EFFECTS OF HABITAT EXTENT AND FOREST DISTURBANCE ON COMMUNITIES IN LOWLAND

Bhagawan Raj Dahal MSc (Ecology)

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2015 School of Geography, Planning and Environmental Management

Abstract Habitat loss and degradation are recognized as the major contributors to decline and extinction, and therefore represent a key conservation challenge for biodiversity conservation. Key to the protection of biodiversity is acquisition of ecological knowledge about how anthropogenic forest disturbances affect species and how species respond to emergent landscape characteristics. Furthermore, it is also important to assess how different management approaches and land tenures influence retention of the biota of particular sites and of landscapes. However, this crucial ecological knowledge is yet to be obtained for the threatened lowland landscapes of Nepal.

Protected areas cover only a small proportion of forests in lowland Nepal; the majority of forests outside the protected areas (off-reserve) have been managed by the state government. However, in recent years, community forestry programs have been increasingly popular as attempts to protect biodiversity while permitting consumptive forest use by people. It is therefore important to understand effectiveness of different forest management tenures for avifaunal conservation. I compared species richness, abundance, diversity and community composition of among sites in community forests, state forests and protected areas. Although sites in protected areas had the greatest richness of birds, community forests and state managed forests had complementary assemblages, supporting species not represented in protected areas. Vegetation characteristics such as large tree density, tree canopy cover and shrub density were also greater in community forests than in state-managed forests. The findings suggest that the community forestry approach appears to improve habitat quality compared to state-managed forests, and therefore can be an alternative tenure type for conservation of off-reserve forests and avifauna in the region.

Subsistence forestry practices such as logging, lopping, and grazing are sources of forest disturbance in lowland Nepal. Such activities do not reduce forest area, but change habitat characteristics, potentially affecting biodiversity directly, and through interactions with landscape characteristics. I examined effects of forest use practices on species richness and abundance of forest birds, and whether landscape context such as the extent of forest cover moderates disturbance effects on birds. I found that extraction of forest products reduced forest structural complexity and altered the avifaunal community of a site. At the site level, large tree density, tree canopy cover and shrub density were important habitat characteristics, while the extent of forest cover in the landscape had the greatest influence on richness of birds. The effects of forest disturbance (livestock grazing and logging) intensity on birds depended on the extent of forest in the surrounding landscape, with strongest effects in sites with less surrounding forest. Thus, although

ii site-level vegetation structure is important, maintenance of forest extent in the landscape is also key for avifaunal conservation in the region.

Several recent studies have demonstrated that the extent of forest cover and other landscape characteristics significantly influence bird species richness. However, different foraging guilds are likely to respond to landscape characteristics in different ways. Therefore, I examined the strength and magnitude of the relationships between the extent of forest cover and estimated species richness for overall birds and for each foraging guild separately. I found that landscape-level species richness of birds positively related to the extent of forest cover in the landscape. However, the relationship varied among the foraging guilds, with strong effects for foliage-gleaning insectivores and, to a lesser extent, frugivores, but only weak effects for sallying insectivores. The relationship between estimated species richness and the extent of forest cover in the landscape was nonlinear, with species richness decreasing more steeply below about 20-30% forest cover in the landscape. Importantly, I found that the relationship between richness and forest extent varied among foraging guilds and with landscape characteristics. Therefore, generalizing relationships between species richness and the extent of forest across all species could potentially mask important relationships at the functional level.

The findings of this thesis have important implications for the conservation of avifauna in multiple- use forest landscapes. Although both site-and landscape-scale forest characteristics have important influences on bird communities, the extent of forest in the landscape both directly and indirectly affects persistence of birds in these landscapes. The extent of forest in the landscape can moderate the effects of subsistence forest use practices on bird assemblages. Therefore, conservation benefits for avifauna can be maximized by maintaining both site-level habitat structures such as large trees, and the extent of forest cover at the landscape-level. This can be achieved with appropriate protection measures through reducing human pressure on forests, and restoration of degraded forest habitats, particularly those that are heavily exploited such the state-managed forests. Thus, management approaches such as community forestry for management of off-reserve forests can potentially complement protected areas and maximize conservation outcomes in the region. Such measures will improve habitat quality and increase the chance of maintaining viable populations of the full complement of avifaunal species in the lowland landscape of Nepal.

iii

Declaration by author This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the policy and procedures of The University of Queensland, the thesis be made available for research and study in accordance with the Copyright Act 1968 unless a period of embargo has been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

iv

Publications during candidature Dahal, B. R., C. A. McAlpine, and M. Maron. 2014. Bird conservation values of off-reserve forests in lowland Nepal. Forest Ecology and Management 323:28-38.

Dahal, B. R., C. A. McAlpine, and M. Maron. 2015. ‘Impacts of extractive forest uses on bird assemblages vary with landscape context in lowland Nepal’. Biological Conservation. 186: 167- 175.

Publications included in this thesis This thesis contains three jointly authored manuscripts (two published and one has submitted for peer-review). These papers are reproduced in full as chapters of this thesis (2-4). I conducted the majority (90%) of the work contained within these manuscripts, including: original idea, data collection, data analysis, interpretation, synthesis, drafting and writing. Co-author contributions are indicated below the relevant citations.

v

Dahal, B. R., C. A. McAlpine, and M. Maron. 2014. Bird conservation values of off-reserve forests in lowland Nepal. Forest Ecology and Management 323:28-38 – incorporated as Chapter 2.

Contributor Statement of contribution Bhagawan Raj Dahal, (Candidate) Research idea and design (80 %) Data collection (100 %) Wrote and edited the paper (85 %) Statistical analysis of data (100 %) Clive A. McAlpine Research design (10 %) Edited paper (5%) Martine Maron Research idea and design (10 %) Wrote and edited the paper (10 %)

Dahal, B. R., C. A. McAlpine, and M. Maron. 2015. ‘Impacts of extractive forest uses on bird assemblages vary with landscape context in lowland Nepal’. Biological Conservation 186: 167-175. –incorporated as Chapter 3.

Contributor Statement of contribution Bhagawan Raj Dahal, (Candidate) Research idea and design (80 %) Data collection (100 %) Wrote and edited the paper (85 %) Statistical analysis of data (100 %) Clive A. McAlpine Research design (10 %) Edited paper (5 %) Martine Maron Research idea and design (10 %) Wrote and edited the paper (10 %)

vi

Contributions by others to the thesis No contributions by others apart from those listed above

Statement of parts of the thesis submitted to qualify for the award of another degree None

vii

Acknowledgments I would like to take this opportunity to express my sincere gratitude to The University of Queensland for awarding me an International Postgraduate Research Scholarship (IPRS) and UQ Centennial Scholarship (UQCent) that enabled me to undertake PhD study. Without this support, my PhD journey would not have commenced at this renowned University. Many thanks to my enrolling school - the School of Geography, Planning and Environmental Management for all the help and support I received for my research.

I am very much grateful to my supervisors Martine Maron and Clive McAlpine for their consistent support throughout this research project. I am particularly truly indebted and thankful to my primary advisor Martine Maron whose wonderful cooperation, consistent encouragement and support made this thesis possible. Thank you so much for your advice, guidance, ideas, time, help and patience to complete this thesis. I am very grateful and honoured to have you as my principle supervisor.

I thank to the Ministry of Forest and Soil Conservation Nepal for granting research permission. Many thanks to staffs of Himalayan Nature, Koshi Bird Observatory, Chitwan National Park, Parsa Wildlife Reserve and Community forest user committees for their cooperation during the field work.

Many thanks to Suman Acharay, Dinesh Ghimire, Kapil Pokhrel, Tika Sherpa and Arjun Baral for your assistance in the field. I really enjoyed working with you all. I am thankful to Jeevan Bikash Samaj, Arjun Chapagain and Tika Ram Giri for their support in logistic management. I would like to thank to Rudra Kumal, Hira Shrestha and Geeta Ghimire who generously offered me their house to live in during the field work. Many thanks for your all cooperation and help. I am very much grateful to Wayne Heydon for his wonderful hospitality during my stay in Brisbane for thesis correction.

I would like to express my sincere gratitude to Mr Douglas Michael and Mrs Margo Michael for their care and love during my stay in Brisbane. They kindly provided me a free accommodation for two years at their home. Thanks to Meredith Gray for your help in organizing all these things. I am grateful to Dr Roger Jones for his time in brushing up my writing skills and Jessie Wells for her input in statistical modelling.

viii

I would like to thank to GPEM family particularly Judy, Alan, Suhan for their cooperation in administrative work. I am grateful to my colleagues for their excellent company here in University. Thanks to Emma, Kiran, Alvaro, Will, Justus, Andrew, Danielle, Jeremy, Payal, Rowan, and all LEC friends. Special thanks to Kiran for his help in GIS. I thank to Pradeep Gyawali, Gokarn Junga Thapa, Biplav Pokhrel, Rajiv Paudel, Uttam Babu Shrestha and Salma Baral for their support in different stages of my PhD study.

I am grateful to Dr Hem Sagar Baral and Carol Inskipp for their inspiration to work in the field of bird conservation. Their contribution to ornithological research and conservation in Nepal is exceptional. I am truly indebted to Hem Dai for his consistent cooperation and support to build my career in conservation science. I would like to thank to World Pheasant Association (WPA), Dr Narayan Dhakal, Dr Phillip McGowan and Prof John Carroll for their support in my research career. I am also thankful to my thesis reviewers’ Dr Peter Garson and Dr Toby Gardner for their constructive comments, which greatly helped improve this thesis.

Finally, I am thankful to my family for their continuous support and encouragement. I am very grateful with my mother Gaura Devi, late father, all my brothers and sisters for their love, care and consistent support for my studies. They taught me the value of education in my life and society from the beginning of my formal education. Many thanks to my wife Amita and daughter Abha for their love and understanding which was important to accomplish this thesis.

ix

Key words Avian community, conservation value, extractive forest disturbances, forest management, habitat thresholds, interactive effect, landscape context, landscape-level forest cover, multi- use forest landscapes, tenure arrangement

Australian and New Zealand Standard Research Classifications (ANZSRC) ANZSRC code: 050104, Landscape Ecology (50%) ANZSRC code: 050202, Conservation and Biodiversity (30%) ANZSRC code: 050211, Wildlife and Habitat Management (20%)

Fields of Research (FoR) Classification FoR code: 0602, Ecology, (50%) FoR code: 0502, Environmental Science and Management, (30%) FoR code: 0501, Ecological Applications, (20%)

x

Table of contents Abstract ii Declaration by author iv Acknowledgments viii Table of contents xi List of Tables xiv List of Figures xv List of Plates xvii List of Appendices xviii CHAPTER 1: INTRODUCTION AND LITERATURE REVIEW 19 1.1 Background to the problem 20 1.2 Aims and objectives 21 1.3 Theory and concepts underlying this thesis 21 1.3.1 Human modification of landscapes 21 1.3.2 Effects of landscape change on biodiversity 22 1.3.3 Scale and species richness 22 1.3.4 Extent of forest and species richness 23 1.3.5 Conservation of remnant forests 24 1.3.6 Effects of forest use practices on avifauna 26 1.3.7 Forest management in Nepal 27 1.3.8 Threats to lowland forest and birds 29 1.4 Thesis outline 31 CHAPTER 2: BIRD CONSERVATION VALUES OF OFF-FOREST RESERVE FORESTS IN LOWLAND NEPAL 34 2.1 Abstract 35 2.2 Introduction 36 2.3 Material and methods 39 2.3.1 Study area 39 2.3.2 Study sites 40 2.3.3 Bird surveys 41 2.3.4 Explanatory variables 42 2.3.5 Statistical analyses 43 2.4 Results 45 2.4.1 Species richness, abundance and diversity 45

xi

2.4.2 Habitat preference groupings 46 2.4.3 Foraging guilds 47 2.4.4 Bird assemblage structure 48 2.4.5 Habitat characteristics 50 2.5 Discussion 51 2.6 Management implications 53 CHAPTER 3: IMPACTS OF EXTRACTIVE FOREST USES ON BIRD ASSEMBLAGES VARY WITH LANDSCAPE CONTEXT IN LOWLAND NEPAL 55 3.1 Abstract 56 3.2 Introduction 57 3.3 Material and methods 59 3.3.1 Study area 59 3.3.2 Study sites and landscapes 59 3.3.3 Bird surveys 60 3.3.4 Explanatory variables 60 3.3.5 Statistical analysis 62 3.4 Results 65 3.4.1 Vegetation and disturbance 65 3.4.2 Bird communities 65 3.4.3 Effects of site- and landscape-level factors 66 3.4.4. Interactions between forest extent and disturbance intensity 69 3.5 Discussion 72 3.5.1 Effects of forest disturbances on bird communities 72 3.5.2 Moderating effects of landscape context 73 3.6 Management implications 75 CHAPTER 4: RELATIONSHIPS’ BETWEEN LANDSCAPE-LEVEL SPECIES RICHNESS AND FOREST EXTENT VARY AMONG BIRD GUILDS 76 4.1 Abstract 77 4.2 Introduction 78 4.3 Material and Methods 80 4.3.1 Study area 80 4.3.2 Study design 81 4.3.3 Bird surveys 81 4.3.4 Landscape variables 81

xii

4.3.5 Data analysis 82 4.4 Results 83 4.4.1 Relationships between landscape-level species richness and forest cover 83 4.4.2 Relative importance of landscape variables 84 4.4.3 Interactions between landscape characteristics 86 4.5 Discussion 87 4.5.1 Species richness and forest cover 87 4.5.2 Relative importance of landscape characteristics 89 4.5.3 Interactions between landscape characteristics 89 4.6 Management implications 90 CHAPTER 5: SYNTHESIS AND RECOMMENDATIONS 91 5.1 Overview 92 5.2 Off- reserve forests provide complementary habitats for bird conservation 93 5.3 Forest use practices can have detrimental effects on vegetation and associated birds 94 5.4 Effects of forest-use practices on bird assemblages vary with the landscape context 96 5.5 Relationships between species richness and forest extent vary among bird guilds 98 5.6 Management implications 99 5.7 Limitations of my research 107 5.8 Future research 107 5.9 Conclusion 108 REFERENCES 110 APPENDICES 133 Appendix A 133 Appendix B 138 Appendix C 156

xiii

List of Tables Table 2.1 Summary of explanatory variables used to assess influence of stand and landscape- scale variables on bird communities 43 Table 2.2 Mean species richness, average abundance per survey, Shannon diversity index and Pielou evenness index of all birds in three forest management tenures 46 Table 2.3 Mean species richness, average abundance per survey, Shannon diversity index and Pielou evenness index (±SE) of birds of different habitat groups in community managed forests, state managed forests and protected areas 46 Table 2.4 Mean species richness, average abundance per survey, Shannon diversity index and Pielou evenness index of different foraging guilds of birds in three forest management tenures 47 Table 2.5 Results of PERMANOVA and PERMDISP pairwise comparison tests of bird community composition in response to different forest management tenures 48 Table 3.1 Summary of explanatory variables used to assess the influence of stand and landscape-scale variables on bird communities 62 Table 5.1 Summary of study themes and key findings of this thesis 103

xiv

List of Figures Figure 2.1 The location of the study area and bird survey sites: (a) Chitwan forests (b) Parsa forests and (c) Eastern Terai forests 41 Figure 2.2 Bird assemblage NMDS (2D stress = 0.24) plot fitted with environmental vectors. Vectors represent the mean direction and strength of correlation of different environmental variables (Fextent = Forest extent, Scv = Shrub cover, Sden = Shrub density, Mtdn = Mature tree density, Tcv = Tree canopy cover). Vectors with significance P < 0.05 only are shown49 Figure 2.3 Distribution of tree size classes (±S.E.) in sites in community forests, state forests and protected areas 50 Figure 3.1 Location of the three study regions in Nepal: a) Chitwan forest, b) Parsa-Bara forest and c) Eastern forest 59 Figure 3.2 Mean (± S.E.) species richness (dark bars) and average abundance (light bars) of (a) all birds (b) bark-gleaning insectivores and (c) foliage-gleaning insectivores 66 Figure 3.3 Model averaged coefficient estimates (± S.E.) across the 95% confidence set of models for all explanatory variables 67

Figure 3.4 Summed Akaike weights (∑ωі) of final subset of the explanatory variables for (a) overall species richness, (b) average abundance, (c) bark-gleaner species richness (d) bark- gleaner abundance (e) foliage-gleaner species richness, and (f) foliage-gleaner abundance69 Figure 3.5 Relationship between overall bird species richness (a),overall bird abundance (b), bark-gleaning abundance (c), and foliage-gleaning abundance and percent of forest cover in 2.5 km radius of survey site in landscape (heavily disturbed sites (filled circles and heavy dashed line) and lightly disturbed sites (open circles and fine dashed line). For the purposes of displaying the interaction effect, sites were divided into heavily and lightly disturbed based on the value of the disturbance index (from three disturbance types) 71 Figure 4.1 The study landscapes in lowland Nepal (grey shading indicates forest cover) and histogram showing the different extent of forest cover in landscape 80 Figure 4.2 Models of the relationship between estimated species richness and forest extent in the landscape. AIC: Akaike information criterion for each model 84 Figure 4.3 Model averaged coefficient estimates (± S.E.) across the 95% confidence set of models for all explanatory variables: (a) overall bird, (b) foliage-gleaning insectivore, (c) frugivore and (d) sallying insectivore 85

Figure 4.4 Summed Akaike weight (∑ωі) from model averaging of the environmental variables for landscape-level richness of (a) all species, (b) foliage-gleaning insectivores, (c) Frugivores, and (d) sallying insectivores 86

xv

Figure 4.5 Relationship between overall estimated richness and percent of forest cover (a) in more- disturbed landscapes (filled circles and full line) and less-disturbed landscapes (open circles and heavy dashed line) and (b) higher-rainfall landscapes (filled circles and full line) and lower-rainfall landscapes (open circles and heavy dashed line) 87

xvi

List of Plates Plate 1: An example of resource extraction in lowland Terai forests (anti-clockwise from top left: logging for timber, removal of standing tree for firewood, cattle grazing, and tree canopy collection for fodder). 19 Plate 2: Regeneration of forest after implementing a community forestry program in Barandabhar Community forest, lowland Nepal 34 Plate 3: Creekbed in large patch of forest adjacent to an agricultural area in Parsa Wildlife Reserve, lowland Nepal 55 Plate 4: A patch of native forest in Chitwan National Park in lowland Nepal 76 Plate 5: Agricultural intensification in adjacent to a protected area, lowland Nepal 91

xvii

List of Appendices Appendix A 133 Table A.1 Total number of individuals and their respective food guilds and habitat groups, and relative percentage of species observation across sites in lowland tropical landscapes: CF = Community forest, SF = State forest, PA = Protected area 133 Table A.2 Top ten contributing species to dissimilarities between management tenures using SIMPER analysis based on Bray-Curtis dissimilarity 137 Appendix B 138 Table B.1 Summary statistics of habitat characteristics (± SE) across sites in grazing, logging and lopping disturbances in lowland tropical landscape 138 Table B.2 Results of analysis of variance (ANOVA) for the vegetation characteristics across the different disturbance types 139 Table B.3 Results of analysis of variance (ANOVA) for overall species richness, total average abundance, species richness and abundance of bark-gleaning and foliage-gleaning insectivore 142

Table B.4 Model averaged coefficient and sum of Akaike weights (∑ωі) (fixed effects) for each variable based on AIC. Significant estimates are in bold 144

Table B.5 AIC, ∆value and Akaike weights (ωі) for models of overall species richness, total average abundance, species richness and abundance of bark-gleaning and foliage-gleaning insectivore 146

Table B.6 AIC, ∆value and Akaike weights (ωі) for interaction models of overall species richness, total average abundance, species richness and abundance of bark-gleaning and foliage- gleaning insectivore 152 Table B.7 Mean values of different disturbance types between sites classified as lightly and heavily disturbed in lowland Terai forests 154 Table B.8 Correlation matrix of explanatory variables measured. Coefficients in bold shows pairs highly correlated variables 155 Appendix C

Table C.1 AIC, Δ value and Akaike weights (ωі) for models of overall estimated species, frugivore, foliage-gleaning insectivore and sallying insectivore 156 Table C.2 Model averaged coefficients across the 95% confidence set of models for all explanatory variables 160 Figure C.1 Species accumulation curves of all 28 studied landscapes based on Chao2/ICE estimated richness 161

xviii

CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW

Plate 1: An example of resource extraction in lowland Terai forests (anti-clockwise from top left: logging for timber, removal of standing tree for firewood, cattle grazing, and tree canopy collection for fodder).

19

1.1 Background to the problem In the face of growing demand for food and resources for an ever-increasing human population, managing forests for biodiversity conservation is a challenging task. Rapidly growing demands for food and resources have often been met at the cost of forests and woodlands across the globe (FAO 2005, Gibbs et al. 2010). Approximately 80% of the world’s original forest cover has been cleared, fragmented and degraded (World Resources Institute 1997). Anthropogenic habitat degradation is greater in areas of high population density and poverty, particularly in developing countries (Laurance 2010) where a significant proportion of the population lives near the forests (Hegde and Enters 2000, Millennium Ecosystem Assessment 2005).

Such widespread anthropogenic habitat loss and degradation has significantly changed the pattern and function of the landscape and consequently altered the habitat characteristics in the landscape (Christensen et al. 2009). Loss of habitat is the key threat to biodiversity and is the major contributor to global species extinction (Fahrig 2001, Pereira et al. 2004). Landscape characteristics such as the extent of forest cover in the landscape and forest disturbance type and intensity may be important determinants of species richness in the landscape. Therefore, understanding the response of species to the emergent properties of landscape is crucial for their effective conservation.

Forest loss and degradation as a result of anthropogenic pressure pose the principal threats to forest-dependent birds (Schmiegelow and Mönkkönen 2002). However, effects of disturbance on forest birds may vary among species, depending on their sensitivity to habitat disturbances. Anthropogenic disturbance often has the most adverse effects on forest habitat specialists (Sodhi and Ehrlich 2010) and species of birds with specialised diets (Cleary et al. 2007, Greve et al. 2011). Forest loss and degradation are often associated with disproportional loss of critical habitat components such as large trees, fallen logs and woody debris (Inskipp et al. 2013). Persistence of many species depends on availability of such critical habitat features in the landscape (Kumar et al. 2011). Thus, as habitat loss changes the distribution of habitat resources in the landscape (Andren 1994, Coulson and Tchakerian 2010), understanding how bird communities respond to habitat loss and associated forest disturbance is crucial for making sound conservation management decisions.

20

1.2 Aims and objectives The overall aim of this study was to investigate the effects of habitat characteristics, forest extent and forest use disturbances on forest bird assemblages in human-dominated landscapes in lowland Nepal. A secondary aim was to explore conservation values of off-reserve forests for bird communities. In this thesis, I examined effects of anthropogenic forest use practice such as logging, grazing and lopping on vegetation structure and associated bird communities, and the role of landscape context in moderating the effects of disturbance on birds. Further, I investigated whether the effects of landscape-level forest extent on landscape-level species richness of birds are universal across foraging guilds, or if a particular guild contributes disproportionately to observed patterns between species richness and forest cover in the landscape. The specific research objectives investigated in thesis are to:  Determine conservation values of off-reserve forests for bird communities in comparison to those of protected areas in lowland Nepal.  Investigate effects of site and landscape characteristics on bird species richness and abundance, and explore interactions between disturbance intensity and landscape context in their effects on bird assemblages.  Investigate whether the effects of remnant forest extent in a landscape on the species richness of birds in that landscape are consistent across foraging guilds, or if a particular guild contributes disproportionately to observed relationship between species richness and forest cover in the landscape.

1.3 Theory and concepts underlying this thesis 1.3.1 Human modification of landscapes Native forests are invaluable habitats for the conservation of globally significant biodiversity. These forests encompass important characteristics of natural ecosystems and support important terrestrial ecosystems (World Commission on Forests and Sustainable Development 1999). However, despite the global biological importance of forest ecosystems, anthropogenic activities such as forestry, subsistence and commercial agriculture, and urban development have caused significant loss and degradation of forest landscapes. These processes of habitat loss and modification reduce the total area of suitable habitat, increase the number of patches and increase the isolation of patches from one another (Saunders et al. 1991, Fahrig 2003, Fischer and Lindenmayer 2007). Such fragmentation contributes directly

21 and indirectly to changes in the structure and function of ecological mosaics and their constituent biota (Franklin et al. 2002, Morrison et al. 2006, Fischer and Lindenmayer 2007).

1.3.2 Effects of landscape change on biodiversity Landscape change alters ecological processes and threatens the survival of species and populations in several ways. For example, landscape modification involves the removal of critical habitat components and alters habitat structure and composition, thereby affecting long-term persistence of faunal populations (Flynn et al. 2009, Halley and Iwasa 2011, Lindenmayer and Cunningham 2013). Anthropogenic modification of habitat can disproportionately affect particular vegeation types, which reduces native plant species richness, in turn reducing habitat diversity in the landscape (Fischer and Lindenmayer 2007). The reduction in habitat extent and diversity significantly alters the faunal assemblages that landscapes can support (Fahrig 2003, Bennett et al. 2006, Morrison et al. 2006). As a consequence, many species of fauna are severely threatened and some are on the verge of local or even global extinction (Jetz et al. 2008, IUCN 2009). Therefore, understanding the relationships between patterns of species occurrence and landscape properties in modified landscapes is critically important for developing effective conservation strategies (Radford and Bennett 2007, Gardner et al. 2009, Balmford et al. 2012).

Landscape modification not only reduces the type and amount of habitat but also changes the configuration of remaining habitat patches (van den Berg et al. 2001, Bennett et al. 2006). Such changes affect resource availability for a wide range of faunal species. Because isolated habitat patches may not contain all of a species’ resource requirements (Bennett and Saunders 2010), many species that require different habitat attributes for different purposes (such as foraging and breeding) must travel between patches to acquire resources (Dunning et al. 1992, Law and Dickman 1998). Increasing isloation of habitat attributes could potentially impede this movement of fauna, affecting the persistence of faunal communties in human- modified landscapes (Hinsley 2000, Fahrig 2003, Ewers and Didham 2006). Thus, the protection of a complementary habitat within close proximity can be important for maintaining viable populations of many species (Dunning et al. 1992).

1.3.3 Scale and species richness Understanding how the landscape affects the species that can occupy a given site does not necessarily reveal the patterns of species diversity expected across different spatial scales.

22

The number of species in an individual site/or patch (alpha diversity) is only part of what generates regional species richness (Whittaker 1972). As the number of sites sampled across an area increases, variation in species composition (beta diversity) across sites/or habitat types in a landscape is also likely to increase (Whittaker 1977, Koleff et al. 2003, Kessler et al. 2009). Because a landscape is composed of a mosaic of different habitat types, the contribution of landscape heterogeneity to species diversity can be important. Thus species richness at the site level might not necessarily represent the pattern of species diversity at larger scales (Kettle and Koh 2014). The rate of change of species composition among sites or habitat patches (also called species turnover) contributes significantly to the regional diversity (γ-diversity) Tscharntke et al. 2005). Thus, a multi-scale approach to understanding the relationships between landscape properties and species richness is important for conservation planning for regional diversity (Bennett et al. 2006, Radford and Bennett 2007).

1.3.4 Extent of forest and species richness There is likely to be a minimum extent of habitat in the landscape required to support the persistence of full species diversity (Andren 1994, Fahrig 2001). Landscapes with more forest cover support a larger species pool (Hu et al. 2012, Taylor et al. 2012) through both sampling effects (Wiens 1992, Whittaker and Fernández-Palacios 2007) and because a greater forest extent offers habitat diversity (Radford et al. 2005, Maron et al. 2012). Several empirical studies have shown that sites in landscapes with more forest support higher densities of reptiles (McAlpine et al. 2015), greater species richness and abundance of birds (Villard et al. 1999, Mortelliti et al. 2010, Martensen et al. 2012, Taylor et al. 2012), and greater richness of small mammals (McAlpine et al. 2006, Estavillo et al. 2013). These findings suggest that a wide range of forest-dependent faunal communities can persist even in modified landscapes, as long as adequate forest cover is retained (Radford et al. 2005, Ochoa‐Quintero et al. 2015). Moreover, the extent of forest in the landscape not only affects faunal communities directly, but may also have a potential role in moderating the effects of anthropogenic disturbance. For example, the intensity of disturbance (livestock grazing, logging and extraction of firewood and fodder) may interact with the amount of forest in the landscape to drive landscape-level species richness. Thus, understanding of interactive effects between forest extent and disturbance intensity is important for biodiversity conservation in multiple-use forest landscapes.

23

The understanding of the nature and magnitude of the relationship between species richness and forest extent is important for conservation of faunal communities in human-modified landscapes. As habitat extent decreases, so do ecological functions and associated biodiversity in the landscape (Swift and Hannon 2010). Species richness may respond non- linearly to habitat loss, and decline sharply when the extent of forest exceeds certain threshold levels (Maron et al. 2012, Ochoa‐Quintero et al. 2015). For example, in the eastern , Maron et al. (2012) found a non-linear relationship between woodland birds and landscape-level forest extent. A sharp decline in species richness occurred when forest extent fell below 40% in the landscape. In southern Australia, Radford et al. (2005) found that the species richness of woodland birds declined sharply in landscapes with less than 10% of habitat cover. Similarly, Martensen et al. (2012) found evidence of thresholds for understory birds at 30-50% of forest cover in Atlantic Forest landscapes, below which species richness declined more steeply. A similar trend in the relationship between species richness and landscape-level forest cover was reported by Ochoa‐Quintero et al. (2015) in the Brazilian Amazon forest. They found evidence of a threshold at 30-40% of forest cover, below which species richness of both mammals and birds declined aburptly.

Although these relationships have typically been considered as general phenomena, the response of species to habitat extent may vary depending on land-use or soil type (Maron et al. 2012), species sensitivity to forest cover change (Martensen et al. 2012), and diversity of food resoures available in the landscape. For example, species richness of frugivorous birds in particular landscapes is likely to be driven by the diversity and availability of fruit resources (Kissling et al. 2007), while, although at site level, abundance of insectivorous birds is affected by the abundance of prey resources such as arthropods in forest patches (Capinera 2011). Therefore, it is likely that the relationship between the extent of forest cover and landscape-level species richness of birds varies among different foraging groups, although this has not hitherto been explored.

1.3.5 Conservation of remnant forests In response to continued loss of habitat and biodiversity, protected areas have been established around the globe. A large body of research has shown that protected areas have been important in maintaining larger extent of forests and their constituent biota (e.g. Rodrigues et al. 2004, Jenkins and Joppa 2009). Protected areas are often targeted at protecting species of high conservation value (e.g. Lee et al. 2007), and therefore reduce the

24 risk of species extinction (Brooks et al. 2004, Jenkins and Joppa 2009, Joppa and Pfaff 2011). However, the protected area network also suffers from lack of representation of certain species and habitats (Tewksbury et al. 2002, Hoekstra et al. 2005, Barr et al. 2011, Butchart et al. 2012). Protected areas may not necessarily contain the full range of habitats required to support all species across the landscape. For example, approximately 43% of Asian Important Bird Areas are unprotected (BirdLife International 2004). In addition to this, protected areas have not been able to prevent habitat loss and associated faunal diversity in many parts of the world (e.g. Clark et al. 2013). These findings indicate that protected areas alone are inadequate for conservation.

In response to the inadequacy of the current protected area network, there has been increasing interest in conservation of off-reserve forests for biodiversity conservation (Bhagwat et al. 2005, Ellis and Porter-Bolland 2008, Persha et al. 2010). As more than 80% of the world terrestrial habitats are outside protected areas (Barr et al. 2011, Bastian et al. 2012), these off- reserve forests can be important complementary habitats to the existing protected area network. For example, off-reserve forests often include patches of remnant native forest of different types to those within protected areas, and regrowth (secondary) forests (Dudley and Phillips 2006). These provide habitat resources for many species that are poorly represented within protected areas (Lindenmayer and Burgman 2005, Mathur and Singh 2008, Cox and Underwood 2011). Thus, appropriate policies for conservation management of off-reserve forests could help maintain a diversity of habitat resources for a range of species across the landscape (Lindenmayer 2009).

Community forestry has emerged as an alternative forest management strategy to commercial forestry or complete reservation, particularly in developing countries (Ellis and Porter- Bolland 2008, Porter-Bolland et al. 2012, Kitamura and Clapp 2013). Community forests offer legal rights to local community and establish a sense of ownership in forest resource management (Hayes and Ostrom 2005, Singh and Chapagain 2006). In developing countries, about 27% of the total forest area is either community-managed or owned (Molnar et al. 2011). Studies have shown that community forest initiatives have lower deforestation and improve forest condition in the landscapes due to restoration of degraded forests (Nagendra and Gokhale 2008, Porter-Bolland et al. 2012, Lambrick et al. 2014). They are also important sources of livelihoods for forest-dependent human communities (Lawrence et al. 2006, Charnley and Poe 2007). For example, forest resources such as fuel wood, fodder, and timber

25 are the primary source of livelihood for local communities (Kumar and Shahabuddin 2005). Thus, managing forests outside protected areas through community participation can be beneficial to both the biodiversity and local community.

1.3.6 Effects of forest use practices on avifauna Anthropogenic forest use practices such as timber and firewood extraction and livestock grazing are key drivers of forest loss and degradation (Millennium Ecosystem Assessment 2005). Yet in many developing countries, these practices constitute the most important aspects of a subsistence livelihood (Chao 2012, World Bank 2012). For example, globally, particularly in developing countries, between one-third and one-half of people rely on wood and other biomass fuels for energy (Bailis et al. 2012). Such widespread exploitation of forest severely alters habitat characteristics, with changes affecting trees, shrubs and associated structures such as hollows (Chettri et al. 2002, Kumar et al. 2011). This structural simplification of habitat can result in significant declines of fauna (Lindenmayer et al. 2012, Lee and Carroll 2014).

In developing countries, forest resource extraction such as timber and firewood, lopping of tree branches for fodder and firewood, and cattle grazing are major forms of disturbance in remnant forests (Chettri et al. 2005, Shahabuddin and Kumar 2007, Thapa and Chapman 2010). Excessive extraction of these resources can lead to significant habitat loss and degradation. This may disproportionally reduce the key suitable habitats for faunal communities in the landscape. Harvesting of resources for human use potentially affects vegetation structure and composition (Sagar and Singh 2004, Kumar and Shahabuddin 2005, Gil-Tena et al. 2008). The removal of standing trees changes stand structure (Marzluff et al. 2000, Moktan et al. 2009) and alters the abundance and density of trees (Lindenmayer 2009). For example, Chettri et al. (2005) found that large tree density, tree basal area and woody biomass significantly lower in sites in heavily degraded forests in , . Therefore, the simplification of vegetation structure as a result of forest resource extraction may pose a serious threat to the persistence of avifauna.

Studies have shown that structural complexity of habitat strongly influences bird communities (Sekercioglu et al. 2002, Maron and Kennedy 2007), especially forest interior specialists and insectivores (Lee et al. 2007, Greve et al. 2011). These groups of birds use large and dead or rotting trees as foraging and nesting habitats (Shahabuddin and Kumar

26

2007). For example, species richness and abundance of bark-gleaning insectivores can be reduced by the reduced density of larger trees (Adams and Morrison 1993), while removal of tree branches and canopy cover may have caused lower richness and abundance of foliage- gleaning insectivores (Leal et al. 2013). However, most studies of the effects of extractive forest practices have focused on the effects of large-scale logging on avifaunal communities (Sekercioglu et al. 2002, Politi et al. 2012, Thinh et al. 2012). The role of repeated extraction of smaller amounts of woody biomass for timber and firewood in affecting birds and bird functional groups is less well-understood.

Livestock grazing is another common and detrimental form of anthropogenic forest disturbance. The disturbance caused by grazing significantly affects forest ecosystem processes and reduces plant diversity (Ludwig et al. 2000, Mayer et al. 2006). Changes in plant species composition and foliage density may adversely affect the resource availability to birds (Alexander et al. 2008), and thus affect their richness and abundance. Grazing often perturb bird species assemblages, with effects varying with grazing intensity (Eyre et al. 2009, Whitehorne et al. 2011). Grazing changes understory composition through reduction or removal of shrubs and herbaceous perennials (Buffum et al. 2009); therefore species that depend on understory vegetation for foraging and nesting may be most negatively affected by livestock grazing (Martin and McIntyre et al. 2007). Grazing is a common practice in community-managed forests, and so understanding its effects in those forests is important for recommendations around forest management prescriptions and tenure.

1.3.7 Forest management in Nepal In Nepal, forests cover nearly 36% of the land area and include many important ecosystems (BCN and DNPWC 2012). The country’s forests are designated as national or private forests, with five sub-categories on the basis of management regimes: state managed forests (also called national forest), community forests, protected forests, leasehold forests and religious forests. However, most forest is managed under three of these management regimes: state managed forests, community managed forests, and protected forests. An estimated total of 2.45 million hectares remains off reserve (Shyamsundar and Ghate 2011). Of the forest outside the protected area, 1.2 million hectares are currently managed by community forest user groups (DoF 2010, BCN and DNPWC 2012), while ~1 million hectares of forests is under the national government management (Shyamsundar and Ghate 2011).

27

Although Nepal has designated about one quarter of its land mass as protected areas, the majority of its protected land is concentrated in the high Himalayas and throughout the less- productive landscapes (HMG/MFSC 2002, Heinen and Shrestha 2006). For example, about 48% of high Himalayas are protected, whereas only 0.8% of the Mild Hills and 5.5% of the Terai zone are protected (Shrestha et al. 2010). Thus, only a small proportion of Nepal’s most productive lowland forest is represented in the current protected area network. Despite the small proportion in size, lowland protected areas are important in conservation of country’s last remaining natural habitats and their constituent biota. They are critical for the conservation of many endangered and globally significant mammalian species (e.g. Greater One-horned Rhinoceros Rhinoceros unicornis, Royal Bengal Tiger Panthera tigris) as well globally threatened avifaunal species such as Buceros bicornis and Kashmir Flycatcher suvrubra.

However, protected areas do not represent large tracts of primary forests in the lowland of Nepal. Several important habitats for birds are still located outside the protected areas and have received little conservation attention despite their great bird conservation values. For example, in recent years, community-managed forests have become increasingly recognised for their conservation values for biodiversity in Nepal. Nepal offers some of the best examples of community-based forest management in the world (Pokharel et al. 2007, Nagendra and Gokhale 2008). The transfer of use and management rights to locally-formed community groups has improved effectiveness in management of resources (Ojha et al. 2009, Nagendra 2007, Shyamsundar and Ghate 2011). Since the implemenation of the Forests Act of 1993 and the Forest Rules and Regulations of 1995, more than one quarter of country’s national forest has become managed by community forest user groups (CFUGs) (MOF 2012). Local communities perceive community forest regime as a secure tenure generating sustainable sources of income from forest products (Gautam 2007, Kanel and Dahal 2008). For example, the annual income of community forests was double the total revenue of the Department of Forests for the fiscal year of 2002 (Kanel et al. 2003). This ownership in resource management and utilization is one of the key drivers of community forestry success. Studies have shown that rates of habitat loss and degradation are reduced in community managed areas compared with state managed forests (Nagendra 2007, Kanel and Dahal 2008). In addition to improving forest health, community forestry has played a crucial role in boosting local livelihoods (Sapkota et al. 2009).

28

Although both protected areas and off-reserve forests comprise a -dominated forest mosaic, habitat characteristics among these differently-managed forest types vary. This is primarily due to different forest management practices. For example, the strict reserve systems are predominantly relatively undisturbed primary forests (Nagendra et al. 2008). Off- reserve forests support both primary forest and regenerating forest (Kanel and Dahal 2008). Such structural and compositional differences in habitat characteristics among tenure types may offer habitat for different assemblages of faunal species in the landscapes. The old growth primary forest in protected areas, for example, may support an assemblage of forest- sensitive species and forest specialists (Inskipp and Inskipp 1991, BCN and DNPWC 2011) while the silviculturally-treated and successional forest habitat within community forests may be expected to support a range of bird species including forest specialists (BCN and DNPWC 2011), open country species (Roman 2001) and mixed-habitat species (Anand et al. 1997, Blake and Loiselle 2001). Thus, the effective conservation of these forests may play an important complementary role in safeguarding regional diversity of birds.

However, the response of faunal communities to community management of forests is poorly known in the region. In Nepal, the majority of faunal studies have been within the protected areas (Inskipp and Inskipp 2001, Baral et al. 2012, Kapfer et al 2012) and are largely focussed at species level (Poudyal et al 2008, Dahal et al 2009, Baral 2012). More recently, there have been some studies in off-reserve forests, in particular on forest regeneration and plant communities (Gautam et al. 2002, Timilsina et al. 2007, Sapkota et al. 2009, Thapa 2010). Studies reported a significant increase in plant diversity and decreased deforestation within the community-managed forests following the tenure change (Nagendra et al. 2008, Sapkota et al. 2009). Yet, to my knowledge, the effect of this vegetation change on fauna is poorly understood in Nepal. There is therefore a need to quantify the contribution of these areas to biodiversity conservation, and in particular, the extent to which the biota they support is complementary to that within formal protected areas.

1.3.8 Threats to lowland forest and birds Forests in lowland landscapes are generally degraded due to anthropogenic forest extraction activities (Webb and Sah 2003, DoF 2009). Before 1950, the region supported continuous dense tropical forest. With the eradication of malaria in the early 1950s, large tracts of the highly productive lowland forests were converted to agriculture (Hrabovszky and Miyan 1987). Consequently, most of the forest was destroyed and the remaining forest areas were

29 subjected to intense human exploitation. Nearly half of Nepal’s population now lives in the 17% of the country that is lowland (Central Bureau of Statistics 2011). Therefore the majority of forests in this region are exploited mainly for subsistence livelihood.

Forest use practices such as logging, grazing and the excessive extraction of fuel wood and fodder are the major activities in these forests (Webb and Sah 2003). Similarly, extraction of litter, grass and removal of dead and logged trees are other examples of extraction of forest- based products in the region. Livestock farming is a major profession and is an essential component of the rural farming system in Nepal. For example, about 41% of the country’s cattle, 39% of buffaloes and about 37% of goats are all farmed in the lowlands (MoAC 2004), that has contributed to the rapid depletion of forest land (HMG/MFSC 2002).

Despite the prevailing threats, these forests can be important habitats for a wide range of avifaunal diversity. Recent assessment study by BirdLife Nepal and Wildlife Department- Nepal for the region reported that these forests harbour more than 50% of nationally threatened species (BCN and DNPWC 2010), over three quarters of the country’s breeding species and 67% of wintering species (Inskipp 1989). Of the total species recorded in Nepal, over 70% of the forest bird species are recorded in tropical and subtropical forests of the lowland (Inskipp and Inskipp 1991, HMG/MFSC 2002). Furthermore, 53% of the country’s nationally threatened bird species inhabit the lowland tropical forest (Bird Life International 2012), of which 56% are only found in the lowland forest (BCN and DNPWC 2010).

Ecological information about the effects of anthropogenic habitat disturbances on avifauna, particularly at multiple scales, is poor in lowland Nepal. This knowledge gap is a challenge when devising strategies for avifaunal conservation for the region. To achieve effective conservation of forests and associated biodiversity in these areas, empirical information about how the remaining extent of forest and other landscape attributes affects to species richness, abundance and community structure of birds across landscapes is required. Thus, understanding the drivers of species richness and assemblages at the landscape-level is critical for appropriate conservation of birds in the region.

30

1.4 Thesis outline As I discussed in the previous sections, the increasing extent and intensity of land-use is a significant cause of biodiversity declines world-wide. In developing countries, forest-use practices in the form of logging, grazing, and lopping are the major forms of habitat disturbances in the remnant forests. However, effects of subsistence forest disturbances on bird assemblages and the role of landscape context in moderating effects of disturbance on birds are poorly understand in lowland Terai forests of Nepal. The effects of properties of landscapes on forest bird assemblages in this region with its particular disturbance regimes are not known. As identified in the previous section, the current understanding of patterns of species richness and forest extent relationships at the functional levels of birds is limited.

This thesis aims to address these knowledge gaps. This section describes how the remainder of this thesis is structured to answer the three research questions related to the overall aim of this study. The next three chapters (2-4) are presented as a set of stand-alone journal articles, each of which is either published or in review. Each chapter addresses a specific research question that relates to the broader research aim of this thesis.

Chapter 1: Introduction and literature review This chapter includes the thesis overview and general introduction and overview of the problem, with aims and specific objectives of the study. I summarise the literature from a broad range of topics relevant to this thesis. I concluded this chapter with a brief overview of threats to birds, knowledge gaps and the need to acquire ecological information for conservation of avifauna in the region.

Chapter 2: Bird conservation values of off-reserve forests in lowland Nepal In this chapter, I evaluated whether off-reserve forest supports bird assemblages complementary to those of protected area. I compared species richness, abundance, diversity and community composition of birds among three management tenures. The findings from this chapter support the hypothesis that off-reserve forests in particular community forests have complementary bird assemblages to that of protected areas. This chapter has been published in Forest Ecology and Management in 2014 (Volume 323).

Chapter 3: Impacts of extractive forest uses on bird communities vary with landscape context in lowland Nepal

31

In Chapter 2, I used site-scale habitat characteristics to evaluate the conservation values of off-reserve forests and protected areas for bird communities. As I highlighted in previous sections, lowland forests are subjected to subsistence forestry practices, which, in turn, may significantly change the habitat structure and associated avifauna. Therefore in Chapter 3, I investigated effects of forest disturbances on bird communities, and whether landscape context—specifically, forest extent within a 500 m and a 2500 m radius of survey sites—can moderate these effects on forest bird assemblages. I then examined the relative importance of site-and landscape-scale forest habitat characteristics on bird species richness and abundance. This study demonstrated that structural features of forest stands such as canopy cover, tree sizes and shrub density were important influences on avifaunal assemblages. I found that while forest use practices significantly affected the avifaunal community of sites, the intensity of these effects can be moderated by maintaining the forest cover in the landscape. This chapter has been accepted for publication in Biological Conservation (accepted 11 February, 2015).

Chapter 4: Relationships between landscape-level bird species richness and forest extent vary among guilds. In Chapter 3, I found strong positive relationships between forest bird assemblages and both site and landscape characteristics. I also found that extent of forest within a 500 m and a 2500 m radius of survey sites was important to moderate impacts of forest disturbance on birds at the site-level. However in Chapter 4, I extended my research approach beyond the site/landscape context design. Here, I investigated the relative importance of forest extent and other landscape characteristics on estimated richness at the landscape-level of all birds, frugivores, foliage-gleaners and sallying insectivores. In this Chapter, I found that although the relationship between species richness and forest extent was strong, the strength and magnitude of the relationship varied considerably among foraging guilds. The relationship between species richness and extent of forest cover in the landscape was non-linear, with landscape-level species richness declining more steeply when forest cover in the landscape fell below 20-30%. This chapter will be submitted to Diversity and Distributions in March 2015.

Chapter 5: Synthesis and conclusions This chapter synthesises the findings of the previous chapters and explores how each contributes to our understandings of multi-scale effects of habitat characteristics and forest

32 disturbances on bird assemblages in human dominated forested landscapes. I have also highlighted future research priorities for the region, and for relevant areas of the field of landscape ecology.

33

CHAPTER 2 BIRD CONSERVATION VALUES OF OFF-RESERVE FORESTS IN LOWLAND NEPAL

Plate 2: Regeneration of forest after implementing a community forestry program in Barandabhar Community forest, lowland Nepal.

Published as: Dahal, B. R., C. A. McAlpine, and M. Maron. 2014. Bird conservation values of off-reserve forests in lowland Nepal. Forest Ecology and Management 323:28-38.

34

2.1 Abstract Although protected areas are central to global biodiversity conservation, off-reserve forests are increasingly recognized as potentially important for the long term conservation of biota, particularly in less-developed countries where communities rely directly on resources from natural areas. We assessed the conservation value of differently managed forests for birds in lowland tropical forests of Nepal. In particular, we explored whether their conservation value was additional or complementary to those of formal protected areas. Using data collected from 112 sites in protected areas (n = 31), state managed forests (n = 37) and community managed forests (n = 44), we assessed how bird species richness, abundance, diversity and community composition varied among tenures. Although sites in protected areas had the greatest species diversity, community managed forests supported a complementary assemblage. Of 124 species recorded, only 45% were common to all management tenures. Overall, the distinctiveness and richness of species in sites in forests outside of protected areas contributed substantially to regional avifaunal diversity. These results highlight the potentially critical role of appropriately managed community forests. The maintenance of diverse bird assemblages in forest regions depends on complementary management of forests both outside and inside the established protected areas.

Key word: Bird community; forest management; conservation value; tenure arrangement; community participation

35

2.2 Introduction In the face of growing pressure on global biological diversity, the protected area network is increasingly important for biodiversity conservation worldwide (Joppa et al. 2008, Jenkins and Joppa 2009). Protected areas have been important in maintaining extensive primary forests and protecting species of high conservation value (Lee et al. 2007). However, there are concerns regarding the adequacy of protected areas in terms of representation of species and their habitats (Rodrigues et al. 2004). Recent work has highlighted limitations of protected areas in maintaining key biodiversity features in landscapes (Laurance et al. 2012, Clark et al. 2013). With the conservation focus primarily on particular areas, biodiversity conservation in surrounding landscapes can be neglected (Bhagwat et al. 2005, Hansen and DeFries 2007).

There has been increasing interest in the importance of forests outside the protected areas for biodiversity conservation (Bhagwat et al. 2005, Persha et al. 2010). Off-reserve forests can be important reservoirs of biodiversity that are complementary to the existing protected area network in several ways. For example, off-reserve forests are often in different vegetation types to those within protected areas, providing habitat resources that are poorly represented within protected areas (Cox and Underwood 2011). For instance, tropical moist deciduous and semi-evergreen forest in south (Persha et al. 2010), evergreen mixed deciduous forests in (Tantipisanuh and Gale 2013), and natural sacred forests in India (Bhagwat and Rutte 2006) are predominantly represented outside of reserves, where community management initiatives appear important for biodiversity conservation. Such forests can also have greater habitat heterogeneity due to different disturbance regimes, therefore supporting species that use various successional stages of habitat (Brawn et al. 2001, Chandler et al. 2012). As no single habitat necessarily provides all the required resources for a given species’ persistence (Saunders et al. 1991, Becker et al. 2007), conservation management of off-reserve forests can be essential for the persistence of many species (Sodhi and Ehrlich 2010). Thus, effective off-reserve conservation policies help ensure a diversity of habitat resources across the landscapes in which protected areas are embedded.

In developing countries, about 22% of the total forest area is either community-managed or owned, compared with only three percent in developed countries (White and Martin 2002). Community forest initiatives have been increasingly successful in preventing deforestation

36 and restoration of forest condition in the landscapes (Klooster and Masera 2000; Nagendra and Gokhale 2008, Porter-Bolland et al. 2012). Nepal offers some of the best examples of community-based forest management in the world (Pokharel et al. 2007, Nagendra and Gokhale, 2008). About one-fourth of forests in Nepal are currently managed by community forest user groups (Kanel and Dahal 2008, Ojha et al. 2009). Rates of habitat loss and degradation are reduced in community managed areas compared with state managed forests (Nagendra 2007, Kanel and Dahal 2008). This success is primarily driven by effective implementation of a decentralized forest management regime (Ojha et al. 2009). This approach has increased active participation of local communities in conservation of resources as they perceive community forest regime as a secure tenure for sustainable sources of forest products (Gautam 2007, Kanel and Dahal 2008).

Protected areas and off-reserve forests differ in their habitat features due to different management approaches, so it is likely that species richness and composition of birds may vary among sites in different forest tenures. For example, landscapes with relatively undisturbed and structurally complex forest habitat within the formal protected area systems may support forest specialists and disturbance-intolerant species (Inskipp and Inskipp 1991, BCN and DNPWC 2011). However, successional habitats of different stages within the community managed forests may support more open country specialists (Roman 2001) and mixed habitat species (Anand et al. 1997, Blake and Loiselle 2001). As differently-managed forests offer habitat heterogeneity, appropriate conservation of these habitats can optimize regional bird diversity. It is therefore important to quantify the contribution of these areas to biodiversity conservation, and in particular, the extent to which the biota they support is complementary to that within formal protected areas.

In this study, we examine the contribution of differently-managed forests to the conservation of forest bird communities in lowland Nepal. The role of alternative forest management tenures in biodiversity conservation is often neglected. In particular, while state-centric forest management approaches tend to have spatially uniform management approaches, community management approaches can be diverse, while also securing the right to resources and embracing a participatory approach to the management of forest resources. Several studies have investigated the effectiveness of the community forestry approach in reducing deforestation and improving plant species richness and density (e.g. Agrawal et al. 2008, Sapkota et al. 2009, Persha et al. 2010), However, it is important to examine whether

37 community managed forests in particular can play an important conservation role for forest bird communities, complementing that of protected areas. We specifically aimed to: 1) determine whether species richness, abundance and diversity of forest bird assemblages varied among sites in community forests, state forests and protected areas, and 2) compare the composition of forest bird assemblages among different management regimes to assess conservation values of variously managed off-reserve forests for avian biodiversity in Nepal’s lowland landscapes.

38

2.3 Material and methods 2.3.1 Study area The study was conducted in the eastern and central Terai of Nepal. Nepal has a total landmass of 147,181 km2 divided among three main geographical regions: the Himalayan region, mid hill region and the Terai region. The lowland Terai encompasses most of the country’s tropical moist forest from the Mechi River in the east to the Narayani River in the centre. The area is characterized by a tropical climate, with average precipitation of approximately 1,800 mm (Springate-Baginski et al. 2003) and mean maximum temperatures of 15-400C (Sah et al. 2002). Before 1950, the region was an uninterrupted patch of dense tropical forest. With the eradication of malaria in the early 1950s, the highly productive lowland zone of the country was settled and subsequently agricultural expansion occurred (Hrabovszky and Miyan 1987). Consequently, most of the forest was destroyed and remaining forest areas were subjected to intense human exploitation. Nearly half of the country’s population lives in the 17% of the country that is lowland (Central Bureau of Statistics 2011).

The government of Nepal introduced and implemented forest legislation in 1978 with the aim of diversifying the management tenures and reducing large-scale clearance of forest (Department of Forest 2009). Thus, forests in Nepal are now managed under three major regimes: as state managed forests (forests managed by the central government), community managed forests (forests managed by local forest user groups), and protected forests (IUCN management categories I-IV). The state-managed forests are those that are managed by the Department of Forests as production forests, and protected areas are managed by Department of National Parks and Wildlife Conservation for conservation. Protected areas in this study corresponded to IUCN protected area management categories II (National Park) and IV (Wildlife Reserve). The main aim of category II is to maintain ecological integrity at ecosystem-scale, while the category IV is aimed to protecting habitats and individual species. Approximately 1.2 million hectares of forests are currently managed by the community forest user groups (>15000 community forest user groups) (Kanel and Dahal 2008, BCN and DNPWC 2012) while ~1 million hectares of forests is directly managed by the central government (Shyamsundar and Ghate 2011).

39

2.3.2 Study sites A total of 112 sites were selected within lowland tropical forest within an elevational range of 90 – 300 m asl. These sites were allocated among three management tenures in approximate proportion to the available area within each. We randomly allocated survey sites within forests of each tenure type using digital vegetation mapping data. Initially, we chose 128 sites using a GIS, but based on accessibility, we ended up with 44 sites within community managed forests, 37 within the state managed forests and 31 within protected areas, including in Chitwan National Park (IUCN category II protected area) and its buffer zone forest of Barandabhar corridor, Parsa Wildlife Reserve (IUCN category IV protected area) which have been managed for conservation for more than twenty-five years (Baral and Inskipp 2005). The southern part of the Barandabhar core forest is managed by the park authority; its peripheral areas are community-managed forests. Geographically, 60 sites were located in the eastern landscapes (Eastern Terai forests) and 52 sites were located in the central lowland landscapes (Parsa and Chitwan forests). The vegetation of the lowland Terai is mainly consisted of Shorea robusta mixed forest. Therefore, all sites were located within the same vegetation type. All sites were located at least 500 m from roads to minimize any road induced variation on bird assemblages. The minimum distance between sites was at least 1000 m so as to reduce the chance of spatial dependence.

40

Figure 2.1 The location of the study area and bird survey sites: (a) Chitwan forests (b) Parsa forests and (c) Eastern Terai forests

2.3.3 Bird surveys Each study site comprised a fixed-width belt transect measuring 200 m x 50 m. Study sites were demarcated by placing visible markers at each site and taking GPS coordinates. Each transect was surveyed on three occasions between November 2012 and May 2013, across two seasons. This arrangement of bird survey allowed us to capture winter visitors, winter migrants and early summer visitors. Each site was visited on two occasions in winter and one in summer. On each visit, the observer (BRD) recorded all birds seen or heard within 25 m of the centreline of the transect while walking along its length over a 10-min period.

A variety of techniques have been employed to describe the characteristics of bird populations. These include radio-telemetry (Powell et al. 2005), colour banding (Powlesland et al. 2000, Rodewald et al. 2013), distance sampling (Buckland 2001), fixed-radius point- counts (Gregory et al. 2004, Buckland 2006) and fixed-width transect-counts (Bibby et al. 2000, Westbrooke et al. 2003, Maron and Kennedy 2007). Both fixed-radius point-count and fixed-width transect-count methods are widely used to describe the species richness, relative

41 abundance and densities of birds (Manuwal and Carey 1991, Buckland 2006Gregory et al. 2004). We therefore used fixed-width transect counts in order to compare the relative abundance and richness of birds per unit area among different sites (Manuawal and Carey 1991, Bibby et al. 2000, Hostetler and Main 2011).

Surveys were conducted only between 0600 and 1100 hours in the morning and 1400 to 1745 in the afternoon. Although we did not test for effects of time of day on bird observation prior to actual field survey; several other studies reported that the detection rate of most bird species is greater in morning (Bried et al. 2011) with another peak in activity in the late afternoon, 2-3 h before sunset (Kessler and Milne 1982). Generally birds tend to avoid the midday heat (Pizo et al. 1997), therefore we surveyed birds within 4 h after sunrise and within 3.45 h before sunset. To avoid possible bias, we standardized the survey protocol in such a way that although not all sites had afternoon surveys, this occurred equally among site categories, and so no bias was introduced due to this. All surveys were conducted by the same observer during fair weather at no heavy rain and wind.

2.3.4 Explanatory variables Data on vegetation and habitat structure were collected at each bird survey transect. Using four randomly-located 20 m x 20 m quadrats, the percentage of tree canopy cover was estimated, the number of trees counted, and their diameters measured within the 20 m x 20 m quadrat. Tree cover was estimated visually (Pattison et al. 2011). We divided the quadrat into quarters, and assessment of tree canopy cover was determined by two observers for each quarter. The cover values for each quarter were then averaged and the four mean values for each quadrat averaged, before a grand mean was calculated for the site. Nested within each of the 20 m x 20 m tree quadrats was a 5 m x 5m quadrat, used to collect understorey vegetation data. The shrub cover and number of individual shrubs were collected within each of these nested quadrat and the grand mean taken for each transect. We calculated the total area of forest habitat (in ha/km2) within a 500 m buffer distance from each bird survey sites in a GIS (using ArcGIS 9.3). We included both primary and old growth regenerating forests to classify forest habitat based on land-cover data provided by WWF Nepal (WWF 2005). Stand and landscape-scale explanatory variables used in analyses are summarised in Table 2.1.

42

Table 2.1 Summary of explanatory variables used to assess influence of stand and landscape-scale variables on bird communities. Variables Unit Description Forest extent Hectares/km2 Amount of forest area in 500 m radius of survey site.

Tree canopy cover Percent Mean percentage cover of all tree crowns in 200 m x 50 m line transects. All trees density Count/m2 Number of trees with >10 cm diameter at breast height (DBH) per square metre. Mature large trees Count/m2 Number of trees with >100 cm DBH per square metre. density Shrub cover Percent Mean percentage cover of all shrubs <2 m tall, in 200 m x 50 m line transect. Shrub density Count/m2 Number of individual shrubs <2 m tall, per square metre.

2.3.5 Statistical analyses 2.3.5.1 Univariate analysis We compared average abundance per survey, species richness, evenness (as measured by Pielou evenness index) and species diversity (as measured by Shannon diversity index) of bird assemblages among sites within the three management tenures using a one-way analysis of variance (ANOVA) followed by Tukey’s Honestly Significant Difference (HSD) test. We also compared abundance, species richness, evenness and diversity of birds with different habitat associations (forest specialists, forest generalists and forest edge specialists) and within different foraging guilds among the management tenures. Due to large number of zero values for diversity and evenness of forest specialists and bark gleaning insectivore, we only included species richness and abundance of these groups in statistical comparisons. Membership of habitat groups and foraging guilds was identified based on primary habitat specialization and diet information compiled from Ali and Ripley (1983) and Grimmett et al. (2009). In addition, we compared habitat characteristics among different management tenures. All analyses were carried out using the R statistical package version 3.1 (R Development Team 2012).

43

2.3.5.2 Multivariate analyses Multivariate data analyses were employed to examine variation in bird assemblage composition among forest management regimes. We used nonparametric, permutational multivariate analysis of variance PERMANOVA (Anderson 2001) based on Bray-Curtis similarity values from species abundance matrix to test whether bird community composition differed across management tenures. In this analysis, management tenures were considered factors and sites were samples. Bird abundances were square-root transformed prior to analysis to reduce the influence of highly abundant species. The test statistic for PERMANOVA is the pseudo F-ratio, where a large pseudo F-ratio indicates that sites in different management tenures are differed in bird community composition in multivariate space. We performed additional pairwise PERMANOVA tests (Anderson et al. 2008) to explore the extent of differences in species composition between the management tenures.

A significant pseudo-F ratio with P-values from the PERMANOVA can indicate a difference in community composition between treatments due either to differences in the location of the treatment communities in multivariate space or to differences in dispersion of communities in multivariate space within the treatments (Anderson et al. 2008). Thus, we used a complementary analysis, the permutational analysis of multivariate dispersions (PERMDISP) (Anderson et al. 2006) to test for differences in the homogeneity of multivariate dispersion among sites in different management tenures. Following a finding of significant differences in dispersion, we performed pairwise tests. We also calculated mean species turnover (i.e. beta diversity) using the vegan package in R (Oksanen et al. 2011) for each forest management tenure. Beta diversity is the rate of change of species composition among sites or habitat patches (Whittaker 1972). Understanding beta diversity among sites across different management tenures allowed us to quantify their contribution to the regional avifaunal diversity. We also evaluated which species were most responsible for differentiating communities using similarity percentage (SIMPER) analysis. SIMPER evaluates the contribution of each species to the Bray-Curtis dissimilarity of all pairs of samples between groups (Clarke and Warwick 2001). These analyses were carried out using PRIMER v6 with PERMANOVA+ add-on software (Anderson et al. 2008).

Non-metric multidimensional scaling (NMDS) ordination was conducted using Bray-Curtis similarities to visualize pattern of bird assemblage among management tenures (Clarke 1993). Ordination serves to summarize community data by producing a low-dimensional

44 ordination space in which distance between species and samples sites reflect the ecological differences between them (Gauch 1982). We performed a vector-fitting routine using the vegan package in R (Oksanen et al. 2011) to examine the relationship between bird communities with environmental variables. A vector-fitting protocol allows us to visualize the most contributing variables to the pattern of bird community composition (Johnson et al. 2007). The length of the fitted arrow is proportional to the correlation between the ordination axis and the environmental variable. Environmental variables were standardized to lie between 0 and 1 prior to analysis.

2.4 Results 2.4.1 Species richness, abundance and diversity A total of 124 bird species from 28 families were recorded across all sites over three seasons (Appendix A: Table A.1). Of all recorded species, 68% were local residents, 16% winter visitors, 2% summer visitors and 4% winter passage migrants. In all three surveys, the most widespread species was the black-hooded oriole Oriolus xanthornus, which occurred at 110 of the 112 survey sites. The grey-headed canary flycatcher Culicicapa ceylonensis, spangled Dicrurus hottentottus, and jungle babbler Turdoides striata, were the next most widespread species, each being recorded 80%, 79%, and 78% of sites, respectively (Appendix A: Table A.1).

Ninety-three species were recorded in community managed forests, 87 species were recorded in the state managed forests, and 80 species were recorded in protected areas. Some species of birds were distinct to each of the forest management regimes. Seventeen species were recorded only in sites in community managed forests, 13 species were in sites in state managed forests, and 15 species were in sites in protected areas. Only 45% of the recorded species were found in all three management tenures.

Among the three forest management tenures, neither the total bird abundance (F2, 109 = 0.83,

P > 0.05), total species number (F2, 109 = 2.44, P > 0.05) nor Pielou evenness index (F2, 109 = 0.26, P > 0.05) differed significantly. However, there was a statistically significant difference in species diversity (Shannon diversity index; F2, 109 = 4.79, P < 0.05) among management tenures (Table 2.2). A post-hoc Tukey test showed that differences in species diversity were between state forests and community forests (P < 0.01), and between state forests and

45 protected areas (P < 0.05). There was no significant difference in species diversity between sites in community forest and protected areas (P > 0.05).

Table 2.2 Mean species richness, average abundance per survey, Shannon diversity index and Pielou evenness index (±SE) of all birds, in community managed forests (n = 44), state managed forests (n = 37) and protected areas (n = 31). Variables Community forest State forest Protected area F value Species richness 19.2 ± 0.5 17.5 ± 1.1 20.0 ± 0.3 2.47 Abundance 23.3 ± 1.3 23.0 ± 1.6 25.2 ± 1.3 0.83 Shannon diversity index 2.6 ± 0.0 2.4 ± 0.1 2.7 ± 0.1 4.79* Pielou evenness index 0.9 ± 0.0 0.9 ± 0.1 0.9 ± 0.1 0.26 *Significant at P < 0.05

2.4.2 Habitat preference groupings We recorded a total of 14 species of birds classified as forest specialists, 30 forest edge species and 75 forest generalist species. Overall, significantly higher mean species richness and abundance of forest specialist birds were observed in sites in protected areas followed by community managed forests and state managed forests (Table 2.3).

Table 2.3 Mean species richness, average abundance per survey, Shannon diversity index and Pielou evenness index (±SE) of all birds in different habitat in community managed forests (n = 44), state managed forests (n = 37) and protected areas (n = 31). Degrees of freedom between groups (management types) = 2 and within groups = 109 in all cases. Habitat Community State Protected Variables F value group forest forest area Species richness 11.5 ± 0.4 10.2 ± 0.4 11.3 ± 0.5 3.22* Forest Abundance 15.3 ± 0.8 15.5 ± 1.0 16.2 ± 0.2 0.58 generalists Shannon diversity index 2.2 ± 0.0 1.9 ± 0.0 2.1 ± 0.1 4.32* Pielou evenness index 0.9 ± 0.0 0.9 ± 0.0 0.9 ± 0.0 1.35 Species richness 6.7 ± 0.3 5.9 ± 0.5 6.4 ± 0.4 0.84 Forest edge Abundance 7.5 ± 0.7 6.6 ± 0.7 7.2 ± 1.7 0.31 species Shannon diversity index 1.6 ± 0.0 1.4 ± 0.1 1.6 ± 0.1 3.02

46

Pielou evenness index 0.9 ± 0.0 0.8 ± 0.0 0.9 ± 0.0 5.28** Species richness 2.2 ± 0.2 2.1 ± 0.3 2.7 ± 0.3 5.91** Forest Abundance 1.3 ± 1.0 0.8 ± 0.2 1.9 ± 0.2 6.64** specialists Shannon diversity index 0.4 ± 0.0 0.3 ± 0.0 0.7 ± 0.1 -

Pielou evenness index 0.9 ± 0.0 0.8 ± 0.0 0.9 ± 0.0 - Significant at *P < 0.05; **P < 0.01

2.4.3 Foraging guilds The sites in protected areas had the greatest abundance, species richness and diversity of bark gleaning insectivore, followed by community forests and then state forests (Table 2.4). However, other foraging guilds varied little among management tenures (Table 2.4). While we identified seven main foraging guilds of birds, we excluded those groups that have large number of zero values in the analyses, leaving only four that could be analysed (Table 2.4).

Table 2.4 Mean species richness, average abundance per survey, Shannon diversity index and Pielou evenness index (±SE) of different foraging guilds of birds in community managed forests (n = 44), state managed forests (n = 37), and protected areas (n = 31). Degrees of freedom between groups (management types) are 2 and within groups are 109 in all cases for food guild. Community State Protected Food guild Variables F value forest forest area Bark Species richness 2.9 ± 0.3 2.5 ± 0.2 3.8 ± 0.3 7.58*** gleaning Abundance 3.5 ± 0.4 2.2 ± 0.3 4.5 ± 0.4 10.37*** insectivore Shannon diversity index 0.7 ± 0.1 0.7 ± 0.0 1.1 ± 0.0 - Pielou evenness index 0.9 ± 0.0 0.8 ± 0.1 0.8 ± 0.0 - Foliage Species richness 2.7 ± 0 .2 3.3 ± 0.3 3.3 ± 0.2 1.93 gleaning Abundance 3.2 ± 0.4 4.7 ± 0.7 4.5 ± 0.5 2.51 insectivore Shannon diversity index 0.7 ± 0.1 0.8 ± 0.1 0.9 ± 0.1 1.86 Pielou evenness index 0.8 ± 0.0 0.8 ± 0.0 0.8 ± 0.0 1.71 Species richness 4.0 ± 0.2 3.3 ± 0.3 3.3 ± 0.2 3.90* Sallying Abundance 3.5 ± 0.4 2.8 ± 0.3 3.0 ± 0.3 2.25 insectivore Shannon diversity index 1.2 ± 0.1 0.9 ± 0.1 1.0 ± 0.1 2.78 Pielou evenness index 0.9 ± 0.0 0.9 ± 0.0 0.9 ± 0.0 2.36 Frugivore Species richness 5.9 ± 0.3 5.4 ± 0.3 5.3 ± 0.3 1.22

47

Abundance 9.5 ± 0.7 8.8 ± 0.8 8.1 ± 0.6 0.8 Shannon diversity index 1.5 ± 0.0 1.4 ± 0.1 1.3 ± 0.1 1.17 Pielou evenness index 0.9 ± 0.0 0.9 ± 0.0 0.8 ± 0.0 1.06 Significance at*P < 0.05; **P < 0.01;***P < 0.001

2.4.4 Bird assemblage structure Significant differences in community composition of birds were observed among sites in all three forest management tenures (PERMANOVA: F = 3.56, P < 0.001). The PERMANOVA pairwise a posteriori comparison tests showed that community composition differed between each pair of tenures (P < 0.05, Table 2.5). The greatest difference in community composition was between sites in community forests and those in protected areas (P < 0.001), and protected areas and state forests (P < 0.001). PERMDISP analyses for the homogeneity of multivariate dispersions also showed significant differences in community composition among the forest management tenures (F = 5.56, P < 0.01). However, no significant difference in multivariate dispersion was observed between protected areas and community forests. Thus, the significant difference in community composition in sites between community forests and protected areas from the PERMANOVA was due to differences in the location of sites of community forests and protected areas in multivariate space rather than differences in dispersion around the mean composition within sites in the community forests and protected areas (Table 2.5). The multivariate dispersion in community composition in sites in state managed forests was significantly larger, having an average Bray-Curtis distance-to-centroid over 44% greater than sites in protected areas (39%) and community forests (39%). A separate beta diversity (i.e. species turnover) analysis showed that the mean variation in species composition was highest among sites in state managed forests (mean β diversity = 0.55 ± 0.1) than among sites in sites in protected areas (0.48 ± 0.1) and community managed forests (0.48 ± 0.1).

Table 2.5 Results of PERMANOVA and PERMDISP pairwise comparison tests of bird community composition in response to different forest management tenures. PARMANOVA test PERMDISP test Management tenures t-value t-value Community forest - Protected area 2.22*** 0.24

48

Community forest - State forest 1.32* 3.23*** Protected area - State forest 2.07*** 2.56*** Significance at*P < 0.05; **P < 0.01;***P < 0.001

Non-metric multidimensional scaling (NMDS) ordination of species composition showed strong clustering of sites according to forest management tenures, although stress was relatively high (Fig. 2.2). Vector fitting of environmental variables to the bird assemblage NMDS ordination showed significant correlation of environmental variables (Fig. 2.2). Among the environmental variables, the amount of forest within a 500 m radius had a significant influence on bird community structure (P < 0.001). Significant correlations were found with mature tree density (P < 0.01), tree canopy cover (P < 0.01), mean shrub cover (P < 0.01) and shrub density (P < 0.5) in the study area.

Figure 2.2 Bird assemblage NMDS (2D stress = 0.24) plot fitted with environmental vectors. Vectors represent the mean direction and strength of correlation of different environmental

49 variables (Fextent = Forest extent, Scv = Shrub cover, Sden = Shrub density, Mtdn = Mature tree density, Tcv = Tree canopy cover). Vectors with significance P < 0.05 only are shown. The SIMPER analyses identified the species responsible for distinguishing community composition between management tenures. The rose-ringed parakeet Psittacula krameri contributed most to the dissimilarity between sites in protected areas and community forests, being more common in sites in community forests, while the jungle babbler Turdoides striata contributed most to the dissimilarities between sites in state forests and community forests, and between protected areas and state forests, as it was commonest in sites in protected areas and community forests. The top ten bird species that contributed to differences between management tenures are presented in Appendix A: Table A.2.

2.4.5 Habitat characteristics There were significant differences in habitat characteristics among different management tenures. Habitat characteristics such as tree density (F2, 109 = 6.47, P < 0.01), shrub cover (F2,

109 = 4.96, P < 0.01), tree canopy cover (F2, 109 = 3.58, P < 0.05) and mature tree density (F2,

109 = 2.91, P < 0.05) differed significantly among sites in all three management regimes.

However, we found no significant difference in shrub density (F2, 109 = 2.05, P > 0.05) among management tenures. Sites in state managed forests had the highest average number of small trees (<20 cm diameter DBH), while community forests and protected areas had highest average number of large mature trees (>100 cm DBH) (Fig. 2.3).

Community forest 0.05 0.05 Protected area 0.04 0.04 State forest 0.03 0.03 0.02 0.02 0.01

Tree density (stems/m2) density Tree 0.01 0.00 >10 dbh 20-49 dbh 50-99 dbh >100 dbh Tree size classes (cm)

Figure 2.3 Distribution of tree size classes (±S.E.) in sites in community forests, state forests and protected areas.

50

2.5 Discussion While sites in protected areas had the greatest richness of birds, community forests and state managed forests had complementary assemblages, supporting species not represented in formal conservation reserves. In this study 17 species of birds were recorded only in community forests, among which Abbott’s babbler Malacocincla abbotti and blue-eared barbet Megalaima australis are nationally threatened (Inskipp 1989, Baral and Inskipp 2004; Inskipp et al. 2013). In total, 24% of species were found only in forests outside the protected areas, and only 45% of species were common to all forest tenures. This distinctness of bird assemblages in off-reserve sites contributes to diversity in Nepal’s lowlands.

Overall bird community composition differed among the three management tenures, indicating that differently managed off-reserve forests support distinctive bird assemblages. While sites in community forests and state forests had complementary bird assemblages, protected areas emerged as particularly valuable, with significantly greater species richness and diversity of forest specialists and bark-gleaners. The greater community indices of forest specialists and bark-gleaners sites in protected areas are likely to reflect the larger extent of mature forests within the Terai protected areas (Wikramanayake et al. 2004). Furthermore, the higher richness and diversity of forest specialists in sites in protected areas may be related to the fact that anthropogenic disturbance is limited in such areas (Baral and Inskipp 2005). Several studies in the region show that extraction of fodder, firewood and non-timber forest product can negatively influence bird communities (Shahabuddin and Kumar 2007, Dahal et al. 2009, Kumar et al. 2011; Inskipp et al. 2013). Disturbance-intolerant species may therefore be benefited by strict forest management that restricts the removal of standing dead trees, fallen timber for firewood and canopies by pruning.

On several measures, including overall Shannon diversity and the species richness and abundance of bark-gleaners and forest specialists, sites in community forests were most similar to those in protected areas. This similarity is likely to reflect the relatively more similar habitat characteristics of sites in protected areas and community forests, with no differences in tree canopy cover, large mature tree density and shrub density between sites in community managed forests and protected areas. Similarities in habitat structure between community forests and protected areas have also been noted by Timilsina et al. (2007), who found that community forests and protected areas had similar tree density in Nepal’s western lowlands, and Nagendra (2002) who reported similar tree and sapling biomass between

51 protected areas and community forests. A recent study by Persha et al. (2010) in three south Asian countries; Nepal, India and found that tree species richness in community forests was equivalent to that of protected areas. This demonstrates the potentially valuable contribution of community forests to provision of habitat of similar quality to that within protected areas.

Although sites in community forests and protected areas had higher bird diversity, state managed forests had the greatest multivariate dispersion. This reflects higher species turnover among sites in state managed forests. A comparatively high species turnover among sites in state managed forests may be due to number of factors. First, forest management practices such as selective logging, thinning and pruning are commonly employed in much of the state managed forests. Such silvicultural practices cause temporary increases in habitat heterogeneity and this can increase spatial and temporal variation in species richness and abundance of birds (De La Montana et al. 2006, Forsman et al. 2010). Second, of 80 species of birds recorded in sites in state managed forests, 28% were singleton records (i.e. species represented by detection of a single individual during sampling) compared to 15% each for protected areas community forests (Appendix A: Table A.1). These results suggest that, though state managed forests had complementary bird assemblages, these forests may not necessarily provide suitable habitats for long-term persistence of populations of several species we detected.

While overall species richness and abundance of birds in sites in community forests and state managed forests were similar, bird assemblage structure was significantly different. In this study only 57% of species were common to sites in community managed forests and state management forests. In particular, forest specialists and bark gleaners were less abundant in sites in state managed forests than those in community managed forests. This is likely to reflect the altered habitat conditions in state managed forests (Kanel and Dahal 2008). Large mature trees are removed from such forests for their timber, leaving a high density of small trees in sites (Fig. 2.3). In contrast, community-managed forests often include arrangements to reforest and restore degraded forest, which can contribute to maintaining habitat for specialist and rare bird species in Terai forests (Baral and Inskipp 2005).

Among the dietary groups, bark gleaners differed most markedly in their richness, abundance and diversity among the management tenures. The greater abundance of bark-gleaning

52 insectivores in sites in protected areas and community forests may be linked to both the greater extent of forest surrounding such sites, and the higher density of large trees. Several other studies have found that bark gleaning birds are particularly sensitive to habitat alteration (Adams and Morrison 1993, Zurita and Bellocq 2012, Inskipp et al. 2013), and are strongly correlated with large tree density (Cleary et al. 2007, Greve et al. 2011). Furthermore, forest resource extraction like harvesting of standing dead trees and fallen timber for firewood may reduce foraging and nesting sites for many bark-foraging species. Fallen timber and dead trees are linked to nesting success of bark foraging birds including the greater Chrysocolaptes lucidus, grey-headed Picus canus and greater yellownape Picus flavinucha (Kumar et al. 2011). Bark-foraging bird communities may therefore be benefited by forest management that retains higher densities of large and dead trees as foraging and nesting habitats.

2.6 Management implications While the lowlands of the Terai in Nepal contain about 28% of country’s forested areas (Sinha 2011), less than 10% of the lowlands forests are formally protected. Nearly half of the country’s population lives in the lowlands (Central Bureau of Statistics 2011) and the majority of them rely directly on resources from their surrounding natural areas. For instance, about 80% of the total energy consumed in the country is produced by fuel wood, extracted from these forests (Gurung et al. 2011).

Although protected areas serve as essential refugia for most species of forest birds, the further extension of protected areas in lowland landscapes is likely to be limited due to economic, political and social factors. Instead, participation of a wide spectrum of stakeholders and institutions in forest management, such as through community forest arrangements, can be used to complement the contribution of protected areas to biodiversity conservation. The community forestry program in Nepal has demonstrated its value for improving forest conservation outside the protected areas (Kanel and Dahal 2008, Nagendra et al. 2008, this study). Yet, there has been reluctance to transfer management rights to local communities in the lowland Terai region due to the prevalence of forests with high economic value (Bhattarai 2006). Thus, only 10% of the Terai forests have transferred to community management (Kanel and Dahal 2008), compared to 24% in the hill regions of Nepal (Bhattarai 2006). Strengthening the community forestry programs across the off-reserve forests in lowland landscapes can not only ameliorate habitat loss and degradation (Gautam et al. 2004, Kanel

53 and Dahal 2008) but also generate livelihood opportunities for surrounding communities and reduce pressure on protected areas (Straede et al. 2002). Hence, the maintaining regional bird diversity in lowland forest landscapes critically depends on complementary management of forests both outside and inside the established protected areas.

54

CHAPTER 3 IMPACTS OF EXTRACTIVE FOREST USES ON BIRD ASSEMBLAGES VARY WITH LANDSCAPE CONTEXT IN LOWLAND NEPAL

Plate 3: Creekbed in a large patch of forest adjacent to an agricultural area in Parsa Wildlife Reserve, lowland Nepal.

Published as: Dahal, B. R., C. A. McAlpine, and M. Maron. 2015. ‘Impacts of extractive forest uses on bird assemblages vary with landscape context in lowland Nepal’. Biological Conservation. 186: 167-175.

55

3.1 Abstract Forest use practices such as logging, lopping of tree branches for fodder, and grazing do not reduce forest area but disturb forest structure and impact biodiversity. Although such forest disturbances can be key determinants of the biota occupying a site, rarely is the interaction between disturbance intensity and landscape context considered, despite its relevance to conservation management. We investigated the influence of site-and landscape-level habitat characteristics on birds, and explored whether the effects of site-level disturbance on bird richness varied with forest extent in lowland landscapes in Nepal. While extractive uses reduced forest structural complexity and altered the avifaunal community of a site, the intensity of such effects depended on the extent of forest in the surrounding landscape (19.6 km2). The extent of forest, large tree density, and tree canopy cover were important predictors for all bird response groups. However, the effect of forest extent on bird richness was stronger for sites with greater disturbance intensity. Managing and restoring landscapes to support greater forest cover may not only have a positive direct effect on bird conservation, but may also help to compensate for site-level disturbance, such as characterises multiple-use forests worldwide.

Key word: Avian community, multi-use forest landscapes, extractive forest disturbances, landscape context, interactive effect, landscape-level forest cover

56

3.2 Introduction In recent decades, anthropogenic activities have been the principal cause of habitat loss and degradation worldwide (Millennium Ecosystem Assessment 2005, Foley et al. 2005, Ellis and Ramankutty 2008). However, anthropogenic habitat degradation is greater in areas of high population density and poverty. Such areas are mainly in developing countries (Laurance 2010) where a significant proportion of the population live near the forests (Hegde and Enters 2000, Millennium Ecosystem Assessment 2005). About one billion people living in developing countries rely on forest-based products, primarily for subsistence livelihoods (Chao 2012). This has resulted in extensive use of forest resources, for timber and firewood, cutting of tree canopy for fodder, livestock grazing and collection of non-timber forest products (Chettri et al. 2002, Shahabuddin and Kumar 2007, Christensen et al. 2009). Thus, managing forests for biodiversity conservation while satisfying human demands for forest products is a major global conservation challenge (Chappell and LaValle 2011).

Anthropogenic activities can reduce the forest area and also cause significant changes in forest structure and composition (Chettri et al. 2002, Sagar and Singh 2004, Shahabuddin and Kumar 2006). Repeated extraction of timber resources reduces tree basal area, tree height, canopy closure, and regeneration capacity (Sundriyal and Sharma 1996, Mishra et al. 2004, Sapkota et al. 2010). For example, removal of live trees increases light levels in the forest, thereby modifying canopy structure (Sekercioglu et al. 2002, Villela et al. 2006), altering tree density and diversity (Moktan et al. 2009), and changing understorey characteristics (Aleixo 1999, Moktan et al. 2009). The logging and forest extraction practices also change tree diversity and composition (Kumar and Shahabuddin 2005, Berry et al. 2008, Sapkota et al. 2009, Borah et al. 2014). Other forms of extraction of woody biomass such as lopping of tree branches affects canopy structure. Livestock grazing also simplifies the understory forest structure and reduces regeneration, foliage density, canopy height, and vegetation cover (Tasker and Bradstock 2006, Piana and Marsden 2014). Such loss of structural components and alteration of floristic diversity in forests ultimately affects populations of many species reliant on forest habitat (Díaz et al. 2005, Berry et al. 2008, Lee and Carroll 2014).

Habitat variables measured at the site-scale (<1 ha), however, may not be sufficient for meaningful prediction of species responses to disturbance type and intensity. Rather, the local effects of such anthropogenic disturbances on forest fauna may also depend on the landscape context (100s-1000s ha) in which a site is embedded. Sites within more forest cover may

57 offer greater habitat heterogeneity (Heikkinen et al. 2004, Kallimanis et al. 2008), which may potentially govern species richness in the landscape (Tews et al. 2004). Furthermore, increasingly, faunal communities are being shown to be affected strongly by the proportion of forest habitat in a landscape (McGarigal and McComb 1995,Trzcinski et al. 1999, Radford et al. 2005, Ewers and Didham 2006 , Smith et al. 2011). Studies in fragmented landscapes suggest that landscape context - in particular, surrounding forest extent - mediates the effects of fragmentation on faunal communities (e.g Graham and Blake 2001, Deconchat et al. 2009). It is therefore plausible that landscape-scale variables, such as the extent of forest, may actually moderate the effects of site-scale anthropogenic impacts such as subsistence forestry practices on faunal communities, and vice-versa. Yet knowledge of such interactions between the extent of forest in the landscape and the impact of forest disturbances remains limited.

In this study, we first investigated the effects of anthropogenic disturbance on vegetation structure and consequences for bird communities in the lowland Terai forests of Nepal. This region is dominated by the highly productive Sal forests, which are facing significant anthropogenic pressure from extractive and grazing uses. The economy of rural communities in the region is based largely on subsistence agriculture, livestock rearing, and selling of firewood and non-timber forest products (Sharma 1990). Such activities have contributed elsewhere to a decline of many forest bird species (Inskipp et al. 2013, Baral et al. 2014), particularly species with small home range and/or other ecological requirements (Inskipp 1989).

Second, we modelled the effect of the interaction between landscape context and disturbance intensity on the bird assemblages of these forests. We hypothesized that forest disturbances will negatively affect vegetation structure and bird communities, but that disturbance intensity will interact with the extent of forest in the landscape to affect the avifauna of a site. Specifically, our main objectives were to: (1) determine whether the vegetation characteristics and species richness and abundance of forest bird assemblages varied with logging, grazing, and lopping intensity; and (2) assess the relative importance of site-and landscape-scale forest habitat characteristics on bird species richness and abundance and the existence of any interaction effects between disturbance intensity and landscape context.

58

3.3 Material and methods 3.3.1 Study area The study was conducted in southern Nepal, also called ‘Terai’ (80° 4’ 30” to 88° 10’ 19” E 26° 21’ 53” to 29° 7’ 43” N, elevation 63 - 330 m ASL). The Terai encompasses most of the country’s tropical moist forest from the Mechi River in the east to the Narayani River in the centre. The annual rainfall decreases from 2,680 mm to 1,138 mm from east to west, and the mean monthly rainfall ranges from 8 mm in November to 535 mm in July (FRA/DFRS 2014). The area is characterized by a tropical climate, with the maximum monthly mean temperature of 35-40°C in April/May and the minimum, 14-16 °C, in January (Jackson et al. 1994). Before 1950, the region supported continuous dense tropical forest. With the eradication of malaria in the early 1950s, large tracts of the highly productive lowland forests were converted to agriculture (Hrabovszky and Miyan 1987). Consequently, most of the forest was destroyed and the remaining forest areas were subjected to intense human exploitation. Nearly half of Nepal’s population now lives in the 17% of the country that is lowland (Central Bureau of Statistics 2011).

Figure 3.1 Location of the three study regions in Nepal: a) Chitwan forest, b) Parsa-Bara forest and c) Eastern forest.

3.3.2 Study sites and landscapes Twenty-eight landscapes, each 5 km x 5 km, and supporting different amounts of forest cover (7.9% - 95.3%), were selected across south-central (Bara-Parsa forest and Chitwan forest). and south-eastern lowland Terai forests (eastern forests) among three tenure types.

59

Geographically, 15 landscapes were located in eastern Terai forests and 13 landscapes were located in central lowland Terai forests. Four survey sites, each measuring 200 m x 50 m, were randomly located in each landscape, resulting in a total of 112 sites (28 landscapes x 4 sites). Of the total 112 sites, 44 sites were in community manage forests, 37 were in state- managed forests and 31 were in protected areas. The non-forested part of landscapes in this region is mainly comprised of a mixed land–use type that includes rural towns, agriculture and agro-forestry. All sites were located at least 500 m from roads to minimize any road- induced variation in bird assemblages. The minimum distance between sites was at least 1000 m to reduce the chance of spatial dependence.

3.3.3 Bird surveys At each study site, birds were surveyed on three occasions between November 2012 and May 2013 allowing us to capture winter visitors, winter migrants and early summer visitors. Each site was visited on two occasions in winter and one in summer. On each visit, the observer (BRD) recorded all birds seen or heard within 25 m of the centreline of the transect while walking along its length over a 10-min period. Prior to the data collection, we tested for visibility of birds within 50 m and 25 m of the transect, and found that visibility beyond 25 m was challenging. To reduce the risk of sampling bias (Järvinen and Väisänen 1975), we used a rangefinder to help ensure all birds counted were within the fixed belt transect.

Surveys were conducted only between 0600 and 1100 hours in the morning and 1400 to 1745 in the afternoon. Although the effects of time of day on bird observation were not tested prior to the actual field survey; several other studies have reported that the detection rate of most bird species is greater in morning (Bried et al. 2011) with another peak in activity in the late afternoon, 2-3 h before sunset (Kessler and Milne 1982). In general, birds tend to reduce activity during the midday heat (Pizo et al., 1997). To avoid possible bias, we standardized the survey protocol in such a way that, although not all individual sites had an afternoon survey, afternoon surveys occurred equally among sites. All surveys were conducted during fair weather when there was no heavy rain and the wind speed was low.

3.3.4 Explanatory variables Data on vegetation and habitat structure were collected at each bird survey transect (Table 3.1). Four 20 m x 20 m quadrats were randomly located on each transect. For each quadrat, we measured the percentage of tree canopy cover, the number of trees, and their diameter at

60 breast height. Tree canopy cover was estimated following Pattison et al. (2011) by dividing the quadrat into quarters, and visually assessing canopy cover for each quarter. The values for each quarter were then averaged and the four mean values for each quadrat averaged to provide an overall mean for the site. Nested within each of the 20 m x 20 m quadrats was a 5 m x 5 m quadrat, used to collect understorey vegetation data. Shrub cover estimates and the number of individual shrubs were collected within each nested quadrats and an overall mean calculated for each transect. Similarly, herbaceous cover was estimated from a 1 m x 1 m quadrat, nested within each of the 20 m x 20 m tree quadrats. The extent of forest habitat (in ha/km2) within a 0.5 km and a 2.5 km buffer distance from each bird survey sites was calculated using GIS (using ArcGIS 9.3). We included both primary and old growth regenerating forests to classify forest habitat based on land-cover data provided by WWF Nepal (WWF 2005). These spatial extents, while arbitrary, were chosen areas as likely to represent the likely daily area of use for many individual birds and the extent over which a species might range throughout the landscape over a year. The density of paved roads within each landscape was also calculated using ArcGIS.

As one of our objectives was to investigate the effects of forest disturbance on birds, indicators of disturbance due to forest-use practices were recorded to reflect the intensity of disturbances for each site. Forest-use practices such as livestock grazing (cattle), logging and lopping are the major forms of anthropogenic disturbance in the lowland Terai forests. Livestock grazing tends to result in changes to understorey species composition and structure; logging involves the removal of trees >20 cm diameter for timber production, house construction and fuelwood, and lopping is usually for fodder and small fuelwood and involves removal of tree branches 5 – 20 cm diameter. In each quadrat, all the logging stumps, lopping trees, and dung piles were counted. The values of each disturbance variable across the four quadrats were averaged for each site.

61

Table 3.1 Summary of explanatory variables used to assess the influence of site and landscape-scale variables on bird communities. Variables Unit Description Site-scale Tree canopy cover Percent Mean percentage cover of all tree crowns in 200 m x 50 m line transects. Large trees density Count/m2 Number of trees with >100 cm DBH per square metre. Shrub density Count/m2 Number of individual shrubs <2 m tall, per square metre. Landscape-scale

Forest extent (0.79 Hectares/km2 Amount of forest area in a 0.5 km radius of km2) survey site. Forest extent (19.6 Hectares/km2 Amount of forest area in a 2.5 km radius of km2) survey site. Road density Meters/hectare Total length of paved roads divided by the area (ha) of each 5 x 5 km study landscape

3.3.5 Statistical analysis We compared overall species richness, average bird abundance per survey, and species richness and mean abundance of bark-gleaning and foliage-gleaning birds among sites that differed in logging, grazing and lopping intensity using a three-way analysis of variance (ANOVA). To do this, we classified sites into heavily logged or lightly logged, heavily grazed or lightly grazed and heavily lopped or lightly lopped based on the extent of disturbance intensity (Appendix B: Table B. 7).. As the livelihoods of the local population partly depend on extraction of adjacent forest resources, these subsistence activities are common forms of forest disturbances and often occur together in the region.

Sites with ≤ 1 cut stumps per quadrat on average were categorized as lightly logged and >1 as heavily logged; sites with ≤ 2 dung clusters on average were categorized as lightly grazed and > 2 as heavily grazed; sites with ≤ 5 lopped tree branches on average were categorized as lightly lopped and > 5 as heavily lopped. We used only two categories for each disturbance

62 type (i.e. lightly vs. heavily disturbed) because the level of disturbance among the categories were very different, indicating distinctness among categories in terms of disturbance intensity. We used Multidimensional Scaling (MDS) ordination to visualize the pattern of disturbances among different disturbance categories. As we included three tenure types in this study, the majority of sites in protected areas and community managed forests experienced less human disturbance and were separately clustered from those in state- managed forests. The state-managed forests were most heavily disturbed. In addition to the comparison of bird responses, we also compared all vegetation variables (Appendix B: Table B.1 and B.2) among these site categories, also using a three-way analysis of variance (ANOVA). These analyses were conducted using the ‘lm’ function in package “car” (Fox et al. 2012) in the R environment (R Core Team, 2014).

We used generalised linear mixed models to evaluate the influence of vegetation covariates, landscape-scale habitat characteristics and road density on the estimates of bird species richness and the abundance (all birds, bark-gleaning and foliage-gleaning birds). For this modelling, we selected a subset of vegetation variables deemed most likely to affect bird assemblages (Table 3.1). Prior to modelling, we tested for collinearity among explanatory variables using Pearson’s correlation coefficient and excluded one of each pair of variables that had coefficients of correlation >|0.5| from the further analyses (Appendix B: Table B.8 for correlation matrix). All explanatory variables were standardised (mean = 0, standard deviation = 1) to allow comparison of model parameter estimates.

Mixed-effects models are a robust statistical method for a nested study design (Pinheiro and Bates, 2000; Zuur et al., 2010). In a hierarchically nested study, where data are clustered at different spatial scales, there is a different error variance associated with each scale (Crawley 2007). As we collected data from sites (1 ha) embedded within larger landscapes (5 km x 5 km), fitting a standard regression may lead to poor model fit (Beck and Katzy 1995). We therefore used generalized linear mixed models to handle random effects as well as non- normally distributed data (Bolker et al. 2009). In this analysis, fixed effects included vegetation variables, forest extent and road density. Because sites were clustered by 5 km x 5 km landscape, landscape identity was used as random factor in the mixed model, with sites nested within landscape. The mixed-effects modelling was performed using the ‘‘lme4’’ package with Poisson error distributions (Bates et al. 2014) in R (R Core Team 2014).

63

Model averaging of the explanatory variables was then conducted for all response groups using the package MuMIn (Bartoń 2012) in R statistical software (R Core Team, 2014). The model averaging approach determines the strength of effects of the subset of explanatory variables of species richness and abundance of each bird group (Burnham and Anderson 2002). Models were ranked according to their Akaike’s Information Criterion (AIC) value and Akaike weight (ωі). The relative importance of all explanatory variables was calculated by summing the Akaike weights (∑ωі) of variables across all the models where the variable occurred. The larger the ∑ωі value the more important the variable (Burnham and Anderson 2002, Symonds and Moussalli 2011). To evaluate the goodness-of-fit of the model to the data, we calculated conditional and marginal R2 values for the best models using the method described by Nakagawa and Schielzeth (2013). The conditional R2 values show the proportion of the variance explained by the global models (i.e. variance explained by fixed and random factors), while the marginal R2 values show the proportion of the variance explained by the fixed factors only (Nakagawa and Schielzeth 2013). Calculations for both R2 values were done using the ‘arm’ package in R statistical software (Gelman et al. 2012). We tested for spatial autocorrelation by constructing spline correlograms of the model residuals of full models for all response variables using functions “spline.correlog” in the “ncf” R package to plot correlograms with 1000 permutations (Bjornstad 2013).

We used generalized linear models to investigate interactive effects between disturbance intensity and forest extent measured within a 0.5 km and 2.5 km radii of the sampling site on richness and abundance of all birds and bird within each foraging guilds. For this, we classified study sites into two broader disturbance categories: lightly disturbed and heavily disturbed (Appendix B: Table B. 7). To generate the classification, we used the mean value of each of the three disturbance types and standardized the range to lie between 0 – 1. We then took the average value for each site across all three disturbance factors. This composite variable for each site was then used as an index of disturbance in the modelling of interactions between different extents of forest cover and disturbance intensity. We ranked the models based on their AIC values. We then compared the models by highest ranking AIC value and calculated the Akaike weight the explanatory variables for each response variable.

64

3.4 Results 3.4.1 Vegetation and disturbance Large tree density, large tree basal area and tree canopy cover were significantly lower in heavily logged than lightly logged sites (large tree density: F2, 102 = 22.36, P < 0.001; large tree basal area: F2, 102 = 9.70, P < 0.01); tree canopy cover: F2, 102 = 4.69, P < 0.05). Total basal area (F1, 102 = 9.04, P < 0.01), percentage of shrub cover (F1, 102 = 7.94, P < 0.01), and shrub density (F1, 102 = 4.02, P < 0.05) were significantly lower in heavily lopped sites compared to lightly lopped sites. Grazing significantly reduced the ground herbaceous cover

(F1, 102 = 8.55, P < 0.01). The detailed results of effects of different disturbance types on vegetation characteristics are presented in Appendix B: Table B.1 and B.2.

3.4.2 Bird communities A total of 124 bird species was recorded across the 112 sites, comprising 68% local residents, 16% winter visitors, 2% summer visitors and 4% winter passage migrants. Overall mean species richness was 19 ± 0.5 (±SE) and average abundance was 23 ± 0.69 (±SE) across all sites. The overall species richness of birds in sites in heavily grazed and heavily lopped areas was significantly lower (grazed: F1, 102 = 10.29, P < 0.001; lopped: F1, 102 = 27.6, P < 0.001) (Fig. 3.2). The average abundance of birds (all species combined) was significantly lower in heavily logged sites (F2, 102 = 3.5, P < 0.05). Heavy lopping had significant negative effects on foliage-gleaning species richness and abundance (richness: F1, 102 = 7.13, P < 0.01; and abundance: F1, 102 = 6.9, P < 0.01). Similarly, heavy logging adversely affected bark-gleaning bird communities (richness: F2, 102 = 3.48, P < 0.05; abundance: F2, 102 = 1.46, P < 0.05) (Appendix B: Table B.3).

65

Figure 3.2 Mean (± S.E.) species richness (dark bars) and average abundance (light bars) of (a) all birds (b) bark-gleaning insectivores and (c) foliage-gleaning insectivores.

3.4.3 Effects of site- and landscape-level factors The extent of forest cover within a 2.5 km radius of each survey site, large tree density, and tree canopy cover had a strong influence on species richness and abundance for all bird response groups (Fig. 3.3 and Appendix B: Table B.4). A high density of large trees was found to be particularly important for bark-gleaning insectivores, while tree canopy cover and shrub density positively influenced richness and abundance of foliage-gleaning insectivores (Fig. 3.3).

66

Figure 3.3 Model averaged coefficient estimates (± S.E.) across the 95% confidence set of models for all explanatory variables.

There were between 6 and 9 models in the 95% confidence set (∑ωі = 0.95) for all response groups (Appendix B: Table B.5). The goodness of fit statistics based on R2 values showed that fit of the best models was sound for overall species richness (marginal R2 = 0.49 and conditional R2 = 0.51) and total abundance (marginal R2 = 0.45 and conditional R2 = 0.66). Model fit was also good for models of richness and abundance of bark-gleaning insectivores (highest marginal R2 = 0.77 and 0.53 for richness and abundance; conditional R2 = 0.75 and 0.60 for richness and abundance) and richness and abundance of foliage-gleaning insectivores (marginal R2 = 0.74 and 0.86 for richness and abundance; conditional R2 = 0.71 and 0.85 for richness and abundance). Although both conditional and marginal R2 values were fairly

67 similar for all response groups, model performance was better for overall species richness and abundance when the random factor was included in the global models. The extent of forest cover within a 2.5 km radius of the survey site was the predictor with the highest rank importance, and was positively correlated with all response groups. Similarly, tree canopy cover and large tree density had a high rank importance and a positive relationship with all response groups (Fig. 3.4).

68

Figure 3.4 Summed Akaike weights (∑ωі) of final subset of the explanatory variables for (a) overall species richness, (b) average abundance, (c) bark-gleaner species richness (d) bark- gleaner abundance (e) foliage-gleaner species richness, and (f) foliage-gleaner abundance.

3.4.4. Interactions between forest extent and disturbance intensity The interaction between the extent of forest cover within a 2.5 km radius of each survey site and disturbance intensity was significant in models of overall species richness (P < 0.01), average abundance (P < 0.01), and abundance of bark-gleaning birds (P < 0.05) and foliage- gleaning birds (P < 0.01). However, interaction effects of forest cover within a 0.5 km radius

69 of survey site were evident only for abundance of bark-gleaning and foliage-gleaning birds. Interaction terms were included amongst the best models for all response groups (<2 ΔAIC) (Appendix B: Table B.6). Thus, the effects of disturbance on bird communities at the site- scale depended on the extent of forest cover in the landscape. The bird assemblages of more- disturbed sites responded more strongly to the percent of forest cover in the landscapes than did those of undisturbed sites (Fig. 3.5). Both bark-gleaning and foliage-gleaning species responded more strongly to landscape-level forest cover in more-disturbed sites. For all two response groups, the higher richness and/or abundance in less-disturbed sites was only evident when there were low to moderate levels of forest cover in the landscape; in the most forested landscapes, bird responses were similar among heavily-and lightly-disturbed sites.

70

Figure 3.5 Relationship between overall bird species richness (a),overall bird abundance (b), bark-gleaning abundance (c), and foliage-gleaning abundance and percent of forest cover in 2.5 km radius of survey site in landscape (heavily disturbed sites (filled circles and heavy dashed line) and lightly disturbed sites (open circles and fine dashed line). For the purposes of displaying the interaction effect, sites were divided into heavily and lightly disturbed based on the value of the disturbance index (from three disturbance types).

71

3.5 Discussion Heavily-extracted sites had less-complex forest structure and depauperate avifaunal communities, indicating potentially widespread deleterious effects of forest disturbances on avifaunal communities in multi-use forests. However, while forest use practices significantly affected the avifaunal community of sites, the intensity of these effects was dependent on landscape context. Similarly, the effect of the extent of forest in the surrounding landscape was weaker when the site was less-disturbed. This study therefore underscores the importance of understanding the potentially interactive effects of disturbances at multiple scales.

3.5.1 Effects of forest disturbances on bird communities All three types of forest disturbance had deleterious effects on bird communities when they occurred at higher intensities. Our results are largely consistent with past findings that reported lower richness and abundance of birds in more disturbed sites (Peh et al. 2005, Shahabuddin and Kumar 2007). Selective removal of large mature trees reduces habitat suitability for many species of birds (Eyre et al. 2009, Touihri et al. 2014) that rely on them for foraging and nesting (Sekercioglu 2002, Vergara and Marquet 2007). For example, bark- gleaning insectivores are often strongly associated with large tree density (Cleary et al. 2007, Greve et al. 2011, Dahal et al. 2014), and so are highly sensitive to habitat alteration (Adams and Morrison 1993, Zurita and Bellocq 2012, Inskipp et al. 2013). Therefore, density of large trees in a forest stand was the most important predictor of the distribution and abundance for many species of birds at the site-scale.

Lopping and logging also had negative impacts on foliage-gleaning bird communities. These species use the canopy layer for foraging, so removing parts of the canopy can have a negative impact on species richness, abundance and composition. In this study, lightly lopped and lightly logged sites had three times more foliage-gleaning insectivores than did heavily lopped and heavily logged sites. Similar patterns of species distribution have also been noted by Shahabuddin and Kumar (2006), who found that foliage-gleaning species such as great tit Parus major, Hume’s warbler Phylloscopus humei and small minivet Pericrocotus cinnamomeus were significantly more common where canopy cover was closed in an Indian reserve. A recent study by Leal et al. (2013) in a Mediterranean region, found that foliage- gleaning species such as great tit and chiffchaff Phylloscopus collybita were most affected by canopy pruning activities.

72

In this study, heavily grazed sites had only half the species richness of lightly grazed sites, indicating a strongly detrimental effect of grazing practices on bird communities. Heavy grazing significantly affects the population of many species of birds (Martin and McIntyre 2007). Grazing and trampling reduces understorey vegetation (Whitehorne et al. 2011, Piana and Marsden 2014), thereby affecting understory-foraging species (Martin and McIntyre 2007) and ground-dwelling bird species (Maron and Lill 2005, Inskipp and Baral 2013). Grazing can also have indirect effects on bird communities through changes in vegetation characteristics such as nest site suitability and food supply (Dennis et al. 2008). For example, in our study area, Abbot’s babbler Malacocincla abbotti, a terrestrial insectivore that nests close to ground or in shrubs (Grimmett et al. 1999), was only recorded in lightly grazed sites, presumably due to loss of understorey vegetation cover.

3.5.2 Moderating effects of landscape context Although both site-and landscape-scale forest characteristics were important predictors in determining species richness and the abundance, landscape characteristics had a consistently positive and strong effect on all groups of birds. At the 19.6 km2 (2.5 km radius of survey site) landscape scale, the forest extent had an important influence on bird species richness and abundance. Strong effects of the proportion of habitat in landscapes surrounding sites have been reported for forest breeding birds (e.g. Trzcinski et al. 1999) and woodland birds (e.g. Maron et al. 2012). The scale at which landscape context is important varies with taxon and habitat type. For example, reptiles respond most strongly to habitat context measured at 0.5 km2 scale (Bruton 2014), koalas at 1 km2 (McAlpine et al. 2006) and land birds (100 km2) (e.g. Radford et al. 2005). However, in our study, the extent of forest within 0.79 km2 of a site was less important, suggesting a greater role for landscape context at larger scales in influencing the structure and composition of avian assemblages.

As we predicted, the impact of site-scale forest disturbance on bird communities depended on the extent of forest cover within the surrounding landscape. Overall bird species richness, average abundance and the abundance of bark-gleaners increased most rapidly with landscape-level forest cover in the most heavily-disturbed sites, which indicates a higher importance of forest extent in human-dominated landscapes where site-level habitat quality is poor (Vergara and Armesto 2009). The benefits to the avifauna of less-disturbed sites were strong at low to moderate levels of forest cover, but in landscapes with higher forest cover (60% - 90%), the avifauna was less sensitive to disturbance. We found that disturbance

73 intensity was negatively related to the extent of forest cover (r = -0.67), with more intense effects in lower-cover landscape. Therefore, our results suggest that the local-scale impact of disturbances on bird communities may be moderated, at least partly, by maintaining a high proportion of habitat surrounding such sites.

The response of bird communities to the interaction between landscape-level forest cover and disturbance intensity varied depending on levels of species mobility. The differences may be due to the different movement strategies that characterise each of the foraging guilds. For example, many species of bark-gleaners (e.g. Chrysocolaptes lucidus, lesser yellownape Picus chlorolophus) have specialised niches and small ranges (Kumar et al. 2014), and this group responded strongly to the amount of forest cover within 0.5 km in the highly disturbed sites. This indicates that forest cover close to disturbed sites may buffer negative effects of disturbance by offering different habitat resources. For highly mobile birds, forest cover within this small extent may be less important, as they range over much larger areas. Furthermore, the highly mobile and widespread generalists such as sallying insectivores (e.g. Microhierax caerulescens, Dicrurus macrocercus) had little response to the extent of forest cover in the landscape at either scale.

Landscapes with more forest cover support a larger species pool (Radford and Bennett 2007, Haslem and Bennett 2008, Taylor et al. 2012) through both sampling effects, and because a greater extent of forest cover in landscapes offers habitat diversity (Radford et al. 2005, Maron et al. 2012), and can serve source habitats for range of species (Pulliam 1988). Birds are a mobile taxon, and their presence at a particular site does not necessarily mean they are resident, nor that they solely use the resources within that site. Thus, complementary resources may be more widely distributed within the occupied landscape. More habitat within the surrounding landscape increases the chance that suitable refugial or complementary habitat exists (Dunning et al. 1992), increasing the likelihood of occupancy within the landscape and thus the chance of detection at a site, even a degraded one. Therefore, maintaining more forest cover has important ecological consequences for the ability of a wide-range of avifaunal species to persist.

74

3.6 Management implications Our results have important management implications in terms of sustainable forest management and biodiversity conservation at the landscape scale. Our study demonstrates the maintenance of larger areas of mature forest in the landscape should be a high conservation priority for bird conservation in highly-disturbed landscapes. This can be achieved with: a) appropriate conservation and restoration of degraded landscapes particularly for those forests that are heavily degraded such as most of the state forests and b) maintenance of existing forest cover in protected areas as protected areas are important natural habitats in the region. Similarly, it is also important to reduce human pressures on forests to maintain vegetation complexity at the site-scale. Since most lowland landscapes in Nepal are multiple-use and are subject to a high degree of anthropogenic pressures, the development and implementation of sustainable forest management plans are urgent to prevent further degradation of habitat and their avifauna in lowland landscapes of Nepal.

75

CHAPTER 4 RELATIONSHIPS BETWEEN LANDSCAPE-LEVEL SPECIES RICHNESS AND FOREST EXTENT VARY AMONG BIRD GUILDS

Plate 4: A patch of native forest in Chitwan National Park in lowland Nepal.

Submitted to Journal of Applied Ecology as: Dahal, B. R., C. A. McAlpine, and M. Maron. 2015. ‘Effects of landscape characteristics and habitat extent on bird communities in lowland Nepal’.

76

4.1 Abstract Anthropogenic habitat modification has dramatically altered the spatial patterning of different habitats, which affects the richness of species that landscapes support. A particular focus of research has been thresholds in extent of preferred habitat, below which richness declines rapidly. However, there is likely to be variation among functional groups in the strength and shape of the relationship between the extent of habitat and landscape-level species richness. I surveyed birds across 28 landscapes (each 5 × 5 km) that differed in the extent of remnant forest in southern lowland Nepal. The estimated species richness of all birds and bird within several foraging guilds at the landscape-level were modelled as a function of forest extent in the study landscapes, and factors that potentially modify the relationship (such as degree of disturbance) were explored. The landscape-level species richness of forest birds was consistently positively related to the extent of forest cover in the landscape. However, the strength of the relationship varied substantially among foraging guilds, with effects strongest for foliage-gleaning and frugivores. As with previous studies, species richness increased most steeply with forest extent in less-forested landscapes, but my findings differed in that richness continued to increase as forest extent approached 100%. The strongest effects of forest cover on overall bird richness occurred in landscapes with a greater extent of disturbance. Managing and restoring forests to maintain forest extent, particularly in more-degraded landscapes, should be a key strategy for landscape-level conservation of birds in the region.

Keywords: Forest extent, forest birds, habitat thresholds, habitat disturbance, landscape change, Nepal

77

4.2 Introduction Anthropogenic land-cover change has dramatically altered the spatial patterning of different habitats (DeFries et al. 2004, Foley et al. 2005) which in turn has affected the distribution and abundance of many species. More than 50% of the world’s species have declined in population size over the last four decades due principally to the loss and degradation of natural landscapes (WWF 2014). As pressure from human land-uses continues and landscape patterns and ecological processes are disrupted (Franklin et al. 2002, Fischer and Lindenmayer 2007), understanding how these novel landscape patterns affects species persistence at large spatial scales is increasingly important for biodiversity conservation (Fahrig 2003, Bennett et al. 2006).

The total extent of forest cover is a key driver of the occurrence and abundance of species, ultimately affecting the total number of species a landscape can support. Landscapes with more habitat support a larger species pool (McGarigal and McComb 1995, Trzcinski et al. 1999, Fahrig 2003, Maron et al. 2012) through both sampling effects (Wiens 1992, Whittaker and Fernández-Palacios 2007), and because habitat extent correlates with habitat diversity (Radford et al. 2005, Kallimanis et al. 2008). Different types of habitat across the landscapes can be refugial or complementary for many species of fauna. For example, habitat diversity can provide critical complementary resources for different activities such as breeding, foraging, and nesting, allowing persistence of more species in the landscape (Dunning et al. 1992, Law and Dickman 1998). Similarly, certain citical habitats and habitat features (e.g. riparian forest, large mature trees) required for particular species are more likely to occurr in landscapes with more habitat. Therefore, the larger extent of habitat is likey to offer both primary and complementary habitats for the peristence of faunal communties within a landscape.

Empirical studies suggest that sites in landscapes with more forest can have higher densities of reptiles (McAlpine et al. 2015), greater species occurrence and abundance of birds (e.g. Villard et al. 1999, Mortelliti et al. 2010, Martensen et al. 2011, Taylor et al. 2012), and greater richness of small mammals (e.g. McAlpine et al. 2006, Estavillo et al. 2013). In particular, studies of both individual species (e.g. Betts et al. 2007, Suarez-Rubio et al. 2013) and landscape-level species richness of birds (e.g. Radford et al. 2005, Hu et al. 2012, Maron et al. 2012) have reported non-linear responses to forest extent. A particular focus has been habitat extent thresholds, whereby if forest extent falls below the threshold, sharp declines in

78 population sizes and species diversity occur within that landscape (Andren 1994). Radford et al. (2005) showed that steep declines of landscape-level species richness of woodland birds below a threshold of 10% remnant habitat cover, and Maron et al. (2012) also reported nonlinearities in this relationship. Since such thresholds would have important consequences for species persistence in modified landscapes (Ewers and Didham 2006), understanding where and for which species groups such thresholds exist is key to developing conservation and restoration strategies (Lindenmayer and Luck 2005, Huggett 2005).

Although several studies have now reported a nonlinear relationship between landscape-level species richness and extent of suitable habitat (Taylor et al. 2012, Maron et al. 2012), the shape and magnitude of bird species response to forest cover may not be universal across the different foraging guilds. It is likely that different groups of species vary in their response to forest cover in the landscape. For example, frugivore richness might be strongly driven by the variety of fruiting plants which diminishes as habitat loss proceeds, where as raptorial carnivores with more generalist diets might be less-affected. Thus, strongly nonlinear relationships with forest cover might be driven by particular groups of species that are most sensitive to forest loss. However, variation in the nature of the relationship between richness of different foraging guilds of birds and the extent of forest habitat in the landscapes remains unexplored.

Further, both species persistence and the effect of forest cover are likely to be influenced by other landscape characteristics, such as levels of anthropogenic disturbance. Many types of forest disturbance, such as logging and livestock grazing, negatively affect species persistence at the site-level (Shahabuddin and Kumar 2007) and such relationships could cloud understanding of how species respond to loss of forest cover. Thus, the relationship between landscape-level species richness and forest cover may be altered by interactions among landscape characteristics. Knowledge of such interactions remains limited.

Here I investigate relationships between landscape characteristics and landscape-level species richness of forest birds across lowland Nepal. Firstly, I examine whether the effects of remnant forest extent in a landscape on the species richness of birds in that landscape are consistent across foraging guilds, or if a particular guild contributes disproportionately to observed patterns between species and forest cover in the landscape. Secondly, I determine whether non-linear relationships exist between landscape-level species richness of birds and

79 forest cover, and examine how different landscape characteristics influence these relationships. Finally, I assess the existence of interaction effects between disturbance intensity and landscape-level forest extent on species richness of birds.

4.3 Material and Methods 4.3.1 Study area The study was conducted in southern Nepal, in the region called ‘Terai’ (80° 4’ 30” to 88° 10’ 19” E 26° 21’ 53” to 29° 7’ 43” N, elevation 63 - 330 m ASL). The Terai encompasses most of the country’s tropical moist forest from the Mechi River in the east to the Narayani River in the centre. The mean annual rainfall decreases from 2,680 mm to 1,138 mm from east to west, and the mean monthly rainfall ranges from 8 mm in November to 535 mm in July (FRA/DFRS 2014). The area is characterized by a tropical climate, with the maximum monthly mean temperature of 35-40°C in April/May and the minimum, 14-16 °C, in January (Jackson et al. 1994). Before 1950, the region supported continuous dense tropical forest. With the eradication of malaria in the early 1950s, large tracts of the highly productive lowland forests were converted to agriculture (Hrabovszky and Miyan 1987). Consequently, most of the forest was destroyed and the remaining forest areas were subjected to intense human exploitation. Nearly half of Nepal’s population now lives in the 17% of the country that is lowland (Central Bureau of Statistics 2011).

80

Figure 4.1 The study landscapes in lowland Nepal (grey shading indicates forest cover) and histogram showing the different extent of forest cover in landscape.

4.3.2 Study design Twenty-eight landscapes, each 5 km x 5 km, and supporting different amounts of forest cover (7.9%–95.3%), were selected across south-central (Bara-Parsa forest and Chitwan forest) and south-eastern lowland forests (eastern forests) (Fig. 4.1). Geographically, 15 landscapes were located in eastern Terai forests and 13 landscapes were located in central Terai forests. Four survey sites, each measuring 200 m x 50 m, were randomly located in each landscape, resulting in a total of 112 sites (28 landscapes x 4 sites). All sites were located at least 500 m from roads to minimize any road-induced variation in bird assemblages. The minimum distance between sites was at least 1000 m to reduce the chance of spatial dependence.

4.3.3 Bird surveys At each study site, birds were surveyed on three occasions between November 2012 and May 2013On each visit, the observer (BRD) recorded all birds seen or heard within 25 m of the centreline of the transect while walking along its length over a 10-min period. Surveys were conducted only between 0600 and 1100 h in the morning and 1400 to 1745 h in the afternoon. Although the effects of time of day on bird observation were not tested prior to the actual field survey, several other studies have reported that the detection rate of most bird species is greater in morning (Bried et al. 2011) with another peak in activity in the late afternoon, 2–3 h before sunset (Kessler and Milne 1982). In general, birds tend to reduce activity during the midday heat (Pizo et al. 1997). To avoid possible bias, we standardized the survey protocol in such a way that, although not all individual sites had an afternoon survey, afternoon surveys occurred equally among site types and landscapes. All surveys were conducted during fair weather when there was no rain and the wind speed was low.

4.3.4 Landscape variables The total area of forest and water (river, permanent lakes) in each study landscape was calculated using GIS (using ArcGIS 9.3). I measured the total area of forest and open water in ha/km2 from land cover data of the region (WWF 2005). The land-cover data were made available by WWF Nepal. Mean rainfall (mm/yr) data for each landscape was obtained from the closest weather station (Department of Hydrology and Meteorology, Nepal). The total numbers of trees were counted within four 20 m x 20 m quadrats to measure the species

81 richness of trees at each site. Similarly, disturbance intensity was assessed at each site using four 20 m x 20 m randomly located quadrats. In each quadrat, indicators of disturbance due to forest-use practices were recorded to reflect the intensity of disturbances in terms of grazing, fodder collection and fuel wood extraction (Shahabuddin and Kumar 2007). These included the proportion of trees showing signs of lopping, number of cut stumps and number of livestock dung piles. The density of livestock dung piles indicates the degree of usage of forest habitat by grazing livestock, stump density reflects logging intensity, while signs of lopping indicate the amount of fodder and fuelwood extraction for each site. The values of each disturbance variable across the four quadrats were averaged for each site. The mean value of lopping (1.7 ± 0.4), grazing (2.0 ± 0.4) and logging (0.5 ± 0.1) in the lightly disturbed sites were significantly lower than mean values of lopping (46.5 ± 3.4), grazing (12.5 ± 0.9) and logging (11.1± 0.9) in heavily disturbed sites. The values from the four survey sites per landscape were then averaged to derive an average disturbance value. In addition, to display the interactions, I classified the landscapes into lightly disturbed and heavily disturbed landscapes. Landscapes with ≤ 5 disturbance index on average were categorized as lightly disturbed, while ≥ 5 as heavily disturbed (Appendix B: Table B.7).

4.3.5 Data analysis Bird survey data were pooled across all sites within each study landscape. Membership of foraging guilds was identified based on primary habitat specialization and diet information compiled from Ali and Reply (1987) and Grimmett et al. (2009). Although several foraging guilds were identified, only those foraging guilds which had more than five species present in each landscape were included in the analyses. These include: foliage-gleaning insectivores, sallying insectivores and frugivores. To account for variation in sample completeness among landscapes, I calculated the estimated species richness for each landscape for all birds and different foraging guilds using the nonparametric species richness estimator Chao2 in the programme Estimates 9.1.0 (Colwell 2013).

I fitted a series of regression models to visualize the relationships between estimated species richness (Chao2) of each species group and forest cover. I compared the performance of linear, exponential, loess and discontinuous piecewise regressions using the Akaike Information Criterion (AIC) and adjusted R2 values. The loess (locally weighted, non- parametric regression) is useful for examining nonlinear relationships (Cleveland and Devlin 1988, Jacoby 2000). For the piecewise regression model, we used the “segmented” package

82 in R (Muggeo 2012) to identify the break-point from the data. When using the segmented package to test for break-points, an initial estimate of the break-point from the data is required as a starting estimate, which we tested for a priori using the Devies test (Davies 1987, Muggeo 2008).

I also modelled the relationship between landscape-level bird species richness and all landscape variables using GLMs with a Poisson distribution in the lme4 package v.1.1-5 in R (R Core Team 2014). All explanatory variables were standardised (mean = 0, standard deviation = 1) to allow comparison of model parameter estimates (Burnham and Anderson 2002). Model averaging of the explanatory variables was then conducted for all response groups using the package MuMIn (Bartoń 2012) in R statistical software (R Core Team, 2014). The model averaging approach determines the strength of effects of the subset of explanatory variables of species richness and abundance of each bird group (Burnham and

Anderson 2002). Models were ranked according to their AIC value and Akaike weight (ωі). A 95% confidence set of models, used for model averaging, was constructed by starting with the model with the highest Akaike weight and repeatedly adding the model with the next highest weight until the cumulative sum of weights exceeded 0.95 (Burnham and Anderson 2002). The Akaike weights in which each predictor variable occurred were summed; the larger the ∑ωі value the more important the variable (Burnham and Anderson 2002, Symonds and Moussalli 2011). As an indication of goodness-of-fit, I calculated R2 values for the global models in R statistical environment (R Core Team 2014).

4.4 Results 4.4.1 Relationships between landscape-level species richness and forest cover The landscape-level species richness of all birds increased with the increase of forest extent in the landscape, with 42.1 % of the variance explained by the best model. However, the strength of the relationship varied substantially among the foraging guilds. Sallying insectivores were only weakly associated with forest cover, but foliage-gleaning insectivore and frugivore richness were strongly related to forest extent, with 63.2 % and 36.2 % of the variance explained by the best models, respectively. The shape of the response to forest cover also varied among foraging guilds (Fig. 4.2). There was limited evidence to support the threshold models. The break points (thresholds) as a function of forest cover ranged from 8.3 % of forest cover (sallying insectivores) to 14.1 % (foliage-gleaning insectivores). However, inspection of the loess models for overall bird richness and foliage-gleaning species richness

83 suggested initial steep increases with forest cover at low levels of cover, before a less-steep, but continuing, increase in landscapes beyond 20-30% forest cover. This was most pronounced for richness of foliage-gleaning species.

Figure 4.2 Models of the relationship between estimated species richness and forest extent in each landscape. AIC: Akaike information criterion for each model

4.4.2 Relative importance of landscape variables The strong positive effect of forest area on estimated species richness was evident from the model averaging, with the summed Akaike weight revealing that forest cover was the most influential parameter in each of the models for all response groups (Fig.4.4). The importance of landscape-level forest cover was greatest for all species (∑ωі = 1.0) and foliage-gleaning insectivores (∑ωі = 1.0) and weakest for sallying insectivores (∑ωі = 0.33). Other important landscape characteristics included the extent of water body, which was the most reliable

84 predictor for frugivorous and all species, while annual rainfall was an important influence on foliage-gleaning insectivores. However, effects of all parameters were weak for sallying insectivores (Fig. 4.3).

Figure 4.3 Model averaged coefficient estimates (± S.E.) across the 95% confidence set of models for all explanatory variables for each of: (a) all species, (b) foliage-gleaning insectivore, (c) frugivore and (d) sallying insectivore.

85

Figure 4.4 Summed Akaike weights (∑ωі) from model averaging of the environmental variables for landscape-level richness of (a) all species, (b) foliage-gleaning insectivores, (c) Frugivores, and (d) sallying insectivores.

4.4.3 Interactions between landscape characteristics The interaction between the extent of forest cover and disturbance intensity was significant in models of estimated richness of all species. The bird assemblages of less-disturbed landscapes responded more weakly to the extent of forest cover in the landscape than did those of more-disturbed landscapes. This interaction effect was not evident when individual guilds were considered. However, for foliage-gleaning insectivores, rainfall interacted significantly with forest cover, such that the response of foliage-gleaning insectivores to the extent of forest was weaker in higher-rainfall landscapes (Fig. 4.5).

86

Figure 4.5 Relationship between overall estimated richness and percent of forest cover (a) in more-disturbed landscapes (filled circles and full line) and less-disturbed landscapes (open circles and heavy dashed line) and (b) higher-rainfall landscapes (filled circles and full line) and lower-rainfall landscapes (open circles and heavy dashed line).

4.5 Discussion Strong relationships between landscape-level forest extent and estimated species richness of birds were evident, and these relationships varied among foraging guilds, with richness of foliage-gleaning species contributing most strongly to the positive relationship. Further, the effect of landscape-level forest extent on bird richness was influenced by disturbance levels and rainfall. The relationship between estimated species richness and the extent of forest cover in the landscape was nonlinear, supporting the hypothesis that the relationship between forest extent and richness is steepest at low levels of forest extent significantly reduces. However, as the relationship between richness and forest extent varied among foraging guilds and with landscape characteristics, generalizing such relationships may mask important elements of the consequences of landscape change.

4.5.1 Species richness and forest cover Most studies of avian responses to landscape change have focused on response variables measured at the site or patch-level, using characteristics of the site or the landscape context surrounding site as explanatory variables (McGarigal and Cushman 2002, Guenette and Villard 2005). However, response variables measured at small scales cannot necessarily reveal emergent properties of whole landscapes (Bennett et al. 2006). More recently, key

87 empirical studies (e.g. Radford and Bennett 2007, Pardini et al. 2010, Maron et al. 2012, Taylor et al. 2012, Bennett et al. 2014, Ochoa‐Quintero et al. 2015) have begun to focus on landscape-scale sampling, examining how the assemblages of entire landscapes respond to landscape characteristics in Australia. My results are consistent with several of these recent studies for birds that showed strong positive relationships between species richness and the extent of forest cover in the landscape (Radford et al. 2005, Maron et al. 2012, Ochoa‐Quintero et al. 2015). However, there was variation among foraging guilds, indicating that not all groups of species respond in the same way to forest cover. The richness of foliage-gleaning insectivores and, to a lesser extent, frugivores, was responsible for the pattern, whereas richness of sallying insectivores were only weakly related to forest cover.

The strong positive relationships between foliage-gleaning insectivores and frugivores and forest cover in the landscape may be due to increased diversity of trees and therefore foraging substrates and fruit sources in more-forested landscape. Haddad et al. (2009), for example, revealed that the species richness and abundance of arthropods was positively related to plant species richness. In my study landscapes, forest cover was positively related to the richness of tree species recorded in the transects (r = 0.54) and tree richness was also related to richness of foliage-gleaners (r = 0.31) and frugivores (r = 0.19). A similar positive correlation— although at a site level—between tree diversity and species richness of avian insectivores and frugivores birds was found in agricultural landscapes (Harvey et al. 2006). Thus, a greater variety of potential food sources is more likely to support a correspondingly large suite of species, whereas diversity of food sources for the more generalist group of aerial insectivores may be less likely to relate to forest extent.

The relationship between species richness and extent of forest cover was somewhat non- linear, with species richness decreasing more steeply below about 20-30% forest cover in the landscape. My findings are largely consistent with the results of other studies that have shown non-linearity in the relationship between species richness and habitat extent (Radford et al. 2005, Maron et al. 2012). However, in contrast to the findings of these studies, my results also showed continued, but less-steep, increase of species richness above the threshold. I also found that the observed responses of birds to forest loss were not consistent across the forest bird community. Instead, several underlying factors such as landscape productivity, vegetation or soil type (Maron et al. 2012) and species sensitivity to forest cover change (Martensen et al. 2012) can affect the response of species richness to forest cover.

88

Considering total species richness alone could potentially obscure variation in relationships at the functional level of birds, particularly when forest cover is confounded with underlying factors (Maron et al. 2012).

4.5.2 Relative importance of landscape characteristics In addition to forest extent, several other factors affected the species richness of forest birds in the landscape. Annual rainfall and extent of water were also positively related to landscape-level species richness of birds. Richness of all birds and of frugivores both increased with the extent of water in the landscape. This association is likely to be driven by differences in primary productivity and resource availability in riparian habitats. Riparian habitats generally support distinct vegetation (Palmer and Bennett 2006) and provide more resources particularly emergent aquatic and aerial terrestrial for many insectivores birds (Iwata et al. 2003). A higher species richness of birds in riparian habitats has been reported in other studies (Knopf and Samson 1994, Schneider and Griesser 2009, Bennett et al. 2014). Furthermore, the presence of permanent water bodies increases landscape heterogeneity, generating habitat diversity (Tews et al. 2004). Habitat diversity in the landscape is also important for persistence of mobile species which can require different habitats for their survival (Saunders 1990, Law and Dickman 1998). Foliage-gleaning insectivore richness was also higher in landscapes with greater annual rainfall. Rainfall influences arthropod abundance and diversity in forest landscapes (Sofaer et al. 2012) and increases carrying capacity for insectivorous birds (Williams and Middleton 2008).

4.5.3 Interactions between landscape characteristics My study revealed that the effects of extent of forest cover on bird species richness depended on the degree of disturbances in the landscapes. Positive effects of forest extent on bird communities in the landscape are often reported (e.g. Andren 1994, Cushman and McGarigal 2003) but the way that intensity of disturbances (livestock grazing and logging) interacts with landscape-level forest extent to drive landscape-level richness has not previously been examined. The strongest effect of forest cover on all species richness occurred in landscapes with a greater extent of disturbance, indicating a role of forest extent in moderating the effects of disturbance. Accordingly, disturbance effects on avifauna of less-disturbed landscapes were strongest at low to moderate levels of forest cover, but in landscapes with greater forest cover (60% – 90%), the avifauna was less affected by disturbance. The species richness of foliage-gleaning insectivores across the landscape was also dependent on the

89 interaction between rainfall and forest cover, suggesting that the extent of forest cover has less effect on foliage-gleaning insectivores in higher-rainfall landscapes, which are likely to be more stable in food availability (Williams and Middleton 2008).

4.6 Management implications This study provides strong evidence of positive association of birds with the extent of forest cover at the landscape level, and therefore has important management implications. The non- linearity of relationships between species richness and forest cover suggest that species richness increases more rapidly with forest cover at low (< 20%) levels of cover, but the positive relationship between richness and forest cover is still evident at high levels of cover. Nepal’s lowland forest landscapes face a high risk of collapse in avifaunal richness if further forest loss and modification occur, particularly in landscapes of lower vegetation cover. For example, the state managed forests in Parsa and Eastern landscapes are particularly vulnerable due to low forest cover. Furthermore, logging, over-extraction of forest resources and cattle grazing are ongoing disturbances in these forests. It is therefore important to place a regulatory mechanism that helps reduce human pressure and maintain vegetation complexity at both site-and landscape-scales. As more than 70% of the country’s forest bird species (of which more than 50% are nationally threatened) inhabit the lowland forests (Inskipp et al. 2013, Baral et al. 2014), maintenance of existing forest cover and interventions to restore degraded habitats to prevent further loss of avifaunal populations in lowland landscape of Nepal are critical.

90

CHAPTER 5 SYNTHESIS AND RECOMMENDATIONS

Plate 5: Agricultural intensification in adjacent to the protected area, lowland Nepal.

91

5.1 Overview The conservation of faunal communities in human-dominated landscapes is a challenging task, particularly in developing countries. About 1.6 billion people living in poverty depend on forests for their livelihoods (World Resource Institute 2005, Chao 2012) and this dependency is likely to increase in the future. This burgeoning demand for food and fibres puts enormous pressure on remnant forests, and has significantly contributed to degrading landscapes and changing habitats for both fauna and flora. Thus, effective conservation of avifaunal assemblages in multiple-use landscapes, for example, requires an understanding of how habitat characteristics at individual sites and also across landscapes affect species composition and persistence (Ewers and Didham 2006, Lindenmayer and Fischer 2006).

In this thesis, I investigated the effects of habitat characteristics on forest bird assemblages at multiple spatial scales in order to draw inferences for conservation of avifauna in the lowland Terai forest of Nepal. In Chapter 2, I examined relationships between site characteristics and bird richness and abundance, with both response and predictor variables measured at the same scale. This site-level study was used to compare the conservation value for birds of differently-managed forests. In Chapter 3, I examined the relationships between forest bird assemblages and both site and landscape characteristics, including the extent of forest within both a 500 m and a 2500 m radius of survey sites. Here, I hypothesized that the occurrence of species and assemblages depends not only on the properties of sites at which birds were sampled, but also on the proportion of forest in the landscape, and its interaction with site- level factors. In Chapter 4, I extended my research approach beyond the site/landscape context, and adopted a whole-of-landscape approach in which both the response variables and predictor variables were measured at the scale of the whole landscape (Bennett et al. 2006). Such an approach is useful in understanding the influence of emergent properties of entire landscapes on faunal assemblages (Mortelliti et al. 2010, Maron et al. 2012, Taylor et al. 2012).

In this final chapter, I summarize the main outcomes of my research in relation to the questions I posed, and discuss the implications for the management of human-dominated landscapes for avifaunal conservation in the lowland Terai forests of Nepal. I also discuss the limitations to this study and propose directions for future research.

92

5.2 Off- reserve forests provide complementary habitats for bird conservation In recent years, forest outside of formal conservation reserves has been increasingly recognized for its potential role in conservation of biota (Persha et al. 2010, Porter-Bolland et al. 2012, Baral et al. 2014). As existing protected area coverage is often biased in terms of the species and habitats that are protected (Tewksbury et al. 2002, Hoekstra et al. 2005), the conservation of off-reserve forests can provide complementary habitats that help to support a larger suite of species across the landscape. Such complementary habitats in the landscape provide critical resources, particularly for species that require a variety of habitats for their persistence (Dunning et al. 1992, Law and Dickman 1998). Thus, off-reserve forests can be critical to maximizing representation of biodiversity features in a landscape. It is therefore important to understand conservation values of off-reserve forests and how such spatially- distributed forest habitats complement existing protected area networks in achieving conservation of avifauna in the region.

In Chapter 2, I investigated whether off-reserve forests (state-managed and community- managed forests) support bird assemblages that complement those of protected areas. I compared the habitat attributes and bird assemblages among sites in each of these three forest tenure categories. Protected area sites had the greatest richness and diversity of birds compared to sites in community forests and state forests. They also had significantly greater species richness and diversity of forest specialists and bark-gleaning insectivores. However, off-reserve forests supported bird assemblages that complement those of protected areas. Many species of birds that were not recorded in sites in protected areas were recorded in sites in off-reserve forests. Only 45% of species detected were common to all three forest management tenures. This distinctness of bird species in the sites in off-reserve forests contributes to maintenance of species diversity across landscapes.

Habitat features such as tree canopy cover and large tree density were similar between sites in community-managed forests and those of protected areas. This indicates that aspects of habitat condition in sites in community forest are relatively good, and potentially provide critical resources for many species of birds at levels similar to protected areas. Furthermore, as off-reserve forests provide structural links among different critical resources in the landscape, species of birds that exploit a variety of habitats, for example, frugivores that must follow the shifting pattern of fruit availability over time, can benefit. Such connectivity is of

93 primary importance to the distribution and abundance of biota (Lindenmayer et al. 2008, Chisholm et al. 2011).

Having a complementary habitat resource within a landscape is likely to increase beta diversity. Beta diversity (species differentiation across sites, also referred as spatial turnover) has important implications for conservation planning (e.g. Condit et al. 2002, Wiersma and Urban 2005). In this study, I found that beta diversity (i.e. species turnover) among sites in off-reserve forest types was higher than among sites in protected areas (Chapter 2). Thus, although alpha diversity is lower in off-reserve forests, richness at larger scales is likely to be relatively similar.

However, these forests outside the protected areas, in particular state-managed forests, are subject to heavy anthropogenic pressure for subsistence livelihood activities. As such, anthropogenic activities have detrimental impacts on forest structure and associated avifauna (Chapter 3), these state-owned forests may not necessarily support suitable habitats for long- term persistence of populations of several species including forest specialists. Thus, Chapter 2 concludes that including state-owned forest into appropriate conservation measures such as habitat restoration and preventing over exploitation to maximize regional avifaunal diversity.

5.3 Forest use practices can have detrimental effects on vegetation and associated birds Forest use practices such as logging, livestock grazing, and lopping of tree branches for fodder and fuelwood are the major forms of forest disturbance in multiple-use landscapes. These extractive activities, mainly for subsistence, may significantly change the forest structure and diversity (Sagar and Singh 2004, Kumar and Shahabuddin 2005). For example, livestock grazing tends to result in changes to species composition and structure in the understorey (Tasker and Bradstock 2006, Whitehorne et al. 2011), and logging for timber production and fuelwood and lopping for fodder and fuelwood simplifies the stand structure (Shahabuddin and Kumar 2007, Thapa and Chapman 2010). However, effects of subsistence forest disturbance on bird assemblages have received little attention despite the fact that such information is essential for effective conservation planning for anthropogenically-disturbed landscapes.

In this study, I found that logging, grazing and lopping activities had deleterious effects on multiple aspects of habitat condition (Chapter 3). As expected, the density and basal area of

94 large trees and tree canopy cover were significantly lower in heavily-logged sites than in lightly-logged sites. Lopping activities also significantly affected shrub density and shrub cover, while grazing activities reduced herbaceous cover. Studies have shown negative effects of logging (e.g.Moktan et al. 2009, Sapkota et al. 2010) and grazing (e.g. Whitehorne et al. 2011) on vegetation structure; this study also underscores negative lopping activities as substantial drivers of simplified vegetation structure. In lowland Terai forests, lopping that involves removal of tree branches 5 – 20 cm diameter is widely practiced usually for fodder and fuelwood (Chapter 3, Thapa and Chapman 2010). Thus, I concluded that as the livelihoods of the local population partly depend on extraction of adjacent forest resources, these subsistence activities are reducing the stand-scale habitat complexity in the lowland Terai forests.

Mirroring the results for vegetation structure, species richness and abundance of birds were also negatively affected by the intensity of logging, lopping and grazing practices in the lowland landscape. The overall species richness of birds in sites in heavily grazed and heavily lopped areas was significantly lower than in lightly grazed and lightly lopped sites. Similarly, the average abundance of birds (all species combined) was significantly lower in heavily logged sites than lightly logged sites. I found that the effects of forest disturbances on forest birds varied with foraging guild (Chapter 3). For example, the average species richness and abundance of bark-gleaning insectivores were significantly lower in heavily logged sites, indicating that this group of birds was strongly affected by the removal of large trees. Likewise, the average species richness and abundance of foliage-gleaning insectivores were significantly lower in heavily lopped sites. This shows that lopping activities that involve the removal of tree branches for fuelwood and the canopy layer for fodder heavily affected foliage-gleaning insectivores (Chapter 3, Shahabuddin and Kumar 2006, Leal et al. 2013). Similarly, birds that forage on the canopy layer were affected by excessive pruning of mid- and upper-storey vegetation, probably because of the loss of foraging substrates. I concluded that the several species including forest-specialist species tended to be more severely affected by forest disturbance.

In this study, I found that overall richness and abundance of birds positively related to the large tree density, tree canopy cover and shrub density. However, the relationship between bird species and local-scale habitat characteristics differed among foraging guilds (Chapter 2, Chapter 3). For example, species of birds such as bark-gleaning and foliage-gleaning

95 insectivores that require large trees for their feeding and nesting were strongly related to the density of large tress and tree canopy cover for their survival (Chapter 2, Chapter 3, Adams and Morrison 1993, Laiolo et al. 2004, Galitsky and Lawler 2015). The tree canopy cover and shrub density was positively influenced the richness and abundance of foliage-gleaning insectivores (Chapter 3).

As I outlined in Chapter 3, excessive resource extraction is the major cause of habitat degradation in the lowland Terai forests. As richness and abundance of forest birds are related to local-scale habitat structure (e.g Chettri et al. 2002, Guenette and Villard 2005), the simplification of habitat as a result of forest disturbance is likely to affect the bird assemblages within a site (Cleary et al. 2007, Maron and Kennedy 2007, Greve et al. 2011, Chapter 3). However, in this study, I found that the effects of forest disturbance on birds were not restricted only to the site level, as higher intensities of disturbance also negatively affected the landscape-level species richness of birds (Chapter 4). This may have a significant effect on the population of many threatened species which are already in decline in the region (Inskipp et al. 2013). Simplification of habitat structure further threatens the last remaining population of species such as the blue–eared barbet Megalaima australis, Abbott’s babbler Malacocincla abbotti and the greater flameback Chrysocolaptes lucidus in the lowland Terai forests. These bird species were not recorded in sites in heavily-disturbed areas during my survey period. It is therefore critical to introduce effective conservation measures that help reduce the current rate of exploitation of forest resources in the lowland Terai forests.

5.4 Effects of forest-use practices on bird assemblages vary with the landscape context A large body of research has focused on the effects of site-scale habitat characteristics on species richness and abundance (Chapter 2, Moktan et al. 2009, Khanaposhtani et al. 2012). These studies demonstrated the importance of local habitat features for species and abundance of bird communities. However, in recent years, increasing numbers of studies have shown that species occurrence at a particular site depends not only on site characteristics, but also on characteristics of the landscape in which the site is located (Radford et al. 2005, Haslem and Bennett 2008, Döbert et al. 2014). In this thesis (Chapter 3), I therefore examined the relative importance of site-and landscape characteristics on forest birds in the lowland Terai forest of Nepal, and whether the effects of forest use practices on the forest bird community depend on the extent of forest cover surrounding the sites.

96

As outlined in 5.3 above, the structural features of forest stands such as canopy structure, tree sizes, and shrub density are important influences on avifaunal assemblages (Chapter 2, Chapter 3, Guenette and Villard 2005, Bakermans et al. 2012). However, the extent of forest cover on species richness and abundance of birds was a still more important predictor and had a positive effect on all response groups (Chapter 3 and Chapter 4). However, the forest extent was found strongly important to the forest sensitive species. For example, I found that species richness and abundance of bark-gleaners and foliage-gleaner insectivores were significantly related to the extent of forest cover (Chapter 3 and Chapter 4).

Among the most important findings of my thesis is that the extent of forest cover in the landscape not only has direct effects on birds, but also has a significant role in moderating the effects of disturbance on birds at the site-level (Chapter 3) as well as at the landscape-level (Chapter 4). To my knowledge, this is the first study to investigate interactive effects between intensity of disturbances and species richness at the site scale as well as at the landscape scale. I found that effects of disturbance on site-level species richness and abundance of all birds, and site-level species richness and abundance of bark-gleaning and foliage-gleaning insectivores, depended on the extent of forest cover in the surrounding landscape. Furthermore, analysis of landscape-level bird species richness indicated a role of forest extent in moderating the effects of disturbance at the landscape level (Chapter 4). For example, I found that landscape-level species richness of birds was less-affected by disturbance in sites with a greater extent of forest cover in the landscape (Chapter 4). Thus, the richness of birds across landscapes depends on both the extent of forest cover and the level of disturbance in that landscape.

These findings are of significance for conservation management of avifauna, particularly in multiple-use forest landscapes where disturbance levels vary across space. The remnant forests of the region not only support the flora and fauna, but also meet the subsistence demands of food and shelter for people residing near forests. My results show that the effects of subsistence forest resources extraction on avifauna assemblages can be compensated for by maintaining the extent of forests in the landscape (Chapter 3, Chapter 4) which would be a potential win-win outcome for forest-dependent human communities. Therefore, the focus should be on the restoration of forest habitats through a participatory approach to forest conservation that seeks to benefit both biodiversity and the interests of the local people. I

97 believe that these findings point a way forward for resolving conflicts among different stakeholders in forest resources management in lowland Terai forests.

5.5 Relationships between species richness and forest extent vary among bird guilds There has been an increasing interest in understanding the relationship between landscape properties and faunal assemblages in human-dominated landscapes (Radford and Bennett 2007, Maron et al. 2012, Taylor et al. 2012). Studies that focus on site-level responses may not sufficient to characterise the influences on avifauna across the landscape. In Chapter 4, I adopted a whole-of-landscape approach in characterising the relationships between forest birds and extent of forest cover, measuring both response and predictor variables at the landscape-scale (Bennett et al. 2006). I investigated the relative importance of forest extent and other landscape characteristics such as disturbance levels, rainfall and the extent of water bodies in the landscape on estimated richness of all birds, frugivores, foliage-gleaners and sallying insectivores.

The extent of forest cover in a landscape is an important predictor of species-rich assemblages in the landscape (Haslem and Bennett 2008, Zuckerberg and Porter 2010, Maron et al. 2012, Moura et al. 2013, Ochoa‐Quintero et al. 2015, Chapter 3, Chapter 4). In the lowland forests, I found a consistent positive response of landscape-level species richness of birds to the extent of forest cover in the landscape. Moreover, the relationships between landscape-level species richness and the extent of forest cover were not uniform across the foraging guilds. For example, I found that the richness of foliage-gleaning insectivores and, to a lesser extent, frugivores, were strongly related to the extent of forest cover, whereas richness of sallying insectivores was only weakly related to forest cover. The strong positive relationships between foliage-gleaning insectivores and frugivores, and forest cover in the landscape, may be due to the greater diversity of tree species in landscapes with more forest cover, which in turn means more diverse foraging substrates and fruit sources.

The nonlinearity of the response of landscape-scale species richness of bird to the extent of forest cover in the landscape was another important finding of this study (Chapter 4). As the extent of forest cover in the landscape decreased, species richness also decreased in the landscape, but this was steepest where forest cover in the landscape was below 20-30%. However, in contrast to other empirical studies (e.g. Martensen et al., Maron et al. 2012), I

98 found that species richness of birds continued to steadily increase up to 100% forest cover, rather than plateauing (Chapter 4).

Chapter 4 provides clear evidence of the role of forest extent on persistence of forest bird in lowland Terai forest. As I found, a major change in species richness occurred when forest cover in the landscape declined to approximately 20-30% of the landscape. Although I did not find evidence of thresholds in relationships between species richness and forest cover, the non-linearity response of birds to loss of forest cover in the landscapes suggests that lowland forest landscapes face a high risk of collapse in avifaunal richness if further forest loss and modification occur. This information may help guide forest managers and other relevant authorities to set a conservation target for the protection and restoration of forest in the landscape to avoid any further loss and decline of species from the region.

5.6 Management implications My study illustrated potential drivers of forest disturbances and consequent effects on habitat structure and avifaunal communities at different spatial scales in lowland Terai forests. On the basis of my findings, I have three major management recommendations. These relate to: 1) the maintenance of complexity of forest stand structure, in particular, retention of large trees, canopy cover and shrub density; 2) the protection and restoration of forest extent in the landscape; and 3) off-reserve forest conservation through community forestry approaches.

At the local scale, large tree density, tree canopy cover and shrub density were the most important habitat features that determined the species richness and abundance of all birds, bark-gleaners and foliage-gleaning insectivores (Chapter 2, Chapter 3). These habitat features were key for forest specialist species such as green-billed Phaenicophaeus tristis, grey-headed woodpecker, Abbot’s babbler and the greater flameback. As I outlined in Chapter 3, forest-use practices such as logging, grazing and lopping have contributed to a reduction of the structural complexity of vegetation, but particularly in state-managed forests. To optimize species richness at the site level, extraction of forest biomass such as standing trees, snags, woody debris and large tress should be reduced and strictly regulated. Similarly, retention of larger trees and tree canopy cover through habitat restoration activities should be strategy focus for forest conservation.

99

At the landscape-scale, the extent of forest cover was an important predictor of species richness across the landscape. The maintenance of forest in the landscape supports not only species-rich assemblages, but also potentially reduces the impacts of disturbance on forest bird assemblages (Chapter 3, Chapter 4). Furthermore, the protection of a minimum extent of forest cover in the landscape reduces the chance of abrupt species declines (Andren 1994). Native remnants of Sal -dominated forest represent the last remaining habitats for much of biodiversity in lowland landscapes. Although this study was carried out in the lowland Terai forest of Nepal, my findings are relevant to Sal-dominated lowland landscapes of south Asian region, and extend our understanding of landscape-level species-area relationships.

In addition to the forest cover, the extent of water bodies (rivers, creeks, lakes) also influenced species richness. Greater areas of water bodies resulted in higher species richness of all birds and frugivores (Chapter 4). Thus protection of riparian habitat is likely to contribute to maintaining bird species richness and abundance in the landscape. Nepal’s lowland rivers face increased extraction of gravel, sand and boulders, which severely affects water discharge from the systems (Dahal et al. 2012). These activities also pose a threat to the vegetation surrounding extraction sites (BCN and DNPWC 2010) and associated bird communities. It is therefore important that conservation of water resources receives greater emphasis by implementing sustainable river bed extraction planning in the region.

While Nepal has designated about 23% of its land mass as protected areas, the majority of its protected land is concentrated in the high Himalayas and throughout the less-productive landscapes (HMG/MFSC 2002, Heinen and Shrestha 2006). For example, about 48% of high Himalayas are protected, whereas only 0.8% of the Mild Hills and 5.5% of the Terai zone are protected (Shrestha et al. 2010). Thus, only a small proportion of Nepal’s most productive lowland forest is represented in the current protected area network. A large tract of lowland Terai forest is located adjacent to protected areas. It is therefore critical to adopt landscape planning and management strategies focused not only on protected areas, but also on off- reserve forests, which are the last remaining natural habitats in lowland Nepal. Effective management of these forests should not only improve the biodiversity in those forests, but also in the adjacent protected areas (Kindlmann 2011).

Protected areas are subject to a relatively strict management regime, with habitat management and monitoring done by the Department of National Parks and Wildlife

100

Conservation (DNPWC). The DNPWC has its sector offices and range posts in the field, and management is controlled from these field offices. Forest management activities include regular burning, ploughing and uprooting of unwanted tree species for grassland management, control of invasive species in particular management of Mikania micrantha, regular patrolling to prevent illegal logging and hunting of wildlife species. The National Park and Wildlife Conservation Act in 1992 made provision for buffer zones and provided limited rights to local communities to manage forest adjacent to national park boundaries (Paudel et al. 2007), however the buffer zone management committee has no role in management of the protected areas themselves. Protected areas may be more effectively managed if the community living in and using the buffer zone participates in conservation planning and decision-making processes.

State-managed forests occupy a significant proportion of off- reserve forests in lowland Nepal and are centrally managed. The centralized forest management approach limits participation of surrounding communities in management and conservation of forest resources. Unlike protected areas, state-managed forests are managed for production rather than conservation by the Department of Forest (DoF). The harvesting of forest products, in particular, collection of logs for timber and fuelwood, is the primary objective of management of these forests. Excessive extraction of forest products for timber and fuelwood have negative effects on forest condition (e.g. Kanel and Dahal 2008), so it is important to develop and implement a sustainable forest management plan that reduces human pressures on forests to maintain vegetation complexity both at the site- and landscape-scale.

Unlike state-managed and protected area forest tenure types, community managed forests are based on a participatory approach where communities have the rights of access to resources and their management. This approach focuses on the collective management of forest resources to improve both human well-being and biodiversity conservation. Under this management framework, Community Forest User Groups (CFUGs) are formed and management authority is given to these user groups for protection, management and utilization of forest products. Forest restoration and management activities, for example, plantation of trees and silvicultural operations such as thinning and pruning, removal of unwanted weeds and forests patrolling to prevent illegal logging and grazing (Nagendra et al. 2005, Kanel and Dahal 2008), are currently undertaken by the community forest user

101 committees. This initiative does not only contribute to better forest condition, but also offers significant benefits to the local community from forest products (e.g. Kanel and Dahal 2008).

In recent years, community-managed forests have become increasingly recognised for their conservation values for biodiversity (Nagendra and Gokhale 2008, Birendra et al. 2014, Chapter 2). They are important breeding habitats for many threatened bird species (Chapter 2, Inskipp 1989). For instance, the last remaining population of blue–eared barbet (<50 individuals) and Abbott’s babbler (<250 individuals) are in the community-managed forests in eastern Nepal (Inskipp et al. 2013). Moreover, the greater flameback that requires larger trees to breed (Kumar et al. 2011) was recorded only in the community-managed forests during this study (Chapter 2).

However, despite the importance of community forests in faunal conservation, the roll out of community forestry programs in the lowland Terai are progressing slowly because these forests are commercially valuable and also a major source of government revenue (Kanel and Dahal 2008). Thus, only 10% of the Terai forests have transferred to community management (Kanel and Dahal 2008), compared to 24% in the hill regions of Nepal (Bhattarai 2006). Strengthening community forestry programs can not only ameliorate habitat loss and degradation (Gautam et al. 2004; Kanel and Dahal 2008) but also generate livelihood opportunities for surrounding communities and reduce pressure on protected areas (Straede et al. 2002). I therefore recommend that the extension of community forestry programs should be prioritized and given strong and urgent Government support in order to minimise further habitat deterioration.

102

Table 5.1 Summary of study themes and key findings of this thesis

Theme Key findings Chapter Conservation 1. Off-reserve forests, in particular, community forests, have complementary bird assemblages to 2 values of off- protected areas, supporting species not represented in formal conservation reserves. reserve forests 2. Habitat features of sites in protected areas and community forests were relatively more similar, with no 2 differences in tree canopy cover, large mature tree density and shrub density. 3. Beta diversity (i.e. species turnover) among sites in off-reserve forest tenures was higher than among 2 sites in protected areas. 4. Local habitat characteristics such as tree density, shrub cover, tree canopy cover and mature tree 2,3 density had strong influence on bird communities. Management implications 1. Strengthen community forestry programs across off-reserve forests in lowland landscapes. 2. Provide technical, managerial and organizational support for institutionalization of community forests in particular those they are newly formed. 3. Develop and implement sustainable forest extraction guidelines across the off-reserve forests to prevent further degradation of local habitat characteristics. Effects of forest 1. Logging, grazing and lopping activities had deleterious effects on vegetation structure and associated 2,3 use practices on avifaunal communities in the lowland landscape. vegetation and 2. The density and basal area of large trees and tree canopy cover were significantly lower in heavily- associated logged sites than in lightly-logged sites. Lopping activities significantly affected shrub density and 3

103 avifaunal shrub cover, while grazing activities reduced herbaceous cover. communities 3. The overall species richness of birds in sites in heavily grazed and heavily lopped areas was significantly lower than in lightly grazed and lightly lopped sites. Similarly, the average abundance of 3 birds was significantly lower in heavily logged sites than lightly logged sites. 4. The effects of forest disturbance on birds were not restricted only to the site level, as higher intensities of disturbance also negatively affected the landscape-level species richness of birds. 4 Management implications 1. Develop and implement a sustainable forest management plan that reduces human pressures on forests to maintain vegetation complexity at the site-scale. 2. Excessive grazing, over extraction of fire wood and fodder, lopping and removal of tree canopy should be control with a regulatory mechanism. 3. Retention of larger trees and tree canopy cover in the landscape through habitat restoration activities should be a focus for forest conservation. Relative effects of 1. The structural features of forest stands such as canopy structure, tree sizes, and shrub density had a 2,3 site-and landscape- significant positive influence on avifaunal assemblages. level factors on 2. Species of birds such as bark-gleaning and foliage-gleaning insectivores that require large trees for bird communities their feeding and nesting were strongly related to the density of large trees and tree canopy cover for 2,3 their survival. 3. The extent of forest cover in the landscape had a strong positive influence on species richness and abundance for all bird response groups. 2,3,4

104

Management implications 1. The extraction of forest biomass such as standing trees, snags, woody debris and large tress should be reduced and strictly regulated to optimize species richness at the site level. 2. Development and effective implementation of policies to restore degraded forests in lowland Terai in particular Parsa and eastern landscapes is urgent. 3. Maintain existing forest cover as it boosts landscape-level species richness in the region. Effects of 1. The extent of forest cover within a 500 m and a 2.5 km radius of each survey site had a strong positive 3,4 landscape context influence on species richness and abundance for all bird response groups. and interaction 2. The strongest effect of forest cover on all species richness occurred in landscapes with a greater extent 3,4 effects on bird of disturbance, indicating a role of forest extent in moderating the effects of disturbance. communities 3. The species richness of foliage-gleaning insectivores across the landscape was also dependent on the 4 interaction between rainfall and forest cover, suggesting that the extent of forest cover has less effect on foliage-gleaning insectivores in higher-rainfall landscapes. Management implications 1. Habitat restoration and maintenance of forest extent should be prioritized in areas of low forest cover particularly in landscapes of eastern lowland. 4. Develop and implement a sustainable forest management plan that reduces human pressures on forests to maintain vegetation complexity both at the site- and landscape-scale.

105

Relationship 1. The landscape-level species richness of forest birds was consistently positively related to the extent of 3,4 between species forest cover in the landscape. richness and forest 2. However, the strength of the relationship varied substantially among foraging guilds, with effects 4 extent vary with strongest for foliage-gleaning and frugivores. 4 foraging guilds 3. The relationship between species richness and extent of forest cover was somewhat nonlinear, with species richness decreasing more steeply below about 20-30% forest cover in the landscape. Management implications 1. Habitat restoration and maintenance of forest extent should be prioritized in areas of low forest cover and high degree of deforestation to prevent further loss of avifaunal populations in the region. 2. Supporting community forestry program across the off-reserve forests of lowland region may prevent deforestation and increase forest extent through active restoration and tree planting.

106

5.7 Limitations of my research There are number of limitations affecting this study. One limitation was the difficulty in selection of accessible sites and landscapes proportionally across state forests, community forests, and protected areas in lowland Terai forests. However, accessibility to all parts of protected areas was impossible due to transportation challenges and also risks from wild , particularly the greater one horned-rhinoceros Rhinoceros unicornis and the Asian wild elephant Elephas maximus. This reduced the opportunity for study landscapes within the protected areas. To address this limitation, some sites were selected from within buffer zone forests. The buffer zone forests are forests at the immediate periphery of protected area that have been managed by the protected area management authority and habitat conditions are largely similar to those of protected areas.

A second limitation of this study was number of bird survey sites that I was able to visit within each landscape. I established a total of four bird survey sites in each landscape. This number may be inadequate for full characterisation of the avifauna of landscapes, particularly in the context of large landscape size (5 x 5 km). However, survey effort was equal among landscapes, with at least one site located in each quadrant of the landscape, and species richness estimation was used to correct for differences in survey completeness among landscapes. Another limitation was that I was only able to survey birds from November to May, which covered only winter and early-summer bird assemblages. Thus, species that were summer visitors were poorly captured by this study. These limitations should be considered in future research.

5.8 Future research This study revealed consistent relationships between the extent of forest cover and most groups of birds. However, the observed pattern of consistent positive relationships may not be due solely to independent effects of forest cover on the landscape. Such relationships may be due to a correlation between forest extent and the diversity of resources available. Previous studies have shown that larger areas of forest in the landscape support diverse vegetation types and therefore provide an array of resources across a range of biodiversity including avifauna (Miller et al. 1997, Williams et al. 2002). Although lowland forest is dominated by the Shorea robusta species, it also contains other vegetation types such as riverine vegetation, mixed hardwood forest, and different gradients of habitat conditions (Joshi et al. 2003). Thus, the lowland landscape is a heterogeneous mosaic of different

107 vegetation types that in turn, may have strong influences on forest bird assemblages. Therefore future studies on bird assemblages in the lowland forests should examine these potential mechanisms behind the patterns I observed.

In this study, I found that the relationship between species richness and extent of forest cover was non-linear, with species richness decreasing more steeply below about 20-30% forest cover in the landscape (Chapter 4). However, recent empirical research by Maron et al. (2012) has shown that the response of species to habitat cover may vary across different landscape types. The study landscapes differed in terms of rainfall, soil types and topography. Thus further research is required to investigate the relationships between landscape-scale forest extent and forest bird assemblages across different landscape types/or landscape productivity. Such studies would require greater replication of landscapes, but would further inform managers in setting management targets for maintenance of forest cover to provide effective protection of biodiversity.

I compared species richness and abundance and community composition of forest birds among differently-managed forest such as protected areas, state forests and community forests (Chapter 2). Although the main question here was to explore whether the off-reserve forest support complementary bird assemblages in the lowland Terai forests, a secondary interest was to investigate the role of community forestry in bird conservation. I found that community forests support richer assemblages than state-managed forests (Chapter 2). However, in this study, I did not examine the effects of different durations of community forest management on forest bird assemblages. The protected areas and state-managed forests often have a longer history of similar forest management than the community-based forests. The forest tenure transfers to community have occurred at different intervals over time in lowland Terai forests since 1990. It is therefore important to evaluate the effectiveness of participatory forest management practices for bird conservation across different intervals of time since the tenure rights transferred to the local communities. This would allow documentation of the conservation outcomes of community managed forests across time and space.

5.9 Conclusion The remnant lowland Terai forests are subject to intense anthropogenic pressures due to agricultural intensification and forestry practices. These human-induced disturbances

108 threatened ecological communities. However, the influence of landscape modifications on habitat features and their constituent biota at different spatial scales is poorly understood in the region. Biodiversity conservation in such landscapes depends on the effective conservation of remnant forests and associated habitat attributes. Thus, identifying how species respond to habitat characteristics at different spatial is crucial for conservation of avifauna in the region.

This thesis makes an important contribution to our understanding of relative effects of habitat characteristics on bird assemblages at multiple spatial scales in human-dominated multiple- use landscapes. It provides a broader picture about the effects of forest disturbances on species assemblages as result of subsistence forest-use practices, reveals a significant role of landscape-level forest extent to reduce these effects on bird assemblages, and describes the nature and shape of species and forest extent relationships for birds of different functional groups in the lowland Terai forests. This study therefore provides important ecological information for forest managers and other stakeholders seeking to achieve landscape-level conservation of avifauna in the region.

109

REFERENCES Adams, E.M., Morrison, M.L., 1993. Effects of Forest Stand Structure and Composition on Red-Breasted Nuthatches and Brown Creepers. The Journal of Wildlife Management, 57(3): 616-629. Aleixo, A. 1999. Effects of selective logging on a bird community in the Brazilian Atlantic forest. The Condor 101:537-548. Alexander, J. D., J. L. Stephens, and N. E. Seavy. 2008. Livestock utilization and bird community composition in mixed- forest and oak woodland in southern Oregon. Northwest Science 82:7-17. Ali, S., Ripley, S.D., 1983. Handbook of the birds of the Indian subcontinent. Compact edition. Oxford University Press, Delhi. Ali, S., S. D. Ripley, and J. H. Dick. 1987. Compact handbook of the birds of India and Pakistan. Anand, M. O., J. Krishnaswamy, and A. Das. 2008. Proximity to forests drives bird conservation value of coffee plantations: implications for certification. Ecological Applications 18:1754-1763. Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology 26: 32-46. Anderson, M.J., Ellingsen, K.E.,McArdle, B.H., 2006. Multivariate dispersion as a measure of beta diversity. Ecology letters, 9(6): 683-93. Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA for PRIMER: guide to software and statistical methods.PRIMER–E Ltd., Plymouth, United Kingdom, 214 pp. Andren, H. 1994. Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. Oikos:355-366. Bailis, R., J. L. Chatellier, and A. Ghilardi. 2012. Ecological sustainability of woodfuel as an energy source in rural communities. Pages 299-325 Integrating Ecology and Poverty Reduction. Springer. Bakermans, M. H., A. D. Rodewald, and A. C. Vitz. 2012. Influence of forest structure on density and nest success of mature forest birds in managed landscapes. The Journal of Wildlife Management 76:1225-1234. Balmford, A., R. Green, and B. Phalan. 2012. What conservationists need to know about farming. Proceedings of the Royal Society B: Biological Sciences:rspb20120515.

110

Baral, H. S., A. K. Ram, B. Chaudhary, S. Basnet, H. Chaudhary, T. R. Giri, and D. Chaudhary. 2012. Conservation status of Bengal Florican Houbaropsis bengalensis bengalensis (Gmelin, 1789)(Gruiformes: Otididae) in Koshi Tappu Wildlife Reserve and adjoining areas, eastern Nepal. Journal of Threatened Taxa 4:2464-2469. Baral, H., B. Sahgal, S. Mohsanin, K. Namgay, and A. Khan. 2014. Species and habitat conservation through small locally recognised and community managed Special Conservation Sites. Journal of Threatened Taxa 6:5677-5685. Baral, H.S., Inskipp, C., 2005. Important Bird Areas in Nepal: key sites for conservation. Bird Conservation Nepal and BirdLife International,Kathmandu and Cambridge.

Baral, H.S.,Inskipp, C., 2004. The State of Nepal’s Birds 2004. Department of National Parks and Wildlife Conservation, Bird Conservation Nepal & IUCN Nepal, Kathmandu, 64pp. Barr, L. M., R. L. Pressey, R. A. Fuller, D. B. Segan, E. McDonald-Madden, and H. P. Possingham. 2011. A new way to measure the world's protected area coverage. PLoS One 6:e24707. Bartoń, K. 2012. MuMIn: multi-model inference. R package version 1. Bastian, O., D. Haase, and K. Grunewald. 2012. Ecosystem properties, potentials and services–the EPPS conceptual framework and an urban application example. Ecological Indicators 21:7-16. Bates, D., M. Maechler, B. Bolker, and S. Walker. 2014. lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1-7, http://CRAN.R- project.org/package=lme4. BCN and DNPWC. 2010. The State of Nepal’s Birds 2010. Bird Conservation Nepal and Department of National Parks and Wildlife Conservation, Kathmandu. BCN, DNPWC, 2012. Conserving biodiversity and delivering ecosystem services at Important Bird Areas in Nepal. Kathmandu and Cambridge, UK: Bird Conservation Nepal,Department of National Parks and Wildlife Conservation, and BirdLife International. Beck, N. and J. N. Katz. 1995. What to do (and not to do) with time-series cross-section data. American political science review 89:634-647. Becker, C. G., Fonseca, C. R., Haddad, C. F. B., Batista, R. F. and Prado, P. I. 2007. Habitat split and the global decline of amphibians. Science, 318(5857), 1775-1777.

111

Bennett, A. F. and D. A. Saunders. 2010. Habitat fragmentation and landscape change. Conservation Biology for All 93:1544-1550. Bennett, A. F., D. G. Nimmo, and J. Q. Radford. 2014. Riparian vegetation has disproportionate benefits for landscape‐scale conservation of woodland birds in highly modified environments. Journal of Applied Ecology 51:514-523. Bennett, A. F., J. Q. Radford, and A. Haslem. 2006. Properties of land mosaics: implications for nature conservation in agricultural environments. Biological Conservation 133:250-264. Berry, N. J., Phillips, O. L., Ong, R. C., and Hamer, K. C. 2008. Impacts of selective logging on tree diversity across a rainforest landscape: the importance of spatial scale. Landscape ecology, 23(8), 915-929. Betts, M. G., G. J. Forbes, and A. W. Diamond. 2007. Thresholds in songbird occurrence in relation to landscape structure. Conservation Biology 21:1046-1058. Bhagwat, S.A., Kushalappa, C.G., Williams, P.H., Brown, N.D., 2005. The Role of Informal Protected Areas in Maintaining Biodiversity in the of India. Ecology and Society, 10(1): 1-42. Bhagwat, S.A.,Rutte, C., 2006. Sacred Groves: Potential for Biodiversity Management. Frontiers in Ecology and the Environment, 4(10): 519-524. Bhattarai, B. 2006. Widening the gap between terai and hill farmers in Nepal: the implications of the New Forest Policy 2000. In: Mahanty, S., Fox, J., Nurse, M.,Stephen, P., McLees, L., (Eds.), Hanging in the Balance: Equity in Community- Based Natural Resource Management in Asia, RECOFTC, Bangkok, Thailand, and East-West Center, Honolulu, Hawaii, USA. Brawn, J.D., Robinson, S.K., Thompson, F.R., 2001. Bhusal, N. P. 2014. Buffer Zone Management System in Protected Areas of Nepal. The Third Pole: Journal of Geography Education 11:34-44. Bibby, C. J. 2000. Bird census techniques. Elsevier. BirdLife International. 2004. Threatened birds of the world 2004 CD-ROM. Cambridge, UK: BirdLife International. Birendra, K., A. J. Mohammod, and M. Inoue. 2014. Community Forestry in Nepal’s Terai Region: Local Resource Dependency and Perception on Institutional Attributes. Environment and Natural Resources Research 4:p142.

112

Bjornstad, O. N. 2013. spatial nonparametric covariance functions. Package ‘ncf’. URL http://onb.ent.psu.edu/onb1/R. Blake, J. G. and B. A. Loiselle. 2001. Bird assemblages in second-growth and old-growth forests, Costa Rica: perspectives from mist nets and point counts. The Auk 118:304- 326. Bolker, B. M., M. E. Brooks, C. J. Clark, S. W. Geange, J. R. Poulsen, M. H. H. Stevens, and J.-S. S. White. 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends in ecology & evolution 24:127-135. Borah, N., Athokpam, F. D., Garkoti, S., Das, A. K., and Hore, D. 2014. Structural and compositional variations in undisturbed and disturbed tropical forests of Bhuban hills in south Assam, India. International Journal of Biodiversity Science, Ecosystem Services & Management, 10(1), 9-19. Brawn, J.D., Robinson, S.K., Thompson, F.R., 2001. The role of disturbance in the ecology and conservation of birds. Annual Review of Ecology and Systematics, 32: 251-276. Bried, J. T., K. E. Langwig, A. A. Dewan, and N. A. Gifford. 2011. Habitat associations and survey effort for shrubland birds in an urban pine barrens preserve. Landscape and Urban Planning 99:218-225. Brooks, T. M., M. I. Bakarr, T. Boucher, G. A. Da Fonseca, C. Hilton-Taylor, J. M. Hoekstra, T. Moritz, S. Olivieri, J. Parrish, and R. L. Pressey. 2004. Coverage provided by the global protected-area system: is it enough? BioScience 54:1081-1091. Bruton, M. J. 2014. Multi-scale patterns of habitat use by reptiles in regenerating dryland landscapes. doi:10.14264/uql.2014.84. Buckland, S. T. and C. Handel. 2006. Point-transect surveys for songbirds: robust methodologies. The Auk 123:345-357. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 2005. Distance sampling. Wiley Online Library. Buffum, B., G. Gratzer, and Y. Tenzin. 2009. Forest grazing and natural regeneration in a late successional broadleaved community forest in Bhutan. Mountain Research and Development 29:30-35. Burgman, M., D. Lindenmayer, and J. Elith. 2005. Managing landscapes for conservation under uncertainty. Ecology 86:2007-2017. Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Springer.

113

Butchart, S. H., J. P. Scharlemann, M. I. Evans, S. Quader, S. Arico, J. Arinaitwe, M. Balman, L. A. Bennun, B. Bertzky, and C. Besancon. 2012. Protecting important sites for biodiversity contributes to meeting global conservation targets. PLoS One 7:e32529. Capinera, J. 2011. Insects and wildlife: arthropods and their relationships with wild vertebrate animals. John Wiley & Sons. Central Bureau of Statistics. 2011. Population census 2011.National report. Central Bureau of Statistics, Government of Nepal, Kathmandu, Nepal. Chandler, C. C., D. I. King, and R. B. Chandler. 2012. Do mature forest birds prefer early- successional habitat during the post-fledging period? Forest Ecology and Management 264:1-9. Chao, S. 2012. FOREST PEOPLES: Numbers across the world. Forest Peoples Programme. Chappell, M. J. and L. A. LaValle. 2011. Food security and biodiversity: can we have both? An agroecological analysis. Agriculture and Human Values 28:3-26. Charnley, S. and M. R. Poe. 2007. Community Forestry in Theory and Practice: Where Are We Now?*. Annual Review of Anthropology 36:301. Chettri, N., E. Sharma, D. Deb, and R. Sundriyal. 2002. Impact of firewood extraction on tree structure, regeneration and woody biomass productivity in a trekking corridor of the Sikkim Himalaya. Mountain Research and Development 22:150-158. Chisholm, C., Z. Lindo, and A. Gonzalez. 2011. Metacommunity diversity depends on connectivity and patch arrangement in heterogeneous habitat networks. Ecography 34:415-424. Christensen, M., S. Rayamajhi, and H. Meilby. 2009. Balancing fuelwood and biodiversity concerns in rural Nepal. Ecological Modelling 220:522-532. Clark, N. E., E. H. Boakes, P. J. McGowan, G. M. Mace, and R. A. Fuller. 2013. Protected Areas in South Asia Have Not Prevented Habitat Loss: A Study Using Historical Models of Land-Use Change. PLoS One 8:e65298. Clarke, K.R., 1993. Non-parametric multivariate analysis of changes in community structure. Australian Journal of Ecology 18: 117-143. Clarke, K.R., Warwick, R.M., 2001. Change in marine communities: an approach to statistical analysis and interpretation. 2nd Edn. (PRIMER-E: Plymouth.). Cleary, D. F., T. J. Boyle, T. Setyawati, C. D. Anggraeni, E. E. V. Loon, and S. B. Menken. 2007. Bird species and traits associated with logged and unlogged forest in Borneo. Ecological Applications 17:1184-1197.

114

Cleveland, W. S. and S. J. Devlin. 1988. Locally weighted regression: an approach to regression analysis by local fitting. Journal of the American Statistical Association 83:596-610. Colwell, R. 2013. Estimates 9.1. 0 User's Guide. Storrs, CT: Department of Ecology and Evolutionary Biology, University of Connecticut. Condit, R., N. Pitman, E. G. Leigh, J. Chave, J. Terborgh, R. B. Foster, P. Núnez, S. Aguilar, R. Valencia, and G. Villa. 2002. Beta-diversity in tropical forest trees. Science 295:666-669. Cox, R.L., Underwood, E.C., 2011. The Importance of Conserving Biodiversity Outside of Protected Areas in Mediterranean Ecosystems. PLoS ONE, 6(1): e14508. Crawley, M. J. 2007. Mixed‐Effects Models. The R Book, Second Edition:681-714. Cushman, S. A. and K. McGarigal. 2003. Landscape-level patterns of avian diversity in the Oregon Coast Range. Ecological Monographs 73:259-281. Dahal, B. R., C. A. McAlpine, and M. Maron. 2014. Bird conservation values of off-reserve forests in lowland Nepal. Forest Ecology and Management 323:28-38. Dahal, B. R., P. J. McGowan, S. J. Browne. 2009. An assessment of census techniques, habitat use and threats to Swamp Francolin Francolinus gularis in Koshi Tappu Wildlife Reserve, Nepal. Bird Conservation International 19:137-147. Dahal, K. R., S. Sharma, and C. M. Sharma. 2012. A review of riverbed extraction and its effects on aquatic environment with special reference to Tinau River, Nepal. Hydro Nepal: Journal of Water, Energy and Environment 11:49-56. Davies, R. B. 1987. Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika 74:33-43. De La Montana, E., Rey-Benayas, J.M.,Carrascal, L.M., 2006. Response of bird communities to silvicultural thinning of Mediterranean maquis. Journal of Applied Ecology, 43(4): 651-659. Deconchat, M., Brockerhoff, E., and Barbaro, L. 2009. Effects of surrounding landscape composition on the conservation value of native and exotic habitats for native forest birds. Forest Ecology and Management, 258, S196-S204. Deconchat, M., E. Brockerhoff, and L. Barbaro. 2009. Effects of surrounding landscape composition on the conservation value of native and exotic habitats for native forest birds. Forest Ecology and Management 258:S196-S204.

115

DeFries, R. S., J. A. Foley, and G. P. Asner. 2004. Land-use choices: balancing human needs and ecosystem function. Frontiers in Ecology and the Environment 2:249-257. Dennis, P., J. Skartveit, D. I. McCracken, R. J. Pakeman, K. Beaton, A. Kunaver, and D. M. Evans. 2008. The effects of livestock grazing on foliar arthropods associated with bird diet in upland grasslands of Scotland. Journal of Applied Ecology 45:279-287. Díaz, I. A., J. J. Armesto, S. Reid, K. E. Sieving, and M. F. Willson. 2005. Linking forest structure and composition: avian diversity in successional forests of Chiloé Island, Chile. Biological Conservation 123:91-101. Döbert, T. F., B. L. Webber, A. D. Barnes, K. J. Dickinson, R. K. Didham, C. Kettle, and L. Koh. 2014. 4 Forest Fragmentation and Biodiversity Conservation in Human- dominated Landscapes. Global Forest Fragmentation:28. DoF. 2009. Community forestry database.Ministry of Forest and Soil Conservation. Kathmandu, Nepal. Dudley, N. and A. Phillips. 2006. Forests and Protected Areas: Guidance on the use of the IUCN protected area management categories. IUCN, Gland, Switzerland and Cambridge, UK. x + 58pp. Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect populations in complex landscapes. Oikos:169-175. Ellis, E. A. and L. Porter-Bolland. 2008. Is community-based forest management more effective than protected areas?: A comparison of land use/land cover change in two neighboring study areas of the Central Yucatan Peninsula, Mexico. Forest Ecology and Management 256:1971-1983. Ellis, E. C. and N. Ramankutty. 2008. Putting people in the map: anthropogenic biomes of the world. Frontiers in Ecology and the Environment 6:439-447. Estavillo, C., R. Pardini, and P. L. B. da Rocha. 2013. Forest Loss and the Biodiversity Threshold: An Evaluation Considering Species Habitat Requirements and the Use of Matrix Habitats. PLoS One 8:e82369. Ewers, R. M. and R. K. Didham. 2006. Confounding factors in the detection of species responses to habitat fragmentation. Biological Reviews 81:117-142. Eyre, T. J., M. Maron, M. T. Mathieson, and M. Haseler. 2009. Impacts of grazing, selective logging and hyper‐aggressors on diurnal bird fauna in intact forest landscapes of the Brigalow Belt, Queensland. Austral Ecology 34:705-716. Fahrig, L. 2001. How much habitat is enough? Biological Conservation 100:65-74.

116

Fahrig, L. 2003. Effects of habitat fragmentation on biodiversity. Annual review of ecology, evolution, and systematics:487-515. FAO. 2005. Food and Agriculture Organization of the United Nations; Rome, Italy: 2005. Global forest resource assessment. Fischer, J. and D. B. Lindenmayer. 2007. Landscape modification and habitat fragmentation: a synthesis. Global Ecology and Biogeography 16:265-280. Flynn, D. F., M. Gogol‐Prokurat, T. Nogeire, N. Molinari, B. T. Richers, B. B. Lin, N. Simpson, M. M. Mayfield, and F. DeClerck. 2009. Loss of functional diversity under land use intensification across multiple taxa. Ecol Lett 12:22-33. Foley, J. A., R. DeFries, G. P. Asner, C. Barford, G. Bonan, S. R. Carpenter, F. S. Chapin, M. T. Coe, G. C. Daily, and H. K. Gibbs. 2005. Global consequences of land use. Science 309:570-574. Forsman, J. T., P. Reunanen, J. Jokimäki, M. Mönkkönen. 2010. The effects of small-scale disturbance on forest birds: a meta-analysis. Can. J. For. Res. , 40: 1833-1842 Fox, J., S. Weisberg, D. Bates, and M. J. Fox. 2012. Package ‘car’. R Foundation for Statistical Computing, Vienna, Austria. FRA/DFRS. 2014. Terai Forests of Nepal (2010 – 2012). Babarmahal, Kathmandu: Forest Resource Assessment Nepal Project/Department of Forest Research and Survey. Department of Forest Research and Survey, P.O. Box 3339, Babarmahal, Kathmandu, Nepal. Franklin, A. B., B. R. Noon, and T. L. George. 2002. What is habitat fragmentation? Studies in Avian Biology 25:20-29. Galitsky, C. and J. J. Lawler. 2015. Relative influence of local and landscape factors on bird communities vary by species and functional group. Landscape Ecology:1-13. Gardner, T. A., J. Barlow, R. Chazdon, R. M. Ewers, C. A. Harvey, C. A. Peres, and N. S. Sodhi. 2009. Prospects for tropical forest biodiversity in a human‐modified world. Ecol Lett 12:561-582. Gauch, H.G., 1982. Multivariate Analysis in Community Ecology. Cambridge University Press, Cambridge, England. 298 pages. Chinese edition 1989. Gautam, A. P. 2007. Group size, heterogeneity and collective action outcomes: Evidence from community forestry in Nepal. International Journal of Sustainable Development & World Ecology 14:574-583.

117

Gautam, A. P., G. P. Shivakoti, and E. L. Webb. 2004. Forest cover change, physiography, local economy, and institutions in a mountain watershed in Nepal. Environmental Management 33:48-61. Gautam, A., G. Shivakoti, and E. Webb. 2004. A review of forest policies, institutions, and changes in the resource condition in Nepal. International Forestry Review 6:136-148. Gelman, A., Y.-S. Su, M. Yajima, J. Hill, M. Grazia Pittau, J. Kerman, and T. Zheng. 2012. Arm: data analysis using regression and multilevel/hierarchical models. R package v. 1.5–02. Gibbs, H. K., A. Ruesch, F. Achard, M. Clayton, P. Holmgren, N. Ramankutty, and J. Foley. 2010. Tropical forests were the primary sources of new agricultural land in the 1980s and 1990s. Proceedings of the National Academy of Sciences 107:16732-16737. Gil-Tena, A., O. Torras, and S. Saura. 2008. Relationships between forest landscape structure and avian species richness in NE Spain. Ardeola 55:27-40. Graham, C. H. andBlake, J. G. 2001. Influence of patch-and landscape-level factors on bird assemblages in a fragmented tropical landscape. Ecological Applications, 11(6), 1709-1721. Gregory, R. D., D. W. Gibbons, and P. F. Donald. 2004. Bird census and survey techniques. Bird ecology and conservation:17-56. Greve, M., S. L. Chown, B. J. van Rensburg, M. Dallimer, and K. J. Gaston. 2011. The ecological effectiveness of protected areas: a case study for South African birds. Animal Conservation 14:295-305. Grimmett, R., C. Inskipp, and T. Inskipp. 1999. Guide to the birds of India, Pakistan, Nepal, , Bhutan, , and the Maldives. Grimmett, R., C. Inskipp, and T. Inskipp. 2000. Birds of Nepal. Christopher Helm. Grimmett, R., Inskipp, C. and Inskipp, T., 2009. Birds of Nepal. Om Books International. Guenette, J. S. and M. A. Villard. 2005. Thresholds in forest bird response to habitat alteration as quantitative targets for conservation. Conservation Biology 19:1168- 1180. Gurung, A., Bryesecon, I., Oh, S., 2011. Micro hydropower: a promising decentralized renewable technology and its impact on rural livelihoods. Scientific Research and Essays 6(6): 1240-1248. Haddad, N. M., G. M. Crutsinger, K. Gross, J. Haarstad, J. M. Knops, and D. Tilman. 2009. Plant species loss decreases arthropod diversity and shifts trophic structure. Ecol Lett 12:1029-1039.

118

Halley, J. M. and Y. Iwasa. 2011. Neutral theory as a predictor of avifaunal extinctions after habitat loss. Proceedings of the National Academy of Sciences 108:2316-2321. Hansen, A.J.,DeFries, R., 2007. Ecological mechanisms linking protected areas to surrounding lands. Ecological Applications, 17(4): 974-988. Harvey, C. A., A. Medina, D. M. Sánchez, S. Vílchez, B. Hernández, J. C. Saenz, J. M. Maes, F. Casanoves, and F. L. Sinclair. 2006. Patterns of animal diversity in different forms of tree cover in agricultural landscapes. Ecological Applications 16:1986-1999. Haslem, A. and A. F. Bennett. 2008. Birds in agricultural mosaics: the influence of landscape pattern and countryside heterogeneity. Ecological Applications 18:185-196. Hayes, T. and E. Ostrom. 2005. Conserving the world's forests: Are protected areas the only way. Ind. L. Rev. 38:595. Hegde, R. and T. Enters. 2000. Forest products and household economy: a case study from Mudumalai Wildlife Sanctuary, Southern India. Environmental Conservation 27:250- 259. Heikkinen, R. K., M. Luoto, R. Virkkala, and K. Rainio. 2004. Effects of habitat cover, landscape structure and spatial variables on the abundance of birds in an agricultural– forest mosaic. Journal of applied ecology 41:824-835. Heinen, J. T. and S. K. Shrestha. 2006. Evolving policies for conservation: an historical profile of the protected area system of Nepal. Journal of Environmental Planning and Management 49:41-58. Hinsley, S. A. 2000. The costs of multiple patch use by birds. Landscape Ecology 15:765- 775. HMG/MFSC. 2002. Nepal Biodiversity Strategy. Supported by the Global Environment Facility (GEF) and UNDP. Kathmandu, Nepal: HMG. Hoekstra, J. M., T. M. Boucher, T. H. Ricketts, and C. Roberts. 2005. Confronting a biome crisis: global disparities of habitat loss and protection. Ecol Lett 8:23-29. Hrabovszky, J. P. and K. Miyan. 1987. Population growth and land use in Nepal" the great turnabout". Mountain Research and Development:264-270. Hu, G., J. Wu, K. J. Feeley, G. Xu, and M. Yu. 2012. The effects of landscape variables on the species-area relationship during late-stage habitat fragmentation. PLoS One 7:e43894. Huggett, A. J. 2005. The concept and utility of ‘ecological thresholds’ in biodiversity conservation. Biological Conservation 124:301-310.

119

Inskipp, C. 1989. Nepal’s Forest Birds: Their Status and Conservation. ICBP. Monograph No 4. International Council for Bird Preservation, Cambridge, UK. Inskipp, C., Baral, H., Inskipp, T.,Stattersfield, A., 2013. The state of Nepal birds 2010. Journal of Threatened Taxa, 5(1): 3473-3503. IUCN. 2009. Red List of Threatened Species. International Union for Conservation of Nature, Cambridge, UK. Iwata, T., S. Nakano, and M. Murakami. 2003. Stream meanders increase insectivorous bird abundance in riparian deciduous forests. Ecography 26:325-337. Jackson, J. K., C. Stapleton, and J.-P. Jeanrenaud. 1994. Manual of afforestation in Nepal. Jacoby, W. G. 2000. Loess:: a nonparametric, graphical tool for depicting relationships between variables. Electoral Studies 19:577-613. Järvinen, O. and R. Väisänen. 1975. Estimating relative densities of breeding birds by the line transect method. Oikos:316-322. Jenkins, C. N. and L. Joppa. 2009. Expansion of the global terrestrial protected area system. Biological Conservation 142:2166-2174. Jetz, W., C. H. Sekercioglu, and J. E. Watson. 2008. Ecological correlates and conservation implications of overestimating species geographic ranges. Conservation Biology 22:110-119. Johnson, M., Reich, P., Nally, R.M., 2007. Bird assemblages of a fragmented agricultural landscape and the relative importance of vegetation structure and landscape pattern. Wildlife Research, 34: 185-193. Joppa, L. N. and A. Pfaff. 2011. Global protected area impacts. Proceedings of the Royal Society B: Biological Sciences 278:1633-1638. Joppa, L.N., Loarie, S.R.,Pimm, S.L., 2008. On the protection of “protected areas”. Proceedings of the National Academy of Sciences, 105(18): 6673-6678. Joshi, A. R., M. Shrestha, J. L. Smith, and S. Ahearn. 2003. Forest classification of Terai Arc Landscape (TAL) based on Landsat7 satellite data. A Final Report submitted WWF- US. Kallimanis, A. S., A. D. Mazaris, J. Tzanopoulos, J. M. Halley, J. D. Pantis, and S. P. Sgardelis. 2008. How does habitat diversity affect the species–area relationship? Global Ecology and Biogeography 17:532-538. Kallimanis, A. S., A. D. Mazaris, J. Tzanopoulos, J. M. Halley, J. D. Pantis, and S. P. Sgardelis. 2008. How does habitat diversity affect the species–area relationship? Global Ecology and Biogeography 17:532-538.

120

Kanel, K. R. and G. R. Dahal. 2008. Community forestry policy and its economic implications: an experience from Nepal. International Journal of Social Forestry 1:50- 60. Kessler, M., S. Abrahamczyk, M. Bos, D. Buchori, D. D. Putra, S. R. Gradstein, P. Höhn, J. Kluge, F. Orend, and R. Pitopang. 2009. Alpha and beta diversity of plants and animals along a tropical land-use gradient. Ecological Applications 19:2142-2156. Kessler, W. B. and K. A. Milne. 1982. Morning versus evening detectability of southeast Alaskan birds. Condor 84:447-448. Kettle, C. J. and L. P. Koh. 2014. Global Forest Fragmentation. CABI. Khanaposhtani, M. G., M. Kaboli, M. Karami, and V. Etemad. 2012. Effect of habitat complexity on richness, abundance and distributional pattern of forest birds. Environ Manage 50:296-303. Kindlmann, P. 2011. Himalayan biodiversity in the changing world. Springer Science & Business Media. Kissling, W. D., C. Rahbek, and K. Böhning-Gaese. 2007. Food plant diversity as broad-scale determinant of avian frugivore richness. Proceedings of the Royal Society B: Biological Sciences 274:799-808. Kitamura, K. and R. A. Clapp. 2013. Common property protected areas: Community control in forest conservation. Land Use Policy 34:204-212. Klooster, D. and O. Masera. 2000. Community forest management in Mexico: carbon mitigation and biodiversity conservation through rural development. Global Environmental Change 10:259-272. Knopf, F. L. and F. B. Samson. 1994. Scale perspectives on avian diversity in western riparian ecosystems. Conservation Biology 8:669-676. Koleff, P., K. J. Gaston, and J. J. Lennon. 2003. Measuring beta diversity for presence– absence data. Journal of Animal Ecology 72:367-382. Kumar, R. and G. Shahabuddin. 2005. Effects of biomass extraction on vegetation structure, diversity and composition of forests in Sariska Tiger Reserve, India. Environmental Conservation 32:248-259. Kumar, R., G. Shahabuddin, and A. Kumar. 2011. How good are managed forests at conserving native woodpecker communities? A study in sub-Himalayan dipterocarp forests of northwest India. Biological Conservation 144:1876-1884.

121

Kumar, R., G. Shahabuddin, and A. Kumar. 2014. Habitat determinants of woodpecker abundance and species richness in sub-Himalayan dipterocarp forests of north-west India. Acta Ornithologica 49:243-256. Laiolo, P., E. Caprio, and A. Rolando. 2004. Can forest management have season-dependent effects on bird diversity? Biodiversity & Conservation 13:1925-1941. Lambrick, F. H., N. D. Brown, A. Lawrence, and D. P. Bebber. 2014. Effectiveness of Community Forestry in Prey Long Forest, . Conservation Biology 28:372- 381. Laurance, W. F. 2010. Habitat destruction: death by a thousand cuts. Conservation Biology for All 1:73-88. Laurance, W.F. et al., 2012. Averting biodiversity collapse in tropical forest protected areas. Nature, 489(7415): 290-4. Law, B. and C. Dickman. 1998. The use of habitat mosaics by terrestrial vertebrate fauna: implications for conservation and management. Biodiversity & Conservation 7:323- 333. Lawrence, A., K. Paudel, R. Barnes, and Y. Malla. 2006. Adaptive value of participatory biodiversity monitoring in community forestry. Environmental Conservation 33:325- 334. Leal, A. I., R. A. Correia, J. M. Palmeirim, and J. P. Granadeiro. 2013. Does canopy pruning affect foliage-gleaning birds in managed cork oak woodlands? Agroforestry systems 87:355-363. Lee, M.-B. and J. P. Carroll. 2014. Relative importance of local and landscape variables on site occupancy by avian species in a pine forest, urban, and agriculture matrix. Forest Ecology and Management 320:161-170. Lee, T. M., N. S. Sodhi, and D. M. Prawiradilaga. 2007. The importance of protected areas for the forest and endemic avifauna of Sulawesi (). Ecological Applications 17:1727-1741. Lindenmayer, D. and G. Luck. 2005. Synthesis: thresholds in conservation and management. Biological Conservation 124:351-354. Lindenmayer, D. and M. Burgman. 2005. Practical conservation biology. Csiro Publishing. Lindenmayer, D. B. 2009. Forest wildlife management and conservation. Annals of the New York Academy of Sciences 1162:284-310. Lindenmayer, D. B. and J. Fischer. 2006. Habitat fragmentation and landscape change: an ecological and conservation synthesis. Island Press.

122

Lindenmayer, D. B. and S. A. Cunningham. 2013. Six principles for managing forests as ecologically sustainable ecosystems. Landscape Ecology 28:1099-1110. Lindenmayer, D. B., W. F. Laurance, and J. F. Franklin. 2012. Global decline in large old trees. Science 338:1305-1306. Ludwig, J. A., R. W. Eager, A. C. Liedloff, J. C. McCosker, D. Hannah, N. Y. Thurgate, J. C. Woinarski, and C. P. Catterall. 2000. Clearing and grazing impacts on vegetation patch structures and fauna counts in eucalypt woodland, Central Queensland. Pacific Conservation Biology 6:254. Magurran, A. E. 2004. Measuring biological diversity. Manuwal, D. A. and A. B. Carey. 1991. Methods for measuring populations of small, diurnal forest birds. Maron, M. and A. Lill. 2005. The influence of livestock grazing and weed invasion on habitat use by birds in grassy woodland remnants. Biological Conservation 124:439-450. Maron, M. and S. Kennedy. 2007. Roads, fire and aggressive competitors: determinants of bird distribution in subtropical production forests. Forest Ecology and Management 240:24-31. Maron, M., M. Bowen, R. A. Fuller, G. C. Smith, T. J. Eyre, M. Mathieson, J. E. Watson, and C. A. McAlpine. 2012. Spurious thresholds in the relationship between species richness and vegetation cover. Global Ecology and Biogeography 21:682-692. Martensen, A. C., M. C. Ribeiro, C. Banks‐Leite, P. I. Prado, and J. P. Metzger. 2012. Associations of forest cover, fragment area, and connectivity with Neotropical understory bird species richness and abundance. Conservation Biology 26:1100-1111. Martin, T. G. and S. McIntyre. 2007. Impacts of livestock grazing and tree clearing on birds of woodland and riparian habitats. Conservation Biology 21:504-514. Marzluff, J. M., M. G. Raphael, and R. Sallabanks. 2000. Understanding the effects of forest management on avian species. Wildlife Society Bulletin:1132-1143. Mathur, P. and P. Sinha. 2008. Looking beyond protected area networks: a paradigm shift in approach for biodiversity conservation. International Forestry Review 10:305-314. Mayer, A., V. Stöckli, W. Konold, and M. Kreuzer. 2006. Influence of cattle stocking rate on browsing of Norway spruce in subalpine wood pastures. Agroforestry systems 66:143-149. McAlpine, C. A., J. R. Rhodes, J. G. Callaghan, M. E. Bowen, D. Lunney, D. L. Mitchell, D. V. Pullar, and H. P. Possingham. 2006. The importance of forest area and

123

configuration relative to local habitat factors for conserving forest mammals: a case study of koalas in Queensland, Australia. Biological Conservation 132:153-165. McAlpine, C., M. Bowen, G. Smith, G. Gramotnev, A. Smith, A. L. Cascio, W. Goulding, and M. Maron. 2015. Reptile abundance, but not species richness, increases with regrowth age and spatial extent in fragmented agricultural landscapes of eastern Australia. Biological Conservation 184:174-181. McGarigal, K. and S. A. Cushman. 2002. Comparative evaluation of experimental approaches to the study of habitat fragmentation effects. Ecological Applications 12:335-345. McGarigal, K. and W. C. McComb. 1995. Relationships between landscape structure and breeding birds in the Oregon Coast Range. Ecological Monographs 65:235-260. Millennium Ecosystem Assessment, 2005. Ecosystems and human well-being. Island Press Washington, DC. Miller, J. N., R. P. Brooks, and M. J. Croonquist. 1997. Effects of landscape patterns on biotic communities. Landscape Ecology 12:137-153. Mishra, B., O. Tripathi, R. Tripathi, and H. Pandey. 2004. Effects of anthropogenic disturbance on plant diversity and community structure of a sacred grove in Meghalaya, northeast India. Biodiversity & Conservation 13:421-436. MoAC. 2004. Statistical Information on Nepalese Agriculture. His Majesty's Government, Ministry of Agricultuure and Cooperative. Agribusiness Promotion and Statistics Division, Singh Durbar, Kathmandu, Nepal. Moktan, M. R., G. Gratzer, W. H. Richards, T. B. Rai, and D. Dukpa. 2009. Regeneration and structure of mixed conifer forests under single-tree harvest management in the western Bhutan Himalayas. Forest Ecology and Management 258:243-255. Molnar, A., M. France, L. Purdy, and J. Karver. 2011. Community-Based Forest Management. The Extent and Potential Scope of Community and Smallholder Forest Management and Enterprise. Morrison, M. L., B. Marcot, and W. Mannan. 2006. Wildlife-habitat relationships: concepts and applications. Island Press. Mortelliti, A., S. Fagiani, C. Battisti, D. Capizzi, and L. Boitani. 2010. Independent effects of habitat loss, habitat fragmentation and structural connectivity on forest‐dependent birds. Diversity and Distributions 16:941-951.

124

Moura, N. G., A. C. Lees, C. B. Andretti, B. J. Davis, R. R. Solar, A. Aleixo, J. Barlow, J. Ferreira, and T. A. Gardner. 2013. Avian biodiversity in multiple-use landscapes of the Brazilian Amazon. Biological Conservation 167:339-348. Muggeo, V. 2012. segmented: segmented relationships in regression models with breakpoints/changepoints estimation: R package. See: http://cran. r-project. org/web/packages/segmented/index. html. Muggeo, V. M. 2008. Segmented: an R package to fit regression models with broken-line relationships. R news 8:20-25. Nagendra, H. 2007. Drivers of reforestation in human-dominated forests. Proceedings of the National Academy of Sciences 104:15218-15223. Nagendra, H. 2008. Do parks work? Impact of protected areas on land cover clearing. AMBIO: A Journal of the Human Environment 37:330-337. Nagendra, H. and Y. Gokhale. 2008. Management regimes, property rights, and forest biodiversity in Nepal and India. Environ Manage 41:719-733. Nagendra, H., 2002. Tenure and forest conditions: community forestry in the Nepal Terai. Environmental Conservation, 29(04). Nagendra, H., M. Karmacharya, and B. Karna. 2005. Evaluating forest management in Nepal: views across space and time. Ecology and Society 10:24. Nagendra, H., S. Pareeth, B. Sharma, C. M. Schweik, and K. R. Adhikari. 2008. Forest fragmentation and regrowth in an institutional mosaic of community, government and private ownership in Nepal. Landscape Ecology 23:41-54. Nakagawa, S. and H. Schielzeth. 2013. A general and simple method for obtaining R2 from generalized linear mixed‐effects models. Methods in Ecology and Evolution 4:133- 142. Ochoa‐Quintero, J. M., T. A. Gardner, I. Rosa, S. F. Barros Ferraz, and W. J. Sutherland. 2015. Thresholds of species loss in Amazonian deforestation frontier landscapes. Conservation Biology. Ojha, H., Persha, L., Chhatre, A., 2009. Community Forestry in Nepal: A Policy Innovation for Local Livelihoods and Food Security. Working Paper No. W09I-02. International Food Policy Research Institute, Ann Arbor, Michigan, 34pp. Oksanen, J. et al., 2011. vegan: Community Ecology Package. R package version 1. 17–6. Palmer, G. C. and A. F. Bennett. 2006. Riparian zones provide for distinct bird assemblages in forest mosaics of south-east Australia. Biological Conservation 130:447-457.

125

Pardini, R., Bueno, A. d. A., Gardner, T. A., Prado, P. I. and Metzger, J. P. 2010. Beyond the fragmentation threshold hypothesis: regime shifts in biodiversity across fragmented landscapes. PLoS ONE, 5(10), e13666. Pattison, R. R., C. M. D'Antonio, T. L. Dudley, K. K. Allander, and B. Rice. 2011. Early impacts of biological control on canopy cover and water use of the invasive saltcedar tree (Tamarix spp.) in western Nevada, USA. Oecologia 165:605-616. Paudel, N. S., P. Budhathoki, and U. R. Sharma. 2007. Buffer zones: New frontiers for participatory conservation. Journal of Forest and Livelihood 6:44-53. Peh, K. S. H., J. d. Jong, N. S. Sodhi, S. L. H. Lim, and C. A. M. Yap. 2005. Lowland rainforest avifauna and human disturbance: persistence of primary forest birds in selectively logged forests and mixed-rural habitats of southern Peninsular . Biological Conservation 123:489-505. Persha, L., H. Fischer, A. Chhatre, A. Agrawal, and C. Benson. 2010. Biodiversity conservation and livelihoods in human-dominated landscapes: forest commons in South Asia. Biological Conservation 143:2918-2925. Piana, R. P. and S. J. Marsden. 2014. Impacts of cattle grazing on forest structure and raptor distribution within a neotropical protected area. Biodiversity and Conservation:1-14. Pinheiro, J. C. and D. M. Bates. 2000. Mixed effects models in S and S-PLUS. Springer. Pizo, M.A., Simao, I., Galetti, M., 1997. Daily variation in activity and flock size of two parakeet species from southeastern Brazil. The Wilson Bulletin, 109(2): 343-348. Pokharel, B. K., P. Branney, M. Nurse, and Y. B. Malla. 2007. Community forestry: Conserving forests, sustaining livelihoods and strengthening democracy. Journal of Forest and Livelihood 6:8-19. Politi, N., M. Hunter Jr, and L. Rivera. 2012. Assessing the effects of selective logging on birds in Neotropical piedmont and cloud montane forests. Biodiversity and Conservation 21:3131-3155. Porter-Bolland, L., E. A. Ellis, M. R. Guariguata, I. Ruiz-Mallén, S. Negrete-Yankelevich, and V. Reyes-García. 2012. Community managed forests and forest protected areas: An assessment of their conservation effectiveness across the tropics. Forest Ecology and Management 268:6-17. Poudyal, L. P., P. B. Singh, and S. Maharajan. 2008. Bengal Florican Houbaropsis bengalensis in Nepal: an update. Birding Asia 10: 43 47. Powell, L. A., J. D. Lang, D. G. Krementz, and M. J. Conroy. 2005. Use of radio-telemetry to reduce bias in nest searching. Journal of Field Ornithology 76:274-278.

126

Powlesland, R., J. Knegtmans, and A. Styche. 2000. Mortality of North Island tomtits (Petroica macrocephala toitoi) caused by aerial 1080 possum control operations, 1997-98, Pureora Forest Park. New Zealand Journal of Ecology:161-168. Pulliam, H. R. 1988. Sources, sinks, and population regulation. American naturalist:652-661. R CoreTeam, R. 2014. A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. . ISBN 3- 900051-07-0. Radford, J. Q. and A. F. Bennett. 2007. The relative importance of landscape properties for woodland birds in agricultural environments. Journal of Applied Ecology 44:737-747. Radford, J. Q., A. F. Bennett, and G. J. Cheers. 2005. Landscape-level thresholds of habitat cover for woodland-dependent birds. Biological Conservation 124:317-337. Raman, T. 2001. Effect of Slash‐and‐Burn Shifting Cultivation on Rainforest Birds in Mizoram, Northeast India. Conservation Biology 15:685-698. Rodewald, A. D., L. J. Kearns, and D. P. Shustack. 2013. Consequences of urbanizing landscapes to reproductive performance of birds in remnant forests. Biological Conservation 160:32-39. Rodrigues, A. S., S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, R. M. Cowling, L. D. Fishpool, G. A. da Fonseca, K. J. Gaston, and M. Hoffmann. 2004. Effectiveness of the global protected area network in representing species diversity. Nature 428:640-643. Sagar, R. and J. S. Singh. 2004. Local plant species depletion in a tropical dry deciduous forest of northern India. Environmental Conservation 31:55-62. Sah, S.P., Sharma, C.K., Sehested, F., 2002. Possible role of the soil in the Sissoo forest (Dalbergia sissoo, Roxb.) decline in the Nepal terai. In: W.J. Horst et al. (Editors), Plant Nutrition. Developments in Plant and Soil Sciences. Springer , pp. 930-931. Sapkota, I. P., M. Tigabu, and P. C. Odén. 2009. Spatial distribution, advanced regeneration and stand structure of Nepalese Sal (Shorea robusta) forests subject to disturbances of different intensities. Forest Ecology and Management 257:1966-1975. Sapkota, I. P., M. Tigabu, and P. C. Odén. 2010. Changes in tree species diversity and dominance across a disturbance gradient in Nepalese Sal (Shorea robusta Gaertn. f.) forests. Journal of Forestry Research 21:25-32.

127

Saunders, D. 1990. Problems of survival in an extensively cultivated landscape: the case of Carnaby's cockatoo< i> Calyptorhynchus funereus latirostris. Biological Conservation 54:277-290. Saunders, D. A., R. J. Hobbs, and C. R. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18-32. Schmiegelow, F. K. and M. Mönkkönen. 2002. Habitat loss and fragmentation in dynamic landscapes: avian perspectives from the Boreal forest. Ecological Applications 12:375-389. Schneider, N. A. and M. Griesser. 2009. Influence and value of different water regimes on avian species richness in arid inland Australia. Biodiversity and Conservation 18:457- 471. Sekercioglu, C. H. 2002. Effects of forestry practices on vegetation structure and bird community of Kibale National Park, Uganda. Biological Conservation 107:229-240. Sekercioglu, C. H., P. R. Ehrlich, G. C. Daily, D. Aygen, D. Goehring, and R. F. Sandi. 2002. Disappearance of insectivorous birds from tropical forest fragments. Proc Natl Acad Sci U S A 99:263-267. Shahabuddin, G. and R. Kumar. 2006. Influence of anthropogenic disturbance on bird communities in a tropical dry forest: role of vegetation structure. Animal Conservation 9:404-413. Shahabuddin, G. and R. Kumar. 2007. Effects of extractive disturbance on bird assemblages, vegetation structure and floristics in tropical scrub forest, Sariska Tiger Reserve, India. Forest Ecology and Management 246:175-185. Sharma, U. R. 1990. An overview of park-people interactions in Royal Chitwan National Park, Nepal. Landscape and Urban Planning 19:133-144. Shrestha, U. B., S. Shrestha, P. Chaudhary, and R. P. Chaudhary. 2010. How representative is the protected areas system of Nepal? A gap analysis based on geophysical and biological features. Mountain Research and Development 30:282-294. Shyamsundar, P., Ghate, R., 2011. Rights, Responsibilities and Resources: Examining Community Forestry in South Asia. Working Paper, SANDEE, 59-11: 1-24. Singh, B. and D. Chapagain. 2006. Trends in forest ownership, forest resources tenure and institutional arrangements: are they contributing to better forest management and poverty reduction Sinha, D.R., 2011. Betrayal or'Business as Usual'? Access to Forest Resourecs in the Nepal Terai. Environment and History, 17: 433-460.

128

Smith, A. C., L. Fahrig, and C. M. Francis. 2011. Landscape size affects the relative importance of habitat amount, habitat fragmentation, and matrix quality on forest birds. Ecography 34:103-113. Sodhi, N. S. and P. R. Ehrlich. 2010. Conservation biology for all. Oxford University Press Oxford, United Kingdom. Sofaer, H. R., T. S. Sillett, S. I. Peluc, S. A. Morrison, and C. K. Ghalambor. 2012. Differential effects of food availability and nest predation risk on avian reproductive strategies. Behavioral Ecology:ars212. Springate-Baginski, O., Dev, O.P., Yadav, N.P., Soussan, J., 2003. Community Forest Management in the Middle Hills of Nepal: The Changing Context. Journal of Forest and Livelihood, 3(1): 5-20. Straede, S., G. Nebel, and A. Rijal. 2002. Structure and floristic composition of community forests and their compatibility with villagers' traditional needs for forest products. Biodiversity & Conservation 11:487-508. Suarez-Rubio, M., S. Wilson, P. Leimgruber, and T. Lookingbill. 2013. Threshold responses of forest birds to landscape changes around exurban development. PLoS One 8:e67593. Sundriyal, R. and E. Sharma. 1996. Anthropogenic pressure on tree structure and biomass in the temperate forest of Mamlay watershed in Sikkim. Forest Ecology and Management 81:113-134. Symonds, M. R. and A. Moussalli. 2011. A brief guide to model selection, multimodel inference and model averaging in behavioural ecology using Akaike’s information criterion. Behavioral Ecology and Sociobiology 65:13-21. Tantipisanuh, N., Gale, G.A., 2013. Representaion of thretened vertebrates by a protected area system in : The importance of non-forest habitats. The raffles bulletin of zoology: National University of Singapore, 1: 359–395. Tasker, E. M. and R. A. Bradstock. 2006. Influence of cattle grazing practices on forest understorey structure in north‐eastern New South Wales. Austral ecology 31:490-502. Taylor, R. S., S. J. Watson, D. G. Nimmo, L. T. Kelly, A. F. Bennett, and M. F. Clarke. 2012. Landscape‐scale effects of fire on bird assemblages: does pyrodiversity beget biodiversity? Diversity and Distributions 18:519-529. Tewksbury, J. J., D. J. Levey, N. M. Haddad, S. Sargent, J. L. Orrock, A. Weldon, B. J. Danielson, J. Brinkerhoff, E. I. Damschen, and P. Townsend. 2002. Corridors affect

129

plants, animals, and their interactions in fragmented landscapes. Proceedings of the National Academy of Sciences 99:12923-12926. Tews, J., U. Brose, V. Grimm, K. Tielbörger, M. Wichmann, M. Schwager, and F. Jeltsch. 2004. Animal species diversity driven by habitat heterogeneity/diversity: the importance of keystone structures. Journal of Biogeography 31:79-92. Thapa, S. and D. S. Chapman. 2010. Impacts of resource extraction on forest structure and diversity in Bardia National Park, Nepal. Forest Ecology and Management 259:641- 649. Thinh, V. T., P. F. Doherty, and K. P. Huyvaert. 2012. Effects of different logging schemes on bird communities in tropical forests: a simulation study. Ecological Modelling 243:95-100. Timilsina, N., Ross, M.S., Heinen, J.T., 2007. A community analysis of sal (Shorea robusta) forests in the western Terai of Nepal. Forest Ecology and Management, 241(1-3): 223-234. Touihri, M., M.-A. Villard, and F. Charfi. 2014. Cavity-nesting birds show threshold responses to stand structure in native oak forests of northwestern Tunisia. Forest Ecology and Management 325:1-7. Trzcinski, M. K., L. Fahrig, and G. Merriam. 1999. Independent effects of forest cover and fragmentation on the distribution of forest breeding birds. Ecological Applications 9:586-593. Tscharntke, T., A. M. Klein, A. Kruess, I. Steffan‐Dewenter, and C. Thies. 2005. Landscape perspectives on agricultural intensification and biodiversity–ecosystem service management. Ecol Lett 8:857-874. van den Berg, L. J., J. M. Bullock, R. T. Clarke, R. H. Langston, and R. J. Rose. 2001. Territory selection by the Dartford warbler (< i> Sylvia undata) in Dorset, England: the role of vegetation type, habitat fragmentation and population size. Biological Conservation 101:217-228. Vergara, P. M. and J. J. Armesto. 2009. Responses of Chilean forest birds to anthropogenic habitat fragmentation across spatial scales. Landscape Ecology 24:25-38. Vergara, P. M. and P. A. Marquet. 2007. On the seasonal effect of landscape structure on a bird species: the thorn-tailed rayadito in a relict forest in northern Chile. Landscape Ecology 22:1059-1071.

130

Villard, M. A., M. K. Trzcinski, and G. Merriam. 1999. Fragmentation effects on forest birds: relative influence of woodland cover and configuration on landscape occupancy. Conservation Biology 13:774-783. Villela, D. M., M. T. Nascimento, L. E. O. Aragão, and D. M. Da Gama. 2006. Effect of selective logging on forest structure and nutrient cycling in a seasonally dry Brazilian Atlantic forest. Journal of Biogeography 33:506-516. Webb, E. L. and R. N. Sah. 2003. Structure and diversity of natural and managed sal (< i> Shorea robusta Gaertn. f.) forest in the Terai of Nepal. Forest Ecology and Management 176:337-353. White, A., Martin, A., 2002. Who Owns the World's Forests? Forest Tenure and Public Forests in Transition, Forest Trend, Washington, D.C.: 2-29. Whitehorne, I., M. Harrison, N. A. Mahony, P. Robinson, A. Newbury, and D. J. Green. 2011. Effects of Cattle Grazing on Birds in Interior Douglas-Fir (Pseudotsuga Whittaker, R. H. 1972. Evolution and measurement of species diversity. Taxon:213-251. Whittaker, R. H. 1977. Evolution of species diversity in land communities. In Evolutionary Biology. Vol. 10 pp. 1–67. Edited by M.K. Hecht and B.W.N.C. Steere. Plenum Press, New York. Whittaker, R. J. and J. M. Fernández-Palacios. 2007. Island biogeography: ecology, evolution, and conservation. Oxford University Press. Wiens, J. A. 1992. The ecology of bird communities. Cambridge University Press. Wiersma, Y. F. and D. L. Urban. 2005. Beta diversity and nature reserve system design in the Yukon, Canada. Conservation Biology 19:1262-1272. Wikramanayake, E. et al., 2004. Designing a Conservation Landscape for Tigers in Human- Dominated Environments. Conservation Biology 18:839-844. Williams, S. E. and J. Middleton. 2008. Climatic seasonality, resource bottlenecks, and abundance of rainforest birds: implications for global climate change. Diversity and Distributions 14:69-77. Williams, S. E., H. Marsh, and J. Winter. 2002. Spatial scale, species diversity, and habitat structure: small mammals in Australian tropical rain forest. Ecology 83:1317-1329. World Bank. 2012. World Development Indicators 2012. World Bank Publications. World Commission on Forests and Sustainable Development.1999. Our forests our future. Report of the World Commission on Forests and Sustainable Development. Cambridge University Press, Cambridge, England.

131

World Resource Institute. 1997.The Last Frontier Forests: Ecosystems and Economies on the Edge. Washington, DC.pp54. World Resource Institute. 2005. World resources 2005: the wealth of the poor – managing ecosystems to fight poverty. World Resources Institute, Washington, DC, USA. WWF. 2014. Living planet report. Species,spaces,people and places. WWF International Gland,Switzerland. Zuckerberg, B. and W. F. Porter. 2010. Thresholds in the long-term responses of breeding birds to forest cover and fragmentation. Biological Conservation 143:952-962. Zurita, G. A. and M. I. Bellocq. 2012. Bird assemblages in anthropogenic habitats: identifying a suitability gradient for native species in the Atlantic forest. Biotropica 44:412-419. Zuur, A. F., E. N. Ieno, and C. S. Elphick. 2010. A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1:3-14.

132

APPENDICES

Appendix A Table A.1 Total number of individuals and their respective food guilds and habitat groups, and relative percentage of species observation across sites in lowland tropical landscapes: CF = Community forest, SF = State forest, PA = Protected area.

Habita Sightin Species CF SF PA Food guild t group g%

Abbot's Babbler Malacocincla abbotti 5 Ins Fs 2 Alexandrine Parakeet Psittacula eupatria 26 34 29 fru-gra Fg 25 Dicrurus leucophaeus 1 Ai Fg 1 Ashy Minivet Pericrocotus divaricatus 2 Fgi Fg 1 Ashy Artamus fuscus 16 Ai Fg 2 Asian Koel Eudynamys scolopacea 1 1 Ins Fg 1 Asian Palm Swift Cypsiurus balasiensis 9 Ai Fg 4 Asian Pied Sturnus contra 2 ins-fru-gra Fg 1 Bar-winged Flycatcher-shrike Hemipus picatus 2 8 Fgi Fes 4 Black Bulbul Hypsipetes leucocephalus 1 fru-ins-nec Fs 1 Black Drongo Dicrurus macrocercus 56 61 11 Si Fg 50 Black Eagle Ictinaetus malayensis 1 Ca Oc 1 Black-crested Bulbul Pycnonotus melanicterus 21 14 32 fru-ins Fs 24 Black-hooded Oriole Oriolus xanthornus 142 132 153 fru-nec-ins Fg 98 Black-naped Monarch Hypothymis azurea 6 2 4 Si Fes 4 Black-rumped Flameback benghalense 33 47 12 Bgi Fg 46 Black-winged Coracina melaschistos 1 6 2 Fgi Fes 4 Blue-bearded Bee-eater Nyctyornis athertoni 3 2 Si Fes 4 Blue-eared Barbet Megalaima australis 2 Fru Fs 1 Blue-throated Barbet Megalaima asiatica 9 1 Fru Fg 6 Blyth's Leaf Warbler Phylloscopus reguloides 2 Fgi Fes 1 Bronzed Drongo Dicrurus aeneus 28 9 12 Si Fes 20 Brown Shrike Lanius cristatus 1 ins-ca Fes 1 Chestnut-bellied Nuthatch Sitta castanea 176 126 171 Bgi Fg 67 Chestnut-headed Bee-eater Merops leschenaulti 2 Si Fg 1 Chestnut-tailed Starling Sturnus malabaricus 100 44 5 fru-ins-nec Fg 13 Collared Falconet Microhierax caerulescens 3 2 Si Oc 4 Common Hawk Hierococcyx varius 4 1 6 Fgi Fg 9 Common Hoopoe Upupa epops 1 Ins Fg 1

133

Common Aegithina tiphia 2 1 6 Fgi Fes 4 Common Myna Acridotheres tristis 12 30 Omn Fg 8 Common Tailorbird Orthotomus sutorius 30 43 31 Fgi Fg 52 Tephrodornis pondicerianus 9 5 29 Fgi Fg 9 Crested Serpent Eagle Spilornis cheela 6 3 6 Ca Fg 13 Crow-billed Drongo Dicrurus annectans 1 1 Si Fs 2 Drongo Cuckoo Surniculus lugubris 3 1 2 Fgi Fs 4 Dusky Warbler Phylloscopus fuscatus 8 Ins Fg 5 Emerald Dove Chalcophaps indica 4 gra-fru Fs 4 Eurasian Collared Dove Streptopelia decaocto 19 6 8 Gra Fg 14 Eurasian Wryneck Jynx torquilla 1 Ins Fg 1 Fulvous-breasted Woodpecker Dendrocopos macei 21 25 18 Bgi Fes 33 Golden-fronted Leafbird Chloropsis aurifrons 110 47 23 nec-ins-fru Fg 37 Golden-spectacled Warbler Seicercus burkii 1 Fgi Fg 1 Great Tit Parus major 62 72 134 Fgi Fg 57 Greater Coucal Centropus sinensis 3 2 Omn Fg 4 Greater Flameback Chrysocolaptes lucidus 2 Bgi Fg 1 Greater Racket-tailed Drongo Dicrurus paradiseus 92 57 81 Si Fes 78 Greater Yellownape Picus flavinucha 2 2 Bgi Fes 4 Green Bee-eater Merops orientalis 12 3 7 Si Fg 4 Green-billed Malkoha Phaenicophaeus tristis 1 2 5 Fgi Fs 4 Greenish Warbler Phylloscopus trochiloides 26 28 16 Fgi Fg 36 Grey-backed Shrike Lanius tephronotus 1 ins-ca Fg 1 Grey-bellied Tesia Tesia cyaniventer 1 Ins Fg 1 Grey-capped Pygmy Woodpecker Dendrocopos canicapillus 9 2 17 Bgi Fg 1 Grey-headed Canary Flycatcher Culicicapa ceylonensis 104 64 53 Si Fg 80 Grey-headed Woodpecker Picus canus 7 2 40 Bgi Fs 21 Grey-sided Bush Warbler Cettia brunnifrons 1 Fgi Fg 1 Himalayan Bulbul Pycnonotus leucogenys 14 7 1 fru-ins-nec Fg 6 Himalayan Flameback Dinopium shorii 27 14 48 Bgi Fs 38 Indian Roller Coracias benghalensis 1 Ins-ca Fg 1 Hume's Warbler Phylloscopus humei 3 Fgi Fes 1 Cuculus micropterus 3 1 Fgi Fg 4 Indian Grey Hornbill Ocyceros birostris 1 Fru Fg 1 Indian Peafowl Pavo cristatus 3 15 Omn Fg 6 Jungle Babbler Turdoides striatus 264 215 236 Ins Fg 79 Jungle Myna Acridotheres fuscus 2 fru-gra-nec Fg 1 Jungle Owlet Glaucidium radiatum 6 3 13 ins-ca Fg 15

134

Large Cuckooshrike Coracina macei 61 40 30 ins-fru Fes 48 Tephrododornis gularis 35 18 18 Fgi Fes 8 Large-billed Crow Corvus macrorhynchos 8 3 20 Omn Fg 17 Lesser Racket-tailed Drongo Dicrurus remifer 2 1 1 Si Fs 4 Lesser Yellownape Picus chlorolophus 49 44 66 Bgi Fes 54 Lineated Barbet Megalaima lineata 20 8 3 Fru Fes 21 Little Pied Flycatcher Ficedula westermanni 1 Fgi Fs 1 Long-tailed Shrike Lanius schach 4 5 1 ins-ca Fg 7 Olive-back Pipit Anthus hodgsoni 3 1 Ins Fg 2 Orange-breasted Green pigeon Treron bicincta 6 9 Fru Fs 3 Orange-headed Zoothera citrina 5 2 22 ins-fru Fes 11 Oriental Magpie Robin Copsychus saularis 1 1 ins-nec Fg 2 Oriental Pied Hornbill Anthracoceros albirostris 23 5 Fru Fs 7 Oriental Turtle Dove Streptopelia orientalis 2 Gra Fg 1 Oriental White-eye Zosterops palpebrosus 22 5 nec-ins Fg 4 Osprey Pandion haliaetus 2 Pis Oc 1 Pale-chinned Flycatcher Cyornis poliogenys 14 10 1 Ai Fg 18 Pallas's Fish Eagle Haliaeetus leucoryphus 1 Pis Oc 1 Plain Prinia Prinia inornata 30 Fgi Fg 6 Plum-headed Parakeet Psittacula cyanocephala 150 84 52 fru-gra Fes 45 Puff-throated Babbler Pellorneumru ficeps 1 10 Ins Fg 4 Purple Sunbird Nectarinia asiatica 2 nec-ins Fg 1 Red Collared Dove Streptopelia tranquebarica 1 Gra Fg 1 Red Junglefowl Gallus gallus 32 11 24 Gra Fes 36 Red-billed Blue Magpie Urocissa erythrorhyncha 4 12 20 Omn Fes 10 Red-throated Flycatcher Ficedula parva 104 74 73 Si Fg 78 Red-vented Bulbul Pycnonotus cafer 68 41 33 fru-ins-nec Fg 22 Red-whiskered Bulbul Pycnonotus jocosus 4 8 fru-ins-nec Fg 4 Rose-ringed Parakeet Psittacula krameri 307 350 206 fru-gra Fg 17 Rosi Minivet Pericrocotus roseus 6 Fgi Fg 3 Rufous Treepie Dendrocitta vagabunda 113 83 93 Omn Fg 17 Scaly Thrush Zoothera dauma 1 1 3 ins-fru Fs 4 Scarlet Minivet Pericrocotus flammeus 184 175 105 Fgi Fes 59 Short-toed Snake Eagle Ciraetus gallicus 1 Ca Oc 1 Sirkeer Malkoha Phaenicophaeus leschenaultii 2 Ins Fg 1 Slaty-blue Flycatcher Ficedula tricolor 1 1 Ins Fg 2 Slaty-headed Parakeet Psittacula himalayana 2 fru-gra Fg 1 Slender-billed Oriole Oriolus tenuirostris 1 1 1 fru-nec-ins Fg 3 Small Minivet Pericrocotus cinnamomeus 26 65 48 Fgi Fg 17 Spangled Drongo Dicrurus hottentottus 178 126 135 nec-si Fes 79

135

Spotted Dove Streptopelia chinensis 95 42 13 Gra Fg 52 Spotted Owlet Athene brama 1 Si Fg 1 Streak-throated Woodpecker Picus xanthopygaeus 9 Ins Fg 1 Striped Tit Babbler Macronous gularis 2 1 Fgi Fes 3 Thick-billed Warbler Acrocephaluss aedon 4 1 1 Fgi Fes 4 Tickell's Leaf Warbler Phylloscopus affinis 2 Fgi Fes 1 Tree Pipit Anthus trivialis 4 ins-gra Fes 1 Velvet-fronted Nuthatch Sitta frontalis 54 36 20 Bgi Fes 24 Verditer Flycatcher Eumyias thalassina 3 3 Si Fg 3 White-bellied Drongo Dicrurus caerulescens 48 23 27 Si Fes 46 White-rumped Shama Copsychus malabaricus 25 25 13 Ins Fes 29 White-throated Fantail Rhipidura albicollis 4 Ins Fg 1 White-throated Kingfisher Halcyon smyrnensis 2 ins-ca Fg 2 Yellow-bellied Warbler Abroscopus superciliaris 1 Ins Fg 1 Yellow-crowned Woodpecker Dendrocopos mahrattensis 3 Bgi Fg 2 Yellow-footed Green Pigeon Treron phoenicopterus 1 6 Fru Fg 2 Zitting Cisticola Cisticola juncidis 1 Ins Fg 1 Food guilds: ins = insectivore, si = sallying insectivore, ai = aerial insectivore, fgi = foliage gleaning insectivore, bgi = bark gleaning insectivore, fru = frugivore, gra = granivore, nec = nectarivore, , ca = carnivore, omn = , pis = piscivore. Habitat groups: fg = forest specialist, fes = forest edge specialist, fs = forest specialist, oc = open country species

136

Table A.2 Top ten contributing species to dissimilarities between management tenures using SIMPER analysis based on Bray-Curtis dissimilarity.

Mean abundance Contribution to Species Community Forest Protected Area dissimilarity %

Rose-ringed Parakeet Psittacula krameri 7.21 5.58 4.54 Jungle Babbler Turdoides striatus 6.20 6.69 3.97 Plum-headed Parakeet Psittacula cyanocephala 3.08 2.88 3.74 Chestnut-bellied Nuthatch Sitta castanea 4.07 5.04 3.73 Great Tit Parus major 2.25 5.22 3.50 Scarlet Minivet Pericrocotus flammeus 4.27 3.24 3.44 RufousTreepie Dendrocitta vagabunda 4.53 4.27 2.75 Spotted Dove Streptopelia chinensis 3.59 1.01 2.67 Spangled Drongo Dicrurus hottentottus 5.64 5.05 2.56 Golden-fronted Leafbird Chloropsisaurifrons 2.79 1.00 2.47 Species Community Forest State Forest Contribution to dissimilarity % Jungle Babbler Turdoides striatus 6.20 5.70 5.45 Rose-ringed Parakeet Psittacula krameri 7.21 8.68 5.06 Chestnut-bellied Nuthatch Sitta castanea 4.07 3.17 3.50 Plum-headed Parakeet Psittacula cyanocephala 3.08 3.42 3.50 Scarlet Minivet Pericrocotus flammeus 4.27 4.21 3.41 Black Drongo Dicrurus macrocercus 2.36 4.52 3.36 Spangled Drongo Dicrurus hottentottus 5.64 4.12 3.36 RufousTreepie Dendrocitta vagabunda 4.53 4.81 3.16 Spotted Dove Streptopelia chinensis 3.59 2.36 2.98 Black-hooded Oriole Oriolus xanthornus 5.20 6.99 2.85 Species Protected Area State Forest Contribution to dissimilarity % Jungle Babbler Turdoides striatus 6.69 5.70 5.16 Rose-ringed Parakeet Psittacula krameri 5.58 8.68 4.97 Plum-headed Parakeet Psittacula cyanocephala 2.88 3.42 3.69 Chestnut-bellied Nuthatch Sitta castanea 5.04 3.17 3.68 Black Drongo Dicrurus macrocercus 0.71 4.52 3.45 Great Tit Parus major 5.22 2.95 3.39 Scarlet Minivet Pericrocotus flammeus 3.24 4.21 3.31 RufousTreepie Dendrocitta vagabunda 4.27 4.81 3.13 Spangled Drongo Dicrurus hottentottus 5.05 4.12 2.98 Black-hooded Oriole Oriolus xanthornus 6.09 6.99 2.48

137

Appendix B

Table B.1 Summary statistics of habitat characteristics (± SE) across sites in grazing, logging and lopping disturbances in lowland tropical landscapes

Variables Grazed Light Logged Light Lopped Light lopped grazed logged Total tree density 623.1 643.9 639.9 627.6 645.1 623.5 (ha-1) (24.67) (28.73) (34.14) (19.25) (33.47) (20.49) Large tree density 65.7 109.2 51.4 117.9 67.8 103.2 (ha-1) (5.06) (7.59) 6(4.07) (6.26) (7.45) (5.95) Total basal area 87.5 110.0 85.1 110.2 79.7 114.3 (m2ha-1) (5.68) (5.59) (5.86) (5.36) (5.55) (5.19) Large tree basal area 38.5 72.6 29.4 77.7 37.4 70.2 (m2ha-1) (3.93) (6.20) (3.15) (5.44) (4.41) (5.63) Tree canopy cover 48.4 56.7 46.5 57.7 47.5 56.7 (%) (1.49) (1.40) (1.47) (1.26) (1.55) (1.32) Shrub density 2714.3 3101.4 2609.7 3159.8 2492.2 3249.0 (ha-1) (218.95) (191.81) (240.23) (172.23) (213.19) (191.83) Shrub cover 28.9 33.5 28.7 33.2 26.7 34.8 (%) (1.79) (1.71) (2.02) (1.52) (1.73) (1.66) Herbaceous cover 28.6 34.2 31.5 31.2 30.6 31.9

(%) (1.65) (1.80) (1.56) (1.89) (1.65) (1.83)

138

Table B.2 Results of analysis of variance (ANOVA) for the vegetation characteristics across the different disturbance types. Disturbance types Variables F value P value

Grazing Tree density 0.80 0.37 Large tree density 2.71 0.10 Basal area 0.84 0.36 Large basal area 2.25 0.14 Tree canopy cover 2.13 0.15 Shrub density 0.01 0.91 Shrub cover 0.57 0.45 Herbaceous cover 8.55 0.00 Logging Tree density 0.04 0.96 Large tree density 22.36 0.00 Basal area 0.25 0.77 Large basal area 9.70 0.00 Tree canopy cover 4.69 0.05 Shrub density 1.10 0.34 Shrub cover 0.73 0.48 Herbaceous cover 1.89 0.16 Lopping Tree density 0.16 0.68 Large tree density 0.73 0.39 Basal area 9.04 0.00 Large basal area 0.29 0.59 Tree canopy cover 2.33 0.13 Shrub density 4.02 0.05 Shrub cover 7.94 0.00 Herbaceous cover 0.03 0.85 Grazing*logging Tree density 9.75 0.00 Large tree density 10.32 0.00 Basal area 0.03 0.85 Large basal area 2.69 0.10 Tree canopy cover 4.28 0.04 Shrub density 0.21 0.65 Shrub cover 1.01 0.32

139

Herbaceous cover 27.83 0.00 Grazing*Lopping Tree density 0.01 0.93 Large tree density 4.80 0.03 Basal area 0.37 0.54 Large basal area 0.02 0.96 Tree canopy cover 0.85 0.36

140

Table B.2(continued) Disturbance types Variables F value P value Shrub density 1.09 0.29 Shrub cover 1.62 0.20 Herbaceous cover 0.26 0.60 Logging*Lopping Tree density 0.66 0.52 Large tree density 1.29 0.28 Basal area 0.05 0.94 Large basal area 0.35 0.70 Tree canopy cover 0.60 0.55 Shrub density 0.11 0.89 Shrub cover 0.54 0.58 Herbaceous cover 0.72 0.49 Grazing*Logging*Lopping Tree density 0.48 0.48 Large tree density 2.40 0.12 Basal area 0.63 0.43 Large basal area 0.24 0.62 Tree canopy cover 0.66 0.42 Shrub density 0.58 0.45 Shrub cover 9.11 0.00 Herbaceous cover 0.43 0.51

141

Table B.3 Results of analysis of variance (ANOVA) for overall species richness, total average abundance, species richness and abundance of bark-gleaning and foliage-gleaning insectivore Disturbance types variable F-value P-value

Grazing Overall species richness 10.29 0.00 Average abundance 3.05 0.08 Bark gleaner species richness 0.89 0.35 Bark gleaner abundance 3.28 0.07 Foliage gleaner species richness 10.80 0.00 Foliage gleaner abundance 4.91 0.03 Logging Overall species richness 2.74 0.07 Average abundance 3.50 0.03 Bark gleaner species richness 3.47 0.03 Bark gleaner abundance 4.46 0.01 Foliage gleaner species richness 1.68 0.19 Foliage gleaner abundance 0.30 0.74 Lopping Overall species richness 27.60 0.00 Average abundance 2.90 0.09 Bark gleaner species richness 1.01 0.32 Bark gleaner abundance 0.01 0.93 Foliage gleaner species richness 7.14 0.01 Foliage gleaner abundance 6.89 0.01 Grazing * logging Overall species richness 0.66 0.42 Average abundance 0.19 0.60 Bark gleaner species richness 0.51 0.47 Bark gleaner abundance 0.18 0.67 Foliage gleaner species richness 0.28 0.60 Foliage gleaner abundance 0.10 0.76 Grazing * lopping Overall species richness 0.43 0.51 Average abundance 0.29 0.59 Bark gleaner species richness 0.15 0.70 Bark gleaner abundance 0.11 0.74 Foliage gleaner species richness 0.05 0.82 Foliage gleaner abundance 0.06 0.81 Logging * lopping Overall species richness 0.63 0.53

142

Average abundance 0.24 0.79 Bark gleaner species richness 0.21 0.81 Bark gleaner abundance 0.05 0.95 Foliage gleaner species richness 0.19 0.83 Foliage gleaner abundance 0.02 0.98 Grazing * logging * lopping Overall species richness 0.64 0.42 Average abundance 0.41 0.52 Bark gleaner species richness 0.00 1.00 Bark gleaner abundance 1.18 0.28 Foliage gleaner species richness 0.27 0.60 Foliage gleaner abundance 0.14 0.70

143

Table B.4 Model averaged coefficient and sum of Akaike weights (∑ωі) (fixed effects) for each variable based on AIC. Significant estimates are in bold.

Response Variable Coefficient SE Lower Upper ∑ ωі group CI CI Overall bird Large trees 0.06 0.02 0.03 0.12 0.84 species Forest extent (2.5 km radius) 0.14 0.03 0.05 0.17 1.00 richness Tree canopy cover 0.08 0.03 0.04 0.14 0.95 Shrub density 0.06 0.03 0.00 0.10 0.59 Forest extent (0.5 km radius) 0.06 0.03 0.00 0.10 0.55 Road density -0.02 0.03 -0.06 0.04 0.25 Average Large trees 0.06 0.02 0.02 0.09 0.99 abundance Forest extent (2.5 km radius) 0.10 0.04 0.05 0.20 0.98 Tree canopy cover 0.15 0.02 0.12 0.18 1.00 Shrub density 0.02 0.01 -0.02 0.05 0.32 Forest extent (0.5 km radius) 0.02 0.01 -0.04 0.04 0.30 Road density 0.03 0.04 -0.06 0.09 0.32 Bark-gleaning Large trees 0.16 0.06 0.02 0.27 0.82 insectivores Forest extent (2.5 km radius) 0.23 0.08 0.01 0.31 0.80 richness Tree canopy cover 0.15 0.07 0.00 0.27 0.68 Shrub density 0.12 0.06 -0.02 0.22 0.66 Forest extent (0.5 km radius) 0.19 0.08 0.02 0.32 0.99 Road density -0.02 0.07 -0.16 0.11 0.25 Bark-gleaning Large trees 0.21 0.05 0.10 0.31 1.00 insectivores Forest extent (2.5 km radius) 0.28 0.12 0.06 0.51 0.91 abundance Tree canopy cover 0.14 0.04 -0.07 0.13 0.97 Shrub density 0.03 0.03 0.05 0.22 0.26 Forest extent (0.5 km radius) 0.10 0.04 -0.04 0.18 0.58 Road density 0.02 0.11 -0.20 0.23 0.26 Foliage- Large trees 0.10 0.06 0.00 0.24 0.58 gleaning Forest extent (2.5 km radius) 0.28 0.07 0.07 0.46 1.00 insectivores Tree canopy cover 0.15 0.07 0.01 0.26 0.87 richness Shrub density 0.13 0.06 -0.02 0.25 0.57 Forest extent (0.5 km radius) 0.11 0.05 0.06 0.16 0.74 Road density -0.07 0.07 -0.21 0.08 0.27 Foliage- Large trees 0.09 0.05 -0.01 0.17 0.63 gleaning Forest extent (2.5 km radius) 0.33 0.07 0.09 0.36 1.00

144 insectivores Tree canopy cover 0.13 0.04 0.05 0.21 0.81 abundance Shrub density 0.10 0.04 0.03 0.20 0.79 Forest extent (0.5 km radius) 0.09 0.03 -0.03 0.15 0.55 Road density -0.08 0.10 -0.27 0.11 0.24

145

Table B.5 AIC, ∆value and Akaike weights (ωі) for models of overall species richness, total average abundance, species richness and abundance of bark-gleaning and foliage-gleaning insectivore

Response group Model AIC ∆ AIC ωі Overall species Large trees density + forest extent (2.5 km radius) + 621.10 0.00 0.27 richness shrub density + tree canopy cover + forest extent (0.5 km radius) Large trees density + forest extent (2.5 km radius) + 621.45 0.34 0.23 tree canopy cover + forest extent (0.5 km radius) Large trees density + forest extent (0.5 km radius) + 623.42 2.32 0.08 shrub density + tree canopy cover + forest extent (0.5 km radius) + road density Large trees density + forest extent (2.5 km radius) + 623.49 2.39 0.08 tree canopy cover + forest extent (0.5 km radius) + road density Large trees density + forest extent (2.5 km radius) + 623.74 2.64 0.07 shrub density + tree canopy cover Forest extent (2.5 km radius) + tree canopy cover + 624.19 3.09 0.06 forest extent (0.5 km radius) + shrub density

146

Table B.5 (continued)

Response group Model AIC ∆ AIC ωі Overall Large trees density + Forest extent (2.5 km radius) + 969.5 0.00 0.39 abundance Tree canopy cover 1 Large trees density + Forest extent (2.5 km radius) + 971.0 1.53 0.18 Tree canopy cover + road density 4 Large trees density + Forest extent (2.5 km radius) + 971.7 2.25 0.13 Tree canopy cover + shrub density 7 Large trees density + Forest extent (2.5 km radius) + 971.7 2.28 0.12 Tree canopy cover + Forest extent (0.5 km radius) 9 Large trees density + Forest extent (2.5 km radius) + 973.3 3.8 0.06 shrub density + Tree canopy cover + road density 1 Large trees density + Forest extent (2.5 km radius) + 973.3 3.85 0.06 road density + Tree canopy cover + Forest extent (0.5 6 km radius) Large trees density + Forest extent (2.5 km radius) + 974.0 4.57 0.04 shrub density + Tree canopy cover + Forest extent 9 (0.5 km radius)

147

Table B.5 (continued)

Response group Model AIC ∆ AIC ωі Bark gleaning Large trees density + Forest extent (2.5 km radius) + 405.49 0.00 0.18 insectivores Forest extent (0.5 km radius) + Tree canopy cover richness Forest extent (2.5 km radius) + Forest extent (0.5 km 406.36 0.87 0.13 radius) + Tree canopy cover + Shrub density Forest extent (2.5 km radius) + Large trees density + 406.89 1.40 0.09 Forest extent (0.5 km radius) + Shrub density Large trees density + Forest extent (2.5 km radius) + 406.91 1.43 0.09 Forest extent (0.5 km radius) + Shrub density Forest extent (2.5 km radius) + Forest extent (0.5 km 407.68 2.19 0.06 radius) + road density + Tree canopy cover+ Shrub density Forest extent (2.5 km radius) + Tree canopy cover + 408.03 2.54 0.05 Shrub density Large trees density + Forest extent (2.5 km radius) + 408.19 2.71 0.05 Tree canopy cover + Shrub density Large trees density + Forest extent (2.5 km radius) + 408.60 3.11 0.04 Forest extent (0.5 km radius) + Tree canopy cover + Shrub density + road density

148

Table B.5 (continued)

Response group Model AIC ∆ AIC ωі Bark gleaning Large trees density + Forest extent (2.5 km radius) + 726.00 0.00 0.26 insectivores Tree canopy cover + Forest extent (0.5 km radius) abundance Large trees density + Forest extent (2.5 km radius) + 726.53 0.53 0.20 Tree canopy cover Large trees density + Forest extent (2.5 km radius) + 727.63 1.63 0.12 Tree canopy cover+ Shrub density Large trees density + Forest extent (2.5 km radius) + 727.28 2.28 0.08 Tree canopy cover + road density Large trees density + Forest extent (2.5 km radius) + 728.50 2.49 0.08 Forest extent (0.5 km radius) + Tree canopy cover + Shrub density Large trees density + Forest extent (2.5 km radius) + 728.85 2.85 0.06 Forest extent (0.5 km radius) + Tree canopy cover + road density Large trees density + Forest extent (2.5 km radius) + 729.95 3.94 0.04 Tree canopy cover + Shrub density + road density

149

Table B.5 (continued)

Response group Model AIC ∆ AIC ωі Foliage gleaning Large trees density + Forest extent (2.5 km radius) + 395.19 0.00 0.17 insectivores Tree canopy cover + Shrub density richness Large trees density + Forest extent (2.5 km radius) + 395.46 0.27 0.15 Shrub density Large trees density + Forest extent (2.5 km radius) + 396.62 1.43 0.08 Tree canopy cover + Shrub density + road density Forest extent (2.5 km radius) + Tree canopy cover + 396.72 1.53 0.08 Shrub density Large trees density + Forest extent (2.5 km radius) + 397.17 1.98 0.06 Shrub density + road density Large trees density + Forest extent (2.5 km radius) + 397.38 2.19 0.06 Forest extent (0.5 km radius) + Tree canopy cover + Shrub density Large trees density + Forest extent (2.5 km radius) + 397.55 2.36 0.05 Forest extent (0.5 km radius) + Shrub density Forest extent (2.5 km radius) + Tree canopy cover + 398.03 2.84 0.04 Shrub density + road density

150

Table B.5 (continued)

Response group Model AIC ∆ AIC ωі Foliage gleaning Large trees density + Forest extent (2.5 km radius) + 859.47 0.00 0.24 insectivores Tree canopy cover + Shrub density + Forest extent (0.5 abundance km radius) Forest extent (2.5 km radius) + Tree canopy cover + 860.23 0.76 0.16 Shrub density + Forest extent (0.5 km radius) Large trees density + Forest extent (2.5 km radius) + 860.66 1.19 0.13 Tree canopy cover + Shrub density Forest extent (2.5 km radius) + Tree canopy cover + 860.99 1.53 0.11 Shrub density Large trees density + Forest extent (2.5 km radius) + 861.11 1.64 0.11 Tree canopy cover + Forest extent (0.5 km radius) Forest extent (2.5 km radius) + Large trees density + 861.94 2.47 0.07 Tree canopy cover Forest extent (2.5 km radius) + Forest extent (0.5 km 862.37 2.90 0.06 radius) + Tree canopy cover Forest extent (2.5 km radius) + Forest extent (0.5 km 862.76 3.29 0.05 radius) + road density + Tree canopy cover + Shrub density + Large trees density Forest extent (2.5 km radius) + Forest extent (0.5 km 863.21 3.74 0.04 radius) + road density + Tree canopy cover+ Shrub density

151

Table B.6 AIC, ∆value and Akaike weights (Ѡі) for interaction models of overall species richness, total average abundance, species richness and abundance of bark-gleaning and foliage-gleaning insectivore

Response group Model AIC ∆ AIC ωі Overall species Forest extent (2.5 km radius) + disturbance + 623.7 0.00 0.61 richness disturbance * Forest extent (2.5 km radius) Forest extent (2.5 km radius) + Forest extent (0.5 626 2.22 0.2 km radius) + disturbance + disturbance * Forest extent (2.5 km radius) + disturbance * Forest extent (0.5 km radius) Forest extent (0.5 km radius) + disturbance + 626.76 2.99 0.14 Forest extent (2.5 km radius) Forest extent (2.5 km radius) + disturbance + 628.87 5.09 0.05 disturbance * Forest extent (0.5 km radius) Overall Forest extent (2.5 km radius) + Forest extent (0.5 1336.79 0.00 0.56 abundance km radius) + disturbance + disturbance * Forest extent (0.5 km radius) + disturbance * Forest extent (2.5 km radius) Forest extent (2.5 km radius) + Forest extent (0.5 1337.26 0.47 0.44 km radius) + disturbance + disturbance * Forest extent (2.5 km radius) Bark- gleaning Forest extent (2.5 km radius) + Forest extent (0.5 405.72 0.00 0.45 insectivores km radius) richness Forest extent (2.5 km radius) + Forest extent (0.5 407.06 1.34 0.23 km radius) + disturbance + disturbance * Forest extent (0.5 km radius) Forest extent (2.5 km radius) + Forest extent (0.5 407.66 1.94 0.17 km radius) + disturbance Forest extent (2.5 km radius) + Forest extent (0.5 409.29 3.57 0.08 km radius) + disturbance + disturbance * Forest extent (2.5 km radius) + disturbance *Forest extent (0.5 km radius) Forest extent (2.5 km radius) + Forest extent (0.5 409.84 4.12 0.06 km radius) + disturbance + disturbance * Forest extent (2.5 km radius)

152

Table B.6 (continued)

Response group Model AIC ∆ AIC ωі Bark- gleaning Forest extent (2.5 km radius) + Forest extent 898.63 0.00 0.69 insectivores (0.5 km radius) + disturbance + disturbance * abundance Forest extent (2.5 km radius) Forest extent (2.5 km radius) + Forest extent 900.22 1.6 0.31 (0.5 km radius) + disturbance + disturbance * Forest extent (0.5 km radius) + disturbance * Forest extent (2.5 km radius) Foliage- gleaning Forest extent (2.5 km radius) + Forest extent 384.43 0.00 0.24 insectivores (0.5 km radius) + disturbance + Forest extent richness (0.5 km radius) * disturbance Forest extent (2.5 km radius) + Forest extent 384.55 0.11 0.23 (0.5 km radius) + disturbance Forest extent (2.5 km radius) + Forest extent 384.58 0.15 0.22 (0.5 km radius) + disturbance + Forest extent (2.5 km radius) * disturbance Forest extent (2.5 km radius) + Forest extent 385.02 0.58 0.18 (0.5 km radius) + disturbance + Forest extent (2.5 km radius) * disturbance + Forest extent (0.5 km radius) * disturbance Forest extent (2.5 km radius) + Forest extent 386.23 1.79 0.10 (0.5 km radius) Forest extent (2.5 km radius) + disturbance + 391.27 6.84 0.10 Forest extent (2.5 km radius) * disturbance Foliage- gleaning Forest extent (2.5 km radius) + Forest extent 1044.98 0.00 0.96 insectivores (0.5 km radius) + disturbance + disturbance * abundance Forest extent (0.5 km radius) + disturbance * Forest extent (2.5 km radius) Forest extent (2.5 km radius) + Forest extent 1051.45 6.46 0.04 (0.5 km radius) + disturbance + disturbance * Forest extent (0.5 km radius)

153

Table B. 7 Mean values of different disturbance types between sites classified as lightly and heavily disturbed in lowland Terai forests.

Lightly disturbed sites Heavily disturbed sites Number of Number of Number of Number of Number Number of Summary statistics lopped tree dung piles cut stumps lopped tree of dung cut stumps branches branches piles Mean 1.7 2.0 0.5 46.5 12.5 11.1 Standard Error 0.4 0.4 0.1 3.4 0.9 0.9 Standard Deviation 2.8 3.3 0.9 24.8 6.5 6.4 Range 9.6 11.0 5.0 79.8 28.0 25.5 Count 60.0 60.0 60.0 52.0 52.0 52.0

154

Table B.8 Correlation matrix of explanatory variables measured. Coefficients in bold shows pairs highly correlated variables. Explanatory variables 1 2 3 4 5 6 7 8 9 1. Forest extent (2.5 km radius) 1 2.Road density 0.12 1 3. Large tree density 0.33 0.07 1 4. Total basal area 0.36 0.13 0.88 1 5.Large trees basal area 0.35 0.08 0.89 0.93 1 6. Tree canopy cover 0.43 0.14 0.30 0.30 0.21 1 7.Shrub density 0.12 -0.15 0.22 0.08 0.15 0.17 1 8.Shrub cover 0.13 -0.15 0.35 0.19 0.21 0.15 0.57 1 9.Forest extent (0.5 km radius) 0.14 0.02 0.24 0.23 0.12 0.27 0.30 0.26 1

155

Appendix C

Table C.1 AIC, Δ value and Akaike weights (ωі) for models of overall estimated species, frugivore, foliage-gleaning insectivore and sallying insectivore

Response group Model AIC ΔAIC ωі All birds Forest extent +water body 252.85 0.00 0.19 Forest extent +rainfall +water body +forest extent *rainfall 252.97 0.11 0.18 Forest extent +rainfall +water body 253.85 1.00 0.12 Forest extent +water body +rainfall +disturbance 253.86 1.00 0.12 +disturbance*forest extent +forest extent*rainfall Disturbance +forest extent +water body+ 254.62 1.76 0.08 disturbance*forest extent Forest extent +rainfall 255.02 2.16 0.07 Disturbance +forest extent +water body 255.49 2.64 0.05 Disturbance +forest extent +rainfall +water body 255.98 3.13 0.04 +forest extent*rainfall Forest extent +rainfall +forest extent*rainfall 256.49 3.63 0.03 Disturbance +forest extent +rainfall +water body 256.81 3.96 0.03 Disturbance +forest extent +water body +rainfall 256.97 4.11 0.02 +disturbance +disturbance*forest extent Disturbance +forest extent+ rainfall 257.72 4.87 0.02

156

Table C.1 (continued)

Response groups Model AIC ΔAIC ωі Frugivores Forest extent + water body 152.59 0.00 0.54 Forest extent +rainfall +water body 155.33 2.74 0.14 Disturbance + forest extent +water body 155.33 2.74 0.14 Disturbance + forest extent +water body 157.67 5.07 0.04 +disturbance*forest extent Forest extent +rainfall+ water body +forest extent 157.86 5.26 0.04 *rainfall Disturbance + forest extent+ rainfall + water body 158.32 5.73 0.03 Water body 159.37 6.77 0.02 Disturbance + water body 160.6 8.01 0.01 Rainfall + water body 160.78 8.18 0.01

157

Table C.1 (continued)

Response groups Model AIC ΔAIC ωі Foliage-gleaning Forest extent +rainfall + forest extent*rainfall 146.56 0.00 0.39 insectivores Disturbance + forest extent +rainfall +forest 147.12 0.56 0.3 extent * rainfall Forest extent +rainfall+ water body + forest 148.89 2.33 0.12 extent * rainfall Disturbance +forest extent +rainfall +disturbance 149.93 3.37 0.07 * forest extent+ forest extent * rainfall Disturbance +forest extent +rainfall+ water body 149.93 3.37 0.07 + forest extent * rainfall Disturbance +forest extent +rainfall+ water body 152.88 6.31 0.02 + disturbance * forest extent + forest extent * rainfall Forest extent + rainfall + water body 154.49 7.93 0.01

158

Table C.1 (continued)

Response groups Model AIC ΔAIC ωі Sallying Null 148.75 0.00 0.28 insectivores Forest 150.84 2.09 0.1 Water body 150.85 2.10 0.1 Rain 150.92 2.17 0.1 Disturbance 150.97 2.23 0.09 Forest +water body 152.96 4.21 0.03 Rain +water body 152.96 4.21 0.03 Forest + rain 153.12 4.37 0.03 Disturbance + forest + disturbance * forest 153.23 4.48 0.03 Disturbance + water body 153.25 4.51 0.03 Disturbance + rain 153.27 4.53 0.03 Disturbance + forest 153.34 4.59 0.03 Forest + rain + forest * rain 153.71 4.96 0.02 Forest + rain + water + forest * rain 154.38 5.63 0.02 Forest + rain + water 154.86 6.11 0.01 Disturbance + rain + water body 155.44 6.69 0.01 Disturbance + forest + water body 155.69 6.94 0.01

159

Table C.2 Model averaged coefficients across the 95% confidence set of models for all explanatory variables Response variables Explanatory variables Estimate Std error z-value p-value All birds Forest extent 0.18 0.03 4.90 <0.001 Water body 0.09 0.37 2.36 0.018 Rainfall 0.03 0.04 0.62 0.536 Disturbance -0.02 0.03 0.63 0.529 Forest extent*Rainfall 0.52 0.03 1.80 0.071 Disturbance *Forest extent 0.07 0.03 1.93 0.040 Frugivores Forest extent 0.20 0.07 2.71 0.007 Water body 0.21 0.06 3.28 0.001 Rainfall 0.01 0.08 0.16 0.877 Disturbance 0.00 0.07 0.07 0.948 Forest extent *Rainfall -0.05 0.06 0.72 0.474 Disturbance *Forest extent 0.05 0.07 0.76 0.447 Foliage gleaners Forest extent 0.55 0.10 5.02 <0.001 Water body 0.08 0.09 0.78 0.437 Rainfall 0.44 0.11 3.66 0.01 Disturbance 0.11 0.08 1.38 0.168 Forest extent*Rainfall -0.30 0.10 2.94 0.01 Disturbance *Forest extent 0.06 0.08 0.68 0.42 Sallying insectivores Forest extent -0.05 0.08 0.53 0.597 Water body -0.06 0.09 0.61 0.543 Rainfall 0.05 0.08 0.53 0.600 Disturbance 0.02 0.08 0.25 0.802 Forest extent *Rainfall -0.11 0.07 1.45 0.148 Disturbance *Forest extent 0.14 0.08 1.56 0.119

160

Figure C.1 Species accumulation curves of all 28 studied landscapes based on Chao2/ICE estimated richness.

161