FINAL PROGRAMATIC REPORT - SAVE THE TIGER FUND (2003-0087-024)

This Report consists of two Master’s Students Theses:

1) RESOURCES USE AND CONSERVATION ATTITUDES OF LOCAL PEOPLE IN

THE WESTERN TERAI LANDSCAPE, NEPAL - Nabin Baral.

2) ANALYSIS OF FORESTS UNDER DIFFERENT MANAGEMENT REGIMES IN

THE WESTERN TERAI OF NEPAL AND ITS RELATION TO ENVIRONMENT

AND HUMAN USE -Nilesh Timilsina

FLORIDA INTERNATIONAL UNIVERSTIY

Miami, Florida

RESOURCES USE AND CONSERVATION ATTITUDES OF LOCAL PEOPLE IN

THE WESTERN TERAI LANDSCAPE, NEPAL

A thesis submitted in partial fulfillment of the

requirements for the degree of

MASTER OF SCIENCE

in

ENVIRONMENTAL SCIENCE

by

Nabin Baral

2005 To: Dean Mark Szuchman College of Arts and Sciences

This thesis, written by Nabin Baral, and entitled Resources Use and Conservation Attitudes of Local People in the Western Terai Landscape, Nepal, having been approved in respect to style and intellectual content, is referred to you for judgment.

We have read this thesis and recommend that it is approved.

______Dr David B. Bray

______Dr Mahadev G. Bhat

______Dr Joel T. Heinen, Major Professor

Date of Defense: July 13, 2005

The thesis of Nabin Baral is approved.

______Dean Mark Szuchman College of Arts and Sciences

______Dean Douglas Wartzok University Graduate School

Florida International University, 2005

ii

DEDICATION

I dedicate this thesis to my TEACHERS of past, present and future.

iii ACKNOWLEDGMENTS

I am grateful to many individuals and institutions who have contributed to the success of

this research project. I thank the members of my thesis committee. I express my gratitude

to my Major Professor, Dr Joel T. Heinen for his inspiration, guidance, and meticulous

editing of solecisms. His positive attitudes were valuable in times of hardship during the

fieldwork. I am indebted to my committee members: Dr Mahadev G. Bhat and Dr David

B. Bray, who enlightened me with an ocean of knowledge. They were always interested

in the project and provided me invaluable suggestions.

I could not have completed this thesis without financial assistance from numerous

institutions. I thank the National Fish and Wildlife Foundation’s Save the Tiger Fund and

the Disney Conservation Fund for providing research grant that covered most of the field expenditures. I thank the Sigma Xi and the Institute of Asian Study, Florida International

University for awarding a grant-in-aid for research and a travel grant, respectively. The

Department of Environmental Studies awarded graduate assistantships.

I thank the Department of National Parks and Wildlife Conservation for granting permission to carry out research work in Royal Bardia National Park and Royal

Suklaphanta Wildlife Reserve. Many staff of these parks helped me during the fieldwork.

The King Mahendra Trust for Nature Conservation (KMTNC) provided accommodations in their field offices. I am thankful to Dr Shant Raj Jnawali, Naresh Subedi and Chiranjivi

Pokharel, affiliated with the KMTNC, for their support in the field research. I appreciate

iv hard work of my field assistants Kamal Thapa, Birendra Thapa, Janak Chaudhary,

Rajendra Bhatta, and Dhan Chand.

I am indebted to the alacrity of Thaneswar Tiwari and his family members when I was their guest in Mahendranagar. Birendra Tiwari, Dhan Bahadur Chand, Benu Gautam, and

Prakash Tamang helped me in various ways.

I want to thank Dr Paulette Johnson and Oscar Saenz for their help in statistical consulting and Ms Sushila Nepali, WWF-Nepal Program, for providing literature on the

Terai Arc Landscape project.

Nilesh it was great working with you.

I am lucky to have loving parents and wife who were always supportive. Thanks for your patience and perseverance when I was not at home.

v ABSTRACT OF THE THESIS

RESOURCES USE AND CONSERVATION ATTITUDES OF LOCAL PEOPLE IN

THE WESTERN TERAI LANDSCAPE, NEPAL

by

Nabin Baral

Florida International University, 2005

Miami, Florida

Professor Joel T. Heinen, Major Professor

Two protected areas: Royal Bardia National Park (RBNP) and Royal Suklaphanta

Wildlife Reserve (RSWR) in the Western Terai, Nepal, are under threats due to present political turmoil, uncontrolled immigration, inefficient land reform policies and unsustainable resource use. I did a stratified random questionnaire survey of 234 households to determine how resource use patterns and problems influence conservation attitudes. Chi-square, Student’s t, Mann-Whitney and Kruskal-Wallis tests, and multiple regression were used. There was spatio-temporal variability in resource use patterns and dependency. People were collecting eight and seven types of resources in RBNP and

RSWR, respectively. However, people in RBNP were more dependent on resources than

RSWR. In both areas, the problem of firewood is serious. The mean attitude score of

RBNP (8.4 ± 1.44) was significantly higher than the score of RSWR (7.7 ± 1.66; t = 3.24, p = 0.0007). Training received, wildlife damage and satisfaction with participation in user groups were significant predictors of conservation attitudes.

vi TABLE OF CONTENTS

CHAPTER PAGE

1 INTRODUCTION 1

1.1 Evolution of the Conservation Concept 1 1.1.1 Fortress and fine conservation 2 1.1.2 Participatory conservation 3 1.1.3 Landscape conservation 5 1.2 Literature Review 8 1.3 Research Goal, Questions and Hypotheses 12

2 THE STUDY AREA 14

2.1 Ecological Divisions 14 2.2 Location 15 2.3 Geology and Soils 16 2.4 Climate 17 2.5 Vegetation 18 2.6 Wildlife 19 2.7 Settlements 20 2.8 Ethnographies 20

3 MATERIALS AND METHODS 24

3.1 Selection of Villages 24 3.2 Household Survey 24 3.2.1 Household survey team 24 3.2.2 Survey instrument 25 3.2.3 Survey procedure 25 3.3 Sample Size and Sampling 26 3.4 Survey of User Groups (UGs) 27 3.5 Survey Limitations and Constraints 27 3.6 Statistical Analyses 28

4 RESULTS 30

4.1 Demographic Characteristics of Respondents 30 4.1.1 Gender and age 30 4.1.2 Ethnicity 31 4.1.3 Education 31 4.1.4 Occupation 32 4.1.5 Family size 33 4.2 Migration and Economic Status 34

vii 4.3 Resource Use Patterns and Dependency 37 4.4 Participation in Conservation 42 4.5 Conservation Attitudes 43 4.6 Wildlife Harassment and Attitudes towards UGs 46 4.7 Institutional Development and Strengthening 47 4.7.1 Institutional working procedures 47 4.7.2 Resource management and demand fulfillment 49 4.7.3 Attitudes of conservation leaders 50 4.7.4 Content analysis of the work plans 51

5 DISCUSSION 52

6 CONCLUSIONS 69

REFERENCES 86

APPENDICES 96

viii LIST OF TABLES

TABLE PAGE

Table 1. Frequency distribution of ethnicity in two protected areas 73

Table 2. Frequency distribution of education level of respondents of two areas 73

Table 3. Comparison of family size between area and ethnic groups 73

Table 4. Percentage of immigrants in two protected areas 73

Table 5. Average landholdings in hectare among different ethnic groups in two areas 74

Table 6. Percent of respondents meeting need of staple food from their farm 74

Table 7. Average and range of livestock size unit in two protected areas 74

Table 8. Frequency of resources harvested by respondents of two protected areas 74

Table 9. Nonparametric correlation of resource use score with continuous variables 75

Table 10. Frequency distribution of resource dependency in two areas 75

Table 11. Number and percentage of households mentioning the grazing sites 75

Table 12. Number and percentage of households mentioning the source of fodder 75

Table 13. Number and percentage of households mentioning the source of energy 75

Table 14. Number and percentage of respondents suggesting measures to solve the problem of firewood scarcity 76

Table 15. Frequency distribution of households participating in conservation interventions in two areas 76

Table 16. Percent of respondents agreeing or disagreeing with conservation statements 77

ix Table 17. Multiple regression of conservation attitude score on demographic and socioeconomic variables 78

Table 18. Frequency distribution of mode of Users’ Group formation 78

Table 19. Perception of UG chairs on effective institute for resource management 78

Table 20. Frequency distribution of responses whether UGs are fulfilling demands 79

Table 21. UG chairs’ attitudes towards TAL in two areas 79

Table 22. UG chairs attitudes towards the BZMR and Guideline 79

Table 23. Losses of agencies under the Ministry of Forest and Soil Conservation 79

x LIST OF FIGURES

FIGURE PAGE

Figure 1. Map of Nepal depicting spatial distribution of protected areas 80

Figure 2. RBNP and RSWR with sampled households in the buffer zones 81

Figure 3. Average minimum and maximum monthly temperatures of two areas for the period 1987-2001 82

Figure 4. Average monthly precipitation of two areas for the period 1987-2001 82

Figure 5. Percentage of male and female respondents in two protected areas 83

Figure 6. Percent of respondents in five occupation categories 83

Figure 7. Average family size among ethnic groups in two areas 84

Figure 8. Average landholdings among ethnic groups in two areas 84

Figure 9. Frequency of households rearing four types of livestock in two areas 85

Figure 10. Total number of tourist arrival per year since the beginning of the Maoists insurgency 85

xi ACRONYMS AND ABBREVIATIONS

BZMR Buffer Zone Management Regulation

CBC Community-based Conservation

CBS Central Bureau of Statistics

CITES Convention on International Trade in Endangered Species of Wild Fauna and Flora

DNPWC Department of National Parks and Wildlife Conservation

HMG His Majesty Government

IUCN World Conservation Union

LSU Livestock Size Unit

MFSC Ministry of Forests and Soil Conservation

NGO Non-Governmental Organization

NTFP Non-Timber Forest Product

RBNP Royal Bardia National Park

RSWR Royal Suklaphanta Wildlife Reserve

TAL Terai Arc Landscape

UG User Group

UNESCO United Nations Educational Scientific and Cultural Organization ha hectare kg kilogram mm millimeter sq km square kilometer

xii 1 INTRODUCTION

1.1 Evolution of the Conservation Concept

Developing countries are facing the dilemma of balancing conservation and human needs. They have limited natural resources and populations are growing at exponential rates, while productivity gains through improved agricultural production lag behind.

Economic development and population growth in the absence of more sustainable development is resulting in environmental degradation. Setting up a network of protected areas for the conservation of biodiversity is a step to address environmental degradation.

The Nepalese government adopted the concept of protected areas in the early 1970s. The modern era of conservation of both natural areas and species began with the passage of

National Parks and Wildlife Conservation (NPWC) Act 1973 by His Majesty’s

Government of Nepal (HMG 1973; Heinen & Kattel 1992). The Act established the

Department of National Parks and Wildlife Conservation (DNPWC), and also gave authority to DNPWC to declare any piece of land, forest or wetland important from biological, sociological and historical perspectives within the network of protected areas.

A series of protected areas was established to conserve vulnerable habitats and endangered species. At present, there are four types of protected areas established by the law which fall in three categories recognized by World Conservation Union (IUCN).

These are: nine national parks (Category II), three wildlife reserves (Category IV), one hunting reserve (Category VI) and three conservation areas (Category VI; Sharma &

Yonzon 2005). These are scattered throughout the country (Figure 1). The Fourth

Amendment of NPWC Act 1973, passed in 1993, vested the DNPWC the authority to declare buffer zones in the periphery of protected areas (Heinen & Mehta 2000). So far,

1 seven designated buffer zones (Category VI) have added significant areas under special management and protection (Sharma & Yonzon 2005). Two national parks are included as UNESCO World Heritage Sites and four wetlands are designated as Ramsar sites under IUCN-the World Conservation Union. Including buffer zones, about 18% of the total land area (147,181 sq km) of Nepal is under protected status. The evolution of the conservation concept is passing through a ‘learning curve’ phase and there is always an opportunity to improve. Nepal has tried different approaches to conservation and their impacts on ecology, economics and society are briefly summarized.

1.1.1 Fortress and fine conservation: Nepal’s conservation movement began with active protection of species and habitats. The approach was to replicate the prototype of

Yellowstone and local people were excluded from these lands. This conservation philosophy was based on protectionist policies that discouraged every form of resource use. The people who had been using available resources lost their traditional usufruct rights. The goal of management was to restore populations of flagship species, and there have been notable advances in some parks (Mishra 1982). The population of one-horned rhinoceros Rhinoceros unicornis reached 612 individuals in 2000 from all time low of 55 individuals in the 1960s. Similarly, the population of Bengal tiger Panthera tigris increased from 55 individuals in the 1970s to 300 individuals in 2000 (DNPWC 2001).

The control of poaching and protection of habitats allowed wildlife populations to increase. The ‘preservation-oriented’ approach was successful in conserving endangered species of wildlife (Heinen & Yonzon 1994) at a high social cost. The approach was easier to conceptualize and had measurable success (Heinen & Mehta 2000), but some

2 second generation issues such as park-people conflicts, ownership of resources and equitable distribution of benefits have emerged. Reconciling issues of strict protection versus sustainable livelihoods may garner greater support for conservation in developing countries. Soliciting people’s participation in conservation is a well thought strategy to this end.

1.1.2 Participatory conservation: As a general rule, extraction of resources inside protected areas is prohibited (HMG 1973) but if the park authority deems it necessary, then some resources can be extracted for use by local people. In response to the perceived tensions caused by the designation of protected areas and by way of compensation for loss of rights to various resources inside the protected area, the granting of some resource extraction rights, such as the collection of grasses for two weeks per year, developed in

1978, five years after the establishment of protected areas (Mishra 1982). Management authorities realized that conservation is counterproductive if stringent laws are enacted to forbid local people from subsistence resource use. Resources such as thatch and firewood decreased substantially outside the protected areas and local people have no alternative other than harvest illegally from the park.

Exclusion of local people and restrictions on local-level usury rights have made the

‘fortress and fine’ model less appropriate in the context of developing countries such as

Nepal (Heinen 1996). Furthermore, this approach of conservation brings hardships to poor, rural people living in or around the protected areas who are heavily dependent on natural resources for their living. While biodiversity conservation brings global benefits,

3 there can be few benefits and high costs to local communities. Even with the concession

of grass harvesting once a year, many model parks and reserves failed to solicit favorable

attitudes and park-people relationships were poor (Mishra 1982; Heinen 1993; Nepal &

Weber 1995; Studsrod & Wegge 1995; Mehta 1996). After realizing the fact that

protected areas cannot operate as islands in the matrix of human habitation, the

government gradually changed its policy from exclusion to inclusion of local people in

protected areas management.

After 1980, conservation communities worldwide realized that humans are an integral part of ecosystems, so that, for the sustainability of the ecosystem, human dimensions in conservation should be aptly addressed. The publication of the IUCN’s World

Conservation Strategy of 1980 has been a catalyst for more ‘all-encompassing’ conservation thinking (Infield 1988). Multi-national donor agencies, non-governmental organizations (NGO) and foreign governments set criteria of participation by and empowerment of local people for funding in nature conservation (IUCN 1991; Kemf

1993; Gibson & Marks 1995). Lessons learned from past experiences and changes in policies of donor agencies spurred governments to pass legislation allowing for conservation areas in addition to more strictly protected areas such as national parks and wildlife reserves. However, there have been significant dissenting voices that suggest strict protection remains the highest priority for conservation interests (Brandon et al.

1998). Nonetheless, in Nepal public policy has pursued a more conciliatory approach. In conservation areas in Nepal, local people are empowered to some extent in the management and utilization of natural resources (HMG 1996). The ratification of Buffer

4 Zone Management Regulations 1996 provided the DNPWC with legal power to earmark a certain percent (30-50) of revenue generated by parks and reserves to local communities residing in buffer zones for various activities prioritized by local people, and the concession of resource harvest such as firewood, timber, thatch, fodder etc. for subsistence needs was provided within buffer forests. Various integrated conservation and development activities have been carried out in buffer zones and conservation areas to meet the dual goals of environmental protection and economic development. This participatory approach of management bolstered park-people relationships and attitudes towards conservation have improved in some parks (Heinen & Mehta 1999).

1.1.3 Landscape conservation: The size of protected areas in developing countries is in many cases too small to harbor viable populations of megafauna (Dinerstein &

Wikramanayake 1993). Protected areas are islands in human habitation and landuse patterns outside protected areas may not be always compatible with conservation. One of the challenges for conservationists is how to increase the functional size of protected areas in developing countries by increasing habitat even in intensively used areas.

Connecting already existing protected areas with habitat corridors and connectivity in human-dominated landscapes is one obvious approach to this end. Managing the entire landscape as an intact ecosystem will help secure the existence of protected areas in developing countries. The Nepalese government realized the imperative of the landscape approach and targeted the Terai region for implementation. The Nepalese Terai, the lowlands along Nepal’s southern border with India, is very important from the conservation perspective because it harbors internationally endangered megafauna and

5 has intact sub-tropical forests of high conservation value. The region boasts the most productive protected areas of Nepal.

The Terai Arc Landscape (TAL) encompasses one of the richest ecosystems in the eastern Himalaya ecoregion. It has an area of 49,500 sq km and extends from the

Bagmati River in Nepal to Yamuna River in India (MFSC 2004). The goal of the project is to create a single functioning landscape by connecting 11 protected areas of Nepal and

India through corridors. The landscape is very important for the survival and metapopulation dynamics of charismatic megafauna such as the Royal Bengal tiger, one- horned rhinoceros and Asian elephant Elephas maximus. The Nepali portion of TAL covers about half (23,199 sq km) of the total area, and comprises four protected areas, government forests and community forests. There are 86 species of mammals, 550 species of birds, 47 species of herpeto-fauna, 126 species of fishes, and over 2,100 species of flowering plants found in TAL (MFSC 2004). Recognizing the conservation importance of TAL, the government of Nepal identified it as a priority landscape for biodiversity conservation in the 10TH Five Year Plan (2003-7). The major thrust of the program is sectoral integration for planning, implementation, and monitoring and evaluation.

The Nepalese government ventured into an ambitious landscape conservation program that spans over 50 years. The success of the program would be determined by participation and attitudes of local people. Local people would be given incentives to become active stakeholders in conservation initiatives and influence decision-making

6 processes that ultimately determine their livelihood. The Landscape approach to conservation is promising as it integrates social, ecological, and economic sectors and uses synthetic principles for achieving a sustainable future across an entire landscape.

Within a short span of time, Nepal has come a long way in conservation. Although the

Nepalese government has not formally adopted adaptive management, in practice various conservation models have been designed, implemented and evaluated based on 'learning by doing' philosophy to narrow the rift between conservation and sustainable development. Lessons learned from different approaches were incorporated to improvise conservation rules and regulations. Nepal is now considered as a leader among developing nations with regard to conservation programs and legislation (Heinen &

Kattel 1992) and future courses of action for better management of resources will depend on attitudes and participation of local people who are the principal stakeholders in conservation. However, in the last several years, the Maoist People’s War is throttling the past achievements of the conservation sector (Appendix 1). The War started in 1996 and has had adverse consequences in social, political, and economic sectors. Since the inception of the War, the conservation sector has been badly hit and consequences are now appearing. To meet new challenges, conservation agencies should reconsider the present models (Price 2003). To this end, the evolution of landscape conservation is promising. My goal is to evaluate how various conservation models have facilitated resource allocation and influence conservation attitudes, and provide baseline data for social impact assessment to gauge effectiveness of landscape approach in the western

Terai of Nepal.

7 1.2 Literature Review

The conservation attitudes of local people residing around protected areas (PA) determine

the fate of protected areas in the long run. It is important for protected area managers to

explore what factors influence conservation attitudes. The literature suggests that

influential factors can be grouped into demographic, cultural and socioeconomic

phenomena (Ite 1996) that largely determine local support or resentment to PAs.

Although results are specific to geographical areas, demographic variables such as

gender, age, education, occupation and ethnicity are generally found to be significant

predictors of conservation attitudes (Fiallo & Jabcobson 1995; Mehta & Kellert 1998;

Gillingham & Lee 1999; Sah & Heinen 2001). Attitudinal surveys are indispensable tools for social impact assessment and are widely used in the conservation sector. Attitude surveys do not record actual behavior, so do not predict conservation actions. Favorable conservation attitudes may not always ensure desired action on the part of local people; however, probability of conservation actions increases if people have favorable attitudes.

Attitudinal surveys could be conservatively used as an indicator of participation by local people in collective actions.

Culture plays an important role in natural resource management. Traditional rituals and customs of tribal people can impose restrictions on the exploitation of resources

(Anderson 2001). Local people may also participate in conservation programs due to incentives from projects such as ecotourism. The Community Baboon Sanctuary of

Belize is an example where local people are voluntarily participating in howler monkey conservation (Hartup 1994), in the hopes that increased income from ecotourism will

8 compensate for agricultural lands left in corridors for the monkeys. With respect to culture, religion has been suggested to inculcate a high regard for wildlife in some parks of India, and local people have shown favorable attitudes to conservation in spite of substantial economic loss due to wildlife damage (Sekhar 1998). Resource use patterns of ethnic tribes and immigrants draw attention for sustainability of resources. Although there have been no significant differences in deforestation based on cultural and ethnic background (Sierra 1999), non-indigenous people can cause significantly more change in structure and composition of tropical forests (Nepstad et al. 1992). The dependency on resources could be a function of culture. The indigenous Tharus are more dependent on natural resources than other groups in South Asia (Brown 1997). Understanding cultural perspectives of resource use and how they influence conservation attitudes provides insights for strategic management. In recent years, because of the Maoist insurgency many people migrated from the mountain districts to the Western Terai. The large-scale migration has caused ethnic heterogeneity, which will influence resource use patterns and problems (Heinen 1996).

Conservation attitudes are generally influenced by the perceived costs and benefits of

PAs. People tend to have favorable attitudes to conservation when are allowed to use park resources (Newmark et al. 1993), and they tend to be alienated when restrictions are imposed (Heinen 1993). Whether local people should be given access rights over resources within parks remains a contentious issue. Resource use restrictions bring hardships and incur substantial costs to people residing in the periphery of the park. Local people frequently offset costs by harvesting resources from parks illegally (Straede &

9 Helles 2000). Resenting resource restrictions, people demand usufruct rights over resources within parks and reserves (Gullingham & Lee 1999). Success of conservation measures is gauged by tangible and immediate benefits that people derive (McNeely

1993), and this helps to solicit active participation of people in sustainable management of natural resources.

Economic incentives are very important tools to influence conservation attitudes. In general, people who receive goods and services personally have more favorable attitudes than those who don’t (Infield 1988; Newmark et al. 1993). Economic incentives even out incurred costs of local people and provide socioeconomic benefits. However, inequitable distribution of benefits engenders problems. Parry and Campbell (1992) found that local people had negative attitudes in spite of receiving substantial benefits from conservation in Africa because the rich benefited more from tourism. Distinctions should be made regarding how benefits are distributed at community and household levels. These provide guidance to resource managers on the use of economic incentives for sustainable resource management.

Natural resources are the basis of both subsistence and market based economics.

Sustainability of resources and economic development of the area depend on dynamic interactions between two. More emphasis is given to non-timber forest products (NTFP) than timber when concession of resources is granted within the park to meet subsistence needs and to earn cash (Hedge & Enters 2000) because local people are more dependent on NTFPs and they are easier to regulate. Marginal communities use forest resources in

10 varied fashions, timber for domestic use, fuelwood for energy, honey, fruit, mushrooms,

bushmeat for food and medicinal plants for traditional healing, and NTFP harvesting

complements the subsistence economy of poor and underprivileged people (Makwerere

1996). Sustainable harvesting of NTFPs is put forward as a tool for silvicultural

management (Mahapatra & Mitchell 1997; Schreckenberg 1999), and could generate

employment, income and economic development in the periphery of protected areas.

Research has shown that NTFPs contribute about 17% of household’s annual income in

South Asia (Mahapatra & Mitchell 1997). There are caveats; when the subsistence economy is overtaken by the market economy, resources use tends to be unsustainable

(Yonzon & Hunter 1991). Collection of NTFPs is an opportunity to reduce indirect and opportunity costs of conservation borne by local people residing in the periphery of parks. It has also been promoted as an integrated approach to ameliorate park-people relations and improve attitudes (Archabald & Naughton-Treves 2001).

Now it is a common practice of park management to address development needs of peripheral communities. However, some conservationists criticize the development approach of conservation because it may attract large number of immigrants due to greater economic opportunities and agricultural development assistance (Robinson 1993;

Oates 1995). Socially, immigration results in ethnic heterogeneity (Noss 1997) that may result in inter-ethnic conflicts in resources use and management. Ecologically, immigration disrupts ecosystems (Shreckernberg 1999), thus hastening the exploitation of resources to an unsustainable level. Many integrated conservation and development projects failed to achieve their goals when they could not curb development-induced

11 immigration (Southgate & Clark 1993; Oates 1995; Noss 1997). The Western Terai of

Nepal has received large number of immigrants over the past three decades due to social,

political and environmental reasons. Therefore, understanding resource use patterns of

immigrants is very important to form management strategies for sustainable resource use.

People’s attitudes are decisive to achieve conservation goals (Richards 1996), but park

management should not be guided solely by attitudes of local people (Infield & Namara

2001) because attitudes are volatile, do not necessarily reflect actual behavior, and some

events may have strong short-term influences over them. Nevertheless, management

should deem it necessary to incorporate attitudes in management strategies. Periodic

attitudinal surveys form baseline data for social impact assessments of conservation

interventions.

1.3 Research Goal, Questions, and Hypotheses

The Western Terai harbors two of the most productive protected areas of Nepal: Royal

Bardia National Park (RBNP) and Royal Suklaphanta Wildlife Reserve (RSWR). These

areas are cornerstones for biodiversity conservation, but are under threats due to present

political turmoil. Uncontrolled immigration, inefficient land reform policies, unsustainable resource use, and a dearth of research to provide data for sound policies are some pressing problems in the study area. The goal of the project is to assess resource use patterns and problems, and how they relate to conservation attitudes. This attitudinal

12 survey will provide guidance for policy and management decisions involved in design, implementation, review, and monitoring of landscape level conservation in Nepal.

There are subtle differences between RBNP and RSWR. The buffer zone (328 sq km) of

RBNP was declared in 1996 and local people received 30-50% of revenue generated from the park since then. The buffer zone of RSWR was not officially declared. In

RBNP, grassroots institutions were mandatory by conservation legislation while in

RSWR they were instituted for convenience. There were many more non-governmental organizations working for conservation and sustainable development in the buffer zone of

RBNP than RSWR. The status of RBNP is national park so tourism is promoted but that of RSWR is a wildlife reserve. RBNP is in rural setting while RSWR is closer to an urban area. Historically, ethic Tharus have used natural resources of these areas. The present day grasslands were created when people were evicted from the parks. Owing to these differences, I use a comparative study framework to meet the research goal. My research hypotheses are guided by the following questions:

Question 1. What are the demographic structures, socio-economic status, and ethnic composition of the study area?

Question 2. What are the types and patterns of natural resources extracted and used?

Question 3. How do resource use patterns and problems influence attitudes towards conservation?

13 Question 4. To what extent can the distribution of benefits from access to the protected areas ensure the support of local communities for conservation objectives?

Question 5. Do conservation intervention programs solicit more favorable attitudes?

Question 6. To what extent are local institutions strengthened by conservation programs?

Based on primary data collected by household questionnaire surveys, I test the following hypotheses:

Hypothesis 1. The resource use patterns among caste/ethnic groups will be different.

Hypothesis 2. People participating in conservation intervention programs have more favorable attitudes towards conservation than others.

Hypothesis 3. The attitudes towards conservation differ among local people between two protected areas (RBNP and RSWR).

2 THE STUDY AREA

2.1 Ecological Divisions

Nepal is vertically stratified into three ecological regions: the Mountain, the Hills and the

Terai. The Mountain region covers elevations between 4,877 m to 8,848 m above sea

14 level. This region consists of a large number of magnificent peaks of the Himalayas that are sources of perennial rivers. The terrain is rough and inhospitable and the climate is harsh. It occupies about 35% of the total land area (147,181 sq km) of Nepal but harbors only 7.3% of the population. Livestock tending and seasonal businesses are the vocation of local people because only 2% of its land is suitable for cultivation. The Hill region lies between the elevations of 610 m to 4,877 m above sea level. Landscape diversity in medium sized peaks, fertile valleys and basins is characteristics. The region accounts the largest share (42%) of the total area of which about one tenth of its area is suitable for cultivation. About 44.3% of total population resides in the region. The occupations of people include livestock rearing, cottage industries, and cultivation of cereal as well as cash crops. The Terai, being an extension of the Gangetic plains of India, forms a low flat land in the southern part of the country bordering India. It comprises 23% of the land area of the country and 48.4% of the total population. This area includes most agricultural land (40% is suitable for cultivation) and dense forest of the country. This region is also called the ‘granary’ of Nepal because fertile land with irrigation facilities permit the cultivation of a wide variety of crops such as paddy, wheat, maize, sugarcane, tobacco, and vegetables and two to three crops per year. The population of this region is increasing at a faster rate compared to the other two regions; due both to high birthrates and internal migration.

2.2 Location

The study area comprises two protected areas of the Nepalese Terai: Royal Bardia

National Park (RBNP) and Royal Suklaphanta Wildlife Reserve (RSWR; Figure 2). The

15 RBNP (81.46502 E and 28.44479 N) lies in the mid-western Terai adjoining the eastern bank of the Karnali River in Bardia District. It was established in 1976 (originally), and is the largest protected area in the Nepalese Terai (968 sq km and proposed extension of

550 sq km). The RSWR (80.22640 E and 28.84955 N) is located in Kanchanpur District of far-western Terai along the southern border of Nepal. It was gazetted in 1973 and covers an area of 305 sq km.

2.3 Geology and Soils

The Terai is the northern extension of Gangetic Plain. It is alluvial flood plain in the south and tertiary Siwalik in the north. The Siwalik is composed of coarsely bedded stone, crystalline rocks, clays and conglomerates. Soils are young and very shallow and are exposed to a great degree of erosion and landslide, with little potential for cultivation

(HMG 1971). At the base of the Siwalik range is the ‘Bhabar’ zone, consisting of gravelley soil which has been washed down from the foothills and accumulated at their base. The Bhabar is not suitable for agriculture and large tracts of forests remain here.

South of Bhabar is the Terai flatland, which consists of beds of silts, clay and gravel to great depths and is the most productive agriculture land in Nepal. Soils are predominantly brown or yellow brown sandy loams that are mostly calcareous and slightly alkaline

(HMG 1971).

The soil is moreover sandy loam throughout the western Terai. The depth of soil is high in flat lands and low in the hills. Especially in degraded forests, the nitrogen content is

16 poor. Big boulders are characteristic of soil of the Siwalik foothills. The elevation varies

from 152 m in the Terai to 1441 m in the Churia (Siwalik) Range.

RSWR is drained by a number of rivers and streams, including the Mahakali, which

demarcates the western boundary of Nepal with India, the Bauni, Chaudhara and Syali

Rivers. Major wetlands within the reserve are located on the floodplains of these rivers.

Eight oxbow lakes are found in the reserve (Sah 2002) of which Rani Tal (20 ha) and

Kalikich Tal (10 ha) are large and famous. RBNP is drained by the Karnali and Babai

Rivers, and there are many oxbow lakes within the park.

2.4 Climate

The region has a sub-tropical monsoonal climate with three distinct seasons: hot-dry

(March-June), monsoon (July-October), and cold-dry (November-February). For the

period 1987-2001, average monthly maximum and minimum temperatures of 38.3 0C and

9.5 0C were recorded in May 1996 and January1989, respectively in RBNP (Figure 3).

Likewise, average monthly maximum and minimum temperatures of 38.8 0C and 6.0 0C were recorded in May 1995 and January 1997, respectively in RSWR. The highest annual rainfall of 2798 mm, and the lowest annual rainfall of 1592 mm occurred in the year 1990 and 1992 respectively in RBNP. In RSWR, the highest and the lowest annual rainfall of

2375 mm and 1257 mm occurred in the year 1998 and 1992, respectively. The highest mean monthly rainfall occurred in July (680 mm) in RBNP and in August (635 mm) in

RSWR. The lowest mean monthly rainfall occurs in December (21 mm) in RBNP and in

17 March (3 mm) in RSWR (Figure 4). The highest monthly rainfall of 945 mm occurred in

July 1989 in RBNP, and of 1205 mm occurred in August 1995 in RSWR.

2.5 Vegetation

Dinerstein (1979) classified six major vegetation types in RBNP which were later modified by Jnawali and Wegge (1993) to seven major types:

• Sal Forest is characterized by Sal Shorea robusta and covers about 70% of the

total area. The main associated species with Sal are Terminalia tomentosa and

Buchanania latifolia.

• Khair-Sissoo Forest is a pioneer association on riverside gravel. This forest type is

dominated by Khair Acacia catechu and Sissoo Dalbergia sissoo. Two shrub

species, Murraya koenigii and Callicarpa macrophylla form dense under-stories.

• Moist Riverine Forest is distributed in patches along water courses and in

depressions. This forest is characterized by evergreen species such as Syzigium

cumini, Mallotus philippinensis, Ficus racemosa and Bombax ceiba.

• Mixed Hardwood Forests grow in well drained flat land. Adina cordifolia,

Casearia tomentosa, Garuga pinnata, Mitragyna parviflora are some common

tree species of this forest type.

• Wooded Grasslands are grass-covered areas with sparsely distributed trees.

Imperata cylindrica, Saccharum spontaneum, Vetiveria zizanoides, Cyperus

kyllingia are the most common grasses. Tree species such as Bombax ceiba, Adina

cordifolia, Bahunia malbarica and Mallotus philippinesis are also sparsely

distributed in this habitat.

18 • Phanta is short open grassland in previously cultivated fields. Imperata cylindrica

is the dominating grass species in this vegetation type.

• Flooded Grasslands are tall grasslands along floodplains. The dominant species

are Saccharum spontaneum, S. bengalensis, Phragmatis karka and Narenga

phorphyrocoma.

The vegetation of RSWR consists of forests, grasslands and wetlands. Sah (2002) classified forests into three types. Riverine Forest on the banks of rivers and streams are dominated by Acacia catechu and Dalbergia sissoo. In Mixed Deciduous Forest, there is an assemblage of species such as Adina cordifolia, Celtis tetrandra, Mallotus philippinensis, Syzygium cumini, and Trewia nudiflora. Sal Forest is found in relatively well-drained uplands. The species composition of the forest is Shorea robusta,

Lagestroemia parviflora, Terminalia belerica and Terminalia chebula.

RSWR is famous for the most extensive tracts of grasslands within the protected areas network of Nepal. Suklaphanta is the largest grassland which covers an area of 54 sq km.

Imperata cylindrica, Saccharum bengalensis, Saccharum spontaneum, Narenga porphyrocoma, and Desmostachya bipinata are the dominant species of this grassland.

2.6 Wildlife

The main objective of the establishment and management of both RBNP and RSWR was to conserve critical habitats for globally endangered Bengal tiger and its prey species.

The RSWR boasts the largest density of the tiger (Regmi 2000) and RBNP has a breeding

19 population of 35-40 individuals (DNPWC 2003). The RBNP harbors a known total of 53 species of mammals, ca 400 species of birds, 25 species of reptiles and amphibians and

121 species of fishes (DNPWC 2003). Likewise about 43 mammals and 268 birds have been documented so far in RSWR (DNPWC 2005). These protected areas are important habitats for charismatic megafauna such as Bengal tiger, Asian elephant, and re- introduced one-horned rhinoceros. Probably the largest herd of Barasingha Cervus duvauceli in the wild thrives in RSWR and it has the largest population of the endangered

Bengal Florican Houbaropsis bengalensis (Baral et al. 2003).

2.7 Settlements

The buffer zone of RBNP (328 sq km) was declared in 1996 in the west and south. It includes 17 village development committees and some 120,000 people live in 11,504 households within it. The buffer zone of RSWR has not been officially declared yet, but conservation and development activities have been carried out in the proposed buffer zones (Heinen & Rayamajhi 2001).

2.8 Ethnographies

The founding King of modern Nepal, Prithvi Narayan Shah, described ethnic heterogeneity as ‘char jat and chattis varna’ [garden for all types of people] which was later misconstrued in the Civil Code of 1854. The code classified people into three classes: higher castes, touchables and untouchables. Discrimination based on religion, race, sex, caste or ideology was theoretically abolished by the constitution of 1961 and

20 the new Civil Code also ended the practice of punishing offenders based on their caste

(Karki & Bhattarai 2004).

Nepal is a multiethnic, multilingual and a predominantly Hindu state. Although it is a non-secular state, the practice of other religions is common. About 81% of the population are Hindus followed by Buddhists (10.7%) and Muslims (4.2%; CBS 2001). The Kirat,

Jain, Christian, Sikh and Bahai religions are all also practiced in Nepal. There was no systematic record keeping of ethnic groups in the past national censuses. For the first time, an attempt was made to collect various data based on ethnicity in the 2001 national census. According to that census, there are 102 caste/ethnic groups speaking 92 dialects

(CBS 2001). There is a distinct pattern of geographical distribution of caste/ethnic groups in Nepal. People of Tibetan origin are in the high Himalayas, Hindu caste people

(Brahman, Chhetri and Occupational castes) are in middle hill regions and ethnic Tharu are in the lowlands.

Tharus are probably the oldest inhabitant of the Terai region. They are considered indigenous because it is believed that they have lived there for more than 600 years (Cox

1990). Tharus are an aboriginal tribe found in Nepal and India, but large numbers of

Tharus reside in Nepal (Verma 1998). Tharus are found in scattered settlements in the proximity of forests of the Terai from the Koshi River in the east to the Mahakali River in the west. There are several hypotheses regarding the origin of Tharus. Some authors trace their origin to Rajput ancestors who fled the battle described in the epic Mahabharat while others believe that they fled to northern frontiers in the time of the Islamic

21 invasions in India. Some Tharus claim that they are descendants of Rajput women who fled with domestic servants. Although the origin of Tharus is a controversial issue, present day scholars reject the claim that they are from Rajasthan. Tharus are a

Mongoloid tribe who have successfully assimilated non-Mongoloid physical features as well (Bista 1987).

There is no caste-hierarchy among Tharus; Tharus have 32 groups and no group is considered privileged (Bista 1987; Verma 1998). Rana Tharus are isolated among groups and live in Kailali and Kanchanpur Districts while other groups are scattered in the eastern and western lowlands. In the past, Tharu villages were found in enclaves of dense forests and wildlife (Bista 1987). This is not the typical case at present due to shrinking forest cover. They still prefer to live in the proximity of forested areas. They are peasant farmers (Bista 1987; Verma 1998) who keep different types of livestock and who practiced shifting agriculture in the past.

The solidarity among members of Tharu tribe is appreciable. In spite of the modern sociopolitical system, they have their traditional system of leadership known as

‘budghar’. This is one of the most democratic systems in a tribal society. Members of a tribe elect a village head – the budghar, and render services until he quits or villagers oust him for inefficiency. The budghar, in consultation with elders, settles social disputes and impose fines to defaulters. The fine is pooled in a village trust fund and used for collective causes. The trust, solidarity, traditional institutions and graduated sanctions all

22 have harnessed social capital (Krishna 2002). However, this capital is not yet tapped for socioeconomic development of Tharus in the modern era.

Most Tharus live in joint family structures, and in some cases, in extended familes. There is an explicit hierarchy, division of labor, and functional roles of members living in joint and extended families. Tharus are quickly emulating cultural practices of other groups and opting for nuclear families nowadays. Arranged marriages, working for a wife and elopement are common modes of acquiring mates (Bista 1987). Marriage is within the tribe, but not within one’s ‘gotra’ (clan). In India, marriages are solemnized according to the Hindu calendar (Jain 1991). The institution of marriage is sacred, patrilocal and monogamous. Tharu women used to have a superiority complex to their male counterparts, but this is now ending abruptly (Verma 1998). Among ethic groups, Tharus keep their houses exceptionally clean and are connoisseurs of folk art. They have a great deal of knowledge on color and pattern choices for painting, knitting and clay work.

Throughout their population, the literacy rate among Tharu is very low (Bista 1987; Jain

1991).

Tharus have their own tribal religion and worship a number of local spirits and personal deities (Bista 1987). They have also incorporated Hindu deities into their rituals. Tharus who have been following their traditional religion have their own priests called ‘guruva’.

Some Tharus call Hindu priest to perform weddings and other domestic ceremonies. In

India, Tharus demanded Hindu caste status based on similarities in ceremonies of marriage and shradh (a ritual performed to pacify the departed soul; Verma 1998). Tharus

23 are undergoing tremendous changes due to cultural invasion (Bista 1987; Verma 1998).

They are gradually reforming food habits, religious practices, cultural values, and adopting modern education. These cultural changes will likely have impacts on natural resource conservation and management.

3 MATERIALS AND METHODS

3.1 Selection of Villages

I selected 14 settlements of two village development committees of RBNP, and 15 settlements of six wards of RSWR. As per the Buffer Zone Management Regulation of

1996, the park authority formed User Groups (UGs) at the village (hamlet) level. An adult representative from each household from the village get together and select their representatives for the UG. I chose UGs as sampling units. The rationale for selecting these settlements are, (i) they lie within declared or proposed buffer zones, (ii) some integrated conservation and sustainable development programs have been implemented in these areas, (iii) they are politically more stable, and (iv) they are easily accessible.

3.2 Household Survey

3.2.1 Household survey team: The survey team consisted of myself and one local high school graduate belonging to the ethnic Tharu community. Whenever Tharu respondents preferred to answer in their mother tongue, the research assistant translated their answers into the Nepali language for me. Since I administered all household interviews, personal biases were not a problem.

24 3.2.2 Survey instrument: Heinen (1993) developed a survey protocol to study park people interactions in Koshi Tappu Wildlife Reserve, Nepal. This protocol was later modified and used by Sah (2002) and Shrivastava (2002) to study socioeconomic dynamics of protected areas of Nepal and India, respectively. I slightly modify the protocol to meet my objectives and to adjust to local sociopolitical conditions. As the protocol was reliable, I did not pre-test for validation.

3.2.3 Survey procedure: Prior to data collection, a research proposal and draft survey forms were reviewed and approved by the Institutional Review Board on Human

Subjects, Florida International University, Miami, USA. Permission to conduct surveys in the buffer zones of protected areas was obtained from the Department of National Parks and Wildlife Conservation, Nepal. In the field, the survey team visited each respondent in his/her house. The respondents were briefed about the purpose of the visit and verbal consent was taken to participate in interviews voluntarily. The questionnaire was written in Nepali [the official or national language of Nepal] and administered orally. The average time required to complete one interview was about 30-45 minutes.

The geographical position of each household was taken with GPS. Spatial reference and socioeconomic data were fed into the GIS system. Maps of the study area were derived from the input data.

25 3.3 Sample Size and Sampling

From the archive of the UGs, I stratified the sample households by ethnicity (caste

groups). A structured questionnaire survey was administered to a sample of 234

randomly selected households living in the buffer zones of RBNP and RSWR from

February to May of 2004. Taking into account the high illiteracy rate in rural Nepal, questionnaires were written in the Nepali language, but were asked in Nepali or Tharu, depending on the ethnicity of household being surveyed. Local words were used and technical jargon was avoided. One adult person (≥ 19 years old) in each household was interviewed in his/her residence. Usually, household heads (generally male) were interviewed; in their absence, any member willing to participate was interviewed resulting in more male (186) than female (48) respondents. Each questionnaire was divided into seven general parts: (1) ethno-religious background, household characteristics (gender, age and occupation of all household members), education

(illiterate, primary, secondary and college), and migration status; (2) economic activities such as land-holdings, alternative sources of income, annual cash income, (3) agriculture and animal husbandry; (4) natural resources use; (5) conservation awareness; (6)

participation and benefits (memberships, personal benefits, income generating activities

and saving-credit program), and (7) assessment of satisfaction towards UGs and wildlife

conservation issues. Most of the questions were closed-ended, although some open-ended

contingency questions were also included. A sample of survey protocol is appended

(Appendix 2).

26 3.4 Survey of User Groups (UGs)

The chair of 14 UGs of RBNP and 15 UGs of RSWR were also interviewed. They were asked about group formation, frequency of meetings, policies on non-timber forest products and their marketing, distribution of benefits, and attitudes towards the Terai Arc

Landscape project and conservation legislation. Whenever I had an opportunity, I also did content analysis of operational plans, five-year work plans, and annual reports of UGs.

3.5 Survey Limitations and Constraints

Due to the Maoist insurgency in the region rural people were reluctant whenever we visited them. Although we had explicitly explained our purpose and they were willing to be interviewed, they were frequently wary. Taking into account the sensitive political situation, we did not cross-examine responses on land holdings and household incomes.

In some cases, neighbors of respondents gathered out of curiosity, which is normal in rural areas, and gave suggestions to respondents being interviewed. Therefore, some issues reflected collective perception rather than personal opinion. Women were more hesitant than men to be interviewed. Among many ethno-religious groups, women have less power in decision-making. As per social customs, it is not feasible to interview females when males are present at home. This had resulted in the asymmetrical representation of gender. These limitations and constraints may influence interpretation of results.

27 3.6 Statistical Analyses

Before processing data, they were entered in Microsoft Excel. All questionnaires were scrutinized to detect errors and omissions. Accurate data that were consistent with other facts were included for coding and tabulation. Attribute data such as gender, caste, occupation, literacy were assigned numerals. To facilitate analysis, some quantitative data were converted into categorical data. From MS Excel data were imported to STATA

Version 8 and analyzed.

For conservation attitudes, a series of statements was presented and respondents were asked to agree or disagree. Statements covered broad conservation issues such as the status of forests, custodianship of resources, perceptions of open access resources, wildlife populations and depredation, socioeconomic improvements, access to resources, intra and intergenerational equity, existence of parks, and willingness to contribute to conservation were included. If the respondent agreed with the statement one point was given, otherwise no point was given. The reverse was true for a negative statement. The scores of all statements were summed to derive an attitude score that could theoretically range from 0 to 11. The higher the attitude score, the more favorable attitude the respondent had towards conservation.

There were eight types of resources harvested by respondents in RBNP and seven types in RSWR. At first, harvest frequency of each type of resource was calculated. Based on the harvest frequency, resources were assigned importance values. In case of RBNP, the weight of the eight to one score was applied in descending order; eight was assigned to

28 the resource having the highest frequency. Similarly, in RSWR the weight of the seven to one score was applied in descending order; seven was given to the resource having the highest frequency. The weighted scores of types of resources harvested in a household were summed to calculate a resource use score that could theoretically range from 0 to 36 in case of RBNP and 0 to 28 in case of RSWR. The higher the score, the more dependent were respondents on resources. Based on an equal interval, the resource use score was categorized into four categories: not dependent, somewhat dependent, dependent, and most dependent. Frequency-based 'importance' assignment may not truly reflect the impacts of resource use or scarcity of resources, I used it in my analysis for convenience.

Since my analytic framework is a comparative study between two protected areas, I used a parametric two-sample t test when quantitative data were normally distributed and the non-parametric Mann-Whitney test when they were not, to estimate the difference between certain variables of the two independent samples. Means of quantitative variables were presented with one standard deviation.

The Kruskal-Wallis One-Way Analysis of Variance by Ranks was used to test for differences in family size and landholdings among four groups of respondents (Tharus versus non-Tharus and RBNP versus RSWR). The chi-square test of independence was used for testing associations between qualitative variables. The null hypothesis is that the two variables are not related. To measure a relationship between resource use score and quantitative socioeconomic variables, Spearman’s rank correlation coefficient (Rho) was used.

29 To see the effects of two or more independent variables on a single dependent variable, multiple regression is an appropriate statistical method (see Allison 1998 for an introduction to this technique). I built a regression model taking the conservation attitude score as a dependent variable; and demographic, socioeconomic and benefit variables, and resource use score as independent variables. All categorical independent variables in the model were recorded as dummy variables, each with two categories: ‘yes’ and ‘no’.

4 RESULTS

4.1 Demographic Characteristics of Respondents

4.1.1 Gender and age: Of 234 respondents, 79% were male and 21% were female. There was no significant difference in the proportion of male and female respondents between

2 two protected areas (χ 1 = 1.40, p = 0.237). In RBNP, 82% of respondents were male and

18% were female while in RSWR 76% were male and 24% were female (Figure 5).

The range of respondents’ age was 19 to 75 years. The mean age of respondents of RBNP

(41.51 ± 12.31 years) was not significantly different from RSWR (41.67 ± 13.09 years; t

= -0.10, p = 0.925). The mean age of male (41.80 ± 12.16) and female (40.14 ± 13.18) respondents was not significantly different in RBNP (t = -0.58, p = 0.566). However, in

RSWR the mean age of male respondents (44.02 ± 13.24 years) was significantly higher than the mean age of female respondents (34.15 ± 9.43 years; t = 3.53, p = 0.0006).

30 4.1.2 Ethnicity: Although discrimination based on caste and ethnicity was abolished by

law, it is practiced socially. Brahman and Chhetri rank high and Occupational castes

(cobbler, ironsmith, tailors etc.) rank low in the Hindu caste hierarchy. Tharus are

indigenous people of Terai and Hill tribes include ethnic groups of mountain origin such

as Gurung, Magar and Newar.

There was a significant difference in ethnic composition of respondents between the two

2 areas (χ 4 = 48.85, p = 0.000). In RBNP, more than half (51%) were Tharu followed by

Chhetri (21%), Brahman (18%), Occupational castes (6%) and Hill tribes (3%). In

RSWR, the proportion of Chhetri (45%) was highest followed by Brahman (27%),

Occupational castes (17%), Tharu (10%) and Hill tribes (2%; Table 1).

4.1.3 Education: The education level of respondents was categorized into five groups.

Respondents who did not know how to read and write were classified as ‘illiterate’ and

those who could read or write but had no formal education were ‘literate’. Respondents

who had 1 to 5 years of formal education fall into ‘primary’, those who had 6 to 10 years

of formal education belonged to ‘secondary’. Those who had an associate degree or

above were classified as ‘college’.

There was a significant association between the level of education and protected areas

2 (χ 4 = 18.14, p = 0.001). About 38% were illiterate and 22% were literate in RBNP while

24% were illiterate and 17% were literate in RSWR. In RBNP, 19%, 18% and 2% had

primary, secondary, and college level education respectively while in RSWR, 15%, 39%

31 and 14% had primary, secondary and college level education respectively (Table 2). The

ethnic categories of Brahman, Chhetri, Occupational castes and Hill tribes were merged

into one category (non- Tharu) to compare differences in education level. Ethnicity was

2 significantly associated with level of education (χ 4 = 19.04, p = 0.001). The illiteracy rate among Tharus (36%) was significantly higher than non-Tharus (29%). Among

Tharus, 29% were literate, 21% had primary, 12% secondary, and 1% had a college level education. Non-Tharus were better off than Tharus in level of education. Among non-

Tharus, 16% were literate, 15% had primary, 30% secondary, and 11% had a college level education. The literacy rate among women (40%) was significantly lower than men

2 (76%; χ 4 = 25.20, p = 0.000).

4.1.4 Occupation: ‘Agriculture’ is the primary vocation of most respondents. Paddy,

lentil and wheat are staple crops. In addition to subsistence agriculture, people are also

engaged in other vocations. Service oriented activities in public, private and military

sectors as well as teaching were included in one category ‘job’. Seasonal wage labor and

other non-farm income generating activities were categorized as ‘menial work’. The

respondents who were involved in trade and business at the local level were categorized

as ‘business’. ‘Students’ were those who were currently enrolled in formal education.

About 77% of respondents were subsistence farmers while 23% were engaged in some

off-farm activities. There was a significant difference in the proportion of respondents

2 practicing subsistence agriculture and non-farm activities between two areas (χ 1 = 18.55, p = 0.000). In RBNP, 88% and 12% respondents reported agriculture and off-farm

32 activities respectively as their primary vocation. In RSWR, 64% and 36% respondents reported agriculture and off-farm activities respectively as their primary vocation. The off-farm occupations in RBNP were job (6%), menial work (6%) and business (2%), and in RSWR they were job (17%), menial work (5%), business (7%) and student (6%;

2 Figure 6). The ethnic category was also significantly associated with occupation (χ 1 =

9.58, p = 0.002). Most Tharus (89%) were subsistence farmers while 11% were engaged in off-farm activities. Among non-Tharus, 71% were subsistence farmers while 29% were involved in non-farm activities.

4.1.5 Family size: The average family size of RBNP was not significantly different from that of RSWR (z = -1.03, p = 0.305). The mean family size of RBNP was 7.63 ± 4.60 and varied greatly from two to 32 members. In RSWR, the mean family size was 7.64 ± 3.44 and varied from one to 23 members.

The Kruskal-Wallis one-way analysis of variance by rank showed that mean family size

2 of Tharus and non-Tharus was significantly different in the two areas (χ 3 = 17.46, p =

0.001). The Bonferroni procedure of multiple comparisons showed that the family size of

Tharus of RBNP (9.23 ± 5.75) was larger than that of non-Tharus of RBNP (5.95 ± 1.85; z = -3.87, p = 0.000; Table 3). There was no significant difference in family size of other pairs (p > 0.10). The Kruskal-Wallis test also showed that there was no significant

2 association between family size and level of education (χ 4 = 2.61, p = 0.625).

33 4.2 Migration and Economic Status

Tharus are the only indigenous tribe in the study area. All other ethnic groups or castes migrated to the Terai, mostly from the Mountain. There is some internal migration of

Tharus within the Terai.

There was a significant difference in the proportion of immigrants and residents between

2 the two areas (χ 1 = 28.25, p = 0.000). Almost all (96%) respondents had migrated while only 4% were resident people in RSWR. In RBNP, 70% of respondents had migrated while 30% were resident (Table 4). Migration into these places is vertical (hills to plain) as well as horizontal (east - west). More than half (55%) and about two thirds (74%) of respondents of RBNP and RSWR, respectively, had migrated to the study area from mountain districts. Likewise, 25%, 16% and 3% of respondents migrated to RBNP from other Terai districts, villages within the district and other places, respectively. In RSWR,

21% migrated from villages within the district and 5% from other places. The mean years of duration of residency of migrants was not significantly different between RBNP and

RSWR (z = -0.60, p = 0.549). It was 23.37 ± 11.37 years for RBNP and 23.85 ± 10.50 years for RSWR. People started immigrating some 46 years ago in RBNP and 44 years ago in RSWR and the trend is continuing. The reasons given by respondents (N = 183) for migration were insufficiency of fertile land (39%), landlessness (8%), government programs (12%), unemployment (3%), and social factors (37%) such as to be close to relatives, more physical and infrastructural facilities and access to resources.

34 There was no significant difference in mean landholdings between the two areas (z = -

0.80, p = 0.422). The mean landholding of RBNP was 0.687 ± 0.67 ha and ranged from

0.03 ha to 3.15 ha. Likewise, the mean landholdings of RSWR was 0.714 ± 0.67 ha and

ranged from 0.03 ha to 4.74 ha. The Kruskal-Wallis one-way analysis of variance by rank

2 among four groups showed that there was no significant difference in landholdings (χ 3 =

2.50, p = 0.271). Although Tharus of RSWR had slightly more land than the other three groups, this was not statistically significant (p >0.05; Table 5).

Landholding was classified into three groups: small, medium and large, following government standard criteria. Small holders have less than 0.5 ha of land, medium holders have 0.5 ha to 2 ha of land and large holders have more than 2 ha of land (CBS

2 1992). There was no significant association of land categories with area (χ 2 = 3.70, p =

0.157). In RBNP, 53%, 40% and 7% belonged to small, medium and large landholders,

respectively. Similarly, 45%, 51% and 4% belonged to small, medium and large

landholders, respectively, in RSWR.

One encouraging finding was that 87% of respondents had land tenure, and only 13% did

not, but were using lands without title. There was no significant difference in the

2 proportion of respondents with unregistered lands between the two areas (χ 1 = 0.11, p =

0.743). Fourteen percent of RBNP and 12% of RSWR respondents did not have tenured lands.

35 2 There was a significant association of agricultural yield with area (χ 1 = 27.52, p =

0.000). About 73% of respondents of RSWR mentioned that they had enough produce from their farms to sustain, but 61% of respondents of RBNP said that they did not meet their annual ration of staple food from their farms (Table 6). Of those who did not have enough yield from their agricultural lands, on average they had to buy staple foods for

6.21 ± 3.05 months per year (N = 105).

I asked respondents how much they earn in a year from non-farm activities. Of 234 respondents, 86% responded to the question. The annual cash income is highly skewed with a range of US $ 281 to US$ 2366 and an average of 544 ± 476 US dollars. There was a significant difference in annual cash income of respondents in the two areas (z = -2.94, p = 0.0032). The average annual cash income of RSWR was US$ 664 ± 544 (range US$

28 - 2366), which was larger than the average income of US$ 444 ± 386 (range US$ 28 -

1894) of RBNP respondents.

Cattle, buffalo, goat/sheep, pig and poultry were common types of livestock in these areas. The overwhelming number of respondents (96.2%) had one or more kinds of livestock. As the values of different livestock and their impacts on natural resources vary, the number of livestock per household was expressed using the Livestock Size Unit

(LSU). Since a 400 kg steer is equivalent to 1 LSU (Raut 1997), in the present study, 1 adult buffalo (1 LSU) was considered equivalent to 1 steer, and one immature buffalo, cow, calf, pig, and sheep or goat was equivalent to 0.5, 0.8, 0.4, 0.3 and 0.2 steer,

1 1 US $ = NRs 71 in 2004

36 respectively. Poultry were not included in LSU calculation. There was no significant

difference in LSU between the two areas (z = -0.44, p = 0.656; Table 7). In RBNP, the

average LSU was 4.37 ± 4.76, which ranged from 0.2 to 44. In RSWR, the average LSU

was 3.97 ± 2.11 and ranged from 0.4 to 11. In RBNP, 54.4% rear water buffalo, 65.6%

cattle, 57.6% sheep and goat, and 41.6% pig (Figure 9). In RSWR, 68.8% rear water

buffalo, 91.7% cattle, 25.7% sheep and goat, and 2.7% pig. People of RBNP tend to keep

larger numbers of buffalo (z = 3.13, p = 0.0017) and sheep/goat (z = 2.66, p = 0.0078)

than people of RSWR. There was no significant difference in numbers of cattle (z = 1.03,

p = 0.301) and pig (z = 1.26, p = 0.2066) between these areas. Kruskal-Wallis one way

analysis of variance by rank showed that there was no significant difference in mean LSU

2 among the four groups (Tharu versus non-Tharu and RBNP versus RSWR; χ 3 = 2.51, p

= 0.474).

4.3 Resource Use Patterns and Dependency

Firewood is the main source of energy in the study area. Local people use thatch as roofing material and grasses and fodder for livestock feed. Honey, mushrooms, fruits, vegetables are supplementary to staple diet. Local people ferment home brewed liquor with herbs and use timber for house construction and furniture. Green leaves are used to make leaf plates and in religious ceremonies. Dry leaves are harvested to use as bedding material for livestock and which is later on composted. The park authority provides permit to collect thatch once a year inside the parks and reserves. However, in buffer zones local people are allowed to collect any sort of resources as stipulated in the operational plan. Eight types of natural resources were extracted from the park and buffer

37 zone forests of RBNP while seven types were extracted from the reserve in RSWR. Local people in RSWR did not collect timber from the reserve, while local people in RBNP collected timber from the buffer zone. In RBNP, 93% households collected thatch followed by firewood (68%), leaf litter (62%), grasses (52%), edible plants (42%), timber

(41%), tree fodder (20%) and herbs (15%). In RSWR, 78% of household collected thatch followed by firewood (58%), grasses (44%), leaf litter (34%), edible plants (10%), tree fodder (2%) and herbs (1%). Significantly, more households in RBNP extracted thatch, leaf litter, edible plants, tree fodder and herbs in comparison to RSWR (p < 0.05; Table

8). Various types of resource uses have different impacts on natural resource conservation and management.

The nonparametric correlation of continuous variables with the resource use score showed that there was a significant negative correlation between the resource use score and annual cash income (p = 0.0001), and a positive correlation between the resource use score and recency of arrival (p = 0.000). Family size and LSU were positively correlated with the resource use score at a 10% error level (Table 9). The correlation between landholdings and resource use score was negative, but not significant.

The resource use score was categorized into four: not dependent, somewhat dependent, dependent and most dependent, based on equal intervals. There was a significant

2 association of resource dependency of local people between the two protected areas (χ 3 =

35.62, p = 0.000; Table 10). In RBNP, 23% belonged to most dependent, 50% to

dependent, 21% to somewhat dependent, and only 6% to not dependent. In RSWR, 8%

38 belonged to most dependent, 28% to dependent, 39% to somewhat dependent, and 24%

to not dependent. The dependency on resources was also significantly associated with

2 ethnicity (χ 3 = 43.24, p = 0.000). Among Tharus, 57% and 29% of respondents belonged to dependent and most dependent categories, respectively while 32% and 10% of non-

Tharus belonged to those groups, respectively.

2 There was a significant association between resource use patterns and ethnicity (χ 1 =

77.93, p = 0.000). Most Tharus (95%) collected green leaves for religious and social

functions, but a few (5%) collected dry leaf litter from forests to use as bedding material

for livestock. Conversely, non-Tharus most often (88%) collected dry leaf litter to use as

bedding material which is subsequently used in agriculture to enrich soil fertility, while

only 12% of respondents collected green leaves for making leaf plates.

2 There was a significant association between types of leaves collected and area (χ 1 =

47.13, p = 0.000). In RBNP, 74% of household collected green leaves from the park and buffer zone forests while 95% of households collected dry leaf litter from the reserve in

RSWR. People of RBNP collected more firewood than people of RSWR (z = 4.57, p =

0.0000). The reported average amount of firewood collected per household in RBNP (951

± 1143 kg per year) was more than double the amount collected per household (332 ±

366 kg per year) in RSWR. Local people also collected more thatch from RBNP than

from RSWR (z = 3.79, p = 0.0001). About 41% of respondents of RBNP harvested

timber from buffer zone forests but none of the respondents from RSWR reported using

39 timber from the reserve. The average quantity of timber harvested per household from

buffer zone forests was 9.90 ± 10.95 cubic feet (N = 51).

One of the most important aspects of natural resource use is the methods of gathering

feed for livestock. There was a significant difference in the practice of livestock grazing

2 between the two areas (χ 2 = 23.42, p = 0.000). About 64% of households stall fed their livestock in RBNP while 85% of households stall fed their livestock in RSWR. In RBNP,

28% and 8% of households grazed livestock in community pastures and forests, respectively. In RSWR, 4% and 11% households grazed livestock in community pastures and forests respectively (Table 11).

2 There was a significant association of source of livestock feed between two areas (χ 1 =

5.52, p = 0.019). In comparison to RBNP (73%), more households (86%) in RSWR met their fodder requirement from private lands. More households (27%) in RBNP than in

RSWR (14%) collected fodder from the park and buffer zone forests (Table 12).

In both areas, the proportion of respondents mentioning the problem of getting adequate

2 resources such as firewood was not significantly different (χ 1 = 2.38, p = 0.123). In

RBNP, 68% of households mentioned the problem of firewood while in RSWR 72% of

households mentioned the problem. This is because there was no significant difference in

2 the sources of household energy between two areas (χ 2 = 0.02, p = 0.991). Significantly,

a high proportion of households (90%) in both areas still rely on traditional inefficient

mud stoves to cook. Smaller proportions of households either used improved mud stoves

40 (4%) or relied on alternative energy sources (6%) such as kerosene, liquefied petroleum

gas and biogas (Table 13).

There was a significant association of suggestions regarding the firewood problem with

2 area (χ 4 = 35.98, p = 0.000). Only 20% of respondents of RBNP suggested issuing permits for firewood collection from inside the park while more than half of respondents

(55%) of RSWR were of opinion that the reserve should be open for firewood collection with permit. Higher proportions of respondents (34%) of RBNP were suggesting the introduction of alternative energy sources (especially biogas) than of RSWR (21%). More respondents of RSWR (15%) were in favor of private plantations than RBNP (7%; Table

14).

2 There was a significant association of local trade of natural resources with area (χ 1 =

10.74, p = 0.001). In RBNP, 40% of households mentioned that they have engaged in buying or selling resources, while in RSWR only 20% of households mentioned that they were involved in such activities. About 80% and 60% of households in RSWR and

RBNP, respectively were not engaged in local trade of natural resources within their villages.

There was significant association of interest in non-timber forest product (NTFP) farming

2 with area (χ 1 = 8.34, p = 0.004). In RBNP, only 34% households showed interest in

NTFP farming while 66% did not show any interest. In RSWR, 52% households showed

interest while 48% did not show any interest.

41 4.4 Participation in Conservation

People’s participation is a widely used phrase in the conservation lexicon. After the promulgation of a participatory approach, most institutions dealing with resource management and conservation have devised mechanisms for public participation. In theory, local people are empowered and local institutions are harnessed for sustainable management of natural resources through active participation. With the promulgation of community-based conservation legislations, grassroots institutions have emerged in the conservation areas and buffer zones. Local people are general member of these institutions and among them a few are executive members. NGO sponsored trainings were provided to local people to raise their standard of living and to executive members to strengthen the grassroots institutions. NGOs have also initiated income-generating activities such as goat raising, NTFP farming, piggery, sewing-cutting for local communities. Local people got personal benefits such as biogas plant, drinking water, toilet, piglets etc. from NGOs implementing integrated conservation and development programs. In Saving-Credit program local people form a group and deposit a fixed amount of money per week or month and lend that money to a member of the group to undertake income-generating activities. The creditor pays a nominal interest per month.

This is more like a rural bank. All above programs are taken as conservation interventions by NGOs. I used these programs to measure popular participation of local people in conservation.

Significantly more households of RBNP (29%) were members of one of the grassroots

2 institutions than households (13%) of RSWR (χ 1 = 8.82, p = 0.003). More households of

42 2 RBNP (35%) received some sort of training than those of RSWR (19%; χ 1 = 7.36, p =

0.007). More households got direct benefits from conservation interventions in RBNP

2 (36%) than in RSWR (10%), which was also statistically significant (χ 1 = 21.47, p =

0.000). At 10% error level, the saving-credit program of RSWR was more pronounced

2 than that of RBNP (χ 1 = 3.75, p = 0.053). About 66% of households were participating in the program in RSWR while only 54% households of RBNP were in the program.

Households of RSWR were depositing more money in the saving-credit program than those of RBNP (z = 3.00, p = 0.0027). Households of RSWR deposited on average US$

0.45 ± 0.32 (range 0.07 – 2.82) per month in saving-credit while households of RBNP deposited on average US$ 0.34 ± 0.42 per month (range 0.07 – 1.41). Significantly, a higher proportion of households (18%) of RBNP were involved in income generating

2 activities in comparison to households (5%) of RSWR (χ 1 = 9.66, p = 0.002; Table 15).

4.5 Conservation Attitudes

Five attitude statements were significantly different while six were not significantly different between the two protected areas (Table 16). The statements that differed between the two areas were: perception on forest status, custodianships of resources, trends in wildlife populations, socioeconomic upliftment, and resource use conflicts. The statements that did not differ were: problems with open access, anthropocentric views, existence of protected areas, intergenerational equity, intragenerational equity, and willingness to contribute for conservation causes.

43 A significantly greater proportion of respondents (70%) of RSWR agreed that the forests

in their surroundings were dwindling, but 43% respondents of RBNP disagreed with the

2 statement (χ 1 = 4.37, p = 0.037). This statement had a significant association with

2 migration status (χ 1 = 4.95, p = 0.026). A higher proportion (66%) of immigrants agreed with the statement, while 52% of resident disagreed. There was a significant association

2 of the perception of forest status with gender (χ 1 = 4.89, p = 0.027). A higher proportion

(77%) of female respondents agreed that forests had dwindled in comparison to male

respondents (59%). There was no significant difference between Tharus and non-Tharus

2 (χ 1 = 2.10, p = 0.148).

A significantly higher proportion of respondents (96%) of RBNP than of RSWR (89%) agreed with the statement that local people share responsibility of conserving natural

2 resources (χ 1 = 3.96, p = 0.047). There was no significant association of this statement

2 2 with gender (χ 1 = 0.91, p = 0.340), ethnicity (χ 1 = 0.11, p = 0.743), and landholdings

2 (χ 2 = 0.52, p = 0.772). Although people were harassed by wildlife, they were willing to

2 share responsibilities for conservation (χ 1 = 0.52, p = 0.473). When local people were satisfied with achievements of User Groups they were more likely to agree with sharing

2 responsibilities for conservation (χ 1 = 5.16, p = 0.023).

There was a significant difference in the perception of wildlife population trends among

2 local people between the two protected areas (χ 1 = 58.58, p = 0.000). An overwhelming proportion (91%) of respondents agreed that populations of wildlife in RBNP had increased while more than half (66%) of respondents disagreed that wildlife populations

44 in RSWR had increased. Respondents who had suffered from wildlife damage were more

2 likely to agree that there was an increase in wildlife populations (χ 1 = 16.35, p = 0.000).

Respondents whose main vocation was agriculture were more likely to agree with the

2 statement (χ 1 = 6.26, p = 0.012). An overwhelming proportion (90%) of Tharu agreed

that wildlife populations increased, while only 59% of non-Tharus agreed with the statement.

There was a significant association between the perception on socioeconomic

2 development and area (χ 1 = 14.53, p = 0.000). More respondents (70%) of RBNP agreed that there was improvement in living standard after the establishment of park than RSWR

(45%). There was no significant difference between Tharus and non-Tharus on attitude

2 towards socioeconomic development (χ 1 = 1.57, p = 0.210). The likelihood of men and

2 women agreeing with this statement was the same (χ 1 = 0.01, p = 0.974). Respondents

who were engaged in off farm activities did not associate these opportunities with

2 conservation interventions (χ 1 = 0.56, p = 0.453). Respondents who had not lost livestock and crops to wildlife were more likely to agree that there were improvements in

2 the socioeconomic status of local people (χ 1 = 4.69, p = 0.030).

2 There was a significant association of resource use conflicts with area (χ 1 = 12.74, p =

0.000). In RBNP, 81% of respondents agreed that they did not have problems with resource use after the establishment of the park and its buffer zones. In RSWR, only 59% agreed that they did not have problem of resource use after the establishment of the reserve. Respondents belonging to the dependent and most dependent categories of

45 resource use were more likely to agree with the statement that they did not have a

2 problem of resource access (χ 3 = 13.58, p = 0.004). Perceptions about resource access

2 2 were not related to gender (χ 1 = 0.15, p = 0.703), ethnicity (χ 1 = 0.85, p = 0.358),

2 2 landholdings (χ 2 = 2.96, p = 0.227), and occupation (χ 1 = 0.08, p = 0.778).

The demographic and socioeconomic variables included in the regression model

explained about 25% of the variation in conservation attitudes (p < 0.01). Multiple

regression results revealed that RBNP respondents were more likely to hold favorable attitudes than RSWR respondents (p = 0.049). Similarly, respondents who held more favorable attitudes were likely to be those who were satisfied with UGs activities (p =

0.000), those who participated in the NGO sponsored trainings (p = 0.011), and those

who were not harassed by wildlife (p = 0.024). Other variables did not contribute to

significant variation in conservation attitudes. The coding of each independent variable

is indicated in parenthesis, with the sign of the coefficients showing whether associations

are positive or negative (Table 17).

4.6 Wildlife Harassment and Attitudes towards UGs

2 There was a significant association between wildlife harassment and protected areas (χ 1

= 7.99, p = 0.005). Significantly, a higher proportion (77%) of households of RBNP suffered from wildlife damage than households of RSWR (60%). Twenty three percent of households of RBNP and 40% of households of RSWR did not mention wildlife damage.

There was a significant difference between the two protected areas about the expression

2 of satisfaction towards UGs among local people (χ 1 = 4.44, p = 0.035). In RBNP, 82% of

46 respondents were satisfied with the achievements of their UGs while in RSWR 71% were satisfied. Thus, a higher proportion (29%) of respondents of RSWR were not satisfied with UGs achievements than RBNP respondents (18%).

4.7 Institutional Development and Strengthening

With the passage of the Conservation Area Management Regulation (CAMR) in 1991 and the Buffer Zone Management Regulation (BZMR) in 1996, it became mandatory to solicit local people’s participation in nature conservation and protected areas management. The CAMR and the BZMR vested power to declare conservation areas and to delineate buffer zones in the periphery of national parks and wildlife reserves, respectively. The BZMR stipulated that User Groups (UGs) should be formed at village/settlement level. These groups formed the foundation for a bottom-up approach to resource management within buffer zones and were envisioned as grassroots institutions responsible for conservation and sustainable development of their respective buffer zone units.

4.7.1 Institutional working procedures: Of 29 UGs, 79% were formed unanimously by villagers while 21% were formed by election among local leaders. There was no significant association between mode of user group formation and protected areas (p =

0.390). In RBNP, 71% of UGs were formed by consensus while 29% were formed by election. In RSWR, 87% of UGs were formed by consensus while 13% were formed by election (Table 18). Any adult person above 18 years is eligible for membership and women members are mandatory in UGs.

47 All UGs of RBNP had female members on their executive committees while only three

UGs of RSWR did. The number of female members varied from one to eight.

Disadvantageous Groups (DAGs) were represented in six UGs of RBNP, but only in two

UGs of RSWR. The members from DAGs on the executive committee varied from one to

four. It was surprising that none of the UGs of RSWR had Tharus on the executive

committee but all UGs of RBNP had Tharus members (one to nine) in the committee. In

RBNP, one UG was meeting fortnightly, 12 monthly and one quarterly. In RSWR, three

UGs were meeting weekly, seven fortnightly, three monthly, and two were not meeting at

all.

The Maoist People’s War had less impact on UGs activities in RBNP in comparison to

RSWR (z = -2.50, p = 0.0125). In RBNP, 86% of UGs were holding their regular meetings while only 47% of UGs in RSWR were holding their regular meetings at the time of survey. The reason given by respondents was that due to growing political tension; it was not feasible to call meetings. In many cases, the state’s security forces instructed UGs not to hold meetings as a precaution to prevent Maoist infiltration and extortion. On some occasions, Maoist rebels threatened UG chairs not to call meetings because they thought that UGs were supporting the state. The longest abeyance of meetings (18 months) had occurred in RSWR, but the abeyance was for three months only in RBNP. At the time of the field visit, the average abeyance of meetings for RSWR was 10.62 ± 7.46 months (N = 8) and for RBNP it was 2.5 ± 0.71 months (N = 2).

48 The BZMR requires each UG to prepare a five year work plan and submit it to the Buffer

Zone Management Council for approval and allocation of budget. The Fisher’s exact test

showed that there was a significant association of work plan formulation with protected

areas (p = 0.017). All UGs of RBNP had five year work plans, but only 60% UGs of

RSWR had. There also were more NGOs working in the buffer zones of RBNP than

RSWR. The King Mahendra Trust for Nature Conservation (KMTNC), the Participatory

Conservation Program (PCP), the Terai Arc Landscape (TAL), Center for Aid in Relief

and Emergency (CARE) - Nepal and the Kisan Jagaran (Awareness among Farmers)

were working with local people for conservation and sustainable development in the

buffer zone of RBNP. During the time of fieldwork, only the PCP had significant

presence and the KMTNC had just established its office in RSWR. KMTNC and PCP

facilitated the drafting of five year work plans in RBNP and RSWR, respectively.

4.7.2 Resource management and demand fulfillment: In RBNP, of 14 UGs, 10 had their

own buffer zone forests, but none of RSWR UGs had buffer zone forests. The size of

buffer zone forests of RBNP varied from 10 ha to 150 ha. At the time of the study, no UG

had facilitated the sale of non-timber forest products (NTFP) from their buffer zone forests. Seven UGs of RBNP were contemplating the promotion of NTFP marketing and were looking forward to devise strategic plans to this end.

When asked which institutional arrangement would be most efficient to manage resources in a sustainable manner, a high proportion (64% in RBNP and 60% in RSWR) of UG chairs responded that local people were the most effective institute. About 21%

49 and 33% of UG chairs of RBNP and RSWR, respectively, considered government

agencies as the most effective institution. Some UG chairs (14% in RBNP and 7% in

RSWR) emphasized coordinated efforts of government agencies and local people for

sustainable management of resources in the region (Table 19).

There was no significant difference in UGs’ competence in fulfilling legitimate demands

2 of members in the two areas (χ 1 = 0.29, p = 0.588). In RBNP, 50% of UGs were meeting their subsistence needs of natural resources from their buffer zone forests. Since there was no buffer zone forests in RSWR, funds collected through saving-credit program were the only common property. More than half (60%) of UGs were not able to fulfill loan demands of their members while 40% fulfilled demands (Table 20).

4.7.3 Attitudes of conservation leaders: To implement landscape level conservation in the

Western Terai, the TAL project – a joint undertaking of the Government of Nepal and

WWF-Nepal Program, has been working in these areas since 2001. I asked UG chairs to

express their opinions about the project. In the buffer zones, TAL was more pronounced

in RBNP than in RSWR (p = 0.001). All UG chairs of RBNP were familiar with TAL

while only 60% of UG chairs of RSWR were familiar. I asked the respondents to rank the

satisfaction level of overall TAL activities in the area. In RBNP, 21% were highly

satisfied, 50% moderately satisfied and 29% were not satisfied at all. An overwhelming

proportion (93%) of respondents were not satisfied with TAL in RSWR (Table 21).

50 There was a significant difference in the respondent’s understanding of the Regulations

and Guidelines between the two areas (p = 0.001). About 64% of respondents disagreed

with the present Regulations and Guidelines in RBNP, but 80% of respondents in RSWR

said that there was no problem with these. There were some (20%) respondents who

confessed that they had never studied the legal protocols so were unable to comment

(Table 22).

4.7.4 Content analysis of the work plans: Operational plans were drafted by staff of

NGOs in consultation with UGs. This was reflected in the contents and types of programs implemented. One of the progressive aspects of operational plans was the provision for compensation. UGs can levy nominal tax on resources, impose fine to defaulters, and accept donation from outside sources. Depositing 10-15% income of the UGs into compensation fund is mandatory. The compensation fund is used to reimburse losses due to wildlife depredation, loss of life due to wildlife attack and property damage by natural calamities. Ecological risks such as pesticides and chemical fertilizers were mentioned, but no concrete programs were devised to tackle these problems. Forest management practices mentioned in operation plans were not intensive and based not on current science. The agricultural and livestock development sector got the least priority. The principal focus in this sector was goat raising. Human resource development and income generating programs were traditional and may not serve towards empowerment.

Trainings mentioned in the plans such as sewing-cutting and knitting won’t meet the objectives of income generation because most previous participants were not using the training for income-generating purposes. Another drawback of the plans was not

51 mentioning the percentage of funds allocated for conservation versus development from

the revenue earmarked to and income earned by UGs.

5 DISCUSSION

There were more male (186) than female (48) respondents. This is due to the social status

and level of education. The literacy rate among females is lower in comparison to males

2 (χ 4 = 25.20, p = 0.000). Most female respondents (60%) were illiterate while only 24% of the male respondents were. Socially, women have subordinate roles and less power in decision making so males are usually household heads. Since I targeted household heads for interview and women were reluctant to take part while men were present, there is an asymmetrical representation of gender in the survey. This may have implications on interpretation and generalization of results because women are more involved in forest resource extraction such as firewood, fodder and edibles (Mehta & Kellert 1998).

The literacy rate (68.8%) of the study area was significantly higher than the national average (53.7%; CBS 2002). Significantly, more respondents of RSWR (76.1%) are literate than those of RBNP (62.4%; p = 0.001). This is consistent with the fact that

RSWR is located near an urban center and is close to India, so local people have easy access to educational institutes. Additionally, Tharus constituted about half of the sample in RBNP and the illiteracy rate (36%) among Tharus is higher than non-Tharus (29%).

This may have decreased the overall literacy rate of RBNP. The higher literacy rate in the

buffer zones of RBNP and RSWR compared to their respective districts is attributed to

52 extension programs of various NGOs. As part of awareness campaigns, many NGOs have undertaken adult literacy programs in these areas.

Tharus are the only ethnic tribe of the Terai. The downward spiral of Tharu began with the migration of other people from the mountains that started after the eradication of malaria in the early 1960s. Ethnic heterogeneity caused by migration tends to dilute community solidarity (Ostrom 1990) and may cause inter-ethnic conflicts in resources use (Noss 1997). The plurality of communities should be aptly addressed in management strategies for sustainable management of natural resources. From conservation perspectives, the problems of immigration are: inflated population, higher discount rates, lowered commitments to conservation, and increased pressures on natural resources

(Ostrom 1990; Gadgil et al. 1993; Kremen et al. 1994, Oates 1995). This may ultimately lead to the ‘tragedy of the commons’ (Hardin 1968). Although non-indigenous people are also capable of developing a knowledge base of local environment as they adapt

(Browder 1995, Muchagata & Brown 2000), they significantly alter the structure and composition of resources as they harvest (Nepstad et al. 1992).

The annual population growth rate of Nepal was 2.3% for the period of 1991 to 2001

(CBS 2002). The annual population growth rate of Bardia (3.2%) and Kanchanpur (4.6%)

Districts from which samples were drawn were much higher than the national average.

The mean family size (7.64 ± 4.09) of the study area was also significantly higher than the national average (5.44; t = 8.21, p = 0.000). One of the reasons for the higher growth rate is larger family sizes among Tharus. The average family size of Tharus (8.91 ± 5.42)

53 was significantly higher than non-Tharus (7.04 ± 3.13; z = 2.63, p = 0.0087). Tharus used

to live in extended families so responsibilities of bringing up children were shared, which

may have served as an incentive to have more children, among other factors in traditional societies. The concept of the nuclear family is gaining popularity in the younger generation Tharus. The other reason attributed for rapid growth is migration of people from the Mountains to these areas. In the past, people migrated to these areas to reclaim fertile agricultural lands, to access physical facilities, and to take refuge from environmental hardships. At present, the Maoist insurgency has stoked the fuel of migration. Although the Maoist insurgency was not mentioned by respondents as a cause to migrate, there were camps established in community forests and public lands to provide shelter to people displaced from the Mountains by the armed conflict. The higher annual population growth around forest reserves due to civil strife is also documented elsewhere (Archalbad & Noughton-Treves 2001) which is a main cause of failure of integrated conservation and development programs in such areas (Oates 1995).

The main challenge for resource managers in these areas is to curb forest encroachment.

Freed bonded-laborers, political refugees and opportunists are clearing forests for settlements. The government of Nepal abolished the bonded-labor practice in 2001, but did not bring concrete programs for rehabilitation. Most bonded-laborers are Tharus and they have no alternative livelihood but subsistence agriculture. In a desperate attempt to procure agriculture lands, freed bonded-laborers are reclaiming government forests, community forests, and public lands. Similarly, people displaced by the armed conflict are resettled in forests and the government’s acquiesce to deforestation caused by

54 humanitarian needs did not receive much opposition from civil society. Taking advantage of the lax security situation, many opportunists are clearing forests. There are no studies addressing the encroachment issue so it is premature to conclude the extent and scope of current threats. Without a long lasting peace, reformative actions, and commitments to conservation by the government, these problems seem insurmountable. Most respondents already have tenure rights and this may serve as an incentive for sustainable resource management. One of the components of community natural resource management is to secure land tenure rights (Kellert et al. 2000), and this should be used as a benchmark to curb encroachment of forests which is the pressing challenge at present.

Although respondents have the same average size of land parcels between areas, a significantly high proportion (73%) of RSWR respondents said that they get enough produce from their parcels. This is because there is an irrigation facility in RSWR, which is lacking in RBNP. The size of a land parcel may also influence the livestock size unit.

In the past, Tharus had large parcels of land and used to keep large herds of livestock.

They grazed livestock in forests. Large number of buffaloes in RBNP is due the fact that

Tharus still use these animals as draught power while immigrants use oxen. Culturally they had never practiced stall feeding, but are now gradually adopting the practice because of imposition of conservation laws and influence of other groups. At present, there is no significant difference in the size of land parcels and numbers of livestock between Tharus and non-Tharus (p > 0.10). Most Tharus are still peasant farmers and reduction in the size of land parcels and livestock may have negative impacts on their subsistence economy.

55 Resource use patterns between ethnic Tharus and immigrant non-Tharus are different.

Tharus never collected leaf litter from the forest to use as bedding material for livestock.

However, non-Tharu immigrants used to collect leaf litter in their previous dwellings

because the fertility of soil is very poor in the mountains, and they used to compost leaf

litter to enrich agricultural fields. They are thus still practicing local knowledge acquired

in mountains. Collection of leaf litter from forests may seriously deplete nutrients in the

soil. Sal forests are more vulnerable to nutrient depletion by this practice because they are

comparatively nutrient poor due to frequent fires that expel Nitrogen and dryness that

slows down decay of biological materials. Resource use patterns are a function of

economic status and cultural practices. Tharus earn significantly less annual cash income

(US $ 397 ± 388) than non-Tharus (US $ 619 ± 500; z = -3.15, p = 0.0004) from outside

sources. Tharus most often collect green leaves to make leaf plates for their daily use

while economically better off immigrants do not use leaf plates. Within the Tharu

community, liquor is a must in rituals and social ceremonies so they collect the Dadari2 herb Elephantos scaber to ferment food materials and subsequently distill home brewed liquor. The amount of green leaves collected is insignificant in comparison to dry leaf litter, so the practice of Tharus is superior to immigrants from soil nutrient perspectives.

These differences in resource use patterns have implications on the conservation and management of resources and should be taken into account while formulating management strategies.

2 Root extract of this herb is mixed with rice and is used to make yeast for brewing liquor; medicinally applied for syphilis, rheumatic pain and foot and mouth disease of livestock

56 The dependency of local people on natural resources depends on family size,

socioeconomic status, ethnicity, availability and duration of residency. Tharus are more

dependent on resources than non-Tharus. This result supports the finding of Sah and

Heinen (2001). Economically Tharus are poorer than non-Tharus, and this is in contradiction to the theory that wealthier people would suffer most if restriction on forest resources is imposed (Hegde & Enters 2000). Tharus are the indigenous people of the

Terai, and they have good knowledge about the uses of local resources. Although immigrants quickly learn local knowledge (Browder 1995), their competence in resource

use compared to indigenous people is largely unknown. Additionally, Tharus have larger

family sizes so they collect more resources to meet subsistence needs. When resources

are easily available, people collect more resources and fall into the category of most

dependent. In RBNP, there are buffer zone forests from which local people are allowed to

harvest resources, but in RSWR, there is no buffer zone forest. This is why more

households of RBNP graze and collect fodder for livestock in buffer zone forests and

public lands than households of RSWR. Firewood, thatch and timber are the three most

important resources for local people. People of RBNP are collecting these resources in

greater quantities than people of RSWR because they have places to harvest resources

apart from the park. Due to the availability of resources and low annual cash income,

respondents of RBNP are more dependent on resources.

Traditional trade systems of natural resources such as exchange for wage labor and barter

for food materials, and to some extent conventional trade in local markets, were common

among respondents. Significantly, a higher proportion of respondents of RBNP said that

57 they were engaged in buying and selling of natural resources. Although RSWR is close to

an urban center, the more trade in RBNP could be attributed to more availability and

demand of resources. Irrigation facilities, access to market, technical support, and scarcity of resources all are impetuses to local people to adopt domestic cultivation of non-timber forest products in their agricultural fields. Since these conditions are met in

RSWR, more respondents are interested in the program. No initiation had been taken to promote NTFPs from buffer zone forests during this study. Sustainable extraction of

NTFPs is a silvicultural practice for forest management and conservation that bolsters the local economy by providing income generating activities (Mahapatra & Mitchell 1997).

Present rules and regulations are silent on the commercialization of NTFPs from buffer zone forests in Nepal. There is a need to forge clear guidelines to this end.

Popular participation in conservation can be achieved through various means. Becoming a member of grassroots institutions, participating in trainings, engaging in income generating activities, depositing money in saving-credit programs, and rewarding people for conservation initiatives are some forms of participation. There are many grassroots institutions in buffer zones such as User Groups, Women Groups, Forest Groups etc., and

local people are exclusive members. Sewing-cutting, painting, vegetable farming, goat

keeping, and pig raising trainings are given to local people for income generation, and leadership, conflict resolution and accountancy trainings are given to members of grassroots institutions. During the household surveys, I found that many recipients are not using new skills acquired through training for income generating purposes, but for subsistence use. These may be good for extension work but they are not meeting the

58 objective. This is the reason why income-generating activities are not so prominent, and

they failed to empower local people. Although there was no significant association

2 between ethnicity and participation in training (χ 1 = 0.33, p = 0.506), more Tharus were

2 participating in income generating activities (χ 1 = 5.49, p = 0.019). Biogas, improved

stoves, piglet raising, drinking water, and scholarships are some of the benefits reported by respondents.

2 Significantly more Tharus got personal benefits than non-Tharus (χ 1 = 3.95, p = 0.047)

that is justifiable owing to their lower socioeconomic standing. Wider participation in

RBNP is attributed to the declaration of the buffer zone and subsequent earmarking of revenue to local communities. Many NGOs are working in the buffer zone of RBNP implementing various integrated conservation and development programs. To reflect the bottom-up approach, NGOs solicit local people’s participation while implementing programs. The ultimate goal of participation should be empowerment of local people so that they have more control over resources.

In the periphery of RBNP, public forests exist which are absent in the periphery of

RSWR. Some of these forests were handed over to local people to manage under the buffer zone regime. With the change in management regime, these formally degraded forests are regenerating so well that local people have started harvesting resources. Local people in collaboration with park management can thus be successful in converting open access degraded forests into well protected neo-common property resources. This is the

Government's effort to revive some aspects of common property regimes. Thus buffer

59 zone management could be regarded as what Arnold (1998) calls 'emerging common property regime'. This is the reason why, in RBNP, 43% of respondents disagree with the statement that forests in their area have decreased, and an overwhelming proportion

(96%) agree with sharing responsibilities to protect natural resources. There has been a tremendous change in forest cover within past three decades, but local people’s disagreement suggests that they take the recent time frame at the local scale to form their attitudes. Success stories of community forestry and buffer zone forests have spurred local people of RSWR to demand fringes of the reserve (up to 300 meter from the periphery) as buffer zone forests. Some UGs had submitted a concept paper to this end to the reserve authority. At the time of fieldwork, some UGs had issued grass cutting permits with verbal consent from the reserve authority. Considering overwhelming social pressures and the de facto open access status of the reserve periphery, it will be a strategic management decision to go along with UGs proposal. In the present rules and regulations, there is no provision for handing over part of a park or reserve as a buffer zone forest. Legal mechanism should be explored before initiating such programs.

People’s assessment of wildlife populations was based on sightings during thatch collection inside protected areas, sightings in buffer zone forests, and frequency of crop damage. Over the past decade, populations of one-horned rhinoceros and Asian elephants have increased in RBNP and these animals are causing more problems for local people.

About 91% of respondents said that there has been an increase in wildlife populations in

RBNP. Local people have frequently seen wildlife inside the park and the buffer zone.

According to local people, reasons for the increase are complete protection, decrease in

60 poaching and addition of habitats (buffer zones). In RSWR, people do not go deep into the reserve to collect thatch grass, and the periphery of the reserve is not good habitat due to livestock grazing and illegal resource extraction by local people. These are the reasons why local people do not frequently see wildlife that are abundant inside the reserve. In

RSWR, 56% of respondents said that wildlife populations have decreased. Crop depredation is thus less severe in RSWR and most people have not seen wildlife in their fields. RSWR abuts Laggabaggha Sanctuary of India so local people opine that decreases in wildlife populations are due to migration of wildlife to India and rampant poaching on the border. In fact, there been no records of wildlife decline in recent years in these two protected areas except for blue bull Boselaphous tragocamelus, white-rumped vulture

Gyps bengalensis, slender-billed vulture G. tenurostris and Bengal Florican (Khatri 1993;

Giri & Gharty-Chhetri 2002; Baral et al. 2003). Populations of many other species (e.g. four species of deer, wild pigs, etc.) have increased.

People confound improvements in living condition with resource access, physical facilities and wildlife control. Local people are legally allowed to harvest natural resources from buffer zone forests as stipulated in the management plans. Previously, parks and reserves were the only source for thatch, but now it is found in many buffer zone forests as well. In the case of RBNP, local people are harvesting more thatch from these forests than from the park. Additionally buffer zone forests provided firewood, timber and fodder which are not allowed to be collected from the park. Some NGOs that have been working in the buffer zone of RBNP are the KMTNC, the PCP, the TAL, and the CARE-Nepal. These NGOs introduced many integrated conservation and

61 development programs. Skill enhancement and income generating activities through training and welfare programs for marginalized communities to some extent led to better living standard.

These projects have also invested in community development programs such as road construction, drinking water supply, irrigation in agricultural lands, construction of schools, and establishment of health care centers. Local people praise the development components of these projects. The park authority, with its partner organizations, has addressed the problem of wildlife damage in local communities. The investment in trench and fence maintenance to deter wildlife entering crop fields, and compensation for the loss of life and damage to houses by endangered wildlife species, are taken as positive steps by the park authority. Economic development as a result of these projects is more pronounced in RBNP. Thus, significantly more respondents agreed that there has been an increase in the standard of living with the establishment of park.

The conflict arises when restrictions are placed on resource harvest. In the establishment phase, UGs put restrictions on resources harvested in buffer forests and people felt uncomfortable for a while. In such a situation, protected areas are vulnerable to trespassing because people illegally harvest resources to meet their needs (Straede &

Helles 2000). When local people were allowed to harvest resources from buffer zone forests, the conflict abated. To remediate resource scarcity, respondents in RBNP emphasized buffer zone forest management while in RSWR, they suggested permits to harvest from the reserve. The example of buffer zone forest is given by many respondents

62 to substantiate their views on increased forest cover and shared responsibilities for

conservation. Cursory observation in RBNP suggests that buffer zone forests have

provided, to some extent, social and ecological buffers to the park, but more in-depth

analysis is warranted.

People have stakes in participation and sharing of responsibilities for conservation. These

are guided by the expectation of direct use of resources. They think that the resources that have been conserved will ultimately used by them. They used to ask why the park does not allow us to harvest old, dead and fallen trees. Most respondents think that wildlife

does not have direct use values except for a few who see the tourism potential of the area.

Non-consumptive uses of wildlife are obvious in many protected areas (Bandara &

Tisdell 2003), and this is the reason why it is taken as a revenue generating source for the

government. People vehemently oppose the concept of park extension. According to local

people, agricultural land is more desirable than protected areas. When survival is at stake,

people discount ecological services provided by protected areas. Most productive

protected areas in Nepal are in the Terai, where demand of agriculture lands is very high.

The management authority should choose between quality and quantity. Instead of

increasing size, the government should expand strategies of participatory and intensive

management of existing protected areas and the matrix of forests around them.

Exclusion of outsiders and the free riding problem are challenges for efficient

management of common pool resources. Respondents are skeptical about their

competency in sole management of resources. Local people are helpless when their

63 buffer zone forests were encroached by squatters, freed bonded-laborers, and political refugees. They doubt they would be able to keep common pool resources as closed access in absence of strong support from the government for law enforcement. Experience elsewhere shows that local people cannot fight with international poachers and cannot exclude outsiders free riding their resources (Spinage 1998). Local people are efficient police – a rotational guard system can check irregularities at the local level. The problem of poachers and wildlife traders should be addressed at national and international levels, which is a function of the state. The concerted effort of local people and the state based on participatory principles is important for emerging common property regime.

Respondents citing the example of community forests also hint for collaborative management.

Multiple regression results indicated that RBNP respondents have more favorable attitudes than RSWR respondents do (p = 0.049). This is substantiated by the fact that the mean attitude score of RBNP (8.4 ± 1.44) was significantly higher than the score of

RSWR (7.7 ± 1.66; t = 3.24, p = 0.0007). The buffer zone of RBNP (328 sq km) was declared in 1996, and since then local people have gotten 30-50% of the revenue generated by the park. This has observable impacts on socioeconomic development of the area and empowerment of local people for resource management. During the time of this fieldwork, the proposed buffer zone of RSWR was not officially declared. The presence of NGOs and their programs to provide socioeconomic benefits are pronounced in the buffer zone of RBNP while one NGO was working in the proposed buffer zone of RSWR and its programs are diffuse. Due to the provision of buffer zone forests, more resources

64 are available to RBNP respondents. These are some reasons that help to explain differences in conservation attitudes between two areas.

The objective of buffer zone management is to relieve human pressures on protected areas by providing socioeconomic benefits to local people. To meet the objective, some of the strategies employed are empowerment of local people, management of problem wildlife, and harnessing local institutions. In most integrated conservation and development programs, training is an integral component of core programs. In the study area, training sponsored by NGOs was a significant predictor of conservation attitudes (p

= 0.011). The main objectives of training programs are to enhance skill development of local people and to elevate environmental awareness. Therefore, NGOs take an opportunity to educate participants about conservation during training sessions. Research elsewhere show that people who participated in training programs held more favorable attitudes towards conservation (Mehta & Kellert 1998; Mehta & Heinen 2001). Wildlife damage is strongly associated with negative attitudes towards conservation (Heinen 1993;

Newmark et al. 1993; De Boer & Baquete 1998; Mehta & Kellert 1998). As expected, respondents who suffered from wildlife held more negative attitudes towards conservation (p = 0.024). Mostly people living in the periphery of protected areas are subsistence farmers and research elsewhere shows that people who farm tend to have more negative attitudes than those engaged in off-farm activities (Akama et al. 1995).

Community level benefits do not off set individual costs (Gibson & Marks 1995), and people have negative attitudes when their livelihoods are under threats from wildlife

(Gillingham & Lee 1999). This is corroborated with the fact that people who expressed

65 satisfaction towards UGs were more likely to hold favorable attitudes towards conservation (p = 0.000). The formation of UGs provided the platform for wider participation. When public forests are handed over to UGs as buffer zone forests, local people have control over resources. Loss of local control over resources results in negative attitudes (Mehta & Heinen 2001). By participating in community forestry program, local people have favorable attitudes (Mehta & Kellert 1998). This is because people have control over resources and appreciate the importance of ownership.

Field research conducted in developing countries shows that among demographic variables, age, education, gender, and ethnicity are significant predictor of conservation attitudes (Fiallo & Jabcobson 1995; Mehta & Kellert 1998; Gillingham & Lee 1999; Sah

& Heinen 2001). None of these variables was a significant predictor of attitudes in this study (p > 0.05). According to Sah and Heinen (2001), Tharus held more negative attitudes towards conservation than others in Ghodaghodi Tal, Nepal. This is contrary to the finding that there is no significant difference in conservation attitudes among ethnic groups here (p > 0.05).

The surprising finding of multiple regression is that people benefiting personally tend to have more negative attitudes, however, this is not statistically significant (p = 0.381).

This may be because most personal benefits such as sewing machines, improved stoves, piglets etc. are distributed after the completion of training. People might have attributed benefits to training and not to conservation. Other reasons could be inequitable

66 distribution of benefits (Ferraro & Kramer 1997) and perceptions that benefits are ‘donor provided’ rather than ‘conservation earned’ (Lewis & Phiri 1998).

With the amendment in National Park and Wildlife Conservation Act of 1973, the Buffer

Zone Management Regulation (BZMR) was passed in 1996. The Regulation mentions

User Committee (UC) only. As per the Regulation, the UC should have at least nine members in the executive committee and five years tenure. The UCs are free to hold meetings as required and decisions are made by simple majority. The Regulation is progressive because it mentions hunting concessions of common species in buffer zones and describes methods of reimbursing compensation for wildlife damage. After a lapse of three years, the Buffer Zone Management Guideline (BZMG) came into effect. The

Guideline imposed restrictions on the number (21 at most) of the UCs within a buffer zone. To achieve the goal, a new grassroots institution, the User Group (UG), was introduced which was not in the Regulation. All previously functioning UCs were relegated to UGs. The regressive action alienated local people and many UG chairs do not agree with the present Guideline. This is the reason why many UG chairs (9 of 14) of

RBNP suggested amendments in the Regulation and the Guideline.

As per the Guideline, UCs are formed at the Village Development Committee (VDC

[political constituency]) level with representatives from UGs within the VDC. People can exercise voting rights to elect UGs, but not to elect UCs. Previously, UGs directly communicated with the buffer zone council, but with the addition of one vertical tier,

67 UCs became mediators between them. There is a clear shift from participation to

representation, and UGs lost a great deal of authority in the vertical power distribution.

There are some serious flaws in the Guideline. First, it puts restrictions on the number of

members (at most nine) on the executive committee. The Regulation is flexible on the

size of the executive committee. The Rule 8 (2) mentions at least nine members on the

executive committee. Second, the term of the UGs and the UCs is debatable. Section

12(18) of the Guideline describes the term of the UGs as 2.5 years. However, the term of

officials of the UCs is five years. All officials of the UCs are represented from the UGs.

Therefore, the term of the officials expire before the completion of the UCs tenure. In

such cases, the validity of the UC is questionable on legal grounds. Forming UGs every

two and half year incurs many administrative burdens and may undermine a stability of

grassroots institutions. In theory, the Guideline should be clear and progressive, but in

practice, it is ambiguous and regressive. Reconciling contradictions between the

Regulation and the Guideline is imperative.

The Guideline demands meetings of the UGs. The Section 12 (9) dictates that UGs

should meet once every two weeks. The Regulation is lenient on the UC meetings because UCs could hold meetings as and when required. However, the Section 10 (5) of the Guideline stipulates frequency of the UC meetings at least four times a year not exceeding three months gaps between each meeting. Proposals sent by the UGs are discussed at UC meetings and decisions are made. UGs do not have long agendas to

68 discuss so it is not effective to make fortnightly meetings mandatory because, in practice, most UGs are violating the rule.

The rubric of the programs to be implemented and allocation of the budget are explicitly mentioned in the Guideline. At least, 30% of the total budget should be allocated for conservation programs. However, in the work plans, there is no consistency of budget allocation as per the Guideline. Community development programs received high priority while allocating the budget.

6 CONCLUSIONS

The participation of women in conservation and development programs is limited in

Nepal. This is due to gender-based differences in socioeconomic status and education level. The literacy rate among women was only 40% compared to 76% of men. Similarly, more ethnic Tharus were illiterate (36%) than non-Tharus (29%). The low illiteracy rate among women and ethnic groups may hinder their participation in conservation. There is a need to increase their literacy rate so that it helps in their empowerment and fosters in participation. Still 77% of respondents primarily depend on subsistence agriculture and

86% of respondents had land tenure. The Government’s inefficient land reform policies – letting squatters and refugees settle in forests, and fail to set land tenure as benchmark to prevent encroachment – and lack of off-farm income has increased pressures on natural resources. More than 95% of respondents raised one or more types of livestock, the mean livestock size unit was 4.18 ± 3.75, and none raised improved breeds of livestock.

69 Traditional livestock raising practices are not compatible with conservation objectives.

The Government should give priority to coordination among stakeholders, and make

policy reforms in agriculture and livestock sectors for successful implementation of

landscape conservation.

Eight types of natural resources were extracted from the park and buffer zone forests of

RBNP while seven types were extracted from the reserve in RSWR. Respondents in

RBNP were more dependent on natural resources than in RSWR. There is spatio-

temporal variability in resource dependency and use patterns. There has been slight decrease in resource dependency from historic times, but still most people are heavily dependent on natural resources for subsistence. Poor and ethnic Tharus were more dependent on natural resources which suggests that forest resources are safety net for them. Although Tharus are more dependent on natural resources, their traditional practices are superior to non-Tharus from conservation perspectives. Resource use patterns among ethnic groups should be taken into account for sustainable management of resources. Resource dependency is a function of socioeconomic status, ethnicity, and availability. Income generation activities may help to alleviate pressures on natural resources.

After the promulgation of community-based conservation legislation, the park authority and non-governmental organizations instituted a platform for local people to participate in conservation and sustainable development programs. Local people became members of grassroots institutions, participated in trainings, formed saving-credit groups, and

70 engaged in income generating activities. More NGOs were in the buffer zone of RBNP

than in RSWR and this could be the reason why more respondents of RBNP had

participated in such activities than of RSWR. At present forms, these activities failed to

empower local people but garnered favorable attitudes towards conservation.

The mean conservation attitude score of RBNP respondents (8.4 ± 1.4) was significantly

higher than those of RSWR (7.7 ± 1.6; p = 0.0012). The difference in conservation

attitudes is due to the conservation intervention programs. The significance presence of

NGOs, empowerment of grassroots institutions, and socioeconomic development

contribute to favorable conservation attitudes. These are reasons why local people of

RBNP have more favorable attitudes towards conservation than RSWR. A multiple

regression model showed that participation in trainings, wildlife damage and satisfaction towards User Groups (UGs) are significant predictors of conservation attitudes.

All UGs of RBNP had female members on their executive committees while only three

UGs of RSWR did. In RBNP, 86% of UGs were holding their regular meetings while only 47% of UGs in RSWR did at the time of survey. All UGs of RBNP had five year work plans, but only 60% UGs of RSWR had. About 71% of UGs in RBNP had their own buffer zone forests, but none of RSWR UGs had. When grassroots institutions are legally recognized and management authority is delegated, they are more strengthened.

They are more resilient in times of political instability because of wider support. Taking into account the institutional capacity, demand for land and pressures on natural resources, the government should expand strategies of participatory and intensive

71 management of existing protected areas and matrix of forests around them. The evolution of landscape approach is timely and may help secure sustainability of the entire landscape.

72 TABLES

Table 1. Frequency distribution of ethnicity in two protected areas

Ethnicity/castes RBNP RSWR Total Statistics Tharu 64 (51.2%) 11 (10.1%) 75 (32.1%) χ2 = 48.85 Brahman 23 (18.4%) 29 (26.6%) 52 (22.2%) p = 0.000 Chhetri 26 (20.8%) 49 (44.9%) 75 (32.1%) df = 4 Occupational castes 8 (6.4%) 18 (16.5%) 26 (11.1%) Hill tribes 4 (3.2%) 2 (1.8%) 6 (2.6%) Total 125 109 234

Table 2. Frequency distribution of education level of respondents of two areas

Education Level RBNP RSWR Total Statistics Illiterate 47 (37.6%) 26 (23.85%) 73 (31.2%) χ2 =18.1410 Literate 28 (22.4%) 19 (17.4%) 47 (20.1%) p = 0.001 Primary 24 (19.2%) 16 (14.7%) 40 (17.1%) df = 4 Secondary 23 (18.4%) 33 (30.3%) 56 (23.9%) College 3 (2.4%) 15 (13.8%) 18 (7.7%) Total 125 109 234

Table 3. Comparison of family size between area and ethnic groups

Area_ethnicity N Mean Min Max Mean Rank RBNP and Tharu 64 9.23 2 32 137.92 RBNP and non-Tharu 61 5.95 3 10 88.39 RSWR and Tharu 11 7.00 4 10 118.36 RSWR and non-Tharu 98 7.71 1 23 122.77 Total 234 7.64 1 32

Table 4. Percentage of immigrants in two protected areas

Protected Areas Migrated Not-migrated Total Statistics RBNP 87 (69.6%) 38 (30.4%) 125 χ2 = 28.25 RSWR 105 (96.3%) 4 (3.7%) 109 p = 0.000 Total 192 (82.1%) 42 (17.9%) 234 df = 1

73 Table 5. Average landholdings in hectare among different ethnic groups in two areas

Area_Ethnicity N Mean Min Max Mean Rank RBNP and Tharu 64 0.68 0.07 2.71 111.57 RBNP and non-Tharu 61 0.69 0.03 3.15 116.93 RSWR and Tharu 11 0.96 0.20 2.03 155.14 RSWR and non-Tharu 96 0.67 0.03 4.74 117.50 Total 232 0.69 0.03 4.74

Table 6. Percent of respondents meeting need of staple food from their farm

Protected Areas Sufficient Insufficient Total Statistics RBNP 49 (39.2%) 76 (60.8%) 125 χ2 = 27.5224 RSWR 80 (73.4%) 29 (26.6%) 109 p = 0.000 Total 129 (55.1%) 105 (44.9%) 234 df = 1

Table 7. Average and range of livestock size unit in two protected areas

Livestock Size RBNP (N = 119) RSWR (N = 106) Units (LSU) N Mean Min Max N Mean Min Max Cattle 82 3.52 0.80 38.40 100 2.72 0.40 10.40 Buffalo 68 2.34 0.50 7.00 75 1.81 0.50 5.50 Sheep/Goat 72 0.68 0.20 2.80 28 0.42 0.20 1.00 Pig 52 0.45 0.30 1.80 3 0.30 0.30 0.30

Table 8. Frequency of resources harvested by respondents of two protected areas

Extractive resource RBNP (N = 125) RSWR (N = 109) χ2 categories Yes No Yes No (p value) Firewood 68.0% 32.0% 57.8% 42.2% 0.106 Thatch 92.8% 7.2% 77.9% 22.0% 0.001 Grasses 52.0% 48.0% 44.0% 55.9% 0.224 Leaf litter 61.6% 38.4% 33.9% 66.1% 0.000 Edibles 42.4% 57.6% 10.1% 89.9% 0.000 Tree fodder 20.0% 80.0% 1.8% 98.2% 0.000 Herbs 15.2% 84.8% 0.9% 99.1% 0.000 Timber 40.8% 59.2% - - -

74 Table 9. Nonparametric correlation of resource use score with continuous variables

Variables Spearman’s Rho p value N Significant Family Size 0.11 0.093 234 Yes at p < 0.10 Landholdings -0.08 0.213 232 No Cash Income -0.27 0.001 201 Yes Livestock Size Unit 0.11 0.086 225 Yes at p < 0.10 Recency time 0.32 0.000 234 Yes

Table 10. Frequency distribution of resource dependency in two areas

Dependent scale RBNP RSWR Total Statistics Not dependent 7 (5.6%) 26 (23.8%) 33 (14.1%) χ2 = 35.62 Somewhat dependent 26 (20.8%) 43 (39.4%) 69 (29.5%) p = 0.000 Dependent 63 (50.4%) 31 (28.4%) 94 (40.2%) df = 3 Most dependent 29 (23.2%) 9 (8.3%) 38 (16.2%) Total 125 109 234

Table 11. Number and percentage of households mentioning the grazing sites

Where people graze livestock? RBNP RSWR Statistics Forests 10 (8.4%) 12 (11.3%) χ2 = 23.42 Community pastures 33 (27.7%) 4 (3.8%) p = 0.000 Private and stall feed 76 (63.9%) 90 (84.9%) df = 2 Total 119 106

Table 12. Number and percentage of households mentioning the source of fodder

Where from people get fodder? RBNP RSWR Statistics Private land 87 (73.1 %) 91 (85.8 %) χ2 = 5.51 Public and community forests 32 (26.9 %) 15 (14.2 %) p = 0.019 Total 119 106 df = 1

Table 13. Number and percentage of households mentioning the source of energy

What people use for cooking? RBNP RSWR Statistics Energy inefficient mud stoves 112 (89.6%) 98 (89.9%) χ2 = 0.02 Improved mud stoves 5 (4.0%) 4 (3.7%) p = 0.991 Alternative energy sources 8 (6.4%) 7 (6.4%) df = 2 Total 125 109

75 Table 14. Number and percentage of respondents suggesting measures to solve the problem of firewood scarcity

Suggestions for firewood problem RBNP RSWR Statistics Permit to collect from parks 17 (20.0%) 46 (54.7%) χ2 = 35.98 Buffer zone forests 25 (29.4%) 4 (4.7%) p = 0.000 Private plantation 6 (7.1%) 13 (15.5%) df = 4 Improved stoves 8 (9.4%) 3 (3.6%) Alternative energy 29 (34.1%) 18 (21.4%) Total 85 84

Table 15. Frequency distribution of households participating in conservation interventions in two areas

Conservation RBNP (N = 125) RSWR (N = 109) χ2 interventions Yes No Yes No p value Membership 36 (28.8%) 89 (71.2%) 14 (12.8%) 95 (87.2%) 0.003 Trainings 44 (35.2%) 81 (64.8%) 21 (19.3%) 88 (80.7%) 0.007 Benefits 45 (36.0%) 80 (64.0%) 11 (10.1%) 98 (89.9%) 0.000 Saving-Credit 67 (53.6%) 58 (46.4%) 72 (66.1%) 37 (33.9%) 0.053 Income 22 (17.6%) 103 (82.4%) 5 (4.6%) 104 (95.4%) 0.002 generation

76 Table 16. Percent of respondents agreeing or disagreeing with conservation statements

S. Statements RBNP (N = 125) RSWR (N = 109) χ2 No. Disagree Agree Disagree Agree p value 1 Forests around your village 43.55 56.45 30.28 69.72 0.037 have decreased in recent years. 2 It is responsibility of local 04.13 95.87 11.01 88.99 0.047 people to protect natural resources. 3 If there is unlimited access to 00.83 99.17 00.92 99.08 0.941 forests for fuel wood and fodder, forests will be disappeared soon. 4 There are more wild animals 09.09 90.91 55.96 44.04 0.000 now than ten years ago. 5 What people and their 67.77 32.23 66.06 33.94 0.783 livestock need are more important than saving plants and wild animals. 6 My living condition 30.40 69.60 55.05 44.95 0.000 improved since the protected area’s creation. 7 After the establishment of 19.35 80.65 40.74 59.26 0.000 buffer zone forests/reserve you don’t have problem of access to resources. 8 It is important to set aside a 19.17 80.83 22.22 77.78 0.569 place for the animals and plants to live in. 9 It is important to protect the 21.60 78.40 17.59 82.41 0.443 animals and plants so that our children may know and use them. 10 There is an equitable 17.74 82.26 19.44 80.56 0.739 distribution of common pool resources and benefits. 11 You are willing to contribute 08.26 91.74 02.78 97.22 0.073 for conservation cause.

77 Table 17. Multiple regression of conservation attitude score on demographic and socioeconomic variables

Independent Variables Coefficient Std. Error t p Protected Areas (RSWR = 1) -0.556 0.280 -1.99 0.049* Gender (Male = 1) 0.097 0.289 0.34 0.736 Age (in years) 0.007 0.009 0.75 0.452 Education (formal schooling =1) 0.037 0.097 0.38 0.704 Occupation (non-agriculture = 1) 0.149 0.119 1.25 0.212 Ethnicity (non-Tharus = 1) 0.138 0.114 1.21 0.227 Family Size 0.009 0.034 0.27 0.784 Landholdings (in ha) 0.043 0.195 0.22 0.827 Livestock Size Unit 0.005 0.037 0.12 0.901 Resource Dependency Score 0.012 0.016 0.69 0.489 Annual Income (log transformed) 0.110 0.117 0.94 0.349 Memberships (Yes = 1) 0.207 0.257 0.80 0.423 Trainings (Yes = 1) 0.620 0.241 2.57 0.011* Personal Benefits (Yes = 1) -0.231 0.263 -0.88 0.381 Income Generating Activities (Yes = 1) 0.435 0.301 1.45 0.149 Wildlife Damage (Yes = 1) -0.531 0.234 -2.27 0.024* Satisfaction Towards UGs (Yes =1) 0.941 0.247 3.81 0.000*

F17, 169 = 3.25, p = 0.0000, R-squared = 0.2464, * significant at p<0.05

Table 18. Frequency distribution of mode of Users’ Group formation

Mode of Formation RBNP RSWR Consensus 10 (71%) 13 (87%) Election 4 (29%) 2 (13%) Total 14 (100%) 15 (100%)

Table 19. Perception of UG chairs on effective institute for resource management

Effective Institutions RBNP RSWR Government 3 (22%) 5 (33%) Local People 9 (64%) 9 (60%) Both 2 (14%) 1 (7%) Total 14 (100%) 15 (100%)

78 Table 20. Frequency distribution of responses whether UGs are fulfilling demands

Demands fulfillment RBNP RSWR No 7 (50%) 9 (60%) Yes 7 (50%) 6 (40%) Total 14 (100%) 15 (100%)

Table 21. UG chairs’ attitudes towards TAL in two areas

Satisfaction with TAL RBNP RSWR Not at all 4 (29%) 14 (93%) Moderately 7 (50%) 1 (7%) Highly 3 (21%) - Total 14 (100%) 15 (100%)

Table 22. UG chairs attitudes towards the BZMR and Guideline

Regulations RBNP RSWR Agree 5 (36%) 12 (80%) Disagree 9 (64%) 3 (20%)* Total 14 (100%) 15 (100%)

* Respondents had never studied the rules and regulations so were unable to comment

Table 23. Losses of agencies under the Ministry of Forest and Soil Conservation

Departments District Based Area Range Training Armed Security Total Offices Post Centers Camps DoF 22 39 217 2 2 282 DNPWC 2 4 13 - - 19 DSWC 4 - 1 - - 5 Total 28 43 231 2 2 306

Source: Karki and Bhattarai 2004

79 FIGURES

Figure 1. Map of Nepal depicting spatial distribution of protected areas

Photo Courtesy: Heinen 2001

80 Figure 2. RBNP and RSWR with sampled households in the buffer zones

81 Figure 3. Average minimum and maximum monthly temperatures of two areas for the

period 1987-2001

40 RBNP Max RBNP Min 35 RSWR Max s

u RSWR Min i 30 ls e C 25 ee r g e D

20 n i e

ur 15 at er p 10 m e T 5

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months

Figure 4. Average monthly precipitation of two areas for the period 1987-2001

800 RBNP 700 RSWR

r 600 e t e

m 500 li i M 400 in ll a f 300 in

Ra 200 100 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Months

82 Figure 5. Percentage of male and female respondents in two protected areas

RBNP RSWR Male Male Female Female Female Female 26 22 24% 18%

Male Male 83 103 76% 82%

Figure 6. Percent of respondents in five occupation categories

100 90 RBNP

s 80 RSWR nt

de 70 on

p 60 s e 50 R

of 40 t n

e 30 c r

e 20 P 10 0 Agriculture Jobs Menial work Business Students Occupation Types

83 Figure 7. Average family size among ethnic groups in two areas

10 RBNP 9 RSWR

e 8 z i S

7 ly i 6 m a 5 e F 4 ag r e

v 3 A 2 1 0 Tharu Brahman Chhetri Occupational Hill tribes castes Ethnicity

Figure 8. Average landholdings among ethnic groups in two areas

1.2 RBNP RSWR

e 1 r a t c 0.8 He

n i s 0.6 ing ld 0.4 ho nd

La 0.2

0 Tharu Brahman Chhetri Occupational Hill tribes Average castes Ethnicity

84 Figure 9. Frequency of households rearing four types of livestock in two areas

120 RBNP RSWR 100 s d l o

h 80 e us 60 Ho of

er 40 b m u

N 20

0 Cattle Buffalo Goat/Sheep Pig Livestock Types

Figure 10. Total number of tourist arrivals per year since the beginning of the Maoists insurgency

600

500 ) 0 0

0 400 ('

s l

va 300 Arri t s i

r 200 u o T 100

0 1996 1997 1998 1999 2000 2001 2002 2003 2004 Years

Source: Nepal Tourism Board 2005

85 REFERENCES

Akama, J. S., C. L. Lant, and G. W. Burnett. 1995. Conflicting attitudes toward state wildlife conservation programs in Kenya. Society and Natural Resources 8: 133- 144.

Allison, P. D. 1998. Multiple regression: a primer. The Pine Forge Press, California, USA.

Alvarez, M. D. 2003. Forests in the time of violence: conservation implications of the Colombian war. Pages 49-70 in S. V. Price, editor. War and tropical forests: conservation in areas of armed conflicts. Food Products Press, NY.

Anderson, P. N. 2001. Community-based conservation and social change amongst south Indian honey-hunters: an anthropological perspective. Oryx 35 (1): 81-83.

Archabald, K., and L. Naughton-Treves. 2001. Tourism revenue-sharing around national parks in Western Uganda: early efforts to identify and reward local communities. Environmental Conservation 28 (2): 135-149.

Arnold, J.E.M. 1998. Managing forests as common property. FAO Forestry Paper 136, FAO, Rome.

Bandara, R., and C. Tisdell. 2003. Comparison of rural and urban attitudes to the conservation of Asian elephants in Sri Lanka: empirical evidence. Biological Conservation 110:327-342.

Baral, N., N. Timilsina, and B. Tamang. 2003. Status of Bengal Florican Houbaropsis bengalensis in Nepal. Forktail 19: 51-55.

Bista, D. B. 1987. People of Nepal. 5th edition. Ratna Pustak Bhandar, Kathmandu, Nepal.

Browder, J. 1995. Redepmtive communities: indigenous knowledge, colonist farming systems, and conservation of tropical forests. Agriculture and Human Values 12 (1):17-30.

Brandon, K., K.H. Redford and S.E. Sanderson, editors. 1998. Parks in peril: people, politics and protected areas. The Nature Conservancy Press and Island Press, Washington, D.C., USA.

Brown, K. 1997. Plain tales from the grasslands: extraction, value and utilization of biomass in Royal Bardia National Park, Nepal. Biodiversity and Conservation 6: 59-74.

86 Budhathoki, P. 2003. Nepal’s conservation in crisis: empowering people to secure natural resources. Journal of Forest and Livelihood 2 (2): 72-75.

CBS. 2001. Statistical year book of Nepal 2001. Central Bureau of Statistics, Kathmandu, Nepal.

Chapagain, K. 2002. Thirty-four rhinoceros killed in eight months. The Kathmandu Post, December 20.

Chaudhary, R. P. 1998. Biodiversity in Nepal: status and conservation. Craftsman Press, Bangkok, Thailand.

Collier, P. 2000. Economic causes of civil conflict and their implications for policy. World Bank, Washington, D.C.

Cox, T. 1990. Land rights and ethnic conflict in Nepal. Economic and Political Weekly June 16-23:1318-1320.

Davalos, L. M. 2001. The San Lucas mountain range of Colombia: how much conservation is owed to the violence? Biodiversity and Conservation 10: 69-78.

De Boer, W. F., and D. S. Baquete. 1998. Natural resource use, crop damage and attitude of rural people in the vicinity of the Maputo Elephant Reserve, Mozambique. Environmental Conservation 25: 208-218.

Dinerstein, D., and E. D. Wikramanayake. 1993. Beyond 'hotspots': how to prioritize investments in biodiversity in the Indo-Pacific region. Conservation Biology 7:53- 65.

Dinerstein, E. 1979. An ecological survey of Royal Karnali Wildlife Reserve, Nepal. Part II: Habitat animal interactions. Biological Conservation 16:265-300.

DNPWC. 2001. Annual report (2000-2001). Department of National Parks and Wildlife Conservation, Kathmandu.

DNPWC. 2003. Royal Bardia National Park Management Plan 2001-2005. His Majesty Government of Nepal, Department of National Parks and Wildlife Conservation.

DNPWC. 2005. Revenue generation in protected areas. The Department of National Parks & Wildlife Conservation, Kathmandu, Nepal. Available from http://www.dnpwc.gov.np/ (accessed on February 9, 2005).

Dudley, J. P., J. R. Ginsberg, A. J. Plumptre, J. A. Hart, and L. C. Campos. 2002. Effects of war and civil strife on wildlife and wildlife habitats. Conservation Biology 16(2): 319-329.

87 Eltringham, S. K., and R. C. Malpas. 1993. The conservation status of Uganda’s game and forest reserves in 1982 and 1983. African Journal of Ecology 31: 91-105.

Ferraro, P. J., and R. K. Kramer. 1997. Compensation and economic incentives: reducing pressure on protected areas. Pages 187-211 in R. A. Kramer, C. Van Schaik, and J. Johnson, editors. Last stand: protected areas and the defense of tropical biodiversity. Oxford University Press, New York, USA.

Fiallo, E. A., and S. K. Jacobson. 1995. Local communities and protected areas: attitudes of rural residents towards conservation and Machalilla National Park, Ecuador. Environmental Conservation 22(3): 241-249.

Formoli, T. A. 1995. The impacts of the Afghan-Soviet war on Afghanistan’s environment. Environmental Conservation 22: 66-69.

Gadgil, M., F. Berkes, and C. Folke. 1993. Indigenous knowledge for biodiversity conservation. Ambio 22:151-156.

Gautam, G. 2004. Poachers rule the roost in hunting reserve. The Himalayan Times, July 09.

Gibson, C. K., and S. A. Marks. 1995. Transforming rural hunters into conservationists: an assessment of community-based wildlife management program in Africa. World Development 23 (6): 941-957.

Gillingham, S., and P. C. Lee. 1999. The impact of wildlife-related benefits on the conservation attitudes of local people around the Selous Game Reserve, Tanzania. Environmental Conservation 26(3): 218-228.

Gilmour, D. 2003. Retrospective and prospective view of community forestry in Nepal. Journal of Forest and Livelihood 2: 5-7.

Giri, J. B., and S. Gharty-Chhetri. 2002. Study of vultures in western lowland Nepal. Oriental Bird Club Bulletin 36: 11-13.

Goldstone, J. A. 1996. Saving the environment (and political stability too): institutional responses for developing nations. Environmental Change and Security Report 2: 33-34.

Hamilton, A., A. Cunningham, D. Byarugaba, and F. Kayanja. 2000. Conservation in a region of political instability: Bwindi Impenetrable Foreset, Uganda. Conservation Biology 14(6): 1722-1725.

Hardin, G. 1968. The tragedy of the commons. Science 162: 1243-1248.

88 Hart, T., and J. Hart. 1997. Conservation and civil strife: two perspectives from central Africa. Conservation Biology 11(2): 308-314.

Hartup, B. K. 1994. Community conservation in Belize: demography, resource use, and attitudes of participating landowners. Biological Conservation 69: 235-241.

Hegde, R., and T. Enters. 2000. Forest products and household economy: a case study from Mudumalai Wildlife Sanctuary, Southern India. Environmental Conservation 27 (3): 250-259.

Heinen, J. T. 1993. Park-people relations in Koshi Tappu Wildlife Reserve, Nepal: a socio-economic analysis. Environmental Conservation 20(1): 25-34.

Heinen, J. T. 1996. Human behavior, incentives, and protected area management. Conservation Biology 10: 681-684.

Heinen, J. T., and B. Kattel. 1992. Parks, people, and conservation: a review of management issues in Nepal’s protected areas. Population and Environment 14 (1): 49-84.

Heinen, J. T., and D. P. Chapagain. 2002. On the expansion of species protection in Nepal: advances and pitfalls of new efforts to implement and comply with CITES. Journal of International Wildlife Law and Policy 5: 235-250.

Heinen, J. T., and J. N. Mehta. 1999. Conceptual and legal issues in the designation and management of conservation areas in Nepal. Environmental Conservation 26(1): 21-29.

Heinen, J. T., and J. N. Mehta. 2000. Emerging issues in lagal and procedural aspects of buffer zone management with case studies from Nepal. Journal of Environment & Development 9(1): 45-67.

Heinen, J. T., and S. Rayamajhi. 2001. On the use of goal-oriented project planning for protected area management in Nepal. Environmental Practice 3: 227-236.

Heinen, J. T., and S. Shrestha. 2005. Evolving policies for conservation: an historical profile of the protected area system of Nepal. Environmental and Urban Planning

Heinen, J. T., and P. B. Yonzon. 1994. A review of conservation issues and programs in Nepal: from a single species focus toward biodiversity protection. Mountain Research and Development 14 (1): 61-76.

HMG. 1971. Soil survey of Bardia division. Forest resource survey publication. His Majesty Government of Nepal.

89 HMG. 1973. National Parks and Wildlife Conservation Act 2029. Nepal Gazette 29: 1- 14.

HMG. 1996. Conservation Area Management Regulations of 2053. Nepal Gazette 3, (Suppliment).

Hutt, M., editor. 2004. Himalayan people’s war: Nepal’s Maoist rebellion. Indiana University Press, Indianapolis, USA.

Infield, M. 1988. Attitudes of rural communities towards conservation and a local conservation area in Natal, South Africa. Biological Conservation 45: 21-46.

Infield, M., and A. Namara. 2001. Community attitudes and behaviour towards conservation: an assessment of a community conservation programme around Lake

Ite, U.E. 1996. Community perceptions of the Cross River National Park, Nigeria. Environmental Conservation 23 (4): 351-357.

IUCN. 1991. Caring for the Earth: a strategy for sustainable living. IUCN/UNEP/WWF, Gland, Switzerland.

Jain, B. C. 1991. Tharu tribe and the institution of marriage. Man and Life 17:163-168.

Jnawali, S. R., and P. Wegge. 1993. Space and habitat use by a small re-introduced population of greater one horned rhinoceros (Rhinoceros unicornis) in Royal Bardia National Park in Nepal- A preliminary report. Pages 208-217 in O. A. Ryder, editor. Proceedings of international conference on rhinoceros biology and conservation, Zoological Society, San Diego, USA.

Kaimowitz, D., and A. Faune. 2003. Contras and commandants: armed movements and forest conservation in Nicaragua’s Bosawas Bioshpere Reserve. Pages 21-47 in S. V. Price, editor. War and tropical forests: conservation in areas of armed conflicts. Food Products Press, NY.

Kane, H. 1995. The hour of departure: forces that create refugees and migrants. Paper 125. WorldWatch, Washington, D.C.

Kanyamibwa, S. 1998. Impact of war on conservation: Rwandan environment and wildlife in agony. Biodiversity and Conservation 7: 1399-1406.

Karki, A., and B. Bhattarai, editors. 2004. Whose war? Economic and socio-cultural impacts of Nepal's Maoist-Government conflict. NGO Federation of Nepal.

90 Khatri-Chhetri, S. 2004. Nepal still a hotspot for illegal wildlife traders. The Kathmandu Post, April 27.

Kellert, S. R., J. N. Mehta, S. A. Ebbin, and L. L. Lichtenfeld. 2000. Community natural resource management: Promise, rhetoric, and reality. Society and Natural Resources 13:705-715.

Khatri, T. B. 1993. Status and food habits of nilgai (Boselaphus tragocamelus) in Royal Bardia National Park, Nepal. Unpublished M. Sc. Thesis, Agricultural University of Norway, As, Norway.

Kremen, C., A. M. Merenlender, and D. D. Murphy. 1994. Ecological monitoring: a vital need for integrated conservation and development programs in the tropics. Conservation Biology 8(2): 388-397.

Krishna, A. 2002. Creating and harnessing social capital. In: social capital: a multi- faceted perspectives. Available from http://emedia.netlibrary.com

Lewis, D. M., and A. Phiri. 1998. Wildlife snaring - an indicator of community response to a community based conservation project. Oryx 32 (2): 111-121.

Mahapatra, A., and C. P. Mitchell. 1997. Sustainable development of non-timber forest products: implication for forest management in India. Forest Ecology and Management 94: 15-29.

Mainali, K. P. 2005. Sharp decline in Nepal’s rhino population. Frontiers in Ecology and Environment 5(3): 239.

Martin, P. S., and C. Szuter. 1999. War zones and game sinks in Lewis and Clark’s west. Conservation Biology 13(1): 36-45.

McNeely, J. A. 1993. Economic incentives for conserving biodiversity: lessons from Africa. Ambio 22 (2-3): 144-150.

Mehta, J. N. 1996. Park-people interface in Parsa Wildlife Reserve, Nepal. TRI News: A Journal of the Tropical Resource Institute 15: 11-12.

Mehta, J. N. and J. T. Heinen 2001. Does community-based conservation shape favorable attitudes among locals? An empirical study from Nepal. Environmental Management 28 (2): 165-177.

Mehta, J. N., and S. R. Kellert. 1998. Local attitudes toward community-based conservation policy and programmes in Nepal: a case study of the Makalu-Barun Conservation Area. Environmental Conservation 25 (5): 320-333.

91 MFSC. 2004. Terai Arc Landscape-Nepal: strategic plan 2004-2014. His Majesty Government of Nepal, Ministry of Forests and Soil Conservation, Kathmandu, Nepal.

Mishra, H. R. 1982. Balancing human needs and conservation in Nepal’s Royal Chitwan National Park. Ambio 11 (5): 246-251.

Muchagata, M., and K. Brown. 2000. Colonist farmers' perceptions of fertility and the frontier environment in eastern Amazonia. Agriculture and Human Values 17:371-384.

Mukwekwerere, M.C. 1996. Non-timber forest products: their role in subsistence economics in Zimbabwe. The Zimbabwe Science News 30 (2): 65-67.

Nepal, S. K., and K. E. Weber. 1995. Prospects for coexistence: wildlife and local people. Ambio 24: 238-245.

Nepali, S. 2005. Impacts of a political conflict in conservation and livelihoods. Unpublished report submitted to WWF- Nepal Program, Kathmandu, Nepal.

Nepstad, D., I. Brown, L. Luz, A. Alechandre, and V. Viana. 1992. Biotic impoverishment of Amazonian forest by rubber tapper, loggers, and cattle ranchers. Advances in Economic Botany 9:1-14.

Newmark, W. D., N. L. Leonard, H. I. Sariko, and D. G. M. Gamassa. 1993. Conservation attitudes of local people living adjacent to five protected areas in Tanzania. Biological Conservation 63: 177-183.

Noss, A. J. 1997. Challenges to nature conservation with community development in central African forests. Oryx 31 (3): 180-188.

NTB. 2005. Tourissm Resource Center-Downloads. Nepal Tourism Board, Kathmandu, Nepal. Available from http://www.welcomenepal.com/arrivalstat.asp (accessed on May 20, 2005)

Oates, J.F. 1995. The dangers of conservation by rural development – a case-study from the forests of Nigeria. Oryx 29 (2): 115-122.

Ostrom, E. 1990. Governing the commons: The evolution for collective action. Cambridge University Press, Cambridge.

Parry, D., and B. Campbell. 1992. Attitudes of rural communities to animal wildlife and its utilization in Chobe enclave and Mababe depression, Botswana. Environmental Conservation 19 (3): 245-252.

92 Phuyal, S. 2005. Big cats endangered. The Kantipur Publication, Kathmandu, Nepal. Available from http://www.kantipuronline.com/ (accessed on May 20, 2005).

Phuyal, S., and S. Adhikari. 2003. Makalu Barun: where illegal wildlife traders are defying Maoist rule. The Kathmandu Post, January 30.

Plumptre, A. J., J. B. Bizumuremi, F. Uwimana, and J. D. Ndaruhebeye. 1997. The effects of the Rwandan civil war on the poaching of ungulates in the Parc National des Volcans. Oryx 31: 265-273.

Price, S. V., editor. 2003. War and tropical forests: conservation in areas of armed conflicts. Food Products Press, NY.

Raj, P. A. 2004. Maoists in the land of Buddha: an analytical study of the Maoist insurgency in Nepal. Nirala Publications, New Delhi, India.

Raut, Y. 1997. Concept of stocking rate and carrying capacity: high stocking rate and its effects on pasture production and ecology. Pages 16-34 in A text book of animal husbandry I: pasture production. The Department of Livestock Services, Ministry of Agriculture, Kathmandu, Nepal.

Regmi, U. R. 2000. Status of tiger and livestock depredation in Royal Suklaphanta Wildlife Reserve, Nepal. Unpublished M. Sc. thesis, Agricultural University of Norway, As, Norway.

Richards, M. 1996. Protected areas, people and incentives in the search of sustainable forest conservation in Honduras. Environmental Conservation 23(3): 207-217.

Robinson, J. 1993. The limits to caring: sustainable living and the loss of biodiversity. Conservation Biology 7 (1): 20-28.

Sah, J. P. 2002. Vegetation dynamics and their implications for the management of wetlands in the lowlands of Nepal. Unpublished PhD dissertation, Floirda International University, Miami, USA.

Sah, J. P., and J. T. Heinen. 2001. Wetland resource use and conservation attitudes among indigenous and migrant peoples in Ghodaghodi Lake area, Nepal. Environmental Conservation 28(4): 345-356.

Schreckenberg, K. 1999. Products of managed landscape: non-timber forest products in the parklands of the Bassila region, Benin. Global Ecology and Biogeography 8: 279-289.

93 Sekhar, N.U. 1998. Crop and livestock depredation caused by wild animals in protected areas: the case of Sariska Tiger Reserve, Rajasthan, India. Environmental Conservation 25 (2): 160-171.

Sembajjwe, W. S. G. 1995. Sacred forests in Ganda society. Uganda Journal 42: 32-44.

Sharma, U. R., and P. B. Yonzon, editors. 2005. People & Protected Areas in South Asia. Resources Himalaya Foundation, Kathmandu and IUCN World Commission on Protected Areas, South Asia.

Shrivastava, R. J. 2002. Natural resource use and park-people relations at Kaziranga National Park and World Heritage Site, India. Unpublished M.Sc. thesis, Florida International University, Miami, USA.

Sierra, R. 1999. Traditional resource-use systems and tropical deforestation in a multiethnic region in North-west Ecuador. Environmental Conservation 26 (2): 136-145.

Southgate, D., and H. L. Clark 1993. Can conservation projects save biodiversity in South America? Ambio 22 (2-3): 163-166.

Sperling, L. 2001. The effect of the civil war on Rwanda’s bean seed systems and unusual bean diversity. Biodiversity and Conservation 10: 989-1009.

Spinage, C. 1998. Social change and conservation misrepresentation in Africa. Oryx 32 (4): 265-276.

Straede, S., and F. Helles 2000. Park-people conflict resolution in Royal Chitwan National Park, Nepal buying time at high cost? Environmental Conservation 27(4): 368-381.

Studsrod, J. E., and P. Wegge. 1995. Park-people relationships: the case of damage caused by park animals around the Royal Bardia National Park, Nepal. Environmental Conservation 22(2): 133-142.

Thapa, D., and B. Sijapati. 2003. A kingdom under siege: Nepal's Maoist insurgency, 1996 to 2003. The Printhouse, Kathmandu, Nepal.

The Kathmandu Post. 2004. Huge cache of animal skins, bones recovered. The Kathmandu Post, March 30.

Tripathy, B. 2003. Poaching reduces red panda population. The Kathmandu Post, December 30.

94 The Rising Nepal. 2003. Security negligence cause of rhino deaths. The Rising Nepal, March 1.

The Rising Nepal. 2004. Nepal transit point of wildlife trade in SA. The Rising Nepal, July 15.

Vogel, G. 2000. Conflict in Congo threatens bonobos and rare gorillas. Science 287: 2386-87.

Wunder, S. 2000. Ecotourism and economic incentives - an empirical approach. Ecological Economics 32:465-479.

Yonzon, P.B., and M.L. Hunter Jr. 1991. Cheese, tourists and red pandas in the Nepal Himalayas. Conservation Biology 5(2): 196-202.

95 APPENDICES

Appendix 1. The Maoist People's War and Conservation

1.1 Introduction

Developing countries are confronted with many challenges to conserve their rich biodiversity. Currently, civil war is one of the most serious problems jeopardizing conservation efforts. The frequency of wars has increased during the past 50 years. There were more than three times as many on-going wars in the 1990s than in the 1950s (Kane

1995; Collier 2000). Instances of war, civil strife, and political instability are rife in Asia,

Africa and Latin America (Hamilton et al. 2000; Davalos 2001; Dudley et al. 2002; Price

2003). These wars are endemic to developing countries where biodiversity is rich and pose problems for conservation (Hamilton et al. 2000; Sperling 2001). There is a vicious link between civil war and environmental degradation. War causes serious environmental degradation that ultimately escalates pressures on natural resources, which in turn causes shortages of resources and stokes more fuel for civil war (Dudley et al. 2002).

Conservation is very important in developing countries where rich biodiversity, political instability, a low priority in political agenda, and a dearth of financial resources to meet basic needs of large populations interact in a very intricate fashion, challenging conservation science. Armed conflict is detrimental to conservation (Davalos 2001); during or after the war environmental considerations are not a top priority (Kanyamibwa

1998).

Among developing countries, Nepal has been in the forefront for conservation, however, the Maoist People’s War has been a major setback in recent years. The Maoist insurgency

96 began in 1996, six years after the restoration of multiparty democracy. There is no single

explanatory factor attributed to the cause of insurgency. However, the ignorance of

democratic governments to recognize the plurality of Nepali society has served the

insurgency. A majority of rural people felt that they are discriminated against in ethnicity,

caste, language, and religion by the state. Even after the restoration of democracy,

inequalities in socioeconomic structures such as widespread poverty, caste/ethnic/gender

discrimination, political/social oppression, and corruption were prevalent and they have

stoked fuel for the war (Thapa & Sijapati 2003; Karki & Bhattarai 2004). During the past

nine years, about 12,119 people have died fighting for either the government or the

rebellion (Karki & Bhattarai 2004). The consequences of the war permeated into all

spheres of the nation.

One of the most obvious impacts is on the socio-economic sector. Most development

activities came to an abrupt halt in many parts of rural Nepal. The Government has

slashed development funds and diverted funds to security expenditures. Rebel attacks on

many hydropower plants, access facilities, communication networks and development

projects has caused substantial economic loss. Economic indicators plunged and the war

has had devastating effects on social and cultural aspects of society. Large numbers of

people are fleeing from Maoists strongholds to different parts of the country and, on

some occasions, outside the country. Abandoning huge tracts of agricultural field in

mountain districts has increased the food deficit. The livelihood of many poor people

depends on harvesting non-timber forest products. When the insurgency spread through rural areas, local people are deprived of these resources. They have no alternative means

97 for subsistence so they are forced to seek employment opportunities outside, which causes migration en masse. Many cultural practices were abandoned because of Maoist injunction, or simply out of fear.

There are many studies on socioeconomic consequences of the Maoist People’s War

(Thapa & Sijapati 2003; Hutt 2004; Karki & Bhattarai 2004; Raj 2004). So far, there has been no study on impacts of the war in the environment sector. This is because acquisition of relevant environmental data of war-inflicted areas is very difficult

(Sperling 2001). My analysis is based on a review of published literature, news in the local press, direct observations during field visits, and discussions with local people and conservation practitioners.

1.2 Sabotage of Conservation Agency

The top-level conservation agency of Nepal is the Ministry of Forests and Soil

Conservation. It has four departments: the Department of National Parks and Wildlife

Conservation (DNPWC) administers a network of protected areas (PAs), the Department of Forests (DoF) looks after government forests outside protected areas including community forests, the Department of Soil and Watershed Conservation (DSWC) is in charge of watershed management and control of soil erosion, and the Department of Plant

Resources is concerned with research and policies for plant conservation. Since the inception of civil strife, these agencies are under siege and tremendous damage to infrastructure has threatened institutional stability. To date, the DNPWC manages 16 PAs of four different categories (nine national parks, three wildlife reserves, three

98 conservation areas, and one hunting reserve), which cover about 18% of the country’s

total area. So far, the Maoist rebels have destroyed 47 physical structures of the DNPWC

(Budhathoki 2003), which seriously undermines the integrity of park management.

Rebels vandalized two PAs: Dhorpatan Hunting Reserve (DHR) and Makalu Barun

National Park (MBNP; Phuyal & Adhikari 2003; Gautam 2004). They forcibly evicted all

staff, took valuable communication and other equipment and took charge of the area.

These areas are important habitats for endangered wildlife such as musk deer Moschus

moschiferus and snow leopard Panthera uncia (Chaudhary 1998). The conservation

status of these areas is largely unknown as is the case of resource exploitation in rebel’s

territory. Throughout the world, dedicated park staff have lost their lives for conservation

in times of civil strife (Hart & Hart 1997; Hamilton et al. 2000). Maoist rebels killed park

staff of Royal Suklaphanta Wildlife Reserve and Parsa Wildlife Reserve in ambushes.

There have been reports of skirmishes between rebels and security force in the two World

Heritage sites: Everest National Park and Royal Chitwan National Park. After these incidents, patrolling inside parks and reserves was seriously diminished and so is the visitation by tourists.

Forests outside the network of PAs are important for biodiversity conservation because they provide adjunct wildlife habitats and also serve as corridors for isolated PAs. Still large tracts of forests are under the jurisdiction of the DoF, and it manages these forests through its district and regional offices. Since the advent of the People’s War, the DoF is on the top of the list of rebel targets. As of January 2003, Maoist rebels destroyed 22 district-based offices, 39 area offices, 217 range posts, two training centers and two

99 armed security camps (Table 23). One of the strategies of rebels is to compel the

government to withdraw its presence from forested areas and use them as shelter and

training centers.

Conservation agencies in developing countries have been faced with chronic lack of

funds, equipment and trained staff, and are frequently institutionally unstable (Heinen &

Kattel 1992; Hamilton et al. 2000). The rebel’s brutal act of killing park staff, damaging

physical infrastructure and forcing staff to leave will hamper the institutional

strengthening and stability of such conservation agencies. These will have short and long-

term detrimental impacts on conservation and may jeopardize past achievements.

1.3 Community-based Conservation

During the 1990s, Nepal embarked on community-based conservation (CBC) for natural

resource management and conservation. The CBC approach got momentum when

conservation areas, buffer zones, and community forests were recognized by legislation

and subsequently instituted. CBC is hailed as a critical approach in areas of political

instability (Hamilton et al. 2000) because local people express resentment towards strictly

protected parks and reserves (Heinen 1993). Although the CBC approach is more

resilience than the ‘fortress and fine’ approach, it is still not secure during civil strife. All three conservation areas: Annapurna, Manaslu, and Kanchanjunga, are impacted to varying degrees by civil strife (Budhathoki 2003). The southern flank of Annapurna

Conservation Area (ACA) is essentially under rebel control. Of seven field-based offices, four that lie in southern ACA were bombed and project staff deserted (Dhakal 2004).

100 Maoist rebels mercilessly killed three local conservation leaders who were proponents of

Annapurna Conservation Area Project (ACAP). These incidents intimidated local people, who no longer come forward to participate in conservation and development projects.

The integrated conservation and development projects that ACAP has implemented with active participation of local people, are now virtually ceased. Institutional strengthening in buffer zones of lowland protected areas is also seriously curtailed. Of 29 grassroots institutions surveyed in two buffer zones of the Western Terai, 10 were not holding regular meetings. This has seriously hampered active participation by local people.

Another successful model of CBC is community forestry. Community forests have accomplished the objective of restoring denuded forests in mountain districts, and are taken as an exemplary model for participatory conservation in developing countries. To date, some 12,000 registered Forest User Groups (FUGs) are managing 850,000 ha of forests under a community forests regime in Nepal (Gilmour 2003). The grassroots institution founded on FUGs has been facing many challenges. Most community forests are in hill and mountain districts that are under Maoists control and FUGs must tacitly agree with Maoists rules that are not conservation-friendly. Local press reported that the

Maoist rebels were extorting up to 70% of the revenue generated by community FUGs in those areas. In some instances, rebels infiltrated FUG executive committees and earmarked revenues to support the insurrection. No matter how rebels acquired funds garnered by FUGs, the development fund – a byproduct of the conservation program – is misused for warfare. This has severe impacts on sustainable development of rural areas and survival of FUGs as a viable institution for resource management. There are legal

101 complications when FUGs are dysfunctional. When civil strife challenges the legitimacy

of the Government in rural areas, then present FUGs will become disenfranchised as

legitimate forest managers because they function in collaboration with, and under the

umbrella of, district forest offices (Gilmour 2003). This is a serious problem for

community forestry programs in mountain districts because most are under rebel control.

Community forestry was more successful in the mountain districts of Nepal than in other

areas. No research has been done to date on the impacts of civil strife on community

forestry programs and the functioning of FUGs in Maoist strongholds.

The war pervades all spheres of society and seriously disintegrates the capacity of local

institutions for natural resource management (Goldstone 1996). There have been very few studies on how local institutions thrive in times of civil strife. The pervasive hypothesis is ‘popular participation’ by local people, but research elsewhere shows that local communities are not willing to participate in war-inflicted areas (Davalos 2001). If this is true in all or most war-inflicted areas, the efficacy of the CBC approach will be seriously undermined.

1.4 Wildlife Populations

Some authors suggest that historical warfare has been beneficial for conservation based on the findings that game was abundant in buffer or war zones (Martin & Szuter 1998).

However, there is no dispute in the claim that modern wars and civil strife are detrimental to endangered megafauna. During periods of war, there have been massive declines in populations of elephants and large ungulates in Uganda (Eltringham & Malpas 1993),

102 extirpation of wild ungulate and carnivore populations in Afghanistan (Formoli 1995), and poaching of bonobo Pan paniscus and gorilla Gorilla gorilla in the Republic of

Congo (Vogel 2000). Wildlife populations plummeted directly from opportunistic, deliberate and random shooting by rebels, security forces, and poachers, and indirectly from landmines.

Endangered species such as red panda Ailurus fulgens, snow leopard, musk deer, one- horned rhinoceros, Bengal tiger, Cheer Pheasant Catreus wallichii are falling prey to poachers and rebels in Nepal. There has been a precipitous decline in the blue sheep

Pseudois nayaur population from 2200 individuals in 2002 to 563 individuals in 2004 in the DHR (Gautam 2004). Maoist rebels are shooting these animals to feed their cadres.

Illegal activities of poachers are escalating in the reserve: 53 and 33 traps were recovered in 2002 and 2003 respectively, and one red panda was caught in a trap (Tripathy 2003).

The recent census of one-horned rhinoceros shows a bleak future for these animals. The population of rhinoceros decreased from 544 in 2000 to 372 in 2005 in and around the

RCNP, home to most of Nepal’s rhinoceros (Mainali 2005). The death toll of rhinoceros is steadily increasing, 33 rhinoceros died in 2000, 42 in 2001 and 55 in 2002, with an average mortality rate of 6.32% per year for the period of 2000-05 (Chapagain 2002). At least 94 rhinoceros were killed by poachers since 2000. Tigers are also meeting the same fate: six tigers were killed in 2002 while eight tigers were killed in 2003 (The Rising

Nepal 2003). Escalating of poaching inside protected areas is attributed to lax security and inefficiency of anti-poaching units (APU). The collaborative effort of the park,

NGOs and local people has created APUs in RCNP and RBNP to garner conservation

103 intelligence. Despite hefty donor funding, APUs failed to curb poaching in these areas

(Mainali 2005). At present, wildlife poaching is a serious threat. So far, there have been

no records of wildlife killed by landmines in Nepal.

In the past, elites who had licensed arms hunted wildlife in public forests whenever an

opportunity arose. This hunting seriously depleted wildlife populations outside protected

areas. However, the situation changed when the Maoists usurped licensed arms.

Furthermore, the Government ordered that all licensed arms were to be submitted to

security forces when the plundering of arms escalated. Since hunters lost arms, the

frequency of hunting decreased sharply. This had a positive impact on the resurgence of

common species such as barking deer, common leopard, and pheasant in public forests,

which were once pushed towards local extinction. However, conservationists are

skeptical about population increase of endangered species in rebel-controlled areas.

Scientific data on population trends of wildlife remains obscure. It is not feasible to

undertake scientific research in wildlife habitats that are under guerrilla control (Davalos

2001).

1.5 Law and Regulations

Nepal is one of the few countries in the world to deploy soldiers (Royal Nepalese Army) for the protection of parks and reserves (Budhathoki 2003). They have been effective in providing able protection of these areas (Heinen & Kattel 1992), but are inadequate to maintain law and order in times of political instability. Security of protected areas receives a lower priority since the Government declared a state of emergency in 2001 to

104 curb the Maoist insurrection. Since then, there has been a 70% reduction in guard posts inside protected areas. Of 112 guard posts, only 34 are now occupied to render protection of parks throughout Nepal (Nepali 2005). Many soldiers were deployed to fight against the insurgents rather than poachers. The frequency of patrolling has been reduced in temporal and spatial scales within protected areas. This has emboldened poachers, smugglers and illegal trespassers to abuse protected areas. The withdrawal of army posts has resulted in the widespread breakdown of law and order in and around protected areas.

Nepal is party to many international conservation accords such as the Convention on

International Trade in Endangered Species of Wild Fauna and Flora (CITES), the

Convention on Biodiversity, the World Heritage Convention, and the Ramsar

Convention. Implementation of and compliance with these treaties have been hindered due to a lack of national implementing legislation (Heinen & Chapagain 2002), and in recent years, instances of violations of these treaties have increased because of civil strife. International poachers consider Nepal a safe haven for illegal trade of wildlife parts because the Maoist insurgency paralyzed law-regulating agencies. Kathmandu has became a hub where bones, skins, furs and other body parts of endangered wildlife are traded. Although Nepal is not a major consumer of wildlife parts, poachers are abusing its territory as a transit point for illegal trade with China and India (The Rising Nepal 2004).

On April 23, 2004, police personnel disguised as consignors arrested an illicit trader with

85 pieces of leopard skins and 38 of that of otter in the capital, Kathmandu (Khatry-

Chhetri 2004). He bought these skins in India and intended to export to the China. A huge cache of wildlife parts was recovered on March 29, 2004 near the Nepal-Tibet border.

105 Security personnel seized 172 pieces of rhinoceros skins, seven tiger skins, six skins of unidentified cats and 165 pieces of tiger bone (The Kathmandu Post 2004). Similarly, security personnel seized 32 tiger, 579 leopard and 666 otter skins en route to Tibet in

October 2003 (Khatry-Chhetri 2004). These species are listed in CITES Appendices so are trade-regulated, but their skins or bones fetch hefty price in illicit international markets. Therefore, the illegal trade in wildlife parts is threatening the survival of many endangered species. The recent survey in Sariska Tiger Reserve in India found that not a single tiger was left alive there due to poaching (Phuyal 2005). These activities violate

CITES and national wildlife laws of Nepal, India, and China. Better coordination and cooperation between the three countries will facilitate the enforcement of national laws and CITES. When the host country is grappling with civil strife, the enforcement of international conservation accords and national conservation laws is ineffective.

It is common to include biodiversity and conservation rhetoric in the political agenda by rebels (Alvarez 2003). So far, the Maoists in Nepal have not proclaimed their environmental policies fully. Their environmental policies are guided by short-term benefits accrued from the sale of valuable resources and the protection of forest cover for hideouts. Protecting forests serves their interests because forests are indispensable hideouts in guerilla wars (Davalos 2001). However, this does not reflect any incorporation of sound environmental policy in their political agenda. In the high

Himalayas, the Maoist rebels are facilitating the sale of trade-regulated aromatic and medicinal plants to collect revenue. They are issuing permits to collect Yarsa Gompa

Cordyceps sinensis, Panch Awale Dactylorhiza hatagirea, and Loth Salla Taxus baccata

106 (N. Lama, ACAP officer, personal communication). Over-harvesting of these species is

pushing them towards local extinction. For monetary benefits, Maoists are colluding with

poachers and smugglers in illicit trade of endangered animals and plants. Illegal activities

are common in protected areas and public forests in absence of authorities. Although

there are examples of rebel groups helping to conserve forests (Kaimowitz & Faune

2003), there has been no obvious conservation benefits of the Maoist People’s War in spite of this rhetoric of forest conservation (Heinen & Shrestha 2005).

1.6 Some Economic and Social Impacts

Revenue generated in PAs for the years 2000/01, 2001/02 and 2002/03 were 1796, 912, and 802 thousands US dollars, respectively (DNPWC 2005). There has been a 40-60% decline in park entry fees in that period (Nepali 2005). Mountain parks such as Langtang,

Everest, and SheyPhoksundo are famous for trekking and Terai parks such as RCNP and

RBNP are renowned for wildlife safaris. Since there has been a tremendous decrease in tourist arrival at the national level (Figure 10), these parks also lost substantial revenue.

Trophy hunting was the main source of revenue generation in DHR and mountaineering was in MBNP, but currently these areas are under rebel control so do not contribute to the central treasury. Considering the volatile nature of the tourism market, it is suggested that

income from ecotourism should be complementary, but not substitutive for conservation

(Wunder 2000). The government used to collect substantial revenue through the sale of

timber and non-timber forest products from productive forests of Terai. When Maoists

rebels took over most of these forests, there was a sharp decline in these revenues as well.

The conservation agency is chronically under-funded (Heinen & Kattel 1992), and when

107 revenue generated by parks and productive forests plunged further, the allocation of

government funds for conservation greatly decreased. One of the strategies used by rebels

worldwide is taking control over revenue-generating sources so that the state financially

bankrupts (Davalos 2001). Following this strategy, the Maoist rebels in Nepal have been

generating revenues from the use of natural resources to partially fund the insurgency at a loss for conservation.

War brings about many social problems that ultimately impact natural resource conservation and management. In the early 1990s, Bhutanese refugees were fleeing to the eastern Terai of Nepal, and after 1996, many people who were displaced from mountain districts by the Maoist People’s War migrated to the Western Terai (Heinen & Shrestha

2005). These large-scale influxes of immigrants increased pressure on resources and hasten exploitation. Research elsewhere shows that plundering resources by rebels compel local people to harvest more (Plumptre et al. 1997). Local people are highly discounting the total value of biodiversity due to an uncertain future created by civil strife. Taking advantage of the disorder and confusion, they are exploiting endangered plants and animals for either subsistence or commercial purposes. Through out the world, natural resources are abused by guerrillas, military forces, local people and refugees during and subsequent to periods of war and civil strife (Dudley et al. 2002). Religious faith may garner support for conservation. Sacred forests thrived in Uganda during civil unrest (Sembajjwe 1995). In many places, people have favorable attitudes towards wildlife even though they cause substantial crop and livestock damages (Sekhar 1998).

However, religious faith and cultural harmony slowly deteriorate when civil strife creates

108 political chaos. In the name of social reform, the Maoists of Nepal outlawed many religious and cultural practices that were conservation friendly and previously checked over-exploitation of resources (Karki & Bhattarai 2004).

1.7 Synthesis and Conclusion

A decade of civil strife is beginning to have far-reaching environmental consequences in

Nepal. Damages to physical facilities, take over of protected areas and forests, and killing of staff by Maoist rebels are seriously hampering the stability of the conservation sector.

In the absence of stringent law enforcement, poachers are taking a large toll on many endangered species. The future of one-horned rhinoceros and Bengal tiger is bleak in

Nepal if the current level of poaching remains. Escalating wildlife trade and the

Government’s negligence of the conservation sector violates international conservation accords and results in ineffective implementation of national wildlife laws. The country’s dependence on tourism to fund conservation programs also has deleterious impacts on conservation. The conservation agency has been chronically under funded, and the situation exacerbated when the volatile tourism market suffered from the insurgency. In addition, the Maoist rebels are abusing natural resources to support their insurgency.

Most conservation legislation, rules, regulations and modalities are crafted in periods of political stability. One of the lessons learned from on-going civil strife is that legislation and models are inadequate to address the multifaceted issues of conservation in times of political instability. The deployment of soldiers in protected areas during civil strife has not rendered able protection in Nepal because they were withdrawn from parks and

109 reserves to fight the Maoist rebels. The time is right to contemplate the role of Royal

Nepalese Army in parks protection and explore alternative institutional mechanisms that could be more effective to provide security to parks and reserves.

The CBC approach has been, in general, more resilient than the ‘fortress and fine’ approach for conservation, but is not impervious to civil strife. Rebels are extorting funds from community forest user groups, kill local conservation leaders and bomb NGO offices that support local people. In spite of these adversaries, many grassroots institutions have thrived in civil strife because local people have favorable attitudes towards the CBC approach and extend their support. So far, there is no example of third party involvement in conservation by collaborating with both the rebels and the

Government. The anti-imperialist policy (ban on NGOs and INGOs) of the rebels makes it hard for the involvement of the international conservation community to support conservation.

The fate of Nepalese conservation efforts depends on the duration of this uncertain period and its consequences on the political, social, and economic sectors. The problem is intricate and there are no quick fixes. Although it has had recent negative impacts on conservation efforts achieved over the past three decades, it also provides an opportunity to explore more robust conservation models. Like biodiversity, civil wars and political instability are common in developing countries. To address the issue, strategies should be explored at local levels with national and international support.

110 Appendix 2. A Sample of Household Questionnaire Survey Form

Your participation in this survey is voluntary. You will neither get any direct monetary benefits for participating nor penalized for not answering some or all of the questions. Any information gathered in this survey will be only used for the purposes of research. The interview is completely confidential; your name will not be associated with your answers. The purpose of this survey is to evaluate resource use patterns and conservation attitudes among local people in the Western Terai Landscape, Nepal. Your cooperation will help policy makers and planners make informed decisions.

Name of interviewer: …………………. Date survey: …………………….

Household identification number: ……. Village name: ……………………

Respondent’s Ideographic Data: Gender: ------Male ------Female

Age: ------Ethnicity (Caste): ------

Education: ------Occupation: ------

Information about Household Members: Please, tell about gender, age, education and occupation of your household members S. No. Household members Gender Age Education Occupation 1 2 3 4 5 6 7 8

Migration: Have you migrated to this place from elsewhere? Yes: ------No: ------

From where have you migrated to this place?

A district in mountain ------A district in the Terai ------

Other villages of the same district ------Others ------

When did you migrate here? ------years ago.

What was the reason to migrate here?

111 Landlessness ------Insufficient land ------

Unemployment ------Under Government’s Scheme ------

Others ------

Landholding Size and Tenure Status: Do you have plots for farming? Yes/size ------No ------

What is the status of land?

Private registered land ------Pubic land without title ------

Feudal land ------Others ------

Do you get enough agricultural products to support your family for whole year? Yes ------No ------

If no, how many months do you get shortage of food, not fulfilled from your agricultural products? ------Months.

What are alternate sources of income to fulfill your requirements?

Government employment ------Pension ------

Business ------Paid labor ------

Remittances ------Others (Specify) ------

Livestock Holding: Do you own any livestock? Yes ------No ------

What kind of livestock species and how many of them do you have? S. No. Species Number Local breed Improved breed 1 Cattle 2 Buffalo 3 Calves 4 Goat/Sheep 5 Hens/Ducks 6 Pigs 7 Others (Specify)

Resource Uses: What kind of materials do you bring from the National Park or National Forest? If you bring any thing, please give an estimate of resources harvested.

112 S. No Material Types Quantity in local units 1 Fodder 2 Fuelwood 3 Thatch grass 4 Leaf litter 5 Medicinal herbs 6 Edible items (foods) 7 Timber 8 Others (Specify)

From where you bring fodder for your livestock?

Own farms ------Buffer zone, protected or national forest ------

How many days per week do you go to forests to bring fodder? ------days.

About how may fodder tress do you have on your private land? ------

Where do you take your livestock for grazing?

Forests ------Private farms ------

Stall-fed ------Public grazing lawns

What kind of cooking stove do you use in your house?

Simple mud stove ------Improved mud stove ------

Kerosene stove ------Bio-gas ------

From where do you get fuel wood?

Private farm ------Other forests ------

From market ------Drift wood ------

Others ------

Do you have problems to get enough fuelwood? Yes ------No ------

If yes, what do you think could be the best way to solve the problem?

Free access to existing forest ------Community forests ------Private plantation ------Access to improved stove ------

113 Subsidized kerosene depot ------Others ------

Any sales or purchases of forest products in last 12 months? Forest product sold Amount sold Sale price/unit Where sold?

Forest product bought Amount bought Buying price/unit Where bought?

Conservation Attitudes: Different people who live in this area hold very different opinions about the park or program. Here are a few of the things that people say about the park or program. Will you tell me whether you agree or disagree with them. S. No. Statements Agree Disagree 1 Forests around your village have decreased in recent years. 2 It is responsibility of local people to protect natural resources. 3 If there is unlimited access to forests for fuel wood and fodder, forests will be disappeared soon. 4 There are more wild animals now than ten years ago. 5 What people and their livestock need are more important than saving plants and wild animals. 6 My living condition improved since the protected area’s creation. 7 After the establishment of buffer zone forests/reserve you don’t have problem of access to resources. 8 It is important to set aside a place for the animals and plants to live in. 9 It is important to protect the animals and plants so that our children may know and use them. 10 There is an equitable distribution of common pool resources and benefits. 11 You are willing to contribute for conservation cause.

Participation and Benefits: Are you or any of your family members elected in any grassroots organizations? No ------Yes ------(Mention the Name ------)

Have you or your family members ever received any kind of trainings? No ------Yes ------(Which? ------)

114 Are you benefited from any conservation organizations? (Benefits include free seedlings, vaccination, biogas/toilet construction support etc.) No ------Yes ------(Mention the items ------)

Are you a member of saving-credit group? No ------Yes ------

Are you engaged in income generating activities promoted by NGOs? No ------Yes ------(If yes how much do you earn per month? ------NRs.)

Miscellaneous: Are you satisfied with User Groups? Yes ------No ------

Are you suffering from wildlife damage? Yes ------No ------

Do you have any suggestions for conservation and sustainable development?

115 FLORIDA INTERNATIONAL UNIVERSITY

Miami, Florida

ANALYSIS OF FORESTS UNDER DIFFERENT MANAGEMENT REGIMES IN THE

WESTERN TERAI OF NEPAL AND ITS RELATION TO ENVIRONMENT AND

HUMAN USE

A thesis submitted in partial fulfillment of the

requirements for the degree of

MASTER OF SCIENCE

in

ENVIRONMENTAL STUDIES

by

Nilesh Timilsina

2005

To: Interim Dean Mark Szuchman College of Arts and Sciences

This thesis, written by Nilesh Timilsina, and entitled Analysis of forests under Different Management Regimes in the Western Terai of Nepal and its Relation to Environment and Human Use, having been approved in respect to style and intellectual content, is referred to you for judgment.

We have read this thesis and recommend that it be approved.

Michael E. McClain

Michael S. Ross

Joel T. Heinen, Major Professor

Date of Defense: July 6, 2005

The thesis of Nilesh Timilsina is approved

Interim Dean Mark Szuchman College of Arts and Sciences

Dean Douglas Wartzok University Graduate School

Florida International University, 2005

ii

DEDICATION

I dedicate this thesis to my parents for everything they did to bring me to this level.

iii ACKNOWLEDGMENTS

First, I would like to express my sincere gratitude to my Major advisor Dr Joel T.

Heinen for his help throughout the work. Without his help this work wouldn’t have been possible. Thank you, Dr Heinen for all the help, support and encouragement you have provided me. I have also highly benefited from the experience and knowledge of my committee member Dr. Michael S. Ross. I am grateful for his advices and suggestions in conceptualizing the work and as well as during data analysis. I would also like to thank my other committee member Dr. Michael E. McClain for his time and support. I acknowledge my friend Nabin Baral for all his help and suggestions during the last ten years of our friendship. Financial support as graduate assistanships from FIU Department of Environmental Studies has been valuable. I also thank all my Professors, colleagues and staff at Department of Environmental studies for their warm support and friendship.

I would like to thank National Fish and Wildlife Foundation’s Save the Tiger

Fund and Disney Wildlife Conservation Fund for supporting international conservation works through financial support for this study. I am thankful to Institute of Asian Studies at FIU for giving financial support. I can’t remain without thanking His Majesty’s

Government of Nepal, Department of National Parks and Wildlife Conservation,

Kathmandu office, Bardia office and Shukla office and the staff for giving permission and other help to conduct the field work. I have to thank King Mahendra Trust For

Nature Conservation (KMTNC) project offices both in Bardia and Shuklaphanta for providing a place to stay and other logistic support during the field work. I acknowledge the help of Dr. Shant Raj Jnawali at KMTNC, Kathmandu, Naresh Subedi, Sri Ram

Ghimire, Ramesh Singh and other staff at KMTNC Bardia and Chiranjeevi Pokharel,

iv Suman, Achyut, Parmananda, Bintiram, Shankar, Birchu and other staff at KMTNC

Shukla office. Thank you, Hukum Shahi for good food and warm hospitality during our stay in Bardia. I am thankful to Birendra Tiwari, Thaneswor Tiwari and their family,

Ramesh Bhatta, Benu Gautam, Sher Bahadur and Ram Bilas for their help.

I also thank Dr. R. P. Chaudhary and other people at the Department of Botany,

Tribhuvan University for their help in plant identification. I would like to thank the

National Herbarium and Plant Laboratory at Godavari for plant identification, the

Department of Agriculture, Soil laboratory, Lalitpur for soil analysis and the Department of Forests and its district offices for providing literature and other help. Thanks to everyone whom I forgot to mention and who have helped directly and indirectly for completion of this work.

Thanks are due to Dr J. P. Sah and Susana Stoffella at South East Environmental

Research Center (SERC) for help during data analysis. I would like to thank the

Department of Statistics at FIU for statistical consulting. Help from my field assistants

Janak and Bhatta were invaluable.

Finally, I am expressing my heartfelt gratitude to my brother’s family and my wife for their love, inspiration, support and encouragement.

v ABSTRACT OF THE THESIS

ANALYSIS OF FORESTS UNDER DIFFERENT MANAGEMENT REGIMES IN THE

WESTERN TERAI OF NEPAL AND ITS RELATION TO ENVIRONMENT AND

HUMAN USE

by

Nilesh Timilsina

Florida International University, 2005

Miami, Florida

Professor Joel T. Heinen, Major Professor

This study was done to understand the forest structure, composition and dynamics of the Sal forest, the relationships of forest communities with environmental variables, and ecological differences between community forests and forests inside protected areas in the western Terai of Nepal. Forest sampling was done along transects in two protected areas and two community forests, and sampling locations were established every 200 m to sample trees, saplings, shrubs, seedlings and herbs. Soil samples from each plot were analyzed. Agglomerative cluster analysis, non-metric multidimensional scaling,

ANOVA, Kruskal-Wallis, Mann-Whitney and t-tests were all used to analyze data.

The sampled forest had lower tree diversity, density and lower basal area compared to forests in other areas of India and Nepal. Three different associations of Sal

Forest were identified, but none of the soil variables tested identified the distribution of communities. Community forests were in poorer conditions compared to protected forests and require additional protection to resemble the structure and diversity of protected areas.

vi TABLE OF CONTENTS

CHAPTER PAGE

1. INTRODUCTION 1 1.1 Background 1 1.2 Topography and Physiography 3 1.3 Vegetation Zones of Nepal 4 1.4 Legal Protection of Forests in Nepal 6 1.5 Protected Areas Management and Landscape Approach to Conservation 10 1.6 Forest and Environment Relationship 12 1.7 Forests in the Western Terai 15 1.8 Objectives of the study 18

2. THE STUDY AREA 19 2.1 Royal Bardia National Park 20 2.2 Kanchanpur District 22 2.3 Royal Suklaphanta Wildlife Reserve 23 2.4 Community Forests 25

3. METHODS 25 3.1 Forest Sampling 25 3.2 Soil sampling and analysis 27 3.3 Data Analysis 28

4. RESULTS 32 4.1 Average forest structure and composition 32 4.2 Classification 34 4.3 Forest-environment relationships 38 4.4 Saplings and Shrubs 39 4.5 Canopy closure and Cover percent 41 4.6 Ground vegetation 42 4.6.1 Seedlings 42 4.6.2 Herb layer 44 4.7 Stand dynamics 45 4.8 Comparison between management regimes 47

5. DISCUSSION 50

REFERENCES 91

APPENDICES 101

vii LIST OF TABLES

TABLE PAGE

1 Area (ha) under different land uses in Bardia District 66

2 Importance value of trees for the three different associations identified by cluster analysis. 66

3 Basal area (m2 ha-1) of the different tree species among groups. 67

4 Mean, minimum and maximum values for soil variables tested for three different associations. 67

5 Mean sapling density per ha, Species richness per plot (S), Shannon’s Diversity Index (H’) and Evenness (E) for all forests sampled and the three different associations identified by cluster analysis. 68

6 Mean sapling density of different species per ha in different groups. 68

7 Kruskal-Wallis and Mann-Whitney results showing the difference in sapling density of different species among groups. 69

8 Mean shrub density per ha, Species richness per plot (S), Shannon’s diversity index (H’) and evenness (E) in all forests sampled and in different groups. 69

9 Comparison between groups in terms of shrub density, Shannon’s diversity and species richness. 70

10 Mean, minimum and maximum cover percentages for overstory, understory and ground cover of three groups identified by clusteranalysis 70

11 Comparison of cover percentages between the groups identified by cluster analysis. 71

12 Status of regeneration of different species in the three Groups. 71

13 Species richness, Evenness, Shannon’s diversity index and Density/ha of herbs for the three associations and all sampled forests. 72

viii LIST OF FIGURES

FIGURE PAGE

1a Average monthly precipitation and average annual precipitation for Bardia District (RBNP) and Kanchanpur District (RSWR). Data were taken for fifteen Years, 1987 to 2001. 73

1b Mean monthly maximum temp (0C) and mean monthly minimum temp (0C) for Bardia District (RBNP) and Kanchanpur District (RSWR). Data were taken for fifteen years, from 1987 to 2001. 73

2 Map of the study area 74

3 Diameter distribution of trees (> 5 cm dbh) for RBNP, RSWR and the Community Forests sampled. 75

4 Dendrogram showing the different associations identified by the hierarchial agglomerative cluster analysis based on importance value of trees. Groups are described in the text. 76

5 Diameter distribution of trees (> 5 cm dbh) for Shorea robusta- Buchanania latifolia association (Group 1). 77

6 Diameter distribution of trees (> 5 cm dbh) for Terminalia tomentosa- Shorea robusta association (Group 2) 78

7 Diameter distribution of trees (> 5 cm dbh) for Shorea robusta- Cleistocalyx operculatus association (Group 3). 79

8 Site scores from 2 axis non-metric multidimensional scaling (NMS) ordination based on importance value of trees. 80

9a Relationship between different associations identified by cluster analysis and soil pH. Contour plot of pH is superimposed upon NMS ordination of trees. 81

9b Relationship between different associations identified by cluster analysis and soil organic matter. Contour plot of soil organic matter is superimposed upon NMS ordination of trees. 82

9c Relationship between different associations identified by cluster analysis and soil total nitrogen. Contour plot of total nitrogen is superimposed upon NMS ordination of trees. 83

ix 9d Relationship between different associations identified by cluster analysis and available phosphorous. Contour plot of available phosphorous is superimposed upon NMS ordination of trees. 84

9e Relationship between different associations identified by cluster analysis and potassium. Contour plot of available potassium is superimposed upon NMS ordination of trees. 85

9f Relationship between different associations identified by cluster analysis and soil texture. Contour plot of soil texture is superimposed upon NMS ordination of trees. 86

10 Seedlings of tree taxa present in each group as the percentage of total tree taxa recorded in seedlings. 87

11 Log density of trees (stems ha-1) plotted against log Average Stand Diameter (ASD) for all the plots sampled. Plots are labeled by Stand Density Index. The diagonal line asserts the maximum stocking for the forest sampled. 88

12 Plot of total seedling density against SDI and ASD. 89

13 Scatter plots of Sal seedling density vs Sal tree relative density for Sal Forest in the western Terai of Nepal. 90

x ACRONYMS

TAL Terai Arc Landscape

RBNP Royal Bardia National Park

RSWR Royal Suklaphanta Wildlife Reserve

HMGN His Majesty’s Government of Nepal

ASD Average Stand Diameter

SDI Stand Density Index

IVI Importance Value Index

xi 1. INTRODUCTION

1.1 Background

Forests covered 73.3% of the total area of the Terai (the subtropical lowlands) between central and western Nepal during the 1950s (Joshi 2002). Due to the importance of these forests for both commercial and subsistence purposes (Webb and Sah 2003), in the past few decades, heavy human pressures have reduced the forested area resulting in degradation and fragmentation of historically contiguous landscapes and posing threats to biodiversity conservation and local livelihoods. Between 1958 and 1988, forests cover in the lowland Terai between central and western Nepal declined from 73.3% to 45.8% of the total land area (Joshi 2002). In such a human dominated environment, it is necessary to have baseline ecological information on forests for their management in the future.

A major part of these subtropical areas are covered by seasonal broad leaved forest which are typical of monsoonal climates (Wesche 1997). Sal (Shorea robusta) is the single most important species. Sal Forests are important for both commercial purposes and local livelihoods. Its sustainable management and utilization is necessary to meet the broad objectives of biodiversity conservation and sustenance of the rural economy.

Despite widespread occurrence of Sal Forests and its importance both from economic and ecological points of view, little information exists on the eological aspects of this forest. Past studies on the forests and flora of Nepal (Stainton 1972; Dobremez

1976), a few floristic studies inside protected areas (Dinerstein 1979; Shrestha and Jha

1997; Sharma 1999) and a study in central Nepal (Wesche 1997) provide information on

Sal Forest, but there is a dearth of information in Sal Forests of the western Terai, and

1 especially outside protected areas. The present study is an attempt to fill this gap and provide important information on the structure, composition and dynamics of Sal Forests in protected areas and community forests of the western Terai.

Physical factors such as climate and rainfall, soil physical and chemical properties and existing disturbances play a significant role in the distribution and diversity of plant communities (Tilman 1982; Kozlowski et al. 1991; Swaine and Becker 1999).

Recognition of soil heterogeneity is important for analysis of plant community patterns in tropical forests (Huston 1980; Villers-Ruiz et al. 2003). In the western Terai of Nepal, the length of monsoon, total rainfall, seasonal flooding and soil conditions, and other factors such as grazing, clearing for cultivation, burning, selective cutting, logging and lopping have been considered as factors modifying vegetational composition and succession (Dinerstein 1979). In the present study, I also make an attempt to understand the relationship of forest communities with physical environmental factors.

Forests exist under different conditions and are managed with different objectives.

Forests inside protected areas are managed with protection as the main objective and maintained for environmental services, biodiversity conservation, to provide habitat for wildlife and promoting tourism. Community forests are managed by communities with the goal of sustainable utilization of forests resources such as timber, fodder, firewood and other non timber forest products (NTFPs). Other forests (for example national forests) are open access (Ostrom 1990) and will face the tragedy of the commons (Hardin

1968). Several studies have found that forests managed under different objectives show significant differences in ecological conditions (Shankar et al. 1998; Nagendra 2002;

Web and Sah 2003). Structure and composition of natural forests differ from secondary

2 forests and plantation forests. Forests inside protected areas differ from community forests and national forests.

Protected areas and forests outside protected areas in the western Terai are part of the Terai Arc Landscape (TAL), which is a government initiative to take landscape approach to protect biodiversity in protected areas and outside forests and to link protected areas with forest corridors (HMGN 2004). Since the protected areas in the western Terai are under strict protection for nearly three decades, a comparative study between the two provides important ecological information on the outside forests. The information will be useful in assessing the existing biodiversity and the habitat quality of outside forests, and management interventions required to bring them to the same richness and diversity level as protected areas. This study also compares protected area forests with community forests. For the success of TAL, information on ecological conditions of outside forests in comparison to protected areas will be useful for biodiversity conservation and to secure additional wildlife habitats.

1.2 Topography and Physiography

Here I want to provide general information on topography and physiography of the country to make readers aware of geographical settings of the study location.

Nepal is a Himalayan country situated on the southern slopes of the Himalayas. It is located between the latitudes 26o 22’ and 30 0 27’ N and longitudes 80o 40’ and 88o

12’ E. The shape of the country is roughly rectangular covering an area of 147, 181 sq km. Average east west length of the country is 885 km and average north-south width is

193 km. About 83% of the country is occupied by mountains and nearly 17% by the

3 lowland Terai. The altitude ranges from 60m above the sea level to 8848 m on Mt.

Everest, the highest peak in the world.

Nepal is classified into many different physiographic zones. The Terai and

Bhabar, which are the main focus of the present study, lie between 60-300 m elevation.

The Terai is the northern extension of alluvial gangetic plains and is highly productive in

terms of agriculture. The Bhabar abuts the Terai on the south and the Siwalik on the north and consists of large boulders that have been brought down by the rivers from the mountains to the north. The Churia or Siwalik hills rises to the north of the Bhabar and reach 1500 m in elevation. It extends from east to west. It is mainly composed of sedimentary rocks and big boulders. These areas have been subjected to severe soil erosion due to intensive removal of forest cover. Dun valleys are gently sloping valleys to the north of the Churia. The Mahabharat and Midlands range from 600 m to 3500 m.

The Mahabharat lies between the Churia in the south and midlands in the north. The elevation varies from 1500m to 2700m. The Midlands lie at the base of the Himalayas and north of the Mahabharat range and cover most of the central region of the country.

Elevations range from 600 m to 3500 m with an average of 2000m. The Himalayas lie in the northern part of the country and extend from east to west. They consist of the major peaks of the world and are covered with snow year-around over 5500m in elevation.

Between the greater Himalayas lie several inner Himalayas with a dry and monsoon-less climate.

1.3 Vegetation Zones of Nepal

Since this study deals with forest communities, this section provides an introduction to the past studies on flora and classification of vegetation zones to make

4 readers aware of the existing information on the flora as well as the plant communities that can be found in the country.

The study of vegetation in Nepal was started by Buchanan and Hamilton in 1802 and later continued by Nathaniel and Wallich in 1802-21 (HMGN 1976). Other studies on Nepalese flora were done by Schweinfurth (1957), Stearn (1960), Stainton (1972),

Dobremez (1972) and HMGN (1976). Stainton (1972) did a detailed classification of forest types in Nepal. For the description of vegetation he classified Nepal into Terai,

Dun Valleys and Outer Foothills, the Midlands (West, East, Central and Country to the south of Annapurna Himal), Humla-Jumla area, Dry River Valleys, Inner Valleys and the arid zones. He classified forests of Nepal into six divisions on the basis of ecology and vegetation, namely tropical and subtropical forest, temperate and alpine broad leaved forest, temperate and alpine conifer, minor temperate and alpine association. The factors that determine vegetation distribution are climatic conditions, altitude, geographical location, natural composition of the soil and biotic factors. The major zones are:

• Tropical Zone: This area lies between 200 to 1000m. It consists of the lowland

Terai and Bhabar zone. It is characterized by a hot climate and heavy monsoonal

rain. The major vegetation types of this region are Sal Forest, Tropical Deciduous

Forest and Tropical Evergreen Forest.

• Subtropical Zone: This zone lies between 1000- 2000m. There is no clear

distinction between tropical and subtropical zone in Nepal (Shrestha 1997), but

for convenience these two groups are kept separately. This region encompasses

the Siwaliks, lower Mahabharat ranges, midland areas to 2000 m. This region

5 consists of mixed tropical evergreen and broad-leaved forest. At high elevations

one may find mixed hard wood forest.

• Temperate Zone: This zone lies at an altitude between 2000 to 3000 m. It

includes the southern slopes of the Himalaya and higher elevations of the

Mahabharat range. The forests types consist primarily of temperate mixed

broadleaved and evergreen forest, and upper temperate mixed broadleaved forest.

• Subalpine Zone: This is a transitional zone between the temperate and alpine

zones and consists of part of the greater himalyas between 3000 to 4100 m. Silver

fir and Rhododendron forests are found in this zone. The treeline in western

Nepal is 3850 m and for eastern Nepal it is 4000 m (Chaudhary 1998).

• Alpine Zone: This is the zone above 4100m. It is characterized by strong winds,

cold and snow. Vegetation comprises the association of Juniper-Rhododendron

and alpine meadows.

1.4 Legal Protection of Forests in Nepal

According to Nepal’s Forest Act of 1993, forest has been defined as an area which is completely or partially covered by tree species. In Nepal forest covers approximately 5.6 million hectares, which is about 37% (including forests and shrub) of the total land area of the country (15 million hectares). Land use classes other than forest include agricultural land, eroded land, water, stream, river bed, flooded areas, urban and industrial area, grassland, barren land, snow and ice. Forest resources play a crucial role in socio-economic upliftment of the people of Nepal. About 15% of the country’s GDP comes from forest resources (HMGN 1996) and more than 75% of total energy used is derived from fuelwood from forests and shrublands (HMGN 1998).

6 Forests were managed traditionally until the mid 1950s in the hilly regions of

Nepal (Thapa and Weber 1995). The Forest Nationalization Act of 1957 brought all forested land under government ownership and alienated local communities (Neupane

2000). This resulted in the conversion of limited-access community-controlled forests into open access resources (Ostrom 1990). The National Forest Act of 1976 attempted to return ownership to communities to a certain extent but was unsuccessful largely due to administratively-defined government structures such as village Panchayats (equivalent to parish; Thapa and Weber 1995). Rather than true community involvement, Panchayat representatives and officials had influence in decision making. Realizing the need for community participation in forest management, the Government of Nepal introduced the

Community Forestry Act in 1993 (Varughese 2000). The major thrust of the act was to provide communities with the rights to protect and manage forests, to utilize forest products and to derive income from forests. Over 8500 forest user groups had been formed and about 620,000 ha of forest area had been handed over to user groups by 1999

(Chaudhary 2000). Community forestry has been successful in improving the conditions of forests and people in the midhills, but increased inequalities in the distribution of agrarian resources and greater ethnic diversity because of migration from the hills have been attributed as causes of infeasibility of community forestry in the Terai region

(Chakraborty 2001).

According to the 1993 Forest Act, five categories of forests are identified in Nepal.

These are Government Managed Forest, Protected Forest, Community Forest, Leasehold

Forest and Religious forest.

7 • Government Managed Forest: The forest type is strictly managed by His

Majesty’s Government of Nepal (HMGN) with production as the main objective.

It is illegal to collect any forest resources from this category without permission

of authorized person with HMGN. A certain amount of money should be paid to

HMGN for resource utilized.

• Protected Forest: Under the act, government forests with any cultural,

environmental or scientific importance are declared by HMGN as protected

forest.

• Community Forests: National Forests that are handed over to community “users

groups” for their conservation, management and utilization are community

forests. The major goal of this policy is to initiate community participation in

forest management. Forests are handed over to community user groups, who are

granted the right to manage and protect forests and the right to forest produce and

income derived therefrom.

• Leasehold Forests: These are National Forests handed over to institutions,

industries or communities established under the current law. The main objective

is to provide raw materials needed for forest industries, and to encourage

plantation forestry, ecotourism and agroforestry.

• Religious Forests: These are national forests in and around religious sites that

are handed over to religious groups for their conservation, utilization and

development. Religious groups can utilize resources for religious causes but not

for commercial use.

8 According to the Act, the term national forest includes all forests, excluding

private forests whether the boundaries are delineated or not; it also includes waste or

unregistered or uncultivated lands in or around forests as well as paths, ponds, lakes,

streams or rivers and riverine lands within forests. The term Private Forest denotes any

forested land that is planted, nurtured or conserved on private land owned by an

individual under the current law.

Realizing the need to manage forests sustainably for the long term fulfillment of

local needs, the Government of Nepal endorsed The Master Plan for the Forestry Sector in 1989. The plan, which was prepared in 1988 (HMGN 1988), presents up to date strategies for the management of forests in Nepal for the next 21 years. The primary goal of the Plan was to foster community and private participation and partnership with the

Ministry of Forest and Soil Conservation for the management and sustainable utilization of trees, shrubs, grasses and medicinal plants. Long term strategies for the management of the forestry sector according to the plan were to: meet people’s basic need; increase agricultural production through forest management; protect against land degradation (soil erosion, landslides, desertification and flooding); provide economic upliftment of both local and the national economy; and conserve ecosystems and genetic resources. There were management plans in the early sixties but they were not implemented because of lack of resources and government initiative. Operational Forest Management plans for

18 districts of the Terai were prepared in the late 1980s but were also not implemented

(Kanel and Shrestha 2001). Forest management has been provided great emphasis during planning and documentation phases, but implementation has been insubstantial in Nepal.

9 1.5 Protected Areas Management and Landscape Approach to Conservation

The National Parks and Wildlife Conservation Act was enforced in Nepal in

1973, which provided the legal base for the declaration, conservation and management of protected areas in Nepal. The main purposes of this act are to protect wildlife and its habitats, control hunting, and promote the conservation and management of important natural areas. The 1973 Act and its amendments identified five categories of protected areas in Nepal: strict nature reserves, national parks, wildlife reserves, hunting reserves and conservation areas. There are now nine national parks, three wildlife reserves, one hunting reserve and three conservation areas in the country, which cover approximately

18% of the total area (Heinen and Shrestha in press). Although 18% of the land area of

Nepal is protected, protected areas are not enough to conserve the full array of biodiversity and significant levels of biodiversity exists outside the system (Hunter and

Yonzon 1993).

Forests outside protected areas are being depleted. Between 1978 and 1991, about 99,000 ha of Sal (Shorea robusta) forest in the Terai were cleared with an average deforestation rate of 1.3% (8,300 ha per year; FRISP 1994). If Nepal were to lose its remaining humid tropical forest, there would be a loss of ten species of highly valuable timber, six species of fiber, six species of edible fruit trees and shrubs. This would severely alter habitat for 200 species of birds, 40 species of mammal and 30 species of reptiles and amphibians (HMG/IUCN 1998).

The advent of the concept of landscape ecology (Forman and Gordon 1986) has changed the way that managers, ecologists and conservation biologists think about the conservation of biodiversity. It introduced the idea of protecting whole landscapes rather

10 than protecting species or single ecosystems. It has been realized that protecting biodiversity inside reserves is not a panacea, and it is necessary to extend conservation efforts outside protected areas. Connectivity of historically contiguous landscapes should be maintained to provide more habitats for wildlife outside reserves.

To address the issue of protecting ecosystems and habitats outside reserves, the

Government of Nepal has embarked on a landscape level approach to conservation and endorsed the Terai Arc Landscape program (TAL) in April 2001 (HMGN 2004). TAL is the outcome of the government’s initiative, with support from World Wildlife Fund, to connect four lowland protected areas of Nepal with forests outside protected areas that act as movement corridors for the larger mammals such as tigers, elephants and rhinos.

The broader vision of TAL is to manage larger areas through participatory landscape planning based on the ecological, economic and social needs of the region.

TAL covers 49,500 square kilometers, encompassing 11 protected areas and forest corridors in India and Nepal, and extends from Nepal’s Bagmati River in the east to India’s Yamuna River in the west. The Nepalese portion of TAL is 23,199 sq km and covers 14 Terai districts. TAL is also important for its rich biodiversity. It provides habitat for fascinating megafauna such as the Royal Bengal Tiger (Panthera tigris), One- horned Rhinoceros (Rhinoceros unicornis) and Asian Elephant (Elephas maximus) as well as 11 species of ungulates and 550 species of birds (HMGN 2004). TAL covers

75% of the forests of the Terai and foothills of Churia. The population within TAL borders is 6.7 million, most of whom depend on forests for their livelihood. Sixty percent of the households in TAL rely on agriculture as their main source of income, 69% of households own livestock and depend on forests for fodder and 61% of households use

11 fuelwood for cooking (HMGN 2004). The present study encompasses two districts of

TAL and includes two protected areas, Royal Bardia National Park and Royal

Suklaphanta Wildlife Reserve, and two community forests in between. The outcome of the present study will provide valuable ecological information regarding forest management within TAL.

1.6 Forest and environment relationship

Forests are important natural resources both ecologically and economically. They are repositories of biodiversity and provide habitat for flora and fauna. Several ecosystem services such as nutrient recharges, rainfall, prevention of landslides and soil erosion, watershed services, and more are provided by forests. They are complex ecosystems (Lal 1992) because forests contain many more species per unit area than other ecosystems; are subjected to human disturbances such as fire and grazing; go through successional changes; show geographical variability; and interact and influence other systems such as rivers, lakes, pastures and agricultural land. This complexity makes it difficult to establish cause and effect relationships and predict the results of human intervention in forested ecosystems (Lal 1992).

Plant species show a varied range of requirements of and tolerance to environmental conditions, which is evident from their abundance and distribution along environmental gradients. The establishment of a forest in an area is determined by many factors. The local and regional climate, topographic position, disturbances, environmental factors and biotic interactions determine forest structure and composition

(Spurs and Barnes 1980).

12 The most obvious factor that limits the establishment of vegetation in an area is the amount of solar radiation, which determines the climate of an area. Since light is the main source of photosynthesis in plants, the amount of solar radiation also determines the availability of light for photosynthesis. In the temperate zone, photoperiod affects processes such as dormancy and germination of seeds, leaf fall and flowering (Champion and Seth 1968). The amount of available light also determines the amount of understory in a forest community; shade tolerant species grow under less illuminated conditions but shade intolerant species require more light. Plants of the same species growing under different light conditions show different morphology (Lal 1992). Herb species richness in stands of lower ages can be high because of the large canopy gaps and sufficient amounts of light penetrating the forest floor (Harcombe and Marks 1977).

Temperature, soil nutrients and rainfall are limiting factors for plant function

(Ogutu 1991). Associations between species distribution and average rainfall suggest that the complex moisture gradients underlie vegetation distribution (Ogutu 1996). Other studies also relate floristic characters with amount of rainfall (Belsky 1987; Boutton et al.

1988). The classification of Nepalese forest into topical, subtropical, temperate, sub alpine and alpine (Stainton 1972; Chaudhary 1998) is based on temperature, moisture and rainfall. Several global vegetation mapping systems (Holdridge 1967; Prentice et al.

1993) use temperature and moisture relations as predictors of the type of vegetation that occurs in different areas. Forests are established in soils that are rich in moisture and grasslands are established in drier soils (Spur and Barnes 1998).

The variety of soil types, their structure, parent material, pH, water and moisture holding capacity and nutrient content also limits plant species richness and distribution

13 (Lal 1992). Many soil nutrients (e.g. carbon, nitrogen and zinc) increase from grassland to forest, suggesting the influence of soil nutrients on vegetation types (Barnes et al.

1998). Tree species composition in La Selva, Costa Rica is significantly related to soil type and topographic variation, but both of the environmental variables explained small percentages of variation in species distribution (Clark et al. 1999). Species richness in tropical forest has been related to substrate (Richards 1961; Ashton 1971) and other factors such as chemical fertility and soil moisture (Hart et al. 1989). Low forest diversity and occurrence of few species have been attributed to low nutrient levels in some cases (Ashton 1971; Janzen 1974). Rainfall (Gentry 1982) and soil moisture (Hall and Swaine 1981) are related to an increase in diversity. Tropical rain forest structure and biomass varies with different variables such as soil type (Tuomisto et al. 1995), soil nutrients (Laurance et al. 1999), climate (Gentry 1982), disturbance regime (Lugo and

Scatena 1996), successional status (Saldarriaga et al. 1988), topographic position (Austin et al. 1996), and human impacts (Laurance et al. 1997). A comparison of forest structure and composition on different soil types in La Selva Biological station in Costa Rica showed lower density and larger average tree diameter on the more nutrient rich old alluvial soils than residual soils, but there was no difference in basal area and above- ground biomass (Clark and Clark 2000).

Disturbances are also important factors that shape the structure and composition of forest communities. Different factors cause disturbances; strong winds, fire and land slides destroy vegetation, open up the canopy and remove vegetation and soil cover.

Human induced disturbances include removal of biomass and intentional fires. Other forms of disturbance such as herbivory also affect the natural course of forest

14 regeneration and growth. Maintenance of forest cover is important to prevent the loss of nutrients, because the vegetation adsorps nutrients and holds soil particles, so removal of vegetation and reduced soil fertility will affect the inter-dependence of soil and vegetation. Disturbance is interactive with moisture, only in sites with high moisture

(forests and bushlands) the effect of disturbance is significant (Ogutu 1996). Vegetation disturbance, especially grazing, increases species richness in some situations, due to the greater occurrence of non-endemic species (Green and Kauffman 1995). Low intensity and sustained human disturbance through selective logging, firewood extraction, grazing and land clearing for permanent agriculture may influence plant communities and their successional patterns (Attiwill 1994; Fujisaka et al. 1998). Within forests, disturbances influence the availability of resources such as light, water and nutrients necessary for the survival and growth of seedlings (Marks 1974; Carlson and Groot 1997). Fire can alter both the structure and composition of forests, especially in the understory (Rodgers et al.

1986). In both Corbett and Dudwa National Parks (India), adjacent communities of very different understory structure have arisen due to differential fire frequency (Rodgers et al.

1986). The influences of past human disturbances on forests and ecological processes have been observed even after fifty years (Xiaoming et al. 1995).

1.7 Forests in the Western Terai

The Western Terai is located in the subtropical zone of Nepal. Forest types that occur there are: Sal Forest, Tropical Deciduous Forest and Tropical Evergreen Forest

(Staintion 1972). Some workers have also classified Sal Forest as moist deciduous forest

(Dinerstein 1979). Tropical Deciduous Forest occurs in a variety of climatic conditions, but alternating wet and dry periods favor their establishment. Various factors such as the

15 length of the wet period, total rainfall, latitude, longitude and altitude affect the structure and composition of deciduous forests (Shankar 2001). More than half of the Terai in

Nepal is dominated by Shorea robusta (locally known as Sal; Webb and Sah 2003). It is a light demanding large deciduous tree growing up to 45 m in height on a wide range of soil types (Rautiainen and Suoheimo 1997). It belongs to the Dipterocarpaceae family, which is found in tropical and sub-tropical Asia. In Sal, leaves fall for a very short period of time before the emergence of new foliage, resembling a deciduous species (Pande and

Shukla 2001). The Sal forest develops in well drained soil and is regarded as a climax community (Dinerstein 1979, Banerjee et al 1992). This forest is not inundated during the monsoon period, and the ground water table is very low (Bolton 1976).

In Nepal, Sal is considered to be the most valuable tree species and is used in construction and carpentry work, and is also the main source of fuelwood in the lowland areas. Sal leaves are valuable as fodder and for making disposable plates by local people

(Jackson 1994). Most of the rural communities in Nepal, which constitute 80% of the total population of the country (World Resource Institute 1996), depend on Sal Forest for subsistence needs. Before 1950s, the vast tract of Shorea robusta forests of the Terai remained unutilized. The Forest condition changed after people started migrating into the fertile Terai because of the eradication of malaria, establishment of resettlement offices in the districts, construction of the East-West Highway and political disturbances in the mountains (HMGN1996; HMGN 1998). The districts of Western Nepal (Bardia, Kailali and Kanchanpur) are major recent migration districts. Almost all of the 545,900 ha of forests outside protected areas in the Terai and Siwalik region have been converted into secondary forests because of intensive use by both the government and local people

16 (Kanel and Shrestha 2001). These forests are declining at a fast rate with negative

consequences on surrounding temperatures, land stability, soil and biodiversity, and on

the livelihood of local people who are directly dependent on them (Zomer et al. 2001).

Local people collect fodder, firewood, poles, timber and wild vegetables (ferns,

mushrooms, medicinal plants. etc) from these forests. In addition to primary forests,

conservation and management of these secondary forests is essential to provide most of

the resources for local livelihoods and protect biodiversity.

It is essential to understand the composition, functioning and dynamics of the

system to identify the major determinants of forest health within human dominated

environments and to manage these resources for the future. The present study examines

the structure and composition of Sal Forests in the western Terai of Nepal to provide

valuable ecological information on the forests that exists under variable degree of human

disturbance.

Various biotic and abiotic factors have been recognized that contribute to plant succession in the region. Abiotic factors include the length of the monsoon, total rainfall, seasonal flooding, and soil conditions, whereas biotic factors are previous land-use practices such as grazing of domestic stock, burning, clearing for cultivation, selective cutting of trees, logging, lopping for fodder and thatch grass cutting (Dinerstein 1979).

The study also analyzes the relationship between plant communities with the abiotic environment to identify the major environmental conditions influencing forests.

The main objective of TAL is to conserve biodiversity in protected areas and outside forests and to link protected areas with corridors; for this to be successful, ecological integrity of forests outside protected areas must be taken into consideration.

17 The study compares between protected forests and community forests outside protected areas so that information on the ecological health of community forests, which have been under more severe human pressures, can be incorporated in planning efforts.

1.8 Objectives of the study

The goal of the present study is to provide more ecological information about forest for management of areas under the TAL program. The broad objectives of the study are:

1) To study the structure, composition and dynamics of Sal Forest in the western

Terai of Nepal

2) To understand the relationships between forest communities and environmental

conditions

3) To determine the ecological differences between community forests and forests

inside protected areas.

To address these objectives, I posed several questions 1) What is the structure and composition of the forests sampled? 2) What are the different associations found within these forest types? 3) What are the structural and compositional differences between different associations? 4) What explains the variation between the associations identified?

5) How are the different structural variables such as tree size and density, tree abundance and seedling density and percent cover of different layers related to each other? and, 6)

What are the differences in structure and composition of protected forests and community forests?

18 2. THE STUDY AREA

The study area is located in the south-western part of Nepal in Bardia and

Kanchanpur Districts. Bardia District lies in the Midwestern Development Region and

Kanchanpur in the Far Western Development Region of Nepal. This study was carried

out in Royal Bardia National Park (RBNP) in Bardia District, and Royal Suklaphanta

Wildlife Reserve (RSWR) and Birendra and Mayur Jagdamba Community Forests in

Kanchanpur District (Figure 2). Although these community forests have been officially

demarcated and user groups defined for the last two years, they have not been formally

handed over to communities for management. The elevation of the study area ranges

from 150m to 220 m. The study area is composed of alluvial flat land, commonly referred as Terai.

Bardia district lies between 28o 36’ - 28o 50’ N and 81o 30’- 81o 45’ E. The

district lies in western Nepal and is bordered by Banke District in the east and Kailali

district in the west. The district is 64 km long and 63 km wide. According to the

Operational Forest Management Plan of the district, out of 201,677 ha total area, RBNP covers the majority of the area (87,936 ha). Agricultural land and settlements cover

33.7% (68,075 ha) of total area (Table 1).

The district can be divided into Terai and Siwalik. The Terai extends towards the southern portion of the district. The Terai is composed of flat and highly fertile alluvium deposits. The northern portion of the district is the Siwalik Range, which extends from east to west and is fragile and undulating. The elevation of the district ranges from 150 m in the south to 1457 m in the north.

19 The geological formation of the district is alluvium in the south which is similar

to the Gangetic Plain of India. The northern part of the district is similar to tertiary

Siwaliks. The Siwaliks are composed of coarsely bedded stone, crystalline rocks, clays

and conglomerates (HMGN 1996). The rocks of the Siwalik are fragile and composed of

sandstones. The disintegration of these sandstones forms sandy soils in the region. The

foothills of the Churia or Siwaliks are known as Bhabar. They are comparatively dry,

consisting mainly of boulders, cobbles and coarse sand layers. The Bhabar soils are

relatively well drained and deep. The Terai or the southern extension of Bhabar is

composed mainly of alluvium soils. The soil type is mainly sandy loam (FSRO 1971)

throughout the district with some local variation. Soil depth is great in flat lands and

shallow in the hills.

The climate of the district is sub tropical monsoonal in the Terai and Siwaliks.

Four distinct seasons occur: winter, spring, summer and rainy. The mean minimum temperature varies between 10o C (January) to 26o C (June). The mean maximum temperature ranges from 20o C (January) to 37o C (May). Rainfall recorded in the

Chisapani Station from 1987 to 2001 shows that mean monthly rainfall varies from 23 mm (March) to 635 mm (July). The mean total annual rainfall recorded from 1987 to

2001 was 2100 mm. Most rain occurs between the months of June and September followed by 7 to 8 months of dry season (Figure 1a and Figure 1b).

2.1 Royal Bardia National Park

This study was carried out in the southwestern section of Royal Bardia National

Park (81o 20’ E and 28o 35’ N; Figure 2). The park is 968 sq km not including the proposed extension area and is the largest park in the Terai. It started as a Royal Hunting

20 Reserve in 1969 and later in 1976 was declared Royal Karnali Wildlife Reserve (386 sq

km; HMG 2001). In 1984 the area was extended to the east to include a total area of 968

sq km. It was officially designated as Royal Bardia National Park (RBNP) in 1988. The

park consists of Bhabar, Terai and Riverine Flood Plains. Most of the park lies in the

Bhabar zone which contains rocks, boulders and sand interbedded with clay and silt

driven down from Churia hills. Soils in this zone are young and shallow and are very

prone to erosion. The alluvial soils deposited by the rivers like Orai, Karnali and Babai

are deep.

The vegetation of the park is early successional floodplain grassland in alluvial floodplains to climax Sal community in the relatively dry flat lands. Dinerstein (1979) classified the vegetation of the park into six major types and later modified by Jnawali and Wegge (1993) into seven types. The major vegetation types are Sal Forest, Khair- sissoo Forest, Moist Riverine Forest, Mixed Hardwood Forest, Wooded Grasslands,

Phantas and Tall Alluvial Floodplain Grassland.

Sal Forests are found in relatively well drained upland areas and the major species is Shorea robusta with other associated species such as Buchanania latifolia, Terminalia sps and Lagerstroemia parviflora. Khair-sissoo Forests are established on old floodplains and consists of Dalbergia sissoo and Acacia catechu as dominant tree species. Along the water courses are Moist Riverine Forests with tree species such as

Syzygium cumini, Mallotus phillippnensis and Ficus glomerata. Adina cordifolia,

Casearia tomentosa, Lagerstroemia parviflora and Mitragyna parviflora are common species of Mixed Harwood Forest that grows in well drained areas. It resembles Bombax savannah (Wooded grasslands) due to similarity in composition of the trees and the

21 understory, but consists enough tree density and conspicuous shrub layer to qualify as

forest (Dinerstein 1979). Wooded grasslands are similar to Savannah and contain grasses

with interspersed trees. Trees such as Bomabx ceiba, Adina cordifolia and Mallotus

phillippensis are scattered amongst grasses such as Saccharum spontaneum, Imperata

cylindrica, Desmostachia bipinnata and Vetiveria zizanoides. Previously cultivated lands

where grasses established after the resettlement of villages and farms have been identified

as Phantas. They are dominated by grasses such as Imperata cylindrica, Saccharum

spontaneum and Narenga porphyrocoma. On the river beds are found Tall Alluvial

Floodplain grasslands which are dominated by Saccharum spontaneum, Saccharum

bengalensis, Phragmites karka and Arundo donax.

The park is home to many endangered species of flora and fauna. There are 53

species of mammals including five species of deer, approximately 400 species of birds,

25 species of reptiles and amphibians and 121 species of fishes (Basnet 1995). The

Royal Bengal Tiger (Panthera tigris), One-horned Rhinoceros (Rhinoceros unicornis) and Asian Elephant (Elephus maximus) are also found in Bardia. Endangered avifauna such as Bengal Florican (Houbaropsis bengalensis; Baral et al. 2002) and Giant Hornbill

(HMGN 2001) are present.

2.2 Kanchanpur District

Kanchanpur District is located between 80o 3’ – 80o 33’ E longitude and 28o 32’-

29o 8’ N latitude. The district borders Kailali District in the east and India in the west and

south. It lies in the Mahakali zone of the Farwestern Development Region. The total

geographical area covered by the district is 163, 678 ha. Of the total area, forests cover

88,200 ha (53 %) and agriculture, urban areas and other land use types cover 74,478 ha

22 (46.1% ; FSD/FORESC 1993). The topography of the district is similar to Bardia as

described above. Most of the district lies in the Terai region and the northern part is

covered by Siwaliks and Bhabar. The elevation ranges from 169m in the south to 1528 m

in the north.

The climate of the region is subtropical monsoonal with four distinct seasons:

winter, spring, summer and rainy. Winters are cold and summers are very hot. The mean

maximum temperature varies from 21o C (January) to 37o C (May) and the mean min temperatures ranges from 7o C (January) to 25o C (July). The mean total annual rainfall for the period of 1987 to 2001 was 1579 mm, about 75% of the annual precipitation falling at Bardia (RBNP). The lowest amount of mean monthly rainfall recorded was 3

mm in March and the highest was 636 mm in August. Most of the rainfall occurs in the

four months from June to September. The climatic data from 1987 to 2001 are shown in

Figure 1a and Figure 1b.

2.3 Royal Suklaphanta Wildlife Reserve

RSWR lies in Far Western Nepal in Kanchanpur District and covers an area of

305 sq km (Figure 2). In 1969 it was declared as Royal Hunting Reserve and it was

officially gazetted as Royal Suklaphanta Wildlife Reserve in 1973. The reserve lies

between 28o 45’- 28o 57’ N to 80o 07’ to 80o 21’ E and is bordered by the Chaudhar river

on the east and by forests and cultivated fields on the north. It has a common boundary

with the Indian State of Uttar Pradesh on the south and borders the Mahakali River on the

west. The Reserve has a similar climate, topography and soil as described for

Kanchanpur District above. It consists mostly of Terai and some areas lie in the Bhabar

23 zone. The elevation varies from 162 m to 237 m. A number of rivers such as the

Mahakali, Bauni, Chaudhara and Syauli drain the reserve.

Vegetation types that are found in RSWR are forests, grasslands and wetlands.

Sal Forest is the dominant forest type in the Reserve, and is found in well drained upland

areas. Species such as Shorea robusta, Terminalia tomentosa, Lagerstroemia parviflora

and Mallotus philippensis are found. The other forest types present within the reserve

are: Acacia catechu-Dalbergia sissoo forest in floodplains and Bombax ceiba,

Holarrhena pubescens and Grewia disperma in the interior; Mixed Deciduous Forests are

present in the poorly drained soils and species such as Adina cordifolia, Trewia nudiflora,

Syzygium cumini and Celtis tetrandra are found. The reserve is famous for its large tracts

of grasslands, locally known as Phanta. The name Suklaphanta was derived from one of

the grasslands found inside the reserve. The major phantas are Sukla, Haraiya, Barkaula,

Singhpur etc. Imperata cylindrica, Vetiveria zizanoides and Saccharum spontaneum are

the major species found in the grasslands. Wetlands such as Rani Tal and Sikari Tal are

present and dense grasses such as Phrgamites karka and Saccharum spontaneum predominate.

A total of 267 species of birds has been recorded in the reserve (Inskipp 1989).

The reserve is important for endangered birds such as Swamp Francolin (Francolinus gularis) and Bengal Forican (Houbaropsis bengalensis). This reserve is the last

stronghold of the endangered Swamp deer (Cervus duvauceli) in Nepal. An estimated

population of about 3000 individuals lives in the reserve (Chaudhary 1998). It is home to

endangered megafauna such as tiger, one-horned rhinoceros and Asian elephants.

24 2.4 Community Forests

The community forests sampled were located in Bank area of Kanchanpur district.

Two community forests were studied, namely Birendra and Mayur Jagdamba. Both of

the community forests lie north of east-west highway between 28o 52’ 12’’ to 28o 52’ 54”

N latitude and 80o 24’ 55” to 80o 25’ 32” E longitude. The Banda River runs along the

eastern side of the forests. The forest type was Sal Forest dominated Shorea robusta and

Terminalia tomentosa. The area was heavily used by locals for grazing, fuelwood and

fodder. Although logging and cutting of trees was legally not allowed, locals reported

that there are instances of illegal logging. There were cut stumps and dead trees. Several

species of birds and a hare were observed during the field survey. I did not find any

tracts of nor did I observe any large mammals during the survey.

3. METHODS

3.1 Forest Sampling

The forests of the study area were sampled between February and April 2005.

Woody vegetation was sampled at three levels; trees (> 5 cm dbh), saplings (1-5 cm dbh

and > 1m height) and seedlings (<1m height). Shrubs and herbs of the area were also

sampled. Three sites were selected for the study; two sites were inside protected areas

and the third site included the two community forests. Sites were chosen to encompass

forests under different management regimes. Altogether 7 transects were laid in the

study areas. Three transects were in Royal Bardia National Park (RBNP) in Bardia

District, and three were in Royal Suklaphanta Wildlife Reserve (RSWR) in Kanchanpur

district. A single transect encompassed the two community forests (Birendra and Mayur-

Jagadamba Community Forest) in Kanchanpur District. The two protected areas were

25 approximately 150 km apart and the community forests were more than 10 km east of

RSWR and more than 130 km west of RBNP. The study areas were at a similar elevation and extended from east to west. Sampling locations were established every 200 m within each 2 km transect. Starting points of transects were selected randomly, within the fire line for National Parks and along the east-west highway for community forests determined from a topographic map of the study area. Altogether 30 locations were sampled in RBNP, 30 in RSWR and 10 in the community forests. The continuing Maoist insurgency in Nepal posed security problems and made it difficult to collect enough samples from community forests.

For sampling trees, a plotless technique or variable plot cruising (Grosenbaugh

1952) was used. This technique is more efficient and useful than fixed plots for one-time estimation of forest structure in areas where there are a few large trees and numerous small trees (Zhang et al. 2005). With the sampling point as center, all trees around were observed through a prism of known dioptre. A tree was counted “In” if its diameter at breast height was large enough to subtend the fixed critical angle of the prism, or “Out” if it was not. Each “In” tree was identified to species and diameter at breast height (dbh;

1.3m above the ground) was measured. The height of the three tallest trees was also measured with a clinometer. Using the diameter of “In” trees basal area and density were estimated for each sampling point. Each “In” tree represents a fixed basal area equal to the Basal Area Factor (BAF) associated with the prism. Density was calculated by dividing BAF by the basal area of the tree. The basal area and density associated with each tree was summed to estimate the basal area and density of each sampling point, then later converted into basal area and density per ha.

26 For saplings and shrubs (> 1m and < 0.5 cm dbh), a five meter radius circular plot was established with the tree sampling point as the center. Within each plot, saplings and shrubs were identified to species and the number of individuals of each species was counted. Two 1 sq m circular plots nested within sapling and shrub plots were used for sampling herbs and seedlings. Herbs and seedlings were also identified to species and their numbers within each plot were estimated. Specimens were collected for unidentified species, which were later identified in the Central Department of Botany,

Tribuvan University, Kathmandu and The National Herbarium and Plant Laboratory,

Godavari. Nomenclature follows Hara and Williams (1979), Hara et al. (1978, 1982) and

Press et al. (2000).

Fixing tree sampling point as the center, percentage canopy closure was estimated. Canopy closure percentages were estimated in all the four directions with the help of a densiometer and the average of the four values was used to calculate the final percent closure. Ocular estimation of undertstory and ground cover was done using a modified Braun Blanquet system (< 1%, 1-4%, 4-16%, 16-33%, 33-66% and >66%).

3.2 Soil sampling and analysis

Soil samples (0-15 cm depth) were collected from each plots. The composite sample used for the analysis was a mixture of samples collected from four directions.

Samples were collected in polyvinyl bags and were analyzed at the laboratory of

Department of Agriculture, Lalitpur, Nepal for pH, texture, organic matter (%), total nitrogen (%), available phosphorous (kilogram/hectare) and available potassium

(kilogram/hectare). For the determination of pH, mixtures (1:1) of air dried soil samples and distilled water were measured with a pH meter.

27 Organic matter content was determined by Walkley and Black’s method

following Bray (1945). For the process, a mixture of 10 ml potassium dichromate (1N),

20 ml concentrated sulfuric acid and 1 gm of 0.2 mm sieved soil samples were titrated

with 0.5 N ferrous ammonium sulphate. Total nitrogen was estimated by micro-Kjeldahl method (Jackson 1958). The procedure used was auto digestion and auto distillation, and titration with 0.01 N hydrochloric acid. Available phosphorous was estimated following

Olsen’s method and the extraction solution used was 0.5 N sodium bicarbonate at pH 8.5

(Hesse 1994). For available potassium the soil was extracted with 1 N ammonium acetate at pH 7.0 and potassium was determined by flame photometer method. Soil texture was determined by soil a expert in the Department of Agriculture by feeling the samples between his fingers.

3.3 Data Analysis

Density and basal area per ha were calculated for all tree species. Relative values for frequency, density and dominance of trees was calculated by dividing individual values for frequency, density and basal area by sum of frequencies, densities and basal areas of all species. All relative values were multiplied by 100 to express them as percentages. The Importance value index (IVI) was calculated for all the tree species by

summing the relative frequency, relative density and relative dominance values of

individual species. For the analysis of population structure, individuals were classified

into following dbh classes; 5-10 cm, 10-15 cm, 15-20cm, 20-25cm, 25-30 cm, 30-35 cm

and > 40 cm. For some species, population distribution in two dbh classes, 50-80cm and

> 80 cm, was also calculated. Densities of saplings, shrubs, seedlings and herbs per plot

and per ha were estimated.

28 Following the method described in PC-ORD statistical package (McCune and

Mefford 1999), species richness, eveness and Shannon’s diversity index (H’) were calculated for each plot. These values were calculated separately for different life forms; trees, saplings, shrubs, seedlings and herbs. Numbers of species per plot was taken as a measure of species richness. The Shannon diversity index (Shannon and Weaver 1949) was calculated by the following formula:

H’ = - ∑pi ln pi where pi represents the proportional abundance of ith species in the community.

Pielou’s J (Pielou 1966, 1969) was used as an index of evenness, which is expressed as:

J= H’/ log S where H’ is Shannon’s diversity measure and S is the species richness.

Hierarchical agglomerative cluster analysis (McCune and Grace 2002) was used to define grouping among the 70 plots sampled. Cluster analysis was performed using the importance value of tree species. The agglomerative clustering method builds hierarchically from bottom to top. It calculates a distance between any pair of entities and merges groups with some criteria of minimum distance. Results of a cluster analysis are presented in a dendrogram. The Wishart objective function in a dendrogram indicates the amount of information lost during merging of each group. It is the sum of the error sum of squares from the mean of variables in a cluster to the individual observations of variables in that cluster. For the process, Sorensen (Bray- Curtis) dissimilarity was used as the distance measure and flexible beta was used to calculate relatedness among the groups (McCune and Grace 2000). Species occurring in less than 5% of the samples and two outliers were deleted. The remaining 68 plots and 18 species were used for the

29 analysis. Groups were selected from the dendrogram using information provided by the

Wishart objective function with approximately 40% of the information remaining after

the merging of groups.

I analyzed the interrelationships between plant communities by ordinating sample

plots in species space using non-metric multidimensional scaling (NMS). The

advantages of NMS over other ordination techniques are: 1) it is not based on the

assumption of multivariate normality and, 2) it is robust to large numbers of zero values

(Minchin 1987). The same data set used for cluster analysis with 68 plots and 18 species was used for ordination. Among the 29 species recorded, species present in less than 5% of the plots were deleted. Following this criterion, 18 species were used for the ordination. I tried to explain relationships between plant communities and environmental variables by overlaying and contouring environmental variables on the NMS ordination of plots. Correlations between the ordination and environmental variables were calculated with Pearson’s r (Peterson and McCune 2001). PC-ORD statistical package

was used for the cluster analysis and ordination.

One way analysis of variance was used to test the difference in environmental

variables between the groups if the data were normal. Normality was tested by plotting

histograms and with the Shapiro-Wilk test (Shapiro 1980). If the data were not normal

and the assumption of equal variance was violated, I used the Kruskal-Wallis test

statistic, a non-parametric anova equivalent (Sokal and Rohlf 1995) to determine if the

values of environmental variables between the groups differed. I used Kruskal-Wallis to

test the difference among groups, identified by cluster analysis, in parameters related to

trees, saplings, shrubs, seedlings and ground layer. For the multiple comparisons I used

30 one-tailed Mann-Whitney U test (Nagendra 2002) with the alpha level fixed by dividing

0.05 by number of groups compared. Most of the techniques I used in the analysis are non- parametric except few parametric ANOVAs and two sample t tests. Non-parametric methods are well suited to data that are non-normal, are on discontinuous scale and contain a large proportion of zero values (Peterson and McCune 2001). Variables tested among the groups are sapling density, richness and diversity; shrub density, richness and diversity; seedling density; canopy closure, and understory and ground cover percentages; and herb diversity and richness. I used Spearman’s rank correlation to test the association between different cover percentages. Data on forests inside the two protected areas were pooled together and compared with pooled data from community forests. For comparative purposes, I used two sample t-tests in the case of normal data, and for the non-normal data, I used Mann-Whitney U test.

To understand the stand dynamics, I tested three different kinds of relations:

1) Average tree size and total tree density relationship 2) tree structure (maturity and stocking %) with seedling density 3) and species abundance in tree layer and seedling density.

To test the first relationship, I calculated the stand density index (SDI; Reineke

1933) for all the stands. Reineke (1933) discovered that the most important factor determining stand density is the average stand diameter (ASD- diameter at breast height of a tree with average basal area). He found a straight line relationship with a slope of -

1.6 between log transformed values of tree density and ASD. Based on this, he proposed a stand density index (SDI) which provided information on maxiumum tree density at full stocking of stands in any stages of development having any ASD. He revealed that the

31 number of trees per acre of any pure, even-aged and fully-stocked stand having a certain

average stand diameter is approximately equal to the number of trees per acre of another

pure, even aged and fully stocked stand of the same species having the same average

stand diameter. SDI (stand density index) values provide the full stocking of stands of

certain ASD and are useful to compare stands with different ASD values. Reineke’s SDI is also more useful in comparing of the stocking among uneven-aged stands than is basal area (Stage 1968 cited in Daniel et al. 1979). For the present study the metric equivalent of Reineke’s index provided by Daniel et al. (1979) was used for calculating SDI.

The second relationship was tested by plotting ASD (maturity) and SDI (stocking

%) values of each plot against its seedling density. To test the third relation, I used a scatterplot of Shorea robusta tree relative density and its seedling relative density for each plot. Relative density was calculated as a percentage of maximum observed for the species in all transects.

4. RESULTS

4.1 Average forest structure and composition

Altogether 121 species were recorded: 28 species of trees, 10 species of shrubs, 6

species of climbers and 87 species of herbs. The forests sampled in this study were Sal

Forests, the canopy layer was dominated by Shorea robusta and Terminalia tomentosa.

Occassionally Adina cordifolia and Terminalia bellirica were present in the canopy layer.

The sub canopy was dominated by Buchanania latifolia, Dillenia Pentagyna,

Cleistocalyx operculatus and Lagerstroemia parviflora. The understory was quite sparse

and dominated by Shorea robusta saplings and shrubs such as Flemingia strobilifera,

Clerodendrum viscosum and Indigofera pulchella. The mean density across all plots was

32 220 trees per ha. The highest density was recorded for Shorea robusta (64 stems ha-1), followed by Buchanania latifolia (50 stems ha-1), Cleistocalyx operculatus (25 stems ha-

1), Lagerstroemia parviflora (22 stems ha-1), Terminalia tomentosa (16 stems ha-1) and

Dillenia pentagyna (11 stems ha-1). The density distribution of the most common tree

species among different dbh classes for the seven transects is presented in Figure 3.

Shorea robusta was present in all dbh classes but was in higher density in the

lower 5- 10 cm dbh class and above 30 cm dbh classes. The highest density was in > 40

cm dbh class. Terminalia tomentosa was completely absent from the lowest dbh class,

but was present in other dbh classes in low density. The most abundant species in 5-10

cm dbh class was Buchanania latifolia (47%) followed by Lagerstroemia parviflora and

Cleistocalyx operculatus. The largest dbh class was represented by Shorea robusta and

Terminalia tomentosa and very rarely by other species. Species such as Buchanania

latifolia, Cleistocalyx operculatus, Lagerstroemia parviflora and Dillenia pentagyna

were more abundant in less than 30 cm dbh class and rare in classes above 30 cm. I also

classified trees into 5 cm -25 cm, 25 cm -50 cm, 50cm-80 cm and > 80 cm classes. S.

robusta and T. tomentosa were present in > 50 cm diameter classes. Density of S.

robusta in the 50- 80 cm dbh class was 14 trees/ha and T. tomentosa was 3 trees/ha. S.

robusta was 2 trees/ha and T. tomentosa was 0.5 tree/ha in above 80 cm dbh class. Other

tree species that were present in the dbh class >50 cm were Adina cordifolia, Terminalia

bellirica and Syzygium cumini. All the other tree species had < 50 cm dbh.

The average total basal area across all plots was 13.2 sq m/ha, the minimum was

3.4 sq m/ha (RSWR) and the maximum was 22 sq m/ha (Community Forest). The

highest basal area was for Shorea robusta (9 sq m/ha) followed by Terminalia tomentosa

33 (2 sq m/ha) and other species covered less than 1 sq m/ ha. Tree height was measured for the three tallest trees in each plot. The minimum tree height for the tallest trees was 20.7 m and the maximum was 42 m, with a mean tree height of 28.5 m. More than 80 % of the tallest trees were S. robusta and T. tomentosa. Other trees that were present in the tallest category were Adina cordifolia, Terminalia bellirica, and Syzygium cumini.

There were altogether 3.6 tree species per plot and the Shannon Diversity Index (H’) was

0. 82 for the trees.

4.2 Classification

The hierarchical agglomerative cluster analysis separated three groups (Figure 4) among the seventy plots that were sampled. The three groups or associations were identified based on the presence or absence and importance values of different species in each group. There were 18 plots in Group 1, 13 plots in Group 2 and 37 plots were in

Group 3. All but three plots in Group 1 were from RBNP. Similarly in Group 2, all but three plots were from RBNP. Of 37 plots in Group 3, 6 plots were from RBNP and all others were from RSWR or Birendra and Mayur Jagdamba Community Forests of

Kanchanpur district. Table 2 presents the importance value of the different species for the three groups.

Shorea robusta-Buchanania latifolia association

Group 1 was defined by the Shorea robusta–Buchanania latifolia association based on the importance value (Table 2) of the two species. Other species abundant in this group based on their importance values were Dillenia pentagyna, Terminalia tomentosa and Semecarpus anacardium. Twenty-one species of trees were present in this association. Ten species were abundant and the remaining 11 species had frequencies

34 less than 11%. Engelhardtia spicata, Ficus benghalensis, Holarrhena pubescens, and

Acacia catechu were present exclusively in this association. The minimum density for this association was 21 trees/ha and the maximum was 308 trees/ha with the mean density of 289 trees/ha. B. latifolia (111 trees/ha) had the highest density, followed by S. robusta

(33 trees/ha), L. parviflora (31 trees/ha) and C. operculatus (30 trees/ha). Other species such as Schleichera oleosa, Adina cordifolia, Syzygium cumini, Terminalia bellirica and

Careya arborea were less than 2 trees/ha. The minimum basal area for this community was 5.7 sqm/ha and the maximum was 18.3 sqm/ha with the mean basal area of 12.8 sqm/ha (Table 4). S. robusta was the dominant species with the basal area 6.2 sqm/ha, other species had basal area less than 2 sqm/ha. The basal area for S. robusta varied between 2.2 sq.m/ha to 11.4 sq.m/ha (for details see Appendix 1 and Appendix 2).

For Group 1, the smallest diameter class was represented by Buchanania latifolia and Lagerstroemia parviflora (Figure 5). Shorea robusta was present in both smaller

(very few in number) and larger diameter classes but was most abundant in the largest diameter class. Terminalia tomentosa was present in the larger diameter class (above 20 cm) and the highest density was in > 40 cm dbh class. Dillenia pentagyna was most abundant in the diameter class 15-20 cm. The density of Cleistocalyx operculatus was highest in the diameter class 10-15 cm. As is common in many forests, density distribution was skewed towards the left.

Terminalia tomentosa-Shorea robusta association

Group 2 was Terminalia tomentosa-Shorea robusta association which was explained by the higher importance value (Table 2) of Terminalia tomentosa followed by

Shorea robusta. Other species abundant in the group were Buchanania latifolia,

35 Mallotus philippensis, Myrsine semiserrata, Dillenia pentagyna and Lagerstroemia parviflora. Anogeissus latifolius which was present in the first group was absent in this group. Nineteen species of trees were present in this community. Seven species were abundant and the rest had frequencies of less than 15%. Species such as Picrasama javanica, Desmodium oojeinense and Paruli (local name) were present exclusively in this community. The mean density of trees was 297/ha with density varying between 13 trees/ha to 938 trees/ha. B. latifolia (96 trees/ha) had the highest density followed by T. tomentosa (46 trees/ha), M. semiserrata (46 trees/ha), L. parviflora (42 trees/ha) and S. robusta (25 trees/ha). Adina cordifolia (0.23/ha) had the lowest density in this community. The mean basal area was 13.5 sq m/ha with the basal area ranging from 6.8 sq m/ha to 18.3 sq m/ha. S. robusta and T. tomentosa were the dominant species in terms of basal area, representing 5.1 sq m/ha and 4.5 sq m/ha respectively. Other species comprised less than 1 sq m/ha basal area. The basal area for S. robusta varied between

2.2 sq m/ha to 8 sq m/ha and for T. tomentosa, between 2.2 sq m/ha to 6.8 sq m/ha (Table

3).

The smallest diameter class in Group 2 was represented by Buchanania latifolia,

Myrsine semiserrata and Lagerstroemia parviflora (Figure 6). The density of Terminalia tomentosa was higher in 10-15 cm than in the largest diameter class. T. tomentosa was evenly distributed in the dbh classes > 10 cm. The largest diameter class was represented by Shorea robusta, Dillenia Pentagyna and Terminalia tomentosa. Tree density was higher either in smaller diameter classes or in the largest diameter class. S. robusta was completely absent from < 20 cm dbh classes and was present in 20- 25 dbh class and in larger dbh classes.

36 Shorea robusta-Cleistocalyx operculatus association

The third group was identified by the Shorea robusta-Cleistocalyx operculatus

(Eugenia operculata) association. Shorea robusta was highly dominant in this group as shown by its importance value (Table 2). Other species that were abundant are Mallotus philippensis, Lagerstroemia parviflora, Buchanania latifolia, Dillenia pentagyna,

Terminalia tomentosa. Semecarpus anacardium, which was present in the other two groups, was completely absent from this group. Altogether 18 species of trees were recorded in this community. Four species were abundant and the remaining 14 species had frequencies of less than 9%. Jingar (local name) was present exclusively in this community. The mean density of trees was 163 trees/ha with densities ranging from 13 trees/ha to 567 tress/ha. S. robusta (94 trees/ha) had the highest density of trees followed by C. operculatus (31 trees/ha). Other species had densities of less than 6 trees/ha. The mean basal area was 13.7 sq.m/ha with a range between 3.4 sq.m/ha to 20.6 sq.m/ha

(Table 4). S. robusta was the dominant species with the basal area ranging from 2.2 sq.m/ha to 19.5 sq.m/ha (Table 3). All the other species had basal areas less than 1 sq.m/ha. See Appendix 1, 2 and 3 for detailed descriptions of density, basal area, frequencies and relative frequencies of tree species in different groups.

The density of Shorea robusta in the largest diameter class was the highest for this group compared to other two groups (Figure 7). Shorea robusta was present in all diameter classes but was most abundant in the smallest and largest diameter class.

Cleistocalyx operculatus was well represented in the smallest diameter class and was found up to 35-40 cm dbh class. T. tomentosa was present in most of the dbh classes except the lowest and 15-20 cm dbh class.

37 Tree height was similar in all the three groups. A one-way ANOVA showed no significant difference in the height of the tallest trees between groups (F = 2.10, p =

0.1307).

4.3 Forest-environment relationships

Plots were ordinated using the importance value of individual species. Nonmetric

Multidimensional Scaling (NMS) was used for the ordination (Figure 8). Among 29 species recorded, species present in less than 5% of the plots were deleted. Following this criterion, 18 species were used for the ordination. Two outliers were deleted and minimum stress of 10.2 for two dimensional solutions was used for the final ordination.

Axis 1 separated T. tomentosa-S. robusta (Group 2) association from S. robusta-B. latifolia (Group 1) association and S. robusta-C. operculatus (Group 3) association. Axis

2 separated S. robusta-C. operculatus association from T. tomentosa- S. robusta association and S. robusta-B. latifolia association. I overlaid contour plots of environmental variables on NMS site ordination to explain the variation with the environmental gradients. None of the variables tested in the study explained distribution of the groups in the ordination space (Fig 9a- Fig 9f). Pearson’s correlation showed that axis 1 was significantly correlated with pH and available phosphorous, but the correlation was minimal. None of the environmental variables were correlated with Axis 2. The environmental variables measured were pH, percent organic matter, total nitrogen, available phosphorous, available potash and soil texture (Table 4).

The soil in the forest was acidic, with pH ranging from 4.8 to 7.5 and a mean pH of 5.52. The percent organic matter content of the soil varied from 0.59% to 4.08% with mean organic matter content of 2.19%. Total nitrogen varied from 0.03% to 0.2% with a

38 mean of 0.11%. Available phosphorous (kg/ha) and available potassium (kg/ha) varied

from 3.14 to 370.27 and 104.79 to 456. 51 respectively, with a mean value of 78.06 for

phosphorous and 77.74 for potassium. Soil texture varied from Sandy loam to Clay.

Among the 68 plots, 26% were Sandy loam, 30 % were Clay, 25% were Clay loam and

rest were Loam, Sandy clay and Sandy Clay loam. Soil texture was converted to an

ordinal scale, to generate contour plots, based on the percent of Sand, Silt and Clay.

Sandy loam, which is the most important soil type of the Sal forest (Dinerstein 1979) was

given a score of 6 and Clay soil received a score of 1. All the other soil types received

scores between 6 and 1. One-way ANOVAs and Kruskal-Wallis tests were performed to

see the difference in environmental variables between the groups. None of the

environmental variables measured were statistically different between the groups (P >

0.05).

4.4 Saplings and Shrubs

Altogether saplings of 17 tree species were recorded. Grewia sp., Cassia fistula

and Kaphale (local name) were absent in the tree layer but were present as saplings.

Mean sapling density was 1797 plants/ha (Table 5). Sapling species richness was 2.4

species / plot and Shannon’s diversity index was 0.56 (Table 5), which indicates low

richness and diversity. Mean sapling density per ha for the Shorea robusta-Buchanania

latifolia forest (Group 1) and Terminalia tomentosa-Shorea robusta forest (Group 2)

were close but mean sapling density for the Shorea robusta-Cleistocalyx operculatus

(Group 3) was less than the two other groups (Table 5). A Kruskal-Wallist test showed a

2 significant difference in sapling density between groups (χ 2 = 12.53; p = 0.0057). There was no significant difference in sapling density between Groups 1 and 2, and Groups 2

39 and 3, but Group 1 had higher density than Group 3 (z = 3.48; p = 0.0005). Table 6

presents sapling density of different species in the three groups. A Kruskal-Wallis test

showed a significant difference in Shorea robusta sapling density for the three groups

2 (χ 2 = 15.983, p = 0.0003). A Mann-Whitney (one tailed) pairwise analysis of difference between groups indicated that the difference in sapling density between Group 1 and

Group 2 was insignificant (z = 0.533, p = 0.0003), but the difference was significantly greater in Group 1 than Group 3 (z = 4.685, p = 0.000) and, Group 2 than Group 3 (z =

3.498, p=0.0005). Sapling density of Lagerstroemia parviflora was not significantly different between the groups. Likewise, Buchanania latifolia sapling was insignificant between Group 1 and Group 2. Refer to Table 7 for detailed descriptions of the sapling density difference between groups among species that are present in all three groups.

Sapling species richness was 14 for group 1, 13 for group 2 and 11 for group 3. Sapling species richness was significantly different between the groups (F= 5.58, p = 0.0058).

Results of Bonferroni multiple comparison showed that the difference was not significant between Group 1 and Group 2, and Group 2 and Group 3 ( p = 0.05), but was significantly higher in Group 1 than Group 3 (p = 0.004). Shannon’s diversity index for

2 sapling was not significantly different among the groups (χ 2 = 1.303, p = 0.5214).

Altogether 10 species of shrubs were recorded. Average shrub density across all plots was 337 plants/ha (Table 8). Per plot species richness, Shannon’s diversity index and eveness all were low (Table 8). Six species were present in Group 1, three species in

Group 2 and seven species in Group 3. Shrub species diversity was low in all three groups as shown by the Shannon diversity index (Table 8). Flemingia strobilifera was

the abundant shrub species in Group 1 and Group 2. This is considered to be an

40 indicator shrub species of the Terai-Sal forest (Shrestha and Jha 1997). In Group 1 other shrubs such as Phyllanthus sps and Indigofera pulchella were also abundant. In Group 2, the most abundant shrub species was Flemingia strobilifera followed by Indigofera pulchella and Flemingia chappar. Another indicator species of Terai Sal forest,

Clerodendrum viscosum, was present in Group 1 and Group 3 but absent from Group 2.

Clerodendrum viscosum was the most abundant shrub species in Group 3, followed by

Flemingia strobilifera, Elsholtzia blanda, Pogostemon benghalensis and Indigofera pulchella. Mean shrub density per plot for the three groups were significantly different

(P = 0.05) but Shannon’s index and species richness between the groups were not significantly different (Table 9).

4.5 Canopy closure and Cover percent

The mean percent overstory (Table 10) canopy closure, as measured by densiometer, was highest for Group 2 (69.92) followed by Group 1 (68.9) and Group 3

(62.57). Most of the plots belonging to Group 1 and Group 2 were from RBNP; they did not show much variation in tree canopy closure percentage, in contrast, mean percentage understory cover (ocular estimation) in Group 1 (38.17) was much higher than in Group 2

(34.77) or Group 3 (19.97). The mean percentage ground cover (ocular estimation) in

Group 3 (57.17), which had lower mean canopy closure and understory cover, was highest among the three groups.

Percent canopy closure was significantly different (Table 11) between the groups.

There was no difference between Group 1 and Group 2, but both the Groups had significantly higher canopy closure than Group 3. The difference between groups for both understory and ground cover was non-significant at p = 0.05 but both variables

41 differed siginificantly at p = 0.1. Understory cover in Groups 1 and 2 were significantly

higher than Group 3, but the ground cover was higher in Group 3 (Table 10).

4.6 Ground vegetation

4.6.1 Seedlings

Seedlings were recorded for 25 tree taxa, though three taxa could not be identified to species. The seedlings that were present for the important canopy and sub canopy species were Shorea robusta, Terminalia tomentosa, Buchanania latifolia, Dillenia pentagyna, Cleistocalyx operculatus, Lagerstroemia parviflora and Mallotus philippensis. Seedling density was 79071 per ha. The highest seedling density was found for S. robusta (70462/ha), B. latifolia (1071/ha), M. philippensis (1500/ha), C. operculatus (571/ha), D. pentagyna (214/ha) and L. parviflora (357/ha). Other less abundant species included Picrasma javanica, Syzygium cumini, Schleichera oleosa,

Careya arborea and Zizyphus sps.

Seedlings of only 11 tree species were found in Group 1, with S. robusta having

the highest density (49166/ha). Seedlings of C. operculatus and D. pentagyna were

absent in Group 1, but P. javanica (2222/ha), Engelhardtia spicata (1388/ha) and B.

latifolia (1111/ha), M. philippensis (555/ha), S. oleosa (555/ha), Zizyphus sps (555/ha), T. tomentosa (277/ha) and L. parviflora (277/ha) were present. In Group 2, seedlings of only 7 tree species were found. S. robusta (61538/ha) was the most abundant species.

Seedlings were absent for T. tomentosa, the tree species with highest IVI in this group.

B. latifolia (2692/ha) and C. operculatus (3076/ha), which were associated with S. robusta in Group 1 and Group 3 respectively, were abundant in the seedling category in

Group 2. Besides these, M. philippensis (1923/ha) and S. oleosa (769/ha) were also

42 present. Seedlings of 14 tree species were present in Group 3 with S. robusta (87432/ha)

as the most dominant species in this group. The importance value of S. robusta in the tree category was exceedingly high for this group compared to other two groups. This was also shown in its seedling density (87432/ha), which is the highest for all the groups.

C. operculatus (405/ha), the second most dominant tree species in this group had lower

seedling density compared to M. philippensis (1891/ha) and T. tomentosa (945/ha). Out

of the 25 species of seedlings recorded, Group 3 had the higher percentages of the total

number of species followed by Groups 1 and 2 (Figure 10). A Kruskal-Wallis test

2 showed that the seedling density was not significantly different between groups (χ 2 =

2.734, p = 0.2549). See Appendix 4b for detailed on density and presence and absence of seedlings in different groups.

The status of regeneration was determined using the following criteria (Shankar

2001): a) ‘good’, if seedling > sapling > trees b) ‘fair’, if seedling > sapling ≤ trees c)

‘poor’, if a species survives in only sapling stage but not as seedlings d) ‘none’, if a

species is absent in both in sapling and seedling stages e) ‘new’ if a species has no adults,

but only saplings or seedlings or both. Following the criteria, S. robusta showed good

regeneration in all the three groups. B. latifolia showed fair regeneration in Group 1,

good regeneration in Group 2 and no regeneration in group 3. T. tomentosa showed fair

regeneration in Group 1 and Group 3, but no regeneration in Group 2. C. operculatus

was regenerating well in Group 2 and Group 3 but not in Group 1. Six species are new to

Groups 1, 2 and 3. Table 12 presents the status of regeneration of different species in the

three groups.

43 4.6.2 Herb layer

Eighty-seven different species were found in the ground layer. Among 65 forbs

(non-graminoid herbs), 18 species were distinguished to be different but could not be

assigned to a genus or species because of the lack of reproductive parts at the time of

sampling. Among 20 species of grasses, seven species could not be identified to the

genus and species level. Two species of sedges, one orchid and one pteridophyta were

identified. Species richness was 4.38 species/plot, evenness was 0.68 and Shannon’s

diversity index was 0.983 (Table 13).

Altogether 53 species of herbs were found in Group 1. Thirty-nine species of

forbs, 14 species of grasses, 2 species of sedges, one orchid and one pteridophyta were present in Group 1. In Group 2, among 38 species of herbs found, 22 were forbs, 13 were grasses, 1 was sedge, 1 was an orchid and 1 was a pteridophyte. Sixty-five species of herbs were present in Group 3. Among them 43 species of forbs, 19 grasses, 1 sedge, 1 orchid and 1 pteridophyte were present. A Kruskal-Wallis test showed that there was no

2 significant difference in per plot species richness between groups (χ 2 = 3.274, p =

0.1945). There was also no significant difference in Shannon’s Diversity index between

2 groups (χ 2 = 1.250, p = 0.5353).

Imperata cylindrica was the most abundant grass species in all three groups.

Evolvulus nummularius and Justicia procumbens were the most abundant forbs in Group

1. Among graminoids, Desmostachya bipinnata, an unidentified grass and Cyperus rotundus (sedge) were also abundant. In Group 2, Ageratum houstoniaum was the most abundant forb followed by E. nummularius and J. procumbens. Desmostachya bipinnata,

Veitveria zizanoides and Barhan (local name) were the abundant grasses after I.

44 cylindrica. In Group 3, Evolvulus nummularius was the most dominant forb followed by

Justicia procumbens, Desmodium sp. and Ageratum houstonianum. An unidentified grass species was the most abundant after I. cylindrica, followed by D. bipinnata, Karonj

(local name), V. zizanoides and C. rotundus. Seedlings of shrubs were treated as forbs.

Altogether 20 species of shrub seedlings were present. Seedlings of C. viscosum, I.

pulchella, F. strobilifera were present in all the three groups. Desmodium contertum was

a new addition in seedlings in all the three groups. Murraya koenigii was found only in

Group 2 and Group 3. Apart from S. parviflorus, seedlings of two other climbers were

present. Millettia fruticosa was found only in Group 2 and Bauhinia vahlii only in Group

3.

4.7 Stand dynamics

To understand stand dynamics properly, one should take temporal variation of the

variables into account, but one-time surveys provide information to formulate hypothesis

that can be tested in the future with sufficient temporal data (Zhang et al. 2005).

During stand development (after disturbance), competition among trees and other

associated vegetation for light, water and nutrients causes mortality. This mortality tends

to be concentrated among smaller and slower growing individuals, and may depend on the density of seedlings and being greatest where seedling density is highest per unit area

(Barnes et al. 1998). The mortality of individuals causes less competition, and remaining individuals become larger as smaller ones are continually removed from the population.

This results into a predictable decrease in the density of trees as the average stand diameter (ASD) increases. This relationship was used to derive SDI, which provides stocking level of stands in any stages of development. Calculation of SDI value provides

45 a standard measure to compare densities of stands with varying ASD. The plot of

Average Stand Diameter against density of trees (stems/ha) labeled by SDI values was used to measure and compare the stocking level of stands (Figure 11). SDI ranges from

50 (Plot 46 in Group 3) to 423 (Plot 5 in Group 1). If SDI 423 is taken to represent full stocking in Sal Forests, then stocking in stand varies from 12% to 96 % of full stocking.

Thirty six of the 70 plots had less than 60 % of full stocking. Among them were 7 out of

18 plots in Group 1, 6 out of 13 plots in Group 2 and 23 out of 37 plots in Group 3.

Stands in Group 1 averaged 68%, Group 2 70% and Group 3 38 % of full stocking.

Seedling establishment and survival in a stand depends upon biotic interactions and environmental factors such as light, nutrients, water and space, which themselves may be affected by other community components. The competition from adults for resources, and their capacity to act as a local seed source also play an important role in seedling establishment. In places where disturbances are frequent and environment is harsh, species are known to regenerate mainly by vegetative means (Charpentier et al.

1998 cited in Pandey and Shukla 2001). I examined the relationship between tree abundance (SDI) and tree size (ASD) with seedling density. It was expected that 1) seedling density would be higher in stands of higher ASD because of the availability of mature trees for seed production, and 2) it would also be higher in stands of lower SDI because of incomplete site occupancy, which means less competition (Zhang et al. 2005).

There was no statistically significant relation between ASD value for plots and seedling density, although there was visually negative relation between seedling density and ASD values (Figure 12). The relationship between SDI value for plots and their seedling density was also non-significant. No trend was observed between SDI value for plots and

46 their seedling density (Figure 13). It shows that seedling establishment is not dependent upon the density and presence and absence of mature trees in the area.

Since availability of trees as a seed source affects the seedling density, I also tested the relation. The relationship was tested by looking at the species abundance in tree layer and its seedling density within each plot. Since S. robusta was the most dominant species in the tree and the seedling category, this relationship was tested for S. robusta only. This was done by calculating relative density of tree and seedling relative density, which was expressed as the percentage of maximum observed for the species in all the transects (Zhang et al. 2005), within each sampling location. The two variables were plotted (Figure 12). No statistically significant relationship was observed.

I also tested canopy-understory relationship between all life forms present in the three layers by looking at the association between percentages of canopy closure, and understory and the ground cover. Spearman’s rank correlation showed a negative relationship between percentage canopy closure and ground percentage cover

(Spearman’s rho = -0.51, p < 0.001), showing the importance of canopy openness for the establishment of sufficient ground vegetation. There was positive correlation between overstory cover and understory cover (Spearman’s rho = 0.6, p < 0.001), which suggests that most of the species in the understory are shade tolerant. Ground cover is also negatively correlated with understory cover (Spearman’s rho = -0.32, p = 0.0056) but the association was not strong as with the overstory.

4.8 Comparison between management regimes

To get a general idea about how the forests differ between management regimes, plots from protected areas (RBNP and RSWR pooled together) were compared with plots

47 from the community forests. Although there were 60 samples from the protected areas

and only ten samples from community forests, the comparison gives some indication of

the differences between the forests that are protected for nearly 30 years and forests under

continuous anthropogenic influences. The results derived from this study can be tested

with sufficient sample sizes in the future. Because of security reasons, sufficient

numbers of plots could not be sampled in the community forests.

Altogether 28 tree species and a climber (Spatholobus parviflorus) were present

in protected areas but only 7 species of trees were present in community forests.

Although comparison of species numbers based on different areas sampled should consider rarefactions, the difference in number of species found in the protected area forests and the community forests gives information on the status of community forests.

Comparison of total number of tree species per plot with a Mann-Whitney U test showed that the protected area forests had significantly higher number of species per plot than the community forests (z = 1.956, p = 0.05). Tree species that were common in protected areas such as Dillenia, Buchanania, Myrsine and Lagerstroemia were completely absent in the community forests. A two sample t-test with an assumption of unequal variance indicated that the tree density in protected areas (245 trees/ha) and community forests (71 trees/ha) were significantly different (t = 5.1574, p = 0.00), but the density of a dominant canopy species, Shorea, was not significantly different. The density of Terminalia was higher in protected areas than community forests (t = 2.60, p = 0.005). A two sample t- test indicated that basal area per ha was not significantly different between protected areas and community forests (t = 0.7689, p = 0.4589).

48 A Mann-Whitney test showed that the sapling density for protected areas (2053

plants/ha) was significantly higher than community forests (268 plants/ha) (z = 3.355, p =

0.0008). Community forests had only three species of saplings (Mallotus, Zizyphus sp.

and Kaphale) whereas saplings of 17 species were present in the protected areas. Sapling

species richness per plot for the protected area forests was significantly higher than the

community forests (t = 2, p = 0.02). Although density of shrubs in protected areas (374

plants/ha) was higher than community forests (128 plants/ha), the difference was not

significant at alpha = 0.05. Altogether 10 different species of shrubs were present in the

protected forests compared to a species in the community forests. F. strobilifera (181

plants/ha) was the most abundant shrub in protected areas. C. viscosum (128 plants/ha) was the only shrub species present in community forests and was abundant in community forests compared to protected forests (36 plants/ha).

Seedling density per ha was compared with a Mann- Whitney test of difference.

The test showed that seedling density in protected areas (84000 plants/ha) was significantly higher than the community forests (49500 plants/ha) (z = 2.50, p = 0.01).

Seedlings of 21 species were present in protected areas with S. robusta (75334 plants/ha) as the most abundant species followed by M. philippensis (1417 plants/ha) and B. latifolia (1250 plants/ha). The community forests had seedlings of five species only: S. robusta, M. philippensis, T. tomentosa, Zizyphus sp. and H. pubescens. The seedling density of the dominant tree species, S. robusta, was significantly higher in protected areas than the community forests (42500 plants/ha), but the densities of Mallotus and

Terminalia were higher in community forests compared to the protected areas.

49 In protected areas, 79 different species were recorded in the ground layer, whereas

only fifteen species were recorded in community forests. Protected areas had 61 species of forbs, 15 species of grasses, a sedge, an orchid and a fern. In community forests, 9 species of forbs, 4 species of grasses and a sedge were recorded. Density of ground vegetation was significantly greater in community forests (679500 plants/ha) than protected areas (393584 plants/ha). Forbs such as Oldenlandia corymbosa, Lippia nodiflora and Sida sp. and the grass Dactyloctenium aegypticum, were exclusively present in the community forests. Imperata cylindrica was the most abundant grass species in the protected areas. In the community forests, an unidentified grass was the most abundant followed by Imperata cylindrica. Cyperus rotundus, the only sedge found in the study, was present in both community forests and the protected areas.

A Mann-Whitney test was performed to see the difference in cover percentages for protected forests and community forests. The percentage canopy closure was significantly higher in protected forests (67%) than community forests (55%; z = 2.610, p

= 0.009). The percentage understory cover for protected forests (32%) was significantly higher than community forests (5%; z = 3.481, p = 0.0005) and the ground cover percentage for the protected forests (56%) were also significantly higher than in the community forests (39%; z = 2.004, p = 0.045).

5. DISCUSSION

Shorea robusta (Sal) is considered climatic climax (Champion and Seth 1968).

Sal is an extremely gregarious species (Champion and Seth 1968) and rarely occurs as a

component of other forest types (Stainton 1972). This forest is not rich in other

associated species and epiphytes and climbers are rare (Stainton 1972). The present

50 study recorded 28 species of trees and one species of climber, and a lower overstory

Shannon Wienner Diversity index compared to similar forests in India (Singh et al. 1995,

Pandey and Shukla 2003).

Density of trees (220 trees/ha) was low, which also indicates the openness of the

sampled forests. The openness of these forests is further demonstrated by the SDI results,

which showed that 53% of plots have stocking percentages of less than 60% of full

stocking. Singh et al. (1995), in their study on the woody vegetation of Corbett National

Park, found density to be higher in the S. robusta-dominated communities and lowest in the Anogeissus latifolia-Acacia catechu community; density ranged from 197 to 728 trees/ha. Density of trees of Shorea robusta forest in RBNP was 348 stems/ha (Shrestha and Jha 1997) and the Sal Forest in Gorakhpur, India had 408 trees/ha for the trees ≥ 30 cm dbh (Pandey and Shukla 2003). The stem density of more than 80 year old pure Sal

Forest in the Bhabar-Terai zone of Nepal ranged from 152 to 264 trees/ha (Rautiainen

1999), which is consistent with the findings of the present study. The present forest stands appeared to be mature forest as shown by their low density, tall canopy and presence of trees in the larger diameter classes (>50 cm).

Average basal area of 13 sq m/ha was low compared to Terai Shorea robusta

forest (36 sq.m/ha) of RBNP (Shrestha and Jha 1997). The upper limit for the basal area

(22 sq m/ha) was within the range for Corbett National Park (Singh et al. 1995), where basal area ranged from 16.0 sq.m/ha to 61.1 sq.m/ha. The lower basal area was due to the presence of a higher number of trees in the lowest dbh class. In the historical past, most of the Terai that extends between the Bhabar and the Indian border were covered with S. robusta as homogenous forest stands (Stainton 1972). However, due to selective logging

51 in the past, influences of burning, overgrazing and indiscriminate cutting of firewood and building timbers, old growth S. robusta forests have been reduced leaving a mixed type of Sal Forests with other tree species such as T. tomentosa (Dinerstein 1979; Shrestha and Jha 1997). In most stands, bigger S. robusta trees have been felled resulting into a change in the proportion of S. robusta to other species. The higher density of trees in the lowest and highest dbh classes, and very few in between indicated that the forest is still recovering from past disturbances. These forests were protected after the 1970s, and trees left after selective logging in the past were found in the larger dbh classes.

Sal forest has been divided into three different types Shorea robusta-Buchanania latifolia forest, Dry Sal Forest and Hill Sal Forest (Dinerstein, 1979; Upreti 1994).

Another Sal forest type; Terai Mixed Shorea robusta (Shrestha and Jha 1997) has been identified in RBNP. Among the three communities identified by the hierarchical cluster analysis, the first is similar to Shorea robusta-Buchanania latifolia community and the second community is similar to the Dry Sal Forest community identified by Dinerstein

(1979) in the southwestern section of RBNP. The first and the second community included most of the plots from RBNP (Bardia District) and only a few plots from RSWR and community forests (Kanchanpur District) sampled. The S. robusta-B. latifolia

(Group 1) association has been described as Dense Sal forest and the S. robusta-T. tomentosa association as Open Sal Forest in RBNP (Sharma 1999). The third community identified was the S. robusta-C. operculatus (Group 3) and included most of the plots from RSWR and community forests in Kanchanpur district. The groups were clearly separated as two groups from Bardia District comprising RBNP and a group comprising

52 plots from Kanchanpur District which included RSWR and community forests. This shows the difference in the Sal forest of RBNP and RSWR.

Groups 1 and 2 differed little in tree density, but Group 3 had lower tree density than other two groups. All three groups had similar values for basal area per ha. Group 3 had the highest density of S. robusta in the largest dbh class (Figure 7) and a higher density of S. robusta trees. It also had greater species representation in lower dbh classes.

It indicates that the pure Sal stands at present might turn to mixed forest in the future, or the survival of other species is lower. The diameter distribution of Shorea robusta for

Group 1 and Group 2 shows fewer trees in the smaller diameter classes and higher densities in the largest diameter class indicating the characteristics of an established pioneer successional community. It is also indicative of disturbance, which may have posed recruitment problems. The presence of S. robusta in both the larger and smaller dbh classes in Group 3 indicates ongoing reproduction. The density of T. tomentosa was lowest in Group 1 and its complete absence from lower dbh classes suggests recruitment problems, whereas its presence in most of the diameter classes in Group 2 and Group 3 indicates ongoing reproduction. In all three groups B. latifolia had higher densities in lower dbh classes and a progressive decline in density in higher dbh classes. A “reverse

J” distribution, with few trees distributed among large dbh classes and many smaller- diameter trees, suggests shade tolerance and continuous recruitment (Sugita et al. 1994).

The distribution of C. operculatus also showed progressive decline in density from lower to higher dbh classes in both Group 1 and Group 3, but it was nearly absent in Group 2.

L. parviflora had a similar distribution as shown by C. operculatus.

53 Comparison of SDI values among the three groups showed that all three groups had low percentage stocking, with Group 3 most understocked. This suggests that the sampled forests are recovering from past disturbances with Group 3 (Kanchanpur

District) having suffered higher disturbance compared to Groups 1 and 2.

Some tree species were confined only to certain associations. Four tree species were exclusively present in Group 1 and Group 2 and one species was exclusively present in Group 3. Based on this, it can be said that Engelhardtia spicata, Ficus benghalensis,

Holarrhena pubescens, and Acacia catechu are the species found in Group 1. Those restricted to Group 2 were Picrasma javanica, Desmodium oojeinense and Paruli (local name). Jingar (local name) was recorded in Group 3 only.

Sal forests are found on better-drained and more developed soils (Dinerstein

1979; Banerjee et al.1992). The ground water table in these forests is very low and inundation does not occur during monsoon (Bolton 1976). Dinerstein (1979) attributed variation in physiognomy and composition of Sal forests to differences in topography, drainage and soil conditions. Topography might be the factor that separates the above described types from the hill Sal forest. Since the forests sampled above were all on flat terrain, the effect of topography is negligible. The soil variables tested in the study were not significantly different among the three communities identified. The pH at my sites

(mean = 5.5, range = 4.8 to 7.5) was in line with the pH of Sal forest in the Bhabar-Terai zone of Nepal (pH = 5.7- 5.9; Rautiainen and Suoheimo 1997), of in the Moist Bhabar

Forest in India (pH = 6.0; Yadav and Sharma 1967) and in S. robusta forests in Uttar

Pradesh (Bhatnagar 1965). The percentage of organic matter (0.59 % to 4.08 %) with mean organic matter of 2.19 %, indicates lower fertility of Sal Forest. The value of 1.7-

54 2.33% of organic matter is an indicator of low fertility (Suoheimo 1995). The percent

organic matter content of the present study is in line with 2.5 % organic matter for Terai

Sal Forest (Shrestha 1992). The low amount of organic matter present may be a result of

good drainage and dry conditions in Sal Forest. Total nitrogen for all three groups was

medium and was similar to Sal forests in Uttar Pradesh, India (Bhatnagar 1965). In his

study, Bhatnagar (1965) didn’t find variation in total nitrogen from quality I to quality IV

Sal Forest in both good and poor Sal regeneration areas. Available phosphorous and

available potassium content was high in all three groups.

The soil variables relationships reported here (Table 4, Figure 9a- 9f) suggest that

soil texture and other chemical properties do not determine the difference in the three

forest communities. Distribution of plant communities is generally determined together

by a wide range of factors, including soil moisture, soil nutrients, rainfall, topographic

position and past disturbances (Barnes et al. 1998). Due to logistic problems, other

variables could not be measured and further studies including a broad range of factors are desirable in the future. Since most of the plots in Group 1 and Group 2 are from Bardia

District and Group 3 contains plots from Kanchanpur District, the amount of rainfall (Fig

1a), which is higher in Bardia (RBNP) compared to Kanchanpur District (RSWR), and its subsequent effects on soil moisture conditions could be one factor. Swaine and Becker

(1999) observed similar results in Ghana, where the amount of rainfall explained the variation in structure and compostion of forests better than the other environmental variables. As previously explained, these forests were exposed to different intensities of disturbance in the past and are still burnt annually, past disturbances and fire could also have shaped these communities. Factors that can influence species distributions are local

55 environmental heterogeneity, stand disturbance history (Clark et al. 1999), mass effects

(Shimda and Wilson 1985), and chance (Hubbell and Foster 1986).

Altogether 17 species were present in the sapling layer. The high density of saplings indicated good regeneration, but the per plot species richness and diversity were lower. This shows less representation of species in the sapling layer. New species are being added to these communities because Cassia fistula, Grewia sp and Kaphale (local name) were present in the sapling layer but not on the tree layer. T. tomentosa, an important tree species, was completely absent in the sapling layer. This may be due to the fact that the seeds of this species have low germinating power (Lal 1992). Fifteen tree species were completely absent in the sapling. It might suggests that Shorea,

Buchanania, Terminalia, Dillenia, Lagerstroemia, Mallotus, Holarrhena, Cleistocalyx,

Semecarpus, Schleichera, Engelhardtia and Zizyphus are the tree species associated with

Sal Forest and other species might have been introduced due to the past disturbances or they are not regenerating.

All three communities identified in the study showed good regeneration in the sapling layer. Sapling density was higher than recorded in Corbett National Park, India, where it varied from 90 to 1240 plants/ha (Singh et al. 1995). S. robusta had the higher density of saplings in all three groups. Less representation of S. robusta sapling in Group

3 compared to Groups 1 and 2 shows lower Shorea sapling density in Sal Forest of

RSWR. B. latifolia was the dominant tree species after S. robusta in Group1, but in the sapling, density of M. philippensis, Engelhardtia spicata, Zizyphus, Grewia sp., C. operculatus were higher than B. latifolia. It might be that the survival rate between sapling and tree stages in other species is lower than B. latifolia. In Group 2 also, the

56 most abundant sapling species after S. robusta was M. philippensis followed by

Buchanania, Cleistocalyx, Zizyphus sp. and Lagerstroemia. Absence of Terminalia saplings and higher density of Buchanania saplings might suggest that Group 2 (T. tomentosa–S. robusta association) is a successional stage which might eventually lead to

Group 1 (S. robusta-B. latifolia association). The dominance of Terminalia in moist deciduous forest indicates presence of moisture- retentive heavy soil (Lal 1992). The absence of Terminalia saplings in Group 2 and increase in Buchanania saplings suggest the loss of moisture retentive capacity probably due to low organic matter and clay content. In Group 3, after S. robusta saplings, Grewia sp., Holarrhena, Cleistocalyx and

Mallotus were the most abundant respectively. Group 3 showed lower regeneration in the sapling layer compared to Group1 and Group 2. Although Group 3 was highly dominated by S. robusta (IVI-178.69) in the tree layer, the sapling layer showed a mixed group of different species. These forests are heavily burnt during the dry season

(Dinerstein 1979; Shrestha and Jha 1997) and burning might have significant impacts on the density of saplings. Group 3 consisted most of the plots from RSWR and community forests in Kanchanpur district, the density of saplings indicated the lower regeneration in these areas. All three associations were equally diverse in saplings, but Group 1 was rich in saplings compared to Group 3.

Seedling density (79072 plants/ha) here was higher compared to seedling density

(2360 to 8591 plants/ha) of Corbett National Park, India and natural forests (31,250/ha) in the Darjeeling Himalayas (Singh et al. 1995; Shankar et al. 1998). S. robusta showed good regeneration in the seedlings (70463 plants/ha) and is close to the S. robusta seedling density (73,542 to 91,125 plants/ha) in Makwanpur District of central Nepal

57 (Rautiainen and Suoheimo 1997). S. robusta was the dominant species in the tree, sapling and seedling categories and represented approximately 90% of all seedling density. Although M philippensis was not as abundant in the tree category, it was well represented in sapling and seedling categories. M. philippensis replaces other species as associates in dry areas (Champion and Seth 1968; Singh and Singh 1992).

The analysis of regeneration status showed that the most dominant species S. robusta was regenerating well in all three groups. B. latifolia which is an associate of Sal in Group 1 showed regeneration in both Groups from Bardia, but it was not regenerating in Group 3 which represents RSWR. T. tomentosa was regenerating in Group 1 and

Group 3, but its lack of regeneration in Group 2 confirms the above-mentioned idea that

Group 2 (Terminalia tomentosa-Shorea robusta association) might transform into Group

1 (S. robusta-B. latifolia association) over time. The addition of six new species in all the

groups shows that the communities studied are not stable.

Many species showed low regeneration and only a few species showed good to

fair regeneration. The species that were associated with Sal in all three groups were

regenerating well except for T. tomentosa in Group 2. The higher dominance of Sal in all

three layers is because of its resistance to fire and potential to grow on a wide range of

soil types (Dinerstein 1979; Rautiainen and Suoheimo 1997). Die-back has been observed

during natural regeneration of S. robusta (Troup 1921; Jackson 1994). Although the shoot portion dies during the recruitment phase, the root system remains alive and continuously sends new shoots every year. Finally a strong root stock develops, which eventually sends a shoot that grows further. Ramet producers such as M. philippensis

(Pandey and Shukla 2001) showed good regeneration in all three groups. H. pubescens,

58 another ramet producer, showed poor regeneration in Group 1, was completely absent in

Group 2 and was new to Group 3. Zizyphus sp. that grows on drier areas was

regenerating well in Group 1 and 3 but not in Group 2. The dominance of T. tomentosa

in the tree layer and poor regeneration of Zizyphus sp. in Group 2 characterizes the

community as developing on moist soil. Adina cordifolia, Syzygium cumini and

Lagerstroemia parviflora increased in abundance after conversion to taungya plantation

in dry mixed deciduous forests of the Darjeeling Himalaya (Shankar et al. 1998). Except

for fire, other kinds of disturbances in the present study areas were rare in the recent past,

not including the 10 plots sampled in community forests. The rarity of species such as A.

cordifolia and S. cumini in the tree layer and their complete absence from the sapling and

seedling layer suggests that the communities sampled are successional forests.

Sal forests sampled in Chitwan had sparse understory (Lehmkuhl 1994). The

understory in the Sal forest remains dense in Bardia and visibility remains poor for most of the year, and it improves only in the hot season after the annual fire burns the understory (Dinerstein 1979). The findings of the present study indicated that shrub density (338/ha) was low in Sal forest. Both species richness and diversity of shrubs were also low. This result is contrary to the findings of Pandey and Shukla (2003) in the

Sal forest of Gorakhpur, India, where the density of a single species was 2669/ha. None of the shrubs they described as having high densities were present in this study area. This may be because sampling in the present study took place after annual burning had occurred. Low species diversity and low density (676/ha) of shrubs in the Terai Shorea robusta Forest of RBNP have been attributed to heavy dry season fires (Shrestha and Jha

1997). Rodgers et al. (1986) didn’t find a significant difference between burnt and

59 unburnt Sal forest in Dehradun India. In their study, the shrub density in burnt forest was

33366/ha and unburnt forest was 27865/ha. Their study used a structural definition of shrub rather than a taxonomical definition. Tree saplings of S. robusta and M. philippensis were regarded as shrub. Flemingia strobilifera was the most abundant species in all three groups. Forty-five percent of the total number of individuals per ha was represented by this species. This species, along with C. viscosum has been regarded as an indicator species of Sal forest (Shrestha and Jha 1997). C. viscosum was absent in

Group 2. The presence of C. viscosum indicates favorable condition for Sal regeneration in India (Champion and Seth 1968). Absence of C. viscosum and dominance of T. tomentosa over Sal in Group 2 indicates poor Sal regeneration.

The three groups did not show significant differences in shrub density, although the mean shrub density was higher in Group 1 followed by Group 2 and Group 3. Only three species were present in Group 2. Group 1 had a higher density than Group 3, but the highest number of species occurred in Group 3. Evenness was lower in Group 2 compared to Group 1 and Group 3. Group 1 species are the most evenly distributed followed by Group 3. Diversity was highest in Group 1, but the difference between the groups was not statistically significant. It can be concluded that all three communities are equally diverse and rich in terms of shrubs, although the overall shrub diversity and richness were low. Fifty-five percent of the plots in Group 1, 39% in Group 2 and 54% in Group 3 didn’t have a single species of shrub. Shrub density, diversity and richness were low because the understory was either open or dominated by saplings. Callicarpa macrophylla and Hedyotis sp. were present in Group 1 only. Elsholtzia blanda,

Pogostemon benghalensis and Grewia sp. were present in Group 3 only. Callicarpa was

60 present in all size classes in the unburnt forest of Dehradun, India but was absent on burnt

sites (Rodgers et al. 1986). The presence of Callicarpa in Group 1 and its absence in other groups might suggest more influence of fire in other Groups than Group 1.

The abundance of Imperata cylindrica among the grasses and Sal among seedlings in all groups suggests the effects of fire in these communities. Frequent fire promotes the growth of fire resistant species such as S. robusta and I. cylindrica (Wesche

1997). Group 3 had the highest total species richness in the herb layer with 63 species followed by Group 1 with 53 species. Group 2 had the lowest species richness with 38 species. Percent canopy closure was significantly different between the Groups, and was the highest for Group 1 and the lowest in Group 3. The high species richness of the herb layer in Group 3 suggests low canopy closure and understory cover, and more light to the species growing on the ground. This was also confirmed by the significant negative correlation between canopy closure and ground cover. Per plot species richness between groups was not significantly different and neither was Shannon’s diversity index. There was not much difference in evenness between the groups. The low Shannon’s diversity index is due to the low evenness in distribution and dominance of a few species such as

Imperata cylindrica, Justicia, Evolvulus and Desmostachya. The density of the herbaceous layer (Table 12) was higher in all groups than the herbaceous layer in natural forest (0.165 million/ha) and forest recovering after taungya plantation (0.098 million/ha) of Mahananda Wildlife Sanctuary, West Bengal, India (Shankar et al. 1998).

Following stand initiation after disturbance, self thining occurs as slower growing individuals are replaced and remaining individuals grow in size. This results into predictable decrease in density as the ASD increases. The relationship between ASD and

61 tree density can be used to calculate SDI, which gives full stocking of stands at any ASD.

Calculation of SDI showed that most of the plots sampled are understocked. Stands in

RSWR are the most understocked among the three groups. The reasons for lower stocking might be due to: (1) insufficient recruitment after disturbances in the past (2) subsequent minor disturbances because of fire (3) the stand structure is not even aged

(Zhang et al. 2005). All the three reasons discussed above could explain the lower stocking of stands. As explained previously, these forests are burned annually. The density distribution among dbh classes suggests the uneven-aged structure of stands.

During the field study a number of herbivores such as Axis axis, Cervus unicolor,

Muntiacus muntjak, Cervus duvauceli, Axis porcinus and Boselaphus tragocamelus were sighted in the protected area forests sampled. Herbivory might have affected recruitment processes.

There was no relation between ASD and SDI values and seedling density suggesting no effect of density (stocking %) and presence and absence of mature trees on seedling density. There was also no relation between the availability of seed source (tree relative density) and seedling density. This might also suggests the countervailing effect of seed source and canopy cover. Higher tree density may act as a sufficient seed source but at the same time increase in canopy cover may prevent seedlings from growing because of low light condition. These results are may be due to the fact that most of the species present in these forests regenerate by non-seed methods as well (Pandey and

Shukla 2001). The die-back phenomenon of Shorea robusta seedlings increases the time of establishment of a new generation up to 60 years under irregular systems and without protection, even if the recruitment of new seedlings were satisfactory (Troup 1921). The

62 longer time for germination, recruitment and establishment of Shorea robusta seedlings might have also affected the relationship, although sufficient seed sources were available.

The community forests sampled were declared under community protection only two years ago. Although User Groups and the boundaries of these forests had already been defined, the official hand-over process is still to come. These community forests have been subjected to illegal cutting of timber, non-timber forest products, collection of fuelwood and fodder, grazing and intentional fires. The author observed cattle grazing during field work. Community forests were species poor and 20 tree species and a climber found in protected areas were completely absent in community forests. This shows that community forests are in a highly-degraded condition and need protection for a number of years to reach the species richness found in protected areas. This is also supported by the fact that the per plot species richness in protected area forests were higher than community forests. In their study in the Nepalese Terai, Webb and Sah

(2003) found that a 20 yr successional Sal forest had 84% of the total species richness of natural forest. It has been found that, after clear cutting, 40 years is required for a tropical rain forest tree community to return to its pre-cut diversity levels (Faber-

Langendoen 1991 cited in Webb and Sah 2003). Community forests had 29% of tree density compared to protected areas. Although density was different, the basal area/ha was not significantly different between community forests and protected areas. This may be due to the fact that the higher percentage of basal area in community forests was represented by trees in larger dbh classes.

Absence of trees in lower dbh classes, lower species richness and density of saplings, lower shrub richness and density and lower density of seedlings suggest that the

63 understory in community forests is highly disturbed and regeneration is low due to

frequent disturbances. Low intensity and sustained human disturbance through selective

logging, firewood extraction, grazing and land clearing for permanent agriculture may

influence plant communities and their successional patterns (Attiwill 1994; Fujiska et al.

1998). A similar study done in and around Royal Chitwan National Park (RCNP), Nepal

found community forests to be in poorer conditions than national forests and national

parks (Nagendra 2002).

The higher density of M. philippensis in the community forests compared to protected areas suggests the impact of disturbance. Species such as M. philippensis can produce ramets and regenerate well in areas of high disturbances (Pandey and Shukla

2001). The presence of good amount of S. robusta seedlings in community forests may be due to the nature of the species in response to fire as described above. The lower ground cover and species richness in community forests compared to protected areas also suggests the impacts of grazing. Grazing was common in these forests and had significant impact on the regeneration and establishment of plant communities in the ground layer. This suggests that community forests sampled were highly degraded compared to protected areas. Adequate protection is necessary to ensure that the structure and composition of community forests resemble that of protected areas.

The community forests had poor species richness, density and cover values for all the layers sampled. The poorer condition along with competition from cattle grazing rule out the possibility of supporting sufficient herbivore population, because of lack of food and sufficient cover. Imperata cylindrica, Vetiveria zizanoides, Saccharum spontaneum and Desmostachya bipinnata are the most important grazing species for ungulates in

64 RBNP (Dinerstein 1979a). Absence of all other species except Imperata in community forests also suggests lack of food. Seedlings, fruits and flowers of tree species such as

Acacia catechu, Bombax ceiba, Ficus benghalensis, Syzigium cumini, Schleichera oleosa

etc. are consumed by herbivores, these species were absent in the community forests.

Adequate prey density (herbivores) and sufficient cover is necessary for good quality

tiger habitat (Karanth and Sunquist 1995; Smith et al. 1998). Different vegetation types

inside protected areas provide diverse habitat for mammals, which they utilize according

to seasons and phenology of plants (Dinerstein 1979a). This kind of diversity is lacking

in community forests. At the present condition, the role of these community forests as

additional wildlife habitats outside protected areas and as movement corridors is bleak.

Ecological conditions of community forests outside protected areas are important for

biodiversity conservation and for long- term success of landscape level programs such as

TAL.

65 TABLES

Table 1: Area (ha) under different land uses in Bardia District.

Land Use Area (hectares) Area (%) Agriculture and Settlements 68,075 33.7 Forest 33,452 16.6 Royal Bardia National Park 87,936 43.6 Scrub and Grass 294 0.1 Rivers (Water Bodies) 11,920 6.0 Total 2,01,677 100.0 Source: HMGN 1996

Table 2: Importance value of trees for the three different associations identified by cluster analysis. Species are coded (see Appendix 1 and 2 for detailed description of tree species, their density, basal area, frequency and relative frequency in different groups).

Species Group 1 Group 2 Group 3 Shorob 81.7 74.2 178.69 Dilpen 24.46 19.62 9.76 Tertom 28.07 81.44 22.56 Buclat 37.55 32.53 4.81 Anolat 3.56 0 1.22 Myrsem 16.88 20.08 1.21 Malphi 4.63 9.24 6.92 Lagpar 19.84 13.68 4.74 Schole 4.86 2.05 1.39 Cleope 29.39 2.42 27.32 Adicor 1.58 5.07 4.52 Syzcum 2.67 3.3 3.61 Antchi 2.2 4.09 0.28 Semana 8.25 4.41 5.77 Terbel 1.4 4.49 3.23

66 Table 3: Basal area (m2 ha-1) of the different tree species among groups

Species Groups Obs Mean SD Min Max Shorob 1 18 6.25 2.56 2.29 11.48 2 13 5.12 1.79 2.29 8.04 3 37 10.89 3.55 2.29 19.52 Dilpen 1 18 0.83 0.95 0.00 2.29 2 13 0.53 0.76 0.00 2.29 3 37 0.25 0.72 0.00 3.44 Tertom 1 18 1.47 1.23 0.00 3.44 2 13 4.50 1.28 2.29 6.88 3 37 0.81 1.26 0.00 4.59 Buclat 1 18 0.89 1.22 0.00 3.44 2 13 0.79 1.18 0.00 3.44 3 37 0.12 0.53 0.00 2.29 Myrsem 1 18 0.32 0.53 0.00 1.14 2 13 0.53 0.89 0.00 2.29 3 37 0.03 0.19 0.00 1.15 Lagpar 1 18 0.45 0.69 0.00 2.29 2 13 0.35 0.72 0.00 2.29 3 37 0.06 0.26 0.00 1.15 Cleope 1 18 1.02 1.52 0.00 4.59 2 13 0.09 0.32 0.00 1.15 3 37 0.40 0.67 0.00 2.29

Table 4: Mean, minimum and maximum values for soil variables tested for three different associations.

Variables Group 1 Group 2 Group 3 Mean Min Max Mean Min Max Mean Min Max pH 5.5 5 6.5 5.6 5 6.5 5.4 4.8 7.5 OM (%) 2.28 1.15 3.45 2.18 0.59 4.08 2.15 0.94 3.61 N (%) 0.114 0.06 0.17 0.108 0.03 0.2 0.107 0.05 0.18 P205 114.64 32.67 212.36 83.77 21.77 185.14 132.5 3.14 370.27 (kg/ha) K20 290.04 104.79 445.37 271.44 129.46 406.08 271.47 115.83 456.51 (kg/ha)

67 Table 5: Mean sapling density per ha, Species richness per plot (S), Shannon’s Diversity Index (H’) and Evenness (E) for all forests sampled and the three different associations identified by cluster analysis

Groups Mean Min Max S H’ E All sampled forests 1797 0 10063 2.4 0.56 0.55

Group 1 2851 127 9935 3.2 0.69 0.59

Group 2 2733 0 10063 2.3 0.52 0.51

Group 3 1019 0 5732 2.0 0.52 0.55

Table 6: Mean sapling density of different species per ha in different groups. Species are coded (see Appendix 4 for detailed description of species).

Group 1 Group 2 Group 3 Species Mean SD. Species Mean SD. Species Mean SD Shorob 2045 2589 Shorob 1969 3122 Shorob 278 1139 Lagpar 56 89 Lagpar 78 176 Lagpar 24 84 Buclat 77 132 Buclat 137 204 Cleope 79 148 Cleope 99 135 Cleope 107 293 Picjav 6 29 Picjav 14 41 Picjav 29 76 Malphi 82 175 Malphi 134 372 Malphi 205 741 Holpub 210 475 Holpub 56 186 Zizphy 88 190 Zizphy 89 143 Zizphy 120 328 Baslat 29 105 Grewia sp 213 559 Baslat 113 289 Grewia sp 9 35 Schole 3 21 Semana 21 48 Schole 9 35 Syzcum 3 21 Grewia sp 92 221 Casfis 29 106 Kaphal 27 68 Schole 7 30 Dilpen 9 35 Casfis 7 30 Kaphal 29 106 Kaphal 7 30

68 Table 7: Kruskal-Wallis and Mann-Whitney results showing the difference in sapling density of different species among groups.

Species Kruskal Wallis test Mann Whitney test of difference Shorob χ2 =15.983, p = 0.0003 Group1>Group3, Group2 > Group3 Lagpar ** Cleope ** Picjav ** Malphi ** Zizphy ** Schole ** Kaphal* ** * Unidentified (Local name), ** Insignificant.

Table 8: Mean shrub density per ha, Species richness per plot (S), Shannon’s diversity index (H’) and evenness (E) in all forests sampled and in different groups.

Groups Mean shrub density SD Min Max S H’ E All 337.18 640.79 0.00 3057.12 0.69 0.11 0.14 forests sampled 1 467.06 839.28 0.00 2420.22 0.83 0.346 0.23 2 362.54 418.76 0.00 1146.42 0.80 0.085 0.077 3 265.09 599.26 0.0 3057.12 0.60 0.082 0.118

69 Table 9: Comparison between groups in terms of shrub density, Shannon’s diversity and species richness.

Kruskal-Wallis test Mann-Whitney test Mean shrub density per hectare χ2 = 5.728, p = 0.05 Gr1 < Gr2 > Gr3, Gr1 > Gr3 Mean Shannon diversity per plot χ2 = 0.0570, p = 0.5909

Species richness per plot χ2 = 0.4480, p = 0.7992

Table 10: Mean, minimum and maximum cover percentages for overstory, understory and ground cover of three groups identified by cluster analysis.

Cover types Group 1 Group 2 Group 3

Mean Min Max Mean Min Max Mean Min Max

Overstory 68.9 33.18 82.4 69.92 44.36 79.46 62.57 32.66 78.68

(Trees)

Understory 38.17 0.5 83 34.77 0.5 83 19.97 0.5 83

(Saplings

and Shrubs)

Ground 43.69 10 83 49.38 12.5 83 57.17 6.25 83

(Herbs)

70 Table 11: Comparison of cover percentages between the groups identified by cluster analysis.

Cover Kruskal Wallis test Mann Whitney

Overstory χ2 = 10.533, p = 0.005 Gr 1& Gr 2*, p = 0.857

Gr1& Gr 3, p = 0.005

Gr 2 & Gr 3, p = 0.017

Understory χ2 = 5.081, p = 0.078

Ground χ2 = 5.431, p = 0.0662

*Significant, Gr= Group

Table 12: Status of regeneration of different species in the three Groups.

Status Group 1 Group 2 Group 3

Good 16.6% 15% 18%

Fair 12.5% 15% 18%

Poor 12.5% 10% 4.5%

No 50% 55% 41%

New 6 sps 6 sps 6 sps

71 Table 13: Species richness, Evenness, Shannon’s diversity index and Density/ha of herbs for the three associations and all sampled forests

Group 1 Group 2 Group 3 All forests

sampled

Species richness per plot 4.48 3.96 4.49 4.38

Evenness 0.73 0.69 0 .65 0.69

Shnannon’s diversity (H) 1.04 0.91 0.97 0.98

Density 314722.2 482692.3 470000 434428.6

72 FIGURES

Figure 1a: Average monthly precipitation and average annual precipitation for Bardia District (RBNP) and Kanchanpur District (RSWR). Data were taken for fifteen years, from 1987 to 2001

2500

2000 mm)

( 1500 n RBNP io t a

t RSWR i 1000 ip c e r

P 500

0 l t r r v n n b g p y c y l r Ju Ja Ap Ju Oc No a Fe Ma Au Se De Ma Ye

Figure 1b: Mean monthly maximum temp (0C) and mean monthly minimum temp (0C) for Bardia District (RBNP) and Kanchanpur District (RSWR). Data were taken for fifteen years, from 1987 to 2001. Source: Department of Hydrology and Meteorology, Kathmandu.

RBNP Max 40 RBNP Min 35 RSWR Max 30 RSWR Min C

e 0 25 r

atu 20 er p 15 m e

T 10 5

0 l t r r v n n b g p y c Ju Ja Ap Ju Oc No Fe Ma Au Se De Ma

73 Figure 2: Map of the study area

74 Figure 3: Diameter distribution of trees (> 5cm dbh) for RBNP, RSWR and the Community Forests sampled.

Cleistocalyx operculatus 90 Lagerstroemia parviflora 85 80 My r s ine s emis er r ata 75 Buchanania latifolia 70 Terminalia tomentosa

) 65 a

h 60 Dillenia pentagyna /

m 55

e Shorea robusta t 50 s (

y 45 t i

s 40 n

e 35

d 30

em 25

st 20 15 10 5 0 5-10cm 10-15cm 15-20cm 20-25cm 25-30cm 30-35cm 35-40cm >40cm dbh classes

.

75 Figure 4: Dendrogram showing the different associations identified by the hierarchial agglomerative cluster analysis based on importance value of trees. Groups are described in the text.

0 7.2E-01

100 75 Plot 1 Plot 20 Plot 37 Plot 64 Plot 31 Plot 42 Plot 2 Plot 3 Plot 5 Plot 6 Plot 26 Plot 21 Plot 8 Plot 25 Plot 15 Group 1 Plot 10

Group 2

Group 3

76 Figure 5: Diameter distribution of trees (> 5 cm dbh) for Shorea robusta-Buchanania latifolia association (Group 1).

120

110 Cleistocalyx operculatus 100 Lagerstroemia parviflora Myrsine semiserrata 90

) Buchanania latifolia a

/h 80 Terminalia tomentosa s m

e 70 Dillenia pentagyna t s

( 60 Shorea robusta y it s

n 50 e

d 40 m e t

s 30 20

10 0 5-10cm 10-15cm 15-20 cm 20-25 cm 25-30cm 30-35cm 35-40cm >40 cm

dbh classes

77 Figure 6: Diameter distribution of trees (> 5 cm dbh) for Terminalia tomentosa-Shorea robusta association (Group 2).

130 Cleistocalyx operculatus 120 Lagerstroemia parviflora 110 Myrsine semiserrata 100

a Buchanania latifolia 90 /h

s Terminalia tomentosa

m 80 e t Dillenia pentagyna s (

70

y Shorea robusta it 60 ns

de 50 m e

t 40 s 30

20

10

0 5-10cm 10- 15-20 20-25 25- 30- 35- >40 cm 15cm cm cm 30cm 35cm 40cm dbh classes

78 Figure 7: Diameter distribution of trees (> 5 cm dbh) for Shorea robusta-Cleistocalyx operculatus association (Group 3).

150 140 Cleistocalyx operculatus 130 Lagerstroemia parviflora 120 Myrsine semiserrata 110 Buchanania latifolia a)

h 100 Terminalia tomentosa / s Dillenia pentagyna m 90 e t

s 80 Shorea robusta y ( t i

s 70 n e 60 d

em 50 t s 40 30

20 10 0 5-10cm 10- 15-20 20-25 25- 30- 35- >40 cm diameter classes 15cm cm cm 30cm 35cm 40cm

79 Figure 8: Site scores from 2 axis non-metric multidimensional scaling (NMS) ordination, based on importance value of trees.

52 40 5836 51 56 4539 4453 57 59 63 6046 54 35 38 34 6513 0.5687 62 55 49 4 61 41 69 47 30 66 31 0.3279 9 67 42 48 5 68 14 50 0.0022 18 10 12 2 364 26 20 -0.2966 37 6 25

1 21 1 7 32 s

i -0.5738 8 29 Ax 16 23 -0.8906 70 19 33 15 17 -1.1307 28

24 -1.4930

27 -1.8001

11 -2.1611 S. robusta-B. latifolia T. tomentosa-S. robusta -1.7661 -1.2233 -0.6849 -0.1386 0.3667 0.8796 S. robusta-C. operculatus -1.4784 -0.9445 -0.4507 0.1134 0.6034 Axis2

80 Figure 9a: Relationship between different associations identified by cluster analysis and soil pH. Contour plot of pH is superimposed upon NMS ordination of trees. (▲ Shorea robusta-Cleistocalyx operculatus, ■ Terminalia tomentosa-Shorea robusta, + Shorea robustaBuchanania latifolia).

7.5 7.4 7.3 7.2 7.1 7 6.9 6.8 6.7 6.6 6.5 6.4 6.3 6.2 6.1 6 5.9 5.8 5.7 5.6 5.5 5.4 5.3 5.2 5.1 5 4.9 4.8pH

81 Figure 9b: Relationship between different associations identified by cluster analysis and soil organic matter. Contour plot of soil organic matter is superimposed upon NMS ordination of trees. (▲ Shorea robusta-Cleistocalyx operculatus, ■ Terminalia tomentosa-Shorea robusta, + Shorea robusta-Buchanania latifolia).

0.5

0 3.6

3

-0.5

2.4

-1 1.8

1.2

-1.5

0.6 Organic Matter

-2

-1.5 -1 -0.5 0 0.5

82 Figure 9c: Relationship between different associations identified by cluster analysis and total nitrogen. Contour plot of total nitrogen is superimposed upon the NMS ordination of trees. (▲ Shorea robusta-Cleistocalyx operculatus, ■ Terminalia tomentosa-Shorea robusta, + Shorea robusta-Buchanania latifolia).

0.5

0.19 0 0.17

0.15

-0.5 0.13

0.11

-1 0.09

0.07

0.05 -1.5

0.03

Nitrogen -2

-1.5 -1 -0.5 0 0.5

83 Figure 9d: Relationship between different associations identified by cluster analysis and available phosphorous. Contour plot of available phosphorous is superimposed upon NMS ordination of trees. (▲ Shorea robusta-Cleistocalyx operculatus, ■ Terminalia tomentosa-Shorea robusta, + Shorea robusta-Buchanania latifolia).

0.5

363 0 343 323 303 283 263 -0.5 243 223 203 183 -1 163 143 123 103 83 -1.5 63 43 23 3

-2 Phosphorous

-1.5 -1 -0.5 0 0.5

84 Figure 9e: Relationship between different associations identified by cluster analysis and available potassium. Contour plot of available potassium is superimposed upon NMS ordination of trees. (▲ Shorea robusta-Cleistocalyx operculatus, ■ Terminalia tomentosa-Shorea robusta, + Shorea robusta-Buchanania latifolia).

0.5

440 420 0 400 380 360 340 320 -0.5 300 280 260 240 220 -1 200 180 160 140 -1.5 120 100

Potassium -2

-1.5 -1 -0.5 0 0.5

85 Figure 9f: Relationship between different associations identified by cluster analysis and soil texture. Contour plot of soil texture is superimposed upon NMS ordination of trees. (▲ Shorea robusta- Cleistocalyx operculatus, ■ Terminalia tomentosa- Shorea robusta, + Shorea robusta- Buchanania latifolia).

0.5

6.5 6 0 5.5 5 4.5 4 -0.5 3.5 3 2.5 2 -1 1.5 1 0.5 0 -1.5 -0.5 -1 Soil texture

-2

-1.5 -1 -0.5 0 0.5

86 Figure 10: Seedlings of tree taxa present in each group as the percentage of total tree taxa recorded in seedlings.

60%

50%

40% ge a t

n 30% e c r 20% pe

10%

0% Group 1 Group 2 Group 3 Groups

87 Figure 11: Log density of trees (stems ha-1) plotted against log Average Stand Diameter (ASD) for all the plots sampled. Plots are labeled by Stand Density Index. The diagonal line asserts the maximum stocking for the forest sampled. (Axes are in linear scale)

408 990 356 404

402424 695 396 407 475 353 254 288287286306305326 184 283 320353586 221221 281 385 266 234 ) 287

a 264 278 h 241 375 272

s/ 270 183 216

m 262

e 241 130 301 st 219 217233249262852960 y(

t 211226

i 194 176 252531 s 90 189 220 n 155 199 e 151166

D 163 64 162 205 130 174 209 45 109 177

117 28 129 20 83

50 12 87 Group 1 Group 2 10 15 20 25 30 40 50 63 80 100 Group 3 Average Stand Diameter (ASD)

88 Figure 12: Plot of total seedling density against SDI and ASD.

89 Figure 13: Scatter plots of Sal seedling density vs Sal tree relative density for Sal Forest in the western Terai of Nepal.

SRD= 17.7106-0.0086*TRD r2 = 0.0000

90 y t i 70 dens e v ti a l

e 50 ng r i eedl S 30

10

0 20406080100 Tree relative density

90 REFERENCES

Ashton, P. S. 1971. The plants and vegetation of Bako National Park. Malaysian Natural Journal 24: 151-162.

Attiwill, P.M. 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and Management 63: 247-300.

Austin, M.P., J.G. Pausas, and A.O. Nicholls. 1996. Patterns of tree species richness in relation to environment in southeastern New South Wales, Australia. Australian Journal of Ecology 21: 154-164.

Banerjee, S.K., S. Nath, A. Mukherjee, and D. Namhata. 1992. Ecological status of Shorea robusta in Lateritic region. Indian Journal of Forestry 15: 38-43

Baral, N., B. Tamnag, and N. Timilsina. 2002. Status of Bengal Florican Houbaropsis begalensis in Royal Bardia National Park, Nepal. Journal of Bombay Natural History Society 99 (3): 413-417.

Barnes, B.V., D. R. Zak, S. R. Denton, and S. H. Spurr. 1998. Forest Ecology. 4th edition. John Wiley & Sons, New York.

Basnet, K. 1995. Biodiversity inventory of the Royal Bardia National Park. Department of National Parks and Wildlife Conservation and WWF Nepal Program, Kathmandu, Nepal

Belsky, A. J. 1987. Regeneration of natural and human-caused disturbances in the Serengeti National Park, Tanzania. Vegetatio 70: 51-60.

Bhatnagar, H. P. 1965. Soils from different quality Sal (Shorea robusta) forests of Uttar Pradesh. Tropical Ecology 6:56-62.

Bolton, M. 1976. Royal Karnali Wildlife Reserve Management Plan 1976-1981. UNDP/FAO, Kathmandu, Nepal.

Boutton, T. W., L. L. Tiezen, and S. K. Imbamba. 1988. Biomass dynamics of grassland vegetation in Kenya, East Africa. African Journal of Ecology 26: 89-101.

Bray, R. H., and L. T. Kurtz. 1945. Determination of total, organic and available forms of phosphorous in soils. Soil Science 59: 39-45.

Carlson, D. W., and A. Groot. 1997. Microclimate of clear-cut, forest interior, and small openings in trembling aspen forest. Agric. For. Meteorol. 87: 313-329.

91 Chakraborty, R.N. 2001. Stability and outcomes of common property institutions in forestry: evidence from the Terai region of Nepal. Ecological Economics 36: 341- 353.

Champion, H.G., and S. K. Seth. 1968. Revised Survey of Forest Types of India. Manager of Publications, India.

Chaudhary, R. P. 1998. Biodiversity in Nepal: Status and conservation. Craftsman press, Bangkok, Thailand.

Chaudhary, R. P. 2000. Forest conservation and environmental management in Nepal: a review. Biodiversity and Conservation 9: 1235-1260.

Clark, D. B., M. W. Palmer, and D. A. Clark. 1999. Edaphic factors and the landscape- scale distributions of Tropical Rain Forest Trees. Ecology 80(8): 2662-2675.

Clark, D.B., and D.A. Clark.2000. Landscape-scale variation in forest structure and biomass in a tropical rain forest. Forest Ecology and Management 137: 185-198.

Cornet, A.F., C. Montana., J.P. Delhoume and L. Lopez-Portillo.1992. Water flows and the dynamics of desert vegetation stripes. Pages 327-345 in A.J. Hansen and F. DiCastri, editors. Landscape boundaries: Consequences for Biotic Diversity and Ecological Flows. Ecological Studies 92. Springer Verlag, Berlin.

Daniel, T. W., J. A. Helms, and F.S. Baker. 1979. Principles of silviculture. 2nd edition. McGraw-Hill, New York.

Dinerstein, E. 1979. An ecological survey of the Royal Karnali- Bardia Wildlife Reserve, Nepal. Part I: Vegetation, modifying factors, and successional relationships. Biological Conservation 15: 127-150.

Dinerstein, E. 1979a. An ecological survey of the Royal Karnali-Bardia Wildlife Reserve, Nepal. Part II: habitat/ animal interactions. Biological Conservation 16: 265-300.

Dobremez, J.F. 1972. Les Grandes Divisions Phytogeographiques du Nepal et de l’Himalaya (The Major Phytogeographical Regions of Nepal and the Himalayas). Bull. Soc. Bot. France, 119 (1/2): 111-120.

Dobremez, J. F. 1976. Le Nepal: ecologie et biogeographie. Editions de la Recherche Scientifique, Paris.

Forman, R. T.T., and M. Godron. 1986. Landscape Ecology. John Wiley and Sons, New York.

92 FRISP. 1994. Deforestation in the terai districts 1978/89-1990/91. Publication no 60. Ministry of Forest and Soil Conservation, Kathmandu, Nepal.

FSD/FORESC. 1993. Forest Resources of the Terai Districts 1990/1991. Publication no. 57. Ministry of Forest and Soil Conservation, Kathmandu, Nepal.

FSRO. 1971. Soil survey of Bardia Forest Division. Department of Forests, Kathmandu, Nepal.

Fujisaka, S., G. Escobar, and E. Veneklass. 1998. Plant community diversity relative to human land uses in Amazon forest colony. Biodiversity and Conservation 1: 41- 57.

Gentry, A. H. 1982. Patterns of neotropical plant species diversity. Evolutionary Biology 15: 1-84.

Green, D.M., and J.B. Kauffman. 1995. Succession and livestock grazing in a northeastern Oregon riparian ecosystem. Jouranl of Range Management 48: 307- 313.

Greig-Smith, P. 1983. Quantitative plant ecology. 3rd edition. Blackwell Scientific, Oxford.

Grosenbaugh, L. 1952. Plotless timber estimates-new, fast, and easy. Journal of Forestry 50: 32-57.

Hall, J. B., and M. D. Swaine. 1981. Distribution and ecology of vascular plants in a tropical rain forest: forest vegetation in Ghana. Junk, Boston.

Hara, G., A. O. Chater, and L. H. J. Williams. 1982. An Enumeration of the Flowering Plants of Nepal. Vol 3. British Museum of Natural History, London, UK.

Hara, G., and L. H. J. Williams. 1979. An Enumeration of the Flowering Plants of Nepal. Vol 2. British Museum of Natural History, London, UK.

Hara, G., W. T. Stearn, and L. H. J. Williams. 1978. An Enumeration of the Flowering Plants of Nepal. Vol 1. British Museum of Natural History, London, UK.

Harcombe, P.A., and P.L.Marks. 1977. Understory structure of mesic forest in South-east Texas. Ecology 58: 1144-1151.

Hardin, G. 1968. The tragedy of the commons. Science 162: 1243-1248.

93 Hart, T. B., J. A. Hart, and P. G. Murphy. 1989. Monodominant and species-rich forests of the humid tropics: causes of their co-occurrence. The American Naturalist 133 (5): 613-633.

Heinen, J.T., and S. Shrestha. 2005. Evolving policies for conservation: an historical profile of the protected area system of Nepal. Journal of Environmental Planning and Management (in press).

Hesse, P. R. 1994. A textbook of soil chemical analysis, Chemical Publishing Co-Inc, USA.

Holdridge, L. R. 1967. Life Zone Ecology. Tropical Science Center, San Jose, Costa Rica.

HMGN/IUCN (His Majesty’s Government of Nepal/ World Conservation Union). 1998. The national conservation strategy for Nepal: Building on success. His Majesty’s Government of Nepal and IUCN Nepal, Kathmandu.

HMGN (His Majesty’s Government of Nepal). 1973. National Park and Wildlife Conservation Act 1973. His Majesty’s Government of Nepal, Ministry of Forests and Soil Conservation, Kathmandu.

HMGN. 1976. Catalogue of Vascular Plants. Bulletin Department of Medicinal Plants Nepal No. 7, His Majesty’s Government of Nepal, Ministry of Forests & Soil Conservation, Kathmandu.

HMGN. 1988. Master Plan for the Forestry Sector. Main Report. Ministry of Forests and Soil Conservation, Kathmandu, Nepal.

HMGN. 1993. Forest Act 1993. His Majesty’s Government of Nepal, Ministry of Forests and Soil Conservation, Kathmandu.

HMGN. 1996. Operational Forest Management Plan For Bardia District (1996/1997- 2000/2001). His Majesty’s Government of Nepal, Department of Forests, Kathmandu.

HMGN. 1998. Operational Forest Management Plan For Kanchanpur District (1998- 2002/2003). His Majesty’s Government of Nepal, Department of Forests, Kathmandu.

HMGN. 2001. Royal Bardia National Park: Management Plan 2001-2005. Department of National Parks and Wildlife Conservation, Kathmandu, Nepal.

HMGN. 2004. Terai Arc Landscape-Nepal: Strategic Plan 2004-2014. Ministry of Forests and Soil Conservation, Kathmandu, Nepal.

94

Hubbell, S. P., and R. B. Foster. 1986. Biology, chance, and history and the structure of tropical rain forest tree communities. Pages 314-329 in J. Diamond and T.J.Case, editors. Community ecology. Harper and Row, New York, USA. Hunter, M. L., and P. Yonzon. 1993. Altitudinal distributions of birds, mammals, people, forests and parks in Nepal. Conservation Biology 7: 420-423.

Inskipp, C. 1989. Nepal’s Forest Birds: Their status and conservation. International Council for Bird Preservation. Monograph no.4. Cambridge, UK.

Jackson, J.K. 1994. Manual of Afforestation in Nepal, 2nd edition. Forest Research and Survey Center, Kathmandu, Nepal.

Jackson, M.L. 1958. Soil Chemical Analysis. Prentice Hall, Englewood Cliffs, NJ, USA.

Janzen, D. H. 1974. Tropical blackwater rivers, animals and mast fruiting by Dipterocarpaceae. Biotropica 6: 69-103.

Jnawali, S. R., and P. Wegge. 1993. Space and habitat use by a small re-introduced population of greater one-horned rhinoceros (Rhinoceros unicornis) in Royal Bardia National Park. Pages 208-217 in O. A. Ryder, editor. Rhinoceros Biology and Conservation. Proceedings of an International Conference. San Diego Zoological Society, San Diego, USA.

Joshi, A. R. 2002. Establishing biological monitoring system for Terai Arc Landscape. A report submitted to World Wildlife Fund (WWF) Nepal program. WWF Nepal Program, Kathmandu, Nepal.

Kanel, K.R., and K. Shrestha. 2001. Tropical secondary forests in Nepal and their importance to local people. Journal of Tropical Forest Science 13(4): 691-704.

Karanth, K. U., and M. E. Sunquist. 1995. Prey selection by tiger, leopard and dhole in tropical forests. Journal of Animal Ecology 64: 439-450.

Kozlowski, T., P. J. Kramer, and S. G. Pallardy. 1991. The physiological ecology of woody plants. Academic Press, New York.

Lal, J. B. 1992. Forest Ecology. Natraj Publishers, Dehradun, India.

Laurance, W.F., P.M. Fearnside , S.G. Laurance, P. Delamonica, T.E. Lovejoy, J.M. Rankin-de-Merona, J.Q. Chambers, and C. Gascon. 1999. Relationship between soils and Amazon forest biomass: a landscape study. Forest Ecology and Management 118: 127-138

95 Laurance, W.F., S.G. Laurance, L.V. Ferreira, J.M. Rankin-de- Merona, C. Gascon, and T.E. Lovejoy. 1997. Biomass collapse in Amazonian forest fragments. Science 278: 1117-1118.

Lehmkuhl, J. F. 1994. A classification of subtropical riverine grassland and forest in Chitwan National Park, Nepal. Vegetatio 111: 29-43.

Lugo, A.E., and F. N. Scatena. 1996. Background and catastrophic tree mortality in tropical moist, wet and rain forests. Biotropica 28: 585-599.

Marks, P.L. 1974. The role of pin cherry (Prunus pennsylvanica L.) in the maintenance of stability in northern hardwood ecosystems. Ecological Monograph 44: 73-88.

McCune, B., and J. B. Grace. 2002. Analysis of ecological communities. MjM Software Design, Gleneden Beach, OR, USA.

McCune, B., and M. J. Mefford. 1999. PC-ORD. Multivariate Analysis of ecological data. Version 4.0. MjM Software, Gleneden Beach, OR, USA.

Minchin, P.R. 1987. An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69: 89-107.

Nagendra, H. 2002. Tenure and forest conditions: community forestry in the Nepal Terai. Environmental Conservation 29(4): 530-539.

Neupane, H. P. 2000. A grassroots organizer with a commitment to gender-balanced participation: an interview with the Hari Prasad Neupane, Chairperson, Federation of Community Forestry User in Nepal. Mountain Research and Development 20: 316-319.

Ogutu, Z. A. 1991. Human disturbances and vegetation dynamics in the Narok district, Kenya. Ph. D Thesis. Edinburgh University, UK.

Ogutu, Z. A. 1996. Multivariate analysis of plant communities in the Narok district, Kenya: The influence of environmental factors and human disturbance. Vegetatio 126: 181-189.

Ostrom, E. 1990. Governing the commons: The Evolution of Institutions for Collective Action. Cambridge University Press, New York, USA.

Pandey, S. K., and R. P. Shukla. 2001. Regeneration strategy and plant diversity status in degraded sal forests. Current Science 81(1): 95-102.

96 Pandey, S. K., and R. P. Shukla. 2003. Plant diversity in managed sal (Shorea robusta Gaertn.) forests in Gorakhpur, India: species composition, regeneration and conservation. Biodiversity and Conservation 12: 2295-2319

Peterson, E.B., and B. McCune. 2001. Diversity and succession of epiphytic macrolichen communities in low-elevation managed conifer forests in Western Orgeon. Journal of Vegetation Science 12: 511-524.

Pielou, E. C. 1966. Shannon’s formula as a measure of specific diversity: its use misuse. American naturalist 100: 463-465.

Prentice, I. C., W. P. Cramer, S. P. Harrison, R. Leemans, R. A. Monserud, and A. M. Solomon. 1993. A global biome model based on plant physiology and dominance, soil properties and climate. Journal of Biogeography 19: 117-134.

Press, J. R., K. K. Shrestha, and D. A. Sutton. 2000. Annotated Checklist of the Flowering Plants of Nepal. The Natural History Museum, London, UK.

Rautiainen, O., and J. Suoheimo. 1997. Natural regeneration potential and early development of Shorea robusta Gaertn.f. forest after regeneration felling in the Bhabar- Terai zone in Nepal. Forest Ecology and Management 92:243-251.

Rautiainen, O. 1999. Spatial yield model for Shorea robusta in Nepal. Forest Ecology and Management 119: 151-162.

Reineke, L. H. 1933. Perfecting a stand-density index for even-aged forests. Journal of Agricultural Research 46: 627-638.

Richards, P. W. 1961. The types of vegetation of the humid tropics in relation to the soil. Pages 15-23 in Proceedings of the symposium on tropical soils and vegetation, Abidjan, UNESCO, Paris.

Rodgers, W. A., S.S.R. Bennet, and V. B. Sawarkar. 1986. Fire and Vegetation structure in Sal forests, Dehra dun, India. Tropical Ecology 27: 49-61.

Saldarriaga, J.G., D.C. West, M. L. Tharp, and C. Uhl. 1988. Long-term chronosequence of forest succession in the upper Rio Negro of Colombia and Venezuela. Journal of Ecology 76: 938-958.

Schweinfurth, U. 1957. Die Horizontale and Vertikale Verbreitung der Vegetation im Himalaya (mit vegetationskarte 1: 2,000,000). Bonner Geogr. Abh., H.20.

Shankar, U. S., S.D. Lama, and K. S. Bawa. 1998. Ecosystem reconstruction through ‘taungya’ plantations following commercial logging of a dry, mixed deciduous forest in Darjeeling Himalaya. Forest Ecology and Management 102: 131-142.

97

Shankar, U. 2001. A case of high tree diversity in a sal (Shorea robusta) - dominated lowland forest of Eastern Himalaya: floristic composition, regeneration and conservation. Current Science 81: 776-786. Shannon, C. E., and W. Weaver. 1949. The mathematical theory of communication. University of Illinois Press, Urbana, IL, USA.

Shapiro, S.S. 1980. How to test normality and other distributional assumptions, Milwaukee: American Society for Quality Control.

Sharma, B. K. 1999. Wildlife habitat mapping by using Geographic Information Systems (GIS) in the Karnali floodplain of Royal Bardia National Park at lowland Nepal. Masters Theisis. Agricultural University of Norway, Norway.

Shmida, A., and M.V. Wilson. 1985. Biological determinants of species diversity. Journal of Biogeography 12: 1-20.

Shrestha, A. 1992. Physical and chemical properties of soil in Nepal. A Journal of Forest Information of Nepal 3:27-29.

Shrestha, T.K. 1997. Mammals of Nepal: with reference to those of India, Bangladesh, Bhutan and Pakistan. R.K. Printers, Kathmandu, Nepal.

Shrestha, K.K, and P. K. Jha. 1997. Plant diversity and evaluation of conservation measures in the Royal Bardia National Park (RBNP). A report submitted to World Wildlife Fund Nepal Program. Kathmandu, Nepal.

Singh, A., V.S. Reddy, and J.S. Singh. 1995. Analysis of woody vegetation of Corbett National Park, India. Vegetatio 120: 69-79.

Singh, J.S., and S. P. Singh. 1992. Forests of Himalaya. Gyanodaya Prakashan, India.

Smith, J.L.D., S. C. Ahearn, and C. McDougal. 1998. Landscape analsysis of tiger distribution and habitat quality in Nepal. Conservation Biology 12 (6): 1338- 1346.

Sokal, R. R., and F. J. Rohlf. 1995. Biometry. 3rd edition. W. H. Freeman, New York, USA.

Spurr, H. R., and B. V. Barnes. Forest Ecology. 3rd edition. John Wiley, New York.

Stainton, J.D.A. 1972. Forests of Nepal. John Murray, London.

Stearn, W.T. 1960. Allium and Milula in the Central and Eastern Himalaya. Bulletin of British Museum of Natural History (Bot.) 2: 159-191.

98

Sugita, H., M.N. Subedi, C. Yonebayashi, H. Takayam, and M. Minaki. 1994. Floristic Composition and Structures of Forests in Arun Valley, Nepal. Ecological Review 23 (1): 17-42. Suoheimo, J. 1995. Natural regeneration potential of mixed Sal forest in Nepal IInd. FMUDP working paper-18. National forest Division, Department of Forests, Nepal.

Swaine, M.D., and P. Becker. 1999. Woody-life form composition and association on rainfall and soil fertility gradients in Ghana. Plant Ecology 145: 167-173.

Thapa, G.B., and K.W.Weber. 1995. Natural resource degradation in a small watershed in Nepal: complex causes and remediation measures. Natural Resources Forum 19: 285-296.

Tilman, G. D. 1982. Resources competition and community structure. Princeton University Press, New Jersey.

Troup, R. S. 1921. Silviculture of Indian Trees, Vol II, revised and enlarged by H.B. Joshi in 1980. Clarendon press, Oxford, UK.

Tuomisto, H., K. Roukolainen, R. Kalliola, A. Linna, W. Danjoy, and Z. Rodriguez. 1995. Dissecting Amazonian biodiversity. Science 269: 63-66.

Upreti, B. N. 1994. Royal Bardia National Park. National Conservation Strategy Implementation Project. IUCN, Kathmandu, Nepal.

Varguhese, G. 2000. Population and forest dynamics in the hills of Nepal: institutional remedies by rural communities. Pages 193-226 in C. Gibson, M. Mckean, and E. Ostrom, editors. People and Forests: Communities, Institutions, and Governance. MIT Press, MA, USA.

Villers-Ruiz, L., I. Trejo, and J. Lopez-Blanco. 2003. Dry vegetation in relation to the physical environment in Baja California Penninsula, Mexico. Journal of vegetation science 14: 517-524.

Webb, E. L., and R. N. Sah. 2003. Structure and diversity of natural and managed sal (Shorea robusta Gaertn.f.) forest in the Terai of Nepal. Forest Ecology and Management 176: 337-353.

Wesche, K. 1997. A classification of a tropical Shorea robusta forest stand in southern Nepal. Phytocoenologia 27 (1): 103-118.

World Resources Institute. 1996. World Resources 1996-1997. Oxford University Press, New York.

99

Xiaoming, Z., C. P. Zucca, R.B. Waide, and W.H. McDowell. 1995. Long-term influence of deforestation on tree species composition and litter dynamics of a tropical rain forest in Puerto Rico. Forest Ecology and Management 78: 147-157. Yadav, J.S.P., and D. R. Sharma. 1967. A soil investigation with reference to distribution of sal and teak in Madhay Pradesh. Proceeding of XI Silvicultural Conference, 1967, Dehra Dun, India.

Zhang, K., M. S. Ross, and P. L. Ruiz. 2005. Large-scale assessment of Mangrove landscape changes using enhanced shuttle radar topography mission data, LIDAR and ground observations. Unpublished report. Florida International University, Miami, Florida

Zomer, R., S.L.Ustin, and C. C. Carpenter. 2001. Land cover change along tropical and subtropical riparian corridors within the Makalu Barun National Park and Conservation Area, Nepal. Mountain Research and Development 21: 175-184.

100

Appendix 1

List of trees species with their codes, species names corresponding to the code and density (stems ha-1) of trees species in different groups. Group 1 (Shorea robusta-Buchanania latifolia), Group 2 (Terminalia tomentosa-Shorea robusta) and Group 3 (Shorea robusta- Cleistocalyx operculatus).

Codes Species Density (stems ha-1) Group 1 Group 2 Group 3 Shorob Shorea robusta 33 24 94 Dilpen Dillenia pentagyna 20 16 4 Tertom Terminalia tomentosa 8 46 3 Buclat Buchanania latifolia 110 96 5 Anolat Anogeissus latifolius 0.5 - 0.1 Myrsem Myrsine semiserrata 21 46 0.45 Malphi Mallotus philippensis 5 2 4 Lagpar Lagerstroemia parviflora 31 42 11 Schole Schleichera oleosa 1.3 0.17 0.11 Cleope Cleistocalyx operculatus 30 3 31 Adicor Adina cordifolia 0.15 0.23 0.84 Syzcum Syzygium cumini 1.58 0.59 0.46 Antchi Anthocephalus chinensis - 1.09 0.076 Semana Semecarpus anacardium 20 6 - Terbel Terminalia bellirica 0.85 0.81 0.94 Picjav Picrasma javanica - 0.28 - Baslat Engelhardtia spicata 0.43 - - Desooj Desmodium oojeinense - 0.62 - Paruli* Local name - 1.96 - Cararb Careya arborea 1.05 6 5 Ficben Ficus beghalensis 0.4 - - Butmon Butea monosperma 1.02 - 0.58 Holpub Holarrhena pubescens 2.09 - - Acacat Acacia catechu 1.07 - - Zizyph Zizyphus sps. 0.46 0.58 2.01 Banrit* Local name - - 0.05 Jingar* Local name - - 0.20 Spapar** Spatholobus parviflorus 0.35 2.58 -

‘*’= Local name ‘**’= Climber ‘-‘= absent

101 Appendix 2

List of trees species with their codes, species names corresponding to the codes and basal area (m2 ha-1) of trees species in different groups. Group 1 (Shorea robusta-Buchanania latifolia), Group 2 (Terminalia tomentosa-Shorea robusta) and Group 3 (Shorea robusta- Cleistocalyx operculatus)

Codes Species Basal area (m2 ha-1) Group 1 Group 2 Group 3 Shorob Shorea robusta 6.25 5.12 10.89 Dilpen Dillenia pentagyna 0.83 0.53 0.25 Tertom Terminalia tomentosa 1.47 4.50 0.81 Buclat Buchanania latifolia 0.89 0.79 0.12 Anolat Anogeissus latifolius 0.13 - 0.03 Myrsem Myrsine semiserrata 0.32 0.53 0.03 Malphi Mallotus philippensis 0.13 0.18 0.13 Lagpar Lagerstroemia parviflora 0.45 0.35 0.06 Schole Schleichera oleosa 0.19 0.09 0.03 Cleope Cleistocalyx operculatus 1.02 0.09 0.40 Adicor Adina cordifolia 0.06 0.18 0.09 Syzcum Syzygium cumini 0.06 0.09 0.12 Antchi Anthocephalus chinensis - 0.09 0.03 Semana Semecarpus anacardium 0.13 0.09 - Terbel Terminalia bellirica 0.13 0.18 0.22 Picjav Picrasma javanica - 0.09 - Baslat Engelhardtia spicata 0.06 - - Desooj Desmodium oojeinense - 0.09 - Paruli* Local name - 0.18 - Cararb Careya arborea 0.13 0.09 0.25 Ficben Ficus beghalensis 0.06 - - Butmon Butea monosperma 0.13 - 0.09 Holpub Holarrhena pubescens 0.06 - - Acacat Acacia catechu 0.13 - - Zizyph Zizyphus sps. 0.06 0.09 0.09 Banrit* Ban ritha - - 0.03 Jingar* Jingar - - 0.03 Spapar** Spatholobus parviflorus 0.13 0.18 -

‘*’= Local name ‘**’= Climber ‘-‘= absent

102 Appendix 3

List of trees species with their codes, species names corresponding to the codes and their frequency and relative frequency in different groups. Group 1 (Shorea robusta-Buchanania latifolia), Group 2 (Terminalia tomentosa-Shorea robusta) and Group 3 (Shorea robusta- Cleistocalyx operculatus).

Frequency Relative frequency Codes Species Group 1 Group 2 Group 3 Group 1 Group 2 Group 3

Shorob Shorea robusta 1.00 1.00 1.00 0.20 0.20 0.36 Dilpen Dillenia pentagyna 0.50 0.46 0.14 0.10 0.09 0.05 Tertom Terminalia tomentosa 0.67 1.00 0.41 0.13 0.20 0.14 Buclat Buchanania latifolia 0.44 0.46 0.05 0.09 0.09 0.02 Anolat Anogeissus latifolius 0.11 - 0.03 0.02 - 0.01 Myrsem Myrsine semiserrata 0.33 0.38 0.03 0.07 0.08 0.01 Malphi Mallotus philippensis 0.11 0.15 0.05 0.02 0.03 0.02 Lagpar Lagerstroemia parviflora 0.33 0.31 0.05 0.07 0.06 0.02 Schole Schleichera oleosa 0.17 0.08 0.03 0.03 0.02 0.01 Cleope Cleistocalyx operculatus 0.44 0.08 0.30 0.09 0.02 0.11 Adicor Adina cordifolia 0.06 0.15 0.08 0.01 0.03 0.03 Syzcum Syzygium cumini 0.06 0.08 0.05 0.01 0.02 0.02 Antchi Anthocephalus chinensis 0.11 0.08 - 0.02 0.02 - Semana Semecarpus anacardium 0.11 0.15 0.16 0.02 0.03 0.06 Terbel Terminalia bellirica - 0.08 0.03 - 0.01 0.01 Picjav Picrasma javanica - 0.08 0.03 - 0.01 0.01 Baslat Engelhardtia spicata 0.06 - - 0.01 - - Desooj Desmodium oojeinense - 0.08 - - 0.01 - Paruli Local name - 0.15 - - 0.03 - Cararb Careya arborea 0.11 0.08 0.16 0.02 0.01 0.06 Ficben Ficus beghalensis 0.06 - - 0.01 - 0.00 Butmon Butea monosperma 0.11 - 0.08 0.02 - 0.03 Holpub Holarrhena pubescens 0.06 - - 0.01 - - Acacat Acacia catechu 0.06 - - 0.01 - - Zizyph Zizyphus sps. 0.06 0.08 0.08 0.01 0.01 0.03 Banrit Local name - - 0.03 - - 0.01 Jingar Local name - - 0.03 - - 0.01 Spapar Spatholobus parviflorus 0.11 0.15 - 0.02 0.03 - (- = absent)

103 Appendix 4 a) List of saplings with their codes, species names corresponding to the codes and density (stems ha-1) in different groups. Density (stems ha-1) Codes Species Group 1 Group 2 Group 3 Shorob Shorea robusta 2045.16 1969.49 278.86 Lagpar Lagerstroemia parviflora 56.61 78.39 24.10 Buclat Buchanania latifolia 77.84 137.18 - Cleope Cleistocalyx operculatus 99.07 107.78 79.18 Picjav Picrasma javanica 14.15 29.40 6.89 Malphi Mallotus philippensis 134.46 205.77 82.62 Holpub Holarrhena pubescens 56.61 - 210.00 Zizphy Zizyphus sps. 120.30 88.19 89.51 Baslat Engelhardtia spicata 113.23 29.40 - Semana Semecarpus anacardium 21.23 - - Phorsa Grewia sps. 92.00 9.80 213.45 Schole Schleichera oleosa 7.08 9.80 3.44 Syzcum Syzygium cumini - - 3.44 Casfis Cassia fistula 7.08 29.40 - Dilpen Dillenia pentagyna - 9.80 - Kaphal Kaphale* 7.08 29.40 27.54 (‘*’= Local name; - = absent) b) List of seedlings with codes, species names corresponding to the codes and their density (plants ha-1) in different groups. (- = absent)

Density (plants ha-1) Codes Species Group 1 Group 2 Group 3 Shorob Shorea robusta 49166.67 61538.46 87432.43 Picjav Picrasma javanica 2222.22 - - Schole Schleichera oleosa 555.56 769.23 405.41 Malphi Mallotus philippensis 555.56 1923.08 1891.89 Aegmar Aegle marmelos - 384.62 - Syzcum Syzygium cumini - - 270.27 Tertom Terminalia tomentosa 277.78 - 945.95 Buclat Buchanania latifolia 1111.11 2692.31 - Unid 15 Unidentified - 384.62 - Cleope Cleistocalyx operculatus - 3076.92 405.41 Semana Semecarpus anacardium 1111.11 - - Desooj Desmodium oojeinense 555.56 - - Engspi Engelhardtia spicata 1388.89 - - Dilpen Dillenia pentagyna - - 135.14 Lagpar Lagerstroemia parviflora 277.78 - 540.54 Unid29 Unidentified - - 135.14 Cararb Careya aroborea - - 270.27 Unid37 Unidentified - - 675.68 Zizyph Zizyphus sps. 555.56 - 810.81 Ficrac Ficus benghalensis - - 135.14 Holpub Holarrhena pubescens - - 135.14

104 Appendix 5

List of shrubs with their codes, species names corresponding to the codes and shrub density (plants ha-1) in the different groups.

Density(plants ha-1) Codes Species Group 1 Group 2 Group 3 Flestr Flemingia strobilifera 254.76 215.57 82.62 Calmac Callicarpa macrophylla 14.15 - 0.00 Phupha Indigofera pulchella 35.38 117.58 20.66 Phylan Phyllanthus sp. 148.61 - - Flecha Flemingia chappar - 29.40 6.89 Clevis Clerodendrum viscosum 7.08 - 86.07 Hedyo Hedyotis sp. 7.08 - - Elsbla Elsholtzia blanda - - 37.87 Pogbeg Pogostemon benghalensis - - 27.54 Grewia Grewia sp. - - 3.44

105