. Available at: www. shtml: accessed 15 Dec 2004 tradeport.org/ts/countries/ethiopia/mrr/mark0004. Moran E F & Ostrom E (editors) (2005): Seeing the for- html; accessed 25 Nov 2004 est and trees Human-Environment interactions in VanWey L K, Ostrom E, & Meretsky, V (2005): Theo- Forest Ecosystems. Cambridge, MA: MIT Press. ries underlying the study of human environment Nestel, D (1995): Coffee in Mexico: international mar- interactions In Moran EF and Ostrom E (editors): ket, agricultural landscape and ecology. Ecological Seeing the forest and the trees Human-Environment Economics, 15: 165-178. interaction in Forest Ecosystem. Cambridge, MIT Neumann, R P (2005): Making Political Ecology. Hod- press, p 24-56 der Arnold, New York. Varisco D M, 1986. On the meaning of chewing: the sig- Project Cork (2006): Cork Bibliography: Khat. Available nificance of qat (Catha edulis) in the Yemen Arab at: http://www.projectcork.org/bibliographies/data/ Republic. International Journal of Middle East Bibliography_Khat.html, accessed 12 March 2006 Studies 18, 1 - 13. Robbins, P (2004):Political Ecology. A critical introduc- Walsh S J & Crews-Meyer K A (2002): Linking people, tion. Blackwell, Oxford place and policy. A GIScience approach.: Kluwer Robinson, S (1999): The cost of Catha edulis. Avali- Academic Publishers, Boston able at: http://www.time.com/time/europe/maga- Zewdu & Högberg 2000. Effect of land use on 15N natu- zine/1999/1220/khat.htm/, accessed 12 March ral abundance of soils in Ethiopian highlands. Plant 2006. and soil 222(1): 109-117. Selassie, S & Gebre, A (1996): Rap- Zimmerer K S & Bassett T J (2003): Political Ecology. id assessment of drug abuse in Ethiopia. An integrative approach to Geography and envi- Available at: www.odccp.org:80/bulletin/bulletin_ ronment-development studies. The Guilford Press, 1996_01_01_1_page005.html, accessed 20 Dec New York. 2004.

Paper IV

Ecological and Human Spatial Boundaries and Their Impact on Forest Management and Forest Decline in Wondo Genet, Ethiopia

Gessesse Dessie1,2 and Johan Kleman1

1 Department of Physical Geography and Quaternary Geology, Stockholm University, Sweden 2 Wondo Genet College of Forestry and Natural Resources, Hawassa University, Ethiopia

Abstract

This study investigates how the complex processes driving forest decline are affected by the structure and changes of human spatial and natural, ecologically defined, forest boundaries. The concepts of boundaries, and their importance in this context, are exemplified by conditions in the remaining natural forest of Wondo Genet Ethiopia. Natural forest boundaries in the area delineate an important ecosystem of internal spatial interconnectedness, while the human spatial boundaries delineate jurisdictional and benefit sharing units. We find that more than ten institutional actors are involved in management of the Wondo Genet forest, and that the forest is also divided between two ethnic groups and two regional governments. The forest is divided by a large number of boundaries. Most human spatial boundaries in the area are unrelated to primary ecological boundaries. Our results suggest that the forest management in Wondo Genet is negatively affected by situations of unclear jurisdiction, weak accountability and lack of control related to boundary issues and disputes. Of particular importance is that actions by individuals (local strongmen), occasionally rivals the authority of the institutional actors and are in some cases facilitated by weak control caused by boundary issues. Our findings in the Wondo Genet case confirm that boundary structure and changes in boundary structure can contribute to the forest decline. We define a functional chain that at least partly explains forest management problems in the area, and thus contributes to forest decline: boundary mismatch and boundary changes  “power vacuum” and lack of accountability  intrusion or illegal activities  illegal forest cutting  forest decline. For efficient forest management in the future, it is important to define management units that are a realistic compromise between ecological and human spatial considerations, and also to create relevant foray for discussion and practical collaboration, on the grass root level, as a tool for addressing residual management problems related to boundary issues.

Key words: forest decline, Wondo Genet, Ethiopia, boundary mismatch, human spatial boundary, ecological boundary, social ecological boundary, forest governance

Introduction decline through illegal forest activities continues. This raises a question; why is it so difficult to control the In Wondo Genet, south central Ethiopia, ecologically illegal activities? and economically high value forest is under serious This is not an isolated problem. In fact, illegal threat. During the past three decades, the forest area logging and uncontrolled deforestation are interna- declined from 16% to 2.8% (from 48924 ha to 8600 tionally recognized problems (Gregersen et al 2004). ha) in Awassa watershed, of which Wondo Genet is Studies show that illegal logging is a complex activity a part (Gessesse Dessie and Kleman 2007). Recently, that is influenced by specific social, economic and destructive activities including illegal logging and ille- political factors and reacts and changes in accordance gal settlement (Reporter 2005) and human caused fire with opportunities that transpire (Obidzinski 2000). (WIC 2004) have created substantial forest damage in For example, in Indonesia illegal logging has increased the area. In order to improve the situation, regional after new decentralization laws and it derives envi- states and local governments have instituted various ronmental decline (Mccarthy 2002). protective mechanisms, including consultative and pu- The continuing forest decline, despite the existence nitive measures. In the area, a decentralized resource of a potentially enabling management structure, indi- administration by local governments, run by elected cates that there are other, yet unidentified, problems officials, is in place. Despite the above conditions, which comprises several elements typically considered The meaning is as defined by the local government officials: beneficial to sustainable resource management (Lar- violation of official rules to benefit from logging trees and con- son 2003, Ribot 2003, Cumming et al 2006) the forest verting the forest land to other land uses.

Gessesse Dessie and Kleman, J. 2006. Ecological and Human Spatial Boundaries and their Impact on Forest Management and Forest Decline in Wondo Genet, Ethiopia. (Manuscript). Gessesse Dessie and Johan Kleman of the resource management. On the basis of indivi- social activities and boundaries separate sub-areas dual observed cases of illegal forest cutting, and the over which some aspect of control is (e.g. jurisdiction) temporal clustering of illegal activities to times of applied (Johnston et al 2000; Lovell 2002). From an unrest and government turnovers (Gessesse Dessie ecological perspective, boundaries are distinguished and Christiansson 2007), we here focus our attention by discontinuities of factors such as climate, phy- to the importance of the structure of human spatial siography and vegetation and have a large influence boundaries in relation to ecological boundaries, and on landscape function (Barnes et al 1998; Puth and to changes over time in boundary structure. Wilson 2001). The underlying hypothesis is that mismatch of eco- Power, access, meaning and identity (Lovell 2002; logical and human spatial boundaries, disagreement Walker and Peters 2001; Johanston 2000) are im- over boundaries, and boundary changes over time, portant aspects when human social boundaries are may contribute to disputed claims over areas, poor created, and recreated. Ecological boundaries, alt- accountability, and weak or no control over certain hough reflecting natural factors such as vegetation and areas, all of these possibly being factors conducive to climate, are in some sense also social constructions illegal forest cutting and forest decline. and reflect the particular factors deemed important at The term boundary refers to a wide range of the time when they are defined. Definition of ecosys- real and conceptual structures. It, hence, concerns tem boundaries is strongly influenced by the type of social spatial relations, by which we mean people, research questions posed (Cadenasso et al 2003). space and place links (Johnston 2000). These links The issues of boundaries and their importance in are important when boundaries and areas are defi- the social spatial relations is the subject of a consi- ned, forming landscape units and elements in which derable amount of research. With respect to ecolo- social relations and identities are instituted. One of gical boundaries, a number of authors, for example the important types of human spatial boundaries is Cadenasso et al (2003), Strayer et al (2003), Puth the administrative boundary, while one example of and Wilson (2001), and Fortin et al (2000) have natural boundary can be the delineation of a natural presented scale, functions, characteristics, theories, forest ecosystem. When any form of natural forest classification and strategies for detection. With respect management is implemented, the two boundaries to human spatial boundaries Johnston (2000), Lovell interact and outline the social spatial relation. The et al (2002), Swyngedouw (2004), and Agrawal and location, type and physical manifestation (if any) Goyal (2001) presented definitions, their relevance to of boundaries can influence the form of accounta- integrated natural resources management, concepts in bility, frequency of claim, access types, and state of connection with space and place as well as in terms of coordination and information flow. Natural forest group size and collective action. Others emphasized boundaries function as delineators of ecosystems the problems of boundary mismatch between the eco- of internal spatial interconnectedness, and human logical and human boundaries in the social-ecological social boundaries delineate jurisdictional and benefit systems (e.g. Cumming et al 2006). sharing units. In summary, three general forms of boundaries The study first presents the concept of boundaries stand out: ecological boundaries, human jurisdiction in the human and ecological contexts, then analyses (administrative) boundaries and human-ecological the boundary structure and cases of illegal activities management boundaries (that links administration in the study area, and finally discusses the significance and ecology). With respect to forest decline studies, a of our findings for natural forest management. closer analysis of these three forms of boundaries can The methods employed in this study are largely aid in understanding the social spatial relations and qualitative. We have produced maps showing human- the different types of pressures (alternative land uses, defined as well as ecologically relevant boundaries and logging, etc) to which the forested area is subjected. analyze these in the light of information on particular The delineation of environmental boundaries is an events and processes that have been found important important step in land-use management planning and for the forest evolution in the area (Gessesse Dessie integrated natural resources management (Campbell and Kinlund 2007, Gessesse Dessie and Christiansson 2001et al; Fortin et al 2000), and obviously also re- 2007). Additional information was provided by social quires consideration of human-social and ecological survey methods including interviews and attendance aspects. One of the salient features of boundaries is of public meetings. a frequent mismatch of human spatial and natural/ ecosystem boundaries. Such boundary mismatch can lead to inefficient, inconsistent, wasteful, or destructive resource management (Meidinger 1998). Some Conceptual background of the widely promoted management types including integrated natural resource management (Campbell Boundaries in general are important components of et al 2001), ecosystem management (Lackey 1998) spatially heterogeneous areas and serve the purpose and social-ecological management (Cumming et al of delineating subareas that are in some respect inter- 2006), have similar characteristics that suggest to nally homogeneous (Metzger and Miller 1996). From meet human needs without impairing the integrity of a human perspective, the heterogeneity is defined by underlying systems processes. These approaches, on

Ecological and human spatial boundaries in Wondo Genet, Ethiopia the one hand, encourage system level management to Human spatial boundaries insure spatial integration (connectivity of the system elements) and temporal continuity. On the other hand, Human spatial boundaries specify areas over which these studies also propose decentralized resource certain social rules or activities apply. One of the main management to enable local level participation and origins of human spatial boundary is investigative in equitable distribution of benefits. These two aspects nature depending on the topic of concern such as ad- of the same approach seem to conflict due to mismat- ministrative jurisdiction, resources management and ches where human boundaries could slice the forest settlement pattern. Tangible origin, such as ethnic di- ecosystem into multiple pieces. visions, also exist, often with a strong sense of identity related to place and space. Over time, human spatial boundaries are created and recreated with many varied economic, social, cultural, environmental as Boundaries - concepts, character and well as political consequences (Johnston et al 2000; function Storey 2001; Walker and Peters 2001). Often human spatial boundaries delimit and Boundaries in general can have four properties (Cade- bound certain resources or environments over which nasso et al 2003; Lovell et al 2002; Harvey 1990). social groups establish claims. In such situations 1. Origin: investigative or tangible. Investigative boundaries mark divisions of control over, and re- - arbitrarily placed boundaries for the conve- sponsibility for, resources among individuals, groups, nience of a scientific study or political reasons e.g. organizations and governments. Social boundaries are political boundary. Tangible-structures that can governed by rules and conventions, and they simul- be identified in nature e.g. forest and non forest taneously separate and connect. Boundaries as such boundary. imply a human presence and involvement in shaping 2. Spatial structure: as defined by one or more pro- the nature. Consequently, human spatial boundaries perties including location, scale, geometry. can reshape and sometimes even create ecological ones (Meidinger 1998). 3. Function: transmissions of energy, matter, orga- With respect to natural forest management, human nisms or information. Ecological boundaries are spatial boundaries function to enclose, separate and commonly differently transmissive or permeable; demarcate the given forest unit. This implies at least that is, they may allow only some fraction of four conditions (Fig. 1). Firstly, the management of a material, energy, or organisms to pass. given forest by a given social group is determined by 4. Temporal dynamics: stability/change, age and his- proximity which often is related to right of use and tory. The position of a boundary in the landscape access which often is the outcome of power relations. may be stationary over time, or it may move e.g. Secondly, the management and the mode of extracting change of administrative boundaries, expansion goods and services from the forest depend on the of forests. level of existing/current use, practice or meaning.

0OWER $EFENSEOFVESTEDINTEREST

!CCESS 2ESOURCES )DENTITY 2IGHTOFUSE PROXIMITY .ATURALFOREST 0LACEANDSPACECONCEPT

-EANING SPECIFICUSEPRACTICE

Figure 1. Proposed factors that influence the creation of human spatial boundaries of relevance for resource management.

Gessesse Dessie and Johan Kleman

Figure 2. The study area is located in the south-central Rift Valley of Etiopia. The white ellipse indicates the approximate location of the 2700 ha Wondo Genet forest.

Third, share of the benefit from the forest is primarily as dominating a particular spatial domain. Ecosys- claimed by the social group whose identity is related tems can be distinguished by major discontinuities to the particular place and space where the forest is. of climate, physiography and vegetation (Barnes et Fourth, the actual division of resources depends on al 1998), and are structured by processes and feed- the power relation in defense of vested interest. These backs amongst the organisms constituting the system, four conditions are strongly interrelated and outline including human influences that arise from interaction the social spatial relation. of organisms with their environment (Cumming et al When a given natural forest is divided by human 2006). A given type of natural forest is an ecosystem, boundaries, communities who live on either side of but can also be considered as a natural resource. So the boundary interact over the share of the benefits. far, we have mainly considered the spatial delineation The interaction across human spatial boundaries of ecosystems, but it is important to note that ecologi- can be related to boundary characteristics. Power cal boundaries also are important for understanding over the resources emanates from privileges related the spatial integrity and temporal continuity of the to proximity, access, culture, as well as legal and system. administrative conditions. When boundaries remain In the following, we are particularly concerned consistent/unchanged for long time periods it is likely with the transition of one particular ecosystem, high that communities identify themselves with the space forest, to other vegetation types/land uses, and how and place. This aspect can strengthen the sense of po- this affected by boundary structure. The reasons for wer over the natural forest bounded within the com- the significance attached to high forest is that the munity territory. Furthermore, values and meanings high forest boundaries outline multiple important towards the use and management of the forest develop properties; 1) they delineate areas with valuable over time. Rightful access is related to power, identity indigenous coniferous trees, 2) they delineate areas and meaning. When the interaction over boundaries with important biodiversity, ecosystem and watershed intrudes the power, identity, meaning and access of functions, 3) they have implications for sources and the rightful claimant it is likely that the forest can be transfer of material, water, energy and organisms. exposed to destruction.

Ecological boundaries The case of Wondo Genet forest Ecological boundaries, sharp or indistinct, separate We here present a case example of the human and internally relatively homogeneous areas from each ecological boundaries from Wondo Genet forest and other, allowing us to regard a particular ecosystem Awassa watershed in south central Rift Valley of

Ecological and human spatial boundaries in Wondo Genet, Ethiopia

Ethiopia (Fig 2), and the influence of this boundary Kleman 2007. The recent causes are analyzed from structure on forest management and forest decline. different sources including reports from the college, The Wondo Genet forest; 2700ha, is an upstream personal experience, interviews, and deliberation in forest within Awassa Watershed, 306,000ha, in which public gatherings. Group discussions, in depth inter- over half a million people live. This forest is the partly views with key informants, and participation in public fragmented remnant of a formerly larger (over 16% gatherings revealed that different forms of forest de- of the watershed area during 1972) and more cohe- cline occur in the Wondo Genet forest. The identified rent forest covering the eastern rift flank. It harbors causes are illegal logging, illegal forest settlement, important and rare fauna and flora, and provides illegal forest land transactions and human caused fire watershed, ecosystem, economic, research and edu- (Reporter 2005; WIC 2004). Key informants repor- cation services. The administration/management of ted that the illegal logging is carried out/ordered by the forest is divided between forestry projects (Table powerful and well-connected individuals. They target 1), a Forestry College and kebeles. The importance of high value timber species and charcoal burning. The the forest is well recognized by the federal, regional most affected tree species is Aningeria adolfi-frediric and local governments as well as by forestry experts and as a result this tree species is threatened by local and researchers (Makin et al 1975; WGCF 2004; extinction. Reporter 2005). According to interviews with kebele leaders, illegal The Wondo Genet forest and Awassa watershed forest settlement, i.e establishment of farm plots, gra- present important conditions to study social-spatial/ zing and residence, has increased recently. During the boundary also in perspectives of forest decline because year 2004 alone, about 60 households were identified 1) increasing illegal activities have been reported, 2) as illegal settlers within the forest boundary of one of complex natural and human boundaries exist in the the kebeles. The interviews also revealed illegal selling area, 3) multiple claims and weak accountability of forest land within the kebeles boundaries/jurisdic- have been observed, and 4) it is an area of ethnic and tion particularly in areas along borders. In Ethiopia, cultural diversity, and ethnically related identity and land is state owned and sale of land is officially pro- power struggle have been going on for many years. hibited. However, in Wondo Genet the practice of These characteristics have created a landscape where selling and buying farm land exists between farmers, boundaries of different cause and consequences are and between farmers and immigrants (who come in- entangled and created the background for forest termittently in search of land and employment). Sale management in the area. of forest land is new for Wondo Genet where farm land transaction otherwise is not new. Recent causes of forest decline in Wondo Genet The prevalence of illegal forest destruction activities is recognized by the regional states. Various An assessment of the magnitude of the forest decline measures have been taken to control it, including in the study area was given by Gessesse Dessie and policing and guarding, consultation and mediation,

Figure 3. Complex landscapes, such as this transition zone between the upland-located Wondo Genet forest, and the low- land agricultural area, contain a large number of boundaries, visible as well as invisible. The natural forests are located in the hills and valley pockets. In the foreground, individual fields and plots reflect the smallest human spatial delineations of the landscape, related to the household level. Larger-scale human spatial boundaries, such as those bounding peasant associations (also called kebele, a grass-root level administration unit in Ethiopia with a defined spatial limit of approximately 800 hectare), are not readily visible in the landscape, but nevertheless very important for resource management. Human spatial boundaries in the study area rarely coincide with the ecological boundaries. (Photo Gessesse Dessie)

Gessesse Dessie and Johan Kleman

The ecological and human boundaries

The existing natural and human boundaries in the study area are presented in Fig 3, Fig 4 and Fig 5. The natural boundaries are watershed and ecosystem related. The human boundaries are sociopolitical, cultural and settlement related e.g. urban and rural boundaries. The watershed related boundaries have three forms: . • Primary, topographically determined boundaries between individual watersheds. • Ecosystems defined by varying biophysical condi- KM tions within the watershed e.g. the natural forest ecosystem of Wondo Genet. • A first-order division in upstream and downstream 7ATERSHEDBOUNDARY areas according to distribution and availability of 5PSTREAMDOWNSTREAMBOUNDARY water for agriculture and household purposes. 2EGIONALSTATEANDETHNICBOUNDARY Upstream areas are characterized by high altitude +EBELEPEASANTASSOCIATIONBOUNDARIES and steep slopes while downstream areas are typically associated with gentler slopes. &ORESTRYPROJECTBOUNDARIES

.ATURALFORESTCOREAREA Watershed boundaries have divided the people with regards to resource distribution and access. One 7OSHAKEBELE example is the separation between up-streamers and 7ONDO'ENET#OLLEGEOF&ORESTRY down-streamers. The two groups of people are markedly different with respect to access to agricultural Figure 4. Natural boundaries and human spatial boun- productivity (upstream land is agriculturally less daries in Wondo Genet. The shaded oval area indicates productive than downstream land where fertile soil the location of the core area of the Wondo Genet forest, and water is available). Those who live upstream are which with varying density covers an area that extends in a marginal agricultural environment while “down- from Mount Abaro to the border of the Awassa watershed streamers” are privileged with environmental condi- in the South. tions of high agricultural productivity such as gentle slopes, deep soil and ample water supply. efforts to improve public awareness through public The human boundaries are nested sociopolitical meetings, and even military intervention. However, boundaries, cultural/ethnic boundaries and bounda- no significant progress is registered in relation to ries of management related to value and meaning. The controlling the forest decline. The interview results watershed in general and the forest in particular (Fig are corroborated from other sources, mainly from 4) are divided by the boundary of two regional states the Wondo Genet College. During the last decade, namely Oromia Regional State and South Nations the Aningeria adolfi-fredirci forest of the college was Nationalities and Peoples Regional State (SNNPRS). destroyed beyond recovery (WGCF 2004). In the case Under the regional states the forest falls under the of this particular forest type, a physical confrontation jurisdiction of three zones and three woredas (di- between guards and strongmen involved a gunfight. stricts). Further down, the forest is subdivided into Military has been deployed also on some other oc- kebeles. The area and population of kebeles change casions. Numerous court cases are filed. from time to time due to merging of kebeles. Kebeles According to interviews, forest fires occur fre- are further divided in to sub kebeles. Wondo Genet quently. Respondents believe that most of the fires forest is divided between 10 kebeles. are deliberately set. Most common reasons presented The regional state boundaries more or less follow by the interviewees for the deliberate fire is conflict the cultural/ethnic divide of the two dominant groups over access and use of natural forests among different in the study area, namely Arsi Oromo and Sidama. community groups. Fire seems to have been especially This boundary has persisted in the landscape and frequent in the college controlled forests. During the among the people for many generations and carries past 10 years alone, 5 fires of serious consequences a marked strong territorial identity. Repeated violent were identified. Local and international media out- confrontations have occurred in the past, and various lets have also reported them (WIC 2004; AFP 2004; conflict management measures were implemented by Reporter 2005). the states and local people (Girma 2001). Recently a referendum was conducted to delimit the boundary According to kebele leaders strongmen are financially strong, locally influential and well connected men who are also involved To create a clear-cut upstream and downstream boundary is in illegal timber trade difficult. This is an approximate delineation.

Ecological and human spatial boundaries in Wondo Genet, Ethiopia

Figure 5. a/ Fenced boundaries of parcels belonging to individual households. The area represents grazing management within a formerly continuous Juniperus procera forest. b/ Boundaries of plots farmed by different households. The area represents farming management within a dominantly Podocarpus falcatus forest. (Photos Gessesse Dessie) between these groups (Zerihun 1999), with the result • Land tenure changes; three major land tenure that this culturally defined boundary was somewhat transformations have occurred in the area during modified. the past 100 years (Gessesse Dessie and Christi- Another form of boundary related to the natural ansson 2007). These changes have defined and forest is the boundaries of individual households. redefined the landscape according to the right of Part of the Wondo Genet forest is divided among access and control households (see Fig 5). Some of the forest land is • Sociopolitical changes with administrative boun- distributed among households by the kebeles where daries have been changed repeatedly. individuals change the land use from the natural forest to grazing or farming. Fig 5a shows boundaries/fences • Resource value/meaning changes; the emergence of individual households that have divided a relati- of commercial agriculture. The now dry but vely continuous forest into a piecemeal arrangement previously malaria-prone area was unsuitable of individual plots for grazing and farming (Fig 5b). for human activities but has largely changed Analysis of the Wondo Genet landscape history to commercial agriculture. Forest management revealed three possible causes of boundary creation strategies evolved through different priority ob- and recreation: jectives including plantation forest, farm forest and natural forest.

Gessesse Dessie and Johan Kleman

Table 1. Boundaries, use right, ownership, jurisdiction and temporal aspects. There are no official cadastral maps showing the exact location of each boundary and use right. Most of the boundaries are approximately designated by clearly visible permanent topographic features.

Institutions/ Boundary types Boundary delineation Jurisdiction/ownership Power over forests stakeholders time

The college Ecosystem Late 1970s (Royal • Forest education and • Prepare and – spatially land before early research implement own continuous forest 1960. After • Conserve, utilize and plant forest management area that property of forests. plan. Additionally Norwegian mission) • Develop the land e.g. • Control the income the college cultivation and dairy from forests and operates outside farming land. its boundary by • Report to federal • Protect forest from own initiative government and Debub intrusion. to conserve and University • Has no power protect forest to take punitive measures on intruders Projects Ecosystem and Early 1980s and • Conserve, utilize, plant • Prepare and economic- natural 1990s forests implement own forest and • Report to regional forest management plantation forest. government and regional plan. One of the projects bureau of agriculture • Control the income (Shashemene wood from forests. industry) is one • Protect forest from of national forest intrusion. priority areas in • Have no power Ethiopia. to take punitive measures on intruders Kebeles Smallest public Mid 1970s • Government • Own forests outside administrative unit. representative, public college and project When originally administration. boundaries defined, each • Land administration, • Protect forests from kebele was distribution, intrusion approximately 800 • Report to woreda • Have official power ha in area. Since administrative unit to take punitive then, merging measures on or splitting have intruders changed the original areas Regional states Largely reflects a Early 1990s (The • Own land, decide on land • By default own all language-based ethnic boundary use, redefine administrative land and resources ethnic division exist since at least a units • Manage forests and century, but was not • Report to federal control possible used as administrative government income from forests boundary previously) outside college and project boundaries • Have official power to take punitive measures on intruders Households Use right/ Mid 1970s and • Own the right to use land • Free benefit of ownership e.g. onwards officially allocated to in-forest grazing, fences around farm them. fire wood (dead plots, grazing areas • Report to kebeles wood), non timber products e.g. honey production

Ecological and human spatial boundaries in Wondo Genet, Ethiopia 2EGIONALSTATES

7OREDAS

+EBELES :ONES

+OFFELE

7ONDO'ENET !RSI #OLLEGEOF&ORESTRY /ROMIA 3HASHEMENE

7ONDO'ENET %ASTSHOA 3HASHEMENE 7OOD)NDUSTRY &OREST !WASSA 7ONDO9ANASE 3..02 &ORESTRYPROJECT 3IDAMA

Figure 6. Scheme of administrative units (based on official structure), right, and important actors with economic, research or conservation interests, left, involved in the management of the single ecologically defined unit of the Wondo Genet forest. Recently more administrative boundaries are created. Awassa woreda (an administrative unit composed of several kebeles) is divided into three woredas while Shashemene woreda was elevated to a zone status by merging with four other neighboring woredas.

Our key findings regarding the present boundary Forest governance and boundaries in structure (Fig 4) in the study area can be summarized Wondo Genet as: - The Wondo Genet forest is divided between two Governance in general refers to the structure and regional governments. process by which societies share power (Lebel et al 2006). With respect to natural forest management in - The Wondo Genet forest is divided between two Wondo Genet the power is shared between the two ethnic groups. regional governments and their lower administrative - The Wondo Genet forest straddles an important units (Fig 6), individual forestry projects, the Wondo local water divide. Genet College of Forestry and the kebeles (see Table - The Wondo Genet forest is divided between a 1). Here we emphasize the interaction of the Forestry number of kebeles. College and kebeles. Forest management is the responsibility of kebeles - Forestry projects having both conservation and for two main reasons 1) by virtue of proximity or loca- economic objectives involve parts of the Wondo tion within the forest boundaries, kebeles possess the Genet forest. Their boundaries, which delineate right of forests 2) regional states are by the Ethiopian continuous forest patches, do not coincide with constitution given the power and function to adminis- any other boundaries. ter land and natural resources (Federal Negarit Gazeta - There is a large number of (more than ten) institu- 1995). Kebele is the lowest administrative structure tional actors (college, kebeles, forestry projects, in the regional states, the main power center to run regional governments) involved in management various activities at grass-root level including natural of the Wondo Genet forest. forest management. - The spatial extent of the jurisdiction of the college The College assumed responsibility in the late over the Wondo Genet forest is unclear. 1970s for management of the forest. For this purpose it demarcated most of the Wondo Genet forest, in or- It should also be noted that the Forestry College’s der to exercise the power to protect and utilize it. The management responsibility is based on an ecological forest that is not officially known (outside College’s definition. This is in contrast to the claims or jurisdic- boundary) as the College property is managed by tions of most other actors, and the role of the college nearby kebeles or kebeles located within and partly consequently decreases as a simple function of the within the forest. These two actors, the College and ongoing reduction in area of high forest. the kebeles, demarcate contrasting boundaries where

Gessesse Dessie and Johan Kleman

Table 2. The roles of the college and the kebeles in natural forest governance in Wondo Genet. These two main actors uphold boundaries defined according to widely diverging criteria.

The college Kebeles Objectives are forestry education and forestry research Local government, engaged in development, public service and administration Primary concern is ecosystem services of the forest for the Primary concern is economic benefits and livelihood wider society improvement for the local people Engaged in theoretical forestry and forest-people related Engaged in practical problems related to local peoples problems livelihood improvement Operations guided by global forest management Operations guided by sociopolitical organization of the paradigms and university education policies of the country country and natural resource policies Run by comparatively highly qualified highly educated Run by non professionals and with little formal education, and comparatively highly paid elected/government volunteers or little paid persons elected by the general appointed professionals public Influenced by changes in global environmental Influenced by the change of government, change management narratives/ paradigms and education policy of natural resource management policies and local of the country sociopolitical conditions Information, knowledge, material flow to the kebele in the Information, knowledge, material flow to the college form of extension, mediation, facilitation in the form of indigenous knowledge, community participation Continuous leadership and consistent forest management The short election cycle works against a long time objectives perspective in natural forest management planning

the kebeles divide the forest into artificial units while kebeles and the college try to stop, but with different the college subdivides the forest based on ecological methods, are illegal logging and permanent settlement considerations and spatial integrity. in the forest. The college deploys armed guards and The kebele boundaries confer the normal function police and also initiate participatory measures th- of the state apparatus including implementation of rough its social facilitation unit. The participatory development activities, tax collection and public ser- activity involves mainly elderly people, the general vices, e.g. local courts. The college’s boundaries cover public and elected officials. The use of armed guards multiple kebeles and confer sustenance of the spatial and police are to protect the forest from presumably integrity and temporal continuity of the natural fo- few individuals (e.g. strongmen) involved in illegal rest, mainly for its academic and research purposes. activities. In extreme conditions military is involved. The two types of boundary rarely coincide, instead The kebeles use public militia to enforce rules and law the boundaries defined by the College cut through and use community courts to reprimand violators. regional state and kebele boundaries (Fig 4). Occasionally, general public meetings (Fig 7) are The forest management by the kebeles follow called and the military is involved. The college uses guidelines from the regional states through their Bu- kebeles when a need arises to use community courts reaus of Agriculture. The kebeles determine access, and public militia, but otherwise extreme cases are control and use of the forest demarcated within their referred to the regional state authorities. jurisdiction. Public access to the forest is permit- Some of the major issues with respect to forest ted for grazing, firewood collection for household decline deliberated during public meetings are: consumption and use of non-timber forest products • Natural forests located along modified boundary while cutting live tree and permanent settlement in lines are claimed by both parties. the forest is prohibited. • The boundaries of the college and the kebeles The college, on the other hand, prohibits access have different connotations in the management to the forest, erects fences along the boundaries it of the natural forest (Table 2). upholds, and employ guards patrolling the forest. The spatial mismatch between the two boundary • The natural forests, and particularly valuable types causes problems for both the kebeles and the timber trees are forcefully seized by powerful college. The interviews indicated that in some cases individuals (strongmen) who elude public and the jurisdiction of the kebeles is intruded, while in state rules and regulations. other cases the management interest and objectives • The natural forests are secretly divided between of the college are violated (see Table 1 for jurisdiction few individuals and delineated by covert boun- and power over forests). daries. The two primary illegal activities that both the • Natural forest protection activities by public militia, state police or guards are counteracted The manner of natural forest use by the kebeles is often de- by armed resistance from the strongmen. termined by the regional bureau of Agriculture

10 Ecological and human spatial boundaries in Wondo Genet, Ethiopia

• Court proceeding of violators is time consuming have created wider forest destruction. Our mapping and practically difficult. shows that ecological and human boundaries rarely match. It is also known (Gessesse Dessie and Christi- The kebeles are structured in polycentric adminis- ansson 2007) that in times of political regime changes trative units and report to different higher level units boundaries are often violated and situations characte- (woredas, zones) devolved from two neighboring rized by “power vacuum” and lack of accountability regional states (Fig 7). In the certain circumstances, occur. In those instances, forest decline is accelerated two neighboring kebeles can report to two different as a result of illegal actors exploiting a state of un- states. In such situations violators on the same forest certainty and lack of control. Thus, forest decline is can be taken to different courts spaced up to 200 km partly related to boundary issues and in particular, apart. There is no administrative unit dealing with the to boundary changes. We can define a functional natural forest as a single unit. Individual kebeles are chain with a number of links: boundary mismatch not integrated to sustain their common interest over and boundary changes  “power vacuum” and lack the forest, but have hitherto managed their particular of accountability  intrusion or illegal activities  forest area more or less independently. However, due illegal forest cutting  forest decline. to increased illegal logging, there are now signs of emerging collaboration between kebeles to counter Forest governance and forest decline the illegal activities (WGCF 2004). With respect to the ongoing forest decline, the illegal activities are at least partly related to social spatial Discussion and conclusion relations such as unresolved, overlapping and multiple boundary issues. Possible indicators to this are; The Wondo Genet forest is the partly fragmented 1) continuing violation despite strong government remaining part of a formerly larger and more cohe- presence (policing, military intervention, prosecution rent forest covering the eastern rift flank. It harbors of violators, mediation and awareness of the general important and rare fauna and flora, and provides wa- public) and public readiness 2) failure to solve causes tershed, ecosystem, economic, academic and research of violation despite the decentralized administration services. In fact, a number of different boundary of kebeles that reach the grass root level. types are related to these services. All the mentioned The fact that nested boundaries of the polycentric services are to a varying degree threatened by the organization of administration overlay and crosscut forest decline and thus place the focus on forest ma- an ecologically defined forest structure in Wondo Ge- nagement, which, in turn, is strongly related to the net seems to have contributed to fragmented jurisdic- boundaries upheld by various actors. Despite efforts tion, ownership and management which have led to from several actors, including the kebeles, the college, multiple ownership claims, weak accountability and and the regional states, illegal activities continue and administrative inefficiency. Additional contributing

Figure 7. Photo showing one of the general public meetings in Wosha kebele to deliberate on forest decline in Wondo Genet (Photo Gessesse Dessie).

11 Gessesse Dessie and Johan Kleman factors are likely the lack of common forest manage- to natural resources help manage them, because this ment objectives and sporadic and inconsistent forest improves the chance that the resources will actually protection with complications to enforce legal and be protected 2) the premise of decentralization is that administrative measures. It seem that violators such natural resources should be managed locally to give as illegal loggers and illegal settlers exploit these weak all users the opportunity to participate in management links and are able to easily stifle the elected kebele decisions. However, here we postulate that the need leaders authority. The local structure of boundaries, to involve local people in decentralized management actors, and power and accountability relations deter- possibly divides the ecosystem into too small units mine the success of forest management. for optimal resource management. The existing decentralized administration in Wondo Genet, even though the power and authority reaches local levels, local identities are considered, and Consideration of scale in defining social ecological the kebeles are being run by public elected officials, boundaries seems to be unable to control forest decline. This indicates that decentralization may not always bring Appropriate scales of management are different for about the desired solution in natural forest manage- ecosystem and for decentralized resources. The former ment. Though decentralization is a preferable means require landscape, system or watershed scale (often of ensuring local peoples right, it can unfortunately large area) where spatial integrity is insured while also weaken authority and accountability. Disinte- the latter require small social groups (often small gration of authority, the way each kebele operate area) to insure efficient management. In decentralized by its own in Wondo Genet, hurts ecosystem (trans- management, it has also been noted that group size is boundary) scale management that would require inversely related to successful collective action. The coordinated efforts by several kebeles. community size for decentralized resource manage- Another reason that contributed to prominence ment should be small to medium (Agrawal and Goyal of violators could be the claim of both ethnic groups 2001). The existence of boundary issues regarding over the Wondo Genet and the forests. At face value, the appropriate management scale for certain types the claims seem to have been resolved when a referen- of environmental problems or resources that do not dum (Zerihun, 1999) was held to resolve the ethnic necessarily coincide with political boundaries has boundary issues. However, there are still indications previously been recognized (Walker 1999; Larson that claims related to ethnic identity linger. Due to 2003; Ribot 2003). these unresolved claims, and possibly the history of In Wondo Genet, as in other parts of Ethiopia, the violent conflict, Wondo Genet remain a politically grass root government is at kebele scale. Kebeles can sensitive area. This seems to have created a weak link contain different community types and be of varying that violators can easily exploit. The Wondo Genet size, but are in any case much smaller than the primary case is not unique in an Ethiopian context. Similar ecological unit to be managed and protected. In the conditions of illegal activities as in Wondo Genet are wider Ethiopian context, there are also other accounts rampant in other natural forests and wildlife parks that wildlife parks and patches of forests that do not (Melaku 2003). coincide with political e.g. regional state, cultural e.g. Our results in the case of Wondo Genet indicate ethnic boundaries exist with inherent conflicts and that a decentralized resource management may in management problems (Unruth 2006; Demel and some cases be inefficient in preventing destruction or Mulugeta 2005). The Wondo Genet case supports over-utilization of the forest resource. This indicates that, on one hand, ecosystem management needs to the existence of problems not fully considered in address a scale where elements of the system function works advocating “integrated natural resource ma- together e.g. landscapes and/or watershed scale (large nagement”, as an efficient way to integrate social and areas). On the other hand, to ensure efficient decen- ecological systems for sustainable environment and tralized management small to medium social groups equitable distribution of natural resources (Campbell (small area) is required. Since ecosystems transcend et al 2001; Lackey 1998). On the basis of the Wondo jurisdictional boundaries, ecosystem management Genet case, we argue that boundary structure may be must be built up on cooperative interagency institu- an underlying problem that causes difficulties in forest tional structures and processes. This situation calls protection and resource management. It is important for selection of appropriate scales and boundaries, to note that the boundary issue is not the only poten- leading to efficient management. tial/possible cause of illegal activates. Studies from The optimal area/scale/boundary can be where the Indonesia and elsewhere have shown that weak insti- integrity of the system is insured while local people’s tutions, non-consistent natural resource management rights are protected. The theoretical boundary can be policies, and socioeconomic conditions of the people a divide that separate purpose, objectives, meanings could also lead to a similar outcome (Gregersen et al and areas but at the same time keep the social and 2004; Mccarthy 2002; Obidzinski 2000). natural system integrated. To find the appropriate With respect to decentralized resource manage- scale of management that create efficient management ment according to Agrawal and Goyal (2001) it is and combine ecological and human aspects is one of important that 1) people who live in close proximity the challenges to control forest decline problems.

12 Ecological and human spatial boundaries in Wondo Genet, Ethiopia

.ATURALSYSTEM (UMANSPATIALBOUNDARIES .ATURALFOREST SOCIO POLITICALUNITS %COSYSTEM #ENTRALIZATION -ANAGEMENT

4HEORETICAL

/PTIMALSPATIALSCALE AREA

3USTAINABILITY BOUNDARY

,OCALPEOPLERIGHTS

#OMMODITY -ANAGEMENT $ECENTRALIZATION

Figure 8. From an ecosystem perspective, the optimal condition for efficient management will tend to occur with relatively large and ecologically defined management units. From an equitable resource distribution perspective, efficient management will tend to occur with relatively small management units that are defined on the basis of access and other human considerations. In reality, it may be impractical or impossible to match ecological and administrative boundaries. In those instances it is important to define management units that are a realistic compromise between ecological and human spatial considerations.

Fig 8 outlines factors influencing the definition and sufficient mother trees as a seed source. This and delineation of the social-ecological boundaries. ideal condition needs an optimum area depending Ecosystem management views the land and resource on the particular forest ecosystem for the forest to base in its entirety, as a holistic or integrated entity stay resilient against human and natural pressure to accommodate spatial integrity and temporal con- and continue operating. Need of agricultural land tinuity. Management focuses on entire ecosystems, or settlement land can change the forest boundaries not just individual or commodity resources such as (due to intrusion into the forest/pushing frontiers) timber. This ecological management perspective emp- while selective logging, grazing, fire etc can change hasizes the need for larger area (of management units), the internal structure of the forest. with the consequence that such larger management units often cross jurisdictional boundaries. From a sustainability perspective, holistic management is preferable but from equitable resource distribution Conclusions and suggestions perspective decentralized management is preferable. - The Wondo Genet forest is one ecological unit, but In decentralized management, it is emphasized that there are many actors involved in its management the need to uphold local people’s right requires small and upholding boundaries through the forest. community size (Cumming et al 2006; Larson 2003; Most human spatial boundaries in the area are Agrawal and Goyal 2001; Cambell et al 2001; Ostrom unrelated to primary ecological boundaries. 1999; Lackey 1998). - The forest is primarily divided between two ethnic To bring the ecosystem management and resource groups and two regional governments, and there management together, inter-jurisdictional coordina- is around ten institutional actors (college, kebeles, tion is important. While decentralization can result in forestry projects) involved in management of the effective management, offsite benefits of natural forest Wondo Genet forest. as well as material, energy and information transfer between boundaries need to be considered to ensure - The forest is thus divided by a large number of spatial integration and temporal continuity. human spatial boundaries, some of which interact From a temporal perspective, natural forest mana- in such a way that they give rise to multiple claims gement requires continuity of consistent prescriptions and weak control and accountability, conditions to maintain the dynamics of the ecosystem. It is im- that negatively affect forest management. portant for the forest to have balanced proportions of - The boundary conditions in Wondo Genet seed source, seed germinating environment, seedling contribute to non-coordinated natural forest and sapling, favoring a multi-layer canopy structure management. The college and the kebeles have

13 Gessesse Dessie and Johan Kleman

overlapping claims over forests but the kebeles Forestry. have the authority with regards to enforcing Federal Negarit Gazeta. 1995. The constitution of the administrative measures. Federal Democratic Republic of Ethiopia. Addis - Illegal actions by individuals (local strongmen), oc- Ababa casionally rivals the authority of the institutional Fortin, M.J., Olson, R.J., Ferson, S., Iverson, L., actors and are in some cases facilitated by weak Hunsaker, C., Edwards, G., Levine, D., Butera, control caused by boundary issues. K., and Klemas, V. 2000. Issues related to the detection of boundaries. Landscape ecology 15: - In general, our findings in Wondo Genet confirm 453-466 that boundary structure and changes in boundary Gessesse Dessie, and Christiansson, C. 2007. Forest structure can contribute to the forest decline. decline and its causes in south-central Rift Valley - For efficient forest management, it is important of Ethiopia. Human impact in a one hundred years to define management units that are a realistic perspective. Stockholm University, Department compromise between ecological and human spa- of Physical Geography and Quaternary Geology, tial considerations. (submitted manuscript). - Creation of relevant fora for discussion and prac- Gessesse Dessie, and Kinlund P. 2007. Khat expansion tical collaboration, on the lowest possible level, and forest decline in Wondo Genet, Ethiopia. may be an efficient tool for addressing residual Stockholm University, Department of Physical management problems related to boundary is- Geography and Quaternary Geology, (submitted sues. manuscript). Gessesse Dessie, and Kleman, J. 2007. Pattern and Magnitude of Deforestation in the South Central Rift Valley Region of Ethiopia. Mountain Aknowledgements Research and Development 27(2):162-168. Girma, N. 2001. The Arsi and the Sidama: a history of The authors wish to thank Professor Carl Christi- inter-ethnic relations along the borderlands, ca. ansson and Dr. Peter Kinlund for their constructive 1900-1991. MA thesis, Addis Ababa University, comments and important inputs to this paper. Addis Ababa Gregersen, H., Contreras-Hermosilla, A., White, A., Phillips, L. 2004. Forest Governance in Federal Systems: and Overview of Experiences and References Implications for Decentralization. CIFOR, Bogor, Indonesia awal, A. and Goyal, S. 2001. Group size and Agr Harvey, D. 1990. Between Space and Time: Reflections collective action. Third party monitoring in on the Geographical Imagination. Annals of common-pool resources. Comparative Political Association of American Geographers 80(3): 418- Studies 34(1):63-93 434 AFP (Agence France Presse) 2004. Fire damages one Johnston, R.J., Gregory, D., Pratt, G., and Watts., M. of Ethiopia’s indigenous forests. http://www. 2000. The Dictionary of Human Geography. geocities.com/akababi/wondo1.htm accessed Blackwell, Oxford. August 15, 2006. Lackey, R. T. 1998. Seven pillars of ecosystem Barnes, B.V., Zak, D. R., Deaton, S. R., Spurr, S. H. management. Landscape and Urban Planning 40: 1998. Forest Ecology. John Wiley and Sons: New 21-30 York Larson, A. M., 2003. Decentralisation and Forest Cadenasso, M.L., Pickett, S.T.A., Weathers, K.C., Management in Latin America: Towards a Jones, C.G. 2003. A Framework for a Theory of Working Model. Public Administration and Ecological Boundaries. BioSciences 53(8):750-758 Development 23:211-226 Campbell, B., Sayer, J. A., Frost, P., Vermeulen, S., Lebel, L., Anderies, J.M., Campbell, B., Folke, C., Perez, M. R., Cunningham, A., and Prabhu, R. Hatfied-Dodds, S., Hughes, T. P., and Wilson, J. 2001. Assessing the Performance of Natural 2006. Governance and the Capacity to Manage Resources Systems. Conservation Ecology 5(2): Resilience in Regional Social-Ecological Systems. 22. [online] URL: http://www.consecol.org/vol5/ Ecology and Society 11(1): 19. [online] URL: iss2/art22/ http://www.ecologyandsociety.org/vol11/iss/art19/ Cumming, G. S., Cumming, D. H., and Redman, C. Lovell, C., Mandondo, A., Moriarty, P. 2002. The L., 2006. Scale Mismatches in Social-Ecological Question of Scale in Integral Natural Resource Systems: Causes, Consequences, and Solutions. Management. Conservation Ecology 5(2):25. Ecology and Society 11(1): 14 [online] URL: [online] URL: http://www.consecol.org/vol5/iss2/ http://www.ecologyandsociety.org/vol11/iss/art14/ art/25/ Demel, T., and Mulugeta, L. 2005. Role of Forestry Makin, M.J., Kingham, T.J., Waddams, A.E., Birchall, in Poverty Alleviation in Ethiopia. Unpublished C.J. and Teferra, T. 1975. Development prospects paper. Debub University, Wondo Genet College of in the southern rift valley, Ethiopia (Vol.1&2).

14 Ecological and human spatial boundaries in Wondo Genet, Ethiopia

Land resource study 21. Land Resources Division, 23:53-65 Ministry of Overseas Development: England. Storey, D. 2001. Territory: The Claiming of Space. Mccarthy, J.F. 2002. Turning in Circles: District Prentice Hall, London. Governance, Illegal Logging and Environmental Strayer, D.L., Power, M. E., Fagan, W.F., Picket, Decline in Sumatra, Indonesia. Society and S.T.A., Belnad, J. 2003. A Classification of Natural Resources 15:867-886 Ecological Boundaries. Bioscience 53(8):723-729 Meidinger, E.E. 1998. Laws and Institutions in Cross- Swyngedouw, E. 2004. Scaled Geographies: Nature, Boundary Stewardship. In Knitigt, L.R and Place, and the Politics of Scale. In: Scale & Landres, B.P. Stewardship Across Boundaries, Geographic Inquiry - Nature, Society, and Method. Island Press, London, p87-110 Editors Eric Sheppard and Robert B. McMaster. Melaku, B. 2003. Forest Property Rights, the Role Blackwell, London. p 129-148. of the State, and Institutional Exigency: The Unruth, J.D. 2006. Changing conflict resolution Ethiopian Experience. Doctoral Dissertation. institution in Ethiopian pastoral commons: the role Swedish University of Agricultural Sciences. of armed confrontation in rule making. Geojournal Uppsala, Sweden. 64 (3) 225-237 Metzger, J.P. and Muller, E. 1996. Characterizing the Walker, P.A. 1999. Democracy and environment: complexity of landscape boundaries by remote congruencies and contradictions in southern sensing. Landscape Ecology 11(2): 65-77 Africa. Political Geography 18: 257-284. Obidzinski, K., Suramenggala, I. 2000. Informal Walker, P.A., Peters, P. E. 2001. Maps, Metaphors, and logging in East Kalimantan, social, economic and Meanings: Boundary Struggles and Village Forest political implications. CIFOR, Bogor, Indonesia. Use on Private and State Land in Malawi. Society Ostrom, E. 1999. Self-Governance and Forest and Natural Resources 14: 411-424 Resources. CIFOR, Occasional paper No 20. WGCF (Wondo Genet College of Forestry). 2004. Puth, L. M., and Wilson, K.A. 2001. Boundaries and Proceedings of the international workshop on Corridors as a continuum of Ecological Flow Development-Oriented Interdisciplinary Thematic Control: Lessons from Rivers and streams. Action Research (DOIT-AR), March 7-11, 2004. Conservation Biology 15(1): 21-30 Wondo Genet, Ethiopia. Reporter. 2005. Preservation or profits? Ethiopia’s WIC (Walta Information Center). 2004. Wondo forest dwindling forests. http://www.ethiopianreporter. ablaze: Indigenous trees more than a century old com/modules.php?name=News&file=print&sid=5 under attack. http://www.geocities.com/akababi/ 327 accessed August 15, 2006 wondo.htm , accessed August 15, 2006 Ribot, J.C. 2003. Democratic Decentralisation of Zerihun, M.A. 1999. Natural resources competition Natural Resources: Institutional Choice and and interethnic relations in Wondo Genet, Discretionary Power Transfers in Sub-Saharan South-central Ethiopia. MA thesis, Addis Ababa Africa. Public Administration and Development University, Addis Ababa

DISSERTATIONS IN PHYSICAL GEOGRAPHY 1944–2007

The Department of Physical Geography, Stockholm University Doctor of Philosophy Dissertations before 1994

Sigurdur Thorarinsson, 1944. Tefrokronologiska studier på Island. Þjórsárdálur och dess förödelse. Fakultetsopponent: Prof. Lennart von Post.* Carl Mannerfelt, 1945. Några glacialmorfologiska formelement. Fakultetsopponent: Prof. Veinö Tanner.* O. Harald Johnsson, 1946. Termisk-hydrologiska studier i sjön Klämmingen. Fakultetsopponent: Prof. B. Hellström.* Carl C. Wallén, 1949. Glacial meteorological investigations on the Kårsa Glacier in Swedish Lappland 1942–1948. Fakultetsopponent: Prof. Harald Ulrik Sverdrup.* Carl-Gustav Holdar, 1957. Deglaciationsförloppet i Torneträsk-området. Fakultetsopponent: Doc. Carl C. Wallén. Valter Schytt, 1958. A. Snow studies at Maudheim. B Snow studies inland. C. The inner structure of the ice shelf at Maudheim as shown by core drilling. Fakultetsopponent: Doc. Carl C. Wallén. Gunnar Østrem, 1965. Studies of ice-cored moraines. Fakultetsopponent: Doc. Valter Schytt. P. J. Williams, 1969. Properties and behaviour of freezing soils. Fakultetsopponent: Prof. Gunnar Beskow. Leif Wastenson, 1970. Flygbildstolkning av berghällar och blockmark. Metodstudier av tolkningsmöjligheten i olika flygbildsmaterial. Fakultetsopponent: Prof. Harald Svensson. Lars-Erik Åse, 1970. Shore-displacement in Eastern Svealand and Åland during the last 4,000 years. Fakultetsopponent: Prof. John Norrman. Stig Jonsson, 1970. Strukturstudier av subpolär glaciäris från Kebnekaiseområdet. Fakultetsopponent: Doc. Gunnar Østrem. Bo Strömberg, 1971. Isrecessionen i området kring Ålands hav. Fakultetsopponent: Doc. Erik Fromm. Erik Bergström, 1973. Den prerecenta lokalglaciationens utbredningshistoria inom Skanderna. Fakultetsopponent: Doc. Ragnar Dahl. Kurt Viking Abrahamsson, 1974. Äkäslompolo-områdets glacialmorfologi och deglaciation. Fakultetsopponent: Prof. Veikko Okko. Dietrich Soyez, 1974. Geomorfologiska inventeringar och deglaciationsstudier i Dalarna och Västerbotten. Fakultetsopponent: Prof. Paul Fogelberg. Weston Blake Jr., 1975. Studies of glacial history in the Queen Elizabeth Islands, Canadian arctic archipelago. Fakultetsopponent: Dr. Jan Mangerud. Wibjörn Karlén, 1976. Holocene climatic fluctuations indicated by glacier and tree-limit variations in northern Sweden. Fakultetsopponent: Doc. Jan Lundqvist. Bengt Lundén, 1977. Jordartskartering med flygbildsteknik. En metodundersökning i olika bildmaterial. Fakultetsopponent: Doc. Jan O. Mattsson. Margareta Ihse, 1979. Flygbildstolkning av vegetation. Metodstudier för översiktlig kartering. Fakultetsopponent: Prof. Måns Ryberg. Lill Lundgren, 1980. Soil erosion in Tanzanian mountain areas. Fakultetsopponent: Doc. Lennart Strömquist Kurt-Åke Zakrisson, 1980. Vårflödesprognoser med utgångspunkt från snötaxeringar i Malmagenområdet. Fakultetsopponent: Doc. Thorsten Stenborg. Olle Melander, 1980. Inlandsisens avsmältning i norvästra Lappland. Fakultetsopponent: 1. amanuensis Johan Ludvig Sollid. Ann-Cathrine Ulfstedt, 1980. Fjällen i Västerbotten och södra Norrbotten. Anmärkningar om geomorfologi och isrecession. Fakultetsopponent: Doc. Bo Strömberg. Wolter Arnberg, 1981. Multispectral reflectance measurements and signature analysis. Fakultetsopponent: Prof. Sipi Jaakkola. Carl Christiansson, 1981. Soil erosion and sedimentation in Semi-arid Tanzania. Fakultetsopponent: Doc. Lennart Strömquist. Johan Kleman, 1985. The spectral reflectance of coniferous tree stands and of barley influenced by stress. An analysis of field measured spectral data. Fakultetsopponent: Prof. Dag Gjessing. Lennart Vilborg, 1985. Cirque forms in northern and central Sweden. Studies 1959–1985. Fakultetsopponent: Prof. Anders Rapp. Per Holmlund, 1988. Studies of the drainage and the response to climatic change of Mikkaglaciären and Storglaciären. Fakultetsopponent: Dr. Helgi Björnsson. Henrik Österholm, 1989. The glacial history of Prins Oscars Land Nordaustlandet, Svalbard. Fakultetsopponent: Doc. Harald Agrell. Laine Boresjö, 1989. Multitemporal analysis of satellite data for vegetation mapping in Sweden. Fakultetsopponent: Prof. Friedrich Quiel. Ingmar Borgström, 1989. Terrängformerna och den glaciala utvecklingen i södra fjällen. Fakultetsopponent: Prof. Johan Ludvig Sollid. Christian Bronge, 1989. Climatic aspects of hydrology and lake sediments with examples from northern Scandinavia and Antarctica. Fakultetsopponent: Doc. Stig Jonsson. Kjell Wester, 1992. Spectral signature measurements and image processing for geological remote sensing. Fakultetsopponent: Prof. Friedrich Quiel. Gunhild Rosqvist, 1992. Late Quaternary equatorial glacier fluctuations. Fakultetsopponent: Prof. Chalmers Clapperton. Maj-Liz Nordberg, 1993. Integration of remote sensing and ancillary data for geographical analysis and risk assessment of forest regeneration failures. Fakultetsopponent: Prof. Stein Bie. * Dissertations in Physical Geography from the Department of Geography, Stockholm University College

The Department of Physical Geography, Stockholm University Dissertation Series, 1994–2000

Peter Jansson, 1994. Studies of short-term variations in ice dynamics, Storglaciären, Sweden. Dissertation No. 1. Fakultetsopponent: Prof. Hans Röthlisberger. Elisabeth Isaksson, 1994. Climate records from shallow firn cores, Dronning Maud Land, Antarctica. Dissertation No. 2. Fakultetsopponent: Lektor. Claus Hammer. Karin Holmgren, 1995. Late Pleistocene climatic and environmental changes in Central Southern Africa. Dissertation No. 3. Opponent: Prof. Stephen Trudgill. Pius Zebhe Yanda, 1995. Temporal and spatial variations of soil degradation in Mwisanga Catchment, Kondoa, Tanzania. Dissertation No. 4. Fakultetsopponent: Doc. Jonas Åkerman. Anders Moberg, 1996. Temperature variations in Sweden since the 18th century. Dissertation No. 5. Fakultetsopponent: Dr. Phil D. Jones. Arjen P. Stroeven, 1996. Late Tertiary glaciations and climate dynamics in Antarctica: Evidence from the Sirius Group, Mount Fleming, Dry Valleys. Dissertation No. 6. Fakultetsopponent: Prof. David Sugden. Annika Dahlberg, 1996. Interpretations of environmental change and diversity: A study from North East District, Botswana. Dissertation No. 7. Fakultetsopponent: Prof. Andrew Warren. Helle Skånes, 1996. Landscape change and grassland dynamics. Retrospective studies based on aerial photographs and old cadastral maps during 200 years in South Sweden. Dissertation No. 8. Fakultetsopponent: Prof. Hubert Gulinck. Clas Hättestrand, 1997. Ribbed moraines and Fennoscandian palaeoglaciology. Dissertation No. 9. Fakultetsopponent: Dr. Chris Clark. Guðrún Gísladóttir, 1998. Environmental characterisation and change in South-western Iceland. Dissertation No. 10. Fakultetsopponent: Prof. Jesper Brandt. Jens-Ove Näslund, 1998. Ice sheet, climate, and landscape interactions in Dronning Maud Land, Antarctica. Dissertation No. 11. Fakultetsopponent: Prof. David Sugden. Mats G. Eriksson, 1998. Landscape and soil erosion history in Central Tanzania. A study based on lacustrine, colluvial and alluvial deposits. Dissertation No. 12. Fakultetsopponent: Prof. Michael Thomas. Nkobi M. Moleele, 1999. Bush encroachment and the role of browse in cattle production. An ecological perspective from a bush encroached grazing system, Olifants Drift, Kgatleng District, Southeast Botswana. Dissertation No. 13. Fakultetsopponent: Prof. Norman Owen-Smith. Lars-Ove Westerberg, 1999. Mass movement in East African Highlands: Processes, effects and scar recovery. Dissertation No. 14. Fakultetsopponent: Prof. Ana L. Coelho Netto. Malin M. Stenberg, 2000. Spatial variability and temporal changes in snow chemistry, Dronning Maud Land, Antartica. Dissertation No. 15. Fakultetsopponent: Prof. Jon-Ove Hagen. Ola Ahlqvist, 2000. Context sensitive transformation of geographic information. Dissertation No. 16. Fakultetsopponent: Prof. Peter Fisher.

The Department of Physical Geography and Quaternary Geology, Stockholm University Thesis in Geography with emphasis on Physical Geography 2001–2006

Sara A. O. Cousins, 2001. Plant species diversity patterns in a Swedish rural landscape: Effects of the past and consequences for the future. Dissertation No. 17. Fakultetsopponent: Dr. Roy Haines- Young. Cecilia Richardson-Näslund, 2001. Spatial distribution of snow in Antarctica and other glacier studies using ground-penetrating radar. Dissertation No. 18. Fakultetsopponent: Prof. Robert W. Jacobel. Thomas Schneider, 2001. Hydrological processes in firn on Storglaciären, Sweden. Dissertation No. 19. Fakultetsopponent: Prof. Andrew Fountain. Hans W. Linderholm, 2001. Temporal and spatial couplings between tree-ring variability and climate in Scandinavia. Dissertation No. 20. Fakultetsopponent: Dr. Astrid Ogilvie. Marianne I. Lagerklint, 2001. Marine multi-proxy records of late Quaternary climate change from the Atlantic Ocean. Dissertation No. 21. Fakultetsopponent: Dr. Lloyd H. Burckle Richard Y. M. Kangalawe, 2001. Changing land-use patterns in the Irangi hills, central Tanzania. A study of soil degradation and adaptive farming strategies. Dissertation No. 22. Fakultetsopponent: Prof. William Adams. Anders Clarhäll, 2002. Glacial Erosion Zonation - Perspectives on topography, landforms, processes and time. Dissertation No. 23. Fakultetsopponent: Dr. Chris Clark. Krister N. Jansson, 2002. Glacial geomorphology of north-central Labrador-Ungava, Canada. Dissertation No. 24. Fakultetsopponent: Dr. Andrée Bolduc. Björn E. Gunnarson, 2002. Holocene climate and environmental fluctuations from subfossil pines in central Sweden. Dissertation No. 25. Fakultetsopponent: Prof. Mike G. L. Baillie. Katarina Lövenhaft, 2002. Spatial and temporal perspectives on biodiversity for physical planning. Dissertation No. 26. Fakultetsopponent: Prof. Jan Bengtsson. Anna Allard, 2003. Vegetation changes in mountainous areas. A monitoring methodology based on aerial photographs, high-resolution satellite images, and field investigations. Dissertation No. 27. Fakultetsopponent: Doc. Timo Helle. Per Klingbjer, 2004. Glaciers and climate in northern Sweden during the 19th and 20th century. Dissertation No. 28. Fakultetsopponent: Dr. Georg Kaser. Ola Fredin, 2004. Mountain Centered Icefields in Northern Scandinavia. Dissertation No. 29. Fakultetsopponent: Prof. Jon Landvik. Johan M. Bonow, 2004. Palaeosurfaces and palaeovalleys on North Atlantic previously glaciated passive margins- reference forms for conclusions on uplift and erosion. Dissertation No. 30. Fakultetsopponent: Dr. Adrian Hall. Richard Pettersson, 2004. Dynamics of the cold surface layer of polythermal Storglaciären, Sweden. Dissertation No. 31 Fakultetsopponent: Prof. Helgi Björnsson. Katarina Lundblad, 2006. Studies on Tropical Palaeo-variation in Climate and Cosmic Ray Influx. Geochemical Data from Stalagmites Collected in Tanzania and Northern South Africa. Dissertation No. 32. Fakultetsopponent: Prof. Augusto Mangini. Lena Rubensdotter, 2006. Alpine lake sediment archives and catchment geomorphology; causal relationships and implications for paleoenvironmental reconstructions. Dissertation No. 33. Fakultetsopponent: Prof. Catherine Souch.

Dissertations from the Department of Physical Geography and Quaternary Geology 2006–

Håkan Grudd, 2006: Tree rings as sensitive proxies of past climate change. Dissertation No. 1. Fakultetsopponent: Prof. Brian Luckman. Ulf Jonsell, 2006. Sulfur in polar ice and snow. Interpretations of past atmosphere and climate through glacial archives. Dissertation No. 2. Fakultetsopponent: Dr. Mark Curran. Hanna S. Sundqvist, 2007: Speleothems as environmental recorders - A study of Holocene speleothems and their growth environments in Sweden. Dissertation No. 3. Fakultetsopponent: Prof. Frank McDermott. Patrik Klintenberg, 2007: More water, less grass? An assessment of resource degradation and stakeholders’ perceptions of environmental change in Ombuga grassland, northern Namibia. Dissertation No. 4. Fakultetsopponent: Prof. Stein Bie. Maria Ryner, 2007. Past environmental and climate changes in northern Tanzania. Vegetation and lake level variability in Empakaai Crater. Dissertation No. 5. Fakultetsopponent: Prof. Henry Lamb. Daniel S. Veres, 2007: Terrestrial response to Dansgaard-Oeschger cycles and Heinrich events: the lacustrine record of Les Echets, south-eastern France. Dissertation No. 6. Fakultetsopponent: Prof. John J. Lowe. Yoshihiro Shibuo, 2007: Modelling water and solute flows at land-sea and land-atmosphere interfaces under data limitations. Dissertation No. 7. Fakultetsopponent: Dr. Clifford Voss.