|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
LANDSCAPE CHANGE AROUND KIBALE NAT IONAL PARK, UGANDA: IMPACTS ON LAND COVER, LAND US E, AND LIVELIHOODS By JOEL NATHAN HARTTER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007 1
2007 Joel Nathan Hartter 2
Fr Rachel. Nur mit deinen unendlichen Li ebe und Untersttzung war diese Arbeit mglich. 3
ACKNOWLEDGMENTS I would like to express my deep est thanks to my advisors Dr s. Abraham Goldman and Jane Southworth. My relationship with my advisors has grown to an especially strong one over my three years at UF. They have supported my effort s here at the University of Florida in so many different ways. They have given selflessly of their time and energy and I thank them sincerely for their many contributions, support, and encouragement during my time at UF both in the field and on campus. With his patience, always positiv e attitude, and even-keel demeanor (that is especially important in Uganda), Dr. Abe Goldma n met me in Gainesville and challenged me to think hard about where I wanted to go as a geographer. He was gracious in providing academic and financial support throughout my studies. Weve spent countless hours toge ther in the office, in the classroom, and in Uganda. Ive enjoyed the company as well as the academic stimulation. I thank him for his guidance and valuing my cont ribution to the Landscape Around Parks project. Dr. Jane Southworth has been a great source of positive energy and encouragement. She constantly had her door open, telling me to pop in for a brief conservation or a longer discussion. All of those, whether short or long, were instrume ntal in focusing my interests. Shes been a listening ear and a sounding board for ideas, and th rough that, Dr. Southworth s been especially good at challenging me as a geographer, to look d eeper and to link thos e social and ecological systems. I look forward to leaving the ranks of advisee and entering the new ranks as her colleague. I would also like to express my heartfelt appreciation to my committee members that have spent countless hours discussing th e project, reviewing my work, and collaborating in field data collection. Their contribution and critical evaluation has been invaluable. Thanks to Dr. Colin Chapman, who initially introduced me to research in the Kibale area even before I came to the University of Florida, and who has provided cons tant support in my res earch both in the field 4
and in Gainesville. Its been a wonderful and rewarding learning e xperience working closely with him in the field. I am extremely appreciati ve of his patience and confidence in me. I have learned so much from Dr. Mike Binford in my time at UF. Hes given me great guidance professionally and academically. He has been part icularly insightful with the initial design of this research, in the field, and throughout the co mpletion of this research. Dr. Brian Child has supported me throughout my experi ence here at UF. He has gi ven me great tools to conduct research in Africa and given my interests and goals thoughtful consid eration. I am grateful to all of my committee members for th eir encouragement, criticisms, and resources throughout my time at UF. I thank them for leaving thei r (virtual and literal) doors open to me. I wish to recognize support for my PhD studies from the US Department of Education Title VI Foreign Language and Area Studies Fellowshi p. My research was supported by the National Science Foundation grant (#0352008), a Working Fore sts in the Tropics Field Research Grant, and the Center for African Studies David L. Niddrie Memorial Scholarship. Permission to conduct this research was granted by the Univ ersity of Florida Institutional Review Board, Uganda National Council for Science and Techno logy, Uganda Wildlife Authority, and the local council leaders and their re presentatives of the many communities I visited. I also appreciate the continued cooperation with Makere re University Biological Field Station. I am especially grateful to my research fiel d assistants Agaba Erimosi and Mwesigwe Peace, without both of whom who this work would not have been possible. Many thanks are in order for their patience, respect, and friendship. It will not be forgotten. Mwebale muno. Thanks too to Omeja Patrick for all his hard work in managing the project in my absence and who has become a wonderful colleague and friend through my work in Uga nda. Thanks also to the amazing support in Uganda from the crew at Kiba le NP Dr. Lauren Chapman, Dr. Dennis Twinomugisha, Dr. 5
6 Aventino Kasangaki, Dr. Tom Gillespie, Dr. Tara Harris, Kevin Potts, Sini Skippari, Dr. Gosia Arlet, Dr. Freerk Molleman, Rebecca Chancellor, Tamaini Snaith, Stacey Hodder, Erin Reardon, Mike Wasserman, and so many others. Many thanks also to the Center for Africa Studies at the University of Florida. Drs. Leo Villalon and Todd Leedy have been instrumental in the development and support of the Natural Resource Management in AfricaWorking Group and the facilitation of my Foreign Language and Area Studies Fellowship and my Swahili language courses. I want to also thank Dr. Mark Brenner a nd Dr. Christian Russell of the Land Use and Environmental Change Institute (LUECI) for fu nding and for a physical place to continue my work at campus. Folks at LUECI and the Department of Geography (DoG) have been wonderful. A warm thanks to my cohort (or better, my co-conspirators) in LUECI, DoG, NWAWG, and the at-large UF community Da ve Buck, Liz Binford, Lin Cassidy, Jean-Gael Collomb, Amy Daniels, Maria DiGiano, Hector Costaneda, Andrea Gaughan, Jackie Hall, Natalie Hoyos, Saemi Ledermann, Christine Lucas, Matt Marsik, Forrest Stevens, Gabi Stocks, Miriam Wyman, Keith Yearwood, the ultimate crew (you know who you are), and to many others too countless to name (a nd have forgotten). Thank you! Finally, I wish to express my deepest gratitude to my wife Rachel who has encouraged me these past three years in countless ways. She has been extremely patient during the weeks of long study hours; she has given s upport and a listening ea r in times of frustration; and she has shared in all of the joy that I have experienced he re at UF. I am so thankful for her challenging me to search for answers and am blessed to share in all of lifes adventures with her. I thank her for being by my side every step of the way throu gh graduate school. For all of this and so much more, I am truly indebted to her.
TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ........10 LIST OF FIGURES.......................................................................................................................12 ABSTRACT...................................................................................................................................14 CHAPTER 1 INTRODUCTION................................................................................................................. .16 Introduction................................................................................................................... ..........16 Forests and Fragments.....................................................................................................18 Wetlands..........................................................................................................................20 The Fragmented Landscape.............................................................................................21 Study Area..............................................................................................................................24 Methods..................................................................................................................................25 Diminishing Resources...........................................................................................................26 Resource Use and Household Livelihoods.............................................................................27 Responses to Resource Availability.......................................................................................28 The Impacts of a Forest Park..................................................................................................2 9 Conclusions.............................................................................................................................29 2 DIMINISHING RESOURCES: WE TLANDS AND FOREST FRAGMENTS AROUND KIBALE NAITONAL PARK, UGANDA...........................................................32 Introduction................................................................................................................... ..........32 Study Area..............................................................................................................................35 Methods..................................................................................................................................38 Variation in Interannual Rainfall.....................................................................................39 Land Cover Analysis.......................................................................................................41 Fragmentation Analysis...................................................................................................42 Results.....................................................................................................................................42 Land Cover Analysis.......................................................................................................42 Productivity Assessment.................................................................................................43 Fragmentation Analysis...................................................................................................43 Discussion...............................................................................................................................44 Conclusion..............................................................................................................................52 3 RESOURCE USE ANDHOUSEHOL D LIVELIHOODS: LIVING NEAR WETLANDS AND FOREST FRAGMENT S NEIGHBORING KIBALE NATIONAL PARK........................................................................................................................... ...........62 7
Introduction................................................................................................................... ..........62 Study Region................................................................................................................... .......64 Ecological System...........................................................................................................64 Forest fragments.......................................................................................................66 Wetlands...................................................................................................................66 Social System..................................................................................................................67 Methods..................................................................................................................................68 Results.....................................................................................................................................69 Forest Fragment and Wetland Benefits to Households...................................................69 Problems to Households..................................................................................................71 Discussion...............................................................................................................................74 Benefits from Wetlands and Forests................................................................................74 Environmental Impacts from Social Benefits..................................................................77 Problems..........................................................................................................................78 Conclusion..............................................................................................................................84 4 THE LANDSCAPE SURRO UNDING A PARK: RESPONSES BY HOUSEHOLDS TO RESOURCE AVAILABILITY IN RURAL COMMUNITIES.......................................94 Introduction................................................................................................................... ..........94 Study Region................................................................................................................... .......97 Physical Environment......................................................................................................97 Social Environment.........................................................................................................98 Methods..................................................................................................................................99 Remote Sensing...............................................................................................................99 Landscape Fragmentation..............................................................................................100 Household Interviews....................................................................................................100 Results...................................................................................................................................101 Remote Sensing.............................................................................................................101 Landscape Fragmentation..............................................................................................102 Household Interviews....................................................................................................103 Discussion.............................................................................................................................107 Conclusion............................................................................................................................115 5 IMPOSING WILDERNESS? THE IM PACTS OF A FOREST PARK..............................124 Introduction................................................................................................................... ........124 Study Region................................................................................................................... .....127 Physical Environment....................................................................................................127 Social Environment.......................................................................................................129 Methods................................................................................................................................130 Results...................................................................................................................................131 Discussion.............................................................................................................................133 Conclusion............................................................................................................................143 6 CONCLUSION................................................................................................................... ..150 8
9 Diminishing Resources.........................................................................................................151 Resource Use and Household Livelihoods...........................................................................152 Responses to Resource Availability.....................................................................................153 The Impacts of a Forest Park................................................................................................154 Research Significance and Implications...............................................................................155 Future Research....................................................................................................................158 LIST OF REFERENCES.............................................................................................................160 BIOGRAPHICAL SKETCH.......................................................................................................176
LIST OF TABLES Table page 2-1 Definitions of landscape metrics incorporat ed within this analysis of fragmentation of forest and wetland patches, from 1985, 1995 to 2003...................................................58 2-2 Inside-park fragmentation analysis....................................................................................58 2-3 Forest and wetland productivity as asse ssed by NDVI (NDVI valu es between -1 and +1). Values close to +1 indicate higher productivity.........................................................59 2-4 Fragmentation analysis for surrounding landscape............................................................60 2-5 Fragmentation analysis for east study area........................................................................60 2-6 Fragmentation analysis for west study area.......................................................................61 3-1 Benefits and problems from natural areas.........................................................................88 3-2 Top five benefits derived from wetlands (n=130).............................................................88 3-3 Top five benefits derived from forests (n=130).................................................................89 3-4 Respondents reported their primary sour ces for household water and energy use............89 3-5 Perceived problems due to presence of forest (including KNP) and wetlands..................90 3-6 Most commonly reported problem anim als identified by local landholders.....................91 3-7 Greatest problem animals reported by res pondents and responses relative to distance from the park boundary......................................................................................................92 3-8 Perceptions of respondents whether or not there will be enough wetlands and forest nearby to meet their needs in the future.............................................................................92 3-9 Preferences for living closer or further from natural areas................................................93 4-1 Land cover proportions and class-level metrics outside KNP in the East Study Area (based on 5-class land cover classi fication from 2003 Landsat image)...........................119 4-2 Land cover proportions and class-level metrics outside KNP in the West Study Area (based on 5-class land cover classi fication from 2003 Landsat image)...........................119 4-3 Land cover change detection fo r the east and west study areas.......................................120 4-4 Patch-level statistics for east and west study areas..........................................................120 4-5 Respondents perception of change in resource bases in the last 10 years......................121 10
11 4-6 Reported changes in forest fragments and wetlands........................................................121 4-7 Reported reasons for difficulty in obta ining resources in wetlands and forest fragments...................................................................................................................... ....121 4-8 Resource and wild animal concerns among to all respondents........................................121 4-9 Relationship of response and ethnicit y, wealth, head of household gender, and distance to the park boundary..........................................................................................122 4-10 Household responses to resource sh ortages among all respondents. (Note: households can partake in a variety of combinations of responses, choosing one, several, or none of these responses to shortages)............................................................122 4-11 Relationship of response to resource s hortage and ethnicit y, wealth, head of household gender, and distance to the park boundary.....................................................122 4-12 Questions posed to households regarding problems of and response to crop and/or livestock raiding...............................................................................................................122 4-13 Household responses to crop raiding. (Note: an unknown number of offending animals are killed by landholders)...................................................................................123 4-14 Household responses to resource shor tages among male and female heads of household.........................................................................................................................123 5-1 Reasons reported why KNP should stay or be dissolved.................................................148 5-2 Has the LC zone benefited from KNP and the perceived benefits by the LC1...............148 5-3 Visitor statistics for Uganda national parks, Murchison Falls, Bwindi and Kibale National Parks. Kibale National Park Gr oss Revenue generated and revenue shared with communities is also given........................................................................................149
LIST OF FIGURES Figure page 1-1 Kibale National Park and surr ounding landscape in western Uganda...............................31 2-1 Kibale National Park and surr ounding landscape in western Uganda...............................53 2-2 Annual precipitation collect ed at the Kibale research station (Kanyawara). Annual precipitation totals for image capture years are noted. No data was acquired during September 13-21, 2004 while the broken rain gauge was repaired...................................54 2-3 Monthly precipitation data for the th ree image years (May 1984, January 1995, and January 2003) and the years preceding image capture (1983 data was unavailable) at the Kibale research station (Kanyawara)...........................................................................54 2-4 Precipitation totals for the short a nd long rainy seasons between 1984 and 2005. Mean values and +/1 standard deviations are given for each rainy season.....................55 2-5 Land cover change in-par k and in the surrounding lands cape, includi ng the east and west study areas............................................................................................................... ..56 2-6 Land cover extent in th e surrounding landscape 1984-2003.............................................56 2-7 Connectivity of natural areas in east and west study areas................................................57 3-1 Kibale National Park and surr ounding landscape in western Uganda...............................86 3-2 Distance of households to KNP boundary reporting problems with wild animals for leading crop raiders........................................................................................................... .87 3-3 Distance of households to natural area s reporting problems with wild animals for leading crop raiders........................................................................................................... .87 4-1 Kibale National Park and surrounding landscape in western Uganda.............................117 4-2 Farming along the forest edge. Th e KNP boundary is clearly demarcated by Eucalyptus trees, separating forest (left) and fields of maize (right) (Photo by J. Hartter).............................................................................................................................118 5-1 Kibale National Park and surr ounding landscape in western Uganda.............................144 5-2 Perceived impact of KNP, n = 130..................................................................................145 5-3 Perceived benefits from KNP, n = 79..............................................................................145 5-4 Perceived problems from KNP, n = 44............................................................................146 12
13 5-5 Where would you rather live?, n = 130. Th e definition of closer and further was left to the responde nt to interpret.....................................................................................146 5-6 Should KNP stay or go?...................................................................................................147
Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LANDSCAPE CHANGE AROUND KIBA LE NATIONAL PARK, UGANDA: IMPACTS ON LAND COVER, LAND USE, AND LIVELIHOODS By Joel Nathan Hartter December 2007 Chair: Abraham Goldman Cochair: Jane Southworth Major: Geography Rapid population growth, high population density, and intensive agriculture characterize the landscape surrounding Kibale National Park (KNP) in western Uganda. Although the park itself is completely surrounded by agricultural la nd and large multi-national tea estates, forest fragments and wetland areas are sc attered throughout the landscape. As small rural communities in western Uganda are heavily reliant on land and resources, their livelihoo ds are directly linked to ecological systems. In contrast to most ot her studies that are conc erned with parks, this dissertation research focused on the impacts of parks on its nei ghbors, by examining the changes in wetlands and forest fragments outside th e park and the impacts of those changes on households as cascading effects of pa rk establishment on communities. This research uses KNP as a case study in a two-step an alysis, linking landscapeand household-level data. The first step was quantifyi ng the change in extent, pattern, and quality of wetlands and forest fragments outside KNP over ti me using satellite im age analysis. Land cover analysis shows Kibale National Park as a n early islandized park of forested land cover, surrounded by intensive small-scal e agriculture and some largescale tea plantations, and many unprotected wetlands and forest fragments. I found that the park boundaries remain relatively intact since 1984, while the lands cape surrounding the park has beco me increasingly fragmented. 14
15 As fragmentation outside the park continues, the forests and wetlands in the surrounding landscape not only are becoming increasingly sma ller, but more isolated and with lower productivity In the second step of this analysis, the implications of that change on household livelihoods in communiti es surrounding KNP we re examined using h ousehold interviews. Wetlands and forest fragments outside Kibale Na tional Park serve as important resource bases for local people, but they are also problematic for local farmers, since wild animals come from these natural areas and raid crops. Households also recognize the decreased presence and extent of forests and wetlands outside the park. They have adopting coping strategies not only to address resource shortages, but also the problems of crop raiding. Finall y, my results suggest that not only do most households fe el they benefit from the park (although in ways not widely discussed in the literature on parks), but the majo rity also feel they are not harmed by it, and many feel no effect from the park.
CHAPTER 1 INTRODUCTION Introduction Parks and protected areas are designated as ar eas of biological, geol ogical, or cultural significance and have served as key mechanisms for conservation (Howard et al. 2000). McNeely (1990) sees parks as making fundamental contributions to sustai ning human societies. However, park establishment is not necessarily benign, especially to thos e living in or near a park, and parks should not be viewed in isolati on from the landscape in which they sit. Many parks in Africa have adopted the traditional mode l, with the establishment of hard boundaries and the near to complete exclusion of residents from access, extractive, and settlement privileges (Hayes, 2006). As a result, the landscapes around parks are important. The landscapes around protected areas have become mosaics of natural and human-influen ced patches, and becoming increasingly fragmented. This landscape outside the protected area can be infl uenced by edge effects, population grow th, and changing patt erns of land use, among other factors. Unprotected lands outside the park represent reservoirs of land, resources, and potential income. Increased tourism-based activities or other secondary industries are likely to attract migrants. Since the park boundaries for the most part, prohibit settlement a nd resource extraction, increasing population and land and resource pressu re stimulates intens ified land use surrounding the park, leading to a reduction in forests and wetlands that serve as important resource bases. Patterns of land cover change in many tropical developing countries are closely linked to land use changes and anth ropogenic impacts. La nd use intensification and extensification are common responses not only to the limitation and c onstriction of resource s, but also to the opportunities that unuse d lands, such as forests and we tlands, represent (L ambin and Geist, 16
2001). All methods of intensifying agricultur e concentrate resources in time and sp ace. The varying responses of differen t economic and ethnic groups to economic opportunities, crop loss, and resource needs can have quite different effects on the spatial organization of land use (Stone, 1996; Kagoro-Rugunda, 2004). Under the traditi onal conservation model (a.k.a. fines and fences fortress co nservation) the presence of a protected area is a hard barrier to settlement, extensification, and resource extraction, which affects farmer s land use and livelihood option s. Since resource extraction and other forms of land use are forbidden within the park boundaries, peop le turn to the only available land to me et their needs. Those we tlands and forest fragments that remain continue to be relied upon to produce resources for subsistence and commerci al use. As their use continues, there is potential for overuse, degr adation, reduced size and connectivity with other areas, and elimination of these areas. The landscap e around the park changes and inevitably becomes more fragmented as new land covers and land uses are repl aced with new ones or eliminated. The unprotected fo rests and wetlands, important to household livel ihoods, present both opportunities and challe nges to households a nd the community at larg e. This change in presence, spatial extent, and quality of resour ces will influence the livelihoods of neighboring communities. Estimates of the numbers of people generati ng much of their live lihoods (i.e., income, energy, food) from forests range from one milli on to more than one billion (Byron and Arnold, 1999) and the number of people indi rectly or directly dependent on wetlands worldwide at least one billion (Fischer et al. 1997). Especially on the African continent, ru ral livelihoods continue to be heavily reliant on the land and its resour ces (Scoones, 1998; Abal u and Hassan, 1998). The amount of forest global loss is no less striking. Forest loss worldwide is estimated at 7.3 million 17
hectares (between 2000-2005) (FAO, 2006), where fo rest loss in Africa has been as high as 10.5% (between 1980 and 1995) (Chapman and Peres, 2001) and now 9.7% (between 20002005) (FAO, 2006). As Africas population continues to grow, the number of pe ople dependent on forest and wetland land and resources has the potential to in crease substantially. In Eastern and Southern Africa, 63% of the population is rural with an average 1.8% popul ation increase annually (US Census Bureau, 2006). Uganda is a slightly different story. N early 80% of the 27 million people have rural-based livelihoods and more than 80% of land is used for small-scale farming (Uganda Bureau of Statistics, 2005). The population has increased 240% between 1960 and 2000, with a continued national population increase estimated to be 3% annually (US Census Bureau, 2006). Ugandas forest and wetland resources are esse ntial for the survival of the majority rural population, providing energy, income, and food (B anana and Gombya-Ssembajjwe, 1998). As such, unprotected wetlands and fore st fragments are conv erted to agriculture, grazing land, or woodlots at faster rates. Forests and Fragments Forests and woodlands cover approximately 4.9 million ha in Uganda and nearly 60% remain unprotected and are vulnerable to overe xploitation and agricu ltural encroachment (NEMA, 2001). At the estimated current rate of forest conve rsion, although estimates vary, Uganda continues to lose be tween 0.8% (NEMA, 2001) and 3% (Kayanja and Byarugaba, 2001) annually (about 50,000 ha). Nationwide, unsusta inable domestic tree harvesting for firewood and non-timber forest products continues (K ayanja and Byarugaba, 2001). Closed-canopy tropical forest once covered 20% of the countrys land area, but defo restation has reduced this to just 3% (Howard et al. 2000). Furthermore, Uganda lost 18% of its remaining forest between 1990 and 2000 (Howard et al. 2000). 18
Kabarole District in western Uganda is a good example of the pressure on forest resources and subsequent landscape fragmentation. In the densely populated Kabarole (8109 km2, 91 individuals/km2), only 2501 km2 of forest remain (NEMA, 2001) The once dominant moist, evergreen, closed-canopy forest that dominated the region has fo r the most part succumbed to agricultural expansion at leas t since 1959 (Gillespie and Chapman, 2006). Nearly all of the forests found on potentially arable lands have been converted to small-scal e agriculture, tea, or pasture and soil degradation is a major concern (NEMA, 2001). The remaining forests typically occur in valley bottoms or on the steep rims of crater lakes (Gill espie and Chapman, 2006; Hartter, pers. obs.). These unprotected forests pr ovide a number of ecologi cal services such as stabilization of the local clim ate, erosion control, nutrient uptake, and carbon sequestration (Laurance and Bierregaard, 1997). Fo rest fragments also serve as sp ecies corridors, habitat, and breeding and feeding grounds (Onderdonk and Chapman, 2000; Marsh, 2003). Forests fragments are extremely important to sustain livelihoods in Uganda. Ninety-five percent of all Ugandas energy n eeds are met with fuelwood and char coal, and two-thirds of this amount is used at the household level (Kay anja and Byarugaba, 2001). They provide wood products: fuelwood, timber, building poles, but also non-timber products such as indigenous medicines and food (Banana and Gombya-Ssembajjw e, 1998). Hardwoods within the fragments are especially sought to burn charcoal (Naughton-Treves et al. 2005). Many are also being converted into farmland, pasture, or woodlots (Kayanja and Byarugaba 2001). Despite their socio-economic importance, onl y recently has research addressed tropical forest fragments (Chatelain et al. 1996; Laurance and Bierreg aard, 1997), but most research has focused on South America (Laurance and Bierre gaard, 1997). Previous African studies have addressed the effects of fragm entation on animal and tree sp ecies (Hill and Curran, 2003; 19
20 Chapman et al. 2003; Marsh et al 2003; Stouffer and Bierregaar d, 1995). However, little work has studied the role forest fragments play in providing resources such as fuelwood, timber, poles, thatching, handcraft materials, indigenous medicines, and edible plants to communities (Turner and Corlett, 1996; Gillespie and Chapman, 2005). Wetlands Wetlands include all areas where plants grow and animals live in association with permanent or temporary flooding (Kisamba-Mugerwa and Nuwagaba, 1993). Ugandas wetlands, dominated by papyrus ( Cyperus papyrus L.) comprise about 13% of the countrys total land surface (Muk iibi, 2001a) and fill a critical ecological role in flood abatement, groundwater recharge, and as natu res biological filter. Wetl ands have vital importance in maintaining the quality and volume of surface a nd ground water (Muthuri et al. 1989; Mukiibi, 2001b). They are key spawning grounds for fish (Crisman et al. 2003), and provide other services such as biodiversity, aesthe tic beauty, and cult ural heritage. Wetlands are impor tant to livelihoods of surrounding communities in several ways. In most cases, they have been intensively us ed or drained. Many households are growing fuelwood and timber trees in drained wetlands. Th ey provide a subsistence resource base for water, thatching, hand craft materials, and indigenous medicines. Wetlands are also commercially important, providing a sour ce of revenue from touris ts, and medicinal plant and handcraft sales. Overexploitation occurs in the c ontext of harvesting resources, such as clay for brick building, and pottery and papyrus for that ching houses and making mats (Chapman et al. 2001, MacLean et al. 2003). Similar to forests, Ugandas wetlands represent unclaimed or underutilized land and ar e often encroached to cultivate crops (Crism an et al. 2003; MacLean et al. 2003). Wetlands are of ten drained and converted to ot her land uses such as growing crops and fuelwood particularly Eucalyptus spp ., and expanding grazin g land. Some wetlands
are diverted to irrigate crops. They are also an important source of water for livestock, leading to high nutrient loading and increased turbidity. In addition, as the dry season lingers and temporary wetlands dry up, lives tock are brought to more permanently inundated wetlands (Haack, 1996). Despite maintaining one of the highest ne t primary productivity rates in the world (Muthuri et al. 1 989), people in the past treated wetlands more like wastelands (Mukiibi, 2001b) or as a repository of land for potential agricultural expansion. Several studies have attempted to quantify wetland ecosystem services in terms of economic potential (Costanza et al. 1997; Batagoda et al. 2000; Turner et al. 2001; Balmfo rd et al. 2002). The Fragmented Landscape Landscapes around parks are also affected by wild animals that raid their crops and livestock. Households are affected by crop raid s by primates, elephants, and birds; livestock loss; and even human fatalities occurring in lands near parks (Naughton-Treves, 1997; de Boer and Baquete, 1998; Gillingham and Lee, 1999). Land scarcity has forced many farmers to live and farm at the forest edge (Naughton-Treves, 19 98) and as a result, human-wildlife conflict is a considerable risk for those farms directly adj acent to these natural areas. Although the amount of damage varies, farmers must constantly make decisi ons and create coping strategies to deal with crop-raiding (Goldman, 1996; Naught on-Treves, 1998). Little research has addressed either anthropogenic influences on land-cover change outsid e protected areas, or th e resulting effects of altered ecosystems back onto the communities (Pic kett et al. 1995; Turner et al. 2001). Studies on fragmented landscapes outside protected areas have focused on biodiversity conservation but there has been little research into human-environment linkages (Turner and Corlett, 1996). Moreover, fragmentation literature discusses landscape fragments and biodiversity in terrestrial ecosyst ems, but there is very little discussion of aquatic systems 21
(Pickett et al. 1997). While the use of forest frag ments as a means to sustain livelihoods is well documented elsewhere in the world, East Africa is st rikingly absent from the literature (Marsh et al. 2003), as are studies on the sma ller, interstitial wetla nds and forest patches characteristic of western Uganda. Little research has addresse d anthropogenic influences of land-cover change, or the resulting effects of altered ecosystems back onto the communities. Studies of fragmented systems must include consideration of ongoing human influences (Laurence and Bierregaard, 1997). In order to understand land cover change in the landscape surrounding a park, it is necessary to examine resource consumption patterns of park neighbors. Since most of the resources collected in the wetla nds and forest fragme nts are only obtained locally, households must respond in some way to shortages, in turn shaping the landscap e. Many livelihoods, and thus responses to shortages and problems, are determined through culture s and traditions, and economic, ecological environment, and demogr aphic variables (Chambers and Conway, 1991). Therefore variables such as ethnicity, wealth, an d gender as well as distance from the park may all be related to these responses (Rocheleau and Edmunds, 1997; Hill 1997; Byron and Arnold, 1999; Goebel et al. 2000; Kagor o-Rugunda, 2004). Resource use may vary by ethnicity, wealth, location, and gender (Rocheleau and Edmunds, 1997; Byron and Arnold, 1999; Goebel et al. 2000). Around Kibale National Park (KNP) in western Uganda, what changes have occurred in the natural resource base, how th ese changes have affected de cisions about land and resource use, and how land and resource use varies by ethnicity and wealth has not been addressed in the literature. This research addre sses these questions among others. This research examined the changes in forest fragments and wetlands outside KNP and the adaptation in livelihood strategies and natural resources use by different ethnic and economic 22
groups living within th e densely populated agricultural la ndscape surrounding KNP in western Uganda. Two main ethnic groups, the Bakiga a nd Batoro, dominate the population living in the area, and generally practice distinctive patterns of land and resource use. As in many similar cases around parks, there has been substantial population growth from both in-migration and a high rate of population increase. The area has b ecome a mosaic of diverse agricultural land intermixed with patches of remaining natural area s. These forest fragments and wetlands serve as important resource bases, but are also problematic for farmers. This research links landscape and household levels in a cross-s cale analysis that looks at cha nge in both spatial extent and quality of wetlands and forest fragments, as well as the livelihood consequences of that change. Landscape level changes in wetlands and forest fragments are the product of many decisions made at the household level. Household head gender, wealth, ethnicity, and distance to the park boundary may all be important factor s in contributing to liv elihoods. Sustaining livelihoods have spatial implica tions, and it is important to understand how they physically manifest themselves in land-use (e.g., loss of wetlands and forests). In the surrounding landscape, land cover will change as a result of households coping mechanisms to resource shortages and human-wildlife c onflict. These changes in land cover and the result coping mechanisms are the focus of the research questions for this dissertat ion. Four fundamental questions outline this research. How have the extent and quality of wetlands and forest fragments in the landscape around KNP changed over time? What resources and problems are associated with wetlands and fo rest fragments around KNP; and how have thes e changed over time? How have households ad apted to the declines in the wetlands and forest fragments in the areas around KNP? How are neighboring co mmunities helped or harmed due to the pr esence of KNP? 23
The research presented in this dissertation is divided into several chapters, which in aggregate address how landscapes and households within that landscape are impacted by park establishment. Each of the chapters deals with separate research questions. The last chapter of the dissertation provides broad conclusions and implications of this research. Study Area KNP in western Uganda is illustrative of th e agricultural expansion and intensification surrounding protected areas. Kiba le Forest was demarcated in 1932 as a Forest Reserve and elevated to national park status in 1993 (S truhsaker, 1997). Known for its diverse primate population, KNP ranks fifth in terms of species richness and sixth in overall biodiversity importance among all Ugandan forests (Howard et al. 2000). It is perhaps most well-known because it is home to the largest known community of chimpanzees. The park is an ecological island a dens e, closed-canopy forest surrounded by a large agricultural population, larg e tea estates, and a vast network of wetlands and bottomland forest fragments. KNP is a medium-altitude tr opical moist forest covering about 795 km2 in western Uganda (Figure 1-1) 1 This transitional forest (between lowland rainforest and montane forest) is at an average elevation of 1110-1590m and is a remnant of a previously larger mid-altitude forest region (Struhsaker, 1997). The climate is warm throughout the year, with an average range of 15-23C (Struhsaker, 1997). Although the amount of rainfall and length of season change, the average annual rainfall for the region is 1543mm (avera ge 1903-1999) and 1719mm (1990-2005) (Chapman et al. 2005). 1 While officially KNP now includes a game corridor that connects the so uthern portion of KNP to Queen Elizabeth National Park, this corridor was formally gazette d as part of KNP in 1993 but has not been cleared of human settlement and treated as part of the park until the last year and currently replanting etc. is still in progress. Before that time, it was mixed agricultural land. Therefore, it was not classified as park in this analysis. By doing so, our analysis is conservative since the addition of the corridor would tend to lessen the difference between park and non-park. Excluding the game corridor, the Kibale Forest in this analysis contains 561km2. 24
Rapid population growth, high population densit y, and heavy reliance on agriculture for income characterize the landscape surrounding KN P (Archabald and Naughton-Treves, 2001). Land pressure continues to incr ease (Naughton-Treves, 1996). Nearly 43% of the land within a 5 km periphery of the park is under cultivati on or pasture (Hartter, unpublished data). Tea dominates much of the landscape bordering the northwest portion of KNP, covering nearly 21% within 1 km of the park bounda ry (Hartter, unpublished data). Farm sizes on average are less than five hectares (Hartter, unpublished data) and both ethnic groups interplant more than 30 species of subsistence and cash crops (Naughton-Treves, 1996). Mo st of the crops produced are consumed at the household level, with few bei ng sold outside of imme diate household or local community markets. Agriculturalists in the area bel ong to two dominant ethnic gro ups: the Batoro (west side of KNP) and the immigrant Bakiga (east side of KNP), who came to the Kibale region from southwestern Uganda beginning the 1950s a nd 1960s (Turyahikayo-Rugyema, 1974; NaughtonTreves, 1998). A 1988 census estimated the popul ation within the park and the surrounding communities to be 61,000 (Aluma et al. 1989). Naughton-Treves (1998) estimated the population density has nearly trebled around the park between 1959 and 1990. Methods This research combined exte nsive household interviews w ith an analysis of Landsat satellite imagery over 20 years. As with la ndscape change, it is important to maintain a connection with people across both space and ti me (Walsh et al. 2003). Remote sensing provides a robust and efficient t echnique to observe and monitor the changes that occur within different land-covers over a large area and multiple dates (Lu, 2004). However, while satellite data provide a firm empirical base for measuring the spatial configuration of land-cover, they do not themselves explain the causes for these land -use strategies. Integrating landscape level 25
change detection with household level proc esses requires a sampling scheme that links household surveys to remotely sensed data. Su rveys provide valuable fine-scale, ground-data (Turner, 2003). Household-level decision making is critical to understanding the changes in land-use and land-cover and their effects on liv elihoods (Rindfuss et al. 2003). Combining household and landscape level data in a cohesive analysis can he lp explain both proximate and distal effects of environmental change (Ge oghegan et al. 1998; Rindf uss et al. 2003). Diminishing Resources Protected areas have long serv ed as centerpieces of the cons ervation movement (Howard et al. 2000). They are areas designated as areas of biological, geological, or cultural significance. In Sub-Saharan Africa, early park-based conser vation was derived largel y from Western roots, from the Yellowstone Model (Runte, 1997). Applying the wilderness model after Yellowstone National Park, the Africa continent has seen a dram atic growth in parks, both in number and area protected. This type of c onservation was built on strict ex clusion of humans, the prevention of consumptive use, and minimization of other forms of human impact within park boundaries (Hulme and Murphree, 2001). To do this effectively, park managers believed that humans must be taken out of the equation. Boundaries were er ected and even though the park itself may have been part of a continuous landscape, park mana gers sought to separate it from the landscape surrounding it. People were meant to use resource s outside of the park and plants and animals were meant to stay in the park. Outside parks in Sub-Saharan Africa, the ma jority of the population is dependent on the land for their livelihoods. They depend on the land for growing food, economic opportunity, building materials for their homes, and to meet energy needs. As population outside the park swells and people are prohibited from resource extraction inside its boundary, the demand for resources and agriculture land not only increases, but competes. Since land is a premium with 26
increasing populati on, the remaining unprotected wetlands and fore st fragments are sought after. Those wetlands and forest fragments that remain continue to be relied upon to produce resources for subsistenc e and commercial use. Not only is the presence of forests and wetlands diminishing, but they are also be ing degraded, reducing the quality of resources. As a result, the landscape around a park has become a mosaic of natural and human-influ enced patches. In this chapter, I examine the spatial extent and qu ality of these diminishing resource bases at the landscape level using discrete and continuous data analyses of satellite imagery and place the results within the social context of western Uganda and fortress. Specific research questions addressed in Chapter 2 are How has the productivity and spatial extent of forests and wetlands in the landscape surrounding KNP changed over time? How does this change vary in communities on the east side of the park versus the west side of the park? How does this change relate to change in land under cultivation and tea? Resource Use and Household Livelihoods The livelihoods of rural communities in wester n Uganda are directly linked to ecological systems since they are heavily reliant on land and resources, their livelihoo ds are directly linked to ecological systems. Unprotected wetlands an d forest fragments (i.e., those not gazetted as parks, wildlife reserves, or other forms of prot ected areas) serve as impor tant resource bases for local people as well as biodiversity habitats that harbor many wild plant and animal species. However, they are also problematic for local farmers, since crop raids by primates, elephants, and birds emanate from these fragments. In addition, there has been extensive conversion of wetlands to grazing or cropland, as well as increasing extraction of fuelwood, timber, and other resources from forest fragments, all of which are diminishing thei r area and ecological integrity. This chapter examines the dual character of these natural areas within the agricultural landscape 27
around KNP in terms of the social and environmen tal benefits and problems they represent to local households that vary in ethnicit y, wealth and distance from the park. Specific research questions addressed in Chapter 3 are: What resources and problems are associated with wetlands and forest fragments around KNP? How do these benefits and problems vary by di stance, wealth, gender, and ethnicity from the park boundary? Responses to Resource Availability Although protected areas have become th e primary mechanism for biodiversity conservation, their establishment can have l ong-term impacts on land use, land cover, and livelihoods of people living near them. Since land use and mo st resource access within KNPs boundaries are prohibited, park neig hbors turn to the forest fragmen ts and wetlands to meet their resource needs. In the past with adequate supply and access, households could accommodate their resource needs. However, the decline of wetlands and forest fr agments has an impact on households and their ability to s ecure resources and sustain their livelihoods. Since most of the resources collected in the wetlands and forest fragments are gath ered locally and not purchased, households must respond in some way to shorta ges. Some will purchase fuelwood, others will have to find new sources or trav el longer distances to collect fuelwood and water. Land use and land cover are inextricably linked. Strategies to increase land under cu ltivation and resources from the natural areas not only influence land cover within the landscape (i.e., more wetlands drained and converted to make way for agricultural lands or pastur e), but can also lead to humanwildlife conflict (Naughton-Trev es and Salafsky, 2004). This chapter examines the ways households have adapted to a degraded resource bases of the unprotected wetlands and forests and problems with wildlife in th e landscape surrounding a park. Specific research questions addressed in Chapter 4 are: 28
How have the presence and extent of wetla nds and forest fragments in the landscape surrounding KNP changed over time? How have households adapted to the declines in the wetlands and forest fragments in the landscape surrounding KNP? The Impacts of a Forest Park The fortress conservation model has sparked many debates given its treatment of parkpeople relationships. In many cases, parks have excluded residents fr om resource access and settlement. Certainly, park establishment can be considered an instigator of conflict in rural East Africa, with few benefits accruing at the local level and most at the national and international scales. Those who bear the majority of the costs are the local, rural poor, while most beneficiaries of parks tend to be the wealth ier and foreign visitors. Discontentment and disenchantment of conservation policies in Uganda amongst neighboring communities is a common effect of park establishment under the fortress conservation model. However, is fortress conservation always a negative model to local people? This chapter examines the case of KNP, a forest park, examining how the park has helped and hurt park neighbors. Specific research questions addressed in Chapter 5 are: Do households perceive that KNP has helped, hurt, or had no positive or negative impact on their family? What are the perceived benefits and problems associated with KNP? Conclusions From this dissertation, a number of important conclusions can be drawn. The land cover analysis revealed that since 1984 KNPs bounda ries remain relativel y intact while the surrounding landscape has become increasingly frag mented as land and resource pressure builds outside the park over time. Since the population outside the park is dependent on the land to sustain their livelihoods and because extraction within park boundaries is prohibited, park 29
neighbors turn to the forest fragments and wetlands outside the park to meet their resource needs. The remaining natural areas of forest and wetland s outside the park serve as resource bases and are also believed to be associated with ecosystem se rvices such as fresh air and adequate rainfall. Despite their many benefits, these natural area s are sources of problems. Wild animals were identified as the most prolific problem in the surrounding landscap e as a result of the presence of wetlands and forests. Increasing ex traction of fuelwood, timber, and other resources all lead to diminishing area and ecological integrity of forest fragments. Households continually create coping strategies to a ddress resource shortages as well as crop raiding, cited as a major problem due to wetland and forest presence both closer and further from the park boundary. Lastly, despite the exclusion from access and re sources within park bound aries and the forceful eviction of settlers from the park in 1993, I found that that most people are not overtly hostile towards the park. Results suggest that not only do most househol ds benefit from the park, but the majority is also not hurt and many feel no eff ect from the park. In fact, most respondents say that given the choice and weighing both the bene fits and the costs, th e park should stay. Most work on protected areas has examined at the impact of humans on parks and the impacts of extraction and resource exploitation. This study instead exam ined how parks are not benign within the landscape and have various spatial and temporal impacts on park neighbors. Linking householdand landscape-leve l data is vital to address the complex nature of humanenvironment interactions. This research introduced an innovative methodology for measuring cross-scale linkages. This sampling framework linked both landscape and household level scales used to examine changes in the landscape surrounding the park and how those changes have impacted park neighbors. 30
31 Figure 1-1. Kibale National Park and surrounding landscape in western Uganda
CHAPTER 2 DIMINISHING RESOURCES: WETLAND S AND FOREST FRAGMENTS AROUND KIBALE NAITONAL PARK, UGANDA Introduction Protected areas have long serv ed as centerpieces of the cons ervation movement (Howard et al. 2000). They are areas designated as areas of biological, geological, or cultural significance. In Sub-Saharan Africa, early park-based conser vation was derived largel y from Western roots, from the Yellowstone Model (Runte, 1997). Applying the wilderness model after Yellowstone National Park, the African continent has seen a dramatic growth in parks in the twentieth century, both in number and area protected. This type of conservation was built on strict exclusion of humans, the prevention of consump tive use, and minimization of other forms of human impact within park boundaries (Hulme and Murphree, 2001). Park managers believed that humans must be taken out of the equation. Boundaries were erected and even though the park itself may have been part of a continuous landscape, park managers sought to separate it from the landscape surrounding it. People were m eant to use resources outside of the park, and not to disturb the plants and animals were meant to stay inside the park Outside parks in Sub-Saharan Africa, the ma jority of the population is dependent on the land for their livelihoods. They depend on the land for growing food, economic opportunity, building materials for their homes, and to meet energy needs. For example, in Uganda, over 80% of the land used for small-scale farming and nearly 80% of the population are farmers (Mukiibi, 2001a). As a result of exclusion from hunting, foraging, and gr azing lands or areas of cultural or spiritual significance, park neighbors must go elsewhere to sustain their livelihoods or face punitive measures if caught poaching within park boundaries. Furthermore, the presence of a park represents a barrier to settlement and fu rther extensification, affecting farmers land use and livelihood options. The so aring population charact eristic of most Su b-Saharan countries 32
further intensifies the issue. Th us, land use strategies are altered, often incr easing the pressure on land and reso urces. Patterns of land cover change in most tropica l developing countries are closely linked to these land use changes and anthropogenic im pacts. Land use intensification and extensification are comm on responses not only to the limitati on and constricti on of resources, but also to the opportuni ties that the unused natu ral areas (wetlands and forests in this case) represent (Geist and Lambin, 2 001). All methods of intensif ying agriculture concentrate resources in time and space. However, the varying responses of diffe rent economic and ethnic groups to crop loss and resource needs can have quite different effects on the spatial organization of land use (Stone, 1996; Kago ro-Rugunda, 2004). Landscapes around parks have become mosaic s of natural and huma n-influenced patches and have been modified for thousands of y ears by humans. The once continuous natural habitats are becoming increasi ngly fragmented. This land scape outside protected areas encompasses edge effect s, population growth, and changing patt erns of land use, among other factors. Pressure on residual or unused lands (e.g. forest fragments) continues to mount as local populations increase. In addition, most protected areas are already too small to conserve most species of biodiversity (McNeely, 1994), and the severity and occurrence of human-wildlife interactions are increasing. Th ese aspects will infl uence the livelihoods of the neighboring communities. Studying the effects of human activities on land use and land cover typically involves the analysis of remotely sensed and other spatial da ta (Rindfuss et al. 2003). Remote sensing using satellite imagery provides robust techniques to efficiently observe and monitor the changes that occur within different land cover types over multip le dates at multiple lo cales over a large area 33
using land-cover classifi cation maps (Lu, 2004). Continuous data anal ysis can supplement discrete analyses to examine more subtle, within-class variability (Southworth et al. 2004). In the heterogeneous la ndscapes around many park s, land parcels can be relatively small and land-use is highly divers e, continuous data analyses can provide more revealing spatial analyses and focuses more on biophysical indicators. NDVI captures photosynthetic activity, and thus is used in this study as a proxy of vegetation pr oductivity and quality (Ser neels et al. 2001; ReyBaenayas and Pope, 1995), with higher values of NDVI indicating increased vegetation productivity and health. In add ition remote sensing techniques can be used to analyze spatial patterns that are difficult to capture solely at the household leve l. Fragmentation analyses are useful in understanding the spatial pattern and arra ngement of patches wi thin the landscape, providing critical insights for biodiversity and processes impacting land use change (Forman, 1995; Marsh et al. 2003; Nage ndra et al. 2006; Southwor th et al. in review). This study examines the impact of park establishment on the landscape. Kibale National Park (KNP) in western Uganda is an important case study not only beca use of its biological diversity, but also because over 37 years of c ontinuous research has been conducted within KNP and still little is known about land scape change in th is area. As population outside the park swells and people are prohi bited from resource extraction inside its boundary, the demand for resources and agriculture land not only increas es, but competes. As a result, the remaining unprotected natural areas wetlands and forest fragments become targets for degradation and conversion. We address whether conservation of land inside the park an d exclusion of access is related to fragmentation and res ource degradation. We further examine the spatial extent and quality of these diminishing resource bases at the landscape level using discrete and continuous 34
data analyses of satellite imagery and place the results within the social context of western Uganda and fortress conservation. Despite their ecological impor tance, much of the research addressing forest fragments has focused on Amazonia and So uth America (Chatelain et al. 1996; Laurance and Bierregaard, 1997). Studies on fragmented lands capes outside protected areas have focused on biodiversity conservation to the detriment of research into hu man-environment linkages (Turner and Corlett, 1996). Previous studies have focused on the effects of fragmentation on animal and tree species (Stouffer and Bierregaard, 1995; Chapman and Chapman, 1999; Chapman and Lambert, 2000; Gillespie and Chapman, 2006). While the use of wetlands and forests as a means to sustain livelihoods is well docum ented elsewhere in the world, East Africa is less prominent in the literature (Turner and Corlett, 1996; Marsh et al. 2003) as are studies on the smaller, interstitial wetlands and forest fragments character istic of western Uganda. Little research has addressed anthropogenic influences of land-cover change, or the resulting effects of altered ecosystems back onto the communities. Study Area Kibale National Park in wester n Uganda is illustrative of the agricultural expansion and intensification surrounding protecte d areas. Kibale Forest was de marcated in 1932 as a Forest Reserve and elevated to national park status in 1993 (Struhsaker, 1997). Known for its diverse primate population, KNP ranks fifth in terms of sp ecies richness and sixth in overall biodiversity importance among all Ugandan forests (Howard et al. 2000). It is perhaps most well-known because it is home to the largest known community of chimpanzees. The park is an ecological island a dens e, closed-canopy forest surrounded by a large agricultural population, larg e tea estates, and a vast network of wetlands and bottomland forest fragments. KNP is a medium-altitude tr opical moist forest covering about 795 km2 in western 35
36 Uganda (Figure 2-1) 2 This transitional forest (between lowland rainforest and montane forest) is at an average elevation of 1110-1590m and is a remnant of a previously larger mid-altitude forest region (Struhsaker, 1997). The climate is warm throughout the year, with an average range of 15-23C (Struhsaker, 1997). Forests and woodlands cover approximately 4.9 million hectares in Uganda and nearly 60% remain unprotected (NEMA, 2001). Nearly three million hectares of unprotected forests are vulnerable to overexploita tion and agricultu ral encroachment. Fo rest clearing throughout Uganda became extensive in the last 200 years, but NEMA (2001) reports that the deforestation has been particularly acute in the last 50 year s. At the estimated current rate of forest conversion, Uganda continues to lose 50,000 ha of its forest annually (NEMA, 2001). The Kabarole District in western Ug anda, where KNP lies, is a prime example of this forest loss and subsequent landscape fragmentati on. The once dominant moist, evergreen, closed-canopy forest that dominated the region has for the most part succumbed to agricultura l expansion, especially since 1959 (Gillespie and Chapman, 2006). Nearly all of the forests found on potentially arable lands were converted to small-scal e agriculture, tea, or pasture. The remaining forests typically occur in valley bottoms or on the steep rims of crater lakes (Gill espie and Chapman, 2006; Hartter, pers. obs.). Despite their loss, such unprotected forest fragments have been shown to be important for carbon sequestration, species habitats and corridors (Onderdonk and Chapman, 2000; Laurance and Bierregaard, 1997), as well as for commerciala nd subsistence-base activities. 2 While officially KNP now includes a game corridor that connects the so uthern portion of KNP to Queen Elizabeth National Park, this corridor was formally gazette d as part of KNP in 1993 but has not been cleared of human settlement and treated as part of the park until the last year and currently replanting etc. is still in progress. Before that time, it was mixed agricultural land. Therefore, it was not classified as park in this analysis. By doing so, our analysis is conservative since the addition of the corridor would tend to lessen the difference between park and non-park. Excluding the game corridor, the Kibale Forest in this analysis contains 561km2.
37 Ugandas wetlands comprise about 12% of the countrys total land surface (Mukiibi, 2001a) and fill a vital ecological role in flood abat ement, groundwater recharge, and as natures biological filter. In th e past, people treated wetlands more like wastelands (Mukiibi, 2001b) or as a repository of land for poten tial agricultural expansion. Wetla nds have vital importance in maintaining the quality and volume of surf ace and ground water (Mukiibi, 2001b). Several studies have attempted to quantify wetland ecosy stem services in terms of human welfare (Costanza et al. 1997; Batagoda et al. 2000; Turner et al. 2001; Ba lmford et al. 2002). They are important to livelihoods of surrounding commun ities in several ways. They provide a subsistence resource base for water, thatchi ng, handcraft materials, and local medicines. Wetlands are also commercially important through sales of these materials to community members and also from visiting tourists. They are key spawning grounds for fish (Crisman et al. 2003), and provide other services su ch as biodiversity, aesthetic b eauty, and cultural heritage. Rapid population growth, high population dens ity, and heavy reliance on small-scale agriculture and natural resources to sustain livelihoods ch aracterize the social landscape surrounding KNP (Archabald and Naughton-Treves, 2001) The elevation of Kibale to national park status in 1993 meant that nearly all form s of commercial and s ubsistence based resource extraction and all settlements were prohibited. Resource unavailabi lity inside the park coupled with increasing population has caused land pressure outside the park to increase. Wetlands are under serious threat of conversion due to population growth, in-migration, a nd intensification of agriculture in the surrounding landscape. Agriculturalists in the area bel ong to two dominant ethnic gro ups: the Batoro (west side of KNP) and the immigrant Bakiga (east side of KNP) (Figure 2-1), who came to the Kibale region from southwestern Uganda beginning th e 1950s and 1960s (Turyahikayo-Rugyema, 1974;
38 Naughton-Treves, 1998). A 1988 census estimated the population within the park and the surrounding communities to be 61,000 (Aluma et al. 1989). In the Kabarole District, where KNP lies, 95% of the population sustai ns their livelihoods through agri cultural-based activities. This region is one of the most densely populated areas on the African cont inent (Lepp and Holland, 2006), with a district populati on density of 92 individuals/km2 (NEMA, 2001). NaughtonTreves (1998) reported that population around KNP more than tripled between 1959 and 1990, with the current population estimated to be between 270 and 315 individuals/km2 (Hartter, unpublished data). Since land is a premium with increasing population, the remaining unprotected wetlands and forest fragments are sought after. Those wetl ands and forest fragments that remain continue to be relied upon to produce resources for subs istence and commercial use. Not only is the presence of forests and wetlands diminishing, but they are also being degraded, reducing the amount of large trees (Kayanja and Byarugaba, 2001). Therefore, the landscape surrounding the park must change and the unprotected forests and wetlands outside the park must change as their produce is vital to community liveli hoods. land scarcity has forced many farmers to farm at the forest edge (Naughton-Treves, 1998) Typical farm sizes are less than 5 ha (Hartter, unpublished data) and farmers plant more than 30 species of subsistence and cash crops). Most of the crops produced are consumed at the household level, with few bei ng sold outside of immediate household or local community markets. Methods Landsat TM imagery was chosen because it offers the best combination of spatial, spectral, temporal and radiometric resolutions. Three dry-season images have been acquired: May 26, 1984, January 17, 1995, and January 31, 2003 (path 132, row 060). The first image provides baseline data prior to official park establishment (it had been a protected area as Kibale Forest
39 Reserve prior to this point), the second captures conditions at park establishment, and the third represents current conditions. All images underwent standard preprocessing of image calibration to correct for differences in time, sensor and atmospheric conditions, and were co-registered to within an RMS of <0.5 (below 15 m accura cy) (Southworth et al. in review). Variation in Interannual Rainfall While all images have been geo-rectified and radiometrica lly corrected to account for sensor drift and differences in solar angle and other atmospheric conditions, precipitation trends must also be considered. Satel lite image analysis is especially susceptible to differences in precipitation. The normalized difference vegetation index (NDVI) has been proven to be a robust measure of vegetation attributes (net prim ary productivity, green biomass, and green leaf area index) (Serneels et al. 2001) NDVI has also been found to be well correlated with climate variables, such as evapotra nspiration and precipitation (A nyamba et al. 2001). Healthy vegetation will have higher evapotranspiration rate s. In addition more precipitation can increase vegetation productivity. Richard and Poccar d (1998), Eklundh (1998) and Davenport and Nicholson (1993) have examined the relationship and subsequent sensitivity of NDVI to interannual rainfall. Their work has concluded that there is a high correla tion between vegetation performance measured by NDVI and rainfall in areas where rainfall is limiting (<900mm) and those that have a distinct dry and wet season. Serneels et al. ( 2001) also found that NDVI was a good indicator of the impact of inter-annual climate va riability on vegetation conditions in their study site (weakly bi-modal rain fall pattern, with a gradient of 500-1200mm/yr). However, in his meta-analysis of net primary productivity (N PP) and global climate in the wet tropical forests, Schuur (2003) found that net primary pr oductivity becomes less sens itive to mean annual precipitation at high precipitati on levels. Schuur (2003) found that as precipitation approaches
40 2445 mm annually, NPP levels off a nd then declines. Wang et al. (2001) also report the lowest correlations of NDVI to differences in precipitation with forest. Precipitation in the Kibale region is extremel y local. Although the amount of rainfall and length of season change, the average annual rainfall for the region is 1543mm (average 19031999) and 1719mm (1990-2006) (Chapman et al. 2005). The bi-modal rainfall pattern produces two major rainy seasons. The long rains are betw een late February and early May and the short rains occur between late A ugust to early December. This analysis focuses on two land covers: fo rest and wetlands, both of which are not limited in available soil moisture in this regi on. NDVI is more variab le in more seasonal vegetation such as field crops and pasture grass, but the rema ining unprotected forests outside the park tend to be in bottomlands with enough av ailable soil moisture. Mo st of the wetlands are permanent wetlands as the seasonal ones tend to be converted first by the local population for smallholder agriculture. The papyr us wetlands are not seasonally quiescent and are permanent wetlands, interannual rainfall variation would ha ve little impact. In addition, interannual phenological responses in forests in the Kibale region re sult in differences in fruit production and food availability (Chapman et al. 2005). In this region, the forest canopy remains intact. Three Landsat images are used in this analysis However, an annual time scale may not be the best scale to examine precipitation because of the bi-model rainfall pattern. While annual trends may be identified, a more appropriate time scale is defined by the local social and biophysical conditions. Trees and plants processes are determin ed not by calendar years, but by precipitation cycles in this area. Planting a nd harvesting by the local smallholder agriculture populace are based on rainfall patterns. Therefore, it is appropriate to examine monthly precipitation trends as well as precipitation to tals for both the long and short rainy seasons.
41 Figures 2-2 and 2-3 shows the annual and mont hly precipitation respectively and bi-modal distribution for the park and the years preceding image acquisition. It is important to note that for the image capture dates, the total monthl y precipitation for that month is <50mm. The January 1995 and 2003 images were captured after the long rains (Figure 2-4). The 1995 and 2003 are not anomalous and their monthly precipitation falls within 1 standard deviation above/below the mean. The 1984 image wa s captured following the short rains. It is appropriate instead to examine the trends in th e short rains. The mont hly precipitation for the 1984 image was also within 1 standard deviation of the mean for the short rains. Each of the images were captured following abundant rainfall during the rainy seasons and all three images are not anomalous (not in peak or trough precipita tion months). If the area was precipitation limited, we would expect that ther e would be a distinct seasonal difference. However, the Kibale region, receiving more than 1500mm of rainfa ll annually, is not precipitation limited. Serneels et al. (2001), Richard and Poccard (1998) report the sensitivity of NDVI to inter-annual rainfall in rainfall limited regions (<1200mm/yr and <900mm/yr respectively). There is a dis tinct dry and wet season in the Kibale region, but local farmers report that these seasons are becoming less and le ss distinct and increasingly blend into one another. In addition, seasona lity has little, if any, impact on wetland and forest vegetation. Thus, we can conclude that vegetation c onditions represented in the 1984, 1995, and 2003 Landsat images are comparable, and the use of NDVI will be an appropriate measure of long term vegetation changes, not intera nnual variation in precipitation. Land Cover Analysis During the 2004 and 2005 field seasons 180 traini ng samples were collected and used to construct a supervised classifica tion. The classified image was constructed using a layer stack including all bands 1-5, 7, texture bands of layers 1-5, 7, plus an NVDI layer. The final
42 classification obtained an overa ll accuracy of 89.1% and an ov erall kappa statistic of 0.867 (Southworth et al. in review). Three classes were used in the cl assification: 1) forest, 3) wetland (papyrus [ Cyperus papyrus L.] and elephant grass [ Pennisetum purpureum]); 3) other. The land-cover analysis was al so used to determin e the amount and spa tial distribution of loss or increase of wetla nd and forest classes over time. This allows us to address is sues of the size and location in the landscape of patche s of different land covers across the different dates, as well as to determ ine conversion fro m and to classes, e.g., conversion from forest to agriculture. In addition, NDVI composites from ea ch date were used in a separate analysis of mean NDVI values over time to quantify the amount and type of change in forest and wetland productivity from 1984 to 2003. NDVI was examined by land cover class of interest, so changes in land cover class do not compromise the change in NDVI values. Fragmentation Analysis Landscape metrics were used to identify tren ds in landscape heterogeneity over time. Fragstats 3.3 (McGarigal and Marks, 1995) was used to calculate these metrics. We wanted to focus the analysis on natural forests and not sm all kitchen gardens, spectral confusion perhaps with banana plantations, or small household stands of Eucalyptus spp ., Maesopsis eminii Grevillea robusta and other species grown predominantly fo r fuel and/or sustenance. Therefore, only forests greater than 0.5ha in size were considered. Results Land Cover Analysis Land cover classifications from the three im age dates (Figure 2-5) show the difference between in-park and in the surround ing landscape. Over time, there remains little difference in forest cover inside the park.
43 Figure 2-5 also shows the difference in land cover allocations between the east study area, dominated by the Bakiga ethnic group, and the we st study area, dominated by the Batoro ethnic group. The main difference between these two areas is the presen ce of tea and lower percentage of wetland in the west study area. The east side has about 55% of its land under natural areas (forest + wetland) and the west si de has 45% natural areas. Overall, there is a loss in forest cover and papyrus and elephant grass and a corres ponding increase of la nd under cultivation. Productivity Assessment Table 2-3 shows the NDVI values for forest and wetland across time. Mean NDVI values are decreasing over time in both forests and we tlands. The decreasing trend of NDVI values indicates the decrease in forest and wetland productivity over time both in the park and surrounding landscape. Fragmentation Analysis Tables 2-2 and 2-4 compare the pattern of the different land covers in the surrounding landscape to the park. Within the park, the dom inant land cover is forest. Outside the park however, there are three dominant land covers forest, wetland, and cr ops with the highest proportion of land under cultivation. Compared to the park, the natural areas wetlands and forest are smaller, more plentiful, and have much more edge in th e surrounding landscape. Inside the park, there is a trend of consolid ation; higher clumpiness and connectivity values compared to the surrounding landsca pe that is more fragmented. The trends in the east and west study areas ar e similar to the entire surrounding landscape (Tables 2-4, 2-5, and 2-6). From Tables 25 and 2-6 and Figure 2-6, both study areas are becoming increasingly fragmented. Clumpiness and connectivity values decrease over time while distance to the nearest patch edge increases. Comparing the two study areas, mean patch size is less on the east side, and there are fewer patches of forest and papyrus and they have less
44 edge. In addition, the percentage of the landscape comprised by the largest patch in the east side is more because of the large continuous Magom be Swamp that is protected through a community conservation program Kibale Association for Rural and Economic Development (Lepp and Holland, 2006). The clumpiness, co nnectivity, and cohesion values for both the east and west study areas are similar. In general, the fo rests and wetlands in these areas are becoming increasingly isolated and clumped toge ther in the east and west study areas. Discussion Park boundaries are relatively in tact and the forest cover with in Kibale National Park has not decreased over time and has in fact become mo re consolidated over time (less fragmented) in contrast to what has happened in the landscape surrounding the park. Our results indi cate little change of forest extent inside the park si nce 1984. Mulley and Unr uh (2004) report in their analysis of KNP and the surr ounding landscape from 1955-2001 that most of the park remained forest, while nearly all of the deforestation t ook place outside the park boundaries. Outside the park, there is evidence of continued active conve rsion. The clearing of wetlands and forests is highly variable in each fragment because of lo cal land tenure, resource needs of households, and the number of households de pendent on the particular fore st fragment, human-wildlife interaction, loca l geography and ecolog y, and other factors. Outsid e the park, a higher number of fragments will have increased access by more hous eholds. In turn, more fragments are used. These new fragments become more isolated and edge density increases as they are used. Larger fragments become smaller and smaller as the con tinuous landscape is broken up. This process of increased fragmentation and opening up the interior of remaining fragments leads to degradation, more open/disturbed forest and towards seconda ry vegetation (Schelhas and Greenberg, 1996). Chapman et al. (2006) report a decrease both in ba sal area and stem density (stems>10cm diameter at breast height (d bh) between 2000 and 2003 in forest fra gments located on the
45 western flank of KNP. At the broad landscape level, Mugisha (2002) found that the landscape surrounding Kibale National Park showed a dram atic land cover change from 1955-2000. This landscape showed the largest expansion of smallholder agri culture by approximately 137%, while at the same time there was a decrease of wetlands by 19% and forest by nearly 20%. While this change is notable, Mugisha (2002) does not define the extent of the study area boundary outside the park. Alt hough less dramatic, our analysis shows the overall decrease in forest and wetland and the corresponding rise of cultivated land a nd tea (Figure 2-6). The loss of extent and decrea se in productivity of wetlan ds in the land scape surrounding the park has important implicati ons, both environmental and social. Laurance and Bierregard (1997) describe forest loss and fragmentation as the mo st pressing threat to global biodiversity. While more natural areas are converted to culti vated lands to support the growing population and their resource and econ omic needs, the landscape surrounding KNP has become increasingly fragmented. As a result, the amount and density of edge has likewise increased. Edge effects are widespread, co mplex, and unique to the landscape matrix, but the effects of edge cannot be ignored. Hill and Curran (2005) found that area of a fragme nt was the most important influenc e on tree species and composition in Ghanaian forest fragments. Furthermore, Gascon et al. (2000) report that those fragments le ss than 5000 ha, are especially vulnerable to edge effects. Surrounding KNP th e mean patch size is substantially smaller (10 ha). The matrix surrounding th e remaining natural areas land and resource pressures and the intensified land use can result in edge effects pe netrating further into th e forest (Gascon et al. 2000). As the wetland and forest fragments steadily decrease and edge effects penetrate further, the result is an impoverished interior. This impoverishment has further implications.
46 As the population outside KNP c ontinues to cut forests or dr ain wetlands seeking land and resources and to prevent crop ra iding, long term detriment is caused to the lands. There is anecdotal evidence to suggest that the increased isolation of the forest fragments within the intensely cultivated land scape is related to trees dying out in the fragments. Even though some of the fragments remain, many farmers have noticed that bigger trees are dying and they say that it is taking longer for the endemic species to gr ow to even pole size. Fragmentation not only lowers species number, but also alters comm unity composition and their ecosystem processes because of the reduction in size and the change in shape (Hill and Curra n, 2003). Laurance et al. (2000) found that forest fr agment size and tree species diversit y were directly related. As the amount of edge is increased and interior decreas ed, forest regeneration is slowed. The forest edge becomes degraded and dispersed with weedy species that are useless for home consumption. Fragment size is important. Higher mortality rates in forest fragments are found in smaller fragments in the Amazon (Laurance et al. 2000) and Laurance and Bierregard (1997) found that the number of ra re species increased with fragment size. The loss of extent and decrea se in productivity of wetlands in the landscape surrounding the park has important implications. Not only doe s species richness and diversity increase with loss of wetlands and forest, but the ability of these forests and wetlands to fulfill important ecological processes is severely handicapped by increased fragmentation, loss, and degradation (Tinker, 1997; Laurance, 2004). Bolwig et al. (2 006) suggest that fragmentation can lead to a loss of biodiversity to as high as 2.5% in some agroecosystems. The presence and quality of forest fragments has been found to affect rainfa ll, microclimate, plant and tree species abundance and diversity (Kapos et al. 1997). There is a loss of net primary productivity, carbon
47 sequestration ability, nutrient level maintena nce and ecosystem maintenance and pollutant mitigation (Coops et al. 2004). The degradation, decrease productivity, and in creased fragmentation of the remaining wetlands outside the park have si milar environmental impacts. As wetlands continue to degrade in size and quality, their ability to control erosion and floods a nd to maintain water and volume of water likewise decreases (Chapman et al. 2001). Less papyrus productivity has social and bio-environmental impacts. A degraded wetland will produce less aboveand below-ground biomass, thus limiting the ability to filter and recharge wate r. Jones and Humphries (2002) showed that papyrus swamps have the potenti al to sequester large amounts of carbon (1.6 kg C/m2/yr) when detritus accumulates under water in anaerobic conditions. These swamps also may be a net source of carbon release to the atmosphere (1.0 kg C/m2/yr) when water levels fall to expose detritus and rhizomes to aerobic conditions. The relatively homogenous stands of papyrus often extend over large areas up to tens of kilometers in western Uganda (Chapman et al. 2001). Therefore, wetlands in the KNP regi on represent potentially large carbon sinks. In addition, the continued conversi on producing fewer and less producti ve wetlands is reducing the overall ability of the landscape to sequester carbon and may instead become carbon sources (De la Cruz, 1986). The impacts of size, shape, and isolation of forest fragments are not limited to vegetation and ecological processes. The ecological relationships are complex. As the micro-climate is altered, changes in solar radi ation, humidity and wind pattern that are important for many organisms can impact viability and survival (Ranta et al. 1998). Biological attributes such as predator and prey relationships, habitat, migra tion corridors, and species survival probabilities are also impacted (Laurance et al. 2002). G illespie and Chapman (2006) report that fragments
48 with higher stump densities (those with higher an thropogenic influence) strongly influenced the prevalence of parasitic nematodes in primates. Habitat fragmentation can also cause local extinctions, but also can have long term effects on populations through changes in pollination, predation, and food availability. In forest fragments outside KNP, red colobus abundance ( Procolobus rufomitratus) declined with the loss of food trees (Chapman et al. 2006). Landscape fragmentation and loss and degradat ion of wetlands and forests outside KNP also have important social implications. Th e degradation of habitat, breeding grounds, and nutrition can cause wildlife to seek supplementa l sources. Farms neighboring the park become easy targets because of the abunda nt food supply and the high probability of a successful raid. The increased fragmentation of the landscape causes further detriment because the limited natural areas that act as co rridors through which species such as the vervet monkey ( Cercopithecus aethiops ), the redtail monkey ( Cercopithecus ascanius), and olive baboon ( Papio anubis ) travel. Those fragments that remain act to channel the primat es especially through agricultural lands becaus e of the narrow passages connecti ng remaining wetlands and forest. Furthermore, the wetlands and fragments are ope ned up for human intrusion, thus escalating the number and severity of human-wildlife encounters (Hill, 1997; Naughton-Treves, 1998). Landscape metrics measuring landscape connectiv ity are important tools in understanding the impacts of wildlife. Connectivity is determined based on threshold values. A threshold of 60m was set initially, which is quite small. In real ity, we found that the vervet, redtail, and Lhoesti monkeys ( Cercopithecus lhoesti ) and baboons raided fields over 500 m from the forests and wetlands. Elephants ( Loxodonta africana) were found to raid fields at distances less than 400 m from these natural areas (Hartter, unpublished data). As thre shold values were increased, forest, wetland, and total natural area (tha t is, forest + wetland) connec tivity increases (Figure 2-7).
49 A less productive and smaller wetl and and forest area also means a reduced ability to filter toxins from the drinking supply or to control erosion. Regeneration in forests may be slowed or arrested because of lower nutrient levels, larger more common gaps, human intrusion that lead to topsoil erosion or younger trees may be out-com peted by lower quality and/or weedy pioneer species such as elephant grass and the sub-woody shrub Acanthus pubescens (Chapman et al. 1999; Chapman and Chapman, 1999). As the ability to conduct ecosystem f unctions is reduced, so will the availability of resources to meet subsistence needs. Wetlands dry up and produce less papyrus. Forests are high-graded for their hardwoods, medicinal purposes, building materials, musical instruments, or for other reasons. In the process, the large, stra ight, and vigorous tree species that are important seed trees are removed. Unprotected wetlands in Uganda have intensively been used to extract resources or have been drained to cultivate crops. As the outer edges are cultivated each growing season, the wetland continues to decline in size and water levels recede. Wh en converted, the detritus-based ecosystems of the papyrus swamps initially have fer tile soils. However, as the soil is exposed, it begins to oxidize, leading de pletion of soil nutrients and in creased soil acidity (Cooper 1975, Crisman et al. 1996). Not only does soil moisture and fertility dramatical ly decrease after the first few growing seasons, but al so it is difficult to reclaim abandoned drained papyrus wetlands. Forests have been cut for resources and to make way for crops, pasture, or to reduce crop-raider habitat. Overexploitation occurs in the context of harvesting re sources. In wetlands, clay for brick building, and pottery and papyrus for th atching houses and making mats are commonly extracted (Chapman et al. 2001; MacLean et al. 2003). In forests, indigenous medicines, building poles, timber are taken. Firewood is also procured for fuelwood or for firing bricks.
50 The shortage of fuelwood has also intensifie d the situation. Ninety -five percent of all Ugandas energy needs are met with fuelw ood and charcoal (MacLean et al. 2003). The hardwoods are gone and households must turn to softer exotics to meet their needs. As a result, many households are growing fuelwood and timber trees in drained wetl ands or converted bottomland forests. In the 1950s, Eucalyptus spp. were introduced in many parts of Uganda as a solution to wood shortages, because it is fa st-growing, coppices, and can provide poles and timber as early as three and eight years respectively after planting. While it has been successful at providing resources, Eucalyptus has also become problematic. Eucalyptus water uptake demands are substantial (Shiva and Bandyopadhya y, 1983). Many households in the Kabarole District in western Uganda for instance, complain of decreased water levels corresponding to the increase of Eucalyptus trees planted at the edge of wetlands or within drained areas (Hartter, unpublished data, Richardson 1993). The current situation of populat ion increase and land and res ource pressures has created a negative feedback loop in the landscape su rrounding KNP. The more the areas become exploited and overused to get resources and to cultivate crops and reduce the threat of crop raiding wildlife, the more degraded the forest and wetland remnants become, and in turn their ability to provide abundant and high quality resources in the future is compromised. Despite their social and ecological uses and the broader implications of degradation on livelihoods, forest and wetland conve rsion continues. Chapman et al. (2006) concur, concluding that many of the forest fragments surrounding KNP will disappear in the coming years. Conservation value is low for the small unprote cted natural areas because they cannot support viable populations of primates and other species in the long-term. This has implications on monitoring because it requires the location, extent and pattern of land cover change to be
51 known. Identifying hot-spots is an important pha se in allocating limited resources in wetlands and forest management. A mixed-methodological approach is necessary to examine landscape change from a broad spatial scale, but also to understand the fine-s cale implications on biodiversity and households. Future research is needed to further examin e the changes in land cover extent, pattern, and productivity. It is evident that productivity is decreasing both inside and outside the park and new vegetation indices should be developed desc ribe within-class variation. In addition, Southworth et al. (in review) described the difficulty in diffe rentiating elephant grass and papyrus because of similar spectral signatures. This gives a conservati ve land cover assessment by including elephant grass with papyrus together in one land cover cla ss. Southworth (2004) and Boyd et al. (1996) note the usefulness of usin g the Landsat thermal band for assessing forest regeneration in tropical dry forests. The therma l band can be incorporated in a new analysis to examine within-class variation of forest to investigate loss of productivity within park vs. surrounding landscape. In addition, papyrus appear s exclusively in valley bottoms and elephant grass many times, but not always, is found in th e uplands. New vegetation indices, surface temperatures, and a digital elevation model can be incorporated to create a new rule-based classification to improve overall classification accuracy. KNP is an important park not only because of its biological diversity, but also becau se of the islandized nature of the park. KNP represents an important litmus test for the im pacts of park establishment on the surrounding landscape in mid-elevation moist tropical rainforests, typical more of central Africa (e.g. Gabon, Democratic Republic of Congo). Th ese results should also be compar ed to park landscapes that are more precipitation limited such as in southern Africa savannah landscapes.
52 Conclusion Parks have long served as centerpieces of the conservation movement. However, their establishment is not benign and can have las ting impacts on the landscape and livelihoods of people living near them. As population outside the park swells, the demand for resources and agriculture land not only increases, but compet es. Remaining wetlands and forest become targets for degradation and conversion. Remote sensing techniques and fragmentation analysis were useful tools to address th e amount and location of change a nd also provided the ability to compare within and outside park boundary ch ange. We have shown the landscape surrounding the park has become increasingly fragmented since 1984. As fragmenta tion outside the park continues, the forests and wetlands in th e surrounding landscape not only are becoming increasingly smaller, but increasingly isolated and decreasing in productivity, producing both environmental and social implications. Characteristic of fortress conser vation is its measures of success. To this movement, the status of plants, animals, and the ecological co mmunity as a whole is the ultimate measure of successful conservation (Struhsak er, 2002). KNPs boundaries remain relatively intact, while the surrounding landscape has become increasingly frag mented as land and resource pressure builds outside the park over time. While it could be argued that KNP is a muted success story for fortress conservation, the reduction in forest a nd wetland productivity should not be ignored and further research should address the phenomenon. Parks are associated with landscape change. Whether the exclusive nature of such park s for the promotion of biological diversity conservation can hold out in the face of mounting pressure to sustain human livelihoods is an important question that must be addressed.
53 Figure 2-1. Kibale National Park and surrounding landscape in western Uganda
54 Figure 2-3. Monthly precipitat ion data for the three image years (May 1984, January 1995, and January 2003) and the years preceding image capture (1983 data was unavailable) at the Kibale research station (Kanyawara). Annual Precipitation 1984-20051000 1500 2000 2500 198319851987198919911993199519971999200120032005YearPrecip. (mm) 1536.1 Figure 2-2. Annual precipitation co llected at the Kibale research station (Kanyawara). Annual precipitation totals for image capture year s are noted. No data was acquired during September 13-21, 2004 while the broken rain gauge was repaired. Mon 0 50 100 150 200 250 300 350 JanFebMarAprMayPrecip. (mm thy Precipitation (mm) JunJulAugSepOctNovDec Month ) 2003 2002 1995 1994 1984
55 Rainy Season Monthly Preciptation 1984-20050 200 400 600 800 1000 1200 1984198519861987198819891990199119921993199419951996199719981999200020012002200320042005YearTotal Precip. (mm) Short Long Mean short Mean long +/1 s.d. Jan 17, 1995 Jan 31, 2003 May 26, 1984 Figure 2-4. Precipitation totals for the short and long rainy seasons between 1984 and 2005. Mean values and +/1 standard deviations are given for each rainy season.
Land Cover for Surrounding Landscape0% 20% 40% 60% 80% 100% 198419952003 Year Other Wetland Forest Land Cover for Kibale National Park0% 20% 40% 60% 80% 100% 198419952003 Year Other Wetland Forest Land Cover for East Study Area0% 20% 40% 60% 80% 100% 198419952003 Year Other Wetland Forest Land Cover for West Study Area0% 20% 40% 60% 80% 100% 198419952003 Year Other Wetland Forest Figure 2-5. Land cover change in -park and in the su rrounding landscape, in cluding the east and west study areas Land Cover Change in Surrounding Landscape0 2000 4000 6000 8000 10000 1984 1995 2003YearArea (ha) Forest Wetland Other Figure 2-6. Land cover extent in the surrounding landscape 1984-2003 56
Connectivity of Natural Areas, East Study Area 0 5 10 15 20 6025050010001500 Threshold (m)Connectivity 2003 1995 1984 Connectivity of Natural Areas,West Study Area0 5 10 15 20 6025050010001500Threshold (m)Connectivity 2003 1995 1984 Figure 2-7. Connectivity of natural areas in east and west study areas 57
Table 2-1. Definitions of landscape metrics incorporated within this analysis of fragmentation of forest and wetland patches, from 1985, 1995 to 2003 Landscape Metric Description PLAND Proportion of the landscape covered by each land cover LPI Percent of landscape comprised by largest patch NP Number of patches MPS Mean patch size (ha) ED Edge density is the sum of all edge segments, divided by total area for each class (m/ha) ENN Euclidean distance to nearest patch edge (m) Connectivity (range 0 to 100, expressed as a percent). Measurement of connections between patches of the corresponding patch type divided by the number of possible connections. A value of 0 is when either the class consists of a single patch or none of the patches of the focal class are connected; and a value of 100 indicates that every patch of the class is connected. Cohesion (range 0 to 100). Measuremen t of physical connectedness of the corresponding patch type. Values tend towards 0 as the proportion of the landscape comprised of the class decreases and becomes increasingly subdivided and less physically connecte d. Values increase as patch type becomes more clumped or aggregated in its distribution. Clumpiness (range -1 to 1) Measurement of the aggregation of patche s within a class. A value of -1 is when the patch type is maximally disaggregated; 0 when the focal patch type is distribu ted randomly; and when the value approaches 1 when the patch type is completely aggregated. Note: A complete description of these and other metrics are provided at http://www.umass.edu/landeco/resear ch/fragstats/fragstats.html. Table 2-2. Inside-park fr agmentation analysis Land Cover Year Proportion of Total Land Area (%) Largest Patch Size (ha) No. Patches Mean Patch Size (ha) Edge Density Clumpiness (-1 +1) Connectivity* (0-100) 198486.339.5467273.318.70.940.152 199585.525.6466873.112.850.950.122 200390.141.49412124.913.390.960.000 19846.00.3642670.813.160.630.010 19956.30.2369950.510.480.580.005 20036.60.38326188.8.131.520.000 *threshold at 60m Forest Wetland 58
Table 2-3. Forest and wetland productivity as assessed by NDVI (NDVI values between -1 and +1). Values close to +1 indicate higher productivity. YearLand CoverLocation Mean NDVI NDVI Std. Dev. NDVI min. NDVI max. 1984 ForestKNP 0.5910.0780.3150.805 1995 ForestKNP 0.5480.0550.1430.770 2003 ForestKNP 0.3600.0520.1330.559 1984 ForestSL 0.6440.0630.3380.799 1995 ForestSL 0.5800.0440.3330.778 2003 ForestSL 0.3980.0490.0430.586 1984 ForestEast SA 0.5970.0550.3700.759 1995 ForestEast SA 0.5530.0340.4000.664 2003 ForestEast SA 0.3780.0440.1190.508 1984 ForestWest SA 0.6820.0380.4620.799 1995 ForestWest SA 0.6110.0380.3780.730 2003 ForestWest SA 0.4100.0520.0850.586 1984 WetlandKNP 0.5610.0590.0970.759 1995 WetlandKNP 0.5060.0540.0710.706 2003 WetlandKNP 0.3030.061-0.1330.518 1984 WetlandSL 0.5560.0650.0130.767 1995 WetlandSL 0.4820.0490.1110.647 2003 WetlandSL 0.2630.075-0.2600.511 1984 WetlandEast SA 0.5330.0640.1400.708 1995 WetlandEast SA 0.4720.0360.2460.595 2003 WetlandEast SA 0.2740.057-0.2000.492 1984 WetlandWest SA 0.5980.0540.1010.757 1995 WetlandWest SA 0.5190.0390.1370.647 2003 WetlandWest SA 0.3050.059-0.2600.496 59
60 Table 2-4. Fragmentation anal ysis for surrounding landscape Year Land Cover Type No. Patches Edge Density (m/ha) Mean Patch Size (ha) Clumpiness (-1 +1) Connectivity (0-100) Distance to Nearest Patch Edge (m) 1984 Natural Areas 2733 35.48 14.98 0.8307 0.028 78.3 1995 Natural Areas 2912 34.63 12.70 0.8188 0.028 76.5 2003 Natural Areas 3879 44.68 9.82 0.7939 0.000 72.9 1984 Forest 1514 27.88 14.07 0.7725 0.039 106.2 1995 Forest 1426 29.35 16.08 0.7721 0.041 102.1 2003 Forest 1780 21.39 10.53 0.8235 0.000 118.5 1984 Wetland 9587 38.78 2.05 0.6637 0.008 77.3 1995 Wetland 10200 30.91 1.38 0.6297 0.006 84.5 2003 Wetland 8581 39.16 2.25 0.6868 0.000 80.0 1984 Other 9013 35.48 3.10 0.7714 0.008 76.4 1995 Other 6314 34.63 5.09 0.7981 0.012 75.8 2003 Other 10311 44.68 3.02 0.7582 0.000 74.2 Table 2-5. Fragmentation an alysis for east study area Year Land Cover Type No. Patches Edge Density (m/ha) Mean Patch Size (ha) Clumpiness (-1 +1) Connectivity (0-100) Distance to Nearest Patch Edge (m) 1984 Natural Areas 204 83.67 17.68 0.7422 0.425 72.8 1995 Natural Areas 327 99.14 7.81 0.6455 0.257 72.6 2003 Natural Areas 397 99.44 6.48 0.6475 0.000 69.0 1984 Forest 165 52.41 9.44 0.7359 0.244 97.5 1995 Forest 211 57.51 6.31 0.673 0.230 91.1 2003 Forest 207 35.51 4.34 0.7202 0.000 108.7 1984 Wetland 731 108.00 2.80 0.5612 0.136 69.9 1995 Wetland 884 66.44 1.38 0.5992 0.058 77.5 2003 Wetland 761 86.87 2.20 0.5917 0.000 72.5 1984 Other 949 83.67 1.82 0.6103 0.095 71.1 1995 Other 729 99.14 3.82 0.6605 0.149 66.2 2003 Other 575 99.44 4.81 0.6567 0.000 65.8
Table 2-6. Fragmentation an alysis for west study area Year Land Cover Type No. Patches Edge Density (m/ha) Mean Patch Size (ha) Clumpiness (-1 +1) Connectivity (0-100) Distance to Nearest Patch Edge (m) 1984 Natural Areas 299 62.92 21.58 0.7818 0.359 71.8 1995 Natural Areas 159 61.57 39.05 0.7824 0.661 72.5 2003 Natural Areas 816 63.06 5.13 0.7139 0.000 74.9 1984 Forest 284 61.10 14.97 0.7240 0.284 79.8 1995 Forest 118 71.29 44.46 0.7218 0.826 74.4 2003 Forest 395 41.20 7.25 0.7488 0.000 95.3 1984 Wetland 1890 56.32 1.17 0.5686 0.037 75.8 1995 Wetland 1646 29.62 0.59 0.5161 0.027 92.2 2003 Wetland 1871 39.59 0.71 0.5207 0.000 85.9 1984 Other 1499 62.92 2.56 0.6923 0.057 73.8 1995 Other 1433 61.57 2.85 0.7119 0.054 72.9 2003 Other 861 63.06 7.09 0.7728 0.000 68.7 61
CHAPTER 3 RESOURCE USE ANDHOUSEHOLD LIVELI HOODS: LIVING NEAR WETLANDS AND FOREST FRAGMENTS NEIGHBORING KIBALE NATIONAL PARK Introduction Landscapes around parks are important areas. The matrix outside a park, consisting of edge effects, population growt h, changing patterns of land us e, and others, will influence livelihoods of adjacent communities and shape th e surrounding landscape. They are potential reservoirs of land, resources, and economic opportun ity. Park landscapes ar e likely to attract migrants because of tourism-based activities, increas ed infrastructure for tourists and researchers, or because of other industries directly or indire ctly related to park es tablishment. Increasing population will likewise extend the need for land and resources. In addition, these landscapes are critical for biodiversity (McNeely, 1994). Patches of forests and we tlands outside parks serve as important wildlife corrido rs and critical habitat and f eeding grounds and resource bases for local communities (Barbier, 1993; Byron and Arnold, 1999). To secure their livelihoods, mo st of the rural population in Sub-Saharan Africa turns to the land. Land use intensification and extensific ation are common responses to limitation and constriction of resources. However, protecte d areas are commonly barriers to settlement, extensification, and reso urce extraction, which in turn affects farmers land use and livelihood options (Brockington et al. 2006; Wilkie et al. 2006). Since re source extraction and most forms of consumptive land use are often forbidden w ithin the park boundaries people intensify land use on their own land and turn to the only availa ble land to meet their needs. Unprotected wetlands and forest fragments become targets for conversion to agriculture, grazing land, or woodlots or are harvested at unsustainable levels In order to understand land cover change in the landscape surrounding a park, it is necessary to examine resource consumption patterns of 62
park neighbors. Resource use may vary by ethnicity, wealth, location, and gender (Rocheleau and Edmunds, 1997; Byron and Arnold, 1999; Goebel et al. 2000). Patterns of land cover change and landscape fragmentation in most tropical developing countries relate significantly to anthropogenic impacts (Nage ndra et al. 2003). To date, fragmentation literature has focused heavily on impacts to ecology and survival/degradation of species and habitat (Chatela in et al. 1996; Onderdonk and Chapman, 2000; Hill and Curran, 2003; Chapman et al. 2003), with a heavy geogr aphical preference to Amazonia and South America (Stouffer and Bierregaard, 1995; Laur ance and Bierregaard, 199 7; Marsh et al. 2003, 1997). Only limited work has acknowledged (and commonly, only briefly mentions) the role forest fragments play in pr oviding resources such as fuel wood, timber, poles, thatching, handcraft materials, indigenous me dicines, and edible plants to communities (Turner and Corlett, 1996; Gillespie and Chapman, 2005). Studies on fr agmented landscapes instead have focused on biodiversity conservation and larger fragments, thus to the detriment of research into humanenvironment linkages (Turner and Corlett, 1996). The human use of small (less than 100ha), interstitial fragments characteris tic of agricultural mosaic has not been highlighted. Studies of fragmented systems must include considera tion of ongoing human influences (Laurence and Bierregaard, 1997). Despite these acknowledgements, the impacts of resource base diminishment at the household level remains largely unaddressed in the literature. In Uganda, resources obtained from unpr otected forest fragments and wetlands are essential for survival of the majority rura l population (Banana and Go mbya-Ssembajjwe, 1998; Kayanja and Byarugaba, 2001). At the same time, the unprotected wetlands and forests outside parks and protected areas are problematic. Inte nsified use of forests and wetlands and landscape fragmentation has caused a substantial declin e in habitat and foraging lands for wildlife 63
(Onderdonk and Chapman, 2000; Chapman et al. 2001) Households are commonly affected by crop raids by primates, elephants, and birds; liv estock loss; and even hu man fatalities occurring in lands near these natural ar eas (Tchamba, 1996; Naughton-Treves, 1997; de Boer and Baquete, 1998; Gillingham and Lee, 1999). Little research has addressed either anthropogenic influences on land-cover change outside protected areas, or the resulting effects of altered ecosystems back onto the communities (Pickett et al 1995; Turner et al. 2001). Protected area establishment presents a challenge for households to sustain livelihoods because of exclusion of land and restriction of resources. Social systems and livelihoods of small rural communities are tightly linked to the environment that surrounds them. This paper examines the dual character of these natura l areas within the ag ricultural la ndscape around Kibale National Park in western Uganda as both benefits and hazards. Two questions are addressed, 1) What resources and problems are a ssociated with wetlands and forest fragments around KNP? and 2) How do these benefits and problems vary by distance from the park boundary, gender, wealth, and ethnicity? Study Region Ecological System Kibale National Park (KNP) (Figure 3-1) in western Uganda is illustrative of the agricultural expansion and intensification surrou nding many protected areas. Kibale Forest was demarcated in 1932 as a Forest Reserve and elevat ed to national park status in 1993, leading to the eviction of a few thousand settlers within the park (Struhsaker, 1997; Chapman and Lambert, 2000). Kibale National Park is a medium-a ltitude tropical moist forest covering 795 km2 in western Uganda 3 (Naughton-Treves, 1998). This transitiona l forest (between lowland rainforest 3 While officially KNP now includes a game corridor that connects the so uthern portion of KNP to Queen Elizabeth National Park, this corridor was formally gazetted as part of KNP in 1993. Before that time, it was mixed 64
and montane forest) is ranges in elevation of 11 10-1590m and is a remnant of a previously larger mid-altitude forest region (Struhsaker, 1997). Although the amount of rainfall and length of season change, the average annual rainfall for the region is 1543mm (average 1903-1999) and 1719mm (1990-2005) (Chapman et al. 2005). The landscape outside the park has become a patchwork of small farms (most <5ha in size), tea estates in some areas, and a collection of forest fragments a nd wetlands, effectively isolating the park. In addition, 43% of the land within a 5 km peripher y of the park is under cultivation or pasture (Hartter, unpublished data). Tea dominates much of the landscape bordering the northwest portion of KNP, covering nearly 21% within 1km of the park boundary (Hartter, unpublished data). Tea producers range from large multinational companies with hundreds of hectares, to smaller individual holdi ngs of less than one hectare (Mulley and Unruh, 2004). Forest fragments and wetlands that extend out from the parks boundaries and isolated within the agricultural matrix su rrounding the park vary in size, shape, and resource types and amount. Forest fragments range in size in Bakiga communities (east side of KNP) from 0.5ha to approximately 150ha and 0.5ha to approximately 210ha in Batoro communities (west side of KNP). Wetlands vary in size on both sides of the park with Dura Swamp (~220ha) and Magombe Swamp (~420ha) being th e largest on the west and east side respectively being the largest. Since nearly all of these natural areas occur in bottomland areas, many of the forests and wetlands are combined, effectively extending th e resource bases and wildlife corridors. agricultural land. Therefore, it was not classified as park in this analysis. Since 1993, the total park area is 795km2, but only 561km2 of parkland (excluding the game corridor) is covered in this analysis. 65
66 Forest fragments Forests and woodlands cover approximately 4.9 million hectares in Uganda and nearly 60% of these remain unprotected (NEMA, 2001). These unprotected forests provide a number of ecological services such as stabilization of local climate, erosion control, nutrient uptake, and carbon sequestration (Laurance and Bierregaard, 1997). Forest frag ments also serve as species corridors, habitat, and breeding and feedi ng grounds (Onderdonk and Chapman, 2000; Marsh, 2003). These areas are used for commercial a nd subsistence purposes. They provide fuelwood, indigenous medicines, timber, charcoal, poles, and non-timber products (Banana and GombyaSsembajjwe, 1998). Nearly three million hectares of unprotected forests remain vulnerable to overexploitation and agricultural encroachment. Uganda contin ues to lose 50,000 ha of its forest annually (NEMA, 2001). Kabarole District in western Uganda, where KNP lies, is a good example of forest degradation and loss associated with ra pid population growth and unsustainable harvesting levels, leaving only patches behi nd. NEMA (1997) estimated 21% of the forests in the district have been degraded. As agriculture continues to encroach upon unprotected forests mainly to meet subsistence needs, these forests become progressively degraded and/or isolated. Wetlands Wetlands include all areas where plants grow and animals live in association with permanent or temporary flooding (Kisamba-Mugerwa and Nuwagaba, 1993). Ugandas wetlands, dominated by papyrus ( Cyperus papyrus L.) comprise about 13% of the countrys total land surface (Muk iibi, 2001a) and fill a critical ecological role in flood abatement, groundwater recharge, and as natu res biological filter. Wetl ands have many environmental values including main tenance of water quality and volu me, pollutant filtration, flood abatement, and serve as habitat for a diverse ra nge of aquatic and terrest rial species (Mukiibi,
67 2001b; Bakema and Iyango, 2001; Crisman et al 2003). Despite be ing one of the most productive ecosystems in the world (Muthuri et al. 1989), peopl e in the past treated wetlands more like wastelands (Mukiibi 2001b) or as a repos itory of land for potential agricultural expansion. Wetlands are also importa nt to the livelihoods of surrounding communities. They also serve as resource bases for hous eholdand commercial-based re sources and activities such as water, thatching, handcraf ts, medicinal plants, a nd tourism. Despite their social and ecological benefits however, Ugandas wetla nds are under serious threat of conversion due to population growth, in-migration, a nd intensification of agriculture in the surrounding la ndscape (Chapman et al. 2001; MacLean et al. 2003). Wetlands are often drained and conv erted to other land uses such as growing crops an d fuelwood, particularly Eucalyptus spp ., and expanding grazing land (Crisman et al. 2003). As with forest fragmen ts, wetlands represent un claimed or underutilized land in many cases. As a result, both areas continue to be degraded and converted, while rural communities are challenged to sustain their livelihood s on a day-to-day basis. Social System Rapid population growth, high population densit y, and heavy reliance on agriculture for income characterize the landscape surrounding KNP (Archabald and Naughton-Treves, 2001) and land pressure has steadily increased (Naught on-Treves, 1996). Economic inequality, ethnic disparity, increasing land use in tensity, and a decreasing resource base in these communities serve to illustrate a common trend around protec ted areas in developing c ountries. In Kabarole District, 95% of the population su stains their livelihoods through agricultural-b ased activities. With limited land availability, farm sizes are characteristically. Agriculturalists in the area bel ong to two dominant ethnic gro ups: the Batoro (west side of KNP) and the Bakiga (east side of KNP) (Figure 31). The Batoro are the largest ethnic group in
68 the area (~52% of population), but immigration of Bakiga who came to the Kibale region from southwestern Uganda beginning the 1950s a nd 1960s (Turyahikayo-Rugyema, 1974; NaughtonTreves, 1998) and others have greatly increa sed population growth and demand for agricultural land and resources (MFEP, 1992; NEMA, 1999). Th is region is one of the most densely populated areas on the African continent (Le pp and Holland, 2006), and population around KNP more than tripled between 1959 and 1990 (Naughton-Treves, 1998). The current population (2006) is estimated at 262 individuals/km2 on the west side of the park and 335 individuals/km2 on the east side within 5km of the pa rk boundary (Hartter, unpublished data). Methods I designated a study unit an area both large enough and small enough to include the scales of land cover, biodiversity, and hum an activities and enable us to measure them so that satellite remote sensing images and the on-the-ground meas urements are tied operationally to the same area of land (Southworth et al. in review). To provide this link, sampling units (named superpixels) comprising 9ha areas were used. These superpixels are centered on spatially randomly selected coordinates within 5km of the KNP boundary and within two areas (Figure 31). In all, 95 superpixels were selected. Th ese 95 superpixels are found within 60 villages (a cluster of houses typically within 250m and part of the same administrative unit) and include sample pool of 417 households (that is, households controlling land in one form or another contained within at least 1 of the superp ixels). Between May and August 2006, 130 semistructured household interviews were conducted. At each superpixel, respondents were chosen opportunistically, based on populat ion weighting. The number of respondents for each study was proportional to the number of landholders containing land within the study area. Therefore, study areas with smaller landholders, and thus more landholders, had a higher sampling intensity. During the course of fieldwork, additional inte rviews with village chairmen (LC1), key
69 informants, KNP wardens, tea estate mangers, and representatives from local conservation groups were conducted. All interviews were conducte d in English or in Rutoro or Rukiga using an interpreter. At each respondents house, GPS locations were obtained. GPS points were also taken at access points for the nearest forest fragment a nd wetland for each house. These GPS points were used to calculate the straightline distance from the house to th e nearest wetland and/or forest fragment. Distance from each house to the park boundary was calculated using the KNP boundary polygon. Thus, I mapped the social and eco logical units of interest together. Wealth and distance categories were determined using hierar chical cluster analysis. Wealth indicators used were number of animals (goats and cows) per household, head of household gender, total amount of land hel d, house category (based on a 5-category classification of house construc tion materials). The sample population was broken into three wealth classes more poor, less poor, and not poor. Three distance categories were defined also using hierarchical cluster analysis to relate the straight-line distance of the respondents house to the KNP boundary clo ser (<1000m), moderate (1001-3000m), and further (3001-5000m). I used 2 test for independence to exam ine the relationship between the response and distance to the par k, wealth, gender, and ethnicity. A significance value of 0.1 was used to reduce type I errors and gi ven the use of recalled information. Results Forest Fragment and Wetland Benefits to Households Nearly all of the respondents derived some sort of bene fit from wetlands outside KNP (Table 3-1). While the majority derived benef it in some way from forests outside KNP (78%), the proportion of respondents perceiving benefits was less than for wetlands (92%). Only a small percentage reported they had never derive d any benefit from forests and wetlands. In
70 addition to reporting benefits, the vast majority al so reported that the forests and wetlands were sources of problems to their household (Table 3-1). I found that perceived benefits and problems fr om wetlands and forests and distance to the park were not independent ( 2 test for independence, p>0.1 in all cases). This result suggests households find wetlands and fore sts beneficial and problema tic throughout the surrounding landscape. However, the specific benefit and prob lems and their relations hip to distance to the park had different results. Firewood, water, handcraft materials, indigenous medicines, and ties (vines used to tie building poles together for house and fence construction) were most often cited as benefits from the wetlands (Table 3-2). Wetland res ources: firewood (p=0.070), poles (p=0.042), and medicines (p=0.011) showed trends of increasing use at distances further from the park boundary were found not to be independent of distance from the park. More men reported firewood as a benefit from the wetlands than female res pondents (p=0.078). Ethnicity and wealth class appeared to play less of a role in responses to wetland benefits (p>0.1). Firewood, medicines, building poles, handcraft materials, and water were the most reported benefits of forests (Table 3-3). Ethnic ity and wealth were not significantly related to these resources and gender was only related to building poles. Distance was found to be negatively related to firewood, pol es, medicines, and handcraft materials (p<0.1). These data suggest that more households deri ve benefits from these resources at further distances from the park. Results from Tables 3-2 and 3-3 suggest that crafts, water, and ties can be obtained throughout the surrounding landscape. However, medicines and firewood from wetlands were more often gathered at distances further from the park.
71 In addition, ecosystem services were mentioned as important benefits of wetlands and forests. There is evidence to suggest that responses and ethnicity (p=0.012), gender (p=0.001), and wealth (p=0.029) class were not independent for wetland ecosyst em services. Bakiga, male respondents, and not poor respond ents perceive more often perceived these benefits (Table 32). Surprisingly, only 9% of res pondents reported ecosystem servi ces as benefits from forests, substantially less than the 32% re ported from wetlands (Tables 32 and 3-3). For forests, only ethnicity was found to be related to perceive d ecosystem services from forests (p=0.003). Most respondents fill most of their household water needs fr om an improved source (bore hole, improved well; 51%), while only 24% collect most of their water from the natural areas (Table 3-4). In addition, most households meet their energy needs from the natural areas (38%), while planted firewood supplies en ergy needs for 30% or respondents. Only 10% of respondents fulfilled most of their energy needs by buying fi rewood and charcoal. Local charcoal burners report the shortage of hardwoods in the forest fragments, the fear of punitive measures by the KNP rangers, and the relative abundance of Eucalyptus spp From anecdotal evidence, most of the firewood, charcoal, and trees purchased are not endemic hardwood species, but instead planted species. Problems to Households Wetlands and forests are also sources of hazar ds (Table 3-5). Most respondents (90%, n=130) said they had problems because of the nearby forests and wetlands. Most of the respondents identified wild animals that raid th eir crops and chickens and goats as problems directly related to forest pr esence (including KNP) and wetlands (74%, n=130). However, these responses were not significantly related to et hnicity, gender, wealth, or distance to park, suggesting wild animals do not discriminate betw een demographics or distance from the park boundary. Similarly, the problem of less availa ble resources was unrelated to distance,
72 suggesting that resource scarcity is a problem throughout the surrounding landscape. Maize was the crop that was reported to be raided most, followed by bananas, and then hens. Respondents also identified other problems. Respondents also noted the problem with less available resources (e.g. firewood, handcraft materials, medicines) in the forest fr agments and wetlands that they could gather. They also reported the problem of owner and/ or government stopping them from gathering resources. The 2 test for independence suggests that ow ner and/or government intervention is identified as a problem closer to KNP than furt her away. Responses also varied by demographic category. Ethnicity was found to be significantly relate d to nuisances, w ith Bakiga reporting this problem nearly two times more than Bator o. Nuisances were also significantly related to gender, and wealth class (p<0.1). Males and t hose closer to KNP tended to report the reduction and/or unavailability of resources more than women (p<0.1). Interestingly, mosquitoes, tsetse flies, ticks, and biting ants were also mentioned as key problems because of the presence of wetlands and forests. Although re spondents identifie d this as a problem due to the presence of a forest and/or wetland, it is prevalent in the land scape both in areas closer to and further from natural areas. Nearly 10 % of respondents said that they ha d no problems from the natural areas. Wild animals were reported most often as pr oblems for landholders. Of those that reported animals as a problem, the redtail (Cercopithecus ascanius ) (65%, n=96) and vervet monkeys ( Cercopithecus aethiops ) (51%, n=96) were cited most as pe rpetrators on farms around the park (Table 3-6). Although only a small percentage identifies birds as a crop raiding animal, this group is probably under-reported because there is no distinction in Ruto ro and Rukiga with animal and bird.
73 More Batoro farmers reported problems with elephants ( Loxodonta africana) (p=0.096), than Bakiga farmers. However, a higher propor tion of Bakiga farmers reported problems with redtail monkeys (p=0.068), vervet monke ys (p=0.106), with Lhoesti monkeys ( Cercopithecus lhoesti ) (p=0.001), mongoose (p=0.003), serval cats ( Felis serval ) (p=0.003), and foxes (p=0.003) than Batoro farmers. Male res pondents reported problems with baboons (p=0.049), black and white colobus ( Colobus guereza) (p=0.033), vervet monke ys (p=0.085), red colobus ( Procolobus rufomitratus ) (p=0.096), and birds (p=0.068) than female respondents. Wealth was not related to problem animals (p>0.1). Distance and ethnicity were significantly related to certain crop ra iders. At closer distances to the park boundary, baboons ( Papio anubis) (p=0.000), elephant s (p=0.000), Lhoesti monkey (p=0.016), red colobus (p=0.062), bush pig (Potomochoerus porcus) (p=0.011) were reported. At further distances, a great number of households reported vervet monkeys (p=0.011), mongooses (p=0.024), and serval cats (p=0.068). Of those that reported problems with wildlif e (n=96), 51% of res pondents say that small monkeys (vervet, redtail, black and white colo bus, red colobus, and Lhoesti monkeys) were the greatest problem overall (Table 3-7). Small were greater problems in households located further from the park (p=0.000). Conversely, households cl oser to the park boundary reported elephants and baboons as main problems more than those located further from the park boundary (p=0.000 and p=0.007 respectively). Although small monkeys were reported as th e main crop raiding problem at further distances from the park boundary, farmers at a ll distances reported pr oblems with them. However baboons and elephants were only reported at households less than 3km and less than 2km from the park boundary resp ectively (data from GPS coor dinates of actual incidences;
74 Figure 3-2). Small monkeys and baboons travel ed well beyond the boundaries of the natural areas to raid farmers fields, while elephants te nded to stay closer (<300m) to the natural areas (Figure 3-3). In addition to the problems of wild animals, many house holds report a reduction in the available resources for them to collect from th e forests and wetlands (33% n=130) (Table 3-7). Although not everyone has noticed a reduction, the majority of re spondents believe that there will not be enough wetland (56%, n=130) and forest (71%, n=130) to meet their needs in the future (Table 3-8). Despite the problems, I found that most people would rather live clos er than further from forests and wetlands (Table 3-9). Female respondents overwhelmingly would prefer to live closer to the natural areas, while there was li ttle difference within wealth and ethnicity and distance to the park. Those liv ing further from the park also had the largest proportion of respondents that preferred to li ve closer to the natural areas. Discussion Benefits from Wetlands and Forests Social and ecological systems are tightly li nked in the landscape surrounding KNP. Most people derive benefit in some way from the forests and wetlands both consumptive and nonconsumptive uses. Most report the consumptiv e uses such as firewood, water, handcraft materials, or building poles. However, it is also evident that perceived non-consumptive benefits from forests and wetlands (i.e., ecosystem services ) are important to landhol ders. It is also important to examine benefits a nd their relationships to each category: wea lth, gender, ethnicity, and distance from park. For example, water us e was not significantly re lated to distance from the park. The majority of the respondents have access to bore holes or local or improved wells for their main water sources (71%, n=130; Hartter, unpublished data). Although the sample
75 population was dependent on wood produc ts to meet most, if not a ll, energy needs, wetlands are depended on more than forests. These results suggest that fire wood is more important from the wetlands further from the park. One possible explanation is that the forests and wetlands near the park are so degraded that firewood is less available and must turn to either forest fragments and wetlands further away or go illegally to the park. In addition, people closer to the park have a more limited selection of firewood sources because collection in most cases is prohibited inside park boundaries. Migrants also seeking land found that land nearest the park was available, further diminishing wetlands and forests as they were cleared for fields and pasture. Many people in the area have also begun growing fuelwood, perhaps relieving some of the pressure on the forest fragments. I found that while only 37% (n=130) report the use of forest fr agments and wetlands as their main source for firewood, 71% and 48% report they use the forest fragments and wetlands respectively for at least some of their firewood. People also recognized the environmental benefi ts of forests and wetlands. Benefits such as rain and fresh air were cited most often and believed a direct resu lt of the presence of wetlands and forests in the area (and has been ve rified by other independ ent researchers in the area). Other services such as good and/or maintained soil fertility, soil moisture, and wind breaks were noted. In their mi nds, these areas not only brought ra in, but were responsible to provide enough rain and in a timely fashion for planting, growing, and harvesting crops. Respondents also mentioned that forests and we tlands keep the environment, meaning these areas were important not only for the ecosystem services such as rain and soil moisture, but also climate and temperature regulation. Many respondents also reported these areas keep animals. From their point of view, because the park and other natural areas are around, wild animals do
76 not come to their fields at all or come less. They believe that because these remaining natural areas provide a haven for wildlife food and ha bitat. If these areas were gone, they would surely come to their lands to raid their fields. By the same token, the benefits of rain, fresh air, soil moisture, etc. would not be available. Despite being most often cite d, these non-material benefi ts are based on perception, rather than entirely on fact. Their perceptions, wh ile they may be true ,partly true, or untrue in some cases, are based on conjecture of their own beliefs and schooling, second-hand information, misinformed environmental educators or park outreach repr esentatives, and their own experience. Though both provide many benefits, wetlands and forests are perceived differently. More people derive benefit from the wetlands than th e forests. Three times more respondents cited ecosystem services as an important benefit from wetlands than forests. Ethnicity, gender, and wealth class were significantly related to reported wetland environmental benefits, but oddly only ethnicity was significantly related to reporte d forest ecosystem services. Males and those that were not poor reported more environmental be nefits from the wetlands Males typically stay in school longer, have more opportunities for soci alization and work off the farm and therefore are more exposed to extension services or talk to other farmers. The people who are not poor are also those who tend to be more educated and have more access to printed information or extension agents. In addition, the Bakiga, who are generally poor er and further from population center, and generally have less infrastructure, co uld be less aware of environmental benefits. However, many respondents cited education pr ograms by the community conservation program, KAFRED (Kibale Association for Rural and Econom ic Development) as major contributors to their understanding of the envir onment (Lepp and Holland, 2006).
77 Environmental Impacts from Social Benefits Despite the benefit of environmental serv ices from wetlands a nd forests, population increase and land shortage has led to intens e pressure on the remaining resource bases. Perceived, intangible benefits are difficult to assess and are valued differently by different households. Households tend to forego what w ill not directly and imme diately benefit their short-term gains, even if it does decrease the long-term standard of living or sustainability of resources and livelihoods (Hyden, 1998). The environmental implications of conversion and resource extractions are similar in the remaining forest fragments surrounding KNP. Turner (1996) reported fragmentation causes a local loss of species; and isolated forests have fewer tree species. Fragmentation and loss of trees that serve as key habitats or foods cont ributes to a loss of other taxa, a change in microclimate, and may have long term effects on pollination, predation, and other ecological processes (Ranta et al. 1998). Studies from within KNP show forest recovery on degraded lands is a slow process. This recovery requires soil nutrient s (which are removed with reso urces), the absence of human disturbances and fertile soil and viable seed bank (Duncan and Chapman, 2002). In addition, large canopy gaps that are created from harvesting larger trees are related to limited survival of seedlings and higher mortality (Chapman et al. 1999, Kasenene and Murphy, 1991). In most cases, harvesting in forest fragments is done sele ctively, one tree at a time and processed one tree at a time (e.g. pit-sawing) reducing gap sizes and neighboring tree and seedling mortality (Chapman and Chapman, 1996). However, Chapman and Chapman (1999) and Paul et al. (2004) suggest even though conditions may be favorable to support high le vels of regeneration, low-seedling recruitment and survival ar e threatened by high disturbance and the Acanthus pubescens may dominate because of canopy collapse th at can suppress seedling establishment.
78 Forest fragments, through a long period of hi gh-grading, now are left with few remaining timber-sized trees and lower quality understory species The proliferation of Acanthus pubescens and fast-growing pioneer species in canopy ope nings and forest edge can be used for firewood, but it suppresses the important longe r-growing, hardwood species that are used for building poles. Landholders seek hardwoods because they burn hotter and can withstand termites and rot longer. High qua lity hardwood species used for saw-logs are now more difficult to find in forest fragments, and those that remain are often protect ed by either the National Forest Authority or adjacent landholders. However, pol es-sized trees (typically <20cm diameter at breast height) are deemed acceptable for harvest by many landholders. Many respondents reported a decrease in timber-sized trees, but conti nue to extract poles at a constant or increased rate. This becomes problematic b ecause these trees are typically hardwoods that not only have fought off suppression and established themselves but also grow slow er and are the most productive. Continued extraction of these smaller trees decreases the periodicity of disturbance and nutrient removal, but reduces the juvenile clas s that would eventually grow to larger trees. As a result of continued resource extraction and no effort to replan t trees in species or in number, there are fewer seed trees and reduced habita t important for seed-dispersing frugivores and aviaries. The long-term survival of the forest resource base is in jeopardy. There is also anecdotal evidence to suggest that forest fragme nts are being degraded to a point that the big trees in the fragments are dying (H artter, pers. obs.), consistent with what is reported in the Amazon (Laurance et al. 2000). Problems Several problems related to wetland and fore st fragment (non)presence were identified Nearly one third of the people said there was a noticeable change in available resources that were gathered in the forests and wetlands, and that th eir responses were indepe ndent of distance from
79 the park. As the landscape around the park beco mes increasingly fragmented and the forests and wetlands become smaller and more isolated, the avai lability of high quality resources decreases. As resources become scarcer, landholders begin to intervene and closely monitor and control the use of their forest in order to preserve their ab ility to access resources in to the future. Others regard this control of access a nd stipulations of the amount and type of resources that can be gathered within the forests and wetlands as a problem. With the creation of the National Environment Management Authority in 1995 as the principal agency for the management of the environment (NEMA, 2001), the national govern ment and local councils have been more proactive in wetlands protecti on. Landholders reported that w ithin the last two years the National Environment Management Authority, the District Environmenta l Officer, local council leaders, and other landholders have limited extr action. As a result, local landholders cited government intervention and restri ction of access as one of the main problems they have had with obtaining enough firewood and other resources. Most of the respondents reported that animal s were a problem and these problems were due to the presence of wetlands and forests. I found that distance to the park boundary, ethnicity, wealth, and respondent gender were related to problems and problem animals. However, distance appears to be the most important indi cator for resource consumption and problems. Despite being generally regarded as more inte nsive land users, the Bakiga respondents cited a mean usage of forest and wetlands nearly the same as the Batoro in number of uses (4.5, 5.0 respectively). However, a higher percentage of Bakiga respondents reported problems with animals. Only with respect to elephants di d Batoro respondents repo rt more problems the Bakiga. Despite the differences, I believe that these results are more indicative of location effects than the influence of ethnicity on land use. For example, Lhoesti monkeys and serval cats
80 generally occur on the east side of the park, which is made up predominantly of Bakiga landholders. Anecdotal evidence fr om scientists having work ther e for nearly two decades and local farmers, report only rare sightings, if any of Lhoesti monkeys on the west side of the park (C. Chapman, pers. comm.). Ethnicity is impor tant because ethnic group and the associated culture and land tenure patterns can determine wh ere households will be located (Hill, 1997). Naughton-Treves (1997; 1998) focused on park neighbors and found that proximity to the KNP boundary was the strongest pred ictor of crop raiding. In farms studied close to the park boundary ( 0.5km), baboons were identified as the worst crop raider (Naughton-Treves, 1996; 1997) and small monkeys (cercopithecine) were most prevalent (Naughton-Treves, 1998). Her work showed that the frequency of damage even ts to crops decreased further from the park boundary (maximum 500m). A decade later, the presence of animals as a problem was not significantly related to distance from the park boundary, al so indicating that crop raiding is indiscriminant and happens all over in the surrounding landscape. Instead of targeting specific farms adjacent to the park boundary, I included a larger, geographic random sa mpling of households at multiple locales. I found that distance was related to which animals raided farms (i.e., baboons, elephants, small monkeys). Small monkeys were found to be the mo st prevalent raiders and overall were reported as worse problems than the baboon or elep hant in the surrounding landscape (i.e., 5km from the park boundary). Farmers report these small monke ys are in the fields almost daily, especially during the two harvests in Janua ry and midto late-July thro ugh August, while elephants are infrequent raiders that may come once, twice, or not at all in a season. They typically arrive at night and they can appear unpred ictably throughout the year, but typically from January to May along the western and northern borders and then shifting to the eastern and southern borders
81 from June to December (Chiyo et al. 2005). When they do visit fields, the result is often catastrophic losses to an entire seasons field or crop. They i ndiscriminately pass through fields to feed on bananas, maize, and sweet potatoes, and other crops (Naughton-Treves, 1998; Chiyo et al. 2005; Hartter, pers. obs.). As other studies have shown distance from the park boun dary and distance from the natural areas is also strongly related to farm vulnerability to crop raiding (Hill, 1997, Naughton-Treves, 1998; Chiyo 2000). Landowners report that ba boons and elephants only use the fragments as corridors, but reside inside th e park, while vervets, redtails, and other small primates reside within these same fragments. These results concur with these reports. Vervets are more problematic in farms further from the park than closer. Further from the park, the remaining forests a nd wetlands tend to be islandized (i.e., not connected to other fragments and the park) within the surrounding agricu ltural matrix. These fragments are far away from the continuous habita t of the Kibale forest. Verv ets can live within disturbed habitat (Onderdonk and Chapman, 2000 ), while other primates resi de in the forest, but have larger home ranges. I found that baboons raid ed farms up to 3km from the park boundary, but forays were greater than 0.5km fr om the natural area co rridors. Elephants and baboons tend to raid farms closer to the park boundary, while smal l monkeys are present all over in the surrounding landscape ( 2 p<0.1). The forest fragments and wetlands are then used as corridors to move within the surrounding landscape. Chiyo (2000) and Naughton-Treves and Treves (2005) report ed that elephant damage is tightly confined to distances less than 200m from the park boundary during their studies in the 1990s. However, I found that elephants affect farms much furthe r from the park boundary, traveling up to 2km from the nearest edge.
82 Within that landscape, elephants tend to stay closer to the fo rest fragments, while baboons and small monkeys are much bolder and ve nture through the agricu ltural matrix to the particular fields. The redtail monkey, which wa s the second greatest problem animal, was not significantly related to distance from the park boundary. Distance was also found to relate significantly to reported problem s with red colobus monkey. Bo th of these results confirm Onderdonk and Chapmans (2000) research with fr agments outside KNP. Red colobus dietary preferences may prevent them from living far fr om the continuous forest, while redtail monkeys can reside in disturbed forests and supplement their diets by raiding neighboring farms. These results are also important because they reveal farmers percep tions of threats and origins of those threats. There may be some inhe rent bias to for respondents to say that most of the crop raiders come from the park because of ambivalence toward the park or because the respondent thought the research team had some affiliation with the park. However, these perceptions, especially on farms away from th e park, have implications on land use and land cover. As the population and other pr essures increase, wetlands and forests become increasingly utilized, more intrusion into their interior, incr easingly isolating them within the surrounding matrix of farms, and producing greater edge de nsity (Byron and Arnol d, 1999; Gascon et al. 2000). Many re spondents believed that crop raiding incidence was directly related to the presence of natural areas. As a result, in some cases forests are cut for timber and building poles; to make way for crops or pasture; or even to re duce crop-raider habitat. Many say that they cut forests as a pre-emptive strike to clear their habitat and feeding area s. They believe that if they reduce the forests and wetlands, there will be no wh ere to go and will be forced to retreat within the boundaries of KNP or other fragments. Others say the best defense against crop raiding is to
83 have their farm buffered by at least another fa rmer closer to the fo rest or wetland. This fragmentation of the landscap e not only has impacts on households and farmers seeking resources and land, but it also has detrimenta l impacts on the ability to provide ecosystem services (i.e., water volume maintenance a nd filtration, and carbon sequestration) and the biodiversity (Tinker, 1997). Not only has more edge been created, but also the interior of the forests have become more accessi ble to hunting and other forms of resource extraction. Habitats and food supplies are shrinking. Primates and other animals must travel further distances and ultimately across farm, pasture, and fallow lands to the next natural area (wetland and forest). Nearly one third of the respondents perceived noticeable shortages of resources from some time in the past, but most believe that there will not be enough resources from either forest or the wetlands in the future. Those who said there wi ll be enough resources in the future, live further from the national park. It is because of these perceptions about the problems, benefits, but also about the future availability of resources that the overwhelming majority would rather live closer than further from natural areas. Respondents say these problems are predictable and they can do something about that (e.g. guarding their fields ), but they cannot do anything about resource unavailability. Simply put, their priority is to survive (i.e., maxi mize short-term gains). Forests and wetlands provide resources to do that. Since resource use within KNP is restricted, the fulfillment of the neighboring populations need s place the remaining forest fragments and wetlands in jeopardy and almost certain extinction not only of land and resources, but also of biodiversity. While the boundaries of KNP have remained intact ov er time (Southworth et al. in review), evidence from this study showed that resource shortages are escalating. In the end, can Kibale withstand the social pr essures outside its boundaries? Once remnant forest and wetlands are destroyed or degraded beyond use, then pres sure on protected areas may increase (Chhetri et
84 al. 2003). Many farmers report the positive impact s of the park such as keeping animals and ecosystem services. Future research should a ddress a deeper understanding of attitudes toward the park and the peoples relationship with the en vironment and their land use practices and their impact on different taxa (Twy man, 2001; Brockington et al. 2006) Only then can amenable solutions to crop raiding and resource shor tages between the park and its neighbors be conceived. Conclusion Forests and wetlands outside KNP are both beneficial and problematic to households in the surrounding landscape. I found that while respondent gender, wealth class, and ethnicity do not relate strongly to extracted res ources from wetlands and forest fr agments since the vast majority of rural households depend on the natural area s in one way or another, regardless of demographic. However, the spatial allocation of those resources (i.e., distance from the park) was related. The reduction in the size and number of forest fragments has reduced the available resources from these areas for households, and al so decreased habitat and food for wildlife. Most households perceive res ource shortages and mo st report problems of crop raiding. Wild animals were identified as the main problem in the surrounding landscap e as a result of the presence of wetlands and forests. The vervet an d redtail monkeys are the worst crop raiders, but baboons and elephants are also extremely problema tic. Elephants and baboons tend to raid farms closer to the park boundary, while small m onkeys are present throughout the surrounding landscape. The superpixel methodology was useful to identify the sample population, and will provide a linkage from these househol d-level results to the landscapelevel analysis as part of the larger research effort. The superpixel is an a ppropriate sampling framewor k to collect data from multiple locales and was appropriate for densely settled agricultural landscapes that receive abundant rainfall. Ethnicity is important because it relates to settlement patterns of farms.
85 Though the forests and wetlands are perceived to be sources of both problems and benefits, most landholders say they are worried about the availa bility of resources in the future and would rather live closer to these areas. So while KNP remains intact, heavy re liance on the land and its resources outside the park place remaining forests and wetlands in jeopardy of severe degradation or extinction.
86 Figure 3-1. Kibale National Park and surrounding landscape in western Uganda
87 <500 m<1 km1-1.9 km2-2.9 km3-3.9 km4-5 km Small monkeys Baboon Elephant Figure 3-2. Distance of households to KNP boundary reporting problems with wild animals for leading crop raiders <500 m <1 km1-1.9 km2-2.9 km3-3.9 km4-5 km Small monkeys Baboon Elepha "Further" "Moderate" "Closer" nt Figure 3-3. Distance of households to natural areas reporting pr oblems with wild animals for leading crop raiders
88 Table 3-1. Benefits and problems from natural areas Wetlands useful Forest useful Perceived problems from forests/wetlands Category Yes No Yes No Yes No n All Respondents 92% 8% 78% 22% 90% 10% 130 Mtoro 94 % 6% 78% 22% 90% 10% 72 Mkiga 87 % 13% 78% 22% 93% 7% 46 Ethnicity Other 10 0% 0% 58% 42% 75% 25% 12 Male 92 % 8% 78% 22% 95% 5% 63 Respondent Gender Female 93 % 7% 75% 25% 85% 15% 67 More Poor 90% 10% 76% 24% 93% 7% 58 Less Poor 93% 7% 78% 22% 88% 12% 59 Wealth Class Not Poor 100% 0% 69% 31% 85% 15% 13 Closer 90% 10% 70% 30% 92% 8% 61 Moderate 91 % 9% 76% 24% 91% 9% 33 Distance to KNP Further 97 % 3% 86% 14% 86% 14% 36 Table 3-2. Top five benefits de rived from wetlands (n=130). Wetland Benefit Proportion of Respondents (n=130) Ethnicity Respondent Gender Wealth Class Distance to KNP Crafts 76 % 0.791 0.687 0.646 0.144 Water 65 % 0.917 0.943 0.112 0.167 Firewood 51 % 0.830 0.078* 0.458 0.070* Medicines 49 % 0.641 0.722 0.256 0.011* Ties 38% 0.366 0.927 0.117 0.333 Ecosystem services 32% 0.012* 0.001* 0.029* 0.729 *response and wetland benefit not independent (p<0.1)
89 Table 3-3. Top five benefits derived from forests (n=130). Forest Benefit Proportion of Respondents (n=130) Ethnicity Respondent Gender Wealth Class Distance to KNP Firewood 69 % 0.593 0.884 0.998 0.098* Medicines 49 % 0.145 0.722 0.936 0.034* Poles 46 % 0.796 0.037* 0.838 0.082* Crafts 45 % 0.122 0.753 0.535 0.018* Water 38% 0.929 0.781 0.589 0.348 Ecosystem services 9% 0.003* 0.673 0.782 0.821 *response and forest benefit not independent (p<0.1) Table 3-4. Respondents reported their primary sources for h ousehold water and energy use Most water from: (n = 130) Most energy from: (n = 130) Bore Hole 33.1% Natural forest* 30.0% Wetland 20.8% Planted firewood 30.0% Improved well 17.7% Resting/bush 20.8% Local well 17.7% Buying firewood 7.7% Stream/spring (well) 5. 4% Swamp/forest combined** 5.4% Forest 3.1% Swamp/wetland 2.3% Rain tank 0.8% Buying charcoal 2.3% Lake 0.8% Tea cuttings 1.5% Piped from town 0.8% *Natural forest refers to a forest fragment that is not purposefully planted **Some swamps and forests are combined and difficult to distinguish between the two land cover types ***This value may be under-reported
90 Table 3-5. Perceived problems due to presen ce of forest (including KNP) and wetlands Problem Proportion of Respondents (n=130) Ethnicity Respondent Gender Wealth Class Distance to KNP Animals 74% 0.429 0.323 0.588 0.702 Less available resources 33% 0.683 0.068* 0.774 0.164 Owner/government intervention 25% 0. 022* 0.543 0.717 0.008* Nuisances 63 % 0.000* 0.023* 0.052* 0.944 Mosquitoes 58 % 0.000* 0.138 0.047* 0.848 Flies 34 % 0.000* 0.766 0.178 0.834 Ticks 18 % 0.168 0.394 0.724 0.507 Ants 16 % 0.003* 0.178 0.476 0.062* None 10% n/a n/a n/a n/a Although respondents identified this as a problem due to the presence of a forest and/or wetland; it is prevalent in the landscape both in areas closer to and further from natural areas. *response and distance to KNP boundary not independent (p<0.1)
91 Table 3-6. Most commonly reported problem animals identified by local landholders Problem animals Proportion of Respondents (n=130) Ethnicity Respondent Gender Wealth Class Distance to KNP Olive Baboon ( Papio anubis ) 35.4% 0.611 0.049* 0.468 0.000* Black and White Colobus ( Colobus guereza ) 25.0% 0.171 0.033* 0.494 0.388 African Elephant ( Loxodonta africana ) 32.3% 0.096* 0.155 0.956 0.000* L'Hoesti Monkey ( Cercopithecus lhoesti ) 10.4% 0.001* 0.156 0.548 0.016* Redtail Monkey ( Cercopithecus ascanius) 64.6% 0.068* 0.852 0.917 0.169 Vervet Monkey ( Cercopithecus aethiops ) 51.0% 0.106* 0.085* 0.275 0.011* Red Colobus ( Piliocolobus tephrosceles ) 18.8% 0.828 0.096* 0.570 0.062* Birds 9.4% 0.976 0.068* 0.423 0.135 Bush Pig ( Potomochoerus porcus ) 8.3% 0.404 0.320 0.958 0.011* Mongoose ( unknown sp.) 38.5% 0.003* 0.542 0.391 0.024* Serval Cat ( Felis serval ) 24.0% 0.003* 0.692 0.170 0.068* Fox (unknown sp.) 5.2% 0.003* 0.361 0.705 0.006* Note: When asked about crop raiding animals, most landholders only knew one word for the general term monkey ( enkende in Rutoro and Rukiga). Respondents identified the specific species after being show n pictures of primates endemic to the Kibale region. *response and distance to KNP boundary not independent (p<0.1)
Table 3-7. Greatest problem animals reported by respondents and responses relative to distance from the park boundary Distance to KNP Problem Proportion of Respondents (n=96) n 2 sig. (2-sided) Closer Moderate Further Small monkeys 51.0% 49 0.000* 29.8% 50.0% 92.0% Vervet 28.1% 27 0.000* 8.5% 29.2% 64.0% Redtail 21.9% 21 0.709 21.3% 16.7% 28.0% Black and white colobus 0.0% 0 n/a 0.0% 0.0% 0.0% Red colobus 1.0% 1 n/a 0.0% 4.2% 0.0% L'hoesti 0.0% 0 n/a 0.0% 0.0% 0.0% Elephant 17.7% 17 0.000* 34.0% 4.2% 0.0% Baboon 16.7% 16 0.007* 27.7% 12.5% 0.0% Others 12.5% 12 n/a 6.4% 33.3% 4.0% *response and distance to KNP boundary not independent (p<0.1) Table 3-8. Perceptions of res pondents whether or not there will be enough wetlands and forest nearby to meet their needs in the future Proportion of Respondents (n = 130) Response Enough wetland Enough forest Yes 27.7% 18.5% No 56.2% 70.8% Don't know 13.1% 4.6% N/A 3.1% 6.2% 92
93 Table 3-9. Preferences for living clos er or further from natural areas. Category Sub-Category Closer Further n All Respondents All 66.9% 33.1% 130 Mtoro 66.7% 33.3% 72 Mkiga 65.2% 34.8% 46 Ethnicity Other 75.0% 25.0% 12 Male 57.1% 42.9% 63 Respondent Gender Female 76.1% 23.9% 67 More Poor 67.2% 32.8% 58 Less Poor 66.1% 33.9% 59 Wealth Class Not Poor 69.2% 30.8% 13 Closer 65.6% 34.4% 61 Moderate 63.6% 36.4% 33 Distance to KNP Further 72.2% 27.8% 36 Note: the results are based on the respondents own definition of closer and further
CHAPTER 4 THE LANDSCAPE SURROUNDING A PARK: RESPONSES BY HOUSEHOLDS TO RESOURCE AVAILABILITY IN RURAL COMMUNITIES Introduction Parks, protected areas, and reserves have become the cornerstones of most national strategies to conserve biologica l diversity in developed and deve loping countries (Howard et al. 2000; Wilkie et al. 2006). Arguabl y, exclusionary approaches to park management have become the most effective way to conserve habitats a nd protect biodiversity (T erborgh, 1999). However, parks cannot be viewed in isolation separate from the landscape where they are located (Robinson and Ginsberg, 2004) and often the effects of parks and reserves on local communities are not considered (Newmark et al. 1994). Since neighboring communities have restricted access to land and resources within many parks, they face two choices: use land and resources illegally or use unprotected lands outside the park. Deterred by fear of punitive measures against illegal activity (Struhsaker et al 2004), many households turn to the unprotected resources. In addition, intensive land use and high populati on growth rates around many parks compound the issue. As population increases, pressure mounts on remaining unprotected wetlands and forest fragments to provide resources, but are also converted to othe r land uses to make way for agriculture or pasture (Banana and Gombya-S sembajjwe, 1998). The dynamism of household resource needs impacts the landscape; changing the spatial extent of farmland, forest, wetland, and other land cover types. Landscapes around parks in East and Southern Africa have become mosaics of natural and human-influenced patches. Fragmentation of tropical forests has been called the single greatest threat to global biological diversity (Turner and Corlett, 1996). Hill and Curran (2003) assert that fragmentation negatively impacts specie s composition due to a reduction in forest area and an isolation of the remaining forest fragments. In general, species richness and the number 94
of rare species decline as fragment area decreases (Collinge, 1996; Laurance and Bierregard, 1997; Laurance et al. 1998), and mortality rates of trees in the Amazon are higher in smaller fragments (Laurance et al. 2002). As the wetland and forest fragments steadily decrease and edge effects penetrate further, the result is an impov erished interior (Gas con et al. 2000). As more forest is converted and not replanted, fragments become incr easingly isolated. Therefore, the presence, absence, and qualit y of resources along with the presence of wild animals can impact livelihoods of the adjacent communities. As a result of forest resource exploitation, the conversion of lands to agriculture, such resource bases have become diminished and those that remain continue to be degraded. Edge habit in creases and the ability of large animals to range widely without crossing agricultu re diminishes (Newmark et al. 1994; Naughton-Treves et al. 1998), thereby further increasi ng human-wildlife encounters (Naughton-Treves et al. 1998). Although the amount of damage varies, farmers must c onstantly make decisions and create coping strategies to manage losses due to animal raids (G oldman, 1996; Naughton-Treves, 1998). Land cover and land use and are inextricably linked. The adaptation to resource shortages and crop raiding is manifested through land cover change (i.e., more wetlands drained and converted to make way for agricultur al lands or pasture). At a broader scale, analysis of satellite imagery has been used to examine these changes. Remote sensing is increasingly incorporated into land-use land-cover research that studie s relationships between human societies and the biophysical environment to monitor changes over multiple dates and lo cales (Roughgarden et al. 1991; Turn er, 2003; Kerr and Os trovsky, 2003). While coarse, landscape level effects can be detected using satellite imagery, lower level an alyses applied at the village, or household level may be best be suited to capture the decision making process behind land cover change (Evans 95
and Moran, 2002). Despite improvements in la nd cover characterization through advancing satellite technology, land us es are poorly enumerated (Lambin et al. 2001). What is needed is an enhanced understanding of the causes of change. To do this, we must move beyond separate methodologies examining landscape change through remote sensing methods or social methods (e.g. census data, rapid appraisals, or household su rveys). Research must span temporal and spatial scales. Combining household and landscap e level data in a cohesive analysis can help explain both proximate and dist al effects of environmental change (G eoghegan et al. 1998; Rindfuss et al. 2003). Recent studies on conservation and local people s, have called for a better understanding of local knowledge and perceptions of land and na tural resources (Fairh ead and Leach, 1996; Hulme and Murphree, 2001). However, previo us research has focused on the effects of fragmentation on animal and tree species (Stouffer and Bierrega ard 1995; Chapman and Chapman, 1999; Chapman and Lambert, 2000), and ve ry few have related the resources that forest fragments provide to surrounding commun ities (Turner and Corlett, 1996) and the social implications of exclusionary appr oaches to park management (Sout hworth et al. in review). Since most of the resources collected in the wetlands and forest fragments are only obtained locally, households must respond in some way to shortages, in turn shaping the landscape. Many livelihoods, and thus responses to shortages and problems, are determined through cultures and traditions, and economic, ecological environment, and demographic variables (Chambers and Conwa y, 1991). Therefore variables su ch as ethnicity, wealth, and gender as well as distance from the park may all be related to these responses (Stone, 1996; Hill, 1997; Byron and Arnold, 1999; Goebel et al 2000; Kagoro-Rugunda, 2004). This paper examines the ways households have adapted to a degraded resource ba ses of the unprotected 96
wetlands and forests and problems with wildlife in the landscape surrounding a park by addressing two critical questions : 1) How have the presence and extent of wetlands and forest fragments in the landscape surrounding Kibale National Park changed over time? And 2) How have households adapted to the declines in th e wetlands and forest fragments in the landscape surrounding Kibale National Park? Understandi ng coping mechanisms of households about resource use and wildlife conflict for communities living near protected areas can greatly add to knowledge about and ability to address issues of conservation of forested areas. Study Region Physical Environment Rural livelihoods in Sub-Saharan Africa are heavily reliant on the land and its resources (Scoones, 1998; Abalu and Hassan, 1998). In Uga nda nearly 80% of the 27 million people have rural-based livelihoods and more than 80% of la nd is used for small-scale farming (Mukiibi, 2001). The population has increased 240% betw een 1960 and 2000, with a continued national population increase estimated to be 3% annually (US Census Bureau, 2002). Kibale National Park (KNP) (Figure 4-1) in western Uganda, one of Ugandas 10 national parks, is illustrative of the agricu ltural expansion and inte nsification surrounding forested parks in Africa. Kibale is a medi um-altitude tropical moist forest covering about 795 km2 in western Uganda (Naughton-Treves, 1998). This transitional forest (between lowland rainforest and montane forest) is at an averag e elevation of 1110-1590m and is a remnant of a previously larger mid-altitude forest region (Struhsaker, 1997). Average annual rainfall for the region has ranged from 1543mm (average 1903-1999) to 1719mm (1990-2005) (Chapman et al. 2005). The landscape outside the park has become a patchwork of small farms (most <5ha in size), tea estates in some areas and a vast network of bottomla nd forest fragments and wetlands 97
98 that serve as important resource bases for water and fuelwood, effectively isolating the park. These dendritic networks are typically found in valley bottoms and vary in size, shape, and resource availability. Forest fragments range in size in the communities on the east side of KNP from 0.5ha to approximately 150ha and from 0.5 ha to approximately 210ha in communities on the west side of KNP (Hartter, unpublished data). Wetlands vary in size on both sides of the park with Dura Swamp (~220 ha) and Magombe Sw amp (~410 ha) being the largest on the west and east side respectively being the largest (Hartter unpublished data). Since nearly all of these natural areas occur in bottomland areas, many of the forests and wetlands are combined, effectively extending the resource bases and wildlif e corridors. These areas generally represent wildlife habitats and corridors on the one hand as well as areas of resource extraction on the other. Land scarcity has forced many farmers to farm to the edges of forest s (Naughton-Treves, 1998). In addition, 43% of the land within a 5km periphery of the park is under culti vation or pasture (Hartter, unpublis hed data). Tea covers much of the landscape bordering the northwest portion of KNP. Tea producers range from la rge multinational compan ies with hundreds of hectares, to smaller individual holdings of less than one hectare (Mulley and Unruh 2004). Social Environment Rapid population growth, high population densit y, and heavy reliance on agriculture for income characterize the landscape surrounding KNP (Archabald and Naughton-Treves, 2001) and land pressure has steadily increased (NEMA, 2001). Agriculturalists in the area belong to two dominant ethnic groups: the Batoro (west side of KNP) and th e Bakiga (east side of KNP). The Batoro, the largest ethnic group in the area (~52% of popul ation) (Naughton-Treves, 1998), the Bakiga, and other ethnic gr oups have greatly increased p opulation growth and demand for agricultural land and resources (N EMA, 1999). This region is one of the most densely populated
areas on the African continent (Lepp and Holland 2006). Naughton-Treves (1998) reported that population around KNP more than tripled betw een 1959 and 1990. We estimate the current population estimated to be 262 individuals/km2 on the west side of the park and 335 individuals/km2 on the east side within 5km of the park boundary. Population growth rates in surrounding parishes range between 3 and 4% per year (NEMA, 1999). In th e Kabarole District, where KNP lies, 95% of the population sustains their livelihoods th rough agricultural-based activities. Methods To examine and then link landscapeand hous ehold-level data we designated a study unit area both large enough and small enough to includ e measure land cover, biodiversity, and human activities. Through this linkage satellite remote sensing images and the on-the-ground measurements are tied operationally to the same area of land (Southworth et al. in review). These sampling units (named superpixels) co mprising 9ha areas were centered on spatially randomly selected coordinates within 5km of the KNP boundary and within two areas (Figure 41) representing the two ethnic groups on the east and west side s of the park. In all, 95 superpixels were selected randomly. These 95 superpixels are found within 60 villages and include sample pool of 417 households (that is, h ouseholds that have la nd contained within at least one of the superpixels). Research was conducted through a multiple-scale approach, combining remote sensing techniques and household surveys. Remote Sensing The first step was quantifying the change in extent, pattern, and qua lity of wetlands and forest fragments outside KNP over time using sate llite image analysis. Landsat TM imagery was chosen because it offers the best combination of spatial, spectral, te mporal and radiometric resolutions. Three dry-season images have been acquired: May 26, 1984, January 17, 1995, and 99
January 31, 2003 (path 132, row 060). The first im age provides baseline data prior to park establishment, the second captures conditions at park establishment, and the third represents current conditions. During the 2004 and 2005 field seasons 180 training samples were collected and used to construct a supervis ed classification. The classified image was constructed using the original bands (minus thermal), texture of th e original bands, and a Normalized Difference Vegetation Index (NDVI) layer. The final cla ssification obtained an overall accuracy of 89.1% and an overall kappa statistic of 0.867 (Southworth et al. in review). Five classes were used in the classification: water; bare soil, field crops pasture grass; forest, wetland comprised of papyrus ( Cyperus papyrus L.) and elephant grass ( Pennisetum purpureum ); and tea. Landscape Fragmentation Landscape metrics were used to identify trends in landscape heterogeneity over time. Size, shape, edge, and relative connectivity are importa nt characteristics of forest fragments and wetlands to characterize the extent of landscap e fragmentation. Fragstats 3.3 was used to calculate these metrics. Only forests greater than 0.5ha were considered in the land cover classification to conform to IFRI forest size standards (Hayes 2006), to focus the analysis on natural forests and not small kitchen gardens and household woodlots. Household Interviews We examined the implications of that change on household livelihoods in communities surrounding KNP using household interviews Between May and August 2006, 130 semistructured household interviews were conducted. At each superpixel, respondents were chosen opportunistically, based on populat ion weighting. The number of respondents for each study was proportional to the number of landholders containing land within the study area. Therefore, study areas with smaller landholders, and thus more landholders, had a higher sampling intensity. All interviews were conducted in English, or Rutoro or Rukiga using a local interpreter. At each 100
respondents house, GPS locations were obtained. GPS points were also taken at access points for the nearest forest fragment and wetland for each house. These GPS points were used to calculate the straight-line distan ce from the house to the nearest we tland and/or forest fragment. Distance from each house to the park boundary wa s calculated using the KNP boundary polygon. Within the household-level analysis, data wa s broken into several groups to address key socioeconomic characteristics, re spondent ethnicity, wealth cla ss, and head of household gender and distance from park boundary. In the Bakiga and Batoro culture, the man is traditionally the head of the household. However, if a female respondent is widowed, divorced, or unmarried, then the head of household is a woman. Wealth and distance categories were determined using hierarchical cluster analysis. Wealth indicators used were number of animals (goats and cows), head of household gender, total amount of land held, and house category (based on a 5-category classification of house construction). The sa mple population was broken into three wealth classes more poor, less poor, and not poor. Three distance ca tegories were used to relate the straight-line distance of the respondent s house to the KNP boundary closer (<1000m), moderate (1001-3000m), and fur ther (3001-5000m). We used 2 test for independence to examine the relationship between response and distance to the park, ethnicity, wealth, and gender. We chose a significance value of 0.1 for household data an alysis instead of 0.05 used in the land cover analysis to reduce type I er rors given the use of recalled information. Results Remote Sensing Since 1984, there was a net loss in forest and wetland while there was an increase in land under cultivation and tea (other la nd cover) in the landscape su rrounding KNP in both the east and west study areas (Tables 4-1 and 4-2). However, the east and west study areas differ in their land cover proportions. The east study area ha s a higher proportion of wetland and lower 101
proportion of forested land than the west study area. On the east side of the park, there is no tea, and all other land consisted of cultivated land. Contrasting the two sides of the park, the west side has a lower total percentage (46%) of natural areas but al so a higher proportion of other land cover (54%), which includes cultivated land a nd tea, in 2003 than the east side (56%, 44% respectively). In the east side, nearly 26% of the land is c onverted from natural areas to cultivated land from 1984 to 1995 and 15% converted from 1995-2003 (Table 4-3). From 1984 to 1995 and 1995 to 2003, only 29% and 27% of the land remained as natural areas respectively. The west study area is less intensively farmed and is interspersed with large tea plantations. Comparatively less of the land remains under co ntinuous cultivation and more of the land remains under natural areas. From 1984 to 1995, 18% of the natural areas are converted to crops or tea. From 1995 to 2003, more land remains under cultivation and tea, while a further 25% of the natural areas were converted to crops and tea. Each year both the west and the east side continue to lose more forest and less forest remains as forest at each time step, while land under crops not only increases each time step, but also the amount that stays under cultivation increases. Wetland areas in both the east and the west continue to be converted, bu t interestingly only in the east side is there more land th at remains wetland from 1995 to 2003 than from 19841995. Landscape Fragmentation Forest and wetlands loss is evident (Table s 4-1 and 4-2), and as such the landscape surrounding the park has become increasingly frag mented. In the east and west study areas, the number of patches, edge dens ity, and distance to the nearest same-class patch increased, while the mean patch size has decrease d for natural areas from 1984 to 2003 (Tables 4-1 and 4-2). We found that connectivity decreased over time fo r the forest and wetland areas. Conversely, the 102
land under tea and crops (shown as other land c over) continues to consolidate over time (i.e., connectivity increases). The number of patches d ecreases while the mean patch size increases. Forest and wetland loss varied over time. Two measures of the change and fragment isolation within the agricultural matrix are mean patch size and the distance to the nearest patch (Table 4-4). Mean patch size in both the east and west study areas was not significantly different for forests (p>0.05) for the 1984-1995 time step. However, the mean patch size was significantly different for the 1995-2003 time step for the forest class for both study areas (p<0.05). Distance to the nearest patch was not significantly different in the first time step, but was for the second time step. In the east, the mean patch size and distance to the nearest patch in wetland areas were significantly different in the east (p<0.05) for both time steps. In the east study area, distance to the nearest patch and mean patch area were not significantly different between 1984 and 1995, but was between 1995 and 2003. Conversely, on ly distance to the nearest patch was significantly different (p<0.05) in the west study area. Household Interviews At the household level, most respondents also report changes in the wetlands (88%) and forest (66%) within the last 10 years (n=130) (Table 4-5). Some of the changes that are reported are lower water level in the wetlands, a decrease in overall size, fewer trees in the forests, and increased use from more people (Table 4-6). Not only is tree density lower in the forest fragments, but also nearly 72% (n=130) farmer s report there are tree sp ecies sought specifically for building poles, saw logs, firewood, drums, me dicines, or other uses that are presently unavailable in these areas. Forty-five percent and forty-two percent of respondents said it has become more difficult to obtain resources in the wetlands and forest fr agments respectively (n=130). Most people cited 103
fewer resources as the reason for difficulty ( over 80%), while others reported restricted access because of the government or other landholders, a nd more people competing for resources (Table 4-7). Farmers have recognized the change in fo rest and wetland extent and the wetland and report that resources have become more diffi cult to obtain. These changes in resource availability worry local farmers (Table 4-8). The availability of firewood is an important concern for all classes (85%, n=130). However, mo st believe that there will not be enough forest (71%) and wetland (56%) to meet their future needs. Despite the changes in resource availability, 43% of landholders sa id there are fewer problems with crop raiding compared to the past. Though, the availability of firewood is a c oncern for most respondents, there is no significant relationship between response and ethnicity, wealth, head of household gender, or distance to the park (p>0.1) (T able 4-9). These results sugg est that firewood is a concern throughout the landscape, independent of demographic. Overall, we found that ethnicity and head of household gender were not related to re sponse, but wealth and distance were related (p<0.1) to some responses. Wealth class was significantly related to perception of forest resources for the future (p<0.1). Those who we re more poor (69%, n=58) tended to perceive future shortages than those who were not poor (46%, n=13). We also found that distance was related to perceptions of future forest resource s (p<0.1). More respondent s closer to the park boundary (77%, n=61) felt there would not be enough forest resources in the future than those further from the park (58%, n=36). Those clos er to the park boundary (36%, n=61) felt crop raiding has become worse than those further (17%, n=36). 104
The land cover analysis and results from the household survey indicate resources are becoming scarcer and that they are unsure about resources in the future. Since households depend on these resources, they adapt to shortages in various ways (Table 4-10). The five main responses were: planting trees for fuelwood, bu ilding poles, timber, and/or sustenance (70%, n=130); increasing the travel time to the forest or wetland (n=36%, n=130 ) and also increasing the amount of time it takes to gather all the fire wood, medicines, etc. needed once at the site (n=68%, n=130); going to a different forest or wetland for at least some of the households needs (n=34%); and to supplement the energy needs by purchasing some firewood or charcoal (n=42%). A higher proportion of Bakiga respondent s plant trees (78%, n=46) than Batoro respondents (65%, n=72) (p<0.1) (Table 4-11). Results suggest that a higher proportion of Batoro will go to a different res ource base to obtain firewood (42%, n=72) and buy at least some firewood and/or charcoal and/or trees for en ergy (56%) than Bakiga respondents (20% each) (p<0.1). We also found that of these responses, distance was only signifi cantly related to going to a different place to obtain firewood. More landholders living closer to the park boundary (41%, n=61) looked for a new res ource source than those further (22%, n=36). Overall, wealth and head of household gender were not significantly related to responses to resource shortages. Landholders are particularly susceptible to wild animals, since wild animals emanate from wetlands and forests to raid farmers fields and attack their livestock. Nearly 74% (n=130) of respondents said that they have pr oblems with wild animals (Table 412). Of those that perceive problems from these areas, 78% (n=96) reported th at they take action in some way to mitigate the problems of crop and livestock raiding. 105
In response to crop raiding, landholders reported five methods to mitigate problems with wildlife (Table 4-13). Of those that work to control or mitigate the problems of crop and animal raiding, the most common response is guarding fiel ds with family members, day laborers, or dogs (91%, n=75). A smaller number of households choose not to plant crops that seem to be especially palatable to wildlife (e.g. maize, banana s) in fields closest to the forest fragments and wetlands (49%). The particular crop varies base d on site-specific conditions (i.e., what species visit the farm, presence/absence of a forest or swamp, and other factors). Some people stop growing crops in a field, stop renting land, or m ove away from an area because the crop raiding has been such a hardship for them (29%, n=75) The least common res ponse is to continue farming, but stopping certa in crops (5%, n=75). Household responses to crop raiding varied by ethnicity (Table 4-14). A higher proportion of Bakiga guard (100%, n=33) than the Batoro respondents (75%, n=40) (p<0.1). However, a higher proportion of Batoro stop growing and/or living in an ar ea (40%) or do not plant certain field crops in areas closest to we tlands and forests (63%) because of crop raiding than the Bakiga respondents (18% and 33% respectively) (p<0.1). Responses to crop and animal raiding differe d by wealth class (Table 4-14). Results suggest that a higher proportion of those that are in the more poor class guarded their fields (94%, n=35) than those that are not poor (38%, n=8) (p<0.1). Ov erall, distance from the park boundary and head of household gender were not related to response (p >0.1), suggested that landholders throughout the surrounding landscape ha d to develop measures to mitigate crop raiding. Similarly, both male and female-headed hous eholds had to derive a single or mixture of methods. 106
Discussion Our results show the landscape surrounding KNP has become increasingly fragmented since 1984. Onderdonk and Chapman (2000) reported that the process of forest isolation has been ongoing since at least 1959. We found the numb er and size of forest and wetland patches continues to decline while land under cultivation ha s increased. In addition, as larger forests are broken into smaller fragments, access to these sm aller and more numerous fragments increases. In turn, more fragments are used. These new fr agments have become more isolated and edge density increases as they are used. In this proc ess, larger fragments become smaller and smaller as the continuous landscape is broken up. This process of increased fragmentation and opening up the interior of remaining fragments leads to degradation, more open/ disturbed forest and towards secondary vegetation (Schelhas a nd Greenberg, 1996). Increased landscape fragmentation has cascading effects. A decrease in size of forest can be detrimental to tree species diversity and the presence of rare sp ecies (Laurance et al. 2000; Hill and Curran, 2003), which could be key timber trees or medicinal plants The increase in amount and density of edge makes smaller fragments more vulnerable to e dge effects (Forman, 1995; Gascon et al. 2000). Gascon et al. (2000) report that fragments less than 5000 ha in size are extremely vulnerable, which is especially worrisome in the Kibale landscape since we found the largest forest fragment to be no more than 210 ha. The impacts of landscape fragmentation are no t limited to vegetation. The micro-climate can be altered, changing wind pa tterns, humidity, and solar radiation that are important for organisms (Ranta et al. 1998). The loss of habitat can also cause local dec lines or extinctions of species (Chapman et al. 2006). Th ese conditions lead to a negativ e feedback for households that depend on forest and wetland resources. The incr easing population, intensive land use practices lead to further encroachment into these areas as large converted tracts or as small annual 107
increments of expansion as new ribbons of cultivated land continues. Scherr (2000) asserts that often no concern is shown regarding resource degradation because they do not consider it a serious threat. However, our results show that people in the area do recognize changes, report increased hardship in obtaining resources, and ar e worried about shortages in the future. By converting the forests and wetlands to other land uses, households forego long term resource security for short term survival (Hyden, 1998). Ou r results show though that the majority of the people are responding especially to fu elwood shortages by planting trees. Though the Bakiga were found to respond more than Batoro respondents to shortages by planting trees, this analysis does not account for the amount or type of trees planted per household and the geographic characteristics. Generally, the Bakiga were poorer, held less land, and tended to farm intensively. Conversely, Ba toro respondents were wealthier and therefore had more land and the financial means to grow more trees. In addition, there is also a higher proportion of land as forest and wetland on the east than on the west, bu t nearly 16% of the land on the east side of the park is protected as the Bigodi wetlands (KAFRED). Since many forms of harvesting are prohibited within this we tland sanctuary, the Bakiga may be inclined to grow trees. Both the Batoro and Bakiga recognize the shorta ges and changes in the wetland and forest resources. However, the Bakiga are also thought to use forests more intensively than the Batoro (Chapman et al. 2003), bu t this may not be the most accu rate portrayal. Our analysis showed that more Batoro have had to go to a differ ent place to get fire wood and buy at least some firewood, charcoal, and/or trees. This would suggest that firewo od resources seem somewhat scarcer than on the east side. Future research should address the differences in forest and wetland consumptive practices from the Batoro and Bakiga. 108
We also found that distance to the park did not strongly corr elate to respons e to resource shortage, which suggests th at shortages are not limi ted one area or the othe r or closer or further from the park. In addi tion, these problems impact all wealth classes and households with male and female heads. Households must adapt their livelihood strategies to provide these resources, regardless of where they live or fi nancial status. In addition to the common resp onses discussed in this paper, households also respond to resource shortages in other ways. First, many households substitute one species for another. Since many of the endemic hardwoods have be en harvested from the forest fragments, households use other planted species such as Eucalyptus spp. The once-common ebigoya (vines that are used to tie building poles together) are substituted for dried skin from a banana plant. Second, some households are buying or renting more land, if they have the financial means that is closer or further from wetlands and forest s depending on household needs and concerns. The new land is planted differently or used to grow trees. Third, households also convert more land, especially to grow trees. Fourth, some suppl ement their resource needs by going illegally to KNP or other community forests. Illegal use pa tterns are difficult to track because respondents are fearful of punitive measures. More research is needed to address and quantify the harvesting practices. Many people perceive hardships in obtaining enough re sources and are worried about the future, but still many households cannot do an ything. They do not have enough land to grow trees, live far from wetlands and forests to gather resources there, or do not have the financial means to secure their resources for the future. Mo re research is needed understanding the factors leading to and adoption of the comp lex mix of coping mechanisms. We have shown that connectivity of forest fragments and wetland s has decreased since 1984, while the mean distance to the nearest fore st or wetland has increased. Predator-prey 109
relationships, habitat, an d migration corridors have been und oubtedly impacted (Laurance et al. 2002). Increased edge habitat and the degradatio n of habitat, breeding grounds, and nutrition is detrimental to primates in KNP (Onderdonk a nd Chapman, 2000). The reduction in connectivity between natural areas forces primates to adapt to reduced home ranges (Baranga, 2004) or they must travel within the agricu ltural matrix. Those farms nei ghboring forests and wetlands are vulnerable to attacks on their small-stock and crop raiding. The increased fragmentation of the landscape causes further detriment because the lim ited natural areas that act as corridors through which common crop and animal raiding species such as the vervet monkey ( Cercopithecus aethiops ), the redtail monkey ( Cercopithecus ascanius), olive baboon ( Papio anubis ), and elephants (Loxodonta africana) travel. Those fragments that remain channel the primates and especially elephants through agricultural lands because of the narrow passages connecting remaining wetlands and forest. Our results show a departure to research done a decade earlier. Hill (1997) and Naughton-Treves et al. (1998) assert that as the amount of forest edge continues to grow, humanwildlife encounters will increase. Such studies tended to focus on a single locale or geographic area (i.e., those villages closest to the park boundary) where crop raiding has historically been worst and tensions with the neighboring park ar e highest (Mkanda and M unthali, 1994; de Boer and Baquete, 1998). We found that less than one-thi rd respondents reported that crop raiding was worse than 10 years ago. More landholders reported that problems with crop raiding had actually become better (43%) or at least stay ed the same (10%). Similarly Kagoro-Rugunda (2004) showed that crop damage increases with decreasing population density outside Lake Mburo National Park in Uganda. Landscape frag mentation, intensive land use, and a growing population could have been a good thing in so me ways. Many say that they cut forests 110
because they believe that if they reduce the forests and wetlands, the wild animals will be forced to retreat within the boundaries of KNP or othe r fragments. The increa se in population in the area has also led to increased settlement. As Naughton-Treves (1997) and Hill (1997) found, many farmers still say the best defense against crop raiding is to have their farm buffered by at least another farmer closer to the forest or wetland. Ethnicity is important because it is a key factor because it determines household location. In addition, ethnicity is important in determini ng the neighbors that surro und the farm. A higher proportion of Bakiga respondents were found to guard their fields than Batoro. Wealth is also important in a households ability to cope with animal raiding and the poorest households are the most vulnerable (Naughton et al. 1999). Baki ga respondents were generally poorer and have smaller plots of land and were more intensive land users. Large landho ldings, grazing lands adjacent to the natural areas, and crops plante d far from these areas and near the house are characteristic of the not poor wealth class, who also happen to be Batoro. Households that are wealthier tend to have more land. Not only do they cultivate less land pe r total amount of land owned, leaving the lands cl osest to forest and wetland as pa sture, they can also plant their subsistence crops near the house where they can be closely monitored. Naughton-Treves et al. (1998) recommended th at highly palatable crops should not be planted close to the forest edge in order to ameliorate crop lo sses. Wealthy landholders have enough land so they can buffer thes e crops with those that are less palatable such as tea or they can grow trees. This recommendation is unrealist ic for two reasons. Firs t, most farm sizes are less than 5 ha in size (mean=3.8 ha) in the Kiba le landscape. Poorer landholders have less flexibility and ability to choose which land they would like and which land to farm. More land is allocated to crops and less to pa sture or a woodlot. We also found that given the choice most 111
respondents (67%, n=130) would prefer to live closer rather than further from the forests and wetlands (Hartter unpublished data). Second, some have planted unpalatable species such as tea, coffee, tobacco which serendipitously have only provided limited relief fr om crop raiding. Other field crops such as soybeans and sunflowers have been attempted, but with limited success (Naughton-Treves and Salafsky 2004). While Chi yo et al. (2005) report elephants do not consume Irish potatoes and groundnuts, these crops have been raided by other species (Hartter unpublished data). Though tobacco is not raided, but it proves ineffective in deterring animals because cultivated tobacco is not densely plante d, only planted in small amounts, and is easy for monkeys and baboons to travel through (Hartter, pers. obs.) Many people neighboring tea plantations and even those living within the tea plantation compounds also complain of elephants. It is uncommon for the elephants to tr avel through the tea, but they do frequently use the tea companies foot paths and roads for travel to nearby fields (E. Ijaku, Estate Manager. James Finlays (Uganda) LTD. Interv iew conducted July 14, 2006). Though it may be the most effective measure, guarding also ha s arguably the most dramatic social impacts because of the indirect costs of crop raiding: labor investment, lost opportunities, and increased exposure to mala ria vectors (Tchamba, 1996; Hill, 2004; NaughtonTreves et al. 2005; Linkie et al. 2007). Incr eased time guarding means women must have to spend more time in the fields and away from household duties. Children may be pulled out of school to care for younger brothers an d sisters or to cook at home or to collect resources while the mother is away. Other children are pulled out of school to guard their familys fields. Although fields are guarded year round, crop raids by primates is especially bad during the harvest times, and children can spend three days or more per week in the fields instead of in school. The more well-off househol ds can hire workers to work and guard their fields or may 112
use dogs, while their children stay in school. Hous eholds also cope with crop and animal raiding in other ways. Many households sell their live stock to buy food lost from crop raiding. Others will sell or buy land because of its susceptibil ity to crop raiding. An unknown number of kill wildlife. Farming at the edge or near forests and wetla nds is difficult for local farmers (Figure 4-2). Edge management is important in order to re duce the attractiveness and food availability for would-be raiders. Many practitioners have preferred a hard bounda ry, especially in the case of separating the park and its nei ghbors (Hayes, 2006). KNP has te sted a number of different strategies, such as digg ing a trench along with planting Mauritius thorn ( Caesalpinia decapetala ) along a portion of the boundary on the east side of the park in 2001 (Chhetri et al. 2003). Other methods such as electric fencing has been used in protecting small farm ing areas in Kenya and Zimbabwe (Kassilly, 2002; Osborn and Park er, 2002), but material s, installation, and maintenance for large-scale applications is impractical and unaffordable for most rural communities near KNP. Another typical tactic ha s been for scare-shooting by wildlife rangers, but this has little deterrence upon crop raiders (Osborn and Parker 2002). In most cases, the animals are gone from the field by the time the farmers submit a complaint via the LC1 (village chairperson) or in person to park guards. Collective management has been recommended as one of the most important mechanisms to address crop raiding and res ource shortages (Scherr, 2000; Ch iyo et al. 2005). Such actions would include crop selection, planting, weeding, or defensive strategies such as guarding. In the densely populated landscape outside the park, effectively combati ng crop raiding must be done cooperatively. Prevention measures will be foiled if one or more farmers plants maize at the 113
edge of the wetland, while others plant buffer crops or trees. However, such strategies have cultural issues. In this area, it is uncommon for the Batoro to manage or farm collectively, but Bakiga typically engage in collective fa rming arrangements (Hartter, pe rs. obs.). Group organization for guarding is done rarely in this area. When done so, only the Bakiga will participate (Hartter unpublished data). Others have called for build ing local management institutions to balance conservation objectives (i.e., KNP) and the needs of local farmers (Infield and Namara, 2001; Hill, 2004). However, most respondents believ ed the responsibility of responding to crop raiding fell on KNP (52%) and on the local government (30%) (H artter, unpublished data). Around KNP, no one perfect solution will exist. Instead a combination of measures is appropriate; one that incorporates the dynamic interactions of soci al and natural systems (Berkes, 2004), but also ones that do not rely on fads or solely on anecdotal evid ence (Salafsky et al. 2002). For the estimated up to one billion people w ho generate much of their livelihoods from forests (Byron and Arnold, 1999), more research is therefore needed to address the efficacy of a series of direct and indirect mechanisms to help cope with wildlife and resource shortage problems outside KNP and other protected areas. Overall, the superpixel was effective in identifying a geographic random sampling of households and integrating multiple scales eff ectively for study of KNP and its surrounding landscape. While landscape level approaches may s how more coarse scale trends, household land uses and adaptations over time are useful to verify that this change is happening. Results from the land cover analysis show a net decrease in forests and wetland area outside KNP while land under cultivation has increased. These results co incide with reported trends at the household interviews. Superpixels can reduc e costs because they can provide an accurate representation of 114
landscape change and temporal trends, but more importantly can be used to understand household-level trends. Using such techniques in a linked methodology can lead to an improved examination an d understanding of land change processes and to better informed management decisions (Vogt et al. 2006) and could be a useful monitoring tool for protected area managers. However, while the superpixel methodology ha s been shown to be useful in dense agriculturally-dependent settlements outside a forested park, these results should be compared to communities outside savannah parks, where raid ing by large, migratory animals is more common. The superpixel methodology was appropria te within a landscape where landholdings are relatively small. This methodology could be tested in Latin America where landholdings are much larger or in pastoral landscap es where communal land tenure is common. Conclusion Although protected areas have become th e primary mechanism for biodiversity conservation, their establishment can have l ong-term impacts on land use, land cover, and livelihoods of people living near them. Sin ce land use and resource access within Kibale National Parks boundaries are prohibited, park neighbors turn to the forest fragments and wetlands to sustain their livelihoods. These area s serve as important resource bases for local people as well as biodiversity habitats that harbor many wild plant and animal species. They are also problematic for local farmers, since crop raids by primates and elephants emanate from these natural fragments. While the boundaries of Kibale National Park have remained intact over time (Southworth et al. in review), ther e has been extensive conversion of wetlands and forest fragments which diminishes the supply of resources. Households have coped with resource shortages and crop raiding in many ways. Most people plant tr ees or spend longer to gather enough resources. Guar ding was cited most as a respons e to crop raiding, but avoiding 115
planting crops nearest the forest and wetlands wa s also a common strategy. Our results suggest that wealth, ethnicity, and dist ance from park are important fact ors in determining response to these issues, but head of household gender was no t. While there is no perfect solution to effectively mitigate crop raiding animals and to supply resources, we found that engaging in a mixture of activities is the needed. Finally, the superpixel methodology was useful in linking landscapeand household-level scales to examine the impacts of park establishment. More research is needed however to test this me thodology in areas with dispersed settlements and larger landholdings or savannah landscapes. 116
Figure 4-1. Kibale National Pa rk and surrounding landscape in western Uganda. (Note: while officially KNP now includes a game corri dor that connects the southern portion of KNP to Queen Elizabeth National Park, this co rridor was formally gazetted as part of KNP in 1993. Before that time, it was mixed agricultural land. Therefore, it was not classified as park in this analysis Since 1993, the tota l park area is 795km2, but only 561km2 of parkland (excluding the game corridor) is covered in this analysis. 117
Figure 4-2. Farming along the forest edge. The KNP boundary is clearly demarcated by Eucalyptus trees, separating forest (left) and fields of maize (right) (Photo by J. Hartter) 118
Table 4-1. Land cover proportions and class-level metrics outsi de KNP in the East Study Area (based on 5-class land cover classi fication from 2003 Landsat image) Year Land Cover Proportion of Total Land Area No. Patches Edge Density (m/ha) Mean Patch Size (ha) Connectivity (0-100) Dist. to Nearest Patch Edge (m) 1984 forest 32.0% 165 52.4 9.4 0.244 97.5 1995 forest 28.3% 211 57.5 6.3 0.230 91.1 2003 forest 21.4% 207 35.5 4.3 0.000 108.7 1984 wetland 38.4% 731 108.0 2.8 0.136 69.9 1995 wetland 26.1% 884 66.4 1.4 0.058 77.5 2003 wetland 34.0% 761 86.9 2.2 0.000 72.5 1984 other land cover 29.6% 949 83.7 1.8 0.095 71.1 1995 other land cover 45.6% 729 99.1 3.8 0.149 66.2 2003 other land cover 44.5% 575 99.4 4.8 0.000 65.8 Note: Connectivity: range 0 to 100, expressed as a percen t) Measurement of connections between patches of the corresponding patch type divided by the number of possible connections. A value of 0 is when either the class consists of a single patch or none of the patches of the focal class are connected; and a value of 100 indicates that every patch of the class is connected. Here, we used threshold=60m. Table 4-2. Land cover proportions and class-level metrics outside KNP in the West Study Area (based on 5-class land cover classi fication from 2003 Landsat image) Year Land Cover Proportion of Total Land Area No. Patches Edge Density (m/ha) Mean Patch Size (ha) Connectivity (0-100) Dist. to Nearest Patch Edge (m) 1984 forest 42.8% 284 61.1 15.0 0.284 79.8 1995 forest 52.0% 118 71.3 44.5 0.826 74.4 2003 forest 30.5% 395 41.2 7.2 0.000 95.3 1984 wetland 21.4% 1890 56.3 1.2 0.037 75.8 1995 wetland 9.4% 1646 29.6 0.6 0.027 92.2 2003 wetland 15.4% 1871 39.6 0.7 0.000 85.9 1984 other land cover 35.8% 1499 62.9 2.6 0.057 73.8 1995 other land cover 38.6% 1433 61.6 2.8 0.054 72.9 2003 other land cover 54.1% 861 63.1 7.1 0.000 68.8 119
Table 4-3. Land cover change detec tion for the east and west study areas East SA West SA 1984-1995 1995-2003 1984-1995 1995-2003 Loss of Natural Area Area (ha) % Area (ha) % Area (ha) % Area (ha) % Forest Crops 432 8.3 430 8.3 713 6.9 1503 14.6 Forest Tea n/a n/a n/a n/a 455 4.4 743 7.2 Wetland Crops 961 18.6 338 6.5 702 6.8 277 2.7 Wetland Tea n/a n/a n/a n/a 59 0.6 40 0.4 Addition of Natural Area Crops Forest 190 3.7 240.48 4.6 1067 10.36 422 4.1 Tea Forest n/a n/a n/a n/a 276 2.7 143 1.4 Crops Wetland 345 6.7 632.34 12.2 261 2.5 374 3.6 No Change Forest Forest 843 16.3 625 12.1 3020 29.3 2310 22.5 Wetland Wetland 649 12.6 774 15.0 454 4.4 398 3.9 Crops Crops 971 18.8 1474 28.5 1425 13.8 1821 17.7 Tea Tea n/a n/a n/a n/a 132 1.3 449 4.4 Table 4-4. Patch-level statistics for east and west study areas East Study Area West Study Area Mann-Whitney U Test Mann-Whitney U Test Asymp. Sig. (2-tailed) Asymp. Sig. (2-tailed) Date Land Cover Mean Patch Area Dist. to Nearest Patch Mean Patch Area Dist. to Nearest Patch 19841995 Forest 0.359 0.175 0.242 0.092 19841995 Wetland 0.000* 0.000* 0.545 0.000* 19841995 Other land cover 0.000* 0.025* 0.677 0.259 19952003 Forest 0.036* 0.000* 0.008* 0.000* 19952003 Wetland 0.000* 0.000* 0.000* 0.000* 19952003 Other land cover 0.000* 0.000* 0.060 0.000* 19842003 Forest 0.001* 0.005* 0.059 0.000* 19842003 Wetland 0.000* 0.375 0.000* 0.000* 19842003 Other land cover 0.000* 0.000* 0.021* 0.000* *significant at 0.05 level 120
Table 4-5. Respondents perception of change in resource bases in the last 10 years Perceived changes in resource bases (n = 130) Wetland Forest Yes 87.7% 66.2% No 7.7% 23.8% n/a 4.6% 10.0% Table 4-6. Reported changes in forest fragments and wetlands Reported changes in forest (n = 86) Proportion Reported changes in wetland (n = 114) Proportion More people 73.3% Lower water level 57.7% Smaller size 80.2% More people 53.1% Fewer trees 84.9% Smaller size 52.3% Fewer tree spp. 71.5% Table 4-7. Reported reasons for difficulty in obtaining resources in wetlands and forest fragments Wetland Forest Reason for the difficulty n = 59 n = 55 Less resources 89.8% 83.6% Stopped by owner 13.6% 16.4% Stopped by government 1.7% 3.6% More people 1.7% 3.6% Must travel further 3.4% 0% Table 4-8. Resource and wild animal concerns among to all respondents Response (n = 130) Question Yes No Don't know N/A Firewood a concern? 85.4% 14.6% 0% 0% Enough forest to meet future needs? 18.5% 70.8% 4.6% 6.2% Enough wetland to meet future needs? 27.7% 56.2% 13.1% 3.1% Crop/livestock raiding compared to past? 30.8% 43.1% 10.0% 16.2% 121
Table 4-9. Relationship of re sponse and ethnicity, wealth, he ad of household gender, and distance to the park boundary Relationship of Response and Variable Question Ethnicity Wealth Head of HH Gender Distance to KNP Firewood a concern? 0.362 0.587 0.206 0.345 Enough forest to meet future needs? 0.957 0.030* 0.113 0.009* Enough wetland to meet future needs? 0.312 0.730 0.364 0.600 Crop/livestock raiding compared to past? 0.480 0.402 0.455 0.098* *significant at 0.1 level Table 4-10. Household responses to resource shortages amon g all respondents. (Note: households can partake in a variety of combinations of responses, choosing one, several, or none of these responses to shortages) Reported Responses to resource shortages Proportion (n=130) Plant trees?* 70.0% Increase time to gather resources 67.7% Buy some fuelwood/charcoal/trees* 41.5% Increase travel time to resources 35.4% Go to different place to get fuelwood* 33.8% Table 4-11. Relationship of re sponse to resource shortage a nd ethnicity, wealth, head of household gender, and distan ce to the park boundary Relationship of Response and Variable Question Ethnicity Wealth Head of HH Gender Distance to KNP Plant trees 0.014* 0.537 0.852 0.409 Increase time to gather resources 0.305 0.478 0.983 0.744 Buy some fuelwood/charcoal/trees 0.013* 0.429 0.255 0.442 Increase travel time to resources 0.115 0.225 0.195 0.262 Go to different place to get fu elwood 0.039* 0. 677 0.81 4 0.168 *significant at 0.1 level Table 4-12. Questions posed to households rega rding problems of and response to crop and/or livestock raiding Question Yes Proportion Do you and/or your household have problems with wild animals? 73.8% 96 of 130 respondents do you and/or your household do anything about the wild animals? 78.1% 75 of 96 respondents 122
123 Table 4-13. Household responses to crop raid ing. (Note: an unknow n number of offending animals are killed by landholders) Reported response to crop raiding Proportion (n=75) Guard 90.7% Certain crops not planted close to natural areas 49.3% Stopped growing/living in certain area 29.3% Stopped growing certain crops 5.3% Table 4-14. Household responses to resource shortages amon g male and female heads of household Relationship of Response and Variable Question Ethnicity Wealth Head of HH Gender Distance to KNP Guard 0.004* 0.000* 0.964 0.693 Certain crops not planted close to natural areas 0.066* 0.193 0.308 0.250 Stopped growing/living in certain area 0.064* 0.542 0.471 0.249 Stopped growing certain crops 0.640 0.058* 0.325 0.342 *significant at 0.1 level
CHAPTER 5 IMPOSING WILDERNESS? THE IMPACTS OF A FOREST PARK Introduction In the century and a quarter since the form ation of Yellowstone NP, the worlds first national park, the number of nati onal parks, reserves, and other protected areas around the world has grown substantially. There are now over 100,000 parks and protected areas around the world, covering well over 10% earths land area 4 (Hayes, 2006). Although protected areas have become the most recognizable mechanism for te rrestrial conservation, their establishment and management certainly are not w ithout criticism. As the park system worldwide not only has grown, but also (arguably) matured, the park concept has evolved to encompass multiple meanings and interpretations. The World Conser vation Union defines a protected area as an area of land and/or sea especia lly dedicated to the protection and maintenance of biological diversity, and of natural and a ssociated cultural res ources, and managed through legal or other effective means (as cited in Murphree, 2002). Among other features, this and similar statements illustrate that pr eservation of biological dive rsity has supplanted wilderness preservation as the ultimate objective and justification of conservation oriented parks. Fortress conservation, or more colloquially fin es and fences, is characterized by strict exclusion of humans, the prevention of consump tive use, and minimization of other forms of human impact (Hulme and Murphree, 2001) and or iginally having no linkage with livelihoods or development goals (Salafsky and Wollenberg, 2000). People are meant to use resources outside of the park, and plants and animals are meant to stay in the park. As such, wildlife enthusiasts and policy-makers in East Africa believed that national parks would create a sand harmonious 4 We use the term park to refer to all types of protect ed areas. The World Conservatio n Union (IUCN) prescribes 6 categories. 124
balance between man and wildlife without di sturbing the regions social and economic development (Ofcansky, 2002). However, with the erection of bounda ries, relations of neighboring communities often sour ed and entailed conflict (Weste rn, 1994; Brockington, 2002). Compliance by the local population was based on a fear of punitive measures (Wells and Brandon, 1992). The idea of fortress conservation has sparked many debates over the past few decades, especially since the rise of other conservati on narratives (Hulme and Murphree, 2002; Ferraro and Kiss, 2002). Early parks in Sub-Saharan Africa (SSA) in general were not established for the purpose of biodiversity conservation. Although protected areas primary objectives were protection of wild places and big game for hun ting and timber supplies, it was not until the 1970s, with the growth of the environmental movement that pa rks incorporated biologists attempts to maintain the integrity of wild ecosystems (Child, 2002). Key conservation proponents in SSA including such biologists as J. Terborgh, T. Struhsaker, and J. Oates among others have remain ed steadfast in their support for exclusionary boundaries (Oates, 1999; Terborgh, 1999; Terborgh, 2002; Struhsaker et al. 2005). Struhsaker argued that biodiversity conser vation and humans cannot coexist in a park and that human presence in the park may lead to the introduction of exotic plants, animals, and diseases that may adversely affect indigenous wildlife. (Struhsaker, 2002, p. 100). Moreover, he called other approaches such as extractive reserves, sustainable harvest extremely problematic for conserving the full array of biodivers ity (Struhsaker et al. 2005). Both Oates (2002) and Struhsaker (2002) continue to call for increased prio rity of wildlife protec tion and to assert th at rigorous policing by armed guards in the park is necessary. Terborgh (1999) claims the fate of the rain forest is threatened by constant economic demands and human influence. Unde r this narrative, the 125
survival of biodiversity supersed es the immediate material needs of humans. The benefits of a park to society as a whole are considered to outw eigh the cost to the relatively small number of individuals whose lives may be directly impact (Terborgh and Van Schaik, 2002). Though the fortress conservation approach pers ists in SSA, its support has waned in the last two decades with the intr oduction of other narratives within the conservation community (i.e., community conservation and integrated c onservation and development projects) (Salafsky and Wollenberg, 2000; Hulme and Murphree, 2002). In recent years, it is has been recognized that whole landscapes cannot be isolated from the influence of humans. In order for conservation to succeed, their livelihood needs must be considered (Adams and Hulme, 2001; Brown, 2002), especially those of communities neighboring the pa rk (Salafsky and Wollenberg, 2000). Differentiating human livelihoods from what is deemed as pristin e nature was not only considered unsustainable, but also impo ssible (McNeely, 1994; We stern, 2001; Brown, 2002) and often led to hostility towa rds the park (Brockington, 2004). Kibale National Park (KNP) in western Uganda was formally established as a national park in 1993 (it was elevated from a forest reserve) in the midst of this ongoing conservation debate. Like the other forest parks in SSA, it is surr ounded by smallholder agri culture that depended on the resources outside of the park (Howard et al. 2000). As w ith other parks in Uganda, the creation of KNP was guided by the Uganda W ildlife Statue, 1996, which states, a national park shall be an area of international and national importance becaus e of its biological diversity, landscape or national heritage (Ugand a Wildlife Authority, 2003). It sought to preserve biodiversity within park boundaries first and foremost. In 1993 the park excluded nearly all resource collection and timber concessi ons and forcibly removed settlers from within its borders (Chapman and Lambert, 2000). By 1992, there were approximately 13,000 people 126
living inside the corridor. These people were ev icted and relocated to la nd north of KNP (Chege et al. 2002). In addition, while a national revenue sharing progr am was installed, few benefits have accrued at the local level (A rchabald and Naughton-Treves, 2001). Characteristic of fortress conserva tion is its measures of success. Within this view, it is held that the status of plants, animals, and th e ecological community as a whole is the ultimate measure of successful conservation (Struhsaker, 2002; Salafsky et al. 2002). On the other hand, community conservation proponents argue that w ithout the support of neighboring communities, the protected area cannot be deemed a success or even sustainable (Wells and Brandon, 1992; Hulme and Murphree, 1999). Protected areas around the world are centerpieces for the conservation movement and protec tion of biodiversity, and they have been established often against the wishes of the local population, excluding them from access, resources, and settlement Brockington, 2004). Often, those most impacted by the park most are th e local, rural poor, and most of the beneficiaries seem to be the wealthier and foreign visitors. However, is there a case that a park, which by many, can be considered an ecological and biological success, also can be accepted by the neighboring communities? This pa per examines two questions regarding the perceptions of households in communities surr ounding KNP; 1) Do house holds perceive that Kibale National Park has helped, hurt, or had no positive or negative impact on their family? and 2) What are the perceived benefits and problems associated with KNP? Study Region Physical Environment Kibale National Park is a medium-altitude tropical moist forest covering about 795 km2 in western Uganda and its boundari es correspond to those of th e 1932 forest reserve boundary combined with the former Kibale Forest Corridor Game Reserve (Chege et al. 2002). This transitional forest (between lowland rainforest an d montane forest) is at an average elevation of 127
128 1110-1590m and is believed to be a remnant of a previously larger mid-altitude forest region (Struhsaker, 1997). Although the amount of rainfall and length of season change, the average annual rainfall for the region is 1543 mm (average 1903-1999) and 1719mm (1990-2005) (Chapman et al. 2005). The landscape outside the park has become a patchwork of small farms (most <2ha in size), tea estates in some areas and a network of bottomland fore st fragments and wetlands that serve as a resource base for wate r and fuelwood, effectively isolat ing the park. Land scarcity has forced land scarcity has forced many farmers to farm to the edges of fore sts (Naught on-Treves, 1998). In addition, 58% of the land within a 1.5 km periphery of the park is used for small holder agriculture (Naught on-Treves, 1998). Ugan da continues to lose 50,000 ha (9.8%) of its forest annually (National Envi ronment Management Authority, 2 001, mostly to conversion to agriculture. As agriculture continues to encr oach on unprotected fo rest lands mainly for substance purposes, these forests become more isolated, ra nging in size from 1.2 to 8.7 hectares in size (Gille spie and Chapman, 2006). In addition wetlands continue to be drained and converted into othe r land uses such as growing crops and fuelwood and ex panding grazing land continues to take place (Crisman et al. 2003). Tea dominates much of the landscape bordering the northwest portion of KNP. Tea producers range from large multinational companies with hundreds of hectares, to smaller individual holdings of less than one hectar e (Mulley and Unruh, 2004). Established in the 1950s and re-established in the late 1980s following stagnation during the Amin (1971-1979) and Obote years (1980-1985), the land under tea has been limited ma inly by soil pH (optimal is 5.2) and limited to the west side because of land availability 5 5 A.T. Joseph, General Manager. Rwenzori Comm odities LTD. Interview conducted July 21, 2006.
Social Environment Rapid population growth, high population densit y, and heavy reliance on agriculture for income characterize the landscape surrounding KNP (Archabald and Naughton-Treves, 2001) and land pressure has steadily increased (C hege, 2002; Lepp and Holland, 2006). Economic inequality, ethnic di sparity, intensive s ubsistence agriculture, and a decreasing resour ce base in these communities serves to i llustrate a common trend around protected areas in developing countries. Country-wide, 90% of Ugandas population are directly dependent on natural resources for subsistence needs (Uganda Ministry of Water, Lands, and Environment, 2002), but around KNP, nearly everyone depends on the near by resources to sustain their livelihoods (Hartter, pers. obs.)With limited land availability, farm sizes are characteristically small (< 5 acres on average, Hartter, unpublished data). This region is one of the most densely populat ed areas on the African continent (Lepp and Holland, 2006). Population outside the park has nearly treble d between 1959 and 1990 (Naughton-Treves, 1998) and has grown by 76% between 1980 and 2002 (Uganda Bureau of Statistics, 2005). Density is es timated to be 262 individuals/km2 on the west side of the park and 335 individuals/km2 on the east side within 5km of th e park boundary (Hartter, unpublished data). Agriculturalists in the area belong to two main ethnic gr oups: the Batoro (west side of KNP) and the immigrant Bakiga (east side of KNP), who came to the Kibale region from southwestern Uganda beginning the 1950s a nd 1960s (Turyahikayo-Rugyema, 1974; NaughtonTreves, 1998). Employment opportunities are limited on both the eas t and west side of the park and most can be considered unskilled labor. On the eas t side, some are employed outside the home to 129
work in cultivated fields or as taxi drivers. Others earn inco me from small shops in town, or from selling charcoal, trees, or crops. Another sma ll percentage is employed by KNP or through tourist-based activities. The opportunities are similar on the west side of KNP except for employment with scientists and the presence of the tea companies. Here, some landholders grow small patches of tea on their own land and sell it to the tea companies. The largest percentage is employed by the areas tea-grow ers (Mulley and Unruh, 2004). On both the east and west side, access to external markets is difficult, as the nearest population center is located nearly 20km from the western edge of KNP. Methods The analysis presented in this paper represents a portion of a larger study that investigated the impacts of protected areas on surrounding landscapes. A practic al way to integrate different objectives of our larger study is to designate a study unit an area both large enough and small enough to include the scales of la nd cover, biodiversity, and human activities and enable us to measure them so that satellite remote sensi ng images and the on-the-ground measurements are tied operationally to the same area of land (Southworth et al. in review; Hartter, in review). This sampling strategy differs from previous so cial science studies outside KNP that were limited to one or a small number of locales (Naughton-Treves, 1997; Edmunds, 1997; Lepp, 2004). To provide this link, sampling units (nam ed superpixels) comprising 9 ha areas were used. These superpixels are centered on spatially randomly selected coordinates within 5 km of the KNP boundary and within two areas (Figure 5-1). The West Study Area is dominated by the Batoro ethnic group, while the East Study Ar ea is dominated by the Bakiga. In all, 95 superpixels were selected randomly within these areas. These 95 superpixels include 60 villages and a sample pool of 417 households (i.e., households that hold land contained within at least one of the superpixels). Between May and August 2006, 130 semi-structured household 130
interviews were conducted. At each study area respondents were chosen opportunistically, based on population weighting (i.e., the numbe r of respondents for each study was proportional to the number of landholders holding land within the study area). Superpixels containing more landholders had a higher sampling intensity. In addition, 16 interviews were conducted with village chairmen 6 (LC1s) within these communities. All interviews were conducted in English, or in Rutoro or Rukiga using an interpreter. Results Respondents were asked whether KNP had helped them and/or harmed them. If it had neither helped nor harmed, there was no perceive d impact from the park (Figure 5-2). Most respondents believed that the park helped them in some way (61%). The benefit reported most was that the park keeps the environment, m eaning timely and adequa te rainfall, climate regulation, temperature regulation, moisture levels, fresh air, and also the intrinsic value (Figure 5-3). Another important benefit was that KNP was perceived to keep animals. This expresses the common perception that the park and other natural areas contain most wild animals, which as a result do not come to their fields at all or come less of ten. There seems to be a belief that in the absence of the park and ot her natural areas, animals would surely come to their lands to raid their fields. Only 14% of the respondents mentioned employment as a direct benefit, either as contractual unskilled labor or tourism or the research station maintained by Makerere University. A smaller percentage, 6%, felt their household be nefited in some way from tourism. 6 The local council system (LC) was implemented to serve as institutions for local self-governance. It would enable Ugandans to participate in decision making at a more equa l level regardless of gender, age, ethnicity or political affiliations, while giving local governments autonomy, meaning they have the legislative and executive authority within their listed areas of jurisdiction. The LC system has five levels of local government: village, parish, subcounty, county, and district. For more on the LC system, see Saito (2003). 131
Thirty four percent of responde nts (44 of 130) felt they were harmed in some way by the national park (Figure 5-2 By far, wild animals that trampled, otherwise damaged, or ate their crops, or attacked their livestoc k were perceived as an important problem due to the parks existence (Figure 5-5). Only 11% felt that being denied access to resource s within the park was a significant problem to them. Smaller percen tages still felt that no help from the park, displacement, and land prices were significant pr oblems. In addition to the 66% of respondents who felt that their household was not harmed by the park, 21% of respondents perceived both positive and negative impacts from the parks presence (Figure 5-2). Another question posed to gauge assessments of the park was whether given the choice, they would rather live closer or further from the park boundaries. Sixty-nine percent (n=130) said that they would prefer to live closer to KNP than furthe r from its boundaries (Figure 5-5). Most respondents felt that, if given the choice, KNP should remain (73%, n=130) rather than be abolished (9%) (Figure 56). It is clear that people find environmental benefits and keeping animals away as the two major benefits of the park, and these perceived benefits are major reasons respondents said the park should rema in intact (Table 5-1). Smaller proportions of people reported other reasons such as educati on, employment, and generating revenue for the government (11%, 13%, and 12% re spectively). One important be nefit that was noted was the belief by some respondents that the park may lessen its strict regulation on resource harvesting in the future. Residents say that park manageme nt will see the plight of the poor households surrounding the park and they will help them in the future. Although the vast majority reported that they would like the park to stay, some said that the park caused too many problems and it was best that it be dissolved. Nearly all of respondents who reported the park should be dissolved cited human-wildlife conf lict as the reason (Table 5-2). 132
We also asked local chairmen (LC1) about the pa rk (Table 5-2). Most said that their zone had benefited in some way from the parks existe nce. School infrastructure, including classroom renovation and construction as well as teachers house construction were cited most often as benefits from the park. Keeping the environment and chasing animals from respondents lands were also important be nefits of the park. Discussion The results of the household surveys reveal that the majority of households feel that they benefit in some way from of th e national park. Whether these be nefits (e.g. fresh air, providing rainfall, fertile soil, and keeping animals from th eir fields) are real or only perceived, they are important. When asked, we found that education programs, a well-intentioned, but perhaps misinformed volunteer, even park officials, and local lore were all res ponsible for perpetuating these perceptions of park benefits. Landholders also believe that the park will keep animals away from their fields. In some instances, la ndholders rely on the tact ic of scare-shooting by wildlife rangers. However, this action has litt le deterrence on crop raiders (Osborn and Parker, 2002). In most cases, the animals are gone from the field by the time the farmers submit a complaint via the LC1 or in pers on to park guards. According to them, the park has abundant forage and habitat and by mainta ining the park, the wild animals have little reason to stray far from its boundaries to raid their fields. Only those closer to the pa rk are vulnerable. In fact, several landholders also mentioned that land prices had increas ed as a result of the parks presence, a major contrast to results of a decade earlier in which land near the park was said to be worth less than that more distan t from the park (Naughton-Treves 1997). Many landholders farm near the KNP boundary and nearly all compla in about crop raiding (NaughtonTreves, 1997; Chege, 2002). Hist orically, since the land next to the park was considered more vulnerable to wild animals and limited extractive privileges were allowed within KNP, this land 133
was given or sold to Bakiga immigrants (N aughton-Treves, 1998), or poorer landholders who only could afford a small parcel. The only availabl e or cheaper parcels of land were adjacent to the park. However, recently this land seems to have become more valuable. Although the land continues to be vulnerable to the largest variety of animal s, (Hartter, unpublished data) landholders have come to reali ze that only the wealth ier households will be seeking to buy land here and using them to grow trees or tea, which will not be raided 7 Landholders realize that the only people looking to buy this land have the fi nancial capital to absorb the risk and enough capital to invest in income-gen erating activities and consequentl y, land prices in some instances have increased. Surprisingly the environmental services of pa rks were mentioned more often as benefits than employment, infrastructure, or other material impacts. Very few reported that jobs were benefits. Some in the area are employed th rough timber concessions in cutting and sawing the pinus and cupressus exotics within the park. Another sm all percentage of the local population finds employment as research assistants, sel ling crafts to tourists, game guards, park management and staff, and contractual and shortterm labor. The vast majority of the population remains engaged in activities not directly linked to the park. KNP allows only very limited extraction of downed fuelwood, but only through organized groups and special agreements. Park authorities have often denied re quests by individuals to enter and collect medicines and other resour ces, even those of tr aditional and cultural importance. Park officials stress that commun ities must organize themselves. They not only expect communities to organize to petition for permission to harvest within its boundaries, but also to manage crop raiding. Declining size and increasing isolation of forests and wetlands 7Respondent 37. Interview conducted June 29, 2006. 134
outside the park, increased fr agmentation of the landscape, and increased population and land under cultivation all have had dramatic impacts. Edge habit has increased and the ability of large animals to range widely without crossing agriculture is extremely limited (Newmark et al. 1994; Naughton-Treves et al, 1998), thereby increasi ng human-wildlife enc ounters along the forest boundary. In aggregate though, we l earned that only one third of respondents feel crop raiding has become worse in the last ten ye ars (Hartter, unpublished data). As mentioned above, 31% of the respondents felt that the park harmed their household in some way. Nearly all of these said that they had problems with wild animals. The park acknowledges these problems and many of the la ndholders want compensation for their losses. However, under the Uganda Wildlife Statue 1996, compensation to individual farmers for lost crops due to crop raiding is not permitted in a ny form (e.g. monetary, loans, food replacement, food vouchers). Even though respon dents feel that the park ( 52%, n=130) and the government (31%, n=130) must manage the crop raiding, mo st communities and individuals, especially those not directly neighboring the park, must develop their own deterrence strategies, or develop their own means to cope with losses. Protected areas are celebrated for their pres ervation of local biodiversity and endemic species; mitigation of illegal poaching and timber exploitation ; and protecting potential water sources. Ecotourism derived from park establishment is often touted as the selling point to communities because of its ability to contribute to both conservation and development goals (Dixon and Sherman, 1991; Ashley et al. 2000). As tourism builds, if the communities have a direct economic benefit from tourism, such as re venue sharing program, thei r attitudes are likely have a positive influence on local perception of protected area (NaughtonTreves, 1998; Walpole and Goodwin, 2001; Archabald and Naughton-Tr eves, 2001; Infield and Namara, 2001). 135
Recognizing the needs and pli ght of communities, through the Uganda Wildlife Statue 1996, the Ugandan government mandated that 20% of the gate receipts from national parks were to be shared with local comm unities (Chhetri et al. 2003). Howe ver, the amount of money that returns to communities is small, and its allocation is complicated. Kibale National Park is a relatively small park, and unlike most of the la rge parks in East Africa and Sub-Saharan Africa in general, it is a forest park. Large game speci es generally either do not exist or are difficult to see in forest parks. Kibales main attracti on is the opportunity to see and track chimpanzees ( Pan troglodytes). Other tourists come for bird-watch ing in the nearby wetlands or sometimes for the other primate species in the forest. On the whole, though, forested parks such as Kibale do not have the visible charismatic megafauna, vast open expanses of land, and easily accessed wildlife as in the many famous East African savannah parks such as Serengeti and Ngorongoro National Parks. 5-3 shows park visitor statistic s compared to the Uganda national park system and compared to two parks in Uganda. The first is Murchison Falls (3,860km2), the largest park in Uganda and the second is Bwi ndi Impenetrable National Park (331km2), a popular forest park in southwest Uganda on the border with the Democratic Republic of Congo known for its contingent of endangered mountain gorillas (Gorilla gorilla beringei ). KNP appears to function more as a stop over poi nt for tourists traveling onward to Queen Elizabeth National Park than a primary destinat ion of its own. Each year, a relatively small number of tourists visit the pa rk. Encouraging though is the rapi d growth in visitor numbers and the high fees visitors pay to track chimpanzees, but few tourists spend more than one night, if any at all. On average, KNP represents less th an 5% of the total number of visitors to the Uganda national park system and only park entry fees are disbur sed among neighboring communities. Each visitor pays US$25 (tariff as of July 1, 2006-June 2008 for non-residents for 136
1 day) for park entry. Forested parks do not at tract large numbers of tourists (Table 5-3; Struhsaker et al. 2005) and therefore, the pool of possible revenue to be disbursed to communities is quite small. Expenses exceed ed their revenue generated (Archabald and Naughton-Treves, 2001), which is common for many pa rks in the world (James et al. 2001). In addition, an estimated 80-85% of the tourists go to the popular Kanyanchu tourist site for chimpanzee tracking 8 So while the entry fee of US$25 is c onsidered gate fees and eligible for distribution to the local comm unities, the US $70 (tariff as of July 1, 2006-June 2008 for nonresidents) fee paid by tourists to track chimpanzees is not. Archabald and Naughton-Treves (2001) reported that of the 27 parishes that border KNP and are eligible for disbursement of the funds, onl y five successfully received benefits from the park in 1999. Of the total US$3000, if divided evenly, only (US$600) actually came back to each of the five parishes. Within each parish, the LC1 zones (typically 5-10) must submit their case to the LC3 (sub-county level) for review. Spread over the entire communitys population, the actual benefits at the LC level, the househol d, or at the individual level are miniscule, as demonstrated in villages outside Lake Mburo National Park (Infield and Namara, 2001) and Mgahinga Gorilla National Park in Uganda (A dams and Infield, 2003). The meager amount of money that does come back to the communities is most commonly used in two ways: to build and maintain roads and bridges; and building schoolrooms or staff houses. While the limited funds can be useful, projects are small and securing funds is problematic. Each LC1 zone must reapply the following year if pr ojects have not been completed. Only six of ten LC1s we interviewed said they had applied for funding or planned on applying within the current and future five-year elected term. 8Tumwesigye, C. KNP Chief Warden. Interview conducted June 12, 2007. 137
Some respondents also indicated one of the benefits of being near th e park was that they were situated to take advantage of potential tourist-based activiti es. If they had the capital or could organize a group of community members with common interests, then perhaps they could reserve some forest or wetland outside the park fo r tourists to see birds and primates. Others thought they could establish businesses depende nt on the tourism-based revenue such as restaurants, guest houses, or selli ng firewood to lodges. Table 5-3 illustrates some of the tourism potential in the area. Approxima tely 5,000 visitors come to the pa rk annually, but few stay more than a single day (Obua and Harding, 1996). With such limited economic opportunity derived indirectly or directly from the park, it is unclear how substantial ei ther the individual or community benefits can be. Some residents also felt that the park could be a potential future source of resources. They believe that park management might eventu ally acknowledge the pove rty and struggles of neighboring communities and permit greater extraction of wood and other resources. Even though there was no indication from KNP officials that such harvesting would be permitted, some were still optimistic and wanted to be situat ed close to the park in order to take advantage of the opportunity when (and if ever) it materializes. Protected areas are celebrated for their pres ervation of local biodiversity and endemic species; mitigation of illegal poaching and timber exploitation ; and protecting potential water sources. Ecotourism derived from park establishment is often touted as the selling point to communities because of its ability to contribute to both conservation and development goals (Dixon and Sherman, 1991; Ashley et al. 2000). As tourism builds, if the communities have a direct economic benefit from tourism, such as re venue sharing program, thei r attitudes are likely have a positive influence on local perception of protected area (NaughtonTreves, 1998; Walpole 138
and Goodwin, 2001; Archabald a nd Naughton-Treves, 2001; Infield and Namara, 2001). But in actuality, the amount of revenue shared with communities is small and only small and only a small proportion of households directly benefit fr om employment or tourism revenue outside of Kibale (Table 5-3). The same is true outside Mgahinga Gorilla National Park in Uganda (Adams and Infield, 2003). While some communities may be in favor of pr otected areas, there is still the prevailing feeling of wanting that protecte d area not in my backyard (Marcus, 2001). De Boer and Baquete (1998) found crop raiding to be the most significant variable in fluence attitudes about the protected area. Outside Budongo Forest Reserve in Uganda, 65% of respondents said that they were in favor of elepha nt conservation, but respondents ar e apprehensive of having that conservation area near their lands (Hill, 1998). Similarly 88% of households held a positive attitude towards the Maputo Elephant Reserve in Mozambique, but in areas where crop raiding was highest, the attitudes were lowest (de Boer and Baquete, 1998). Research in communities neighboring other parks indicate th at damage to crops and livestock caused by wildlife leads to negative perceptions of the park (Infield a nd Namara, 2001; Mugisha, 2002). Furthermore, hostility towards the nearby national parks further arises when the people feel that have not been adequately compensated for liv estock or crop damages (Naugh ton-Treves and Salafsky, 2004; Gadd, 2005). There are other key influences that shape negative attitudes towards parks and other conservation initiatives. Lepp and Holland (2006), Mugisha (2002), and Twyman, (2001) suggest that the restrictio n of resources contributes to ambivale nce toward the protected area. In Machalilla National Park in Ecuador, Fiallo and Jacobson (1995) found si milar reactions; most people living within and near the park did not find the park to be beneficial in any way. One of 139
the biggest contributors was the fact that the park prohibited the collection of most forest resources, including the fuelwood they depended on. Intensive community extension and work can also improve community attitudes towa rds the park (Infield and Namara, 2001). KNP is a different case. The landscape surrounding KNP has all the ingredients for hostility between the park and neighboring commun ities. There is little direct benefit to communities from revenue sharing programs. Th ere is continuous crop raiding and no provision for compensation under Ugandan law. KNP ma intains only a relatively small community outreach and environmental education program. We found that most people in the surrounding landscape are not pervasively hostile towards the park despite the losses from crop raiding and other problems they perceive. In places there are negative attitudes about the park as Lepp and Holland (2006) found around Bigodi town, but such negative attitudes are not at all as widespread as Edmunds (1997) suggested. Na ughton-Treves and Sala fsky (2004) found that some people are apprehensive of park interven tion on their land. However, we found that only two of 130 respondents think that the park will expand in the future. Given the high density population and agricultura l saturation of the surrounding land scape, park expansion seems unlikely. Not only did most respon dents say they are not hurt by th e park or feel no effect, the vast majority said that given the choice and weighing both the benefits and the costs, the park should stay and not be abolished. Even for local political elites, 14 of 16 LC1s said that KNP overall has been a good thing for their LC zone. Our results may be different than traditional studies for a number of reasons. The superpixel sampling framework includes a random se lection of the landscape and a much larger spatial extent and did not target specific villages along the park boundary (Naughton-Treves, 1996; Edmunds, 1997) or those that have had problems with crop raiding (Mkanda and Munthali, 140
1994; de Boer and Baquete, 1998). We would expect that farmers in villages directly adjacent to the park boundary were harmed especially by animals. However, we found that the park impacts quickly diminished and by 5 km; park impacts were hardly noticeable. Those far from the park, within the agricultural matrix, are far from the continuous tr ee habitat used especially by primates and thus reduced risk (Hill, 1997). Even though intens ive land use has decreased has decreased forest and wetlands resources, it has also decreased primate habitat and buffered many farms by other farms and tea plantations, thus de creasing the extent and number of crop raids by wildlife in the last 10 year s (Hartter, unpublished data) Often, the value of parks to local people is generally thought of in terms of foreign exchange through tourism (Newmark et al. 1993 ) and many are opposed to abolishment of the park (Newmark and Leonard, 1991), but this lin e of thinking tends to originate from big savannah parks. Conversely, the main benefit and reason the park should remain cited was because of the environmental benefits KNP provides. One respondent explains his rationale, What are we do to if the park is gone? How will we get those [resources and benefits]? We can do something about the animals we can guar d, but we cant do anything about not having rain. 9 Can Kibale National Park be considered a successful example of fo rtress conservation? From the landscape-level, park boundaries remain intact and stable over time, and dense forest continues to dominate the park (S outhworth et al. in review). Healthy and viable populations of many primate species prevail within the park. The largest and only viab le population of the once threatened red colobus monkey ( Procolobus rufomitratus ) resides within KNP (Struhsaker, 9 Respondent 97. Interview conducted July 14, 2006. 141
2005). In addition, one of Struhsakers (2002) indi cators of park success is a decrease in illegal activities. Anecdotal evidence suggests large-scale encroachment into the park has virtually halted and boundaries are well unders tood and maintained. In landscap es with adequate rainfall, dense population, and where forest abuts the agri cultural landscape, a harder edge may be appropriate (Naughton-Treves et al. 1998). Wherever wildlife habitat neighbors agriculture, there will be some risk of crop loss and human-wildlife conflict. It seems people understood this and continue to take measures to mitigate problem s associated with the forest such as guarding their crop fields. There are many approaches to conservation, but there is no single solution for all cases. Is fortress conservation necessarily bad? Is a co mmunity-conservation-based approach invariably preferable? It is not ap propriate to assume that all of the elements of a narrative or received wisdom (e.g. fortress conservati on) are wrong and should disappe ar from policy and practice, while all the elements of a c ounter-narrative (e.g. community conservation) are right and should be adopted. In the case of KNP, while the local people may bear the disproportionate brunt of crop raiding and forego many types of activities in and out of the park, the park appears to be socially acceptable. However, human activ ities are key drivers in landscape change and their impacts must be considered with any park management plan. Therefore, the more pressing contemporary issue is how to relate and a mix of strategies that incorporate elements of fortress conservation and community conserva tion, not to prove that one is al ways better than the other. Threats to parks can rise from inequitable costs and benefits, and limited distribution of benefits to communities (McNeely, 1990). Scale is important in considering where benefits of conservation and management projects will accr ue (Salafsky and Wollenberg, 2000; Jones and Murphree, 2004) and especially when determining the claim of tourist revenue (Adams and 142
Infield, 2003). As such, the scale of benefits will be addressed in future research. Should financial benefits of a national park accrue at the village, parish, regional, or national level? The question remains as to what scale revenues can be allocated to be the most effective to the rural poor majority who are the most vulnerable and have the least capacity to cope. Conclusion Contrary to popular belief of exclusionary approaches to co nservation, our results suggest that not only do most households feel they be nefit from KNP (although in ways not widely discussed in the literature on parks), but the majo rity also feel they are not harmed by it, and many feel no effect from the park. The two main benefits of the park cited were environmental services such as rain and fresh air and keeping animals within the park boundary. The most cited harm predictably was crop and livestock raiding by wild animals. Over 70% of respondents say that given the choice and weighing both the benefits and the costs, the park should stay. Even the large majority of local politi cal leaders also said that KNP overall has been good for their local area. Since the study incl uded a geographic random sample of multiple locales and did not target specific villages, households, or set(s) of demographics, our results are representative of household perceptions about the park in the landscape surrounding KNP Future research however should test the superpixel in other environments such as pastoral landscapes, dry savannah landscapes, or around parks out side of the African continent. 143
Figure 5-1. Kibale National Park and surrounding landscape in western Uganda 144
20.8% 33.8% 60.8%0.0% 25.0% 50.0% 75.0% 100.0% HelpedHurtNo effect Figure 5-2. Perceived impact of KNP, n=130 13.9% 6.3% 41.8% 6.3% 73.4% 11.4% 11.4% 17.7% 0% 25% 50% 75%100% Employment Money to govt/infrastructure Keeps animals Tourism Keeps the environment Education Collect resources Other Figure 5-3. Perceived be nefits from KNP, n=79 145
88.6% 11.4% 4.5% 2.3% 2.3% 0% 25% 50% 75% 100% Wild animals No Access to resources No help from KNP Removed from settlement Land prices increased Figure 5-4. Perceived pr oblems from KNP, n=44 31% 69% Closer to KNP Further to KNP Figure 5-5. Where would you rather live?, n=130. The definition of closer and further was left to the responde nt to interpret 146
73% 9% 18% stay go indifferent Figure 5-6. Should KNP stay or go? 147
Table 5-1. Reasons reported why KNP should stay or be dissolved Why KNP should stay? n Proportion Money to government 11 12% Possibility to collect resources in future 9 10% Protects the environment (air, rain) 50 53% Soil fertility 4 4% B/c the government wants/uses it 2 2% Tourism 7 7% Keeps animals away 31 33% Education about environment 10 11% Employment 12 13% Infrastructure (e.g. building schools, roads) 4 4% Obtain resources 5 5% Spiritual/cultural significance 2 2% No reason 1 1% Why KNP should be dissolved? Crop raiding 11 92% Park will expand in future 2 9% Table 5-2. Has the LC zone benefited from KNP and the perceived benefits by the LC1 Has your LC zone benefited from KNP? n Proportion Yes 11 73% No 5 33% How have they benefited? built/maintained roads 2 18% built bridges 2 18% school classrooms/staff houses 7 64% chase animals/scare shoot 5 45% keep the environment 5 45% jobs for constituents 4 36% increased business 4 36% 148
149 Table 5-3. Visitor statistics fo r Uganda national parks, Murchison Falls, Bwindi and Kibale National Parks. Kibale National Park Gr oss Revenue generated and revenue shared with communities is also given. Annual Visitors KNP revenue shared UWA Parks6,7 MFNP6,7 BINP6,7 KNP5,7 Total KNP Revenue (USD) (20% of gate fees)8 Total KNP revenue shared w/ communities 28,098 6,817 1,106 465 N/A N/A 32,027 7,041 2,461 1,890 N/A N/A 42,783 11,039 3,214 3,640 N/A N/A 2479 39,839 12,099 3,437 1850 $70,238 1 $3,000 1 36,943 12,713 2,100 845 $79,988 2 $269 $3,269 52,161 23,169 3,983 1156 $34,930 2 $865 $3,865 58,004 20,284 4,517 1852 $52,499 2 $1,617 $4,617 90,061 34,241 5,075 4872 $116,300-$130,900 2,3 $7,800-$11,050 3 $12,417-$15,667 6133 $191,6004 $24,5329 5391 212,2004 $21,5649 5639 267,4114 $22,5569 1Archabald and Naughton-Treves (2001) 2Chhetri et al. (2003) 3discrepancy in values reported in Chhetri et al. (2003) 4based on totals from UWA given as shillings and converted at 1 USD=1600 Ug. shillings 5visitor information from Archabald and Naughton-Treves(2001), Chege et al. (2002), Chhetri et al. (2003), UWA (2005) 6UWA (2005) 7National Environment Management Authority (1997) 8The Wildlife Statute 1996 makes provision for shar ing 20% of gate receipts with local communities 9approximated from US$20 park entry fee
CHAPTER 6 CONCLUSION Though protected areas have become the most recognizable mechanism of the worlds interest in conservation, their establishment and their management certainly have a significant impact both positive and negative on the neighboring communities. This research examined these impacts of park establishment by focusing on forest fragments and wetlands outside Kibale National Park (KNP). As in many similar cases, in the landscape around the park, there has been substantial population grow th from both in-migration and a hi gh rate of natural increase. The park is surrounded by agricultural land and larg e multi-national tea estates and a network of forest fragments and wetland areas is scattered throughout the landscape. Since small rural communities in western Ugan da are heavily reliant on land and resources, their livelihoods are di rectly linked to ecological systems. As population in the area continues to climb, land and resource pressure likewise increase. Thus, the landscape surrounding KNP has become a mosaic of diverse agri cultural land intermixed with patches of remaining natural areas as households construct strategies to sustain their livelihoods. Remaining ungazetted forest fragments and wetlands serve as important resource bases, but are also problematic for farmers. Gender, wealth, ethnicity, and distance to the park boundary can be important factors in contributing to resource use, adaptation to shor tages, problems, and means of mitigating crop raiding. This dissertation resear ch quantified the change in spatial extent and productivity of wetlands and forest fragments. I then examined how that decline in extent and quality and increased fragmentation of these natural areas has impacted households and their corresponding coping mechanisms to resource shortages and human-w ildlife conflict. In this chapter, I will also discuss the broader significance and implications for this research as well as future directions. 150
Diminishing Resources As population outside the park swells and peop le are prohibited from resource extraction inside its boundary, the demand for resources and agriculture land not only increases, but competes. As a result, the remaining unprotected natural areas wetlands and forest fragments become targets for degradation and conversion. The objective of this chapter was to establish what change is occurring in the landscape outside the park and to quantify that change. Specific research questions addressed were: How has the productivity and spatial extent of forests and wetlands in the landscape surrounding Kibale National Park changed over time? How does this change vary in communities on the east side of the park versus the west side of the park? How does this change relate to change in land under cultivation and tea? The primary conclusions from this section were: Since the formation of the park, the landscap e outside the park has become increasingly fragmented and the forests and wetlands out side the park have become increasingly isolated there are more patches within each class, mean patch size for forests and wetlands has decreased, the distance between wetl and and forest patches has increased. Land under cultivation and tea ha s increased, especia lly in the two key study areas (East and West Study Areas), while forest and wetland areas have declined. KNPs boundaries remain relatively intact wi th only minimal encroachment since 1984. There has been a substantial decrease in NDVI in forests an d wetlands both inside and outside the park since 1984, but there is no cl ear explanation fo r this decrease. 151
Resource Use and Household Livelihoods Wetlands and forest fragments outside KNP se rve as important resource bases for local people. However, they are also problematic for local farmers, since crop and livestock raids by primates, elephants, and other animals seem to emanate from these natural fragments. This chapter examined the dual character of these na tural areas within the agricultural landscape around KNP in terms of the social and environmen tal benefits and problems they represent to local households that vary in ethnicity, wealt h, and distance from the park. Specific research questions addressed were: What resources and problems are associated with wetlands and fo rest fragments around KNP? How do these benefits and probl ems vary by distance from the park boundary, gender, wealth, and ethnicity? The primary conclusions from this section were: For the majority of respondents, forests and we tlands serve as important resource bases. Most benefit in some way from wetlands and fo rest, but more claim benefits from wetlands than from forests. Benefits include firew ood, building poles, handcraft materials, water, medicines and other resources. Ot her also claim fresh air, timely rain, and soil moisture as direct benefits of wetland and forest presence. Respondent gender was significantly related to some benefits such as poles, firewood, and fish because of defined cultural gender roles, but gender is not found to be significant with most other resources. Wealth class and ethnic group did not signif icantly relate to ex tracted resources from wetlands and forest fragments. Distance related significantly to most of the benefits from the wetland and forest. Wild animals were identified as the most prolific problem in the surrounding landscape as a result of the presence of wetlands and forest s. The vervet and redtail monkeys are worst crop raiders, but baboons and ele phants are also problematic. Whether or not households ha d problems with crop raiding was not significantly related to distance. However, distance was significantly related to all of the worst crop raiders small monkeys, baboons, and el ephants. Elephants and bab oons are identi fied as the 152
worst problems by farmers cl oser to the KNP boundary, while, small monkeys are identified as the most problem atic animals by farmers furthe r from the park boundary. Wealth and gender were found less significant in relation to thes e problem animals. Ethnicity is important because it dete rmines where farmers will settle. Though the forests and wetland s are perceived to be sour ces of both problems and benefits, most landholders say they are worried abou t the availability of resources in the future and would ra ther live closer to these areas. Responses to Resource Availability Forests and wetlands provide resources, but are also sources of problems. For households that are heavily dependent on the land for its resources to sustai n their livelihoods, the decline of these natural areas has an impact on households a nd their ability to secure resources and sustain their livelihoods. Since mo st of the resources collected in th e wetlands and forest fragments are obtained locally, households must respond in some way to shortages. Some purchase fuelwood, others must find new sources or travel longer distances to co llect fuelwood and water. Landscape fragmentation and decline in forests and wetlands has increased edge and decreased not only resources, but also natural habitat. As a result, human-wildli fe conflict for local landholders is imminent. Landholders response to human-wildlife conflict and resource shortage in relation to four di fferent factors was measured: ethni city, head of household gender, wealth, and distance to the park. This chap ter quantified the loss of wetlands and forest fragments outside KNP, landscape fragmentation, and the ways households have adapted to a degraded resource bases of the unprotected wetlands and forests a nd problems with wildlife in the landscape surr ounding a park. Specific research questions addressed were: How have the presence and extent of wetla nds and forest fragmen ts in the landscape surrounding Kibale National Park chan ged over time? How have households ad apted to the declines in the wetlands and forest fragments in the landscape surrou nding Kibale National Park? 153
154 The primary conclusions from this section were: Forests and wetlands are decrea sing in patch size and those th at remain are increasingly isolated within the landscape since 1984. Most people recognized changes within the la ndscape and perceived difficulty in obtaining resources and current and future resource shortages. Distance from park was an important factor in determining perceptions of resource availability. Distance from the park was not significantly re lated to response to resource shortages. While there was no perfect solution to effectively mitigate crop raiding animals and to supply resources, we recommend engaging in a mixture of activities is necessary. The Impacts of a Forest Park Protected areas around the world are centerp ieces for the conservation movement and protection of biodiversity. The fortress conservation model has sparked many debates given its treatment of park-people relati onships. Often, those most imp acted by the park most are the local, rural poor, and most of the beneficiaries seem to be the wea lthier and foreign visitors. This chapter examined whether a forest park establ ished under the fortress co nservation model can be accepted by neighboring rural commun ities in western Uganda. Specific research ques tions addressed were: Most people say the park has helped or at l east had no effect on them. However, most say that they would rather live further than closer from the park boundary. The most prolific benefits are perceived envi ronmental benefits and containment of wild animals. Relatively few people mention tangible benefits such as employment, infrastructure, or tourism. Of those that perceive problems from the par k, most report problems with wild animals. Despite the many problems mentioned, the ove rall sentiment towards the park is not overtly hostile to parks existence Results suggest that KNP could be a muted success under the fortress conservation model. Although it is perceived to be the source for ra iding wild animals, many households benefit
from KNP. Though many costs are associated with park presence, most respondents agree the park should stay. Research Significance and Implications This research is significan t in a number of ways. First, my research adds another perspective to the often divisive and conflicti ng conservation dialogues in Uganda, East Africa and beyond. This research will fill a gap in th e conservation literature about the effects of changing resource bases on househol d livelihoods. With in conservation and development, there is no clear understanding of how ethnicity and cultural differen ces influence land -use strategies and shape the landscape (Stone, 1996). This study responds to the call in the conservation literature for linking landscape a nd household level data as they pertain to land-use patterns, natural resource use, an d livelihoods. This will contribut e to the growing body of land-use and land-cover change studies being conducted at mul tiple spatial and temporal scales (Geoghegan et al. 1998). Second, Kibale National Park and the surroundi ng landscape represent an important case study that links protected area establishment and the impacts on social-ecological systems. The larger KNP area has already been identified as a site where conflic t over resources occurs. Most work on protected areas has examined at the impact of humans on parks and the impacts of extraction and resource exploita tion (e.g. Liu et al. 2001; 2003). This study instead examined how parks are not benign within the landscape a nd have various spatial and temporal impacts on park neighbors. In a country and discourse where fines and fences conservation agendas have historically dominated, the reduc tion of resources as a result of landscape fragmentation has yet to receive sufficient attention within the conservation arena. Furthermore, KNP is an important park not only because of its biological diversity, but also because of the islandized nature of the park. KNP represents an important litmus test for the 155
impacts of park establishment on the surrounding landscape in mid-elev ation moist tropical rainforests, typical more of central Africa (e.g. Gabon, Democratic Republic of Congo). The population and resource pressures and the subseq uent islandization and th e management of the landscape are closely linked to other regions. Th is study shows how park-based conservation and management decisions, such as the protection of biodiversity, can have lasting effects on adjacent wetlands and forest fragments. The re sults of this research will establish baseline information for measuring the impact of future management practices. Third, this research introduced an innova tive methodology for measuring cross-scale linkages. This sampling framework linked both landscape and household level scales used to examine park impacts on the surrounding landscape Land cover analyses revealed spatial and temporal change in extent and pattern within the entire landscape. Containing 9ha, each superpixel represented an area large enough to link land cover change at the broad landscapelevel to a more manageable, yet an accurate representation of la ndscape change, size to link land cover changes to specific households. Supe rpixels contained a sm all enough area to link household-level decisions for land use (i.e., mechanisms for adapting to resource shortages and crop raiding) and study the impacts on land cover. In turn, since the locations of each superpixel were generated randomly, these results and demographic data can be scaled up to accurately assess park impacts and land use within the landscape. Linking householdand landscape-level data is vital to address the complex nature of human-environment interactions (Geoghegan et al 1998; Rindfuss et al. 2003), especially where access to resources can be limited, contested, and prohibited. The larger KNP area has already been identified as a site where conflict over resources occurs. This study shows how park-based conservation and management decisions, such as the protection of biodive rsity, can have lasting 156
effects on adjacent wetlands and forest fragments. The results of this study will establish baseline information for measuring the impact of future management practices. This spatial approach to understanding human-environment interactions has yet to be implemented in East Africa. Fourth, while the ecological importance of Ugandas wetl ands and forests has been recognized, the linkage between the social and ecological systems remains largely understudied (Chapman et al. 2001; MacLean et al. 2003). As a result, wetlands and ungazetted forests have been underemphasized and less defined in nation al level natural resource management (Bakema and Iyango, 2001). Not only are more legislati on, regulatory measures, infrastructure, and monitoring at the central government leve l needed, but also cooperation, education, management, and monitoring at the local leve l. In the past, co nservation planning and environmental policies have often relied on the subjective assessments of planners and development professionals, leading to the use of incomplete or biased data sets (Fuller et al. 1998). Appropriate environmental policy depends on accurate information on past and current conditions. Satellite imagery analysis can be used to pinpoint important hotspots for conservation and monitoring, extent, and rates of change over time (Haack, 1996). Satellite remote sensing not only can be used to determin e the biophysical change ch aracteristics, but it can also be used to consider human needs and anthropogenic influences. At the same time, the perceptions of ordinary men a nd women are critical in genera ting effective and appropriate environmental planning (Basset and Koli Bi, 200 0). This synergy of biophysical and human change data at multiple scales is imperative fo r the formation of practical conservation policy in Uganda (Schweik and Thomas, 2002). Such analyses may be used to better inform decision makers and land managers and conserva tion-based non-government al organizations. 157
Lastly, geography as a discipline engages with complex questions, which are often compounded by political and moral issues. As su ch, its research topics not only add to our academic understanding about people, but also pr ovide opportunities to influence the processes that affect how people meet fundamental livelihood n eeds. The practical natu re of this research, combined with its innovative spat ial framework addressed both of these dimensions. At the local level, this research can provide land management agencies with detailed information about landuse change since the establishment of KNP. It will add to the dialogue guiding future research into the needs of communities living near protec ted areas. In addition, it will contribute to a more complete understanding of the complex inte ractions between communities and a shrinking resource base related to the presence of a protected area. The future of conservation and the preservati on of biodiversity depend on the relationship between protected areas and the communities su rrounding those areas. Greater knowledge of how these communities respond to challenges po sed by protected areas will allow for improved planning and implementation of conservation go als. Uganda is a region noticeably underrepresented in the conservation and development rese arch agenda. This research is more than a case study. It is a much-needed addition to East African case studies. The results of this study may be applicable to the long-term management of national parks and forest reserves and cooperation with neighb oring communities th roughout the world. Future Research This dissertation provides the basis for future directions for my research to examine parkpeople relationships. Future research will focus on the following areas: Examine the changes in land cover extent, patte rn, and productivity. From my analysis, it is evident that productivity is decreasing both inside and outsi de the park in both forests and wetlands. However, there is no clear expl anation for this decline. Is this decease relating to changes in fruit phenolo gy, elephant density, or is this an indication of a greater shift in climate? 158
159 Southworth et al. (in review) described the difficulty in diffe rentiating elephant grass and papyrus because of similar spectral signatures. While Southworth et al.s (in review) and my analyses are conservative in their land cover assessment by including elephant grass with papyrus together in one land cover class, future resear ch should target the separation of these land covers. Although, they have sim ilar growth and spectral characteristics, they are different in their surface temperatures (H artter, unpublished data). In addition, papyrus is found in low lying areas and valley bottoms, while elephant grass generally, but not exclusively, is found in upland areas. By combining surface temperature analyses and elevation data, a new rule-based cl assification will be constructed. New vegetation indices will be developed to bett er describe within-class variation and to assess productivity change w ithin the important natural areas. Southworth (2004) and Boyd et al. (1996) note the usefulness of us ing the Landsat thermal band for assessing forest regeneration in tropical dry forests. The thermal band can be incorporated in a new analysis to examine within-cla ss variation of forest to in vestigate loss of productivity within park vs. surr ounding landscape. The superpixel methodology has been shown to be useful in dense agriculturallydependent settlements outside a fo rested park. It is also important to compare these results across other forested protected area lands capes surrounding by intensive smallholder agriculture. This methodology will be tested in the future in protected area landscapes with: limited rainfall (e.g. savannah region of southern Africa), larger landholdings, dispersed settlements, and pastoral communities. Much of the community-based conservation literature (e.g. Hulme and Murphree, 2001) purports that parks established under the fort ress conservation model cannot be accepted by neighboring communities. KNP, arguably, is an example of a muted success for fortress conservation. Additional research wi ll address the long-term stability of this perception in communities neighboring KNP. In addition, sample size will be increased and the study area will be widened to incl ude areas outside the Bakiga and Batoro communities studied.
LIST OF REFERENCES Adams, W.M. & Hulme D. (2001) If community c onservation is the answer in Africa, what is the question? Oryx 35 (3): 193-200. Adams, W. & Hulme, D. (2001) Conservation & community: changing na rratives, policies & practices in African conservation. In: African Wildlife and Livelihoods: The Promise and Performance of Community Conservation Hulme, D.& Murphree, M. eds., pp. 9-23. Oxford, UK: James Curry. Adams, W. & Infield, M. (2003) Who is on the gori lla's payroll? Claims on tourist revenue from a Ugandan national park. World Development 31 (1): 177-190. Agresti, A. (1996) An Introduction to Categoric al Data Analysis. New York City, New York, USA: John Wiley & Sons, Inc. Aluma, J., Drennon, C., Kigula, J., Lawry, S., Muwanga-Zake, and J. Were. 1989. Settlement in Forest Reserves, Game Reserves and Nationa l Parks in Uganda. LTC Research Paper 98. Land Tenure Center. Madison, Wisconsi n, USA: University of Wisconsin. Anyamba, A., Tucker, C.J. & Eastman, J.R. (2001) NDVI anomaly patterns over Africa during the 1997/98 ENSO warm event. International Journal of Remote Sensing 22(10): 18471859. Archabald, K. & Naughton-Treves, L. (2001) Tourism reve nue sharing around national parks in western Uganda: early efforts to identify and reward local communities. Environmental Conservation 23: 135-149. Ashley, C. (2000) Applying livelihood approaches to natural resource management initiatives: experiences in Namibia and Kenya. Working Paper 134. London, UK: Overseas Development Institute. London. Ashley, C., Boyd, C., & Goodwin, H. (2000) Pro-poor tourism: putting poverty at the heart of the tourism agenda. Natural Resource Pers pectives 51. London, UK: Department for International Development. Bakema, R.J. & Iyango, L (2000). Engaging local user s in the management of wetland resources: The case of the National Wetlands Programme Uganda. Nairobi, Kenya: IUCN Eastern Africa Regional Office. Balmford, A., Bruner, A., Cooper, P., Costanza, R ., Farber, S., Green, R.E., Jenkins, M. Jeffries, P., Jessamy, V., Madden, J., Munro, K., Myers, N., Naeem, S., Paavola, J., Rayment, M., Rosendo, S., Roughgarden, J., Trumper, K., & Turner, R.K. (2002) Ecology economic reasons for conserving wild nature. Science 297 : 950-953. 160
Banana, A.Y. & Gombya-Ssembajjwe, W. (1998) Successful forest management: the importance of security of tenure and rule enforcement in Ugandan forests. In: Forest resources and institutions. Gibson, C., McKean, M. & Ostrom, E., eds. Working Paper 3. Rome, Italy: Food and Agricultural Organiza tion of the United Nations. Baranga, D. (2004) Forest fragmentation and primat es survival status in non-reserved forests of the Kampala area, Uganda. African Journal of Ecology 42(1): 70-77. Barbier, E.B. (1993) Sustainable use of wetlands valuing tropical we tland benefits: Economic methodologies and applications. The Geographical Journal 159 (1): 22-32. Basset, T.J. & Koli Bi, Z. (2000) Environm ental discourses and the Ivorian savanna. Annals of the Association of American Geographers 90(1):67-95. Batagoda, B.M.S., Turner, R.K., Tinch, R., & Brown, K. (2000) Towards policy relevant Ecosystem services and natura l capital values: Rainforest non-timber products. Global Environmental Change Working Paper 200006. London, UK: Centre for Social and Economic Research into the Global Environm ent. University of East Anglia and University College London. Bedunah, D. & Schmidt, S. (2004) Pastoralism and protected area management in Mongolias Gobi Gurvansaikhan National Park. Development and Change 35(1): 167-191. Berkes, F. (2004) Rethinking co mmunity-based conservation. Conservation Biology 18 (3): 621630. Bolwig, S., Pomeroy, D., Tushabe, H., & Mush abe, D. (2006) Crops, trees, and birds: biodiversity change under agricultural intens ification in Ugandas farmed landscapes. Danish Journal of Geography 106(2): 115-130. Boyd, D.S., Foody, G.M., Curran, P.J., Lucas, R.M., & Honzak, M. (1996) An assessment of radiance in Landsat TM middle and thermal infrared wavebands for the detection of tropical forest regeneration. International Journal of Remote Sensing 17(2): 249-261. Brockington D (2002) Fortress Conservation: The Preservation of the Mkomazi Game Reserve, Tanzania. Oxford, UK: International African Institute. James Curry, Mkuki Na Nyota, Indiana University Press. Brockington, D. (2004) Community conservation, in equality and injustice: myths of power in protected area management. Conservation & Society 2(2):411-432 Brockington, D., Igoe, J., & Schmidt-Soltau, K. (2006) Conservation, human rights, and poverty reduction. Conservation Biology 20(1): 150-252. Brown, K. (2002) Innovations for conservations and development. The Geographical Journal 168(1): 6-17. 161
Byron, N. & Arnold, M. (1999) What future s the people of the tropical forests? World Development 27(5): 789-805. Chambers, R. & Conway, G. (1991) Sustainable Rural Livelihoods: Practical Concepts for the 21st Century. Discussion Paper 296. Brighton, UK: Institute of Development Studies. Chapman, C.A. & Chapman, L.J. (1996). Exotic tr ee plantation and the regeneration of natural forests in Kibale National Park, Uganda. Biological Conservation 76: 253-257. Chapman, C.A. & Chapman, L.J. (1999) Forest re storation in abandoned agricultural land: a case study from East Africa. Conservation Biology 13 (6): 1301-1311. Chapman, C.A., Chapman, L.J., Kaufman, L., & Zanne, A.E. (1999) Potentia l causes of arrested succession in Kibale National Park, Uganda : Growth and mortality of seedlings. African Journal of Ecology 37: 81-92. Chapman, C.A. & Lambert, J. (2000) Habitat alteration and the cons ervation of African primates: a case study of Ki bale National Park, Uganda. American Journal of Primatology 50: 169-186. Chapman, C.A. & Peres, C.A. (2001) Primate c onservation in the new millennium: the role of scientists. Evolutionary Anthropology 10: 16-33. Chapman, C.A., Chapman, L.J., Vullnec, K., & Lawes, M.J. (2003) Fragmentation and alteration of seed dispersal processes: an initial evaluation of dung bee tles, seed fate, and seedling diversity. Biotropica 35 (3): 382-393. Chapman, C.A., Lawes, M.J., Naughton-Treves, L., & Gillespie, T. (2003) Primate survival in community-owned forest fragments: are meta population models useful amidst intensive use? In: Primates in Fragments: Ecology and Conservation ed. Marsh, L.K., pp. 63-78. New York City, NY: Kluwer Academic/Plenum Publishers. Chapman, C.A., Chapman, L.J., Zanne, A.E., P oulsen, J.R., & Clark, C.J. (2005) A 12-year phenological record of fruiting: implications for frugivore populations and indicators of climate change. In: Floristics, Phenology and Frugivore Communities: A Pan Tropical Comparison eds. Dew J.L. & Boubli, J.P., pp. 75-92. Dorddrecht, The Netherlands: Springer. Chapman, C.A., Struhsaker, T.T., & Lambert, J.E. (2005) Thirty years of research in Kibale National Park, Uganda, reveals a complex picture for conservation. International Journal of Primatology 26(3): 539-555. Chapman, C.A., Wasserman, M.D., Gillespie, T.R., Speirs, M.L., Lawes, M.J., Saj, T.L., & Zieglers, T.E. (2006) Do food availability, pa rasitism, and stress have synergistic effects on red colobus populations living in forest fragments?. American Journal of Physical Anthropology 131: 525-534. 162
Chapman, L.J., Balirwa, J., Buge nyi, F.W.B., Chapman, C., & Cris man, T.L. (2001) Wetlands of East Africa: biodiversit y, exploitation, and po licy perspectives. In: Biodiversity in wetlands: Assessment, Function and Conservation. Vol. 2. Gopal, B., Junk, W.J., & Davis, J.A., eds. Pp. 101-131. Leiden, The Netherlands: Backhuys Publishers. Chapman, L.J., Chapman, C.A., Crisman, T.L., & Kaufman, L.S. (2003) The conservation and management of African inland waters. In Conservation, ecology, and management of African fresh waters Crisman, T.L., Chapman, L.J., Chapman, C.A., & Kaufman, L.S., eds. pp. 476-488. Gainesville, Florida, US A: University Press of Florida.. Chatelain, C., Gautier, L, & Spichiger, R. (1996) A recent history of forest fragmentation in southwestern Ivory Coast. Biodiversity Conservation 5: 37-53. Chege, F., Onyango, G., Drazu, C., Mwandha, S. (2002) Kibale and Semuliki Conservation and Development Project. End-Of-Phase II/End-of -Project Evaluation Re port. Dar es Salaam, Tanzania, World Conservation Unio n. East Africa Regional Office. Chhetri, P., Mugisha, A., & White, S. (2003). Co mmunity resources use in Kibale and Mt. Elgon National Parks, Uganda. Parks 13(1): 28-49. Child, G. (2002) An historical perspective of wildlife conservation in Southern Africa. Symposium on Wildlife and Land Us e. Paarl, South Africa: Southern African Sustainable Use Specialist Group. Chiyo, P.I. (2000) Elephant ecology and crop depr edation in Kibale National Park, Uganda. MSc Thesis. Makerere University. Kampala, Uganda. Chiyo, P.I., Cochrane, EP., Naughton, L., & Basuta G.I. (2005) Temporal patterns of crop raiding by elephants: a response to changes in forage quality or crop availability? African Journal of Ecology 43: 48-55. Collinge, S.K. (1996) Ecological consequences of habitat fragmentation: implications for landscape architecture and planning. Landscape and Urban Planning 36 (1): 59-77. Cooper, J.P., ed (1975). Photosynthesis and Productivity in Different Environments London, UK: Cambridge University Press. Coops, N.C., White, J.D., & Scott, N.A. (2004) Estimating fragmentati on effects on simulated forest net primary productivity de rived from satellite imagery. International Journal of Remote Sensing 25(4): 819-838. Costanza, R., dArge, R., de Groot, R., Farber S., Grasse, M., Hannon, B., Naeem, S., Limburg, K., Paruelo, J., ONeill, R.V., Raskin, R., Su tton, P., & van den Felt, M. (1997) The value of the worlds ecosystems and natural capital. Nature 387: 253-260. Crisman, T. L., Chapman, L.J., & Chapman, C. A. (1996) Conserving tropical wetlands through sustainable use. Geotimes 41: 23-25. 163
164 Crisman, T.L., Chapman, L.J., Chapma n, C.A., & Kaufman, L.S. (2003) Conservation, Ecology and Management of African Fresh Waters. Gainesville, Florida, USA: University Press of Florida. Davenport, M. L. & Nicholson, S. E. (1993) On the relation between rainfall and the normalised difference vegetation index for diverse vegetation types in East Africa. International Journal of Remote Sensing 14: 2369-2389. de Boer, W. & Baquete, D. (1998) Natural res ource use, crop damage and attitudes of rural people in the vicinity of the Maputo Elephant Reserve, Mozambique. Environmental Conservation 25(3): 208-218. De la Cruz, A.A. (1986) Tropical wetlands as a carbon source. Aquatic Botany 25: 109-115. Dixon, J.A., & Sherman, P.B. (1991) Economics of protected areas. Ambio 20(2): 68-74. Duncan, R.S. & Chapman, C.A. (2002) Limitations of animal seed dispersal for enhancing forest succession on degraded lands. In: Seed Dispersal and Frugivory : Ecology, Evolution and Conservation. Levey, D.J., Silva, W.R., & Galetti, M., eds. Pp. 437-450. Oxon, UK: CABI Publishing. Edmunds, D. (1997) Continuity and change in the resource management institutions of communities bordering the Kibale forest Park, Uganda. PhD Dissertation. Clark University, Worcester, Massachusetts, USA. Eklundh, L. (1998) Estimating relations between AVHRR NDVI and rainfall in East Africa at 10-day and monthly time scales. International Journal of Remote Sensing 19: 563-568. Evans, T.P. & Moran, E.F. (2002) Spatial integrat ion of social and biophysical factors related to landcover change. ACT Publication N o. 02-01. A supplement to vol. 28, 2002. Population and Development Review : 165-186. Fairhead, J. & Leach, M. (1996) Misreading the African Landscape. Society and Ecology in a Forest-Savanna Mosaic. Melbourne, Australia: Cambridge University Press. Ferraro, P.J. & Kiss, A. (2002) Direct payments to conserve biodiversity. Science 298 : 17181719. Fiallo, E.A., & Jacobson, S.K. (1995) Local comm unities and protected areas : attitudes of rural residents towards conservation and Machalilla National Park, Ecuador. Environmental Conservation 22(3): 241-249. Fischer, G., Heilig, G.K., Young, A., Vlek, P., & Tinker, B. (1997) Population momentum and the demand on land and water resources and discussion. Philosophical Transactions: Biological Sciences 352(1356): 869-889. Food and Agriculture Organization of the Un ited Nations 2006. Global Forest Resources Assessment 2005. Rome, Italy.
Forman, R.T.T. (1995) Land Mosaics. The Ecology of Landscapes and Regions. Cambridge, UK: Cambridge University Press. Fuller, R.M., Groom, G.B., Mugisha, S., Ipul et, P., Pomeroy, D., Katende, A., Bailey, R., & Ogutu-Ohwayo, R. (1998) The integration of field survey and remote sensing for biodiversity assessment: A case study in the tropical forests and wetlands of Sango Bay, Uganda. Biological Conservation 86: 379-391. Gadd, M.E. (2005) Conservation outside of parks: Attitudes of local peopl e in Laikipia, Kenya. Environmental Conservation 32(1):50-63. Gascon, C., Williamson, G.B., & da Fonseca, G.A.B. (2000) Receding forest edges and vanishing reserves. Science 288: 1356-1358. Geist H.J., & Lambin, E. (2001) What drives tropical deforestation? A meta-analysis of proximate causes of deforestation based on subnational case study ev idence. LUCC Report Series No. 4. http://www.geo.ucl.ac.be/LUCC/lucc.html Geoghegan, J., Pritchard Jr., L., Ogneva-Himmelb erger, Y., Roy, Chowdhury, R., Sanderson, S., & Turner II, B.L. (1998) Socializing the pixe l and pixelizing the so cial in land-use and land-cover change. In: People and Pixels: Linking Remote Sensing and Social Science Liverman, D., Moran, E., Rindfuss, R., & St ern, P., eds. pp. 51-69. Washington DC, USA: National Academy of Science Press. Gillespie, T.R., Chapman, C.A. (2006) Prediction of parasite infection dynamics in primate metapopulations based on attr ibutes of forest fragment ation. Conservation Biology 20 (2): 441-448. Gillingham, S. & Lee, P. (1999) The impact of wildlife-related benefits on the conservation attitudes of local people around the Selous Game Reserve, Tanzania. Environmental Conservatio n 26(3): 218-228. Goebel, A., Campbell, B., Mukamuri, B, & Veema n, M. (2000) People, value, and woodlands: A field report of emergent themes in in terdisciplinary research in Zimbabwe. Agriculture and Human Values 17: 385-396. Goldman, A. (1996) Pest and disease hazards and sustainability in African Agriculture. Experimental Agriculture 32: 199-211. Goodchild, M.F., Anselin, L., Appelbaum, R.P ., & Hawthorn, B. (2000) Toward spatially integrated social science. International Regi onal Science Review 23(2): 139-159. Haack, B. (1996) Monitoring wetland changes with remote sensing: an East African example. Environmental Management 20: 411-419. Hamilton, A. (1984) Deforestation in Ugand a. Nairobi, Kenya: Oxford University Press. 165
Hartter, J. (In review) Social and ecological aspects of landscape change: forest fragments around Kibale National Park, Uganda. In: Linking Social and Ecologi cal Systems: Theory, Applications and Future Directions Southworth, J., Perz, S., Chowdhury, R.R., & Nagendra, H., eds. Hayes, T.M. (2006) Parks, people, and forest protection: an institutional assessment of the effectiveness of protected areas. World Development 34 (12) : 2064-2075. Hill, C. (1997) Crop-raiding by wild vertebrates: the farmers perspective in an agricultural community in western Uganda. International Journal of Pest Management 43(1): 77-84. Hill, C.M. (2004) Farmers' perspectives of conflict at the wildlife-agriculture boundary: some lessons learned from African subsistence farmers. Human Dimensions of Wildlife 9(4): 279-286. Hill, J., & Curran, P. (2003) Area, shape and isolation of tropical forest fragments: effects on tree species diversity and imp lications for conservation. Journal of Biogeography 30(9): 13911403. Hill, J.L., Curran, P.J. (2005) Fragment shape and tree species composition in tropical forests: A Landscape Level Investigation. African Journal of Ecology 43 : 35-43. Howard, P. Davenport, T. Kige nyi, F., Viskanic, P., Baltzer, M., Dickinson, C., Lwanga, J., Matthews, R., & Mupada, E. (2000) Protecte d area planning in the tropics: Ugandas national system of fore st nature reserves. Conservation Biology 14(3): 858-875. Huising, E.J. (2000) Wetland monitoring in Uganda. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences XXXIV, Part 6/W6: 127-134. Hulme, D. & Murphree, M. ( 1999) Communities, wild life and new conservation in Africa. Journal of International Development 11:277-285. Hulme, D. & Murphree, M., eds (2001) African Wildlife and Li velihoods: The Promise and Performance of Community Conservation Oxford, UK: James Curry. Hulme, D., Murphree, M. (2001) Community c onservation in Africa: an introduction. In: African Wildlife and Livelihoods: Th e Promise and Performance of Community Conservation Hulme. Hulme, D. & Murphree, M., eds. pp. 1-8. Oxford, UK: James Curry. Hyden, G., ed (1998) Governance issues in c onservation and development. CDF Discussion Paper. Gainesville, Florida, USA: Conservation & Development Forum. Infield, M. & Namara, A. (2001) Community attitu des and behaviour towards conservation: an assessment of a community conservation pr ogramme around Lake Mburo National Park, Uganda. Oryx 25: 48-60. 166
James, A., Gaston, K.J., & Balmford, A. (2001) Can we afford to conserve biodiversity? BioScience 51(1):43-52. Jensen, J.R. (2000) Introductory Digital Image Processing Upper Saddle River, New Jersey, USA: Prentice Hall. Jones, B., Murphree, M. (2001) The Evolution of policy on community conservation in Namibia & Zimbabwe. In: African Wildlife and Livelihoods: The Promise and Performance of Community Conservation. Hulme. Hulme, D. & Murphree, M., eds. pp. 160-176 Oxford, UK: James Curry. Jones, M.B. & Humphries, S.W. (2002) Impacts of the C4 sedge Cyperus papyrus L. on carbon and water fluxes in an African wetland. Hydrobiologia 488(1-3): 107-113. Kagoro-Rugunda, G. (2004) Crop raiding ar ound Lake Mburo National Park, Uganda. African Journal of Ecology 42: 32-41. Kasenene, J.M., & Murphy, P.G. (1991) Post-logging tree mortality and major branch losses in Kibale Forest, Uganda. Forest Ecology and Management 46: 295-307. Kassilly, F.N. (2002) The fence as a moderator of the wildlife menace in Kenya. African Journal of Ecology 40 (4): 407-409. Kayanja, F.I.B. & Byarugaba, D. (2001) Disa ppearing forests of Uganda: The way forward. Current Science 81(8): 936-947. Kerr J.T. & Ostrovsky, M. (2003) From space to species: ecological applications for remote sensing. TRENDS in Ecology and Evolution 18(6): 299-305. Kisamba-Mugerwa, W. & Nuwagaba, A. (1993) Comparative and multidimensional analysis of communal and private property resources (land) tenure and rural development in Uganda. Kampala, Uganda: Makerere Institute of Social Research. Lambin, E.F., & Ehrlich, D. (1996) The surf ace temperature-vegetation index space for land cover and land-cover change analysis. International Journal of Remote Sensing 17(3): 463487. Lambin, E.F. & Ehrlich, D. (1997) Land-cove r changes in Sub-Saharan Africa (1982-1991): Application of a change index based on re motely sensed surface temperature and vegetation indices at a continental scale. Remote Sensing of Environment 61(2), 181-200. Lambin, E.F.& Geist, H.J. 2001. Global Land-Us e and Land-Cover Change: What Have We Learned So Far?. Global Change Newsletter N 46. LUCC Publications. 167
Lambin, E.F., Turner, B.L., Geist, H.J, Agbola, S.B., Angelsen, A., Bruce, J.W., Coomes, O.T., Dirzo, R., Fischer, G., Folke, C., George, P.S., Homewood, K., Imbernon, J., Leemans, R., Li, X., Moran, E.F., Mortimore, M., Ramakris hnan, P.S., Richards, J.F., Sknes, Steffen, W., Stone, G.D., Svedin, U. Veldkamp, T.A., Vogel, C., & Xu, J. (2001) The causes of land-use and land-cover change: moving beyond the myths. Global Environmental Change 11: 261-269. Laurance W.F. & Bierregaard Jr., R.O., eds (1997) Tropical Forest Remnants Ecology, Management, and Conservation of Fragmented Communities. Chicago, Illinois, USA: The University of Chicago Press. Laurance, W.F., Ferreira, L.V., Rankin-De Mer ona, J.M., Laurance, S.G., Hutchings, R.W., & Lovejoy, T.E. (1998) Effects of forest fragme ntation on recruitment patterns in Amazonian tree communities. Conservation Biology 12(2): 460-464. Laurance, W.F., Delamnica, Laurance, S.G., Vasconcelos, H.L., & Lovejoy, T.E. (2000) Rainforest fragmentation kills big trees. Nature 404 : 836. Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L ., Bruna, E.M., Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G ., & Sampaio, E. (2002) Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology 16(3): 605618. Laurance, W.F., 2004. Forestclimate interacti ons in fragmented tropical landscapes. Phil. Trans. R. Soc. Lond. B: 359: 345-352. Lepp, A.P. (2004) Tourism in a rural Ugandan village: impacts, local meaning and implications for development. PhD Dissertation. University of Florida, Gainesville, Florida, USA. Lepp, A. & Holland, S. (2006) A comparison of at titudes toward state-led conservation and community-based conservation in the village of Bigodi, Uganda. Society and Natural Resources 19 : 609. Linkie, M., Dinata, Y., Nofrianto, A., & Leade r-Williams, N. (2007) Patterns and perceptions of wildlife crop raiding in and around Keri nci Seblat National Park, Sumatra. Animal Conservation 10(1): 127-135. Liu, J., Linderman, M., Ouyan, Z., An, L., Yang, J., & Zhang, H. (2001) Ecological Degradation in protected areas: the ca se of Wolong Nature Reserve for Giant Pandas. Science 292 : 98101. Liu, J., Daily, G.C., Ehrlich, P.R., & Luck, G.W. (2003) Effects of household dynamics on resource consumption and biodiversity. Nature 421: 530-533. Lu, D., Mausel, P., Brondizio, E., & Moran, E. (2004) Change detection techniques. International Journal of Remote Sensing 25(12): 2365-2407. 168
MacLean, I., Tinch, R., Hassall, & Boar, R. (2003) Social and economic use of wetland resources: a case study from La ke Bunyonyi, Uganda. Norwich, UK: Centre for Social and Economic Research on the Global Environment. Marcus, R.R. (2001) Seeing the forest for the tr ees: integrated conser vation and development projects and local perceptions of conservation in Madagascar. Human Ecology 29(4): 381397 Marsh, L.K., C.A. Chapman, M. Norconk, J. Wallis and G. Umapathy, J.C. Bicca-Marques, K. Gilbert, S. Ferarri, & Scott, S. (2003) Fragment ation: specter of the fu ture of the spirit of conservation? In: Primates in Fragments: Ecology and Conservation Marsh, L.K., ed pp 381-398. New York City, New York: USA: Kluwer Academic/Plenum Publishers. McGarigal, K. & Marks, B.J. (1995) FRAGSTATS: spatial pa ttern analysis program for quantifying landscape structure. Gen. Tec h. Rep. PNW-GTR-351. Portland, Oregon, USA: U.S. Department of Agriculture, Forest Se rvice, Pacific Northwes t Research Station. McNeely, J.A. (1990) The future of national parks. Environment :16-41. McNeely, J.A. (1994) Protected areas for the 21st century: working to provide benefits to society. Biodiversity and Conservation 3: 390-405. Ministry of Finance and Economic Planning (1991) Population and housing census report. Kampala, Uganda: Government of Uganda. Kampala. Mkanda F.X. & Munthali, S.M. (1994) Pub lic attitudes and need s around Kasungu National Park, Malawi. Biodiversity and Conservation 3: 29-44. Mugisha, A. (2002) Evaluation of community-bas ed conservation approaches: management of protected areas in Uganda. PhD Dissertation. Univ ersity of Florida, Gainesville, Florida, USA. Mugisha, S. (2002) Patterns and root causes of land cover/use cha nge in Uganda: An account of the past 100 years. Land Use Change Imp acts & Dynamics (LUCID) Working Paper Series No. 14. Kampala, Uganda: Makerere University Institute of Environment and Natural Resources, International Livestock Research Institute, and United Nations Environment Programme. Mukiibi, J.K., ed. (2001a) Agriculture in Uganda. General Information.. Vol. 1. National Agricultural Research Organisation. Kampala, Uganda: Fountain Publishers Ltd. Mukiibi, J.K., ed (2001b) Agriculture in Uganda. Forestry Vol. 3. National Agricultural Research Organisation. Kampala, Ug anda: Fountain Publishers Ltd. Mulley, B.G. & Unruh, J.D. (2004) The role of off-farm employment in tropical forest conservation: labor, migration, and small holde r attitudes toward la nd in western Uganda. Journal of Environmental Management 71: 193-205. 169
Murphree, M. (2002) Protected areas and the commons. The Common Property Resource Digest 60: 1-4. Muthuri, F., Jones, M., & Imbamba, S. (1989) Primary productivity of papyrus ( Cyperus papyrus ) in a tropical swamp; Lake Naivasha, Kenya. Biomass 18: 1-14. Nagendra, H.J., Southworth, J., & Tucker, C. (20 03) Accessibility as a determinant of landscape transformation in western Honduras : linking pattern and process. Landscape Ecology 18(2): 141-158. Nagendra, H., Pareeth, S., & Ghate, R. (2006) People within parks forest villages, land-cover change and landscape fragmentation in th e Tadoba Andhari Tige r Reserve, India. Applied Geography 26: 96-112. National Environment Management Authority (1997) State of environmen t report for Uganda 1996. Kampala, Uganda. National Environment Management Authority ( 1999) State of Environment Report for Uganda 1998. Kampala, Uganda. National Environment Management Authority ( 2001) State of Environment Report for Uganda 2000/2001. Kampala, Uganda. Naughton, L., Rose, and R., & Treves, A. (1999) The social dimensions of human-elephant conflict in Africa: a li terature review and case studies from Uganda and Cameroon. Gland, Switzerland: IUCN. Naughton-Treves, L. (1996) Un easy neighbors: wildlife and fa rmers around Kibale National Park, Uganda. PhD Dissertation. University of Florida, Gainesville, Florida, USA. Naughton-Treves, L. (1997) Farming the fore st edge: vulnerable pl aces and people around Kibale National Park, Uganda. Geographical Review 87(1): 27-46. Naughton-Treves, L. (1998) Predicting pattern s of crop damage by wildlife around Kibale National Park, Uganda. Conservation Biology 12 (1): 156-168. Naughton-Treves, L., Treves, A., Chapman, C., & Wrangham, R. (1998) Temporal patterns of crop-raiding by primates: linking food availabi lity in croplands and adjacent forest. Journal of Applied Ecology 35: 596-606. Naughton-Treves, L. & Salafsky, N. (2004) Wildlife conservation in agroforestry buffer zones: opportunities and conflict. In: Agroforestry and Biodiversity Conservation in Tropical Landscapes Schroth, G., da Fonseca, A.B., Harvey, C.A., Gascon, C., Vasconcelos, H.L., Izac, A.-M.N., eds. pp. 319-345. Washington DC, USA: Island Press. 170
Naughton-Treves, L. & Treves, A. (2005) Socioecological factors shap ing local support for wildlife: crop-raiding by elephants and other wildlife in Africa. In: People and Wildlife: Conflict or Coexistence? Woodroffe, R., Thirgood, S., & Rabinowitz A. eds. pp. 253277. Cambridge, UK: Cambridge University Press. Cambridge. Naughton-Treves, L., Kammen, D.M., & Chapman, C. (2007) Burning biodiv ersity: woody biomass use by commercial and subsistence groups in western Ugandas forests. Biological Conservation 134 : 232-241. Neumann, R.P. (1998) Imposing wilderness. Struggl es over livelihood and na ture preservation in Africa. University of California Press, London. Newmark, W.D. & Leonard, N.L. (1991) The attitu des of local people toward Kilimanjaro National Park and Forest Reserve. In: Ne wmark WD (ed). The conservation of Mount Kilimanjaro IUCN, Gland. Newmark, W.D., Leonard, N.L., Sariko, H.I., & Gamassa, D.-G.M. (1993) Conservation attitudes of local people living adjacent to five protected areas in Tanzania. Biological Conservation 63: 177-183. Newmark, W., Manyanza, D., Gamassa, D.-G., & Sariko, H. (1994) The conflict between wildlife and local people living adjacent to prot ected areas in Tanzania: human density as a predictor. Conservation Biology 8(1): 249-255. Oates, J.F. (1999) Myth and Reality in the Rain Forest: Ho w Conservation Strategies are Failing in West Africa Berkeley, California, USA: Univ ersity of California Press. Oates, J.F. (2002) West Africa: tr opical forest parks on the brink. In: Making Parks Work: Strategies for Preserving Tropical Nature Terborgh J, Van Schaik C, Davenport L, Rao M., eds. pp. 57-75. Washington DC, USA: Island Press. Obua, J. & Harding, D.M. (1996) Visitor characte ristics and attitudes Towards Kibale National Park, Uganda. Tourism Management 17(7): 495-505. Ofcansky, T.P. (2002) Paradise Lost: A History of Ga me Preservation in East Africa Morgantown, West Virginia, USA: West Virginia University Press. Onderdonk, D.A. & Chapman, C.A. (2000) Coping with forest fragmentation: the primates of Kibale National Park, Uganda. International Journal of Primatology 21 (4): 587-611. Osborn, F.V. & Parker, G.E. (2002) Community-bas ed methods to reduce crop loss to elephants: experiments in the communal lands of Zimbabwe. Pachyderm 33: 32-38. Paul, J.R., Randle, A.M., Chapman, C.A., & Chapman, L.J. (2004) Arrested succession in logging gaps: Is tree seedling growth and survival limiting? African Journal of Ecology 42: 245-251. 171
Pickett, S.T.A., Burch Jr., W.R., Dalton, S.E., Foresman, T.W., Grove, J.M., & Rowntree, R. (1997) A conceptual framework for the st udy of human ecosystems in urban areas. Urban Ecosystems 1: 185-199. Ranta, P., Blom, T., Niemelae, Joensuu, E., & Si itonen, M. (1998) The frag mented Atlantic rain forest of Brazil: size, shape and distribution of forest fragments. Biodiversity and Conservation 7: 385-403. Rey-Benayas, J.M. & Pope, K.O. (1995) Lands cape ecology and divers ity patterns in the seasonal tropics from Landsat TM imagery. Ecological Applications 5(2): 386-394. Richard, Y. & Poccard, I. (1998) A statistical stud y of NDVI sensitivity to seasonal and interannual rainfall variations in southern Africa. International Journal of Remote Sensing 19: 2907-2920. Richardson, B.J. (1993) Environmen tal management in Uganda: the importance of property law and local government in wetlands conservation. Journal of African Law 37(2): 109-143. Rindfuss, R.R., Walsh, S.J., Turner II, B.L., F ox, J., & Mishra, V. (2004) Developing a science of land change: challenges and methodological issues. PNAS 101(39): 13976-13981. Ringrose, S., Vanderpost, C., & Matheson, W. (1996) The use of integrated remotely sensed and GIS data to determine causes of vegeta tion cover change in southern Botswana. Applied Geography 16(3): 225-242. Robinson, J. & Ginsberg, J. (2004) Parks, people, and pipelines. Conservation Biology 18 (3) : 607-608. Rocheleau, D. & Edmunds, D. (1997) Women, men and trees: Gender, power and property in forest and agrarian landscapes. World Development 25(8): 1351-1371. Roughgarden, J, Running, S.W., & Matson, P.A. (1991) What does remote sensing do for ecology? Ecology 72(6): 1918-1922. Runte, A. (1997) National Parks. The American Experience 3rd Ed. Lincoln, Nebraska, USA: University of Nebraska Press. Saito F (2003) Decentralization and Development Pa rtnerships: Lessons from Uganda. Tokyo, Japan: Springer. Salafsky, N. & Wollenberg, E. (2000) Linking livelihoods and conservation: a conceptual framework and scale for assessing the inte gration of human needs and biodiversity. World Developmen t 28(8): 1421-1438. Salafsky, N., Margoluis, R., Re dford, K.H., & Robinson, J.G. (2002) Improving the practice of conservation: a conceptual framework and research agenda for conservation science. Conservation Biology 16 (6):1469-1479. 172
Schelhas, J., & Greenberg, D., eds (1996). Forest Patches in Tropical Landscapes Washington, DC, USA: Island Press. Scherr S.J. (2000) A downward spiral? Research evidence on the relationship between poverty and natural resource degradation. Food Policy 25: 479-498. Schuur, E.A.G. (2003) Productivity a nd global climate revisited: the sensitivity of tropical Forest growth to precipitation. Ecology 84(5): 1165-1170. Serneels, S., Said, M.Y., & Lambin, E.F. ( 2001) Land cover changes around a major East African wildlife reserve: The Mara ecosystem (Kenya). International Journal of Remote Sensing 22(17): 3397-3420. Shiva, S., & Bandyopadhyay, J. (1983) Eucal yptus a disastrous tree for India. Ecologist 13(5): 184-187. Southworth, J. (2004) An assessment of Landsat TM band 6 thermal data for analysing land cover in tropical dry forest regions. International Journal of Remote Sensing 25(4): 68706. Southworth, J., Munroe, D., & Nagendra, H. (2004) Land cover change and landscape fragmentation comparing the ut ility of continuous and disc rete analyses for a western Honduras Region. Agriculture, Ecosystems and Environment 101: 185-205. Southworth, J., Binford, M.W., Hartter, J., Go ldman, A., Chapman, C.A., Chapman, L.J., Omeja, P., & Binford, E. (In review) Landscape fr agmentation in and around Kibale National Park, Uganda The footprint of the agricultural landscape. Applied Geography. Stone, G.D. (1996) Settlement Ecology: The Social and Spatial Organization of Kofyar Agriculture. Tucson, Arizona, USA: The University of Arizona Press. Stouffer, P. & Bierregaard, R. (1995) Use of Amazonian forest fragments by understory insectivorous birds. Ecology 76: 2429-2445. Struhsaker, T.T. (1997) Ecology of an African Rain Forest: Logging in Kibale and the Conflict Between Conservation and Exploitation Gainesville, Florida, USA: University Press of Florida. Struhsaker, T.T. (2002) Strategies for Conserving forest nati onal parks in Africa with a case study from Uganda. In: Making Parks Work: Strategies for Preserving Tropical Nature Terborgh J, Van Schaik C, Davenport L, Rao M., eds. pp. 97-111. Washington DC, USA: Island Press. Struhsaker, T.T. (2005) Conservation of red colobus and their habitats. International Journal of Primatology 26(3): 525-538. Struhsaker, T.T., Struhsaker, P. J., & Siex, K.S. (2005) Conserving Africas rain forests: problems in protected areas and possible solutions. Biological Conservation 123:45-54 173
Tchamba, M.N. (1996) History a nd present status of the human/e lephant conflict in the WazaLogone Region, Cameroon, West Africa. Biological Conservation 75: 35-41. Terborgh, J. (1999) Requiem for nature. Washington DC, USA: Island Press. Terborgh, J. & Van Schaik, C. (2002) Why the world needs parks. In: Making Parks Work: Strategies for Preserving Tropical Nature Terborgh J, Van Schaik C, Davenport L, Rao M., eds. pp. 3-14. Washington DC, USA: Island Press. Tinker, P.B. (1997) The environmental implicat ions of intensified land use in developing countries. Phil. Trans. R. Soc. Lond B 352: 1023-1033. Turner, I. & Corlett, R. (1996) The conservation value of small, isolated fragments of lowland tropical rain forest. Trends in Ecology & Evolution 11(8): 330-333. Turner, I.M. (1996) Species loss in fragments of tropical rain forest: a re view of the evidence. Journal of Applied Ecology 33: 200-209. Turner, M. D. (2003) Methodologic al reflections on the use of remote sensing and geographic information science of hu man ecological research. Human Ecology 31(2): 255-278. Turner, M.G., Gardner, R.H., & ONeill, R. V. (2001) Landscape Ecology in Theory and Practice. New York City, New York, USA. Springer. Turyahikayo-Rugyema, B. (1974) The History of the Bakiga in Southwestern Uganda and Northern Rwanda, 1500. Ph.D. Dissertation. University of Michigan. Ann Arbor, Michigan, USA. Twyman, C. (2001) Natural resource use and li velihoods in Botswanas wildlife management areas. Applied Geography 21: 45-68. Uganda Bureau of Statistics (2005) 2002 Uga nda population and housing census, Main Report. Entebbe, Uganda. Uganda Ministry of Water, Lands, and Enviro nment (2002) National progress assessment report on the implementation of Agenda 21. Kampala, Uganda. Uganda Wildlife Authority (2004) Kibale Na tional Park general management plan 2003-2013. Kampala, Uganda. Uganda Wildlife Authority (2005) Annua l Report for 2004-2005. Kampala, Uganda. US Census Bureau (2006) Uganda. Inte rnational Data Base (IDB). Population Division/International Programs Center. Washington DC, USA. http://www.census.gov/ipc/www/idbnew.html 174
175 Vogt, N., Bahati, J., Unruh, J., Green, G., Banana, A., Gombya-Ssembajjwe, W., & Sweeney, S. (2006) Integrating remote sensing data and rapid appraisals for land-cover change analyses in Uganda. Land Degradation & Development 17 : 31-43. Walpole, M.J. & Goodwin, H.J. (2001) Local attitudes towards conservation and tourism around Komodo National Park, Indonesia. Environmental Conservation 28(2):160. Walsh, S.J., Bilsborrow, R.E., McGregor, S.J ., Frizzelle, B.G., Messina, J.P., Pan, W.K.T., Crews-Meyer, K.A., Taff, G.N., & Baquero, F. (2003) Integration of Longitudinal Surveys, Remote Sensing Time Series, and Spatial Analyses. In: People and the Environment: Approaches for Linking H ousehold and Community Surveys to Remote Sensing and GIS. Fox, J., Rindfuss, R., Walsh, S., & Mishra, V., eds. pp. 91-130. Boston, Massachusetts, USA: Wang, J., Price, K.P., & Rich, P.M. (2001) Spatial patterns of NDVI in re sponse to precipitation and temperature in the central Great Plains. International Journal of Remote Sensing 22(18): 3827-3844. Wells, M. & Brandon, K. (1992) People and park s. Linking protected area management with local communities. Washington DC USA: World Bank, WWF, USAID. Western, D. (1994) Ecosystem conservation and ru ral development: the case of Amboseli. In: Natural Connections. Perspectives in Community-Based Conservation. Western, D. & Wright, R.W., eds. pp. 15-52. Washi ngton DC, USA: Island Press. Wilkie, D.S., Morelli, G.A., Demmer, J., Starkey, M., Telfer, P., & Steil, M. (2006) Parks and people: assessing the human welfare effects of establishing protected areas for biodiversity conservation. Conservation Biology 20 (1): 247-249.
BIOGRAPHICAL SKETCH From 1995-2000, Joel Hartter spent four years at the University of Michigan in Ann Arbor pursuing a B.S.M.E. (mechanical engineering) and a B.S. in German. During the 1996-97 school year, he studied abroad in Germany through Wayne State University's Junior Year in Munich program at both the Technische Universitt Mnchen and Ludwig s-Maximilians-Universitt. Following two brief years working as a mechanical engineer, he returned to graduate school to pursue a masters in forest engineering at Or egon State University. While there, Joel concentrated on making the transition to a natu ral resources-based disc ipline. At OSU, his emphasis was twofold: 1) to make the transition into the natural resource field, and 2) to build up his academic knowledge of natural resource management. His thesis research was focused on integrating new technology to reduce environmenta l impacts resulting from forest harvesting. While at OSU, he also became interested in na tural resource management issues in Sub-Saharan Africa. Following completion of his M.S. in 2004, Joel began to pursue a PhD at the University of Florida in geography funded through a Unite d States Department of Education Foreign Languages and Area Studies Grant. At the Univer sity of Florida his academic emphasis has been on human and environment interactions more specifically the impacts of landscape change on livelihoods. Joel continues to be interested in the interaction of social and ecological systems and developing synergistic remote sensing and fiel d techniques to study the impacts of landscape change at multiple spatial and temporal scales His PhD was completed in December 2007 with a minor in natural resource management a nd a certificate in African studies. 176