Mammalian Diversity Patterns: Effects of Bias and Scale
Total Page:16
File Type:pdf, Size:1020Kb
Mammalian diversity patterns: effects of bias and scale Ben Collen A thesis submitted for the degree of Doctor of Philosophy at Imperial College London, University of London. September 2005 Abstract Abstract In recent years, conservation research has dramatically increased appreciation of the underlying mechanisms of changing biodiversity patterns. More specifically, phylogenetic comparative analyses have provided conservation biologists with a more rigorous tool to explore and understand the underlying processes and patterns of contemporary extinction. I use methods that control for phylogenetic non-independence of taxa, to examine several aspects of mammalian diversity patterns. The first section addresses a number of issues that may bias our current views of patterns of biodiversity. Species description is ongoing, even among well-studied taxa, so how will currently undiscovered species alter our view of biodiversity? Results from chapter two suggest that within groups, we can make some robust generalisations about what undescribed species may be like. A key tool for conservation is the assessment of extinction risk. As the impact and use of threatened species lists become wider reaching, the imperative to ensure they are robust, objective and reliable becomes greater. In chapter three I find that Red List assessments among multiple users of varying experience are remarkably consistent. The final category on the IUCN Red List is ‘Extinct’. Though apparently straightforward, applying the category can be problematic, as the disappearance of the final individual of a species is rarely observed. Conservation science must seek more robust and objective means to classify species for when data are sparse. A potential solution to the problem, which utilises a species sighting record, is tested in chapter four. The second section of this thesis addresses extinction risk and the effect of scale. Global species extinction typically represents the end point in a long sequence of population declines and local extinctions, but there is still little appreciation of how local processes scale up to global patterns. Regional scale analyses presented in chapters five and six, provide some insight and unlike the majority of global analyses, are more able to account for specific threatening processes which, mediated by species biology, are driving extinction. Making progress in understanding the processes 2 Abstract governing species extinction may require robust generalisations about the relationship of extinction between the population and species level. I synthesise these results in chapter seven, with published analyses of extinction risk and show some that correlates are really quite robust to changes in scale. 3 Declaration Declaration All the work presented in this thesis is my own, with the following acknowledgements: Chapter 2 This chapter is published in Global Ecology and Biogeography as a senior authored paper with Andy Purvis and John Gittleman. I was responsible for all aspects of the design, analysis and writing of the paper. The other two authors provided data and thoughtful comments on manuscript structure and clarity. Chapter 3 The idea for this chapter came from Georgina Mace. Many people at the Global Mammal Assessment, African small mammals workshop, and student groups from Imperial College London and Manchester Metropolitan University completed Red List assessments to enable me to write this chapter. All aspects of the design, analysis and writing are my own. Chapter 4 Part of the data presented in this chapter was collated whilst working at the University of Sydney in Chris Dickman’s lab. Dave Roberts and Andy Solow answered my incessant questions about the technique used. Dave Orme and Marcus Rowcliffe helped me implement it. Analyses and writing are my own. Chapter 5 This chapter is to be published in Biodiversity and Conservation in a special issue on extinction. Data for this chapter came from the INTAS funded project 99-1483 from E.J. Milner-Gulland and Stephen Ling. Elena Bykova was instrumental in collating and validating information for the database. Andy Purvis, E.J. Milner-Gulland and Stephen Ling provided comments on the structure and clarity of the final write up – they are all coauthors on the final paper. The design, analysis and write up were my own work. Chapter 6 Data for this chapter was collated during work with Chris Pavey at the Parks and Wildlife Commission of the Northern Territory, Alice Springs, Australia. Marcel Cardillo generated the species lists from Mammal SuperTeam range maps. Data on the environmental variables came from the Pantheria database. All these are fully acknowledged in the text. Ben Collen 4 Acknowledgements Acknowledgements My supervisors Georgina Mace and Andy Purvis have offered advice and support throughout – they have been invaluable and have played a large part in constructing my thoughts on conservation biology. Above all, working with them has been great fun. Jon Bielby, Marcel Cardillo and Sarah Adamovicz have all read and improved multiple versions of my frequently grammar free manuscripts, and I thank them for their invaluable input. Jonathan Baillie, Nick Isaac, Sam Turvey, Craig Hilton-Taylor, Rich Grenyer, John Gittleman, Konrad Dolphin, Zoe Cokeliss, Ian Owens and Tim Barraclough have all offered discussion and advice for which I am grateful. Chris Pavey from the Parks and Wildlife Commission for the Northern Territory and the various members of the Dickman lab at the University of Sydney hosted me during those arduous long hot days of ‘field work’. Chris in particular made me feel very welcome whilst in Alice Springs and for a brief moment I wished we hadn’t beaten you so resoundingly in the Ashes…no, hang on. Andy Solow, Dave Roberts, Andy Beet, Marcus Rowcliffe and Dave Orme answered my questions and offered practical help regarding computing. I thank E.J. Milner-Gulland, Stephen Ling, Lena Bykova, Mike Hoffmann and the Mammal SuperTeam for data and advice. A large group of people, who I won’t name, took the time and trouble to complete Red List assessments necessary for one of my chapters, for which I am very grateful. NERC has provided the funding. Finally, countless friends and family have almost unfailingly acted interested in what I’m doing – the glazed over eyes gave you away. 5 Table of contents Table of Contents Abstract 2 Declaration 4 Acknowledgements 5 Table of Contents 6 List of Tables 8 List of Figures 10 Chapter 1 12 Introduction & chapter rationale 12 Bias in diversity patterns 12 Scaling of extinction 15 Chapter 2 17 Biological correlates of description date in carnivores and primates 17 Introduction 17 Methods 19 Data 19 Phylogenetic analysis 20 Species analysis 22 Results 22 Contrasts analysis 22 Species analysis 23 Discussion 23 Tables and Figures 29 Chapter 3 34 Consistent conservation: examining the reliability of IUCN Red List assessments 34 Introduction 34 Methods 37 Species information 37 Threat assessment protocol & application 37 Analysis 38 Results 39 Application of categories & criteria 39 Error – occurrence, direction, magnitude and type 40 Discussion 41 Consistency of classification 41 In the context of other risk assessment protocols 44 Recommendations 44 Tables and Figures 46 Chapter 4 54 When is a species really extinct? Inferring extinction from a sighting record to inform the IUCN Red List 54 Introduction 54 Methods 56 Data 56 6 Table of contents Analysis 57 Results 59 Tests of robustness 59 Estimating date of extinction 59 Discussion 60 Concerns over effort 62 Levels of certainty 62 Tables and Figures 64 Chapter 5 71 Regional extinction risk in central Asian vertebrates 71 Introduction 71 Methods 72 Data 72 Analysis 73 Results 75 Discussion 77 Tables and Figures 81 Chapter 6 87 Bioregional patterns of extinction in Australian mammals 87 Introduction 87 Methods 89 Database & data collection 89 Analysis 92 Results 94 Distribution of extinction 94 Predictors of extinction pattern 94 Discussion 95 Patterns of extinction 95 Inferring causes of extinction 97 Methodological issues & implications for risk scaling 99 Tables and Figures 101 Chapter 7 108 Synthesis: does extinction risk scale from a local to global level? 108 Summary 108 Introduction 109 Methods 111 Data collation 111 Analysis 113 Results 114 Discussion 115 Evidence of an effect of scale 115 Methodological concerns 116 Processes & measures 118 Conclusions 119 Tables and Figures 120 Chapter 8 126 Conclusions 126 Bias in diversity patterns 126 The scaling of extinction risk 127 Appendices 129 References 152 7 List of Tables List of Tables Table 2.1. Wilcoxon’s signed rank tests of categorical variables predicting description date. 29 Table 2.2. Results of single predictor regressions of contrast data and species data for each order. 30 Table 2.3. Multiple regression model for carnivores predicting description date. 31 Table 2.4. Multiple regression model across carnivores and primates. 31 Table 3.1. IUCN Red List categories and criteria. 46 Table 3.2. Variation in classification amongst the assessors. 47 Table 3.3. Variation within IUCN Red List category. 48 Table 3.4. Analysis of Deviance for the proportion of errors made by assessors in each level of experience. 49 Table 3.5. Proportions of different types of errors made by assessors at each level of experience. 49 Table 4.1. Evaluation of Type I error rate. Date of extinction sequentially estimated for each successive number of sightings starting from the most distant. 64 Table 4.2. Estimated dates of extinction for 10 species listed as Regionally Extinct in NSW. 65 Table 4.3. Estimated extinction dates for 10 Asian bird species considered globally extinct. 66 Table 5.1. Assessment of the contribution of phylogeny to each of the traits. 81 Table 5.2. Single predictor regressions for the phylogenetic analysis of correlates of threat on regional IUCN threat rating. 82 Table 5.3. Single predictor regressions of species data of correlates of threat on regional IUCN threat rating. 83 Table 5.4. Single predictor regressions for the phylogenetic analysis of correlates of threat on regional IUCN threat rating in mammals: game species and non-game species.