Phd Thesis Jonas Geldmann

UNIVE RSI T Y OF COPE NHAGEN FACULT Y OF S C IENCE

CENTER FOR M A CROECOL OGY, EVOLUT ION A ND CLIMAT E

PhD thesis Jonas Geldmann

Evaluating the effectiveness of protected areas for maintaining biodiversity, securing habitats, and reducing threats

Academic advisor: Professor Neil D. Burgess

Submitted: August 2013

UNIVERSITY OF COPENHAGEN FACULTY OF SCIENCE

CENTER FOR MACROECOLOGY,

EVOLUTION AND CLIMATE

PhD thesis Jonas Geldmann

Evaluating the effectiveness of protected areas for maintaining biodiversity, securing habitats, and reducing threats

Academic advisor: Professor Neil D. Burgess

This thesis has been submitted august 2013 to the PhD School of The Faculty of Science, University of Copenhagen

Institutnavn: Biologisk Institut (Center for Makroøkologi, Evolution og Klima)

Name of department: Department of Biology (Center for Macroecology, Evolution and Climate)

Author: Jonas Geldmann

Titel og evt. undertitel: Beskyttede områders evne til at bevare landskaber og biodiversitet samt reducere menneskelig trusler

Title / Subtitle: Evaluating the Effectiveness of Protected Areas for maintaining Biodiversity, securing habitats, and reducing threats

Subject description: This PhD. is part of the conservation theme at the Center for Macroecology, Evolution and Climate (CMEC). The main focus has been to understand how terrestrial protected areas help protect nature and reduce human impact by evaluating changes in state or pressure as a consequence of protected areas as a conservation response.

Academic advisor: Professor Neil D. Burgess, Center for Macroecology, Evolution and Climate, UNEP World Centre for Monitoring of Conservation, Cambridge, United Kingdom, and World Wildlife Fund, USA.

External advisor: Dr. Lauren Coad, Environmental Change Institute, University of Oxford, United Kingdom

Submitted: August 2013

Grade: PhD thesis

Cover photo: elephant: wallcloud. rhino: Brent Stirton,

4 Preface

This thesis is the product of a three year PhD project at the faculty of Science, University of Copenhagen, based at the Center for Macroecology, Evolution and Climate (CMEC). The thesis has been supervised by Professor Neil D. Burgess, but also Dr. Lauren Coad, Environmental Change Institute, University of Oxford, United Kingdom has functioned as an external supervisor though not officially affiliated with the project. While the base has been the Center for Macroecology, Evolution and Climate all work has involved international collaborators of which I have been able to spend time at many of their institutions. In total I have had 11 different office spaces between; The University of Copenhagen, University of Cambridge, United Kingdom, University of Oxford, United Kingdom, the United Nations Environmental Programme’s World Centre for Monitoring Conservation (UNEP-WCMC), and the Zoological Society of London (ZSL), United Kingdom. In total I have spent a little less than a year between Cambridge, Oxford and London including a four month stay at ZSL. Besides these institutions, this thesis has also been part of the IUCN WCPA/SSC joint taskforce on Biodiversity and Protected Areas. In this group we have had many and very fruitful workshops and meeting in, Australia, Canada, New Zealand, United Kingdom, and United States of America.

The thesis consists of two parts. First, a synopsis giving the background, overview, objectives, main findings of the thesis, and some perspectives. The second consist of six research chapters on the effectiveness of protected areas using different approaches and covering ‘state’, ‘pressure’, and ‘response’. Uniting them is the use of temporal data to explore the performance of protected areas. Three of the chapters are either published or accepted for publication. The remaining three are written as scientific research papers.

Besides the included research chapters I have been heavily involved in another research project not at the stage to be included in the final PhD. Through the IUCN WCPA/SSC taskforce I have also been involved in a report for the Global Environmental Facility on their protected areas as well as contributed to the 2012 Protected Planet report. While being a PhD student I have also acted as co-supervisor on two master theses, two bachelor projects as well as assisted with the teaching in i) International Nature Conservation (graduate level), ii) Experimental design and applied statistics (graduate level), and iii) organismal diversity (undergraduate level),

Jonas Geldmann Copenhagen, Denmark, August 2013

6 Table of Contents

Acknowledgement ...... 9 Summary ...... 11 Sammenfatning ...... 13 Synopsis ...... 15 Introduction ...... 15 Evaluating protected area effectiveness ...... 18 Objectives ...... 26 Main findings and perspectives ...... 27 References ...... 30 Evidence of protected area effectiveness ...... 39 Abstract ...... 41 Introduction ...... 41 Methods ...... 42 Results ...... 42 Discussion ...... 46 Acknowledgements ...... 47 References ...... 47 Commonalities and complementarities in Management and Evaluations ...... 51 Introduction ...... 54 Concepts and terminology ...... 55 Approaches to Conservation M&E ...... 56 Discussion ...... 62 Conclusion ...... 68 Acknowledgements ...... 68 Literature cited ...... 68 Management effectiveness and global commitments ...... 73 Abstract ...... 75 Introduction ...... 75 Methods ...... 77 Results ...... 79 Discussion ...... 80 References ...... 84 Changes in Management Effectiveness ...... 87 Introduction ...... 90 Methods ...... 91 Results ...... 93 Discussion ...... 96

7 Acknowledgements ...... 99 References ...... 99 Protected Areas ability to reduce pressure ...... 103 Abstract ...... 105 Introduction ...... 106 Methods ...... 107 Results ...... 113 Discussion ...... 117 Acknowledgements ...... 120 References ...... 120 The effect of management in protected areas on species populations ...... 125 Abstract ...... 127 Introduction ...... 128 Methods ...... 129 Results ...... 132 Discussion ...... 136 Acknowledgement ...... 140 References ...... 140 Appendix I (chapter I) ...... 149 Appendix II (chapter V) ...... 160

Appendix III (chapter VI) ...... 173

Acknowledgements I first and foremost want to thank my supervisor Neil Burgess for his incredible support, supervision, and guidance throughout my PhD. Even if physically located in a different country I never felt forgotten or far from help or advice. His support and commitment to my project and always encouragement to pursuit my ideas have been cardinal in making it through. I particularly want to emphasize my gratitude to him for introducing me the field of conservation science as being more than just academic and giving me a look in to all of the components it takes to preserve and protect nature and biodiversity. I also want to thank my other supervisor, Lauren Coad, who even if not officially associated with the PhD-project has been deeply involved in all phases. Also thank you for welcoming me into your home when I was in Oxford (and of course thanks to Tom C. too).

I want to thank Carsten Rahbek, who has been a great support and inspiration and who, through the Center for Macroecology, Evolution and Climate (CMEC) has created an incredible environment both academically inspiring and socially engaging. Everyone at CMEC reserves great thanks and everyone has been part of making this PhD enjoyable, providing help and advice when needed as well as a social atmosphere surely envied by many PhD students elsewhere. I particularly want to thank Peter Søgaard Jørgensen for his helps with statistics and discussion of results and ideas, as well as many inspiring discussions on scientific challenges as well as the state and politics of the university (and academia in general) – most stretching back before either of us started our PhDs. A big thanks to Jacob Heilmann-Clausen with whom I have shared an office throughout, and who have had to put up with my sometimes too excellent sense of humor, but also for some interesting discussions on conservation science which has helped me bridge my own international academic focus with my interest in Danish nature conservation, and which has led to two master theses and two bachelor projects co-supervised. Also a great thanks to Bjørn Hermansen for his patient and much needed help with ArcGIS as well as Erik Mousing for help with statistics. And of course a great many thanks to the rest of the Section for Ecology and Evolution who has helped make this such a pleasant place to work doing the PhD – and actually for the five years prior to the PhD, where Universitetsparken 15 has been my workplace.

For extended periods of time I have worked outside of Denmark. In total I have spent more than five month at the Institute of Zoology, at the Zoological Society of London (ZSL). I particularly want to thank Ben Collen and Louise McRae for support with data collections and analysis as well as providing a fantastic and welcoming environment. Also great thanks to Monika Böhm, Ellie Dyer, Jon Bielby, and Sarah Whitmee with whom I shared offices with. And thanks to Martina De Fonzo, Lucie Bland and countless others for making the social and academic setting inspiring during my stays in London – which of course includes Jim and the social club as well as the awesome IoZ/CP football team.

Similarly I have spent many months in Cambridge at the university and UNDPs World Conservation Monitoring Centre. I want to thank Wendy Foden who has hosted me for most of my stays and who has always made we fell welcome in her fantastic house – always making me look forward to my next visit. Thanks also to Jonathan Green for sharing his office at the university and everyone at WCMC who have made every visit pleasant and inspiring. And thanks to Lucas Joppa who has helped and guided me when the match got tricky.

This PhD has been possible to a large extent through collaborative efforts where my closest colleagues have often been on the other side of the world. Particularly the IUCN WCPA/SSC joint taskforce on Biodiversity and Protected Areas has been a very important part of this PhD. I especially want to thank Stephen Woodley who has been an incredible support throughout the PhD and has to a large extent made this international collaboration possible – which has been and still is very productive. Your efforts, Stephen, has opened doors and generated possibilities that only few PhD student gets to experience, and you have always shown a respect that has been inspirational and empowering in itself – thanks. I also want to thank Tom Brooks, Marc Hockings and Kent Redford who are also been involved as seniors in the taskforce and who has been very supportive and inspirational. And of course many thanks to Megan Barnes, Ian Craigie, and Luke Harrison, for discussions, support, and many long work-hours in exotic locations around the world. Though located across the world you have often been my closest colleagues.

This work would not have been possible without the enormous efforts by staff an interns working on three databases that has been cardinal to this PhD. I thank everyone who through the times has been involved in collation and entering of data in the METT database. Especially thanks to Fiona Leverington, Christoph Nolte, Mark Zimsky, Jessie Mee, Nanna Vansteelant, and many many more. I also thank the armies of interns as well as staff at ZSL who have helped collect the information in the Living Planet Database as well as the thousands investing entire carries in collecting the population data in the field over decades and decades. I also thank the people working hard making the World Database on Protected Areas available; Bastian Bertzky, Amy Milan, and Naomi Kingston.

I also want to thank all the people around me, friends and family, who has always supported me and often given me the much needed breaks away from my work. I especially want to thank my mother Mette Geldmann, who has always supported me and helped me when needed. I also want to thank Søren Grimstrup, who has been an invaluable resource on statistics and the wonderful language R – having you so close at hand has made a difference.

I thank the Danish National Research Foundation for financial support. I also thank the Hanne and Torkel Weis-Fogh Fund, the International Union for Conservation of Nature (IUCN), the World Bank’s Global Environmental Facility (GEF), and the World Wildlife Fund, USA for financial support doing the PhD.

Summary Protected areas are amongst the most important conservation responses to halt the loss of biodiversity and cover more than 12.7% of the terrestrial surface of earth. Likewise, protected areas are an important political instrument and a key component of the Convention for Biological Diversity (CBD); seeking to protect at least 17% of the terrestrial surface and 10% of the coastal and marine areas by 2020.

Protected areas are expected to deliver on many different objectives covering biodiversity, climate change mitigation, local livelihood, and cultural & esthetic values. Within each of these categories a suite of relevant success matrices exist including; coverage, quality, and performance.

This PhD thesis deals with the effectiveness of terrestrial protected areas using temporal data to explore whether protected areas have had a positive effect as a consequence of their establishment. The aim of this thesis has been to evaluate the performance and effectiveness of protected area in securing biodiversity, by evaluating their ability to either improve conservation responses, the state of biodiversity, or alternatively to reduce the human pressures responsible for the loss of biodiversity. The scope of the thesis has been exclusively terrestrial.

Through the six chapters making up the core of this PhD I have explored the effectiveness of protected areas looking at general patterns transcending individual case- studies. This has involved using large scale existing global data sets, systematic literature reviews and participating in developing global databases. Combined, it explores protected area effectiveness covering changes in i) pressures, ii) state, and iii) responses , as well as setting protected area effectiveness in a topological and political context.

Chapter I is a systematic literature review on the effectiveness of protected areas in delivering either reduced habitat loss or improvements for species populations. Reviewing more than 2,500 articles I find that there are few studies and little evidence for protected areas effect on species populations, making it difficult to draw strong conclusion from already published works. For protected areas ability to reduce habitat loss the evidence is stronger, suggesting that protected areas have been able to reduce the rate of habitat loss compared to a counterfactual scenario. In chapter II we evaluate the different types of methods to assess management, discussing five different generic categories: i) ambient monitoring, ii) management assessment, iii) performance management, iv) impact evaluation, and v) systematic reviews and their differences in application and scope. Chapter III looks at the extent to which the global community is living up to their obligations for the 2020 Aichi target of protecting 17% of their terrestrial surface as well as ensuring that these are effectively and equitably managed. We show that globally 29% of PA area has been assessed, and 23% of countries have reached the 60% target, but also find strong positive biases towards larger protected areas and national parks. Chapter IV is an analysis of how management changes over

11 time in protected areas that has conducted management effectiveness evaluations. On average management improves in protected areas and especially formulations of management plans, monitoring activities, and development of protected area objectives are responsible for the observed improvements. The results also suggests that the reporting usually conducted by local area managers is to some extent a reflection of true improvements and not only motivated by the need to report conservation successes. Chapter V looks at the extent to which protected areas have been able to reduce human pressure. I first developed a global spatial and temporal precise map at 5 km2 resolution of changes in human pressure between 1995 and 2010. This map shows an average increase in human pressure inside protected areas globally, but less so in IUCN management categories with more stick management prescriptions. In chapter VI I look at the correlation between management (using the Management Effectiveness Tracking Tool) and changes in species populations in 53 protected areas across 17 countries, finding that the interaction between Human Development Index and implementation of management plans significantly affects the trend of vertebrate populations, so that areas with has a management plan have are correlated positively to increasing human development.

The results presented in this PhD thesis has contributed to show that protected areas are effective compared to no protection, but also elutes to them not being a panacea for halting the loss of biodiversity. Both populations and habitats are overall decreasing while human pressure both inside and outside protected areas increases. However the results also suggest that management in protected areas do have an effect and that many protected areas have seen an improvement in management over time. A second and perhaps equally important conclusion is that strong empirically based evidence on the effectiveness of protected areas is impeded by the lack of good data to measure change compared to a counterfactual scenario. Too often, data generation has been driven by the need to inform individual studies or fueled by the desire to improve methods thus leaving the conservation community in the quandary of an increasing amount of data continuously being available but still lacking state of the art data to answer the most relevant questions in conservation science.

12 Sammenfatning Internationalt er ’beskyttede områder’ blandt de vigtigste naturbeskyttelsesinstrumenter til at stoppe tabet af biodiversitet. Disse dækker i dag omkring 12,7% af jordens overflade. Ligeledes er beskyttede områder indskrevet i Konventionen for biologisk mangfoldighed ”The Convention on Biological Diversity” (CBD), der forpligter alle underskrivende nationer til at beskyttede minimum 17% af deres landjord og 10% af deres marine områder inden 2020.

Beskyttede områder etableres med en forventning om at levere mere end bare ren beskyttelsen af natur. De forventes at forebygge klimaforandringer, understøtte lokal udvikling samt sikre kulturelle værdier. For hvert af disse formål findes der en lang række relevante mål for, hvor vidt de beskyttede områder er succesfulde dækkende både deres placering, kvalitet og deres effektivitet.

Hovedfokus i denne ph.d.-afhandling har været at undersøge effektiviteten af terrestrisk beskyttede områder ved at bruge tidseriedata til at belyse den forskel etablering og forvaltning af disse områder har haft. Formålet har været at belyse beskyttede områders effektivitet i forhold til at stoppe tabet af biodiversitet ved at evaluere deres evne til at forbedre forvaltningsindsatser, forholdene for biodiversiteten samt reducere menneskelig skadelig påvirkning.

Gennem de seks kapitler, der udgør kernen af denne ph.d.-afhandling, har jeg forsøgt at udforske den allerede eksisterende evidens for om beskyttede områder virker, brugt globale dataset til at analysere beskyttede områder effektivitet samt sat beskyttede områder i en emnemæssig og politisk kontekst.

Kapitel I er et ’systematisk review’ af den eksisterende litterature om beskyttede områders evne til at reducere tabet af habitater og stoppe tabet af biodiversitet. Gennem en systematisk gennemgang af mere end 2500 artikler fandt jeg, at der kun i ringe grad er dokumentation for at beskyttede områder sikrer bestande af dyr. For beskyttede områders evne til at stoppe skovfældning er der langt bedre evidens, der viser at beskyttede områder har været effektive til at forhindre tabet af skov. I kapitel II undersøger og diskuterer vi forskellige metoder til at evaluere forvaltning i conservation og kommer op med fem dinstinkte kategorier for forvaltning og evaluering. Kapitel III er en global evaluering af individuelle landes efterlevelse af CBD’ens mål om at 17% af deres areal ikke bare skal være beskyttet inden 2020 men også forvaltet. Vi finder, at globalt set er 29% af landenes beskyttede områder evalueret for deres forvaltning og at 23% af alle lande allerede har nået målet for 2020, om at 60% af deres beskyttede områder skal være evalueret for deres forvaltningseffektivitet. Kapitel IV undersøger, hvor vidt forvaltning i beskyttede områder i sig selv bliver bedre over tid. Gennemsnitligt blev forvaltningen bedre; særligt forvaltningsplaner, monitering og målsætninger blev forbedret mellem evalueringer af forvaltningseffektiviteten. Kapitel V analyser udviklingen i menneskelige påvirkninger globalt såvel som i beskyttede områder. Først udviklede vi et globalt kort over menneskelige påvirkninger med en

13 opløsning på 5 km2. Dette kort viste, at der globalt set var mere negativ menneskelig påvirkning i 2010 i forhold til 1995. Det viste også, at selv om alle typer af beskyttede områder gennemsnitlig havde det dårligere i 2010 end i 1995 så var påvirkningerne mindre i områder der ifølge IUCN burde være bedre forvaltede. Kapitel VI undersøgte sammenhængen mellem kvaliteten af forvaltning og udviklingen i dyrebestande ved hjælp af ’Management Effectiveness Tracking Tool’ udviklet af verdensbanken og WWF i 53 beskyttede områder i 17 forskellige lande. Beskyttede områders kvalitet var dels bestemt af landenes Human Development Index samt kvaliteten af en forvaltningsplan.

Resultaterne fra denne ph.d.-afhandling har biddraget til at vise, at beskyttede områder har en effekt sammenlignet med områder uden beskyttelse, men har også vist at etableringen af beskyttede områder i sig selv ikke er nogen vidunderkur. Både habitater og populationer er gået tilbage i beskyttede områder, alt imens negativ menneskelige påvirkninger bliver stærkere. En anden og måske lige så væsentlig konklusion er, at vores viden omkring beskyttede områderes effektivitet stadig er meget begrænset. Ofte er dataudvikling drevet af behovet for at forstå små lokale sammenhænge eller af behovet for at udvikle nye og bedre produkter. Dette har efterladt conservation forskere i et dilemma: på den ene side bliver stadig mere datamateriale tilgængeligt, på den anden side, er data til at besvare nogle af de mest fundamentale spørgsmål stadig ikke eksisterende, fordi fokus ikke er på gentagelser men på nyudvikling.

14 synopsis

Synopsis Introduction The biological world is dynamic, constantly changing, governed by processes of ecology and evolution; species go extinct, new species evolve, and ecosystems and habitats disappear even without the assistance of humans. However it is increasingly being recognized that humans are having a profound impact on the earth, unparalleled by any other single species, leading scientist to suggest we have entered a new geological era: the anthropocene (Vitousek et al. 1997; Crutzen 2002; Steffen et al. 2007).

Dramatic human impacts on earth go back for millennia (Balter 2013), long before the industrial revolution – often associated with the onset of the epic present pressure on most natural systems seen today (Steffen et al. 2011) – and have resulted in massive megafauna extinctions (Barnosky et al. 2004; Lorenzen et al. 2011), and loss of natural habitats (Ellis et al. 2013). However, particularly over the last centuries, there have been rapid and intense environmental changes caused by increasing human numbers and technological advances (United Nations Environment Programme 2012). Today more than 75% of the terrestrial surface is impacted by humans (Ellis et al. 2010) and the oceans have experienced dramatic biodiversity declines (Jackson et al. 2001; Worm et al. 2006; Halpern et al. 2012), particularly well-documented for commercial fish stocks (Myers & Worm 2003; Food and Agricultural Organization of the United Nations 2009).

These intense and easily observable impacts as well as the continued declines in nature has been the foundation for conservation science which seeks to act to stop the negative impact of humans in a scientifically sound way building on evidence and an understanding or human actions – both positive and negative.

Pressure-State-Response The Pressure-State-Response (PSR) framework (OECD1993), modified to include benefits (PSRB) (Butchart et al. 2010; Sparks et al. 2011), is a simple way to understand the interaction between threats (pressures), conservation actions (responses), and the natural conditions of the world (state) (Figure 1).

State includes all elements of the environment, nature, biodiversity, and the resources linked to these. As such state reflects the fundamental objective of conservation efforts and policy making. The 1992 Rio summit and the Convention on Biological Diversity (CBD) definition of biodiversity is central to understand state. This covers:

“the variability among living organisms from all sources, including, 'inter alia', terrestrial, marine, and other aquatic ecosystems, and the ecological complexes of which they are part: this includes diversity within species, between species and of ecosystems” (Convention on Biological Diversity 1993)

15 synopsis and makes no distinction between species, or ecosystems, based on their relevance or cultural value for humans. Based on this definition, the objective of conservation is therefore to protect and preserve all species and all ecosystems.

Pressure - State - Response Framework

SERUSSERP ETATS SESNOPSER

Information - Research

State of Biodiversity Economic and Human Activities and the Environment Environmental Agents

Energy Ecosystems Protected areas

Transport Threats Actions Management Industry Species Governance Agriculture Legislation Others Genes Other

Decisions - Actions Figure 1. The Pressure-State-Response framework. Pressure comprises all processes induced directly and indirectly on nature and the environment. State represent all part of the natural system, from ecosystems, and genes, over natural processes such as the nitrogen cycle. Responses are all processes and action by humans, which reduces human pressure and moved the state of nature closer to it natural conditions. Modified from OECD (1993).

There are no reliable estimates of the exact number of species in the world, and extreme estimates ranges from three to 100 million species (May 2010) of which most are still not described, nor even known. Newer studies suggest the true total number is closer to 5-10 million species (Hamilton et al. 2010; Mora et al. 2011). For ecosystems or habitats, there is no universally agreed definition (Gaston 2010). Based on biogeographical distribution data on flora and fauna representing distinct biotas Olson et al. (2001) developed a detailed global terrestrial map consisting of 867 ecoregions within 14 biomes and eight realms. This work builds on older coarser maps with 193 different biotas (Udvardy 1975) and even older work dating back to Alfred Wallace zoogeographic regions (Wallace 1876). Similar systems have been developed for the marine realm consisting of 232 marine ecoregions (Spalding et al. 2007) and 426 fresh water ecoregions (Abell et al. 2008).

Other systems exist like the European Unions (EU) ‘habitat directive’ which operates with 220 distinct European habitat types (European Union 1992). Newer works, building on Wallace (1876) have included genetic information on the phylogenetic history of species introducing an evolutionary aspect in to the present distribution of species, however at a much coarser spatial scale (Holt et al. 2013).

16 synopsis

However protecting species is not only a numbers game of most and many. Nature and species provides services sustaining human livelihood (i.e. ecosystem services) (Balmford et al. 2002; Mace et al. 2012), and these services, also falling in the category of state are instrumental to preserve. But these can be even harder to measure and categorize (Millinium Ecosystem Assessment 2005; Wallace 2007; Fisher & Turner 2008; Tallis et al. 2012) and the interaction between changes in nature and changes in ecosystem services are extremely complex (Tallis et al. 2008). However the subject of ecosystem services has achieved both political and scientific momentum (Perrings et al. 2011), and is increasingly being included in research agendas, economic decision making (Bateman et al. 2013) as well as being integrated in the evaluation of protected areas (Burgess et al. 2009; Balmford et al. 2011).

Pressure constitutes all activities and actions undertaken by humans which threatens and/or reduces the state of nature. Several categorization schemes exist, reflecting the complicated nature of pressures. First, pressures can be perceived differently depending on whether the focal point is the cause or the effect. Second, pressures works both directly and indirectly, and third, pressures rarely work alone but are composite and intertwined. Systems focusing on the cause have divided pressures into five keystone threats (Baldwin 2010) or nine primary threat categories (Salafsky et al. 2008), while systems focusing on the effect have divided threats based on their sources and mechanism (Balmford et al. 2009). Which approach is the most relevant often depends on the specific question being addressed and the data available.

Responses – covers all activities and actions which are undertaken to reduce the effects of human pressure and improve conditions for nature and the environment. Responses are undertaken at both the very local level as well as globally. A suite of different responses to conserve nature exists, covering legislation (e.g. the US endangered species act, and EU REACH), economic incentives (e.g. REDD+), educational and awareness programs (e.g. the UN International Day for biological Diversity – 22 May), and technological innovations (e.g. renewable energy and recycling programs).

Protected areas Perhaps the most far reaching response to the biodiversity crisis has been the development of protected areas, of which more than 170,000 have now been designated, covering more than 12.7% of the terrestrial surface and 1.6% of the world’s oceans (Figure 2) (Bertzky et al. 2012). Protected areas are also included in the Convention for Biological Diversity (CBD) as one of the most prominent instruments to halt the loss of biodiversity, initially by 2010 (Convention on Biological Diveristy 2004), protracted to 2020, by which time 17% of the terrestrial and inland water areas, and 10% of coastal and marine areas should be protected, and effectively and equitably managed (Convention on Biological Diversity 2010).

17 synopsis

Figure 2. Map of the ca. 170,000 protected areas globally. Extracted from the World Database on Protected Areas February 2012.

No “one size fits all” definition for protected areas exists, but it is generally accepted that protected areas should aim to reduce the impact of human activities and preserve and protect natural processes and biodiversity within their borders. The International Union for Conservation of Nature’s (IUCN) has coined what is today one of the most widely accepted definition:

“A protected area is a clearly defined geographical space, recognised, dedicated and managed, through legal or other effective means, to achieve the long term conservation of nature with associated ecosystem services and cultural values.” (Dudley 2008) as well as system defining different management categories depending on the specific aim, naturalness, and human activities allowed within the protected area boundaries (Table 1). However these categories have been criticized for focusing more on the protected area management objectives and the activities allowed than their biodiversity values, making then of little value for evaluation the quality of the protected areas (Boitani et al. 2008).

Evaluating protected area effectiveness Considering the role protected areas play in the efforts to protect and preserve biodiversity understanding how they work is one of the most central questions in conservation. Evaluation of protected areas has been advocated in response to cost- effective use of resources (Mace et al. 2000; Sutherland et al. 2004; Ferraro & Pattanayak 2006), lacking or ineffective management (Hockings & Phillips 1999), failed or ineffective reserves (Fuller et al. 2010; Mascia & Pailler 2011), to promote accountability (Christensen 2003; Jepson 2005), and to ensure natural values and objectives of the reserves (Convention on Biological Diversity 1993, 2010).

18 synopsis

Table 1. IUCN management categories for protected areas Category Name Description Strictly protected areas set aside to protect biodiversity and also possibly geological/geomorphological features, where human Strict Nature Ia visitation, use and impacts are strictly controlled and limited to Reserve ensure protection of the conservation values. Can serve as indispensable reference areas for scientific research and monitoring.

Usually large unmodified or slightly modified areas, retaining their Wilderness natural character and influence, without permanent or significant Ib Area human habitation, which are protected and managed so as to preserve their natural condition

Large natural or near natural areas set aside to protect large‐scale ecological processes, along with the complement of species and II National Park ecosystems characteristic of the area, which also provide a foundation for environmentally and culturally compatible spiritual, scientific, educational, recreational and visitor opportunities.

Are set aside to protect a specific natural monument, which can be a Natural landform, sea mount, submarine cavern, geological feature such as a III Monument or cave or even a living feature such as an ancient grove. They are Feature generally quite small protected areas and often have high visitor value.

Aim to protect particular species or habitats and management reflects Habitat/Species this priority. Many category IV protected areas will need regular, IV Management active interventions to address the requirements of particular species Area or to maintain habitats, but this is not a requirement of the category.

Area where the interaction of people and nature over time has Protected produced an area of distinct character with significant ecological, V Landscape/Seas biological, cultural and scenic value: and where safeguarding the cape integrity of this interaction is vital to protecting and sustaining the area and its associated nature conservation and other values.

Conserve ecosystems and habitats, together with associated cultural Protected Area values and traditional natural resource management systems. They with are generally large, with most of the area in a natural condition, VI sustainable use where a proportion is under sustainable natural resource of natural management and where low‐level non‐industrial use of natural resources resources compatible with nature conservation is seen as one of the main aims of the area. The third version of the IUCN management categories. Adopted from Dudley (2008).

Protected areas can have many different objectives and it is therefore relevant to evaluate their effectiveness on many different parameters. Using the PRSB frameworks (Butchart et al. 2010) protected areas could be expected to i) improve the state of nature, ii) decrease pressures and threats, iii) improve inputs and management and iv) improve benefits and ecosystem services. Ideally protected areas should be effective at all four things, but the weighting and balance between them will vary from site to site, recognizing that optimizing one could reduce the effectiveness or even negatively impact other factors. Understanding which factors are being optimized and which are being examined is thus essential to understand protected area success.

19 synopsis

Figure 3. Grouping of analysis based on their ability to capture the link between input and outcomes. Ranging from gap analysis able only to inform on the planning of conservation based on representing the highest biodiversity (lower left corner, adopted from Rodrigues et al. (2004)) over baseline studies documenting the general trend of biodiversity without link to conservation input (lower right corner adopted from Craigie et al. (2010)) and studies only examining the effectiveness of management not linked to changes in conservation outcomes (upper left corner adopted from Jachmann (2008)) and studies able to explain differences in observed conservation outcomes based on management input (upper right corner adopted from Western et al. (2009)).

When evaluating protected area effectiveness understanding how and why they work (input) as well as what they deliver (outcomes) can help frame what elements of effectiveness are being assesses. Figure 3 divides protected area effectiveness analyses based on their ability to capture the input and outcome elements of protected area effectiveness. There will be many analyses which fall in the cracks between these categories but the framework of understanding to what extent the input and outcomes are being measured is useful as it illustrates a way to understand protected areas as an instrument which has to be evaluated ones established.

All types of analysis can be valid depending on the question being posed and the objectives of the evaluation. However, ensuring that the type of information being collected is aligned with the focal question being posed in the evaluation is central and can improve and direct evaluation efforts (Mascia et al. accepted).

20 synopsis

Effective coverage Traditionally protected areas have been evaluated based on their location and coverage, in relation to ecotypes (Barr et al. 2011) or species representativeness (Rodrigues et al. 2004), threatened or unique species (Knight et al. 2007), or to target areas of high human pressure (Leroux et al. 2010).

These analysis assumes that protected areas are successful if located in areas of high biological value (Pressey et al. 1993), or to identify gaps in the present network of protected areas (Scott et al. 1993). This thinking has formed the backbone of ‘Systematic Conservation Planning’ which involves strategies to select and prioritize areas for conservation based on pre-set criteria (Margules & Pressey 2000). Increasingly, the importance of limitations in resources have been appreciated, and methods to include cost and opportunities are being included in planning where and how to allocate land for protection (Wilson et al. 2006; Wilson et al. 2011).

Including and integrating multiple different factors in the prioritization of areas to protect often involves combining data and making decision across factors that are not easily comparable (Williams & Araújo 2002). Several methods for multi-criteria analyses has been developed which aim to balance the importance of biological and socio-economic factors with opportunities for selecting the most appropriate areas to protect (Araújo & Williams 2000; Araujo & Williams 2001; Williams & Araújo 2002). These methods are based on having set objectives or multiple outcomes against which to evaluate the present system. Whether the final selection is appropriate or sufficient, depend on idiosyncratic decisions on what is most important. Thus, deciding whether the present coverage of species in protected areas, or the number of red listed species falling outside is too high or too low has no one answer.

Reserve selection approaches assume that protected areas, once established, improve conditions and are not only ‘paper parks’ without effective management on the ground (Dudley & Stolton 1999). However both habitats (Joppa & Pfaff 2011) and species (Craigie et al. 2010; Laurance et al. 2012) are generally declining inside protected areas, suggesting that designation and gazzetment alone isn’t enough to secure that protected areas are an “effective means, to achieve the long term conservation of nature”. This recognition has also led to suggestions of replacing underperforming protected areas to optimize the allocation of land and money (Fuller et al. 2010; Mascia & Pailler 2011).

Management Effectiveness Protected Areas Management Effectiveness (PAME) schemes have served as a way to evaluate protected area performance at a more functional level than only species representation. PAME evaluations uses a standardized approach that can be compared across protected areas capturing key information on management of the individual parks (Hockings & Phillips 1999). More than 50 methodologies for assessing management effectiveness has been developed (Leverington et al. 2010), however most of the methodologies are used in only a small number of project or are adaptations of more

21 synopsis standardized approached tailored to local conditions (Hockings 2003). More than 10,000 assessments of PAME have been conducted across more than 6,700 sites globally, and these evaluations are a standard feature of projects supported by many of the large conservation organizations (e.g. the World Bank, the Global Environmental Facility (GEF), and the World Wildlife Fund for Nature (WWF)) (Coad et al. 2013a). The IUCN has developed a common framework to guide management effectiveness evaluations and help harmonize efforts around the world (Hockings et al. 2004) and tools to translate and compare efforts across schemes has also been developed (Leverington et al. 2010). Two of the most common and widely used methodologies are Rapid Assessment and Prioritization of Protected Area Management (RAPPAM) developed by WWF (Ervin 2003) and the Management Effectiveness Tracking Tool (METT), developed by WWF and the World Bank (Stolton et al. 2007). Both identify threats and pressures as well as report on the quality and implementation of management activities. However RAPPAM is targeted entire protected area or project systems, whereas METT is targeted specific protected areas, reporting on the implementation of more than 30 specific management parameters (Table S1, appendix III).

The METT evaluation is divided in six primary reporting elements i) context, ii) planning, iii) input, iv) process, v) outputs, and vi) outcomes (Hockings 2003; Leverington et al. 2010) with the majority of the evaluation focusing on planning, inputs and process. Management effectiveness evaluations have been criticized for not capturing outcomes related to the motivation and objectives of the protected areas and mainly capture those inputs expected to relate to protected area success (Mascia et al. accepted). Nevertheless PAME assessments have become one benchmark used to evaluate the CBD Aichi target 11 which calls for all countries to:

“expand and institutionalize management effectiveness assessments to work towards assessing 60 per cent of the total area of protected areas by 2015 using various national and regional tools and report the results into the global database on management effectiveness” (Convention on Biological Diversity 2010). and successful management is expected to be a prerequisite for successful protection of biodiversity and natural habitats.

Today around 29% of the global protected area has been assessed with Africa and Latin America having more than 40% assessed (Coad et al. 2013a). However analysis of the quality of the management in sites assessed shows that more than 40% of the evaluated protected areas showed major deficiencies in their management operations (Leverington et al. 2010). Further management alone is not guarantee that protected areas are actually successful in preserving biodiversity within their boundaries.

22 synopsis

Change in state – outcomes Many analyses have investigated how a conservation outcome changes within protected areas. These studies traditionally compared outcomes in protected areas to areas outside (buffer analysis), or between types of protected area, but without including or analyzing drivers and interventions which affect the observed outcomes (Ferraro 2004; Geldmann et al. 2013). The majority of such studies have used remote-sensed data to track changes in conservation outcomes. Remote-sensing techniques are best at detecting changes in habitat extent and are thus largely limited to forest cover (forest / no forest), and are less effective in capturing seasonality or subtle changes which can be of more importance in non-forested areas (Nagendra & Rocchini 2008). Three different types of buffer analysis have been used: i) compare to areas of a radius immediately outside the protected area (e.g. Nepstad et al. 2006), ii) compare to a comparable area outside the protected area (e.g. Tabor et al. 2010), and iii) compare to all areas outside the protected area within a given region or country (e.g. Clark et al. 2013). With some exceptions (Liu et al. 2001; Mapaure & Campbell 2002; Ingram & Dawson 2005), these studies generally find deforestation to be lower inside than outside (Geldmann et al. 2013). Studies comparing different types of protected areas show areas with lower IUCN management categories (more restrictive management prescriptions) to perform better than higher management categories (Scharlemann et al. 2010; Joppa & Pfaff 2011), while studies investigating differences between indigenous reserves or community managed reserved against other protected areas are more divided (Bowler et al. 2010; Geldmann et al. 2013).

For remote-sensed data, two different approaches have been used to include the effect of contextual drivers: i) regression and ii) matching analysis (Geldmann et al. 2013). These methods are able to include information on topography, distance to human settlements and population densities to explain differences between protected areas and inside vs. outside, thus using spatially explicit measures to explain some of the variation in effectiveness.

Across all methods and across the world results show that deforestation is lower inside protected areas than outside, though protected areas locations is an important component of their effectiveness (Joppa et al. 2008; Geldmann et al. 2013) accounting for as much as half of the observed reduced deforestation (Joppa & Pfaff 2011).

Fewer studies have been conducted on other biodiversity matrixes than deforestation; such as species richness or changes in species populations in protected areas. Craigie et al. (2010) found animal populations to decline by as much as 65% since 1970 across 89 protected areas in Africa with significant regional differences, where Western African protected areas experienced the worst declines of around 85%. A number of case studies have also found populations to decline in protected areas but less so than outside, however no general pattern have emerged (Geldmann et al. 2013). Using proxies for effective management studies such as protection status (Taylor et al. 2011), type of reserves (Tranquilli et al. 2012) or large scale policy decisions (Maiorano et al. 2007)

23 synopsis results are suggesting that protected areas positively influence species persistence. But none of these directly link the interventions, governance structures, or practices setting protected land aside from land not under formal protection by linking the input to the observed outcomes.

Linking input and outcomes Evaluating the effectiveness of protected areas, linking inputs and outcomes is not easy for several reasons. First; evaluating protected area performance requires a reference or counterfactual scenario against which to evaluate the observed changes inside protected areas (Ferraro 2009). The BACI principle (Before/After and Control/Intervention) serves as an experimental approach to evaluate whether changes inside protected areas are different from before the designation of the area or compared to a suitable reference area outside. Second; even with a counterfactual scenario, setting a baseline is not trivial matter. Whether to accept declines inside protected areas as long as these are smaller than outside or before, or whether to only accept no losses as a measure of success is an idiosyncratic choice that greatly affects the evaluation of the protected area performance. Third; biodiversity inside protected areas will change over time regardless of protection, and while some species will benefit from successful protection others might decline from changes in habitat structure (Fellers & Drost 1993), predator prey interactions (Tambling & Toit 2005; Sinclair et al. 2007), or changes in abiotic factors and diseases (Woinarski et al. 2010; Scholte 2011). Forth, and perhaps most importantly; the requirements for data to evaluate the effectiveness of protected area performance are much greater than for evaluating their location. Effectiveness evaluations, requires not only data on both input and outcomes, but also temporal data on how input affects changes in outcomes compared to a counterfactual scenario. This makes the evaluation of protected area performance extremely challenging, especially at large scales. However establishing and examining this link is highly relevant and can help guide conservation funds to be used more effectively by focusing on implementation of cost-effective solutions.

PAME evaluations can serve as a valuable data source for linking observation of change in biodiversity to the responses and interventions designed to reduce the loss of biodiversity and preserve habitats. However the degree to which PAME evaluations are able to capture management elements which reflect the performance of protected areas is debated. Further thought these evaluations are often mandatory for protected areas receiving funds reviews have shown that compliance is often very low (Global Environmental Facility, internal document 2013).

Finding a suitable outcome matrix, which should i) capture a relevant index of biodiversity (i.e. information on native and naturally occurring species inside the protected or extent or quality of natural habitats), ii) capture changes in the biodiversity matrix over time, iii) be spatially and temporally linked to a specific protected areas, and iv) be of sufficient quality to be a reliable estimator of biological change inside the

24 synopsis

protected area is no trivial matter. Ideally the matrix should also have a counterfactual comparator from before the establishment of the protected areas or from a comparable area outside the protected area (i.e. BACI).

Input

Management Resources actions Legislation

Protected Area Figure 4. The conceptual model for evaluating Outcomes protected area effectiveness in delivering improved conditions for state or STATE PRESSURE reductions in pressure. Improved responses of Reduced threats to Protected areas are dependent conservation target conservation target(s) on inputs in the form of legislative frameworks, management actions, and Context resources to be effective. At Soci-economics Climate change Values the same time protected area do not work independently of Governance Pollution Ethics the context of their establishment which can Ecological Evolutionary Landscape work to improve or reduce processes processes Topography their effectiveness.

While remotely sensed habitat studies have been able to show that protected area are effective in reducing deforestation as well as describe certain factors often associated with increased or decreased performance these studies lack information on on-ground activities which affect the observed patterns (Geldmann et al. 2013). For species, the evidence of protected area effectiveness is even more elusive and very few studies have been able to establish whether changes observed in species populations are linked to protection or other factors (Geldmann et al. 2013). When trying to understand what elements of protection helps protect biodiversity the evidence is even sparser. Two global analysis using questionnaires have shown that protected areas are effective as well as what elements of management contributed to the observed effectiveness (Bruner et al. 2001; Laurance et al. 2012). Such studies point to enforcement, boundary demarcation, and direct compensation to local communities as effectively improving reserves (Bruner et al. 2001) as well as pollution, hunting, mining, and burning both inside and outside the reserves as the major threats causing declines (Laurance et al. 2012). But these methods have been criticized for not using independent data to measure the outcome variable making them vulnerable to the critique for their non- independence in data collection on input and outcomes (Vanclay 2001). However with the lack of good independent data for both inputs and outcomes, these analyses can be valuable in indicating what and where to target efforts and research. Still, understanding if and how protected areas are effective using quantitative data will potentially present a

25 synopsis strong case for how to manage these to ensure their maximum contribution to preserving biodiversity.

Objectives The aim of this doctorate-thesis has been to test the hypotheses one that protected areas are an effective conservation instrument for protecting biodiversity and reducing human threats as a consequence of their establishment, and two that increased management investments into protected areas improve their performance. I also discuss whether the rate of change in either state or pressure is a useful matrix to evaluate protected area performance. This requires understanding protected areas in an experimental framework, asking whether protected areas have made a difference compared to a counterfactual scenario (i.e. before their establishment or outside the protected area (Ferraro 2009)), and what elements of protection are responsible for the observed change.

I have addressed this question using large-scale datasets to look for effects that transcend case-studies of individual protected areas. My project therefore builds on data collected by numerous conservation organizations; especially the World Database on Protected Areas (WDPA) (World Conservation Monitoring Centre 2013), the Living Planet Database (LPD) (Loh et al. 2005) and the METT Database (Coad et al. 2013b). For the first two, I primarily used the existing databases, while for the METT database, I was heavily involved in the procurement of new data from original sources, helped develop and construct the METT database, and assisted in the training and supervision of people entering new data into the database, as well as entered data myself for a number of months.

Besides this introduction, this thesis consists of six research chapters which cover different aspects of protected area performance.

Chapter I is a systematic review on the evidence of protected areas delivering conservation outcomes, particularly focusing on populations and habitat changes. This chapter outlines many of the concepts used throughout the thesis on the effectiveness and performance of protected areas. Chapter I is a shorter version of the original systematic review which is available at the Collaboration for Environmental Evidence website: www.environmentalevidence.org/SR10007.html. Chapter II examines and discusses commonalities and complementarities of the different management and evaluations approaches used to evaluate in conservation efforts, and highlights the roles of PAME in the context of conservation evaluations. Chapter III analyses how management effectiveness changes over time in protected areas where multiple evaluations have been conducted through a project period. Chapter IV looks at the role of evaluations of management effectiveness in achieving the biodiversity Aichi targets, which call for at least 30% and 60% of all protected areas to adequately managed by 2020. Chapter V maps changes in human pressure between 1995 and 2010 and analyses the extent to which protected areas have helped reduce human pressure and whether

26 synopsis

IUCN management categories are good proxies for their effectiveness. Chapter VI examines how specific management investments help explain patterns of changing animal populations inside protected areas by adding detailed information on the quality of management in protected areas using the METT.

Main findings and perspectives The evidence base This thesis is part of a larger research agenda evaluating protected areas performance by examining changes in outcomes (Kapos et al. 2008; Kapos et al. 2009) compared to a counterfactual scenario (Joppa et al. 2008; Ferraro 2009). A large body of work has already addressed this topic and particularly for habitat extent, has shown that protected areas are more effective than no protection (Geldmann et al. 2013). These analysis focusing on remotely sensed outcome matrices (such as deforestation (Joppa & Pfaff 2011) or occurrences of fire (Nelson & Chomitz 2011)) have primarily been able to correlate changes in forest-cover to drivers available from similar large scale datasets; such as human population or topography. However, by systematically reviewing the existing peer-reviewed and grey literature we have documented a knowledge gap resulting in limited evidence from other types of habitats as well as more direct measures of biodiversity (e.g. changes in species richness, species ranges, community shifts, population changes etc.). Further, information on direct management input is still erratic (especially in large scale analysis) impeding a better understanding of what makes protected areas effective. For habitats, remotely sensed data are constantly improving, increasing the range of habitats which can be classified and tracked over time (Nagendra & Rocchini 2008). Methods combining remotely sensed data, ground proofing and expert opinions are even extending analysis into savannahs (Riggio et al. 2012). For data at the species level, the high cost involved in collecting these and the investment in time (both work-hours and years to obtain useful data) have limited the availability.

Our systematic review (Geldmann et al. 2013), using correlations between input and outcomes, helped identify a lack of a clear terminology and methodology for evaluating protected areas effectiveness. In our review we found countless studies using the same terms to mean different things or studies excluding key variables (such as biotic and abiotic interactions, the effects of diseases, civil conflicts etc.), forcing us to reject these from our analysis. Many of these studies suggested protected area success and indicated elements of protection which correlated to better performing reserves, but without the experimental framework necessary to investigate the actual contribution of the reserve or its management (e.g. Jachmann (2008). This is potentially a serious problem. One, because the lack of consistent terminology precludes effective collation of existing studies and evidence. Two, because a lack of a strong and coherent methodology prevents an easy comparison across studies. These caveats can potentially lead to information being undetectable for the research community, either because it does not

27 synopsis appear in search engines, or because it is not comparable to other studies. In the medical sciences initiatives such as the Cochrane methodology have illustrated the power of standardizing experiments as well as provided guidelines for doing so (Cochrane 1972). Conservation Initiatives such as the Collaboration for Environmental Evidence (http://www.environmentalevidence.org/) are trying to mimic such approaches, and could potentially be important not only for the studies they provide, but also for the methodological changes they may inspire in study designs (Pullin & Knight 2009). Likewise, initiatives such as ConservationEvidence.com (http://www.conservationevidence.com/), which seeks to collate and summarize literature on conservation interventions for use by conservationist and researches, may help. Enterprises such as these which are targeted at researchers and practitioners alike are sorely needed to build the evidence base and improve the methods used to construct case-experiments. At present most attempts to systematically build prescriptive solutions based on all available evidence fail, which was documented in a recent review of systematic reviews in conservation. This showed that both the uptake of systematic reviews to inform conservation as well as the quality of literature to support systematic collation of data is limited (Cook et al. 2013).

Correlations and causality A main objective of this PhD has been to understand how input and intervention affect protected area performance. We found a significant and positive correlation between the increase in wildlife population size and country-level human development (i.e. Human Development Index) in protected areas with implemented management plans, while the opposite is the case in protected areas without management plans. These results point towards a complicated relationship between the socio-economic context of protected areas and the activities conducted to manage them. In collaboration with the IUCN’s Species Survival Commission (SSC) and World Commission on Protected Areas (WCPA) task force on biodiversity and protected areas, we are working on an analysis of more than 1,900 time-series across more than 500 protected areas globally to examine drivers of protected area performance (Barnes et al. in prep). Here we found similar dependencies of biodiversity on socio-economic factors, suggesting that these patterns are more general. Our global map of change in human pressure on earth shows that between 1995 and 2010 most land areas have experienced increased human impact, with disproportionally larger impacts in areas already under threat. Likewise, we show that protected areas have also experienced increased in human pressure despite them being set up to achieve the opposite. However these studies remain pattern-driven and more detailed knowledge is needed to understand the underlying causal drivers. Thus, there is still a great need to better understand and detangle the causal link between protection, management interventions and the observed changes in protected areas.

The analyses presented here are correlative, illuminating patterns which can help point to causal links between protected area effectiveness and the preservation of biodiversity. However, understanding the mechanisms driving the observed patterns is essential to

28 synopsis fully understand how and why protected areas are effective (Caro et al. 2009). Some of the challenges can undoubtedly be overcome by better and more sophisticated analysis while others (and most) require better, rather than more, data. For remote sensing analyses we need to understand how non-forest protected areas perform– as well as how changes in habitat quality affect biodiversity (i.e. the empty forest syndrome; Redford 1992). Evidence suggests that biodiversity values are lower in secondary forest compared to primary forest (Gibson et al. 2011), but how management and protected area performance affects these patterns is more difficult to detangle. For more direct measures of biodiversity, such as changes in species populations, there is a need to systematically collect and collate data on management, capacity and monetary resources in protected areas, in sync with biodiversity data collection. This will require methods to dissect often complicated multi-lateral funding schemes spanning administrative boundaries, and governmental and NGO jurisdictions. James et al. (1999) estimated the global budgets and staffing for protected areas. However, this information is outdated, inadequate, and is not informative about specific protected areas. A very recent study has assembled a global database on conservation spending (Waldron et al. 2013) which could potentially be a great step forward. But these data are scaled at the national level and do not discriminate between conservation activities (e.g. protected areas, cleaning of pollution etc.). Thus, we still need better data on resources flowing into the protected areas, and our analyses suggest that evaluations of management effectiveness, alone, seem insufficient for the purpose.

Perhaps more important than the lack of good data on resources and conservation input, is a better inclusion of ecological processes as well as stochasticity in helping to explain the observed patterns of change in a conservation context. These processes are often of critical importance to the observed changes in the few studies where they are actually measured (e.g. Mduma et al. 1999; Sinclair et al. 2007; Adams et al. 2008; Western et al. 2009), but are difficult to include in large scale studies (Craigie et al. 2010; Scholte 2011; Taylor et al. 2011). The integration of ecological theory in building large scale models would allow for a better understanding of the observed patterns and may help explain the large variation between parks and regions which I also observed in my analyses. However these steps are not easy. Additional studies using the LPD data are trying to understand what shapes the specific trajectory of animal populations (Di Fonzo et al. 2013), but linking these observation to large scale analyses of protected area effectiveness has still not been done.

These analyses illustrate the challenges in combining data collected for under a variety of protocols to answer different questions. Thus, while my research has taken advantage of the largest global databases on biodiversity change and management input, the varying quality and independence in the data collection have severely hampered the detection of strong patterns. Thus, while I have taken a different and more quantitative approach compared to other global analyses (i.e. Bruner et al. 2001; Laurance et al. 2012), I have not been able to get much closer to answering fundamental questions on

29 synopsis how management helps shape protected areas – only approached it from a different angle.

Untangling the causal connection between observed changes in biodiversity and conservation efforts will undoubtedly require data and methods suitable for both local scale and larger correlative studies, as well as include both quantitative and qualitative approaches. Well-designed case-studies can help improve our understanding, as long as these are designed so that the setup allows for generating results relevant outside the individual studies. Similarly, large-scale studies combining data from various sources are often successful in pointing to general patterns which merit further investigation, as well as help to indicate where the most important knowledge gaps are. Thus, if we are to live up to the global commitment of stopping the loss of biodiversity by 2020, we must utilize both local- and large-scale studies as well as ensuring an exchange of methods, ideas, concepts, approaches, and philosophies between them.

References Abell, R., M. L. Thieme, C. Revenga, M. Bryer, M. Kottelat, N. Bogutskaya, B. Coad, N. Mandrak, S. C. Balderas, W. Bussing, M. L. J. Stiassny, P. Skelton, G. R. Allen, P. Unmack, A. Naseka, R. Ng, N. Sindorf, J. Robertson, E. Armijo, J. V. Higgins, T. J. Heibel, E. Wikramanayake, D. Olson, H. L. Lopez, R. E. Reis, J. G. Lundberg, M. H. S. Perez, and P. Petry. (2008). Freshwater ecoregions of the world: A new map of biogeographic units for freshwater biodiversity conservation. Bioscience 58:403-414. Adams, L. G., R. O. Stephenson, B. W. Dale, R. T. Ahgook, and D. J. Demma. (2008). Population dynamics and harvest characteristics of wolves in the Central Brooks Range, Alaska. Wildlife Monographs:1-25. Araujo, M. B., and P. H. Williams. (2001). The bias of complementarity hotspots toward marginal populations. Conservation Biology 15:1710-1720. Araújo, M. B., and P. H. Williams. (2000). Selecting areas for species persistence using occurrence data. Biological Conservation 96:331-345. Baldwin, R. F. (2010). Identifying Keystone Threats to Biological Diversity. Pages 17-32 in S. C. Trombulak, and R. F. Baldwin, editors. Landscape-scale Conservation Planning. Springer Netherlands. Balmford, A., A. Bruner, P. Cooper, R. Costanza, S. Farber, R. E. Green, M. Jenkins, P. Jefferiss, V. Jessamy, J. Madden, K. Munro, N. Myers, S. Naeem, J. Paavola, M. Rayment, S. Rosendo, J. Roughgarden, K. Trumper, and R. K. Turner. (2002). Economic Reasons for Conserving Wild Nature. Science 297:950-953. Balmford, A., P. Carey, V. Kapos, A. Manica, A. S. L. Rodrigues, J. P. W. Scharlemann, and R. E. Green. (2009). Capturing the Many Dimensions of Threat: Comment on Salafsky et al. Conservation Biology 23:482-487. Balmford, A., B. Fisher, R. E. Green, R. Naidoo, B. Strassburg, R. K. Turner, and A. S. L. Rodrigues. (2011). Bringing Ecosystem Services into the Real World: An Operational Framework for Assessing the Economic Consequences of Losing Wild Nature. Environmental & Resource Economics 48:161-175. Balter, M. (2013). Archaeologists Say the ‘Anthropocene’ Is Here—But It Began Long Ago. Science 340:261-262.

30 synopsis

Barnes, M., I. D. Craigie, L. Harrison, J. Geldmann, T. Brooks, N. D. Burgess, B. Collen, M. Hockings, S. Whitmee, and S. Woodley. (in prep). Key correlates of population trends for birds and mammals in terrestrial protected areas. Barnosky, A. D., P. L. Koch, R. S. Feranec, S. L. Wing, and A. B. Shabel. (2004). Assessing the causes of Late Pleistocene extinctions on the continents. Science 306:70-75. 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. Bateman, I. J., A. R. Harwood, G. M. Mace, R. T. Watson, D. J. Abson, B. Andrews, A. Binner, A. Crowe, B. H. Day, S. Dugdale, C. Fezzi, J. Foden, D. Hadley, R. Haines-Young, M. Hulme, A. Kontoleon, A. A. Lovett, P. Munday, U. Pascual, J. Paterson, G. Perino, A. Sen, G. Siriwardena, D. van Soest, and M. Termansen. (2013). Bringing Ecosystem Services into Economic Decision-Making: Land Use in the United Kingdom. Science 341:45-50. Bertzky, B., C. Corrigan, J. Kemsey, S. Kenney, C. Ravilious, C. Besancon, and N. D. Burgess. (2012). Protected Planet report: Tracking progress towards global targets for protected areas. IUCN and UNEP-WCMC, Gland, Switzerland and Cambridge, UK. Boitani, L., R. M. Cowling, H. T. Dublin, G. M. Mace, J. Parrish, H. P. Possingham, R. L. Pressey, C. Rondinini, and K. A. Wilson. (2008). Change the IUCN Protected Area Categories to Reflect Biodiversity Outcomes. PLoS Biology 6:e66. Bowler, D., L. Buyung-Ali, J. R. Healey, J. P. G. Jones, T. Knight, and A. S. Pullin. (2010). The Evidence Base for Community Forest Management as a Mechanism for Supplying Global Environmental Benefits and Improving Local Welfare. Environmental Evidence. Collaboration for Environmental Evidence. Bruner, A. G., R. E. Gullison, R. E. Rice, and G. A. B. da Fonseca. (2001). Effectiveness of Parks in Protecting Tropical Biodiversity. Science 291:125-128. Burgess, N., S. Mwakalila, S. Madoffe, T. Ricketts, N. Olwero, R. Swetnam, B. Mbilinyi, R. M. Marchant, F., S. White, P. Munishi, A. Marshall, R. Malimbwi, G. Jambiya, B. Fisher, G. Kajembe, S. Morse-Jones, K. Kulindwa, J. Green, and A. Balmford. (2009). Valuing the Arc: a programme to map and value ecosystem services in Tanzania. Mountain Research Initiative Newsletter 18:18-21. Butchart, S. H. M., M. Walpole, B. Collen, A. van Strien, J. P. W. Scharlemann, R. E. A. Almond, J. E. M. Baillie, B. Bomhard, C. Brown, J. Bruno, K. E. Carpenter, G. M. Carr, J. Chanson, A. M. Chenery, J. Csirke, N. C. Davidson, F. Dentener, M. Foster, A. Galli, J. N. Galloway, P. Genovesi, R. D. Gregory, M. Hockings, V. Kapos, J. F. Lamarque, F. Leverington, J. Loh, M. A. McGeoch, L. McRae, A. Minasyan, M. H. Morcillo, T. E. E. Oldfield, D. Pauly, S. Quader, C. Revenga, J. R. Sauer, B. Skolnik, D. Spear, D. Stanwell- Smith, S. N. Stuart, A. Symes, M. Tierney, T. D. Tyrrell, J. C. Vie, and R. Watson. (2010). Global Biodiversity: Indicators of Recent Declines. Science 328:1164-1168. Caro, T., T. A. Gardner, C. Stoner, E. Fitzherbert, and T. R. B. Davenport. (2009). Assessing the effectiveness of protected areas: paradoxes call for pluralism in evaluating conservation performance. Diversity and Distributions 15:178-182. Christensen, J. D. (2003). Auditing Conservation in an Age of Accountability. Conservation in Practice 4:12-18. Clark, N. E., E. H. Boakes, P. J. K. 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. Coad, L., F. leverington, N. D. burgess, I. C. Cuadros, J. Geldmann, T. R. Marthews, J. Mee, C. Nolte, S. Stoll-Kleemann, N. Vansteelant, C. Zamora, M. Zimsky, and M. Hockings. (2013a). Progress towards the CBD Protected Area Management Effectiveness Targets. Parks 19:13-24.

31 synopsis

Coad, L., F. Leverington, J. Geldmann, C. Nolte, and M. Hockings. (2013b). Management Effectiveness Tracking Tool, global database, University of Oxford and University of Queensland. Available from Cochrane, A. L. (1972). Effectiveness and Efficiency : Random Reflections on Health Services. Nuffield Provincial Hospitals Trust, London. Convention on Biological Diveristy. (2004). Programme of work on protected areas. http://www.biodiv.org/decisions/default.aspx?dec=VII/28 in Secretariat of the Convention on Biological Diversity, editor. United Nations Environment Programme. United Nations Environment Programme Convention on Biological Diversity. (1993). Convention on Biological Diversity. Secretary- General of the United Nations, Rio de Janeiro, Brazil. Convention on Biological Diversity. (2010). Strategic Plan for Biodiversity 2011-2020 - COP 10, decision X/2. Convention on Biological Diversity Available from http://www.cbd.int/decision/cop/?id=12268 Cook, C. N., H. P. Possingham, and R. A. Fuller. (2013). Contribution of Systematic Reviews to Management Decisions. Conservation Biology: 1-14 (early view). Craigie, I. D., J. E. M. Baillie, A. Balmford, C. Carbone, B. Collen, R. E. Green, and J. M. Hutton. (2010). Large mammal population declines in Africa's protected areas. Biological Conservation 143:2221-2228. Crutzen, P. J. (2002). Geology of mankind. Nature 415:23-23. Di Fonzo, M., B. Collen, and G. M. Mace. (2013). A new method for identifying rapid decline dynamics in wild vertebrate populations. Ecology and Evolution 3:2378-2391. Dudley, N. (2008). Guidelines for Applying Protected Area Management Categories. International Union for Conservation of Nature, Gland, Switzerland. Dudley, N., and S. Stolton. (1999). Conversion of “Paper Parks” to Effective Management – Developing a Target. IUCN, WWF, WCPA. Ellis, E. C., K. Goldweijk, K., S. Siebert, D. Lightman, and N. Ramankutty. (2010). Anthropogenic transformation of the biomes, 1700 to 2000. Global Ecology and Biogeography 19:589-606. Ellis, E. C., J. O. Kaplan, D. Q. Fuller, S. Vavrus, K. Klein Goldewijk, and P. H. Verburg. (2013). Used planet: A global history. Proceedings of the National Academy of Sciences. Ervin, J. (2003). WWF: Rapid Assessment and Prioritization of Protected Area Management (RAPPAM) Methodology. WWF, Gland, Switzerland. European Union. (1992). Council Directive 92/43/EEC on the Conservation of natural habitats and of wild fauna and flora. European Union. Fellers, G. M., and C. A. Drost. (1993). Disappearance of the cascades frog Rana cascadae at the southern end of its range, California, USA. Biological Conservation 65:177-181. Ferraro, P. J. (2004). Targeting conservation investments in heterogeneous landscapes: A distance-function approach and application to watershed management. American Journal of Agricultural Economics 86:905-918. Ferraro, P. J. (2009). Counterfactual thinking and impact evaluation in environmental policy. New Directions for Evaluation 2009:75-84. Ferraro, P. J., and S. K. Pattanayak. (2006). Money for nothing? A call for empirical evaluation of biodiversity conservation investments. Plos Biology 4:482-488. Fisher, B., and R. K. Turner. (2008). Ecosystem services: Classification for valuation. Biological Conservation 141:1167-1169.

32 synopsis

Food and Agricultural Organization of the United Nations. (2009). The State of World Fisheries and Aquaculture 2008. Food and Agricultural Organization of the United Nations, Rome, Italy. Fuller, R. A., E. McDonald-Madden, K. A. Wilson, J. Carwardine, H. S. Grantham, J. E. M. Watson, C. J. Klein, D. C. Green, and H. P. Possingham. (2010). Replacing underperforming protected areas achieves better conservation outcomes. Nature 466:365- 367. Gaston, K. J. (2010). Biodiversity. Pages 27-44 in N. S. Sodhi, and P. R. Ehrlich, editors. Conservation Biology for all. Oxford University Press, Oxford, UK. Geldmann, J., M. Barnes, L. Coad, I. D. Craigie, M. Hockings, and N. D. Burgess. (2013). Effectiveness of terrestrial protected areas in reducing habitat loss and population declines Biological Conservation 161:230-238. Gibson, L., T. M. Lee, L. P. Koh, B. W. Brook, T. A. Gardner, J. Barlow, C. A. Peres, C. J. A. Bradshaw, W. F. Laurance, T. E. Lovejoy, and N. S. Sodhi. (2011). Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478:378-381. Global Environmental Facility. (2013). Overview of tracking tool available fo GEF protected area projects - draft 25 July 2013. Page 7. Global Environmental Facility. Halpern, B. S., C. Longo, D. Hardy, K. L. McLeod, J. F. Samhouri, S. K. Katona, K. Kleisner, S. E. Lester, J. O'Leary, M. Ranelletti, A. A. Rosenberg, C. Scarborough, E. R. Selig, B. D. Best, D. R. Brumbaugh, F. S. Chapin, L. B. Crowder, K. L. Daly, S. C. Doney, C. Elfes, M. J. Fogarty, S. D. Gaines, K. I. Jacobsen, L. B. Karrer, H. M. Leslie, E. Neeley, D. Pauly, S. Polasky, B. Ris, K. St Martin, G. S. Stone, U. R. Sumaila, and D. Zeller. (2012). An index to assess the health and benefits of the global ocean. Nature 488:615-620. Hamilton, A. J., Y. Basset, K.-K. Benke, P.-S. Grimbacher, S.-E. Miller, V. Novotn+½, G. A. Samuelson, N.-E. Stork, G.-D. Weiblen, and J.-D. Yen. (2010). Quantifying Uncertainty in Estimation of Tropical Arthropod Species Richness. The American Naturalist 176:90-95. Hockings, M. (2003). Systems for assessing the effectiveness of management in protected areas. BioScience 53:823-832. Hockings, M., and A. Phillips. (1999). How well are we doing? - some thoughts on the effectiveness of protected areas Parks 9:5-14. Hockings, M., S. U. E. Stolton, and N. Dudley. (2004). Management Effectiveness: Assessing Management of Protected Areas? Journal of Environmental Policy & Planning 6:157-174. Holt, B. G., J.-P. Lessard, M. K. Borregaard, S. A. Fritz, M. B. Araújo, D. Dimitrov, P.-H. Fabre, C. H. Graham, G. R. Graves, K. A. Jønsson, D. Nogués-Bravo, Z. Wang, R. J. Whittaker, J. Fjeldså, and C. Rahbek. (2013). An Update of Wallace’s Zoogeographic Regions of the World. Science 339:74-78. Ingram, J. C., and T. P. Dawson. (2005). Technical Note: Inter-annual analysis of deforestation hotspots in Madagascar from high temporal resolution satellite observations. International Journal of Remote Sensing 26:1447 - 1461. Jachmann, H. (2008). Monitoring law-enforcement performance in nine protected areas in Ghana. Biological Conservation 141:89-99. Jackson, J. B. C., M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner. (2001). Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629-638. James, A. N., M. J. B. Green, and J. R. Paine. (1999). A Global Review of Protected Area Budgets and Staffing. WCMC Biodiversity Series. World Conservation Monitoring Centre, Cambridge, England.

33 synopsis

Jepson, P. (2005). Governance and accountability of environmental NGOs. Environmental Science & Policy 8:515-524. Joppa, L. N., S. R. Loarie, and S. L. Pimm. (2008). On the protection of protected areas. Proceedings of the National Academy of Sciences 105:6673-6678. Joppa, L. N., and A. Pfaff. (2011). Global protected area impacts. Proceedings of the Royal Society B-Biological Sciences 278:1633-1638. Kapos, V., A. Balmford, R. Aveling, P. Bubb, P. Carey, A. Entwistle, J. Hopkins, T. Mulliken, R. Safford, A. Stattersfield, M. Walpole, and A. Manica. (2009). Outcomes, not implementation, predict conservation success. Oryx 43:336-342. Kapos, V., A. Balmford, R. Aveling, P. Bupp, P. Carey, A. Entwistle, J. Hopkins, T. Muliken, R. Safford, A. Stattersfield, M. Walpole, and A. Manica. (2008). Calibrating conservation: new tools for measuring success. Conservation Letters 1:155-164. Knight, A. T., R. J. Smith, R. M. Cowling, P. G. Desmet, D. P. Faith, S. Ferrier, C. M. Gelderblom, H. Grantham, A. T. Lombard, K. Maze, J. L. Nel, P. J. D, G. Q. K. Pence, H. P. Possingham, B. Reyers, M. Rouget, D. Roux, and K. A. Wilson. (2007). Improving the Key Biodiversity Areas Approach for Effective Conservation Planning. BioScience 57:256-261. Laurance, W. F., D. Carolina Useche, J. Rendeiro, M. Kalka, C. J. A. Bradshaw, S. P. Sloan, S. G. Laurance, M. Campbell, K. Abernethy, P. Alvarez, V. Arroyo-Rodriguez, P. Ashton, J. Benitez-Malvido, A. Blom, K. S. Bobo, C. H. Cannon, M. Cao, R. Carroll, C. Chapman, R. Coates, M. Cords, F. Danielsen, B. De Dijn, E. Dinerstein, M. A. Donnelly, D. Edwards, F. Edwards, N. Farwig, P. Fashing, P.-M. Forget, M. Foster, G. Gale, D. Harris, R. Harrison, J. Hart, S. Karpanty, W. John Kress, J. Krishnaswamy, W. Logsdon, J. Lovett, W. Magnusson, F. Maisels, A. R. Marshall, D. McClearn, D. Mudappa, M. R. Nielsen, R. Pearson, N. Pitman, J. van der Ploeg, A. Plumptre, J. Poulsen, M. Quesada, H. Rainey, D. Robinson, C. Roetgers, F. Rovero, F. Scatena, C. Schulze, D. Sheil, T. Struhsaker, J. Terborgh, D. Thomas, R. Timm, J. Nicolas Urbina-Cardona, K. Vasudevan, S. Joseph Wright, J. Carlos Arias-G, L. Arroyo, M. Ashton, P. Auzel, D. Babaasa, F. Babweteera, P. Baker, O. Banki, M. Bass, I. Bila-Isia, S. Blake, W. Brockelman, N. Brokaw, C. A. Bruhl, S. Bunyavejchewin, J.-T. Chao, J. Chave, R. Chellam, C. J. Clark, J. Clavijo, R. Congdon, R. Corlett, H. S. Dattaraja, C. Dave, G. Davies, B. de Mello Beisiegel, R. d. Nazare Paes da Silva, A. Di Fiore, A. Diesmos, R. Dirzo, D. Doran-Sheehy, M. Eaton, L. Emmons, A. Estrada, C. Ewango, L. Fedigan, F. Feer, B. Fruth, J. Giacalone Willis, U. Goodale, S. Goodman, J. C. Guix, P. Guthiga, W. Haber, K. Hamer, I. Herbinger, J. Hill, Z. Huang, I. Fang Sun, K. Ickes, A. Itoh, N. Ivanauskas, B. Jackes, J. Janovec, D. Janzen, M. Jiangming, C. Jin, T. Jones, H. Justiniano, E. Kalko, A. Kasangaki, T. Killeen, H.-b. King, E. Klop, C. Knott, I. Kone, E. Kudavidanage, J. Lahoz da Silva Ribeiro, J. Lattke, R. Laval, R. Lawton, M. Leal, M. Leighton, M. Lentino, C. Leonel, J. Lindsell, L. Ling-Ling, K. Eduard Linsenmair, E. Losos, A. Lugo, J. Lwanga, A. L. Mack, M. Martins, W. Scott McGraw, R. McNab, L. Montag, J. Myers Thompson, J. Nabe-Nielsen, M. Nakagawa, S. Nepal, M. Norconk, V. Novotny, S. O'Donnell, M. Opiang, P. Ouboter, K. Parker, N. Parthasarathy, K. Pisciotta, D. Prawiradilaga, C. Pringle, S. Rajathurai, U. Reichard, G. Reinartz, K. Renton, G. Reynolds, V. Reynolds, E. Riley, M.-O. Rodel, J. Rothman, P. Round, S. Sakai, T. Sanaiotti, T. Savini, G. Schaab, J. Seidensticker, A. Siaka, M. R. Silman, T. B. Smith, S. S. d. Almeida, N. Sodhi, C. Stanford, K. Stewart, E. Stokes, K. E. Stoner, R. Sukumar, M. Surbeck, M. Tobler, T. Tscharntke, A. Turkalo, G. Umapathy, M. van Weerd, J. Vega Rivera, M. Venkataraman, L. Venn, C. Verea, C. Volkmer de Castilho, M. Waltert, B. Wang, D. Watts, W. Weber, P. West, D. Whitacre, K. Whitney, D. Wilkie, S. Williams, D. D. Wright, P. Wright, L. Xiankai, P. Yonzon, and F. Zamzani. (2012). Averting biodiversity collapse in tropical forest protected areas. Nature 489:290-294.

34 synopsis

Leroux, S. J., M. A. Krawchuk, F. Schmiegelow, S. G. Cumming, K. Lisgo, L. G. Anderson, and M. Petkova. (2010). Global protected areas and IUCN designations: Do the categories match the conditions? Biological Conservation 143:609-616. Leverington, F., K. L. Costa, H. Pavese, A. Lisle, and M. Hockings. (2010). A Global Analysis of Protected Area Management Effectiveness. Environmental Management 46:685-698. Liu, J., M. Linderman, Z. Ouyang, L. An, J. Yang, and H. Zhang. (2001). Ecological Degradation in Protected Areas: The Case of Wolong Nature Reserve for Giant Pandas. Science 292:98-101. Loh, J., R. E. Green, T. Ricketts, J. Lamoreux, M. Jenkins, V. Kapos, and J. Randers. (2005). The Living Planet Index: using species population time series to track trends in biodiversity. Philosophical Transactions of the Royal Society B: Biological Sciences 360:289-295. Lorenzen, E. D., D. Nogues-Bravo, L. Orlando, J. Weinstock, J. Binladen, K. A. Marske, A. Ugan, M. K. Borregaard, M. T. P. Gilbert, R. Nielsen, S. Y. W. Ho, T. Goebel, K. E. Graf, D. Byers, J. T. Stenderup, M. Rasmussen, P. F. Campos, J. A. Leonard, K. P. Koepfli, D. Froese, G. Zazula, T. W. Stafford, K. Aaris-Sorensen, P. Batra, A. M. Haywood, J. S. Singarayer, P. J. Valdes, G. Boeskorov, J. A. Burns, S. P. Davydov, J. Haile, D. L. Jenkins, P. Kosintsev, T. Kuznetsova, X. L. Lai, L. D. Martin, H. G. McDonald, D. Mol, M. Meldgaard, K. Munch, E. Stephan, M. Sablin, R. S. Sommer, T. Sipko, E. Scott, M. A. Suchard, A. Tikhonov, R. Willerslev, R. K. Wayne, A. Cooper, M. Hofreiter, A. Sher, B. Shapiro, C. Rahbek, and E. Willerslev. (2011). Species-specific responses of Late Quaternary megafauna to climate and humans. Nature 479:359-U195. Mace, G. M., A. Balmford, L. Boitani, G. Cowlishaw, A. P. Dobson, D. P. Faith, K. J. Gaston, C. J. Humphries, R. I. Vane-Wright, P. H. Williams, J. H. Lawton, C. R. Margules, R. M. May, A. O. Nicholls, H. P. Possingham, C. Rahbek, and A. S. van Jaarsveld. (2000). It's time to work together and stop duplicating conservation efforts. Nature 405:393-393. Mace, G. M., K. Norris, and A. H. Fitter. (2012). Biodiversity and ecosystem services: a multilayered relationship. Trends in Ecology & Evolution 27:19-26. Maiorano, L., A. Falcucci, E. O. Garton, and L. Boitani. (2007). Contribution of the Natura 2000 network to biodiversity conservation in Italy. Conservation Biology 21:1433-1444. Mapaure, I., and B. Campbell. (2002). Changes in miombo woodland cover in and around Sengwa Wildlife Research Area, Zimbabwe, in relation to elephants and fire. African Journal of Ecology 40:212-219 Margules, C. R., and R. L. Pressey. (2000). Systematic conservation planning. Nature 405:243- 253. Mascia, M. B., and S. Pailler. (2011). Protected area downgrading, downsizing, and degazettement (PADDD) and its conservation implications. Conservation Letters 4:9-20. Mascia, M. B., S. Pailler, M. L. Thieme, A. Rowe, M. C. Bottrill, F. Danielsen, J. Geldmann, R. Naidoo, A. S. Pullin, and N. D. Burgess. (accepted). Commonalities and Complementarities among Approaches to Conservation Monitoring and Evaluation Biological Conservation. May, R. M. (2010). Tropical Arthropod Species, More or Less? Science 329:41-42. Mduma, S. A. R., A. R. E. Sinclair, and R. Hilborn. (1999). Food Regulates the Serengeti Wildebeest: A 40-Year Record. Journal of Animal Ecology 68:1101-1122. Millinium Ecosystem Assessment. (2005). Millinium Ecosystem Assessment: Ecosystems and Human Well-being: Biodiversity Synthesis. World Resources Institute, Washington, DC. Mora, C., D. P. Tittensor, S. Adl, A. G. B. Simpson, and B. Worm. (2011). How Many Species Are There on Earth and in the Ocean? PLoS Biol 9:e1001127.

35 synopsis

Myers, R. A., and B. Worm. (2003). Rapid worldwide depletion of predatory fish communities. Nature 423:280-283. Nagendra, H., and D. Rocchini. (2008). High resolution satellite imagery for tropical biodiversity studies: the devil is in the detail. Biodiversity and Conservation 17:3431-3442. Nelson, A., and K. M. Chomitz. (2011). Effectiveness of Strict vs. Multiple Use Protected Areas in Reducing Tropical Forest Fires: A Global Analysis Using Matching Methods. PLoS ONE 6:e22722. Nepstad, D., S. Schwartzman, B. Bamberger, M. Santilli, D. Ray, P. Schlesinger, P. Lefebvre, A. Alencar, E. Prinz, G. Fiske, and A. Rolla. (2006). Inhibition of Amazon Deforestation and Fire by Parks and Indigenous Lands. Conservation Biology 20:65-73. Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. V. N. Powell, E. C. Underwood, J. A. D'Amico, I. Itoua, H. E. Strand, J. C. Morrison, C. J. Loucks, T. F. Allnutt, T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Hedao, and K. R. Kassem. (2001). Terrestrial Ecoregions of the World: A New Map of Life on Earth. Bioscience 51:933-938. Organisation for Economic Co-operation and Development. (1993). OECD core set of indicators for environmental performance reviews. OECD, Paris, France. Perrings, C., A. Duraiappah, A. Larigauderie, and H. Mooney. (2011). The Biodiversity and Ecosystem Services Science-Policy Interface. Science 331:1139-1140. Pressey, R. L., C. J. Humphries, C. R. Margules, R. I. Vane-Wright, and P. H. Williams. (1993). Beyond opportunism: Key principles for systematic reserve selection. Trends in Ecology & Evolution 8:124-128. Pullin, A. S., and T. M. Knight. (2009). Doing more good than harm - Building an evidence- base for conservation and environmental management. Biological Conservation 142:931- 934. Redford, K. H. (1992). The Empty Forest. Bioscience 42:412-422. Riggio, J., A. Jacobson, L. Dollar, H. Bauer, M. Becker, A. Dickman, P. Funston, R. Groom, P. Henschel, H. Iongh, L. Lichtenfeld, and S. Pimm. (2012). The size of savannah Africa: a lion’s (Panthera leo) view. Biodiversity and Conservation:1-19. Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, R. M. Cowling, L. D. C. Fishpool, G. A. B. da Fonseca, K. J. Gaston, M. Hoffmann, J. S. Long, P. A. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W. Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. Watts, and X. Yan. (2004). Effectiveness of the global protected area network in representing species diversity. Nature 428:640-643. Salafsky, N., D. Salzer, A. J. Stattersfield, C. Hilton-Taylor, R. Neugarten, S. H. M. Butchart, B. E. N. Collen, N. Cox, L. L. Master, S. O'Connor, and D. Wilkie. (2008). A Standard Lexicon for Biodiversity Conservation: Unified Classifications of Threats and Actions. Conservation Biology 22:897-911. Scharlemann, J. P. W., V. Kapos, A. Campbell, I. Lysenko, N. D. Burgess, M. C. Hansen, H. K. Gibbs, B. Dickson, and L. Miles. (2010). Securing tropical forest carbon: the contribution of protected areas to REDD. Oryx 44:352-357. Scholte, P. (2011). Towards understanding large mammal population declines in Africa’s protected areas: A West-Central African perspective. Tropical Conservation Science 4:11. Scott, J. M., F. Davis, B. Csuti, R. Noss, B. Butterfield, C. Groves, H. Anderson, S. Caicco, F. D'Erchia, T. C. Edwards Jr., J. Ulliman, and R. G. Wright. (1993). Gap Analysis: A Geographic Approach to Protection of Biological Diversity. Wildlife Monographs 123:3- 41.

36 synopsis

Sinclair, A. R. E., S. A. R. Mduma, J. G. C. Hopcraft, J. M. Fryxell, R. Hilborn, and S. Thirgood. (2007). Long-term ecosystem dynamics in the Serengeti: Lessons for conservation. Conservation Biology 21:580-590. Spalding, M. D., H. E. Fox, G. R. Allen, N. Davidson, Z. A. FerdaÑA, M. A. X. Finlayson, B. S. Halpern, M. A. Jorge, A. L. Lombana, S. A. Lourie, K. D. Martin, E. McManus, J. Molnar, C. A. Recchia, and J. Robertson. (2007). Marine Ecoregions of the World: A Bioregionalization of Coastal and Shelf Areas. Bioscience 57:573-583. Sparks, T. H., S. H. M. Butchart, A. Balmford, L. Bennun, D. Stanwell-Smith, M. Walpole, N. R. Bates, B. Bomhard, G. M. Buchanan, A. M. Chenery, B. Collen, J. Csirke, R. J. Diaz, N. K. Dulvy, C. Fitzgerald, V. Kapos, P. Mayaux, M. Tierney, M. Waycott, L. Wood, and R. E. Green. (2011). Linked indicator sets for addressing biodiversity loss. Oryx 45:411-419. Steffen, W., P. J. Crutzen, and J. R. McNeill. (2007). The Anthropocene: Are humans now overwhelming the great forces of nature. Ambio 36:614-621. Steffen, W., J. Grinevald, P. Crutzen, and J. McNeill. (2011). The Anthropocene: conceptual and historical perspectives. Philosophical Transactions of the Royal Society a- Mathematical Physical and Engineering Sciences 369:842-867. Stolton, S., M. Hockings, N. Dudley, K. MacKinnon, T. Whitten, and L. F. (2007). Reporting Progress in Protected Areas A Site-Level Management Effectiveness Tracking Tool: second edition. World Bank/WWF Forest Alliance, Gland, Switzerland. Sutherland, W. J., A. S. Pullin, P. M. Dolman, and T. M. Knight. (2004). The need for evidence- based conservation. Trends in Ecology & Evolution 19:305-308. Tabor, K., N. D. Burgess, B. P. Mbilinyi, J. J. Kashaigili, and M. K. Steininger. (2010). Forest and woodland cover and change in coastal Tanzania and Kenya, 1990 to 2000. Journal of East African Natural History 99:19-45. Tallis, H., P. Kareiva, M. Marvier, and A. Chang. (2008). An ecosystem services framework to support both practical conservation and economic development. Proc.Natl Acad.Sci.U.S.A 105:9457-9464. Tallis, H., H. Mooney, S. Andelman, P. Balvanera, W. Cramer, D. Karp, S. Polasky, B. Reyers, T. Ricketts, S. Running, K. Thonicke, B. Tietjen, and A. Walz. (2012). A Global System for Monitoring Ecosystem Service Change. Bioscience 62:977-986. Tambling, C. J., and J. T. D. Toit. (2005). Modelling Wildebeest Population Dynamics: Implications of Predation and Harvesting in a Closed System. Journal of Applied Ecology 42:431-441. Taylor, M., P. Sattler, M. Evans, R. Fuller, J. Watson, and H. Possingham. (2011). What works for threatened species recovery? An empirical evaluation for Australia. Biodiversity and Conservation 20:767-777. Tranquilli, S., M. Abedi-Lartey, F. Amsini, L. Arranz, A. Asamoah, O. Babafemi, N. Barakabuye, G. Campbell, R. Chancellor, T. R. B. Davenport, A. Dunn, J. Dupain, C. Ellis, G. Etoga, T. Furuichi, S. Gatti, A. Ghiurghi, E. Greengrass, C. Hashimoto, J. Hart, I. Herbinger, T. C. Hicks, L. H. Holbech, B. Huijbregts, I. Imong, N. Kumpel, F. Maisels, P. Marshall, S. Nixon, E. Normand, L. Nziguyimpa, Z. Nzooh-Dogmo, D. T. Okon, A. Plumptre, A. Rundus, J. Sunderland-Groves, A. Todd, Y. Warren, R. Mundry, C. Boesch, and H. Kuehl. (2012). Lack of conservation effort rapidly increases African great ape extinction risk. Conservation Letters 5:48-55. Udvardy, M. D. F. (1975). A classification of the biogeographical provinces of the world. International Union of Conservation of Nature and Natural Resources, Morges, Switzerland. United Nations Environment Programme. (2012). Global Environmental Outlook 5 - Environment for the fututre we want. United Nations Environment Programme, Valletta, Malta.

37 synopsis

Vanclay, J. K. (2001). The Effectiveness of Parks. Science 293:1007. Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. (1997). Human domination of Earth's ecosystems. Science 277:494-499. Waldron, A., A. O. Mooers, D. C. Miller, N. Nibbelink, D. Redding, T. S. Kuhn, J. T. Roberts, and J. L. Gittleman. (2013). Targeting global conservation funding to limit immediate biodiversity declines. Proceedings of the National Academy of Sciences:PNAS early edition 1-5 pp. Wallace, A. R. (1876). The Geographical Distribution of Animals. Cambridge University Press, Cambridge, UK. Wallace, K. J. (2007). Classification of ecosystem services: Problems and solutions. Biological Conservation 139:235-246. Western, D., S. Russell, and I. Cuthill. (2009). The Status of Wildlife in Protected Areas Compared to Non-Protected Areas of Kenya. PLoS ONE 4:e6140. Williams, P., and M. Araújo. (2002). Apples, Oranges, and Probabilities: Integrating Multiple Factors into Biodiversity Conservation with Consistency. Environmental Modeling & Assessment 7:139-151. Wilson, K. A., M. C. Evans, M. Di Marco, D. C. Green, L. Boitani, H. P. Possingham, F. Chiozza, and C. Rondinini. (2011). Prioritizing conservation investments for mammal species globally. Philosophical Transactions of the Royal Society B-Biological Sciences 366:2670-2680. Wilson, K. A., M. F. McBride, M. Bode, and H. P. Possingham. (2006). Prioritizing global conservation efforts. Nature 440:337-340. Woinarski, J. C. Z., M. Armstrong, K. Brennan, A. Fisher, A. D. Griffiths, B. Hill, D. J. Milne, C. Palmer, S. Ward, M. Watson, S. Winderlich, and S. Young. (2010). Monitoring indicates rapid and severe decline of native small mammals in Kakadu National Park, northern Australia. Wildlife Research 37:116-126. World Conservation Monitoring Centre. (2013). World Database on Protected Areas. WCMC, Cambridge, United Kingdom. Available from http://protectedplanet.net/ Worm, B., E. B. Barbier, N. Beaumont, J. E. Duffy, C. Folke, B. S. Halpern, J. B. C. Jackson, H. K. Lotze, F. Micheli, S. R. Palumbi, E. Sala, K. A. Selkoe, J. J. Stachowicz, and R. Watson. (2006). Impacts of biodiversity loss on ocean ecosystem services. Science 314:787-790.

CHAPTER I Evidence of protected area effectiveness

Jonas Geldmann, Megan Barnes, Lauren Coad, Ian D. Craigie, Marc Hockings, and Neil D. Burgess

Published in Biological Conservation, 2013, volume 161

Selected as editor’s choice Top-10 most downloaded articles in Biological Conservation 2013

Supporting Information refers to Appendix I

40 chapter I

Biological Conservation 161 (2013) 230–238

Contents lists available at SciVerse ScienceDirect

Biological Conservation

journal homepage: www.elsevier.com/locate/biocon

Systematic review Effectiveness of terrestrial protected areas in reducing habitat loss and population declines ⇑ Jonas Geldmann a, , Megan Barnes b,c, Lauren Coad d, Ian D. Craigie e, Marc Hockings b, Neil D. Burgess a,f a Center for Macroecology, Evolution and Climate, Department of Biology, University of Copenhagen, Denmark b School of Geography, Planning and Environmental Management, University of Queensland, Australia c Environmental Decisions Group, Australia d Environmental Change Institute, School of Geography, University of Oxford, Oxford OX1 3QY, United Kingdom e ARC Centre of Excellence for Coral Reef Studies, James Cook University, Australia f UNEP, World Conservation Monitoring Centre, Cambridge, United Kingdom article info abstract

Article history: Protected Areas (PAs) are a critical tool for maintaining habitat integrity and species diversity, and now Received 27 June 2012 cover more than 12.7% of the planet’s land surface area. However, there is considerable debate on the Received in revised form 24 February 2013 extent to which PAs deliver conservation outcomes in terms of habitat and species protection. A system- Accepted 27 February 2013 atic review approach is applied to investigate the evidence from peer reviewed and grey literature on the Available online 3 May 2013 effectiveness of PAs focusing on two outcomes: (a) habitat cover and (b) species populations. We only include studies that causally link conservation inputs to outcomes against appropriate counterfactuals. Keywords: From 2599 publications we found 76 studies from 51 papers that evaluated impacts on habitat cover, Effectiveness and 42 studies from 35 papers on species populations. Three conclusions emerged: ﬁrst, there is good evi- Habitat loss Management dence that PAs have conserved forest habitat; second, evidence remains inconclusive that PAs have been Population trend effective at maintaining species populations, although more positive than negative results are reported in Protected area the literature; third, causal connections between management inputs and conservation outcomes in PAs Systematic review are rarely evaluated in the literature. Overall, available evidence suggests that PAs deliver positive outcomes, but there remains a limited evidence base, and weak understanding of the conditions under which PAs succeed or fail to deliver conservation outcomes. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction ecosystem services (Campos and Nepstad, 2006), or for cultural and social reasons (Coad et al., 2008). Understanding the conditions Protected Areas (PAs) have long been regarded as an important under which PAs deliver conservation benefits for habitats and spe- tool for maintaining habitat integrity and species diversity (Brooks cies is essential for policy makers, managers and conservation advo- et al., 2004; Butchart et al., 2010; Coad et al., 2008; Rodrigues et al., cates (Brooks et al., 2004; Kleiman et al., 2000; Margules and 2004), covering more than 12.7% of the planet’s land surface (Bert- Pressey, 2000). zky et al., 2012). However, there is considerable debate on the ex- The success of PAs has generally been evaluated using measures tent to which PAs deliver conservation outcomes in terms of such as the representativeness of PA networks in terms of their spe- habitat and species protection (Brooks et al., 2006; Ferraro and cies diversity, or coverage of endemic and threatened species Simpson, 2002; Meir et al., 2004). It has been suggested that many (Rodrigues et al., 2004), assuming that PAs provide effective protec- of the world’s PAs exist only as ‘paper parks’ (Dudley and Stolton, tion once established. Alternatively, by investigating management 1999), lacking effective management capacity, and unlikely to de- ‘inputs’ – e.g. whether PAs have management plans, boundaries, liver effective conservation (Joppa et al., 2008). staffing, and other management systems and processes (Jachmann, PAs are often treated as a single conservation strategy. However, 2008), assuming that increased levels of management equates to in reality they are established for a variety of reasons, with very dif- successful protection. However, these analyses are not able to de- ferent objectives and criteria for success. PAs have been set up for the scribe how conditions inside PAs change over time (Craigie et al., conservation of ecosystems and their constituent species (Dudley, 2010), or evaluate the effectiveness of protection, by combining 2008), protection of specific threatened species (Liu et al., 2001), measures of inputs and measures of outcomes in a temporal framework; thus measuring how biodiversity outcomes change over time in relation to protection or implementation of management actions. ⇑ Corresponding author. Tel.: +45 3523 1230; fax: +45 3532 2128. The objective of this paper is to use a ‘systematic review’ meth- E-mail address: [email protected] (J. Geldmann). odology (Pullin and Knight, 2009) to review the evidence that PAs

J. Geldmann et al. / Biological Conservation 161 (2013) 230–238 231 deliver a positive change in two conservation outcomes: (a) habitat Third, we extracted information on other biological and geo- cover and (b) species populations, i.e. the ability of PAs to maintain graphical variables, and study biases. These effects had not been or improve native habitat integrity, or native species populations, measured using appropriate counterfactuals, but were mentioned over time respectively. We further consider the impact of different in the publications as having potentially affected biodiversity PA management interventions, or characteristics, where measured, outcomes. on biodiversity outcomes. zWhere multiple publications evaluated the same site using the same data, sites were only included once to avoid double counting. 2. Methods However, for habitat studies, PA effectiveness was evaluated at different scales (i.e. globally, regionally, nationally or site-level). In this 2.1. Search strategy case both studies were included, as results for one level is not simply part of the result of another. Thus, the results presented at different To locate relevant literature, we searched 14 databases, eight levels contribute different information on PA effectiveness. specialist sources and 13 websites in English (Table S1). We identified a list of relevant search terms and used Boolean operators and multi term searches (Table S2). Anonymous reviewers ap- 3. Results praised the list of relevant search terms and the search strategy. The search was conducted between July and August 2010, covering Of the 2599 publications selected through the systematic search all publications available up to that point. For a full description of strategy, we found 51 publications on habitat cover and 35 publi- the search strategy, search-terms, and inclusion criteria see Geld- cations on species population trends that fulfilled the inclusion mann et al. (2012). criteria. Within 13 of the 51 habitat change publications there were 2.2. Study inclusion criteria multiple counterfactual scenarios. When separated these yielded a total of 76 studies. Three population trend publications covered Two main criteria were used to determine study inclusion. First, more than one evaluation of PA effectiveness, yielding 42 studies we reviewed whether the publication assessed conservation in total across the 35 publications. Detailed descriptions of the data interventions and biodiversity outcomes. We only included publi- extracted from individual studies are presented for habitats (Tables cations that measured the effectiveness of PAs targeting biodiver- S5 and S6) and populations (Tables 1 and S7). sity conservation of native habitats/species. We excluded publications that looked at changes in alien species, or species 3.1. Protecting habitats not expected to improve with successful protection. Second, we only included publications that used suitable coun- Of the 76 studies on the effectiveness of PAs in retaining habitat terfactuals (controls), following the BACI (before/after or control/ cover, four were global, 35 evaluated regional, national or subna- intervention) framework. Counterfactuals were defined as: (a) be- tional networks of PAs, and 34 evaluated five or fewer PAs. There fore vs. after: e.g. PA establishment/implementation or PA manage- was a strong bias in study location; 35 were from Latin America, ment intervention, or (b) control vs. intervention: e.g. PAs 18 from Africa, 16 from Asia, two from Oceania, and one each from compared to their immediate surroundings or to non-protected Europe, and North America. There was also a strong bias in habitat areas with similar characteristics. focus. Sixty-eight of the 76 studies (89%) investigated changes in forest cover only, 67 (88%) of which were for tropical forest. The 2.3. Study characterization and quality assessment remaining eight evaluated multiple land-use types of which all but one (Alodos et al., 2004) included forests. For publications where multiple PAs were examined against dif- To determine changes in habitat cover, 63 studies (83%) used ferent counterfactuals, such that the publication contained more satellite remote sensing techniques, three used aerial photos, and than one examination of PA effectiveness, we divided these based five used a combination of both. The remaining five used in situ on the type of counterfactual. All summaries and estimations of data collection, either estimation of disturbance across plots (Ble- impact are based on this subdivision of results from publications her et al., 2006; Liu et al., 2001; Tole, 2002), or interviews and that are henceforth referred to as: ‘‘studies’’. questionnaires (Bruner et al., 2001; Mwangi et al., 2010). To ana- For each study we first extracted detailed information on biodi- lyze PA effectiveness in protecting habitat, 36 of the 76 studies versity outcome variables. This included information on the meth- used buffer analyses (comparing changes in habitat cover inside ods used to measure habitat or species population change (i.e. PAs to a surrounding buffer), 21 compared to similar areas outside remote sensing, transect surveys, etc.), the rates of change, and the PAs, and 10 used matching estimator methods (Table S5). the units of measurement. For studies that did not report the rate Sixty-two of the 76 studies of habitat change (82%) found habitat of change, we noted the given direction of change (improving/no loss to be higher outside PAs than inside, nine studies found habitat effect/declining) compared to the counterfactual. For all studies loss to be higher inside PAs than outside, and five could not detect an of species populations we also estimated the fraction of species effect of protection (Tables 2 and S5). The three global studies were that did better inside the PA compared to the counterfactual, and generally in agreement, finding that PAs were effective in reducing also noted any reported trophic impacts (such as population habitat loss. DeFries et al. (2005) compared PAs to their buffer, and changes due to predator–prey dynamics). found rates of habitat loss for 198 PAs to be 2.6 times lower inside Second, we extracted information on PA management interven- compared to outside. Scharlemann et al. (2010) found that PAs lost tions and characteristics, as well as external drivers of habitat or about half as much carbon as forest outside PAs globally (ca. 2 times species change. We recorded either the given effect size of the var- lower than outside PAs), and the loss in Oceania, the Neotropics, and iable, or where this was not given, noted the direction of change in Tropical Asia to be higher outside PAs than inside. Joppa and Pfaff (improving/no effect/declining) compared to the counterfactual. (2011), found that rates of habitat loss in PAs were 1.08 times lower The management interventions and PA characteristics identified than the counterfactual. were then grouped into categories (with separate categories for In 52 of the 76 studies the results reported, we were able to cal- habitat and species) that were defined post hoc (details of catego- culate the ratio of the habitat change in the PA compared to their ries are provided in Tables S3 and S4). counterfactual (Table S5). Where PAs had lower habitat loss com- 42 232

Table 1 Detailed data from the 42 studies evaluating PA effectiveness for species populations.

Source Countries Protected area Monitoring Taxa Counterfactual BACI Background Effect of period trend PA Adams et al. (2008) USA Arctic NP 1986–1992 Mammals Implementation of regulation BA Stable + Balme et al. (2010) South Africa Phinda–Mkhuze 2002–2007 Mammals PA compared to buffer CI Increase + Bhattacharya (1993) India Kaziranga NP 1908–1991 Mammals Introduction of stafﬁng BA Increase + Blake et al. (2008) Congo 6 PAs 2003–2005 Mammals PA compared to buffer CI Increase + Brereton et al. (2008) England Multiple 1981–2000 Insecta Establishment of PA BA Increase + Caro (1999) Tanzania Katavi NP 1995–1996 Mammals PA compared to buffer CI Increase + Caro (1999) Tanzania Katavi NP 1995–1996 Mammals Game controlled area CI Increase + Caro (1999) Tanzania Katavi NP 1995–1996 Mammals Forest reserve CI Increase + Carrillo et al. (2000) Costa Rica Corcovado NP and Golfo Dulce FR 1990–1994 Mammals Different levels of protection CI Stable + Catry et al. (2009) Portugal Castro Verde 1996–2007 Aves Introduction of artiﬁcial nests CI Increase + Devictor et al. (2007) France All protected areas 1989–2003 Aves National estimates outside PA CI Increase + Eberhardt et al. (2007) USA Yellowstone NP 70 years Mammals Implementation of regulation BA Increase + Fellers and Drost (1993) USA Lassen Volcanic NP 1978–1991 Amphibian Establishment of management BA Decrease À 230–238 (2013) 161 Conservation Biological / al. et Geldmann J. Gough and Kerley (2006) South Africa Addo Elephant NP 1931–2002 Mammals Introduction of fence BA Increase + Harrington et al. (1999) South Africa Kruger NP 1977–1993 Mammals Closing of waterpoints BA Decrease + Herremans and Herremans-Tonnoeyr Botswana Multiple 1991–1995 Aves PA compared to buffer CI Increase + (2000) Hilborn et al. (2006) Tanzania Serengeti NP 1955–2005 Mammals Implementation of regulation BA Decrease + Ma et al. (2009) China Yancheng 1982–2003 Aves Different zones of PA CI Decrease + Mduma et al. (1999) Kenya, Tanzania Serengeti NP 1958–1998 Mammals Establishment of PA BA Increase +

Meijaard and Nijman (2000) Indonesia Pulau Kraget 1997 Mammals Translocation of population BA Decrease - chapter I Metzger et al. (2010) Tanzania Serengeti NP 1970–2008 Mammals Implementation of regulation BA Decrease +

43 Ottichilo et al. (2000) Kenya Masai Mara 1977–1997 Mammals PA compared to buffer CI Decrease 0 Pedrono et al. (2009) Vietnam Yok Don, Cat Tien, Ea So, and Vinh 1990–2005 Mammals Non-protected land within species CI Decrease + Cuu range Pettorelli et al. (2010) Tanzania 5 NPs, 3FR and 3 other PAs 2004–2007 Mammals Different levels of protection CI Increase + Schlicht et al. (2009) USA Multiple 1988–1996 Insecta Areas not managed with ﬁre CI Decrease À Sergio et al. (2005) Spain Doñana NP 1989–2001 Aves Populations outside PA CI Stable 0 Sinclair et al. (2007) Tanzania Serengeti NP 1955–2005 Mammals Implementation of regulation BA Decrease + Stoner et al. (2007) Tanzania Burigi-Biharamulo NP 1980s–2000s Mammals PA compared to buffer CI Decrease + Stoner et al. (2007) Tanzania Greater Ruaha NP 1980s–2000s Mammals PA compared to buffer CI Decrease + Stoner et al. (2007) Tanzania Tarangire NP 1980s–2000s Mammals PA compared to buffer CI Decrease + Stoner et al. (2007) Tanzania Selous-Mikumi NP 1980s–2000s Mammals PA compared to buffer CI Decrease + Stoner et al. (2007) Tanzania Ugalla NP 1980s–2000s Mammals PA compared to buffer CI Decrease + Struhsaker et al. (2005) 11 African 16 PAs 1966–2000 Biodiversity PA compared to buffer CI N/A + countries Suárez et al. (1993) Spain Las Amoladeras and Layna Paramos 1989 Aves PA compared to similar habitat outside CI Decrease À Tambling and Toit (2005) South Africa Pilanesburg NP 1995–2001 Mammals Introduction of fence BA Decrease À Theberge et al. (2006) Canada Algonquin 1988–1999 Mammals PA compared to buffer CI Decrease + Wegge et al. (2009) Nepal Bardia NP 22 years Mammals Establishment of PA BA Increase + Western et al. (2009) Kenya Tsavo NP 30 years Mammals PA compared to buffer CI Decrease 0 Western et al. (2009) Kenya Mara NP 30 years Mammals PA compared to buffer CI Decrease 0 Western et al. (2009) Kenya Amboseli NP 30 years Mammals PA compared to buffer CI Decrease 0 Western et al. (2009) Kenya Meru NP 30 years Mammals PA compared to buffer CI Decrease 0 Whitehead et al. (2008) New Zealand Fiordland NP 2000–2006 Aves Managed section compared to CI Increase + unmanaged

Key: NP = National Park, FR = forest reserve, PA = protected area, BA = before/after, CI = control/intervention. See Table S7 for further information on the individual studies. Counterfactual deﬁnes the comparator which the PA was evaluated against and BACI whether the comparison was before/after or control/intervention. Background trend deﬁnes the overall direction of the majority of the populations (see ratio in Table S7) which can be decreasing even in successful PAs. Effect of PA describes whether protection was better than counterfactual (+) worse than counterfactual ( À), or no difference could be detected (0). chapter I

J. Geldmann et al. / Biological Conservation 161 (2013) 230–238 233

Table 2 Effectiveness of terrestrial protected areas in reducing habitat loss and population declines.

Region Counterfactual Impact Habitat Buffer Regional Matched Positive Negative No effect % Positive Mean difference Forest Multiple Africa 4 11 2 11 4 3 61% 4.7 14 4 Asia 10 3 2 14 1 1 88% 2.4 16 0 Europe 1 0 0 1 0 0 100% – 1 0 Latin America 19 7 5 30 4 1 86% 6.0 33 2 North America 1 0 0 1 0 0 100% – 0 1 Oceania 0 2 0 2 0 0 100% – 2 0 Global 1 2 1 4 0 0 100% – 2 1 Summary 36 24 10 62 9 5 82% – 68 8 pared with the counterfactual (43 studies), ratios ranged from 1.25 factual. Relative performance was worse with protection than (Curran et al., 2004) to 22.7 (Nepstad et al., 2006) times lower loss, without in five studies, and six studies found no effect of protection with an mean of 5.4 (S.D. = 4.9). For the nine studies where PAs had (Table 1 Individual study details, Table 3 Summary data). higher rates of habitat loss compared with the counterfactual, the The largest number of PAs included in any of the 42 studies was difference ranged between 1.15 (Brower et al., 2002) and 3.97 (Liu 16, spread across 11 African countries (Struhsaker et al., 2005). Se- et al., 2001) times higher loss. Differences between inside and out- ven of the 42 studies were at regional or national scale, and 35 side were generally larger for Latin America and Africa, compared to (83%) were of five or less PAs. Like habitat studies, population studies Asia, suggesting that Latin American and African PAs are better at also exhibited geographic bias with 57% from Africa, as well as a tax- reducing deforestation within their borders (Table 2). onomic bias with 74% studying mammals (Table 3). Thirty-four of Studies using a buffer analysis reported higher levels of PA the 42 studies measured changes in species population abundance, effectiveness (mean = 5.2, S.D. = 5.0) than studies which used three measured changes in occurrence, and five used other measures regression modeling (mean = 4.2, S.D. = 5.4). For studies using such as spot counts, questionnaires, or nest mortality (Table S6). matching estimators only one reported a PA/counterfactual ratio, Counterfactuals varied across studies. Fifteen of the 42 used a be- finding 2 times more deforestation outside PAs compared to inside fore/after (BA) counterfactual: Three of those compared the same (Mas, 2005). Similarly, Joppa and Pfaff (2011) comparing the re- area before and after establishment of the PA, and the other 12 com- sults of matching estimators and buffer analyses, also found rates pared the same populations within a PA before and after the of habitat loss in PAs to be smaller using matching. Such results implementation of specific management actions. The remaining 27 show that methods used to evaluate PA effectiveness can alter of the 42 population studies used a control/intervention (CI) counter- the apparent effect size. factual: 16 of those compared populations from one or several PAs to Three global studies examined deforestation rates between re- populations with the PAs immediate surroundings, five compared serves under different IUCN reserve management categories (Joppa trends in PAs to non-protected land with similar characteristics but and Pfaff, 2011; Nelson and Chomitz, 2009; Scharlemann et al., not adjoining the reserve, and six compared populations between 2010), all finding that PA effectiveness increased with IUCN catego- PAs with varying legislation or management (Table 1). ries that infer stricter protection. However, Joppa and Pfaff (2011) In addition to the effect of protection per se, species populations showed this effect to be partly explained by the larger size of cat- in all studies were also affected by specific management actions egory I and II reserves. All seven studies investigating the effective- (Table S6). Consequently, the impacts of protection and individual ness of indigenous protected lands found positive impacts management actions are confounded. In addition, impacts of man- compared to non-protected areas. In the eight studies that com- agement and protection were evaluated using a range of dissimilar pared indigenous or community managed reserves with state man- methods. It is therefore inappropriate and uninformative to calcu- aged PAs, three found community reserves to perform better (Bray late effect sizes. Instead, we report direction of change (improving/ et al., 2008; Ellis and Porter-Bolland, 2008) and five found them to no effect/declining) compared to the counterfactual, as this was the perform worse (Armenteras et al., 2006; Bleher et al., 2006; Gaveau only measure of success which could be justifiably compared be- et al., 2007; Nelson et al., 2001; Nepstad et al., 2006). tween studies (Table 1). Twenty studies included the effect of PA management; ranging The most commonly reported management actions were those from implementation of management plans and staff numbers, to aimed at reducing poaching (12 of 42 studies). Eleven of the 12 involvement of local NGOs. None of the studies could estimate studies reported improved biodiversity outcomes linked to man- the explicit effect of management. Of the 20 studies, eight calcu- agement actions, though of variable magnitude (Fig. 1B). Species lated the difference between inside and outside, showing 2.4 were typically mammals and six of the studies examined large (S.D. = 1.5) times lower deforestation inside PA boundaries. African herbivores. In Serengeti NP, reestablishment of anti-poach- 2Where studies used regression modeling to control for the ef- ing efforts resulted in a shift from large declines in buffalo popula- fect of exogenous biological and geographical variables on habitat tions to increasing population in a short time period (Metzger loss, the effect of these variables was often reported. Remote areas et al., 2010). Similarly, elk populations in Yellowstone exhibited of higher or steeper terrain were generally reported to suffer less large population increases following anti-poaching regulations. In habitat loss (Joppa and Pfaff, 2011). Areas with high human popu- Costa Rica mammals in less strictly guarded reserves were 6–28% lation densities, located in areas with high demand for land, or the relative density of that in reserves with strict anti-poaching with high fire frequency were reported more frequently to suffer regulation (Carrillo et al., 2000). More subtly, in Vietnam banteng greater habitat loss (Fig. 1A). populations (Bos javanicus birmanicus) declined only slightly more slowly when guard numbers increased (Pedrono et al., 2009). Three 3.2. Protecting species populations studies examined the use of fences. One noted stable roan antelope (Hippotragus equinus) populations compared to decreases outside The relative impact of protection and management was positive fencing (Harrington et al., 1999) and a second noted no difference in 31 of the 42 studies; in 12 of these, species populations still with bird populations remaining stable before and after fencing exhibited declined under protection, but less than in the counter- (Sergio et al., 2005). The third showed negative changes, with ca. 44 chapter I

234 J. Geldmann et al. / Biological Conservation 161 (2013) 230–238 A

Fig. 1. Effects of drivers and management interventions of (A) habitat change and (B) species population change. The x-axis represents the number of studies including specific drivers or interventions. To the right of the center line are studies where drivers and interventions contributed positively to the effectiveness of PAs and to the left are studies where drivers or interventions had no effect (grey) or contributed negatively (black) to PA effectiveness. ‘‘No effect’’ is lumped with negative contributions to reflect the aim of the review to identify which drivers and interventions that could improve PA effectiveness. Total scores exceed the number studies, as some studies reported multiple driver and interventions. For habitat studies (1A) the figure includes 67 studies: 33 inside–outside, 22 regression, eight matching, two ground based, and two using questionnaires. For population time series (1B) the figure includes 35 studies: 13 of those compared populations from one or several PAs to populations with the PAs immediate surroundings, 12 compared the same populations within a PA before and after the implementation of specific management actions, five compared populations between PAs with varying legislation or management, three compared trends in PAs to non-protected land with similar characteristics, two examined the same area before and after establishment of the PA.

Table 3 Summary results derived from analysis of 42 studies of the impact of protected areas on species populations.

Region Counterfactual Impact Taxa Buffer PA establ. Strictness of protection Intervention Positive Negative No effect Mammals Birds Other Africa 14 1 3 6 18 1 5 22 1 1 Asia 1 1 1 2 4 1 0 4 1 0 Europe 3 1 0 1 3 1 1 0 4 1 Latin America 1 0 1 0 2 0 0 2 0 0 North America 2 0 0 3 3 2 0 3 0 2 Oceania 0 0 1 0 1 0 0 0 1 0 Summary 21 3 6 12 31 5 6 31 7 4

50% declines in wildebeest (Connochaetus taurinus) populations Ten studies evaluated speciﬁc conservation interventions de- after fencing, while lion (Panthera leo) populations increased to signed to target threats or challenges in the PA (Fig. 1B). Types of an estimated three times their natural carrying capacity (Tambling actions include: Burning (Schlicht et al., 2009), grazing (Fellers and Toit, 2005). and Drost, 1993; Herremans and Herremans-Tonnoeyr, 2000; 45 chapter I

J. Geldmann et al. / Biological Conservation 161 (2013) 230–238 235

Wegge et al., 2009), predator and invasive species exclusion (Suár- data, habitat change is often used as a proxy for overall PA perfor- ez et al., 1993; Whitehead et al., 2008), and involvement of NGOs mance. However, remote sensing generally precludes the estima- (Struhsaker et al., 2005). In four cases management was targeted tion of changes in degradation and quality, and thus risks at specific species, including provision of feeding and breeding overestimating the value of remaining habitat (Redford, 1992; Wil- sites for lesser kestrel (Falco naumanni)(Catry et al., 2009) and kie et al., 2011). Quantification of the relationship between habitat red-crowned crane (Grus japonensis)(Ma et al., 2009), animal-vac- and other outcomes in PAs would be valuable. cination programs for buffalo and wildebeest (Sinclair et al., 2007), The use of remote sensed data also means that most studies of and a failed translocation of proboscis monkeys (Nasalis larvatus) habitat change are able to estimate some measure of relative im- (Meijaard and Nijman, 2000). pact by comparing the rate of change inside PAs to a counterfac- Thirty-eight of the 42 studies measured one or more additional tual. However, although such estimates are valid for individual variables that might be influencing population trends, such as im- studies, care should be taken when comparing between studies. pact of diseases (n = 4), weather (n = 18), inter and intraspecific Further, multiple studies from the same publication are potentially competition (n = 3 and n = 16), food availability (n = 10) or habitat not independent from one another. The overall summary statistics properties (n = 17) (Table S6). No studies were able to control for (e.g. number of studies reporting positive or negative outcomes) the impact of these variables when evaluating the effect of protec- thus need to be interpreted carefully. As demonstrated by Joppa tion, but in all cases the authors indicated they were unlikely to and Pfaff (2010) the counterfactual chosen (such as buffer analyses have affected the overall direction of the results. vs. matching estimators), and to what extent models control for Fourteen studies considered the impact of predator–prey inter- biases in PA placement (i.e. isolation and topology) greatly influ- actions on population under protection. Of these, seven did not re- ences estimates of the relative impacts of protection. Studies con- port any effect of protection on interactions, four reported trol for potential confounding effects to varying degrees, and this increases in both prey and predator species (Carrillo et al., 2000; influences the resulting impact ratio. In addition, sample size var- Eberhardt et al., 2007; Sergio et al., 2005; Wegge et al., 2009), ies widely between studies, and was often not described suffi- one reported increases in predator species and declines in prey ciently well to weight in quantitative comparison. species (Sinclair et al., 2007) (although declines were smaller com- Conversely, in species abundance studies, which require long- pared to the counterfactual), and two studies reported population term field monitoring inside and outside of PAs, 29 of the 42 stud- declines within PAs greater than the counterfactual, possibly due ies included measures of endogenous drivers (i.e. staffing, fencing to increased predation (Suárez et al., 1993; Tambling and Toit, or management plans), but studies generally lacked the coverage 2005). required to draw generic and robust conclusions. Further, a large number of natural ecological processes also influence population 4. Discussion changes, which makes quantifying the precise effect of protection difficult due to substantial background variance. Events such as This review highlights the limited availability of evidence on droughts and floods, diseases and inter-specific competition affect the impact of PAs on habitats and fauna. Further, and more alarm- population numbers, and these events are usually not controlled ingly, there is very little quantitative understanding of how, and for in time-series studies (Owen-Smith et al., 2005; Sinclair et al., under what conditions various PA management interventions im- 2007; Western et al., 2009). prove PA effectiveness. Unfortunately, collection of population time-series data is Analysis of 76 studies across local, regional, and global scales costly (in time, money and human capacity) It is therefore not sur- indicates that PAs experience lower rates of habitat loss than areas prising that studies are skewed towards ‘charismatic’ species, and that are not protected. However, the majority of habitat studies PAs where conservation has a high and direct monetary value suggest that the effect size of protection is small: PAs may be (Balmford et al., 2009). While cost is a major factor influencing reducing the rate of decline compared to counterfactuals, but the implementation of population monitoring, and PA financing where external threats are high PAs are still experiencing dramatic generally does not extend to monitoring outside PAs, nor do man- habitat losses within their borders. agement agencies usually have the desire, will or capacity to For species populations, the effect of protection is unclear and undertake such activities within constrained budgets. Further, that this review found only 35 publications with appropriate coun- many species found within PAs are extirpated outside park bound- terfactuals highlights the lack of sufficient evaluation in PA man- aries (Metzger et al., 2010), or have migratory ranges extending agement. The 42 studies compiled lend some support for PAs outside PA boundaries (Thirgood et al., 2004), making even the being effective, but are not unanimous. This highlights the impor- identification of unprotected control populations challenging. tance of monitoring in PA management and decision-making – For the majority of species population studies identified in this without monitoring we cannot manage effectively (Stem et al., review, PA aims were broad, and management objectives were to 2005). The majority of publications do show at least some positive protect native ecosystems and their constituent species. Less than impact of protection, but poor sample size, and bias in geography 12% of studies investigated single-species management interven- and taxonomy make generalization unwise. Investment in anti- tions. However the importance of broad site-level management poaching appears to be very effective; however given the limited even for single-species conservation has been highlighted by sev- sample, it is impossible to tell whether publication bias has re- eral studies (Forrest et al., 2012; Liu et al., 2001; Palomares et al., sulted in only positive or complex outcomes being reported, bias- 2000) and site-level interventions constitute the majority (80%) ing this conclusion. of suggested interventions for the conservation of threatened spe- Habitat studies generally use remote sensed data, and can gen- cies (McCarthy et al., 2012). The effectiveness of PAs is likely to erate huge sample sizes across the planet. Unfortunately sufficient vary with how appropriately specific conservation interventions accuracy and resolution is primarily limited to forests. Habitat are tailored to individual species. studies have therefore been able to statistically correlate patterns of habitat loss with various exogenous drivers (e.g. Mas, 2005; 4.1. Moving forward Mertens et al., 2004), However, few studies have gone beyond spatial layers to study the more subtle impacts of governance struc- The Convention on Biological Diversity’s Aichi target 11 calls for ture or management interventions on the ground (Oestreicher 17% of terrestrial land surface area to be protected and effectively et al., 2009). Due to the relatively low cost of using remote sensed and equitably managed by 2020. Even if the coverage goal is 46 chapter I

236 J. Geldmann et al. / Biological Conservation 161 (2013) 230–238 achieved, effective and equitable management is unlikely without References site-level monitoring and adaptation. Conservation scientists and practitioners can improve understanding of PA based conservation Adams, L.G., Stephenson, R.O., Dale, B.W., Ahgook, R.T., Demma, D.J., 2008. Population dynamics and harvest characteristics of wolves in the Central by: (a) adopting a more experimental approach for the implemen- Brooks Range. Alaska. Wildlife. Monogr. 170, 1–25. tation of conservation activities; emphasizing the causal link be- Alodos, C.L., Pueyo, Y., Barrantes, O., Escós, J., Giner, L., Robles, A.B., 2004. Variations tween interventions and the outcomes being measured and (b) in landscape patterns and vegetation cover between 1957 and 1994 in a promoting sharing and publication of data in a standardized for- semiarid Mediterranean ecosystem. Landscape Ecol. 19, 543–559. Alsheikh-Ali, A.A., Qureshi, W., Al-Mallah, M.H., Ioannidis, J.P.A., 2011. Public mat, thus facilitating the use and collation of data from across availability of published research data in high-impact journals. PLoS One 6, studies. e24357. Most of the 2599 studies considered for this review did not fol- Armenteras, D., Rudas, G., Rodriguez, N., Sua, S., Romero, M., 2006. Patterns and causes of deforestation in the Colombian Amazon. Ecol. Indic. 6, 353–368. low a BACI design. BACI design is being increasingly demanded by Balme, G.A., Slotow, R., Hunter, L.T.B., 2010. Edge effects and the impact of non- conservation scientists (Ferraro, 2009; Joppa and Pfaff, 2010), but is protected areas in carnivore conservation: leopards in the Phinda-Mkhuze still rarely implemented. For too long, past practice and theory Complex, South Africa. Anim. Conserv. 13, 315–323. Balmford, A., Beresford, J., Green, J., Naidoo, R., Walpole, M., Manica, A., 2009. A have been used to guide decision making in conservation, and in global perspective on trends in nature-based tourism. PLoS Biol. 7, e1000144. particular in PA decision making. Existing initiatives to collate data Bertzky, B., Corrigan, C., Kemsey, J., Kenney, S., Ravilious, C., Besancon, C., Burgess, on population time-series such as the Living Planet database (Loh N.D., 2012. Protected planet report: tracking progress towards global targets for protected areas. IUCN and UNEP-WCMC, Gland, Switzerland and Cambridge, et al., 2005) facilitate the ‘scaling-up’ of multiple small-scale stud- UK. ies by making data freely available in standardized format. Simi- Bhattacharya, A., 1993. The status of the Kaziranga Rhino population. Tiger Papers 1, larly, more recent efforts are underway to collate data on PA 1–6. Blake, S., Deem, S.L., Strindberg, S., Maisels, F., Momont, L., Isia, I.-B., Douglas- management (Leverington et al., 2010) which will aid further anal- Hamilton, I., Karesh, W.B., Kock, M.D., 2008. Roadless wilderness area yses of PA effectiveness. Conservation journals could facilitate determines forest Elephant movements in the Congo Basin. PLoS One 3, these efforts by adopting routine policies for data reporting and e3546. sharing following publication, and/or ensuring that existing data Bleher, B., Uster, D., Bergsdorf, T., 2006. Assessment of threat status and management effectiveness in Kakamega Forest, Kenya. Biodivers. Conserv. 15, policies are consistently followed by researchers; currently only 1159–1177. 9% of the raw data from high impact publications are made avail- Bray, D.B., Duran, E., Ramos, V.H., Mas, J.-F., Velazquez, A., McNab, R., Barry, D., able online (Alsheikh-Ali et al., 2011). Radachowsky, J., 2008. Tropical deforestation, community forests, and protected areas in the Maya forest. Ecol. Soc. 13, 56. Recent studies illustrate the potential for meta-analyses to Brereton, T.M., Warren, M.S., Roy, D.B., Stewart, K., 2008. The changing status of identify patterns in population changes: amongst regions of Africa the Chalkhill Blue butterfly Polyommatus coridon in the UK: the impacts of (Craigie et al., 2010), differences in species recovery correlated conservation policies and environmental factors. J. Insect Conserv. 12, 629– 638. with increased management activity across Australia (Taylor Brooks, T.M., Bakarr, M.I., Boucher, T., Da Fonseca, G.A.B., Hilton-Taylor, C., Hoekstra, et al., 2011), or correlation between species persistence inside J.M., Moritz, T., Olivier, S., Parrish, J., Pressey, R.L., Rodrigues, A.S.L., Sechrest, W., West African PAs and management resources (Tranquilli et al., Stattersfield, A., Strahm, W., Stuart, S.N., 2004. Coverage provided by the global protected-area system: is it enough? Bioscience 54, 1081–1091. 2012). Studies such as these, which go beyond case-based results, Brooks, T.M., Mittermeier, R.A., da Fonseca, G.A.B., Gerlach, J., Hoffmann, M., help bridge the gap between conservation practitioners working on Lamoreux, J.F., Mittermeier, C.G., Pilgrim, J.D., Rodrigues, A.S.L., 2006. Global the ground and the policy processes, setting the stage for further biodiversity conservation priorities. Science 313, 58–61. Brower, L.P., Castilleja, G., Peralta, A., Lopez-Garcia, J., Bojorquez-Tapia, L., Diaz, S., investments and engagement in biodiversity conservation. As Melgarejo, D., Missrie, M., 2002. Quantitative changes in forest quality in a studies increasingly document dramatic declines in habitat extent principal overwintering area of the monarch butterfly in Mexico, 1971–1999. and biodiversity, both inside and outside PAs, the conservation Conserv. Biol. 16, 346–359. community needs to move beyond asking ‘what works’ to ‘when’ Bruner, A.G., Gullison, R.E., Rice, R.E., da Fonseca, G.A.B., 2001. Effectiveness of parks in protecting tropical biodiversity. Science 291, 125–128. and ‘why’. This will require further effort to measure reserve effec- Butchart, S.H.M., Walpole, M., Collen, B., van Strien, A., Scharlemann, J.P.W., Almond, tiveness, and the linkages between input and management mea- R.E.A., Baillie, J.E.M., Bomhard, B., Brown, C., Bruno, J., Carpenter, K.E., Carr, G.M., sures, and species and habitat outcomes. The continuing reliance Chanson, J., Chenery, A.M., Csirke, J., Davidson, N.C., Dentener, F., Foster, M., Galli, A., Galloway, J.N., Genovesi, P., Gregory, R.D., Hockings, M., Kapos, V., on PAs as instruments for the protection of biodiversity means that Lamarque, J.F., Leverington, F., Loh, J., McGeoch, M.A., McRae, L., Minasyan, A., testing how and why they are effective is of critical importance to Morcillo, M.H., Oldfield, T.E.E., Pauly, D., Quader, S., Revenga, C., Sauer, J.R., conservation science. Skolnik, B., Spear, D., Stanwell-Smith, D., Stuart, S.N., Symes, A., Tierney, M., Tyrrell, T.D., Vie, J.C., Watson, R., 2010. Global biodiversity: indicators of recent declines. Science 328, 1164–1168. Campos, M.T., Nepstad, D.C., 2006. Smallholders, the Amazon’s new Acknowledgements conservationists. Conserv. Biol. 20, 1553–1556. Caro, T.M., 1999. Densities of mammals in partially protected areas: the Katavi We thank Professor A. Pullin and the Collaboration for Environ- ecosystem of western Tanzania. J. Appl. Ecol. 36, 205–217. Carrillo, E., Wong, G., Cuaron, A.D., 2000. Monitoring mammal populations in Costa mental Evidence for handling the systematic review, and the five Rican protected areas under different hunting restrictions. Conserv. Biol. 14, reviewers, especially Dr. D. Dawson for invaluable feedback on 1580–1591. the original manuscript. We also thank the three anonymous Catry, I., Alcazar, R., Franco, A.M.A., Sutherland, W.J., 2009. Identifying the reviewers of this manuscript for valuable contributions and effectiveness and constraints of conservation interventions: a case study of the endangered lesser kestrel. Biol. Conserv. 142, 2782–2791. suggestions. Coad, L., Burgess, N.D., Fish, L., Ravillious, C., Corrigan, C., Pavese, H., Granziera, A., We thank the Danish National Research Foundation for finan- Besançon, C., 2008. Progress towards the convention on biological diversity cial support. We also thank the IUCN SSC/WCPA Joint Task-Force terrestrial 2010 and marine 2012 targets for protected area coverage. Parks 17, 35–42. on Biodiversity and Protected Areas, UNEP–WCMC, WWF, and Craigie, I.D., Baillie, J.E.M., Balmford, A., Carbone, C., Collen, B., Green, R.E., Hutton, the University of Queensland for financial and institutional J.M., 2010. Large mammal population declines in Africa’s protected areas. Biol. support. Conserv. 143, 2221–2228. Curran, L.M., Trigg, S.N., McDonald, A.K., Astiani, D., Hardiono, Y.M., 2004. Lowland forest loss in protected areas of Indonesian Borneo. Science 303, 1000–1003. DeFries, R., Hansen, A., Newton, A.C., Hansen, M.C., 2005. Increasing islolation of Appendix A. Supplementary material protected areas in tropical forests over the past twenty years. Ecol. Appl. 15, 19– 26. Supplementary data associated with this article can be found, in Devictor, V., Godet, L., Julliard, R., Couvet, D., Jiguet, F., 2007. Can common species benefit from protected areas? Biol. Conserv. 139, 29–36. the online version, at http://dx.doi.org/10.1016/j.biocon.2013.02. Dudley, N., 2008. Guidelines for Applying Protected Area Management Categories. 018. International Union for Conservation of Nature, Gland, Switzerland. 47 chapter I

J. Geldmann et al. / Biological Conservation 161 (2013) 230–238 237

Dudley, N., Stolton, S., 1999. Conversion of ‘‘Paper Parks’’ to Effective Management – Nelson, A., Chomitz, K.M., 2009. Protected Area Effectiveness in Reducing Tropical Developing a Target. IUCN, WWF, WCPA. Deforestation. A Global Analysis of the Impact of Protection Status, The World Eberhardt, L.L., White, P.J., Garrott, R.A., Houston, D.B., 2007. A seventy-year history Bank, Washington DC. of trends in Yellowstone’s northern elk herd. J. Wildlife Manag. 71, 594–602. Nelson, G.C., Harris, V., Stone, S.W., 2001. Deforestation, land use, and property Ellis, E.A., Porter-Bolland, L., 2008. Is community-based forest management more rights: empirical evidence from Darién, Panama. Land Econ. 77, 187–205. effective than protected areas?: a comparison of land use/land cover change in Nepstad, D., Schwartzman, S., Bamberger, B., Santilli, M., Ray, D., Schlesinger, P., two neighboring study areas of the Central Yucatan Peninsula, Mexico. For. Ecol. Lefebvre, P., Alencar, A., Prinz, E., Fiske, G., Rolla, A., 2006. Inhibition of Amazon Manage. 256, 1971–1983. deforestation and fire by parks and indigenous lands. Conserv. Biol. 20, Fellers, G.M., Drost, C.A., 1993. Disappearance of the cascades frog Rana cascadae at 65–73. the southern end of its range, California, USA. Biol. Conserv. 65, 177–181. Oestreicher, J.S., Benessaiah, K., Ruiz-Jaen, M.C., Sloan, S., Turner, K., Pelletier, J., Ferraro, P.J., 2009. Counterfactual thinking and impact evaluation in environmental Guay, B., Clark, K.E., Roche, D.G., Meiners, M., Potvin, C., 2009. Avoiding policy. New Dir. Eval. 2009, 75–84. deforestation in Panamanian protected areas: an analysis of protection Ferraro, P.J., Simpson, R.D., 2002. The cost-effectiveness of conservation payments. effectiveness and implications for reducing emissions from deforestation and Land Econ. 78, 339–353. forest degradation. Global Environ. Change 19, 279–291. Forrest, J.L., Wikramanayake, E., Shrestha, R., Areendran, G., Gyeltshen, K., Ottichilo, W.K., De, L.J., Skidmore, A.K., Prins, H.H.T., Said, M.Y., 2000. Population Maheshwari, A., Mazumdar, S., Naidoo, R., Thapa, G.J., Thapa, K., 2012. trends of large nonmigratory wild herbivores and livestock in the Masai Mara Conservation and climate change: assessing the vulnerability of snow leopard ecosystem, Kenya, between 1977 and 1997. Afr. J.Ecol. 38, 202–216. habitat to treeline shift in the Himalaya. Biol. Conserv. 150, 129–135. Owen-Smith, N., Mason, D.R., Ogutu, J.O., 2005. Correlates of survival rates for 10 Gaveau, D.L.A., Wandono, H., Setiabudi, F., 2007. Three decades of deforestation in African ungulate populations: density, rainfall and predation. J. Anim. Ecol. 74, southwest Sumatra: have protected areas halted forest loss and logging, and 774–788. promoted re-growth? Biol. Conserv. 134, 495–504. Palomares, F., Delibes, M., Ferreras, P., Fedriani, J.M., Calzada, J., Revilla, E., 2000. Geldmann, J., Barnes, M., Coad, L., Craigie, I.N., Hockings, M., Burgess, N., 2013. Iberian lynx in a fragmented landscape: predispersal, dispersal, and Effectiveness of terrestrial protected areas in reducing biodiversity and habitat postdispersal habitats. Conserv. Biol. 14, 809–818. loss. CEE 10-007, Collaboration for Environmental Evidence. Available from: Pedrono, M., Ha, M.T., Chouteau, P., Vallejo, F., 2009. Status and distribution of the . Endangered banteng Bos javanicus birmanicus in Vietnam: a conservation Gough, K.F., Kerley, G.I.H., 2006. Demography and population dynamics in the tragedy. Oryx 43, 618–625. elephants Loxodonta africana of Addo Elephant National Park, South Africa: is Pettorelli, N., Lobora, A.L., Msuha, M.J., Foley, C., Durant, S.M., 2010. Carnivore there evidence of density dependent regulation? Oryx 40, 434–441. biodiversity in Tanzania: revealing the distribution patterns of secretive Harrington, R., Owen-Smith, N., Viljoen, P.C., Biggs, H.C., Mason, D.R., Funston, P., mammals using camera traps. Anim. Conserv. 13, 131–139. 1999. Establishing the causes of the roan antelope decline in the Kruger Pullin, A.S., Knight, T.M., 2009. Doing more good than harm - Building an evidence- National Park, South Africa. Biol. Conserv. 90, 69–78. base for conservation and environmental management. Biol. Conserv. 142, 931– Herremans, M., Herremans-Tonnoeyr, D., 2000. Land use and the conservation 934. status of raptors in Botswana. Biol. Conserv. 94, 31–41. Redford, K.H., 1992. The Empty forest. Bioscience 42, 412–422. Hilborn, R., Arcese, P., Borner, M., Hando, J., Hopcraft, G., Loibooki, M., Mduma, S., Rodrigues, A.S.L., Akcakaya, H.R., Andelman, S.J., Bakarr, M.I., Boitani, L., Brooks, T.M., Sinclair, A.R.E., 2006. Effective enforcement in a conservation area. Science 314, Chanson, J.S., Fishpool, L.D.C., Da Fonseca, G.A.B., Gaston, K.J., Hoffmann, M., 1266. Marquet, P.A., Pilgrim, J.D., Pressey, R.L., Schipper, J., Sechrest, W., Stuart, S.N., Jachmann, H., 2008. Monitoring law-enforcement performance in nine protected Underhill, L.G., Waller, R.W., Watts, M.E.J., Yan, X., 2004. Global gap analysis: areas in Ghana. Biol. Conserv. 141, 89–99. priority regions for expanding the global protected-area network. Bioscience 54, Joppa, L., Pfaff, A., 2010. Reassessing the forest impacts of protection. The challenge 1092–1100. of nonrandom location and a corrective method. Ann. N.Y. Acad. Sci. 1185, 135– Scharlemann, J.P.W., Kapos, V., Campbell, A., Lysenko, I., Burgess, N.D., 149. Hansen, M.C., Gibbs, H.K., Dickson, B., Miles, L., 2010. Securing tropical Joppa, L.N., Loarie, S.R., Pimm, S.L., 2008. On the protection of protected areas. Proc. forest carbon: the contribution of protected areas to REDD. Oryx 44, Nat. Acad. Sci. 105, 6673–6678. 352–357. Joppa, L.N., Pfaff, A., 2011. Global protected area impacts. Proc. R. Soc. B – Biol. Sci. Schlicht, D., Swengel, A., Swengel, S., 2009. Meta-analysis of survey data to assess 278, 1633–1638. trends of prairie butterflies in Minnesota, USA during 1979–2005. J. Insect Kleiman, D.G., Reading, R.P., Miller, B.J., Clark, T.W., Scott, M., Robinson, J., Wallace, Conserv. 13, 429–447. R.L., Cabin, R.J., Felleman, F., 2000. Improving the evaluation of conservation Sergio, F., Blas, J., Forero, M., Fernandez, N., Donazar, J.A., Hiraldo, F., 2005. programs. Conserv. Biol. 14, 356–365. Preservation of wide-ranging top predators by site-protection: black and red Leverington, F., Costa, K.L., Pavese, H., Lisle, A., Hockings, M., 2010. A global analysis kites in Donana National Park. Biol. Conserv. 125, 11–21. of protected area management effectiveness. Environ. Manage. 46, 685–698. Sinclair, A.R.E., Mduma, S.A.R., Hopcraft, J.G.C., Fryxell, J.M., Hilborn, R., Thirgood, S., Liu, J., Linderman, M., Ouyang, Z., An, L., Yang, J., Zhang, H., 2001. Ecological 2007. Long-term ecosystem dynamics in the Serengeti: Lessons for degradation in protected areas: the case of Wolong Nature Reserve for Giant conservation. Conserv. Biol. 21, 580–590. Pandas. Science 292, 98–101. Stem, Margoluis, R., Salafsky, N., Brown, M., 2005. Monitoring and evaluation Loh, J., Green, R.E., Ricketts, T., Lamoreux, J., Jenkins, M., Kapos, V., Randers, J., 2005. in conservation: a review of trends and approaches. Conserv. Biol. 19, The living planet index: using species population time series to track trends in 295–309. biodiversity. Philos. Trans. R. Soc. B: Biol. Sci. 360, 289–295. Stoner, C., Caro, T., Mduma, S., Mlingwa, C., Sabuni, G., Borner, M., 2007. Assessment Ma, Z.J., Li, B., Li, W.J., Han, N.Y., Chen, J.K., Watkinson, A.R., 2009. Conflicts between of effectiveness of protection strategies in Tanzania based on a decade of survey biodiversity conservation and development in a biosphere reserve. J. Appl. Ecol. data for large herbivores. Conserv. Biol. 21, 635–646. 46, 527–535. Struhsaker, T.T., Struhsaker, P.J., Siex, K.S., 2005. Conserving Africa’s rain forests: Margules, C.R., Pressey, R.L., 2000. Systematic conservation planning. Nature 405, problems in protected areas and possible solutions. Biol. Conserv. 123, 45–54. 243–253. Suárez, F., Yanes, M., Herranz, J., Manrique, J., 1993. Nature-reserves and the Mas, J.-F., 2005. Assessing protected area effectiveness using surrounding (buffer) conservation of Iberian shrubsteppe passerines – the paradox of nest predation. areas environmentally similar to the target area. Environ. Monit. Assess. 105, Biol. Conserv. 64, 77–81. 69–80. Tambling, C.J., Toit, J.T.D., 2005. Modelling Wildebeest Population Dynamics: McCarthy, D.P., Donald, P.F., Scharlemann, J.P.W., Buchanan, G.M., Balmford, A., implications of predation and harvesting in a closed system. J. Appl. Ecol. 42, Green, J.M.H., Bennun, L.A., Burgess, N.D., Fishpool, L.D.C., Garnett, S.T., Leonard, 431–441. D.L., Maloney, R.F., Morling, P., Schaefer, H.M., Symes, A., Wiedenfeld, D.A., Taylor, M., Sattler, P., Evans, M., Fuller, R., Watson, J., Possingham, H., 2011. What Butchart, S.H.M., 2012. Financial costs of meeting global biodiversity works for threatened species recovery? an empirical evaluation for Australia. conservation targets: current spending and unmet needs. Science 338, 946–949. Biodivers. Conserv. 20, 767–777. Mduma, S.A.R., Sinclair, A.R.E., Hilborn, R., 1999. Food regulates the Serengeti Theberge, J.B., Theberge, M.T., Vucetich, J.A., Paquet, P.C., 2006. Pitfalls of applying Wildebeest: a 40-year record. J. Anim. Ecol. 68, 1101–1122. adaptive management to a wolf population in Algonquin Provincial Park. Meijaard, E., Nijman, V., 2000. The local extinction of the proboscis monkey Nasalis Ontario. Environ. Manage. 37, 451–460. larvatus in Pulau Kaget Nature Reserve, Indonesia. Oryx 34, 66–70. Thirgood, S., Mosser, A., Tham, S., Hopcraft, G., Mwangomo, E., 2004. Can parks Meir, E., Andelman, S., Possingham, H.P., 2004. Does conservation planning matter protect migratory ungulates? the case of the Serengeti wildebeest. Anim. in a dynamic and uncertain world? Ecol. Lett. 7, 615–622. Conserv. 7, 113–120. Mertens, B., Kaimowitz, D., Puntodewo, A., Vanclay, J., Mendez, P., 2004. Modeling Tole, L., 2002. Habitat loss and anthropogenic disturbance in Jamaica’s Hellshire deforestation at distinct geographic scales and time periods in Santa Cruz, Hills area. Biodivers. Conserv. 11, 575–598. Bolivia. Int. Reg. Sci. Rev. 27, 271–296. Tranquilli, S., Abedi-Lartey, M., Amsini, F., Arranz, L., Asamoah, A., Babafemi, O., Metzger, K., Sinclair, A., Hilborn, R., Hopcraft, J., Mduma, S., 2010. Evaluating the Barakabuye, N., Campbell, G., Chancellor, R., Davenport, T.R.B., Dunn, A., Dupain, protection of wildlife in parks: the case of African buffalo in Serengeti. J., Ellis, C., Etoga, G., Furuichi, T., Gatti, S., Ghiurghi, A., Greengrass, E., Biodivers. Conserv. 19, 3431–3444. Hashimoto, C., Hart, J., Herbinger, I., Hicks, T.C., Holbech, L.H., Huijbregts, B., Mwangi, M.A.K., Butchart, S.H.M., Munyekenye, F.B., Bennun, L.A., Evans, M.I., Imong, I., Kumpel, N., Maisels, F., Marshall, P., Nixon, S., Normand, E., Fishpool, L.D.C., Kanyanya, E., Madindou, I., Machekele, J., Matiku, P., Mulwa, R., Nziguyimpa, L., Nzooh-Dogmo, Z., Okon, D.T., Plumptre, A., Rundus, A., Ngari, A., Siele, J., Stattersfield, A.J., 2010. Tracking trends in key sites for Sunderland-Groves, J., Todd, A., Warren, Y., Mundry, R., Boesch, C., Kuehl, H., biodiversity: a case study using Important Bird Areas in Kenya. Bird Conserv. 2012. Lack of conservation effort rapidly increases African great ape extinction Int. 20, 215–230. risk. Conserv. Lett. 5, 48–55.

48 chapter I

238 J. Geldmann et al. / Biological Conservation 161 (2013) 230–238

Wegge, P., Odden, M., Pokharel, C.P., Storaas, T., 2009. Predator-prey relationships Whitehead, A.L., Edge, K.A., Smart, A.F., Hill, G.S., Willans, M.J., 2008. Large scale and responses of ungulates and their predators to the establishment of predator control improves the productivity of a rare New Zealand riverine duck. protected areas: a case study of tigers, leopards and their prey in Bardia Biol. Conserv. 141, 2784–2794. National Park, Nepal. Biol. Conserv. 142, 189–202. Wilkie, D.S., Bennett, E.L., Peres, C.A., Cunningham, A.A., 2011. The empty forest Western, D., Russell, S., Cuthill, I., 2009. The status of wildlife in protected areas revisited. Ann. N.Y. Acad. Sci. 1223, 120–128. compared to non-protected areas of Kenya. PLoS One 4, e6140.

CHAPTER II Commonalities and complementarities in Management and Evaluations

Michael B. Mascia, Sharon Pailler, Michele L. Thieme, Andy Rowe, Madeleine C. Bottrill, Finn Danielsen, Jonas Geldmann, Robin Naidoo, Andrew S. Pullin, and Neil D. Burgess

Accepted for publication in Biological Conservation

52 chapter II

Commonalities and Complementarities among Approaches to Conservation Monitoring and Evaluation

Michael B. Mascia,1*∞ Sharon Pailler1∞, Michele L. Thieme,1 Andy Rowe,2 Madeleine C. Bottrill, 3 Finn Danielsen, 4 Jonas Geldmann,5 Robin Naidoo,1 Andrew S. Pullin, 6 Neil D. Burgess1,5

1World Wildlife Fund, 1250 24th St NW, Washington, DC 20037, USA, 2 ARCeconomics Inc., 160 Westcott, Salt Spring Island, British Columbia, Canada V8K 1C2, 3 Conservation International, 2011 Crystal Drive, Suite 500, Arlington, VA 22202 USA, 4Nordisk Fond for Miljø og Udvikling, Skindergade 23-III, DK-1159 Copenhagen, Denmark, 5 Center for Macroecology, Evolution and Climate, Department of Biology, University of Copenhagen, Ole Maaløes Vej 5, DK-2200 Copenhagen, Denmark, 6 Centre for Development Evidence-Based Conservation, School of Environment, Natural Resources and Geography, Bangor University, Bangor Gwynedd LL54 2UW, UK *corresponding author, ∞ equal authorship

Abstract Commonalities and complementarities among approaches to conservation monitoring and evaluation (M&E) are not well articulated, creating the potential for confusion, misuse, and missed opportunities to inform conservation policy and practice. We examine the relationships among five approaches to conservation M&E, characterizing each approach in eight domains: the focal question driving each approach, when in the project cycle each approach is employed, scale of data collection, the methods of data collection and analysis, the implementers of data collection and analysis, the users of M&E outputs, and the decisions informed by these outputs. Ambient monitoring measures status and change in ambient social and ecological conditions, independent of any conservation intervention. Management assessment measures management inputs, activities, and outputs, as the basis for investments to build management capacity for conservation projects. Performance measurement assesses project or program progress toward desired levels of specific activities, outputs, and outcomes. Impact evaluation measures the intended and unintended consequences of conservation interventions, with emphasis upon long-term impacts on ecological and social conditions. Systematic review examines existing research findings to assess the state of the evidence regarding the impacts of conservation interventions, and to synthesize the insights emerging from this evidence base. Though these five approaches have some commonalities, they complement each other to provide unique insights for conservation planning, capacity-building, adaptive management, learning, and accountability. Ambient monitoring, management assessment, and performance measurement are now commonplace in conservation, but opportunities remain to inform conservation policy and practice more fully through catalytic investments in impact evaluations and systematic reviews.

Keywords: ambient monitoring, management assessment, performance measurement, impact evaluation, systematic review Accepted for publication in Biological Conservation, Elsevier B.V., ISSN: 0006-3207

53 chapter II

Introduction Monitoring and evaluation (M&E) has a long history in conservation, with diverse approaches developed for diverse purposes (Stem et al. 2005). In recent years, scholars have advocated M&E as a means to facilitate the wise use of scarce conservation funds (Ferraro and Pattanayak 2006; Sutherland et al. 2004), respond to the environmental implications of ineffective management (Hockings et al. 2006), promote accountability (Christensen 2003; Jepson 2005), and track progress towards broader conservation goals (Gratwicke et al. 2007). These aspirations for widespread and effective use of conservation M&E have become increasingly codified in policy (e.g., DANIDA 2006; USAID 2011), technical guidance (e.g., Hockings et al. 2006; Kapos et al. 2008; Margoluis and Salafsky 1998), and practice (Miteva et al. in press), as scholars, practitioners, and donors have sought to target conservation investments, track progress, foster adaptive management, ensure accountability, and catalyze learning within the conservation sector.

Despite previous reviews (e.g., Birnbaum and Mickwitz 2009; Kapos et al. 2008; Stem et al. 2005), the commonalities and complementarities among current approaches to conservation M&E are not well articulated. This absence of clarity creates the potential for confusion, misuse, and missed opportunities to inform conservation policy and practice through M&E (Stem et al. 2005). Misuse of M&E tools and approaches poses a number of risks (Oral History Project Team 2007), including misallocation of M&E resources, unreasonable expectations of M&E activities, inaccurate assessments of conservation interventions, and misguided allocation of conservation resources (Ferraro and Pattanayak 2006; Stem et al. 2005). If conservation M&E fails to realize its potential because of misuse and missed opportunities, the perceived value of M&E may decline and its use may be limited further. Effective use of M&E, by contrast, may catalyze a virtuous cycle of wider adoption across the conservation community, as predicted by social theories of innovation diffusion (Rogers 1995).

To foster more effective application of M&E approaches within the conservation sector, we examine commonalities and complementarities among five approaches to conservation M&E: ambient monitoring, management assessment, performance measurement, impact evaluation, and systematic review. We define each approach and characterize each in eight domains: the focal question driving the approach, when in the project cycle the approach is employed, the scale of data collection, the methods of data collection and analysis, the implementers of data collection and analysis, the users of M&E outputs, and the decisions informed by these outputs. We then explore the relationship of these five approaches to established frameworks for conservation planning and analysis, and highlight the implications for conservation science and policy. By providing conservation scholars and practitioners with a framework for understanding the relationships among approaches to conservation M&E, we hope to empower better informed “consumers” and “producers” of M&E.

54 chapter II

Concepts and terminology The abundance of jargon, much of it ill-defined, contributes to confusion regarding conservation M&E. In this analysis, we adhere to the concepts and terminology within the established literature on program evaluation (Table 1), as studied and practiced by members of the American Evaluation Association (www.eval.org). Monitoring is an ongoing function that systematically collects data on specified indicators, whereas evaluation is the systematic and objective assessment of an ongoing or completed project, program, or policy, often in order to determine the merit or worth of the intervention (DAC 2002). (Merit is the impact attributable to the intervention; worth is the value of these changes to decisionmakers and key stakeholders.) The distinctions among inputs, activities, outputs, outcomes, and impacts allow clear differentiation among approaches to conservation M&E (below). Similarly, projects, programs, and policies represent distinct scales of human action at which conservation M&E may occur.

Table 1. Definitions of key concepts in conservation monitoring and evaluation. Monitoringa An ongoing function that systematically collects data on specified indicators.

Evaluationa The systematic and objective assessment of an ongoing or completed project, program, or policy, often in order to determine the merit or worth of the intervention.

Inputa The financial, human, and material resources used for an intervention.

Activitya Actions taken or work performed through which inputs are mobilized to produce specific outputs.

Outputa The products, goods, and services that result from an intervention.

Outcomea The desired ends that intervention outputs are intended to induce (i.e., changes in knowledge and attitudes, behaviors, and/or social and environmental conditions).

Impacta The intended and unintended consequences (i.e., changes in knowledge and attitudes, behaviors, and/or social and environmental conditions) that are directly or indirectly caused by an intervention.

Projectb A discrete set of planned activities collectively trying to achieve a specific outcome or set of outcomes, often as a component of a program and as a manifestation of a policy.

Programb A suite of projects collectively organized to achieve a specific outcome or set of outcomes, often serving as a tactical means of achieving policy ends.

Policyc A broad or strategic statement of intent to accomplish specific aims, often implemented through one or more programs.

Interventiona Specific action (project, program, or policy) designed to effect a specific desired change; may be manifestation of broader strategy

Theory of changed An articulation and frequently graphical illustration of the assumed logical, causal relationships between intervention (project, program, policy) inputs, activities, outputs, and outcomes. a Definitions adapted from DAC (2002). b Definition adapted from Bartlett (1994). c Definition adapted from Brewer and DeLeon (1983). d Definition adapted from Weiss (1995). Sometimes referred to as a logic model, logframe, logical framework approach, or results chain.

55 chapter II

Some approaches to conservation M&E strive to document and measure progress against an intervention’s theory of change, which articulates and graphically illustrates the assumed logical and causal relationship between an intervention and its anticipated outcomes (Weiss 1995). Of the five approaches to conservation M&E that we examine, impact evaluation and systematic review attempt to explicitly test or examine the validity of these theories of change. Ambient monitoring, management assessment, and performance measurement, by contrast, do not explicitly test – and often assume – the validity of the underlying program logic (i.e., implementation of an intervention will lead to desired outcomes).

Approaches to Conservation M&E Ambient monitoring Focal question: What is the state of ambient social and/or environmental conditions, and how are these conditions changing over time and space?

Ambient monitoring is the process of systematically observing the state of social and/or environmental conditions over time, independent of any conservation intervention. Sometimes referred to as “status assessment” (Stem et al. 2005) or “surveillance monitoring” (Nichols and Williams 2006), ambient monitoring is not intended to measure the attributes or consequences of conservation interventions, but, rather, to characterize the broader social and ecological context within which conservation occurs. Depending upon the spatial and temporal scale of ambient monitoring, however, data derived from ambient monitoring efforts can be repurposed to inform M&E efforts that directly examine conservation interventions. Ambient monitoring may measure variables such as human demography (e.g., Hobbs and Stoops 2002), human health (e.g., ZIMSTAT & ICF 2012), patterns of natural resource use and other human behaviors (e.g., Ticheler et al. 1998), wildlife population size (e.g., Mysterud et al. 2007), and the condition of important habitats (e.g., Hansen et al. 2008). Government censuses of human populations, which date to perhaps the 16th century B.C. (Missiakoulis 2010), were likely the first formal attempts at ambient monitoring; farmers, fishers, and forest users have informally monitored resource conditions for even longer, their observations influencing survival strategies and resource use (Danielsen et al. 2000). Formal ecological monitoring initially focused on monitoring populations of various species over time. For instance, on the basis of fish catch records, Bertram documented the (declining) population of inshore Scottish herring Clupea harengus in the 19th century (Bertram 1865). Methods for ambient monitoring have since diversified and become more sophisticated (Gardner 2010).

Ambient monitoring provides valuable information for conservation planning and priority-setting, complementing knowledge of history, culture, political dynamics, and other contextual factors. Information regarding spatial patterns and temporal trends in the status of social and ecological conditions helps conservation decisionmakers to

56 chapter II identify locations for future conservation interventions, to set priorities among these locations, and to set management targets for these sites (Gardner et al. 2008; Stephanson and Mascia in review; Yoccoz et al. 2001). Ambient monitoring data also helps decisionmakers to identify socially and ecologically appropriate interventions in a given location (Funder et al. in review; Stephanson and Mascia in review). Ambient monitoring may also provide the data required to explore socio-ecological relationships and, as the social and ecological context for conservation shifts, foster adaptive management (Stephanson and Mascia in review). Thus, in a landscape context, ambient monitoring data from a national census and an ecoregional forest monitoring program could help conservation decisionmakers to (a) identify priority sites for conservation interventions; (b) choose among potential strategies (e.g., national parks v. community forests) in these locations; (c) understand the dynamic relationship between human populations and forest cover; and (d) revisit conservation strategies as human populations and forest cover change with time.

Ambient monitoring provides information relevant to decision making across scales, informing senior decisionmakers and local resource users alike. However, ambient monitoring can be difficult or expensive to undertake (Gardner et al. 2008; Nichols and Williams 2006; Stem et al. 2005). Locally-based ambient monitoring schemes, when properly designed and carefully tailored to local issues, provide a low-cost alternative that simultaneously builds capacity among local constituents (Danielsen et al. 2005). Such locally-based monitoring schemes may prompt practical and effective management interventions by increasing the speed of decision-making and by providing information to address environmental problems at operational scales of management (Danielsen et al. 2010). More generally, novel methods of participatory monitoring (e.g., crowd-sourcing data) present opportunities for through citizen science to expand monitoring of social and ecological conditions (Dickinson et al. 2010).

Management assessment Focal question: What are the management inputs, activities, and outputs associated with a conservation intervention, and how are these changing over time?

Management assessment is the process of measuring the management inputs, activities, and outputs associated with a conservation intervention, in order to identify management strengths, weaknesses, and needs (e.g., NOAA 2011). Management assessments are not linked to specific performance goals or an explicit program logic, but are instead predicated on the assumption that conservation interventions with sufficient management capacity and appropriate activities are more likely to deliver positive conservation outcomes than interventions with low capacity and misaligned activities (Ervin 2003; Leverington et al. 2010a). Thus, management assessments allow one to know if an intervention is “well-managed” (i.e., has a robust management presence on the ground) or exists solely on paper (e.g., “paper park”). Management assessments originated in the late 1990s, when it became clear that (a) declaration of

57 chapter II protected areas did not necessarily result in adequate management inputs, and (b) biodiversity was declining, despite the increasing number and spatial extent of protected areas (Ervin 2003; Hockings and Phillips 1999). Today, management assessments are still primarily employed by governments and international organizations to assess protected areas and protected area systems (NOAA 2011; Stolton et al. 2007), though this approach is applicable to other conservation interventions. (Management assessment is distinct from “management effectiveness evaluation” and its associated tools, though data generated by the latter are often used to assess the adequacy of management inputs, activities, and outputs [see Discussion]).

Management assessments vary in complexity, but the most commonly used methods are relatively fast, simple, and inexpensive to implement (e.g., NOAA 2010, 2011). Management assessments often employ a standardized, self-administered questionnaire to measure intervention inputs (e.g., funding, personnel), activities (e.g., enforcement, boundary demarcation), and outputs (e.g., management plans, regulations) (Ervin 2003; NOAA (National Oceanic and Atmospheric Administration) 2011; Stolton et al. 2007). Project managers usually complete the self-administered questionnaires and then compile the results using a standardized scorecard; information requirements for such management assessments are typically modest, and largely rely on (a) accessible information that is available on site and (b) the knowledge of those undertaking day-today management (Cook and Hockings 2011). Because management assessments are generally self-administered by project managers, scholars have questioned the validity and comparability of the resultant data (Cook and Hockings 2011; Stoll-Kleemann 2010). At the same time, self-assessment are perhaps more likely to foster clarification of management objectives, use of qualitative data that might otherwise be overlooked, and the integration of results into management decisions (Cook and Hockings 2011).

To promote public reporting and transparency, many donors include management assessment as a mandatory component of protected area M&E (e.g., Global Environment Facility, World Bank, Critical Ecosystem Partnership Fund). Management assessments have been used in more than 6,200 protected areas around the world (Leverington et al. 2010a), and are increasingly being used to evaluate national and international management and conservation strategies (e.g., Butchart et al. 2010; Pavese et al. 2007; Quan et al. 2011). Despite these M&E investments, it remains unclear whether “well-managed” interventions lead to more successful conservation, since (a) management assessments do not directly measure biodiversity or human well- being, and (b) researchers have not yet widely tested the assumption that protected area inputs, activities, and outputs foster positive conservation impacts (but see Nolte and Agrawal 2013; Nolte et al. in press).

Performance measurement Focal question: To what extent is a conservation intervention making progress toward its specified objectives for activities, outputs, and outcomes?

58 chapter II

Performance measurement is the process of measuring progress toward specified project, program, or policy objectives, including desired levels of activities, outputs, and outcomes (DAC 2002). Sometimes referred to as “performance monitoring” (Rich 1998) or “performance evaluation” (USAID 2011), performance measurement rose to prominence in the 1980s and early 1990s, as governments and private sector actors responded to a perceived need for greater accountability regarding the performance of public and private sector program investments in education, public health, social services, and other fields (Rich 1998; Wholey 1997). The conservation sector was a relative latecomer to performance measurement, with concerted efforts widely implemented only since the 1990s (Stem et al. 2005). Government agencies, nongovernmental organizations (NGOs), and academia subsequently developed numerous performance measurement methodologies (e.g., Kapos et al. 2008; Margoluis and Salafsky 1998; UNDP/GEF 2005). Though the term “performance measurement” has sometimes been used interchangeably with “performance-based,” “results-based,” or “outcomes-based” management, we follow Wholey (1996) in recognizing performance measurement as a mechanism to provide information necessary for management (i.e., decisionmaking).

Performance measurement tracks the progress of a single project or program over time against intended levels of accomplishment, in order to provide managers, organizational leadership, donors, and the public with information about project or program performance. Indicators are defined to track progress toward both interim objectives (e.g., activities completed, policies changed, threats diminished) and ultimate objectives or goals (e.g., improved status of a species or ecosystem, enhanced human well-being). Tensions exist among rigor, simplicity, and cost-effectiveness when identifying data collection methods and analyses for performance measurement (Margoluis et al. 2009), particularly for outcome indicators that may not be reliably measured via expert judgment or existing secondary sources. Because performance measurement usually relies heavily upon existing information and expert judgment, it is often relatively inexpensive, and does not require specialized expertise or complex research design.

Performance measurement is widely applied among governmental and nongovernmental conservation organizations to monitor project progress, inform adaptive management, and foster reporting and accountability. Performance measurement thus provides useful information to managers, especially where financial and human capacities are limited. The findings derived from performance measurement are overstated, however, by those who would attribute observed changes in outcome indicators to intervention activities and outputs (Ferraro and Pattanayak 2006). In fact, because performance measurement approaches do not compare the outcomes of an intervention to a valid counterfactual that isolates the effects of an intervention from the effects of other causal factors (e.g., by monitoring nonintervention comparison groups), performance measurement cannot definitively attribute observed changes in outcome variables to intervention activities and outputs (Rossi et al. 2004). Nonetheless, if indicators are clearly defined from the

59 chapter II outset, performance measurement provides valuable information regarding the progress of a project or program toward its stated goals and objectives.

Impact evaluation Focal question: What intended and unintended impacts are causally induced by a conservation intervention?

Impact evaluation is the systematic process of assessing the causal effects of a project, program, or policy (Gertler et al. 2011). By comparing what actually happened with an intervention to what would have happened without it (i.e., the counterfactual), impact evaluations measure the intended and unintended consequences attributable to a (conservation) intervention (Gertler et al. 2011). In addition to providing evidence regarding positive and negative impacts, well-designed impact evaluations may provide insights into the variation in impacts within and among groups, the attributes of an intervention that foster positive (or negative) impacts, and the contexts in which an intervention is most likely to succeed (or fail) (Miteva et al 2012). Impact evaluations often employ experimental research designs (i.e., random assignment to treatment and non-treatment groups) or quasi-experimental research designs (i.e., statistical identification of appropriate comparison groups); differences in observed changes between the treatment group and non-treatment comparison group represent the impact of the intervention (Rossi et al. 2004). Other research designs are also employed in impact evaluation (e.g., statistical analyses of observational data, in-depth case studies), “though the credibility of their estimates of program effects relies on how well the studies’ designs rule out competing causal explanations” (GAO 2009, p. 1). Impact evaluation has a strong grounding in the field of economics, with widespread use in the health, education, and development sectors (Patton 2008). Interest in impact evaluation per se emerged within the conservation sector in the mid-2000s (e.g., Ferraro and Pattanayak 2006), spurring an increasing number of impact evaluations from government agencies, NGOs, and universities (Miteva et al. in press). Recent impact evaluations have examined the impacts of protected areas on forest fires (Nelson and Chomitz 2011), recovery planning on the status of endangered species (Bottrill et al. 2011), and communal conservancies on human well-being (Glew et al. in review).

Impact evaluations inform decisions associated with curtailing, reforming, and expanding (conservation) initiatives (Gertler et al. 2011). Accordingly, the use of impact evaluations is most appropriate with fully implemented programs or interventions, where the goals and activities of the initiative have been defined, and where potential users of the evaluation are identified and prospects for use are strong (GAO 2009). For emerging or contested interventions, where the theory of change that links interventions to impacts remains unproven, impact evaluation may have particularly high policy relevance and prospects for use by decisionmakers (Gertler et al. 2011; Patton 2003, pp. 219-220). Experimental and quasi-experimental impact evaluations are most easily employed (a) when it is possible to maintain separation

60 chapter II between the treated and untreated populations; (b) where the causal factors that link interventions to impacts are understood and can be measured; and (c) where a good counterfactual can be identified and necessary data obtained (Ferraro 2009; GAO 2009). Since real-world conservation settings often possess unique or rare characteristics, which vary by setting and intervention, opportunities for effective use of experimental and quasi-experimental approaches may be limited (Ferraro 2009; Margoluis et al. 2009). In addition, impact evaluation – especially experimental and quasi-experimental methods – requires substantial expertise, can be costly to implement, and is not always possible for ethical, logistical, or legal reasons (Ferraro and Pattanayak 2006; GAO 2009). Given these constraints, impact evaluation is best employed selectively, in appropriate situations where the additional rigor of impact evaluation is required to inform major policy or program decisions (GAO 2009; Gertler et al. 2011).

Systematic review Focal question: What is the state of the evidence for the impacts of a conservation intervention and what does this evidence say about intervention impacts?

Systematic review is a structured process that collates, appraises, and synthesizes all available empirical evidence of relevance to a specific research question (Pullin & Knight 2009), facilitating the conversion of scientific knowledge into (conservation) practice (Pullin and Knight 2001, 2009; Stevens and Milne 1997). At its simplest, evidence synthesis can take the form of a literature review of two or more studies, but the gold standard methodology is systematic review, which attempts to maximize transparency and objectivity (minimizing bias) in capturing and critically appraising all relevant studies. Systematic review for evidence synthesis began in the 1980s, when clinical medicine sought to interpret and translate the findings from many individual studies to inform decisions on which interventions might be most effective for any given medical problem (Pullin and Knight 2001). First proposed for conservation by Pullin and Knight (2001), the approach is now relatively widespread in the health sector and is also used in the fields of education, social services and environmental management and international development (Petticrew 2001). Recent systematic reviews in conservation have examined the evidence for engineered structures on salmonid abundance (Stewart et al. 2009a), community forest management on biodiversity and poverty (Bowler et al. 2012), and the conservation impacts of temperate marine protected areas (Stewart et al. 2009b).

Systematic review contributes to a shared evidence base for decisionmakers, addressing questions concerning whether an intervention works or not, and its degree of impact (Pullin and Knight 2001, 2009). Systematic reviews characterize the state of the evidence by gathering together and describing diverse sets of data generated by studies with contrasting designs, which is particularly useful for achieving consensus when studies have conflicting findings. Systematic review is, thus, most useful when (a) an intervention has been widely and independently applied to achieve a specific goal and

61 chapter II

(b) multiple well-designed studies of the intervention’s effectiveness or impact exist. Systematic review is less suitable where studies and data are few, since data limitations increase the risk of arriving at null or faulty conclusions. In the absence of a robust scientific literature, systematic reviews can highlight knowledge gaps and inform research priorities (e.g., Bowler et al. 2012).

Systematic reviews are normally conducted by an independent review team that combines subject matter experts with review and synthesis experts (CEE 2010). Systematic reviews require significant resources, time, and technical expertise (CEE 2010). Though systematic reviews are standard in other sectors, they have not been widely used in conservation, despite their potential (but see www.environmentalevidence.org). Moreover, processes are needed to integrate evidence from systematic reviews into useful policy guidance (Segan et al. 2011). Government agencies and NGOs are beginning to commission systematic reviews to help meet their evidence needs and inform decision making (e.g., Bowler et al. 2012). With an increase in the number of studies that measure the impacts of conservation interventions, opportunities for systematic review and its application to conservation policy will grow commensurately.

Discussion Conservation M&E in theory Though ambient monitoring, management assessment, performance measurement, impact evaluation, and systematic review share some methodological commonalities, these five approaches to conservation M&E address distinct questions and inform distinct policy decisions (Table 2). These five approaches are employed at different points in the project/program cycle (CMP2007); examine different aspects of the drivers, pressures, states, impacts, and responses that constitute the relationship between people and the environment (Smeets and Weterings 1999); and assess distinct components of an intervention (W.K. Kellogg Foundation 2004; Figure 1). These complementarities can lead to productive synergies, such as when ambient monitoring data are resampled and repurposed to document impacts through impact evaluation, or when impact evaluation and perhaps performance measurement provide the evidence necessary for systematic reviews.

The five approaches we examined do not represent the full range of approaches to conservation M&E. Impact assessment – the ex ante estimation of likely ecological and social impacts of proposed future public or private sector activities – is widely employed in conservation and environmental management (Stem et al. 2005). Similarly, both spatial and strategic conservation planning (CMP2007; Margules and Pressey 2000; Pressey and Bottrill 2009) sometimes act as forms of formative evaluation (i.e., evaluative activities intended to guide program improvement (Rossi et al. 2004; Scriven 1991)). Needs assessment is an additional, distinct approaches to M&E (Rossi et al. 2004), though it is less commonly applied in conservation.

62 chapter II

Similarly, “management effectiveness evaluation” is a concept widely discussed and applied by conservation practitioners (Hockings et al. 2006; Leverington et al. 2010b). As defined by Hockings et al. (2006, pp. 12-13), “management effectiveness evaluation” examines the context, planning, inputs, processes (i.e., activities), outputs, and outcomes of a protected area, a framework for evaluating protected areas that largely mirrors conventional definitions of program evaluation writ large. Examining the site-specific context for an intervention, for example, is characteristic of needs assessment; intervention inputs, activities, and outputs are the typical focus of management assessment, performance measurement, and formative evaluation; and outcomes are principally the domain of impact evaluation. The primary focus of “management effectiveness evaluation” is rarely effectiveness in the conventional sense of the term (i.e., outcomes or impacts; Table 3) (Cook and Hockings 2011; Leverington et al. 2010b, p. 3).

Figure 1. Relationship between ambient monitoring, management assessment, performance measurement, impact evaluation, and systematic review and three common frameworks for conservation planning and analysis. The Drivers-Pressures-State-Impact-Response (DPSIR) framework (a) elucidates the relationships among Drivers, human activities that exert Pressure on the environment and, as a consequence, may change the State of the environment and society; these Impacts may elicit societal Responses that address these factors (Smeets & Weterings 1999). The conservation project cycle (b) begins with an initial conceptualization phase, followed by planning, implementation, analysis and adaptation, and learning (CMP 2007). The program logic model (c) articulates the hypothesized causal links between project or program inputs, activities, outputs, outcomes, and impacts (W.K. Kellogg Foundation 2004).

63 chapter II

Additional concepts from the field of program evaluation provide further clarity to our understanding of conservation M&E. The academic literature on program evaluation distinguishes between formative and summative evaluation, emphasizing not only what is measured but also the relationships among actors and the means by which findings are interpreted and used.

Formative evaluation is the systematic examination of the ongoing success of an intervention and its processes, which can provide information, insights, and advice about how these can be improved (Rossi et al. 2004; Scriven 1991). Formative evaluation is generally conducted throughout an intervention, in order to assure continual improvement in efficacy, relevance, logic, and efficiency, and to facilitate ongoing adjustments as the intervention matures (Scriven 1991). Data collection methods include literature reviews, focus groups, structured surveys, interviews, and direct observation of inputs, activities, outputs, and outcomes. Management assessment and performance measurement address aspects of formative evaluation, particularly those aspects that involve monitoring inputs, activities, and outputs and modifying activities accordingly to improve program performance. In some situations, impact evaluation may serve as a form of formative evaluation (Scriven 1991).

Effective formative evaluation involves decision-makers and stakeholders to ensure that the evaluation is salient and legitimate, and that the findings and recommendations feed into the program cycle (Patton 2003; Scriven 1991). Though management assessment and performance measurement provide useful insights for program improvement, they do not answer important questions such as “What is the relevance of this intervention to stakeholders?” or “How do we promote uptake and incorporation of recommendations?” Formative evaluation provides tools to engage stakeholders, complete the program feedback loop, and assure continued relevance of the intervention.

Summative evaluation judges the merit and worth of an intervention (Scriven 1996). Summative evaluation typically occurs in the final stages of an intervention, to decide whether to continue, discontinue, or expand the intervention (Scriven 1991). Summative evaluation uses many methods, including empirical experiments, cost- beneft analysis, case studies, interrupted time series, program theory methods, and meta-analysis. Whereas impact evaluation and systematic review similarly document impacts attributable to an intervention, these two approaches to conservation M&E do not generally examine questions related to the value of these impacts to stakeholders and decisionmakers. Conservation M&E in practice The practice of conservation M&E starts by defining the “research” question that will guide data collection, analysis, and use (Figure 2). The subject of evaluation, the type of knowledge desired, and the intended use for results all inform the development of evaluative questions (Rossi et al. 2004).

Table 2. Commonalities and complementarities among five approaches to conservation monitoring and evaluation. Ambient monitoring Management assessment Performance measurement Impact Evaluation Systematic Review What are the management To what extent is a What is the state of ambient What is the state of the inputs, activities, and outputs conservation intervention What intended and social and/or environmental evidence for the impact of an associated with a making progress toward its unintended impacts are Focal question conditions, and how are intervention, and what does conservation intervention, intended objectives for causally induced by a these conditions changing this evidence say about and how are these changing activities, outputs, and conservation intervention? over time and space? intervention impacts? over time? outcomes? Varies; often pre- During and after Post-implementation, with Timing During implementation Post-implementation intervention implementation pre-implementation baseline Any; often state/province, Multiple projects or one or (social), landscape, One or more interventions, more programs, with Multiple projects, programs, Scale Single project or program ecoregion (ecological), or usually protected areas corresponding noninter- or policies country (both) vention comparison group

Project and program Professional researchers, Professional researchers and Implementer managers, government Project managers Professional researcher citizen volunteers evaluators chapter II agencies

65 Adaptive management of Setting priorities among Program reporting & Selecting an intervention; Spatial and temporal existing and future potential capacity-building accountability assessments; scaling up or scaling down Decisions supported priority-setting, selection of interventions, scaling up or investments at one or more Adapt activities & strategies investments in said strategies and objectives down future investments in projects to enhance performance intervention said intervention

Project and program Project & program Project and program Project and program Decision-makers at local to Practitioner audience managers, donors, senior managers, donors, senior managers, senior managers, senior decision- global levels decisionmakers decisionmakers decisionmakers, donors makers, donors

Primary data collection or Primary data collection; manipulation of secondary remote sensing, transects Expert judgment, secondary Expert judgment, secondary source data; remote sensing, Data extraction from Data collection methods (ecological); household sources sources transects (ecological); secondary sources surveys, focus groups household surveys, focus (social) groups, interviews (social) Simple to moderate; may Complex; requires data Moderate to complex; Moderate to complex; may Simple; requires scoring require statistical management and requires sophisticated data Data analysis require data processing and self-administered manipulation of secondary sophisticated statistical extraction and statistical statistical analyses questionnaires source data analyses analyses Table reflects most common characteristics of each approach, but exceptions do occur and, in practice, scholars and practitioners sometimes mix or integrate approaches. chapter II

Once the question has been defined, it is possible to identify the appropriate M&E approach and, subsequently, the appropriate scope, methods, and tools. These choices will also be influenced by the evaluators’ information needs, as well as resource and timing limitations. Matching the appropriate M&E approach to the evaluation question, and context, fosters clear expectations about what will be measured, how it will be measured, and the insights for conservation policy and practice that are likely to emerge as a result.

Table 3. Measurement foci of two assessment tools commonly associated with “management effectiveness evaluation:” Management effectiveness tracking tool (METT) and rapid assessment and prioritization of protected areas management (RAPPAM). Assessment tool Object of inquiry # Questions % Total

METTa Context 1 3.3% Planning 7 23.3% Inputs 8 26.7% Process 11 36.7% Outputs 1 3.3% Outcomes 2 6.7% Total 30 100.0%

RAPPAMb Background information 8 5.3% Pressures & threats 24 15.8% Context 20 13.2% Vulnerability 10 6.6% Planning 15 9.9% Inputs 20 13.2% Processes 15 9.9% Outputs 10 6.6% Protected area system level 10 6.6% Protected area policies 10 6.6% Policy environment 10 6.6% Total 152 100.0% a Based on categorization in Stolton et al. (2007). In addition, the METT includes approximately 30 background questions and 50 contextual questions about pressures and threats that are not considered part of the formal METT assessment process. b Based on categorization in Ervin (2003).

Ambient monitoring, management assessment, performance measurement, impact evaluation, and systematic review each address complementary questions and inform complementary policy decisions, but use of these five approaches to conservation M&E has varied. Ambient monitoring and performance measurement are ubiquitous, and management assessment is commonly applied to protected areas around the world. Impact evaluations and systematic reviews are growing in number, but remain uncommon (Miteva et al. in press; Segan et al. 2011). Why are some approaches used more widely than others?

Though an approach to M&E is ideally selected in accordance with the focal question, M&E is also influenced by information needs, financial resources, timing, and human capacity. In many conservation settings, resources, time, and expertise are often constrained (Bruner et al. 2004; Gardner et al. 2008; Nichols and Williams 2006).

66 chapter II

Figure 2. Decision tree highlighting common information needs in conservation, and the question and approach to monitoring and evaluation that can best respond to those information needs.

Management assessment and performance measurement, by design, are useful M&E approaches in resource-constrained contexts, providing valuable information quickly at a low cost, albeit with less rigor and certainty. Ambient monitoring, impact evaluation, and systematic review, on the other hand, require substantial time, expertise, and financial investments, which are not always readily available. Yet the longstanding tradition and diverse contributions of ambient monitoring (often beyond the conservation sector) frequently ensure continued investments despite its cost, time, and expertise requirements.

Incentives also influence the selection of an approach to conservation M&E. Donors and conservation decisionmakers often prioritize and provide financial resources for information gathered within the scope of a project (InterAction 2010), neglecting alternatives. As a result, M&E is often conducted only within the timeframe of an intervention (i.e., rarely post-project); with existing (and often few) resources; and with data only from within the project area (ignoring conditions at comparison sites outside of the intervention) (Ferraro and Pattanayak 2006). Program or project implementers may also perceive conservation M&E as an unwanted diversion of scarce resources, a threat to their activities, or simply as an exercise unlikely to add inform or advance conservation policy and practice (Ferraro and Pattanayak 2006). These factors currently limit opportunities for impact evaluations and subsequent systematic reviews.

67 chapter II

Conclusion As the conservation community enters its third decade since the Rio Convention of 1992, ambient monitoring, management assessment, performance measurement, impact evaluation, systematic review, and other approaches to conservation M&E each have complementary roles to play in advancing more informed conservation policies and practices. In the years ahead, greater human capacity, financial investments, and organizational incentives for conservation M&E will be required, especially to generate the impact evaluations and systematic reviews that will enable us to better understand what works, what doesn’t, and why. Knowledge alone, however, will not ensure that the conservation community replicates successes, reforms failures, and avoids repeating the mistakes of the past. Effectively addressing the enduring challenge of biodiversity conservation will require a transformation of conservation policy and practice, through integrated investments in conservation M&E that advance knowledge, inform policy, build evaluation capacity in the developed and developing world, and catalyze a culture of evidence-based decisionmaking.

Acknowledgements We thank K. Haisfield for her help compiling this diverse literature, and L. Glew, E. McKenzie, R. Krithivasan, the Editor, and two anonymous reviewers for valuable feedback. Financial support was provided by the Kathryn S. Fuller Science for Nature Fund.

Literature cited Bartlett, R. V. 1994. Evaluating Environmental Policy Success and Failure. Pages 167-187 in N. J. Vig, and M. E. Kraft, editors. Environmental Policy in the 1990s. Congressional Quarterly Press, Inc., Washington, DC. Bertram, J. G. 1865. The harvest of the sea: a contribution to the natural and economic history of the British food fishes. John Murray, London. Birnbaum, M., and P. Mickwitz. 2009. Environmental Program and Policy Evaluation: Addressing Methodological Challenges. New Directions for Evalution 2009:1-112. Bottrill, M. C., J. C. Walsh, J. E. M. Watson, L. N. Joseph, A. Ortega-Argueta, and H. P. Possingham. 2011. Does recovery planning improve the status of threatened species? Biological Conservation 144:1595-1601. Bowler, D. E., L. M. Buyung-Ali, J. R. Healey, J. P. G. Jones, T. M. Knight, and A. S. Pullin. 2012. Does community forest management provide global environmental benefits and improve local welfare? Frontiers in Ecology and the Environment 10:29-36. Brewer, G. D., and P. deLeon 1983. The Foundations of Policy Analysis. Dorsey Press, Homewood, Illinois. Bruner, A. G., R. E. Gullison, and A. Balmford. 2004. Financial costs and shortfalls of managing and expanding protected-area systems in developing countries. Bioscience 54:1119-1126. Butchart, S. H. M., M. Walpole, B. Collen, A. van Strien, J. P. W. Scharlemann, R. E. A. Almond, J. E. M. Baillie, B. Bomhard, C. Brown, J. Bruno, K. E. Carpenter, G. M. Carr, J. Chanson, A. M. Chenery, J. Csirke, N. C. Davidson, F. Dentener, M. Foster, A. Galli, J. N. Galloway, P. Genovesi, R. D. Gregory, M. Hockings, V. Kapos, J.-F.

68 chapter II

Lamarque, F. Leverington, J. Loh, M. A. McGeoch, L. McRae, A. Minasyan, M. H. Morcillo, T. E. E. Oldfield, D. Pauly, S. Quader, C. Revenga, J. R. Sauer, B. Skolnik, D. Spear, D. Stanwell-Smith, S. N. Stuart, A. Symes, M. Tierney, T. D. Tyrrell, J.-C. Vie, and R. Watson. 2010. Global Biodiversity: Indicators of Recent Declines. Science 328:1164-1168. CEE (Collaboration for Environmental Evidence). 2010. Guidelines for Systematic Review in Environmental Management. Collaboration for Environmental Evidence, University of Bangor, Bangor, Wales. Christensen, J. 2003. Auditing Conservation in an Age of Accountability. Conservation in Practice 4:12-18. CMP (Conservation Measures Partnership). 2007. Open Standards for the Practice of Conservation, Version 2.0. Page 39. Conservation Measures Partnership, Washington, DC. Cook, C. N., and M. Hockings. 2011. Opportunities for improving the rigor of management effectiveness evaluations in protected areas. Conservation Letters 4:372-382. DAC (Development Assistance Committee Working Party on Aid Evaluation). 2002. Glossary of Key Terms in Evaluation and Results Based Management. Organisation for Economic Co-operation and Development, Paris. DANIDA (Danish International Development Agency). 2006. Evaluation Guidelines. Ministry of Foreign Affairs of Denmark, Copenhagen. Danielsen, F., D. S. Balete, M. K. Poulsen, M. Enghoff, C. M. Nozawa, and A. E. Jensen. 2000. A simple system for monitoring biodiversity in protected areas of a developing country. Biodiversity and Conservation 9:1671-1705. Danielsen, F., N. Burgess, and A. Balmford. 2005. Monitoring Matters: Examining the Potential of Locally-based Approaches. Biodiversity and Conservation 14:2507-2542. Danielsen, F., N. D. Burgess, P. M. Jensen, and K. Pirhofer-Walzl. 2010. Environmental monitoring: the scale and speed of implementation varies according to the degree of peoples involvement. Journal of Applied Ecology 47:1166-1168. Dickinson, J. L., B. Zuckerberg, and D. N. Bonter. 2010. Citizen Science as an Ecological Research Tool: Challenges and Benefits. Annual Review of Ecology, Evolution, and Systematics 41:149-172. Ervin, J. 2003. Rapid Assessment and Prioritization of Protected Area Management (RAPPAM) Methodology. WWF, Gland, Switzerland. Ferraro, P. J. 2009. Counterfactual Thinking and Impact Evaluation in Environmental Policy. New Directions for Evaluations 122:75-84. Ferraro, P. J., and S. K. Pattanayak. 2006. Money for nothing? A call for empirical evaluation of biodiversity conservation investments. Plos Biology 4:482-488. Funder, M., F. Danielsen, Y. Ngaga, M. R. Nielsen, and M. K. Poulsen. in review. Reshaping conservation: The social dynamics of participatory monitoring in Tanzania’s community-managed forests. Ecology and Society. GAO (Government Accountability Office). 2009. Program Evaluation: A Variety of Rigorous Methods Can Help Identify Effective Interventions. Page 45. U.S. Government Accountability Office, Washington, DC. Gardner, T. A. 2010. Monitoring Forest Biodiversity: improving conservation through ecologically responsible management. Earthscan, London. Gardner, T. A., J. Barlow, I. S. Araujo, T. C. Ávila-Pires, A. B. Bonaldo, J. E. Costa, M. C. Esposito, L. V. Ferreira, J. Hawes, M. I. M. Hernandez, M. S. Hoogmoed, R. N. Leite, N. F. Lo-Man-Hung, J. R. Malcolm, M. B. Martins, L. A. M. Mestre, R. Miranda- Santos, W. L. Overal, L. Parry, S. L. Peters, M. A. Ribeiro-Junior, M. N. F. Da Silva, C.

69 chapter II

Da Silva Motta, and C. A. Peres. 2008. The cost-effectiveness of biodiversity surveys in tropical forests. Ecology Letters 11:139-150. Gertler, P. J., S. Martinez, P. Premand, L. B. Rawlings, and C. M. J. Vermeersch. 2011. Impact Evaluation in Practice. The World Bank, Washington, DC. Glew, L., M. D. Hudson, P. E. Osborne, J. King, T. Lalampaa, C. Leisher, and M. Rice. in review. Social outcomes of community conservation in northern Kenya. Proceedings of the National Academy of Sciences. Gratwicke, B., J. Seidensticker, M. Shrestha, K. Vermilye, and M. Birnbaum. 2007. Evaluating the performance of a decade of Save The Tiger Fund's investments to save the world's last wild tigers. Environmental Conservation 34:255-265. Hansen, M. C., S. V. Stehman, P. V. Potapov, T. R. Loveland, J. R. G. Townshend, R. S. DeFries, K. W. Pittman, B. Arunarwati, F. Stolle, M. K. Steininger, M. Carroll, and C. DiMiceli. 2008. Humid tropical forest clearing from 2000 to 2005 quantified by using multitemporal and multiresolution remotely sensed data. Proceedings of the National Academy of Sciences 105:9439-9444. Hobbs, F., and N. Stoops 2002. Demographic Trends in the 20th Century: Census 2000 Special Reports. U.S. Census Bureau, Washington, DC. Hockings, M., and A. Phillips. 1999. How well are we doing? Some thoughts on the effectiveness of protected areas. Parks 9:5-14. Hockings, M., S. Stolton, F. Leverington, N. Dudley, and J. Courrau 2006. Evaluating Effectiveness: A Framework for Assessing Management Effectiveness of Protected Areas. IUCN, Gland, Switzerland. InterAction. 2010. Evaluation and Program Effectiveness Working Group, Feedback and Recommendations on USAID's Monitoring and Evaluation Practices. InterAction, Washington, DC. Jepson, P. 2005. Governance and accountability of environmental NGOs. Environmental Science & Policy 8:515-524. Kapos, V., A. Balmford, R. Aveling, P. Bubb, P. Carey, A. Entwistle, J. Hopkins, T. Mulliken, R. Safford, A. Stattersfield, M. Walpole, and A. Manica. 2008. Calibrating conservation: new tools for measuring success. Conservation Letters 1:155-164. Leverington, F., K. Costa, H. Pavese, A. Lisle, and M. Hockings. 2010a. A Global Analysis of Protected Area Management Effectiveness. Environmental Management 46:685-698. Leverington, F., K. Lemos Costa, J. Courrau, H. Pavese, C. Nolte, M. Marr, L. Coad, N. Burgess, B. Bomhard, and M. Hockings. 2010b. Management effectiveness evaluation in protected areas - a global study. Second edition 2010. University of Queensland, Brisbane, Australia. Margoluis, R., and N. Salafsky 1998. Measures of Success: Designing, Managing, and Monitoring Conservation and Development Projects. Island Press, Washington, DC. Margoluis, R., C. Stem, N. Salafsky, and M. Brown. 2009. Design Alternatives for Evaluating the Impact of Conservation Projects. New Directions for Evalution 122:85-96. Margules, C. R., and R. L. Pressey. 2000. Systematic conservation planning. Nature 405:243- 253. Missiakoulis, S. 2010. Cecrops, King of Athens: the First (?) Recorded Population Census in History. International Statistical Review 78:413-418. Miteva, D. A., S. K. Pattanayak, and P. J. Ferraro. in press. Evaluation of Biodiversity Policy Instruments: What works and what doesn't? Oxford Review of Economic Policy.

70 chapter II

Mysterud, A., E. L. Meisingset, V. Veiberg, R. Langvatn, E. J. Solberg, L. E. Loe, and N. C. Stenseth. 2007. Monitoring population size of red deer Cervus elaphus: an evaluation of two types of census data from Norway. Wildlife Biology 13:285-298. Nelson, A., and K. M. Chomitz. 2011. Effectiveness of Strict vs. Multiple Use Protected Areas in Reducing Tropical Forest Fires: A Global Analysis Using Matching Methods. PLoS ONE 6:e22722. Nichols, J. D., and B. K. Williams. 2006. Monitoring for conservation. Trends in Ecology and Evolution 21:668-673. NOAA (National Oceanic and Atmospheric Administration). 2010. Marine Protected Area Management Assessment Checklist. U.S. National Oceanic and Atmospheric Administration, Washington, DC. NOAA (National Oceanic and Atmospheric Administration). 2011. User’s Guide for the NOAA Coral Reef Conservation Program MPA Checklist U.S. National Oceanic and Atmospheric Administration, Washington, DC. Oral History Project Team. 2007. The Oral History of Evaluation, Part 5: An Interview with Michael Quinn Patton. American Journal of Evaluation 28:102-114. Patton, M. Q. 2003. Utilization-Focused Evaluation. Pages 223-244 in T. Kellaghan, and D. L. Stufflebeam, editors. International Handbook of Educational Evaluation: Part One. Kluwer Academic Publishers, Norwood, MA. Patton, M. Q. 2008. Utilization-focused evaluation. Sage Publications, Thousand Oaks, California. Pavese, H. B., F. Leverington, and M. Hockings. 2007. Global Study of Protected Areas Management Effectiveness: the Brazilian perspective. Natureza & Conservacao 5:152- 162. Petticrew, M. 2001. Systematic reviews from astronomy to zoology: myths and misconceptions. BMJ 322:98-101. Pressey, R. L., and M. C. Bottrill. 2009. Approaches to landscape- and seascape-scale conservation planning: convergence, contrasts and challenges. Oryx 43:464-475. Pullin, A. S., and T. M. Knight. 2001. Effectiveness in Conservation Practice: Pointers from Medicine and Public Health. Conservation Biology 15:50-54. Pullin, A. S., and T. M. Knight. 2009. Doing more good than harm - Building an evidence-base for conservation and environmental management. Biological Conservation 142:931- 934. Quan, J., Z. Y. Ouyang, W. H. Xu, and H. Miao. 2011. Assessment of the effectiveness of nature reserve management in China. Biodiversity and Conservation 20:779-792. Rich, R. F. 1998. Program Evaluation and Environmental Policy: The State of the Art. Pages 23-41 in G. J. Knaap, and T. J. Kim, editors. Environmental Program Evaluation: A Primer. University of Illinois Press, Chicago. Rogers, E. M. 1995. Diffusion of Innovation. Free Press, New York. Rossi, P. H., M. W. Lipsey, and H. Freeman 2004. Evaluation: a systematic approach. Sage Publications, Thousand Oaks, California. Scriven, M. 1991. Evaluation thesaurus. Sage Publications, Newbury Park, CA. Scriven, M. 1996. Types of Evaluation and Types of Evaluator. American Journal of Evaluation 17:151-161. Segan, D. B., M. C. Bottrill, P. W. J. Baxter, and H. P. Possingham. 2011. Using conservation evidence to guide management. Conservation Biology 25:200-202.

71 chapter II

Smeets, E., and R. Weterings. 1999. Environmental indicators:Typology and overview. Technical report No 25. European Environment Agency, Copenhagen. Stem, C., R. Margoluis, N. Salafsky, and M. Brown. 2005. Monitoring and Evaluation in Conservation: A Review of Trends and Approaches. Conservation Biology 19:295-309. Stephanson, S. L., and M. B. Mascia. in review. Putting People on the Map: An Approach to Integrating Social Data in Conservation Planning. Conservation Biology. Stevens, A., and R. Milne. 1997. The effectiveness revolution and public health. Progress in public health. Royal Society of Medicine Press, London:197-225. Stewart, G. B., H. R. Bayliss, D. A. Showler, W. J. Sutherland, and A. S. Pullin. 2009a. Effectiveness of engineered in-stream structure mitigation measures to increase salmonid abundance: a systematic review. Ecological Applications 19:931-941. Stewart, G. B., M. J. Kaiser, I. M. Côté, B. S. Halpern, S. E. Lester, H. R. Bayliss, and A. S. Pullin. 2009b. Temperate marine reserves: global ecological effects and guidelines for future networks. Conservation Letters 2:243-253. Stoll-Kleemann, S. 2010. Evaluation of management effectiveness in protected areas: Methodologies and results. Basic and Applied Ecology 11:377-382. Stolton, S., M. Hockings, N. Dudley, K. MacKinnon, T. Whitten, and F. Leverington. 2007. Management Effectiveness Tracking Tool: Reporting Progress at Protected Area Sites: Second Edition. WWF International, Gland, Switzerland. Sutherland, W. J., A. S. Pullin, P. M. Dolman, and T. M. Knight. 2004. The need for evidence- based conservation. Trends in Ecology & Evolution 19:305-308. Ticheler, H., J. Kolding, J. , and B. Chanda. 1998. Participation of local fishermen in scientific fisheries data collection: A case study from the Bangweulu Swamps, Zambia. Fisheries Management and Ecology 5:81–92. UNDP/GEF. 2005. Measuring and demonstrating impact. UNDP/GEF Resource Kit (No.2). United Nations Environment Department/Global Environment Facility, Washington, D.C. USAID (U.S. Agency for International Development) 2011. USAID Evaluation Policy. U.S. Agency for International Development, Washington, DC. W.K. Kellogg Foundation. 2004. Using Logic Models to Bring Together Planning, Evaluation, and Action: Logic Model Development Guide. W.K. Kellogg Foundation, Battle Creek, Michigan. Weiss, C. H. 1995. Nothing as Practical as Good Theory: Exploring Theory-Based Evaluation for Comprehensive Community Initiatives for Children and Families. Pages 65-92 in J. I. Connell, A. C. Kubisch, L. B. Schorr, and C. H. Weiss, editors. New Approaches to Evaluating Community Initiatives: Concepts, Methods, and Contexts. Aspen Institute, Washington, DC. Wholey, J. 1996. Formative and Summative Evaluation: Related Issues in Performance Measurement. American Journal of Evaluation 17:145-149. Wholey, J. S. 1997. Trends in Performance Measurement: Challenges for Evaluators. Pages 124-133 in E. Chelimsky, and W. R. Shadish, editors. Evaluation for the 21st Century: A Handbook. SAGE Publications, Inc., Thousand Oaks, California. Yoccoz, N. G., J. D. Nichols, and T. Boulinier. 2001. Monitoring of biological diversity in space and time. Trends in Ecology & Evolution 16:446-453. ZIMSTAT (Zimbabwe National Statistics Agency), and ICF International. 2012. Zimbabwe Demographic and Health Survey 2010-11. ZIMSTAT and ICF International, Inc., Calverton, Maryland.

CHAPTER III Management effectiveness and global commitments

Lauren Coad, Fiona Leverington, Neil D. Burgess, Ivon C. Cuadros, Jonas Geldmann, Toby Marthews, Jessie Mae, Christoph Nolte, Susanne Stoll- Kleemann, Nanna Vansteelant, Camilo Zamora, Mark Zimsky and Marc Hockings

Published in Parks, 2013, volume 19, issue 1

74 chapter III

PARKS 2013 Vol 19.1

PROGRESS TOWARDS THE CBD PROTECTED AREA MANAGEMENT EFFECTIVENESS TARGETS

Lauren Coad1,2,3*, Fiona Leverington2, Neil D. Burgess3, Ivon C. Cuadros2, Jonas Geldmann4, Toby R. Marthews1, Jessie Mee5, Christoph Nolte6, Susanne Stoll-Kleemann7, Nanna Vansteelant4, Camilo Zamora2, Mark Zimsky8, and Marc Hockings2,3

* Corresponding author, [email protected] 1 Environmental Change Institute, School of Geography, University of Oxford, OX1 3QY, UK 2 School of Geography, Planning and Environmental Management, University of Queensland, Brisbane, QLD 4072, Australia 3 The United Nations Environment Programme – World Conservation Monitoring Centre, Cambridge. UK 4 Center for Macroecology, Evolution and Climate, Department of Biology, University of Copenhagen, Denmark 5 Environment and Energy Group, United Nations Development Programme, New York, USA 6 School of Natural Resources and the Environment, University of Michigan, Ann Arbor, MI 48109, United States 7 Institute for Geography and Geology, University of Greifswald, D-17487 , Germany 8 Natural Resources Division, The Global Environment Facility, Washington DC, USA

ABSTRACT The management effectiveness of protected areas is a critically important consideration for their conservation success. Over 40 different protected area management effectiveness (PAME) data collection tools have been developed to systematically assess protected area management effectiveness. Many of these assessments have recently been collated into the Global IUCN Protected Area Management Effectiveness (PAME) database. We use the PAME database together with and the World Database on Protected Areas (WDPA) to assess current progress towards the Convention on Biological Diversity’s (CBD) 2010 and 2015 targets for PAME, which call for at least 30 per cent and 60 per cent of the total area of protected areas to have been assessed in terms of management effectiveness, respectively. We show that globally 29 per cent of the area protected has been assessed and 23 per cent of countries have reached the 60 per cent target. In addition 46 per cent of countries have reached the 30 per cent target. However, analytical results show that there are biases in the type of protected area assessed; protected areas with larger areas, and protected areas designated as National Parks (IUCN category II) are much more likely to have conducted a PAME assessment. In addition there is a paucity of PAME assessments from Europe and North America, where assessments of protected area management may already be integrated into protected area planning and monitoring systems, creating a challenge for reporting to the CBD. We further discuss the potential and limitations of PAME assessments as tools for tracking and evaluating protected area management, and the need for further assessment tools to address the ‘equity’ elements of Target 11 of the CBD.

INTRODUCTION By 2020, at least 17 per cent of terrestrial and inland Protected areas have long been regarded as an important water, and 10 percent of coastal and marine areas, tool for biodiversity conservation (e.g. WCED, 1987), and especially areas of particular importance for are used as indicators of progress in the protection of biodiversity and ecosystem services, are conserved biological diversity by a number of international through effectively and equitably managed, agreements, including the Convention on Biological ecologically representative and well connected systems Diversity (CBD). The CBD Aichi Biodiversity Targets, of protected areas and other effective area-based agreed on by Parties to the Convention in October 2010, conservation measures, and integrated into the wider include the following target for protected areas: landscapes and seascapes’. Target 11, CBD (emphasis (www.cbd.int/sp/targets/): added).

PARKS VOL 19.1 MARCH 2013

75 chapter III

Lauren Coad et al 2

This new Aichi target was developed from the earlier their management regime. In some cases assessments CBD Target 1.1 (set in 2003), which called for: ‘at least 10 are undertaken in response to donor requirements per cent of each of the world's ecological regions [to be] associated with project support for a protected area or as effectively conserved’. Target 1.1, Decision VII/30, CBD part of an NGO sponsored assessment and improvement project (Hockings et al., 2004a, Leverington et al., Analyses of progress towards Target 1.1 have to date 2010b). Assessments are also undertaken in response to tended to measure protected area coverage (Chape et al., central government requirements to monitor and report 2005, Coad et al., 2008, Spalding et al., 2008, Coad et on protected area management (e.g. NSW Audit Office, al., 2009a, Coad et al., 2009b, Jenkins and Joppa, 2009) 2004, Auditor General of Queensland, 2010). In 2000, and ecological representativeness (Rodrigues et al., the IUCN World Commission on Protected Areas 2004, Spalding et al., 2007, Schmitt et al., 2009, Herbert (WCPA) developed an overarching framework to guide et al., 2010) facilitated by the availability of open-access assessment of management effectiveness that has been global datasets on protected area locations (e.g. The widely used around the world (Hockings et al., 2000, World Database on Protected Areas – WDPA) and global Hockings et al., 2006). According to this framework, the frameworks of ecological regions and key areas for evaluation of management effectiveness can be carried biodiversity (Olson et al., 2001, Eken et al., 2004). In out for a variety of reasons, including providing better terms of global protected area coverage, Parties have management in a changing environment, effective made significant progress towards achieving Target 1.1 resource allocation, improved accountability and for terrestrial biodiversity: over 50 per cent of terrestrial transparency, community involvement, and promotion of ecoregions have 10 per cent or more of their area within protected area values. protected areas, although marine ecosystems are still severely under-represented (Spalding et al., 2008, Coad The WCPA framework was developed to provide overall et al., 2009b). guidance for the evaluation of management, the selection of appropriate indicators and the analysis and However, protected area coverage alone is not a application of assessment results. It has been used to sufficient indicator for meeting global biodiversity develop over 40 different protected area management targets. There has been a growing concern amongst effectiveness (PAME) data collection tools to protected area managers and conservation scientists that systematically assess protected area management many protected areas around the world are not achieving effectiveness at the individual protected area level and at the conservation objectives for which they were a national system level (Leverington et al., 2010a; also established, because of a lack of effective management see www.wdpa.org/me). (Hockings et al., 2004b, Dudley & Stolton, 2009). In response to this concern, in 2004 the CBD established A global study into management effectiveness evaluation the Programme of Work on Protected Areas (PoWPA) was launched in late 2005 and completed in 2010 and set a preliminary global target for 30 per cent of the (Leverington et al., 2008, Leverington et al., 2010a, world’s protected areas to have assessed the effectiveness Nolte et al., 2010). The aim of the study was to obtain a of their management by 2010 (Goal 4.2, CBD PoWPA) global picture of protected area effectiveness and to track (see Coad et al., 2009). This targeted was updated at the CBD targets and reporting needs on behalf of the CBD’s COP 10, when addition to introducing the call for international conservation community. To achieve this ‘effective and equitable’ management of protected areas aim, all existing PAME assessments were collated into a in Target 11, the CBD Aichi targets expanded the single database. The resulting database has since been mandate for management effectiveness assessment. updated as part of a collaborative research effort between Inviting “...Parties to…expand and institutionalize the University of Queensland and the University of management effectiveness assessments to work towards Oxford, with inputs from various other NGO, assessing 60 per cent of the total area of government and intergovernmental partners 1. The protected areas by 2015 using various national and database contains PAME assessments from 1991 to 2012. regional tools and report the results into the global There are likely to be recent assessments that have not database on management effectiveness…” CBD Aichi yet been located and added to the PAME database, Targets, COP 10 Decision X/31, 19a (emphasis added). despite the authors’ best efforts. However, we believe that as a result of the high level of outreach to protected Undertaking an assessment of management effectiveness area managers, donors, NGOs, government and allows conservation agencies to understand better their intergovernmental partners and the wider conservation strengths and weaknesses and to adapt and improve community during the Global Study, which has been

PARKS VOL 19.1 MARCH 2013

76 chapter III

3 www.iucn.org/parks

Gathering data for a management effectiveness assessment in Bwindi Impenetrable National Park and World Heritage Site in Uganda © Marc Hockings followed with regular updates from partners such as  WDPA IUCN, The World Wide Fund for Nature (WWF) and the We used the December 2012 version of the WDPA for Global Environment Facility (GEF), the majority of analysis (IUCN & UNEP, 2012). The WDPA is provided assessments up to 2010 are now contained in the as two separate GIS shapefiles: ‘WDPA polygons’ for database. protected areas where the boundary and shape of the protected area is known, and ‘WDPA points’ for In this paper we use the updated IUCN PAME database, protected areas where only the point location is known. together with the UNEP WCMC / IUCN WDPA (IUCN & Where sites only existed in the WDPA as a point location, UNEP, 2012), to conduct a spatial analysis of national we used the ‘buffer’ tool in ArcGIS to create a circular and global progress towards the ‘effectiveness’ element of polygon of the same size as the given area of the Aichi Target 11 and the PoWPA. We ask specifically protected area (as recorded in the WDPA), with the point whether countries have achieved the CBD 60 per cent location as its centroid. We then used the ‘Merge’ tool to Aichi target for management effectiveness assessments of add the buffered points to the existing WDPA polygon nationally designated protected areas. We then explore shapefile. We included protected areas with a the protected area characteristics that significantly designation status of ‘adopted’, ‘designated’, ‘inscribed’ predict whether a protected area has been evaluated. We and ‘not reported’, and excluded ‘proposed’ protected discuss the results in terms of the future work required to areas. All reserves with international designations measure progress toward the CBD Aichi Target for 17 per (World Heritage, Ramsar and Man and Biosphere) were cent of the world’s protected areas to be effectively and removed leaving only nationally designated reserves, as equitably managed. most international designations either duplicate national reserves or may not meet the requirements for full METHODS protected area status (selection of nationally designated  Data preparation areas has also been applied in previous analyses of All spatial analyses were carried out using the ESRI protected area coverage: see Jenkins & Joppa, 2009, and ArcGIS 10.1 programme (ESRI, 2012). We used the Schmitt et al., 2009, among others). The final version of Mollweide Equal Area projection for all analyses. Results the WDPA for analysis contained 168,054 nationally are displayed in the Robinson projection. designated protected areas, of which 12 per cent were

PARKS VOL 19.1 MARCH 2013

77 chapter III

Lauren Coad et al 4

Figure 1: The location of protected areas that have conducted a PAME assessment. Marine and terrestrial nationally designated protected areas are included. buffered points. Where detailed polygons in the ‘WDPA  Calculating assessed area per country

polygon’ shapefile exist, this results in large numbers of GIS overlay analyses (assessments with WDPA vertices in the shapefile, which can produce ID): We followed the analyses steps outlined by Bubb et geoprocessing errors during analysis. To avoid these al. (2008) for global protected area coverage analyses. errors we used the ArcGIS ‘Repair Geometry’ tool to We linked the WDPA shapefile with the list of assessed check and correct for any further geometry errors (ESRI, PAs, by WDPA ID, using the ‘join’ tool. From this, we 2012). then created a new shapefile of all assessed PAs. We used

the ‘dissolve’ tool to dissolve all assessed protected area  PAME data polygons within each country. We repeated this dissolve Management effectiveness assessments have been for the total WDPA. This resulted in two final shapefiles: systematically collated in the IUCN PAME database, one providing the total area of assessed nationally which is maintained and hosted by the University of designated protected areas (for those with WDPA IDs) Queensland (UQ). Data held in the database includes for each country, and a second providing the total area of protected area name, WDPA Unique Identifier (WDPA all nationally designated protected areas for each ID), year of assessment, methodologies, indicators and country. assessment tools used and, where available, assessment results. In this analysis we used all assessments entered Assessments without WDPA ID: The area (km2) of into the IUCN PAME database up until 30th November 2012. The November 2012 version of the PAME database assessed protected areas without a WDPA ID was holds 10,501 assessments for 6,741 sites. summed for each country, using the area of the protected area provided in the IUCN PAME database. This area In the IUCN PAME database, for each PAME assessment was then added to the total area of protected areas we recorded the WDPA ID for the appropriate national assessed for each country, and the total area of protected protected area record in the WDPA. For those areas for each country. In total, 232 nations were assessments where no WDPA ID existed we noted the assessed, using the International Organisation for area of the protected area in hectares, either from the Standardisation (ISO) 3166-1 A3 list to define nations. original PAME assessment, or from a reputable Dependent territories were added to their parent nations. government or NGO data source. We only included countries that had protected areas

PARKS VOL 19.1 MARCH 2013

78 chapter III

5 www.iucn.org/parks

Figure 2: Regional progress towards the CBD 30 per cent and 60 per cent targets for PAME assessments. Progress is measured by the percentage of the total area of the nationally designated protected area network that has been assessed in each region. recorded in the WDPA; Countries with no recorded predominantly established for strict biodiversity protected areas were excluded from the analyses. conservation, and those which allow for some level of sustainable use and/or human intervention. These  Calculating assessed area globally and per groupings have previously been used in analyses of region protected area coverage (see Scharlemann et al., 2010 Countries were grouped into regions according to the and Joppa & Pfaff, 2011 for examples). We included UN United Nations geoscheme. The area of assessed and region and UN Human Development Index (HDI) as unassessed protected areas for countries within each regional and country-level predictors. region was summed to find the percentage of assessed area for each region. All statistical analyses were carried out using the R statistical package (R Development Core Team, 2012). Surprisingly, given the heterogeneity of the regions  Identifying predictors of PAME analysed, the data were not overdispersed (dispersion assessment parameter = 1) so no correction for this was necessary To identify which protected area characteristics (Gelman & Hill, 2007). significantly predict whether a PAME assessment had been carried out in a protected area, we used a generalized linear model (GLM) with a binomial error RESULTS: GLOBAL, REGIONAL AND NATIONAL structure (i.e. multivariate logisitic regression, Pinheiro PROGRESS TOWARDS THE 60 PER CENT AICHI & Bates, 2000). At the level of an individual protected TARGET area we were limited in our predictors to those with Globally, 29 per cent of the area of nationally designated characteristics that have been routinely documented by protected areas has been assessed for PAME. The the WDPA: area (in km2), IUCN management category location of assessed and unassessed protected area s is and year of establishment (converted into ‘age of shown in Figure 1. Regionally, Africa has assessed the protected area (years)’ for the purposes of these largest proportion by area (44 per cent). Latin America, analyses). We grouped IUCN categories (Dudley, 2008) Asia and Europe have also reached the 2010 CBD into two factor levels category I – II and III – VI, to PoWPA target of 30 per cent assessed (Figure 2). Oceania distinguish between protected areas which have been has not yet met the 30 per cent target, with 17 per cent of

PARKS VOL 19.1 MARCH 2013

79 chapter III

Lauren Coad et al 6

Figure 3: National progress towards the CBD 30 per cent and 60 per cent targets for PAME assessments. Progress is measured by the percentage of the total area of the nationally designated protected area network that has been assessed the protected area assessed. Northern America has the all sites assessed (Figure 6). There was also a significant least assessed area of all regions, with less than 3 per effect of protected area age (year of establishment) on the cent of its area assessed, according to PAME records probability of assessment, with younger protected areas currently held in the database. slightly less likely to have been assessed, although the effect was very small (Table 1). Protected areas in Nationally, 46 per cent of the countries listed (90 developing countries were more likely to be assessed countries in total) met the 2010 target of 30 per cent, than those in more developed countries, the frequency of with 23 per cent (45) already achieving the 60 per cent assessment declining significantly with increasing HDI target of 2015 (Figures 3 and 4). However, for 52 scores (Table 1). However, there were significant regional countries (26 per cent) no assessments have been biases in the results in addition to the differences in recorded in the PAME database. terms of development between nations. In relation to African protected areas in general, Latin American, PREDICTORS OF ASSESSMENT Caribbean and Oceanian protected areas were also more Wald test statistics, which indicate the relative weights of likely to have carried out a management assessment with the explanatory variables in the model, showed that the Asian, European and, especially Northern American, size of the protected area was the most significant protected areas were less likely. predictor of whether an assessment had been carried out; followed by IUCN protected area management category DISCUSSION (Table 1). Larger protected areas were significantly more In this paper we measured progress towards the CBD likely to have conducted a PAME assessment (Figure 5 2010 and 2015 PAME targets. The results of our analyses Table 1). Protected areas with an IUCN protected area are encouraging, suggesting that for over 23 per cent of management category of I - II were also significantly countries the 60 per cent target for 2015 has already more likely to have been assessed than protected areas been achieved, according to the PAME assessments with another management category, even when currently held in the database. A much higher proportion controlling for area (Table 1). National Parks (category (46 per cent) has achieved the 30 per cent target for II) had the highest assessment rate, with 30 per cent of 2010. In addition, we continue to receive data from a

PARKS VOL 19.1 MARCH 2013

80 chapter III

7 www.iucn.org/parks

Figure 4: The number of countries reaching the CBD 30 per cent and 60 per cent targets for PAME assessments

number of sources, including regular updates from the and Central Africa reaching the 60 per cent target. This is GEF, and hence the number of assessments in now likely to a large extent due to the strong efforts of IUCN in that to be greater than that held in the November 2012 region through the PAPACO project 2 (Leverington et al., version of the PAME database. 2010b), which has collated and conducted evaluations as part of a targeted programme. Latin America and Asia However, progress towards the targets is not evenly have also assessed a large proportion of their total spread across the globe. Africa has the highest protected areas by area. Additionally, protected areas percentage area assessed, with many countries in West were more likely to be assessed if they were from

Table 1: Parameter estimates of a Generalized Linear Model (GLM) with binomial error structure, showing the significant predictors of whether an individual protected area has conducted a management effectiveness assessment

Predictor Variables (minimal model) Estimate S.E. z p Intercept -2.42 0.19 -12.82 <0.001 Ln (protected area in km2) 0.96 0.02 48.62 <0.001 Protected area IUCN Category I - II 1.62 0.04 38.55 <0.001 Protected area age (years) 0.01 0.00 10.65 <0.001 Country Human Development Index -2.35 0.30 -7.87 <0.001 Region: Asia -0.82 0.11 -7.61 <0.001 Europe -1.48 0.14 -10.68 <0.001 Latin America and the Caribbean 0.77 0.12 6.57 <0.001 Northern America -3.64 0.25 -14.43 <0.001 Oceania 1.04 0.15 6.77 <0.001 Notes : N = 168,054, of which 4,922 protected areas (with WDPA ID) had a management effectiveness assessment. Reference level for UN Region is Africa, and for IUCN category is III – VI. Note that all these predictors were highly significant in the full model (p- values very close to zero), therefore no model selection step was required (Pinheiro & Bates, 2000), z values are Wald test scores showing the degree of association between the predictor and the probability of having had a management assessment (= square roots of χ2 statistics).

PARKS VOL 19.1 MARCH 2013

81 chapter III

Lauren Coad et al 8

Figure 5: Boxplot showing the median area (and IQ range) of assessed and unassessed protected areas. Median area of assessed protected areas = 74.7 km2, median area of unassessed protected areas = 0.30 km2 countries with a lower HDI score. The role of many large CBD via these intergovernmental organizations. For donor organizations, which predominantly work in example, at a national level, Canada undertakes developing countries, in carrying out PAME assessments assessments through their State of the Parks systems (Belokurov and Besancon, 2009) could partly explain and, where available, these assessments are included in this geographic bias in reported assessments. For the PAME database. example, all protected area targeting projects funded by the GEF since 2004 have been required to complete the These analyses did not consider the different Management Effectiveness Tracking Tool (METT) organizations undertaking PAME assessments, but this (Stolton et al., 2007) at least three times for each topic warrants further investigation. Although many targeted protected area. As the single largest source of PAME assessments may be carried out on a protected finance for biodiversity and ecosystem management area-by-protected area basis, in some countries globally, the GEF makes a significant impact in achieving assessments have been integrated into regional and PAME targets through this reporting requirement in national management of protected area systems (for partner developing and in-transition countries; more example, NSW DEC, 2005). The case of Australia, which than 300 protected areas in approximately 100 countries as a country has achieved the 30 per cent target (Figure around the world are currently required to regularly 1), clearly shows a regional difference in assessments, complete METTs in line with the GEF reporting with eastern Australia accounting for the majority of requirement. Australian assessments (of which the Great Barrier Reef assessment accounts for a significant area). In Victoria, Our results also show that only few assessments on New South Wales and Queensland, PAME assessments PAME have been undertaken for protected areas in have been adopted as a planning tool for state protected North America and Western Europe, despite a dedicated area management and are conducted every few years. effort, particularly for Europe (Nolte et al., 2010) to bring together all PAME information. This may not imply that As well as a geographical bias, we also found a bias in the these countries do not evaluate the effectiveness of their type of protected area being assessed. National Parks protected area networks; they may already have were much more likely to have been assessed (30 per systematic assessments of effectiveness as part of their cent of protected areas assessed) than those with another internal protected area monitoring systems, independent IUCN management category (1 – 7 per cent of protected from the IUCN or donor networks. Even where these areas assessed). Protected areas with a larger area were data exist in North America and Europe, they may not be also more likely to have been assessed. This bias towards available through IUCN or UNEP WCMC networks and larger protected areas and National Parks is not this creates a challenge for a seamless reporting to the surprising; National Parks could be described as the

PARKS VOL 19.1 MARCH 2013

82 chapter III

9 www.iucn.org/parks

Figure 6: The percentage of protected areas that have undertaken a PAME assessment, by IUCN management category

‘charismatic mega fauna’ of protected areas. They are current sample of assessed protected areas is strongly often designated for their high biodiversity value or biased towards large protected areas and National Parks. spectacular landscapes, but also for their recreation and/ Some or all of these limitations in the data can be or spiritual value, and are therefore likely to attract more overcome; however, they must be considered when using funding and attention (and more likely to have PAME assessments to track progress towards monitoring and assessment structures in place, or have international biodiversity targets. been given funding which requires a PAME assessment to be completed) than smaller areas with less emphasis The PAME database, and the kind of information it on visitation and tourism. Older protected areas were contains, is valuable, but not in itself sufficient, for also slightly more likely to have been assessed. This effect tracking CBD Target 11. To address the “equity” element is possibly driven by the low rate of assessment in very of the Target 11, there is an urgent need for more detailed recently designated protected areas, in which protected and systematic assessment of the social and governance area management is more likely to be in the preliminary aspects of protected area management. IUCN and others stages and management effectiveness assessments may are currently working to improve both the social not yet be a priority, and/or the time lag between an indicators of management effectiveness and to create assessment being completed and its entry into the PAME additional tools for the social assessment of protected database. areas (IUCN TILCEPA, 2010). Information on biodiversity outcomes is captured, in part, in Target 11 of the CBD’s Aichi targets calls for ‘effectively management effectiveness assessments but will be better managed’ protected areas and protected area networks to informed by the work of the IUCN WCPA-SSC Task be conserved. PAME evaluations, although not designed Force on Biodiversity Outcomes of Protected Areas 3. as a tool for collecting scientific data, may provide the With these initiatives currently in the design stages, the first global-scale sample of data on protected area time is ripe for a discussion within the wider providing data for over 6,700 protected areas on core conservation community as to how we evaluate protected management inputs, context, process, outputs and area management at local, regional and global levels, outcomes. However, most PAME assessments were not what we are hoping to achieve with these evaluations, primarily designed to track CBD target progress, but and which tools might help us best achieve our aims. rather as a tool to help protected area managers start the process of adaptive management at a site and system level. Most of the assessments are completed by protected area managers, and this may introduce reporting biases. In addition, as these analyses show, the

PARKS VOL 19.1 MARCH 2013

83 chapter III

Lauren Coad et al 10

NOTES Hockings, M., S. Stolton, and N. Dudley. (2000). Evaluating Effectiveness: A framework for assessing management of 1 Some records in the dataset were provided on the basis that protected areas. Gland, Switzerland: IUCN. they were only used for global analyses and access to site Hockings, M., S. Stolton, and N. Dudley. (2004b). data is restricted. For information on the database, contact Management Effectiveness - assessing management of Marc Hockings at [email protected] protected areas. Journal of Environmental Policy and 2 For more information see: http://cms.iucn.org/fr/papaco/ Planning 6: 157 - 174. 3 For more information see: http://www.iucn.org/about/ Hockings, M., S. Stolton, F. Leverington, N. Dudley, and J. work/programmes/gpap_home/gpap_biodiversity/ Courrau. (2006). Evaluating Effectiveness: A framework gpap_wcpabiodiv/gpap_pabiodiv/ for assessing management effectiveness of protected

areas. Second edition. Gland, Switzerland: IUCN.

IUCN and UNEP. (2012). The World Database on Protected REFERENCES Areas (WDPA). Cambridge, UK: UNEP-WCMC. http:// Auditor General of Queensland. (2010). Sustainable www.protectedplanet.net. management of national parks and protected areas: A IUCN TILCEPA. (2010). Joint PAEL-TILCEPA workshop on performance audit. Report to Parliament No 9 for 2010. Protected Areas Management Evaluation & Social Brisbane, Australia: Auditor General of Queensland Assessment of Protected Areas. Gland, Switzerland: IUCN. Belokurov, A. and C. Besancon. (2009). New resources for Jenkins, C. N. and L. Joppa. (2009). Expansion of the global assessing the effectiveness of management in protected terrestrial protected area system. Biological Conservation areas. Oryx 43:14 - 14. 142:2166-2174. Bubb, P., L. Fish, and V. Kapos. (2008). Coverage of protected Joppa, L. N. and A. Pfaff. 2011. Global protected area impacts. areas. Guidance for national and regional use. Cambridge, Proceedings of the Royal Society B-Biological Sciences UK: UNEP-WCMC 278:1633-1638. Chape, S., J. Harrison, M. Spalding, and I. Lysenko. (2005). Leverington, F., K. Costa, J. Courrau, H. Pavese, C. Nolte, M. Measuring the extent and effectiveness of protected Marr, L. Coad, N. D. Burgess, B. Bomhard, and M. areas as an indicator for meeting global biodiversity Hockings. (2010a). Management effectiveness evaluation targets. Philosophical Transactions of the Royal Society B- in protected areas: a global study. Second edition. St. Biological Sciences 360:443-455. Lucia, Queensland, Australia: University of Queensland, Coad, L., N. D. Burgess, B. Bomhard, and C. Besancon. (2009a). IUCN- WCPA, TNC, WWF Progress Towards the Convention on Biological Diversity’s Leverington, F., K. Costa, H. Pavese, A. Lisle, and M. Hockings. 2010 and 2012 Targets for Protected Area Coverage. (2010b). A global analysis of protected area management Cambridge, UK: UNEP-WCMC effectiveness. Environmental Management 46:685 - 698. Coad, L., N. D. Burgess, L. Fish, C. Ravilious, C. Corrigan, H. Leverington, F., M. Hockings, and K. Costa. (2008). Pavese, A. Granziera, and C. Besancon. (2008). Progress Management effectiveness evaluation in protected areas - towards the Convention on Biological Diversity terrestrial a global study. Brisbane, Australia: University of 2010 and marine 2012 targets for protected area Queensland coverage. PARKS 17. Gland, Switzerland: IUCN Nolte, C., F. Leverington, A. Kettner, M. Marr, G. Neilsen, B. Coad, L., N. D. Burgess, C. Loucks, L. Fish, J. P. W. Bomhard, S. Stolton, S. Stoll-Kleemann, and M. Hockings. Scharlemann, L. Duarte, and C. Besancon. (2009b). The (2010). Protected area management effectiveness ecological representativeness of the global protected assessments in Europe. A review of application, methods areas estate in 2009: progress towards the CBD 2010 and results. Bonn, Germany: Federal Ministry of the target. Cambridge, UK: UNEP-WCMC Environment, Nature Conservation and Nuclear Safety Dudley, N. (ed.) (2008). Guidelines for Applying Protected Area NSW Audit Office. (2004). Performance audit: managing Management Categories. Gland, Switzerland: IUCN. natural and cultural heritage in parks and reserves: Dudley, N. and S. Stolton (eds). (2009). Protected area National Parks and wildlife service. Sydney, Australia: The management effectiveness: METT. NORAD. Audit Office of New South Wales Eken, G., L. Bennun, T. M. Brooks, W. Darwall, L. D. C. NSW DEC. (2005). State of the Parks 2004. Sydney, Australia: Fishpool, M. Foster, D. Knox, P. Langhammer, P. Matiku, E. NSW Department of Environment, Conservation (NSW DEC) Radford, P. Salaman, W. Sechrest, M. L. Smith, S. Spector, Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. and A. Tordoff. (2004). Key biodiversity areas as site Burgess, G. V. N. Powell, E. C. Underwood, J. A. D'Amico, I. conservation targets. Bioscience 54:1110-1118. Itoua, H. E. Strand, J. C. Morrison, C. J. Loucks, T. F. Allnutt, ESRI. (2012). ArcGIS Desktop. Release 10.1. Environmental T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Systems Research Institute, Redlands, CA. Hedao, and K. R. Kassem. (2001). Terrestrial ecoregions of Gelman, A. and J. Hill. (2007). Data Analysis Using Regression the worlds: A new map of life on Earth. Bioscience 51:933- and Multilevel/Hierarchical Models. Cambridge, 938. UK:Cambridge University Press Pinheiro, L. and D. Bates. (2000). Mixed-effects models in S Herbert, M. E., P. B. Mcintyre, P. J. Doran, J. D. Allan, and R. and S-Plus. Springer Verlag, New York, USA. Abell. (2010). Terrestrial Reserve Networks Do Not R Development Core Team. (2012). R: A language and Adequately Represent Aquatic Ecosystems. Conservation environment for statistical computing. Vienna, Austria: R Biology 24:1002-1011. Foundation for Statistical Computing Hockings, M., J. Ervin, and G. Vincent. (2004a). Assessing the Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L. Boitani, T. management of protected areas: the work of the World M. Brooks, R. M. Cowling, L. D. C. Fishpool, G. A. B. da Parks Congress before and after Durban. Journal of Fonseca, K. J. Gaston, M. Hoffmann, J. S. Long, P. A. International Wildlife Law and Policy 7:31 - 42. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W.

PARKS VOL 19.1 MARCH 2013

84 chapter III

11 www.iucn.org/parks

Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. of science to address conservation questions from the Watts, and X. Yan. (2004). Effectiveness of the global local to global scales. protected area network in representing species diversity. Nature 428:640-643. Ivon Cuadros is research assistant in the School of Scharlemann, J. P. W., V. Kapos, A. Campbell, I. Lysenko, N. D. Geography, Planning and Environmental Management at Burgess, M. C. Hansen, H. K. Gibbs, B. Dickson, and L. Miles. (2010). Securing tropical forest carbon: the the University of Queensland. Her research interests are contribution of protected areas to REDD. Oryx 44:352 - centred on monitoring the effectiveness and 357. management of protected areas. Schmitt, C. B., N. D. Burgess, L. Coad, A. Belokurov, C. Besancon, L. Boisrobert, A. Campbell, L. Fish, D. Gliddon, Jonas Geldmann is a PhD student at the Center for K. Humphries V. Kapos, C. Loucks, I. Lysenko, L. Miles, C. Macroecology, Evolution and Climate, University of Mills, S. Minnemeyer, T. Pistorius, C. Ravilious, M. Steininger, and G. Winkel. (2009). Global analysis of the Copenhagen, working on linking temporal biodiversity protection status of the world's forests. Biological data and management effectiveness in protected areas. Conservation 142:2122-2130. Spalding, M. D., L. Fish, and L. J. Wood. (2008). Toward Dr. Toby Marthews has a PhD in Plant Science and representative protection of the world's coasts and works as a research Post-doc at the Environmental oceans-progress, gaps, and opportunities. Conservation Letters 1:217-226. Change Institute, University of Oxford. His research Spalding, M. D., H. E. Fox, B. S. Halpern, M. A. McManus, J. involves all aspects of Tropical Forest Ecology, from Molnar, G. R. Allen, N. Davidson, Z. A. Jorge, A. L. conservation science to carbon cycle analysis to land Lombana, S. A. Lourie, K. D. Martin, E. McManus, J. surface modelling and micrometeorology. Molnar, C. A. Recchia, and J. Robertson. (2007). Marine

ecoregions of the world: A bioregionalization of coastal and shelf areas. Bioscience 57:573-583. Jessie Mee (MSc) is a programme analyst with the Stolton, S., Hockings, M., Dudley, N., MacKinnon, K., Whitten, UNDP Ecosystems and Biodiversity team. Her work T. and F. Leverington. (2007). Reporting progress in includes the review and analysis of METTs completed by Protected areas. A site level Management Effectiveness UNDP supported biodiversity projects in developing Tracking tool: second edition. Gland, Switzerland: World countries around in the world. Bank/WWF Forest Alliance and WWF. http:// www.wdpa.org/ME/PDF/METT.pdf WCED. (1987). Report of the World Commission on Christoph Nolte is a PhD candidate and research Environment and Development: Our Common Future. associate with the International Forestry and Resource Oxford, UK: UN World Commission on Environment and Institutions (IFRI) network at the School of Natural Development (WCED) Resources and Environment, University of Michigan. His research empirically evaluates the ecological and social ABOUT THE AUTHORS impacts of protected area management in the Andes- Dr. Lauren Coad is a research fellow with the Forest Amazon region. Governance Group, Oxford University and an honorary fellow with UNEP WCMC. Her research focuses on Professor Susanne Stoll-Kleemann is a Professor of understanding how management inputs and processes Sustainability Science and Applied Geography at the influence the effectiveness of protected areas in Institute of Geography and Geology at the University of conserving forests, and on the socio-economic drivers of Greifswald. She leads several research projects on bushmeat hunting in African tropical forests. Sustainable Land Management and Ecosystem Services, Dr. Fiona Leverington currently works as manager of especially in protected areas and biosphere reserves. planning and strategy at the Queensland Parks and Wildlife Service and is an adjunct senior fellow at the Nanna Granlie Vansteelant is an MSc student at the University of Queensland (UQ). Together with Marc Freshwater Biology Section at Copenhagen University. Hockings, she led the project investigating the global Her thesis looks at the effects of deforestation on picture of management effectiveness at UQ between 2006 macroinvertebrate diversity and leaf litter breakdown in and 2010. Her research interests also include reserve streams in the Udzungwa Mountains, Tanzania. planning, management planning and community relations. Camilo Zamora is biologist and research assistant in Professor Neil Burgess works in the Science the School of Geography, Planning and Environmental Programme at UNEP-WCMC in Cambridge, on practical Management at the University of Queensland. His field conservation projects with WWF and UNDP GEF in research interests are focused on the design and Africa, and as a part time staff member of Copenhagen establishment of marine protected areas, and their University. His research interests cover the practical use effectiveness in biological conservation.

PARKS VOL 19.1 MARCH 2013

85 chapter III

Lauren Coad et al 12

Mark Zimsky is a Senior Biodiversity Specialist at the Professor Marc Hockings is Professor of Global Environment Facility (GEF) and the Coordinator Environmental Management in the School of Geography, of the biodiversity program at the GEF. His work Planning and Environmental Management at the includes monitoring of GEF's biodiversity portfolio and University of Queensland. He is a long-term member of GEF support to protected area management the IUCN WCPA where he leads the global program on effectiveness. Most recently, he led two missions to India Science and Management of Protected Areas. and Zambia to investigate the evolution of the METT at the national level in these two countries with the aim of revising the METT for use throughout GEF's protected area portfolio (www.thegef.org/gef/BIO_results_learning)

RESUMEN La eficacia de la gestión de áreas protegidas es una consideración de importancia crítica para el éxito de los esfuerzos de conservación. Se han desarrollado más de 40 instrumentos de recolección de datos relacionados con la eficacia de la gestión de áreas protegidas (PAME) para la evaluación sistemática de la eficacia de la gestión de áreas protegidas. Muchas de estas evaluaciones han sido recogidas recientemente en la base de datos mundial sobre la Efectividad del Manejo de las Áreas Protegidas de la UICN (PAME). Utilizamos la base de datos de PAME junto con la Base de Datos Mundial de Áreas Protegidas (WDPA) para evaluar el progreso actual hacia las metas sobre PAME para 2010 y 2015 del Convenio sobre la Diversidad Biológica (CDB), que requieren que al menos el 30 y el 60 por ciento, respectivamente, de la superficie total de áreas protegidas haya sido evaluada en términos de efectividad de la gestión. Señalamos que a nivel mundial el 29 por ciento de las áreas protegidas han sido evaluadas y el 23 por ciento de los países han alcanzado la meta del 60 por ciento. Además, el 46 por ciento de los países han alcanzado la meta del 30 por ciento. Sin embargo, los resultados analíticos reflejan la existencia de sesgos en torno al tipo de áreas protegidas evaluadas; las áreas protegidas con áreas más grandes y las áreas protegidas designadas como Parques Nacionales (Categoría II de la UICN) tienen mayor probabilidad de haber realizado una evaluación de PAME. Por otra parte, hay pocas evaluaciones de PAME de Europa y América del Norte, donde las evaluaciones sobre la gestión de áreas protegidas pueden estar ya integradas en los sistemas de planificación y monitoreo de áreas protegidas, lo que dificulta el suministro de información al CDB. También analizamos con detenimiento las posibilidades y limitaciones de las evaluaciones de PAME como instrumentos para el seguimiento y la evaluación de la gestión de áreas protegidas, y la necesidad de nuevos instrumentos de evaluación para abordar los aspectos relativos a la “equidad” de la meta 11 del CDB.

RÉSUMÉ Pour garantir le succès de la conservation des aires protégées, il est extrêmement important de prendre en compte l’efficacité de leur gestion. Plus de 40 outils différents de collecte de données sur l’efficacité de la gestion des aires protégées ont été élaborés pour évaluer de façon systématique cette dernière. Un grand nombre de ces évaluations ont récemment été réunies dans la base de données mondiale de l’UICN sur l’efficacité de la gestion des aires protégées (PAME). Nous avons utilisé la base de données PAME ainsi que la Base de Données Mondiale sur les Aires Protégées (WDPA) pour évaluer les progrès réalisés quant aux objectifs de la Convention sur la diversité biologique pour 2010 et 2015 sur l’efficacité de la gestion des aires protégées. Selon ces objectifs, au moins 30 et 60 pour cent respectivement de la superficie totale des aires protégées doivent être évalués en termes d’efficacité de leur gestion. Nous démontrons ainsi que, à l’échelle mondiale, 29 pour cent des aires protégées ont été évaluées, et 23 pour sont des pays ont atteint l’objectif de 60 pour cent. En outre, 46 pour cent des pays ont atteint l’objectif de 30 pour cent. Cependant, les résultats analytiques montrent certaines limites – notamment dans le type d’aire protégée évaluée. Les aires protégées les plus vastes, ainsi que les aires protégées classées Parc National (catégorie II de l’UICN) sont beaucoup plus susceptibles d’avoir mené une évaluation PAME. En outre, on observe un déficit d’évaluations PAME provenant d’Europe et d’Amérique du nord, ce qui s’explique probablement par le fait que les évaluations sur la gestion des aires protégées sont déjà intégrées dans des systèmes de planification et de suivi des aires protégées – et il est donc plus compliqué de demander à ces acteurs de faire état de la situation auprès de la Convention sur la diversité biologique. Enfin, nous examinons le potentiel et les limites des évaluations PAME en tant qu’outils de suivi et d’évaluation des aires protégées, et étudions l’importance de mettre en place d’autres outils d’évaluation pour aborder les éléments liés à l’équité mentionnés dans l’Objectif 11 de la Convention sur la diversité biologique.

PARKS VOL 19.1 MARCH 2013

CHAPTER IV Changes in Management Effectiveness

Jonas Geldmann, Lauren Coad, and Neil D. Burgess

88 chapter IV

A global analysis of changes in Protected Area Management Effectiveness assessments over time

Jonas Geldmanna, Lauren Coadb, and Neil D. Burgessa,c

Abstract

Protected areas are amongst the most important conservation tools and today cover more than 12.7% of the terrestrial surface of the earth. Most of our knowledge on protected area performance comes from understanding their location in relation to biodiversity values. However, such analyses do not demonstrate whether protection is effective. In the absence of data on biological outcomes, changes in management input can serve as a proxy for conservation objectives, by tracking the investments into the protected area.

Here we use detailed information collected on management effectiveness at 223 protected areas across 48 countries to investigate if management effectiveness scores change over time in protected areas. We find that management effectiveness improves in 63.2% of the protected areas, while 36.8% of the areas show no improvements or even declines in management effectiveness. The main factors responsible for improvements were improved management plans, monitoring, regulations and investments in research. We also find indications that the changes observed are true, and not driven by local managers inflating scores to attract more resources.

Keywords: Management effectiveness, Management plan, METT, PAME, Protected area

89 chapter IV

Introduction Protected areas are one of the most important conservation tools for protecting biodiversity, and Ecosystem Services, as well as secure a sustainable local livelihood (Naidoo et al. 2006; Rodrigues 2006; Klein et al. 2007; Coad et al. 2008). This has led to an impressive global network of protected areas covering more than 12.7% of the terrestrial surface (twice the size of China) (Bertzky et al. 2012).

To what extent protected areas are effective at achieving conservation goals is hotly debated and largely depends on the measure of effectiveness being evaluated. Building from systematic conservation planning (Pressey et al. 1993; Margules & Pressey 2000) many studies have focused on whether protected areas are located in the right places; covering the most species (Rodrigues et al. 2004), the most intense human pressure (Leroux et al. 2010), as well as to identify an asymmetric growth in protected areas, favoring areas of less importance over areas of high biodiversity value (Joppa & Pfaff 2009; Butchart et al. 2012). However, such analyses do not inform on whether protection promotes positive changes in habitat or species inside the protected area. This requires understanding whether some matrix of conservation output or outcome improves compared to if protection had not been established (Ferraro 2009), ideally measuring and tracking changes directly related to the objectives of the protected area (Kapos et al. 2009). Few studies have been able to measure such changes in protected areas, and even fewer been able to correlate these with conservation responses related to the protected area (Geldmann et al. 2013).

Measuring the quality of conservation input can serve as a proxy for conservation performance as well-managed protected areas have a higher likelihood of delivering conservation outcomes than less well managed (Hockings et al. 2004a). This principle has become an important element of the international agreement to halt the loss of biodiversity (i.e. the 2020 Aichi targets), which calls for at least 30% of all protected area and 60% of all protected land to be evaluated by 2020 (Convention on Biological Diversity 2010). A target which has still not been reached (Coad et al. 2013a).

Protected Area Management Effectiveness (PAME) schemes are used in many places of the world to evaluate the strength and weaknesses of protected area management systems. The IUCN World Commission on Protected Areas (WCPA) has developed a framework to guide Management Effectiveness efforts and assist the comparison between different methodologies (Hockings et al. 2006). Management Effectiveness includes three main components: i) design and planning issues, ii) appropriateness of management systems and processes, and iii) delivery of protected area objectives (Hockings 2003). Most often the assessments are conducted by protected area managers, governmental employs and/or donor/NGO’s in the field as questionnaires covering some to all of six generic categories: i) contex, ii) planning, iii) input, iv) process, v) outputs, and vi) outcomes (Hockings 2003). To date more than 10,000 evaluations have been conducted in over 6,700 protected areas (Leverington et al.

90 chapter IV

2010), and these have become a benchmark for evaluating management performances, particularly in many developing countries driven by large international donors investing huge amounts into protected areas (Hockings et al. 2004b). However, where these assessments have been conducted, analyses have shown that many sites achieve low scores, with more than 50% of the sites having major or significant shortcoming in the quality of management (Leverington et al. 2010), suggesting that the implementation of PAME alone is not enough to secure effective reserves. Results confirmed by emerging studies correlating the quality of management effectiveness and protected areas ability to reduce the risk of fires inside their boundaries (Nolte & Agrawal 2013; Nolte et al. 2013).

However, while the implementation of PAME evaluations might not be a panacea for the target protected areas, they do deliver valuable information on challenges and opportunities for improved performance. Further, PAME evaluations are often implemented in areas where a need for outside investments have been identified (Hockings et al. 2004a) suggesting that these areas are not initially state of the art reserves. For the same reason many organizations require multiple evaluations doing a project period. This enables measuring whether protected areas improve their management effectiveness over time as well as which elements of managements improves. While not a direct measure of conservation outcomes, improvements in management effectiveness, or the opposite is likely a surrogate for protected areas potential to deliver true conservation outcomes and helping to halt the loss of biodiversity.

Here we examine the temporal changes in one of the most widely used PAME tools, the Management Effectiveness Tracking Tool (METT) (Stolton et al. 2007) in a global analysis investigating whether protected areas improve their management effectiveness over time. We also attempt to determine the factors driving improving or decreasing METT scores over time.

Methods We developed a database containing detailed information from more than 3,000 METT evaluations from which we extracted information on change in management effectiveness (Coad et al. 2013b). This database has been created through collecting and collating data from the original METT documents solicited from site managers, donor organizations and governmental agencies.

METT is one of the most widely applied PAME tools and has been used worldwide particularly for protected area project supported by the World Bank, the Global Environmental Facility (GEF), the United Nations Development Programme (UNDP), and the World Wildlife Fund (WWF). METT collects information on i) objectives, ii) threats, iii) budgets, iv) staffing, v) size, and vi) designations of the protected areas. METT also documents the status of 30 specific questions ranging from legal status, equipment over outreach programs. For each of the 30 specific questions, local

91 chapter IV evaluators assign scores from 0 to 3 depending on the compliance of the specific management element (Stolton et al. 2007).

From the METT database we first identified all METT evaluations for which we had multiple entries from different years (n = 425). The 425 METT assessments were subsequently carefully reviewed and only protected areas where all assessment were conducted using the same methodology for all years were included in the final analysis. Changes in i) dollar budget, ii) staff numbers, iii) objectives and iv) threats could not be tracked over time and were excluded. Thus, we used only the scores from the 30 questions to compare differences in management between years.

Subsequently, we extracted the specific scores for all evaluations from all sites for the 30 questions. For all sites only matched questions were compared, so that questions assigned scores in on year and not another were not included in the comparison. This was done for each site individually to maximize the information included for all protected areas. We calculated a Standardized Accumulated METT Index for each point in time: for all sites we divided the accumulated score by the number questions filled out, multiplied by the potential maximum number of questions (n = 30), as a proxy for the overall performance of a given site at a given time (Dudley et al. 2007). Scores are assigned based on specific descriptions for each question, (for example law enforcement: 0 = The staff have no effective capacity/resources to enforce protected area legislation and regulations, 1 = There are major deficiencies in staff capacity/resources to enforce protected area legislation and regulations (e.g. lack of skills, no patrol budget), 2 = The staff have acceptable capacity/resources to enforce protected area legislation and regulations but some deficiencies remain, and 3 = The staff have excellent capacity/resources to enforce protected area legislation and regulations) of many different categories of interventions. However, all follow the same pattern ranging from not existing, through insufficient and then to adequate. So while cumulative scores cannot be used to inform on whether specific elements of the PAME is responsible for the improvement in METT scores, changes in accumulated METT scores within the same project across time, can be indicative of improved or decreasing management input.

For all protected areas we identified their Human Development Index (HDI) score (United Nations Development Programme 2011), based on their country of origin as well as their biogeographical realm based on WWF’s ecoregions (Olson et al. 2001). We extracted information on protected area size, age, and IUCN management categories from the World Database on Protected areas (WDPA) (World Conservation Monitoring Centre 2013).

Statistical analysis All analysis was conducted in R 2.14.1 (R Development Core Team 2011). A linear regression model was used to compare the changes in METT against the initial accumulated METT scores as well as changes in METT scores and number of years

92 chapter IV between first and last assessment. A paired students t-test was used to compare the accumulated METT scores across site difference between first and last evaluation as well as to compare the difference between individual questions. An Analysis of Covariance (ANCOVA) was used to test the correlation between changes in METT scores against HDI, IUCN category, protected area size, year of designation, country, and WWF realm.

Results We identified 223 protected areas from 48 countries globally, which had conducted METT evaluations more than once doing the project period (Figure 1). Seventy-eight percent of the sites (n = 175) had only two METT evaluations entered in the METT database, while the remainder had either five (n = 1), four (n =8), or three (n = 39). The mean number of years between first and last evaluation was 4.4 years (S.D. =1.8).

Figure 1. Global map of protected areas and sites for which multiple METT assessments has been conducted. Green areas are the individual protected areas. Countries for which METT evaluations were extracted have been highlighted.

For the 30 specific questions composing the main element of the METT assessment, 116 (23%) had complete information for all questions, 179 lacked information for one question, 78 from two questions, 39 from three questions, 35 from four questions, 17 from five questions, and 42 lacked information from more than five questions of which three had less than 10 questions filled in.

Sites from all WWF realms on average increased their performance between first and last assessment with neotropic sites performing significantly better than sites from any other WWF realm (Figure 2a). Neotropic protected areas also had lower initial mean scores, through this difference was not significant (Figure 2b). We also found a significant effect of countries, suggesting that some countries performed better than others. But this was not reflected by the countries Human Development index. Changes

93 chapter IV in accumulated METT scores also varied significantly between IUCN categories. However variation within the individual categories was great, and only Ib could be separated from the rest as performing significantly better than all other categories (Table 1).

a b 14 50 B 12 40 10

8 30 A 6 A 20 A

ChangeMETT inscores 4

10 2 Initial accumulated METT scores

0 0 Afrotropic IndomalayaNeotropics Palearctic Afrotropic IndomalayaNeotropics Palearctic

Figure 2. METT scores across WWF biogeographical realms (a) shows the average change between first and last assessment, and (b) the average initial accumulated METT scores.

Change in METT scores Accumulated scores increases in 141 (63.2%) out of 223 protected areas, remained stable in 21 (9.4%), and decreased in 61 (27.4%). On average, across all protected areas, accumulated scores were significantly higher in the last evaluations (mean = 44.45, S.D. = 13.15) compared to the first (mean = 39.63, S.D. = 12.37), using a paired students t- test (t = 3.99, p < 0.0001).

Table 1. model output from ANCOVA Variable D.f. F‐value p‐value WWF Realm 3 7.58 0.0001 *** Country 37 3.39 <0.0001 *** Protected area size 1 1.29 0.25 Year of establishment 1 1.42 0.24 IUCN category 7 4.63 0.0001 *** Human Development Index 1 0.91 0.341

When evaluating the 30 questions separately: 28 increased, and two decreased between first and last evaluation. Of these, 15 increased significantly (Table 2). Increases were seen particularly for management plans, monitoring, protected area objectives, protected area regulations, and investments in research increased between METT evaluations.

There was a significant positive effect of the number of years between first and last assessment, and the change in METT scores (t =4.52, p < 0.0001), suggesting that improvements in management effectiveness increased with the number of years the project had been running (Figure 3).

94 chapter IV

Table 2. List of the 30 questions included in the METT questionnaire. PAME First Last Category Difference t value p value element assessment assessment 1 Legal status Context 2.87 2.85 ‐0.03 ‐0.57 0.569 2 Protected area regulations Planning 1.68 1.91 0.23 2.99 0.003 * 3 Law enforcement Input 1.35 1.53 0.18 2.50 0.013 * 4 Protected area objectives Planning 1.81 2.09 0.27 3.43 0.001 ** 5 Protected area design Planning 1.82 1.95 0.13 1.94 0.052 6 Protected area boundary Process 2.10 2.20 0.11 1.43 0.155 7 Management plan Planning 1.07 1.42 0.34 3.58 <0.001 *** Planning/ 1.57 1.77 0.20 2.12 0.035 * 8 Regular work plan output 9 Resource inventory Input 1.69 1.88 0.19 2.47 0.014 * 10 Research Process 1.50 1.72 0.23 2.42 0.016 * 11 Resource management Process 1.53 1.73 0.20 2.47 0.014 * 12 Staff numbers Input 1.57 1.72 0.15 1.99 0.047 * Input/ 1.39 1.54 0.14 1.74 0.083 13 Personal management process Input/ 1.40 1.60 0.20 3.01 0.003 ** 14 Staff training process 15 Current budget Input 1.20 1.35 0.15 2.19 0.029 * 16 Security of budget Input 1.26 1.39 0.13 1.72 0.087 17 Management of budget Process 1.52 1.47 ‐0.04 ‐0.57 0.567 18 Equipment Input 1.36 1.53 0.17 2.16 0.031 * 19 Maintenance of equipment Process 1.43 1.56 0.13 1.38 0.168 20 Education program Process 1.24 1.44 0.20 2.30 0.022 * 21 State and comm. neighbors Process 1.55 1.56 0.01 0.11 0.911 22 Indigenous people Process 0.97 1.12 0.15 1.49 0.138 23 Local communities Process 0.97 1.15 0.18 2.16 0.031 * 24 Visitor facilities Outputs 0.94 1.09 0.15 1.63 0.103 25 Commercial tourism Process 0.84 0.87 0.02 0.24 0.814 Input/ 1.33 1.39 0.06 0.48 0.631 26 Fees Process 27 Condition assessment Outcome 1.93 2.06 0.13 1.88 0.061 28 Access assessment Output 1.37 1.50 0.13 1.59 0.113 29 Economic benefit assessment Outcome 1.55 1.67 0.12 1.48 0.138 Planning/ 1.23 1.54 0.31 3.87 <0.001 *** 30 Monitoring and evaluation process PAME element is from the groupings defined by WCPA. First and last assessments is the average scores from the individual questions as well as the difference between these. T and p values are based on individual pairwise t-tests conducted for each question separately.

95 chapter IV

Figure 3. Partial-effects plot from model of the correlation between changes in METT scores and the number of years between first and last assessment. There is a significant positive relationship between the length doing which the project is running and the improvement in METT scores (t =4.52, p < 0.001).

There was a highly significant negative correlation between the initial first METT scores and the change in METT scores for the individual sites (R2 =, 0.17, t = -6.86, p < 0.0001), so that sites with initial high accumulated METT values changes less than sites with initial lower accumulated METT values (Figure 4).

0 Figure 4. Correlation between the initial METT score for individual

Changes in METT scores in Changes sites and the change between the start and end METT score. There −20 was a highly significant negative relationship (R2 =, 0.17, t = -6.86, p < 0.0001). 10 20 30 40 50 60 70

METT scores at first evaluation

Discussion We show that METT scores generally improve over time, both looking at the overall performance of the protected areas as well as the individual elements. This finding is perhaps not surprising as 1) the time between first and last evaluation often represents a monetary and resource investment in management activities and 2) the evaluations are

96 chapter IV conducted on site often by people depending on showing improvements to secure a continued influx of set resources. However while speculations have suggested that the role of the second part significantly impairs the usability of PAME evaluations, our results suggests the first part being a component of the observed changes. First, accumulated scores were significantly correlated to the time between the first and the last assessment suggesting implementation time positively affects changes in METT. We argue that this result would be expected, if improvements in METT were caused by true changes, as the implementation of management actions would generally be dependent on sufficient time to be successful, contrary to inflated scores which is an artifact of false reporting independent of implementation time. Second, our findings that protected areas with lower initial accumulated scores improve more than initially better performing areas also suggest that areas with a great true potential for change are indeed the areas which improves the most. Third, almost 37% of the sites experienced no change or even declines in overall scores which is a considerable proportion, had there been a systematic manipulation of scores in protected areas.

Across the 30 questions scores increased in 28, of which 15 were significant. The largest increases were observed for questions within the planning and process elements of the METT evaluation while context and outcome elements increased the least. Interestingly management of budgets, an activity largely dependent on protected area managers, decreases between first and last assessment. This question is alone a reflection of local management and would be a likely candidate question for manipulation if scores were inflated by area managers. Of all the management effectiveness questions management plans improved the most. Management plans has previously been shown to improve conditions for biodiversity (Sergio et al. 2005; Pettorelli et al. 2010) and are an instrumental part of successful management of protected areas. Previous studies looking at the correlation between management and protected areas ability to deliver conservation outcomes (i.e. improvements for biodiversity) have also found staffing to be an important predictor (Caro 1999; Carrillo et al. 2000; Sinclair et al. 2007) as well as the subsequent training and education of staff (Jachmann 2008a, b). While these are not amongst the METT questions which increase the most, we do find a significant increase for both, particularly staff training suggesting that protected areas are improving their ability to protect biodiversity with improving METT scores.

Our study is limited by the availability of data, which comes mainly from countries in South America, Africa, Eastern Europe and Asia. Almost no data on METT are available from North America, Western Europe, and Australia. Evaluations of PAME are conducted in these regions, but often using different national systems (Auditor General of Queensland 2010; Nolte et al. 2010; Parks Canada 2013). Our study is therefore restricted to countries and protected areas using METT. This biases the sample towards developing countries, and protected areas receiving donor resources, especially

97 chapter IV those from the GEF where the METT tool is mandatory for all protected area projects (Stolton et al. 2007).

We found significant differences between countries as well as larger geographical regions (i.e. WWF realms) where projects in Latin America performed significantly better than any other region. Similar results are found for deforestation, where Latin American protected areas have been shown to reduce deforestation in protected areas better than Africa and Asia (Scharlemann et al. 2010; Geldmann et al. 2013).

We observed very large variations in our results for both the initial accumulated METT values as well as the observed changes. Still our model explaining changes in METT scores explains ca.s 43% of the variation observed. Many other factors not captured in our model, or in the METT reporting, may still contribute to how well protected areas are able to improve their management. A recent study looking at the adequacy of funding for conservation at a national level found many of the countries where METT have been widely applied to be amongst the most underfunded (Waldron et al. 2013). We haven’t been able to utilize the information on direct monetary input into protected areas also collected in the METT process but expect this would be a significant contributor to explaining the changes in management effectiveness. Studies on local participation in conservation project has also shown this to be a of importance, but the relationship is complicated and the effects variable (Bowler et al. 2010; Porter-Bolland et al. 2012) suggesting that these factors play are more complicated role than can be extracted through METT evaluations.

Overall, we find that management effectiveness does improve over time for protected areas conducting sequential evaluations. However, the large variability in project suggests that many other factors not captured by METT contributed to the observed changes. Further, while we presents results which suggests changes in management effectiveness are in part derived from true improvements, our results is not a direct test of on-ground changes in management. Detailed analyses of factors affecting the assessments as well as the influence of protected area managers on evaluations and scores are needed to fully interpret the link between true improvements and reporting bias. The amendment in the newest version of the METT to include information on the parties present at the evaluations is a step in the right direction. However these were not available for METT evaluations going back in time.

Finally our study only evaluates the changes in METT and is not able to capture whether improved management effectiveness performances over time is correlated to better performing protected areas. However improvement in management over time is vital to ensure reserves adaptable to new situations and capable of addressing pressures and threats more effectively. We see a need to evaluate whether improvements in METT scores are also correlated to improved condition for biodiversity or reduced threats. However the data requirement for such an effort will require independently

98 chapter IV collected data on a relevant outcome matrix, overlapping with the low number of sites for which there has been conducted multiple METT evaluations.

Acknowledgements We thank M. Hockings, F. Leverington, C. Nolte, M. Zimsky, J. Mee, N. Vansteelant, and many others who collected, compiled, and provided METT data. We thank the Danish National Research Foundation for primary financial support. We also thank the IUCN SSC/WCPA Joint Task-Force on Biodiversity and Protected Areas, UNEP- WCMC, WWF and the GEF for financial and institutional support.

References Auditor General of Queensland. (2010). Sustainable management of national parks and protected areas: A performance audit. Report to Parliament No 9 for 2010. Auditor General of Queensland, Brisbane, Australia. Bertzky, B., C. Corrigan, J. Kemsey, S. Kenney, C. Ravilious, C. Besancon, and N. D. Burgess. (2012). Protected Planet report: Tracking progress towards global targets for protected areas. IUCN and UNEP-WCMC, Gland, Switzerland and Cambridge, UK. Bowler, D., L. Buyung-Ali, J. R. Healey, J. P. G. Jones, T. Knight, and A. S. Pullin. (2010). The Evidence Base for Community Forest Management as a Mechanism for Supplying Global Environmental Benefits and Improving Local Welfare. Environmental Evidence. Collaboration for Environmental Evidence. Butchart, S. H. M., J. P. W. Scharlemann, M. I. Evans, S. Quader, S. Aricò, J. Arinaitwe, M. Balman, L. A. Bennun, B. Bertzky, C. Besançon, T. M. Boucher, T. M. Brooks, I. J. Burfield, N. D. Burgess, S. Chan, R. P. Clay, M. J. Crosby, N. C. Davidson, N. De Silva, C. Devenish, G. C. L. Dutson, D. F. D. z. Fernández, L. D. C. Fishpool, C. Fitzgerald, M. Foster, M. F. Heath, M. Hockings, M. Hoffmann, D. Knox, F. W. Larsen, J. F. Lamoreux, C. Loucks, I. May, J. Millett, D. Molloy, P. Morling, M. Parr, T. H. Ricketts, N. Seddon, B. Skolnik, S. N. Stuart, A. Upgren, and S. Woodley. (2012). Protecting Important Sites for Biodiversity Contributes to Meeting Global Conservation Targets. PLoS ONE 7:e32529. Caro, T. M. (1999). Densities of mammals in partially protected areas: the Katavi ecosystem of western Tanzania. Journal of Applied Ecology 36:205-217. Carrillo, E., G. Wong, and A. D. Cuaron. (2000). Monitoring mammal populations in Costa Rican protected areas under different hunting restrictions. Conservation Biology 14:1580-1591. Coad, L., N. Burgess, L. Fish, C. Ravilious, C. Corrigan, H. Pavese, A. Granziera, and C. Besancon. (2008). Progress towards the Convention on Biological Diversity terrestrial 2010 and marine 2012 targets for protected area coverage. NatureBureau, UK, Gland, Switzerland. Coad, L., F. leverington, N. D. burgess, I. C. Cuadros, J. Geldmann, T. R. Marthews, J. Mee, C. Nolte, S. Stoll-Kleemann, N. Vansteelant, C. Zamora, M. Zimsky, and M. Hockings. (2013a). Progress towards the CBD Protected Area Management Effectiveness Targets. Parks 19:13-24. Coad, L., F. Leverington, J. Geldmann, C. Nolte, and M. Hockings. (2013b). Management Effectiveness Tracking Tool, global database, University of Oxford and University of Queensland. Available from Convention on Biological Diversity. (2010). Strategic Plan for Biodiversity 2011-2020 - COP 10, decision X/2. Convention on Biological Diversity Available from http://www.cbd.int/decision/cop/?id=12268

99 chapter IV

Dudley, N., A. Belokurov, L. Higgins-Zogib, M. Hockincs, S. Stolton, and N. Burgess. (2007). Tracking progress in managing protected areas around the world. World Wildlife Fund, Gland, Switzerland. Ferraro, P. J. (2009). Counterfactual thinking and impact evaluation in environmental policy. New Directions for Evaluation 2009:75-84. Geldmann, J., M. Barnes, L. Coad, I. D. Craigie, M. Hockings, and N. D. Burgess. (2013). Effectiveness of terrestrial protected areas in reducing habitat loss and population declines Biological Conservation 161:230-238. Hockings, M. (2003). Systems for assessing the effectiveness of management in protected areas. BioScience 53:823-832. Hockings, M., J. Ervin, and G. Vincent. (2004a). Assessing the management of protected areas: the work of the World Parks Congress before and after Durban. Journal of International Wildlife Law and Policy 7:31-42. Hockings, M., S. Stolton, F. Leverington, N. Dudley, J. Courrau, P. Valentine, and S. Editor. (2006). Evaluating Effectiveness: A framework for assessing management effectiveness of protected areas, Gland, Switzerland. Hockings, M., S. U. E. Stolton, and N. Dudley. (2004b). Management Effectiveness: Assessing Management of Protected Areas? Journal of Environmental Policy & Planning 6:157- 174. Jachmann, H. (2008a). Illegal wildlife use and protected area management in Ghana. Biological Conservation 141:1906-1918. Jachmann, H. (2008b). Monitoring law-enforcement performance in nine protected areas in Ghana. Biological Conservation 141:89-99. Joppa, L. N., and A. Pfaff. (2009). High and Far: Biases in the Location of Protected Areas. PLoS ONE 4:e8273. Kapos, V., A. Balmford, R. Aveling, P. Bubb, P. Carey, A. Entwistle, J. Hopkins, T. Mulliken, R. Safford, A. Stattersfield, M. Walpole, and A. Manica. (2009). Outcomes, not implementation, predict conservation success. Oryx 43:336-342. Klein, A. M., B. E. Vaissiere, J. H. Cane, I. Steffan-Dewenter, S. A. Cunningham, C. Kremen, and T. Tscharntke. (2007). Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B-Biological Sciences 274:303-313. Leroux, S. J., M. A. Krawchuk, F. Schmiegelow, S. G. Cumming, K. Lisgo, L. G. Anderson, and M. Petkova. (2010). Global protected areas and IUCN designations: Do the categories match the conditions? Biological Conservation 143:609-616. Leverington, F., K. L. Costa, H. Pavese, A. Lisle, and M. Hockings. (2010). A Global Analysis of Protected Area Management Effectiveness. Environmental Management 46:685-698. Margules, C. R., and R. L. Pressey. (2000). Systematic conservation planning. Nature 405:243- 253. Naidoo, R., A. Balmford, P. J. Ferraro, S. Polasky, T. H. Ricketts, and M. Rouget. (2006). Integrating economic costs into conservation planning. Trends in Ecology & Evolution 21:681-687. Nolte, C., and A. Agrawal. (2013). Linking Management Effectiveness Indicators to Observed Effects of Protected Areas on Fire Occurrence in the Amazon Rainforest. Conservation Biology 27:155-165. Nolte, C., A. Agrawal, and P. Barreto. (2013). Setting priorities to avoid deforestation in Amazon protected areas: are we choosing the right indicators? Environmental Research Letters 8:015039. Nolte, C., F. Leverington, A. Kettner, M. Marr, G. Nielsen, B. Bomhard, S. Stolton, S. Stoll- Kleemann, and M. Hockings. (2010). Protected Area Management Effectiveness Assessments in Europe, Bonn, Germany. Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. V. N. Powell, E. C. Underwood, J. A. D'Amico, I. Itoua, H. E. Strand, J. C. Morrison, C. J. Loucks, T. F.

100 chapter IV

Allnutt, T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P. Hedao, and K. R. Kassem. (2001). Terrestrial Ecoregions of the World: A New Map of Life on Earth. Bioscience 51:933-938. Parks Canada. (2013). State of Parks reports for national parks. Parks Canada, Quebec, Canada. Available from http://www.pc.gc.ca/eng/docs/bib-lib/docs5hi.aspx Pettorelli, N., A. L. Lobora, M. J. Msuha, C. Foley, and S. M. Durant. (2010). Carnivore biodiversity in Tanzania: revealing the distribution patterns of secretive mammals using camera traps. Animal Conservation 13:131-139. 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. Pressey, R. L., C. J. Humphries, C. R. Margules, R. I. Vane-Wright, and P. H. Williams. (1993). Beyond opportunism: Key principles for systematic reserve selection. Trends in Ecology & Evolution 8:124-128. R Development Core Team. (2011). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rodrigues, A. S. L. (2006). Are global conservation efforts successful? Science 313:1051-1052. Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, R. M. Cowling, L. D. C. Fishpool, G. A. B. da Fonseca, K. J. Gaston, M. Hoffmann, J. S. Long, P. A. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W. Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. Watts, and X. Yan. (2004). Effectiveness of the global protected area network in representing species diversity. Nature 428:640-643. Scharlemann, J. P. W., V. Kapos, A. Campbell, I. Lysenko, N. D. Burgess, M. C. Hansen, H. K. Gibbs, B. Dickson, and L. Miles. (2010). Securing tropical forest carbon: the contribution of protected areas to REDD. Oryx 44:352-357. Sergio, F., J. Blas, M. Forero, N. Fernández, J. A. Donázar, and F. Hiraldo. (2005). Preservation of wide-ranging top predators by site-protection: Black and red kites in Doñana National Park. Biological Conservation 125:11-21. Sinclair, A. R. E., S. A. R. Mduma, J. G. C. Hopcraft, J. M. Fryxell, R. Hilborn, and S. Thirgood. (2007). Long-term ecosystem dynamics in the Serengeti: Lessons for conservation. Conservation Biology 21:580-590. Stolton, S., M. Hockings, N. Dudley, K. MacKinnon, T. Whitten, and L. F. (2007). Reporting Progress in Protected Areas A Site-Level Management Effectiveness Tracking Tool: second edition. World Bank/WWF Forest Alliance, Gland, Switzerland. United Nations Development Programme. (2011). Human Development Report 2011: Sustainability and Equity: A Better Future for All. UNDP, New York, USA. Waldron, A., A. O. Mooers, D. C. Miller, N. Nibbelink, D. Redding, T. S. Kuhn, J. T. Roberts, and J. L. Gittleman. (2013). Targeting global conservation funding to limit immediate biodiversity declines. Proceedings of the National Academy of Sciences:PNAS early edition 1-5 pp. World Conservation Monitoring Centre. (2013). World Database on Protected Areas. WCMC, Cambridge, United Kingdom. Available from http://protectedplanet.net/

101

102

CHAPTER V Protected Areas ability to reduce pressure

Jonas Geldmann, Lucas N. Joppa, and Neil D. Burgess

Invited for resubmission in Conservation Biology John Wiley & Sons, Inc., ISSN: 1523- 1739

All supplementary material refers to appendix II

103

104 chapter V

Spatial and temporal changes in human threats to wild nature and impacts on protected areas between 1995 and 2010

Jonas Geldmann1, Lucas N. Joppa2,3 and Neil Burgess1,3,4

1 Center for Macroecology, Evolution and Climate, Department of Biology, University of Copenhagen, Denmark, 2 Microsoft Research, Computational Ecology, 21 Station Road, Cambridge UK CB1 2FB, 3 United Nations Environmental Programme World Conservation Monitoring Centre, 4 World Wildlife Fund, 1250 24th St NW, Washington, DC 20037, USA

Abstract It is widely accepted that the main engine for the observed and dramatic decline biodiversity is an increasing human pressure on the earth systems. However the spatially explicit patterns of change in human pressure and its relation to human responses is less known. Here we develop a spatially explicit and temporal map of change in human pressure from 1995-2010 that is applicable globally at a resolution of 5 km2. Our temporal pressure map is based on two global threat-layers that could be compared spatially and temporally, following a comprehensive evaluation of numerous other remote sensing products that relate to seven additional categories of threat to biodiversity. Our temporal pressure map estimates not only accumulated total global human pressure, but also shows how this has changed over the past 15 years at a resolution that allows for comparison between regions and conservation responses. Our results show great geographical differences in pressure changes over time. Moreover, within protected areas, change in pressure varies according to continent and between the International Union for Conservation of Nature (IUCN) protected area categories. Our analysis is an initial step towards mapping change over time in drivers of biodiversity declines, a further development of the ideas first proposed in 2002 by the ‘Human Footprint’ analysis. That we were only able to include two datasets in our spatio-temporal threat map highlights the challenges and limitations to producing robust spatial and temporal measures to inform conservation strategies and highlights a startling paucity of data with respect to threats to biodiversity - one of the world’s most pressing problems.

Keywords: Human footprint; Human population; IUCN management categories; pressure; protected area; Response; Spatial; stable nightlight; State; Temporal

105 chapter V

Introduction Biodiversity is in rapid decline (Barnosky et al. 2011) despite continued international commitment (i.e. the Aichi targets: CBD 2010) linked to increasing efforts to reverse this trajectory (Butchart et al. 2010). It is widely accepted that the decline is caused primarily by growing human impacts (Butchart et al. 2010; Zalasiewicz et al. 2010). Understanding the extent and effects of these impacts on nature is a key question for conservation science.

The Pressure-State-Response (PSR) framework (Organisation for Economic Co- operation and Development 1993) was developed to understand the interaction between human activities and the state of biodiversity. In this system ‘pressure’ characterizes all negative impacts caused by human activities. These pressures can be diverse in origin and impact, and have both short and long-term effects on wild nature that can be local, or extend far beyond their origin. They range from direct effects on animals from bush meat hunting (Wilkie & Carpenter 1999; Peres 2000), to indirect global long term effects of climate change (Intergovernmental Panel on Climate Change 2007). Evaluating human pressures is challenging for two major reasons. First; human pressure is composite and the effects intertwined, making aggregated indices challenging. Second; the impact on ecosystems and species may vary considerably depending on a multitude of factors, so that the specific effect cannot be inferred without understanding the system within which it works. This leads to very different ways of evaluating human pressure, depending on whether the focus is on the cause or the effect. In their review, Salafsky et al. (2008) identified 11 categories of pressures / threats which were considered to directly impact biodiversity – which Baldwin (2010) aggregated into five ‘key-stone threats’ (Baldwin 2010) (Table 1 and SI). These approaches to cataloging threats have been criticized for not effectively separating the ‘source’ and the ‘mechanism’, making it difficult to capture information about how different threats operate (Balmford et al. 2009). Ultimately the validity of a scheme will depend on the questions posed and the type of data available. Based on existing threat-systems (Salafsky et al. 2008; Balmford et al. 2009; Baldwin 2010) and focusing on measuring input, we identified nine categories of globally important threat types which also included socio-economic drivers: i) Climate change, ii) Diseases, iii) Education and livelihood, iv) Human population pressure, v) Invasive species, vi) Land properties/resources, vii) Legislation and corruption, viii) Pollutants, and ix) Transport and access (Table 1).

Only a few studies have mapped human pressure on the environment across the entire world (McCloskey & Spalding 1989). Of these the “human footprint” (Sanderson et al. 2002) has become a benchmark for mapping contemporary human pressure. But the human footprint maps are static, with no updates to the data that comprise them, and map contemporary pressure and not changes in pressure. While such maps can help

106 chapter V conservation planning, they are less useful when evaluating whether human impacts are changing over time, where, and why. Likewise, when evaluating the impact of conservation responses, maps of the accumulated pressure only explain the context in which conservation interventions take place, not the effectiveness of conservation measures, for example protected areas.

Here we present an evaluation of many potential sources that might measure pressures on biodiversity and use the results of this assessment to develop a simple, but methodologically consistent, map of human pressure from 1995 to 2010, building on the tradition from McCloskey and Spalding (1989). We use these maps to produce spatially explicit comparisons of pressure over time. Our purpose is to illustrate a way to measure how human pressures have changed across the world, in ways that are useful for conservation science. Because we evaluate change, we are dependent on data layers that represent pressure in a given year, but which are also comparable in terms of classification systems, spatial resolution, seasonality etc. Thus, our approach is more heavily constrained in terms of potential input data than previous static threat mapping products. We present this effort with the aims of: providing a first assessment of how pressure has changed over time, illustrating a blueprint for future efforts, and as a plea for the conservation community to address the startling paucity of data with respect to threats to biodiversity - one of the world’s most pressing problems.

Methods Identification of pressure layers We completed an extensive review of existing spatial products of human pressure, identifying 15 potentially suitable datasets (Table 1). These dataset were accessed together with their supporting information and evaluated based on four criteria for potential use in our analysis. First; only datasets that had global spatial coverage were included, the exception being datasets only missing data in the Polar Regions which were still considered. Second; only datasets that had repeated measurements over more than 10 years were included. Third; only datasets where repeated measurements over time were conducted using a consistent and comparable methodology were included. Fourth; we only included datasets where the impact of the source of pressure was captured by the spatial resolution of the data (e.g. the negative effects of most pollutants are local to the areas of use, and cannot be extracted from national statistics). We did not include the threats from climate change. The effect of a changing climate cannot be captured by just comparing temperature between two years, and establishing a baseline from which to measure change needs more careful consideration before inclusion (Parmesan & Yohe 2003; Hoffmann & Sgro 2011).

107

Table 1. Potential datasets to measure major threat categories affecting biodiversity at global scale Spatial Temporal Threat category Baldwin Salafsky Products Reference Notes coverage coverage Residential and commercial Land Landuse and landcover Bartholomé et al. Composite data sets based on classifying different remote sensing development and Agriculture GLC2000 Global (1km2) 2000 properties/resourcesa change (2005) product into a series of discrete categories. No temporal repeat and aquaculture Composite data sets based on classifying different remote sensing Residential and commercial Land Landuse and landcover Arino et al. product into a series of discrete categories. Comparing across time development and Agriculture GlobCover300 Global(0.3 km2) 2004; 2009 properties/resources change (2007) is generally not recommended and should only be done after and aquaculture careful evaluation Residential and commercial Composite data sets based on classifying different remote sensing Land Landuse and landcover Klein et al. 1700;1800;1900 development and Agriculture HYDE Global (8 km2) product into a series of discrete categories. Comparing across time properties/resources change (2011) ;2000 and aquaculture is not recommended. Landuse and landcover Transportation and service Vmap0 combines remote sensing data and local inventories. Transport and accessa Vmap0 NIMA (1997) Global Ca. 1995 change corridors Coverage is expected to have variable precision across the globe

Agriculture and aquaculture Near global 1970- 2011 Only on country level for the entire world, and with some gaps. Pollutants Pollutants and Natural system FAO maps FAO (2012) Selected regions 2013 Based on retail values. modifications

Agriculture and aquaculture chapter V Only selected Pollutants Pollutants and Natural system Aura OMI NASA (2013) 2012 Not global, and no temporal repeats regions 108 modifications

Invasive and other IUCN global De Poorter & Based on expert opions, and many species with only descriptive Invasive species Invasive species Variable None problematic species and genes database Browne 2005 ranges For some human and livestock diseases there has been a global Disease Disease N/A N/A N/A N/A N/A mapping, including malaria. Climate change and severe Woldclim Hijmans et al. No agreed form of baseline, and no one interpretation of negative Climate change Climate change Global (1km2) 1965 - weather temperature 2005 and positive changes. Climate change and severe Worldclim Hijmans et al. No agreed form of baseline, and no one interpretation of negative Climate change Climate change Global (1km2) 1965 - weather precipitation 2005 and positive changes. Residential and commercial Human population development and Some products have lower spatial resolution. Data is based on N/A GPWv3 CIESIN (2000) Global (1km2) 1990-2015 pressurea Transportation and service modeling population census data from national inventories. corridors Residential and commercial Some products have lower spatial resolution. Data is based on Human population development and N/A GRUMPv1 CIESIN (2012) Global (1km2) 1990-2015 modeling population census data from national inventories. Also pressure Transportation and service included secondary model parameters. corridors Residential and commercial Human population Stable Elvidge et al. N/A development and Energy Global (1km2) 1992 - pressurea Nightlights 1997 production and mining Legislation and TI corruption Increasing global coverage, some countries still missing. N/A N/A CPI (2012) Global 1980-2012 corruption index Changing methodologies impedes comparison Human Education and Only at country level. Some smaller changes has been N/A N/A development UNDP (2011) Global 1990-2012 livelihood implemented, care should be taken when comparing across time. index a included in the human footprint analysis of Sanderson et al. (2002). chapter V

Included datasets We used the stable night lights product (Elvidge et al. 1997a) to detect human infrastructure and some elements of human influence in natural and agricultural habitats. Stable nightlights have been shown to correlate with economic activities on ground, such that higher light values indicate higher economic outputs (Croft 1978; Elvidge et al. 1997b) and have been linked to on ground change in Gross Domestic Product (GDP) (Henderson et al. 2012). The stable night lights product detects lights from urban settlements, industrial sites, gas flares and wild fires, as well as reflected light from moonlit clouds (Elvidge et al. 1997a). Stable nightlights are measured on a scale between 0-63, ranging from no light (0) to maximum light intensity (63) using a threshold model based on the local conditions (For details on the method see Elvidge et al. 1997a).

We used Gridded Populations of the World version 3 (GPWv3) human population grid data developed by Center for International Earth Science Information Network (2005) to track human population changes over time. Increases in human density are linked to depletion of water resources (Shiklomanov 2000), increases in pollution (de Wit 2002; Newsome et al. 2010), and the increased need for land (Food and Agricultural Organization of the United Nations 2009), thus reducing the area of natural habitat and leading to biodiversity loss (Luck 2007). GPWv3 is the newest population estimate that has used a consistent method for all years and does not use secondary data for smoothing, thus reducing the risk of population estimates varying for reasons other than real population changes (Balk & Yetman 2004). The quality of the human population data is largely dependent on the underlying census data used for the estimation, which is of variable quality and spatial resolution. To assess the quality of the data - we compared census data from 10 countries across a range of Gross Domestic Products (Columbia, Denmark, Ghana, Japan, Laos, Mexico, Peru, Switzerland, USA, and Zambia) which had conducted on the ground populations census’ in either 2005 or 2010 - to the CIESIN data. A Wilcoxon’s rank test showed no significant differences between CIESIN modeled data and the census data (W = 35622; p = 0.99).

Excluded datasets Despite its overwhelming importance as a pressure on biodiversity, land cover change had to be excluded from our analysis. The original ‘human footprint’ analysis (Sanderson et al. 2002) used Land Cover version 2, based on monthly AVHRR-NDVI satellite data from 1992-1993 (Loveland et al. 2000). This layer has not subsequently been updated so a temporal comparison is impossible. Other products estimating global land cover classifications have been developed i.e. GLC2000 using MODIS land cover (Bartholomé & Belward 2005), and GlobeCover 300 (European Space Agency 2006), however these are methodologically different and produce different landcover estimates for the same areas (Fritz & See 2008). Only GlobeCover 300 has been repeated (2004 and 2009), but the rate of change is likely to be lower than the classification error; hence

109 chapter V spatially explicit comparisons between the two years is not recommended by the producers (Bontemps et al. 2010). Further, around 9% of the pixels in the 2009 map are taken directly from the 2004 map, and thus do not show any change at all. We also evaluated the HYDE data model (Goldewijk 2001) which estimates the percent human modified landscape in ca. 8km2 pixels. However despite the HYDE data going back 300 years, it only has a temporal resolution of 100 year intervals which is not appropriate for this use (Ellis et al. 2010).

Because the land cover change issue is so important for assessing the pressure on biodiversity, we still examined the impact of including change in GlobeCover300; finding no significant difference influence on our results when landcover changes were included, or excluded. GC300 is at a scale of ca. 300 m2 at Equator. We used both the original resolution as well as an aggregated version at the same resolution as our THPI (ca. 5km2 at Equator). GC300 consists of 22 land cover classes, describing different types of habitats (Table S1). Following from the system used in the Human Footprint (Sanderson et al., 2002) we reclassified all classes into one of four: i) Natural habitats, ii) rain-fed agriculture, iii) irrigated agriculture, and iv) urban areas.

Using this classification, land cover changes were found to be of little importance with 99.12% or 97.34% of the pixels experiencing no change for the 300m2 and 5km2 resolution maps respectively. The percentage of pixels experiencing changes was marginally higher when only looking at the subset confined to protected areas (Figure S4; Table S2). We also calculated the difference between our THPI scores with and without the inclusion of land cover changes, finding that changes were generally few and not very big (mean = 0.0016, 1st Qu. = -0.0008, 3rd Qu.= 0.0169, median = 0.0003) (Figure S5). Considering the methodological challenges for including land cover changes we therefore feel comfortable excluding it from the final THPI.

Human infrastructure (roads and railways) were also excluded as no temporally comparable layers exist. The original VMAP0 layers used in the 2002 human footprint analysis were based on national and sub-national data sources up till ca. 1995, without information on the date of road construction (National Imagery and Mapping Agency 1997). Attempts to update the VMAP0 have mainly been to expand and improve the coverage of the global road-infrastructure. Newer updates therefore contain roads that were present before 1995, but not included in the older maps, so it is not possible to develop the credible measure of change in road and railway density that would be required for our analysis.

All identified data layers of pollutants, invasive species, and diseases were excluded because they lacked sufficient spatial coverage. The Food and Agricultural Organization (FAO) tabulates statistics on fertilizers and pesticide use for most countries, but only at a country level, which was too coarse for our purposes. New maps are being developed using remote sensing to estimate specific air pollutants such as Nitrogen dioxide, but only at regional and national levels with no global products, and no repeated measures,

110 chapter V planned (National Aeronautics & Space Administration 2013). No global, spatially explicit, maps exist for the distribution and intensity of invasive alien species. IUCN has established an invasive species specialist group and a global invasive species database (De Poorter & Browne 2005), including more than 850 invasive species. However, information on species is mainly narrative and based on reported distributions within administrative units. Diseases pose a potentially significant threat to species and their impact and distribution can be affected by human actions, but there is no global database on the distribution of all or a representative subset of diseases (Hurlimann et al. 2011). Mapping the distribution of diseases is also complicated by their often rapid evolution, the lack of good tools for consistent classification, and the fact that most mapped diseases that affects humans (e.g. malaria, see Snow et al. 2005), are harmless to most other species.

We evaluated the Human Development Index (United Nations Development Programme 2011) which has been calculated globally since 1990, Transparency International’s Corruption Perception Index (CPI), and infant mortality rates (United Nations Children’s Fund, 2011). All were excluded because none were mapped at subnational scale for the entire world. Furthermore, the CPI is based on national inventories that cannot be compared between years.

Estimating pressure change Data from our two global pressure layers were spatially aggregated to a resolution of 2.5 arc minutes (approximately 5km2 at Equator), from their original resolution (ca. 2.8 km2 for stable nightlight and census zones for human population density) to reduce biases from slight miss-alignment of data sources. For each terrestrial pixel we calculated the difference between values in 1995 and 2010. This was done separately for stable nightlights and human population density. Human population density was square-root transformed to account for the reduced impacts of changes in heavily populated areas (see justification and analytical work in appendix II). The result was two maps displaying the change in absolute values of human population density and stable nightlights respectively, such that negative values indicated pixels with decreasing pressure between 1995 and 2010 and positive values denote pixels with increases in pressure. The absolute values for each layer were subsequently standardized on a scale between -1 to 1, giving equal weight to the maximum changes in human population density or stable nightlights. The two layers were then combined and standardized on a scale between -100 and 100, where positive values denote increased human pressure and negative values reduced human pressure. This process of standardization gives our Temporal Human Pressure Index (THPI), which is a value between -100 and 100, and presents changes in human pressure over 15 years for each 5x5 km terrestrial pixel.

Geographical divisions of the world We divided the world into biogeographical realms and biomes within realms following Olson et al. (2001). The original seven realms were reduced to six: i) the Afrotropic

111 chapter V

(AT), ii) the Indo-Malay (IM) (here referred to as South East Asia), iii) the Nearctic (NA) (here referred to as North America), iv) the Neotropic (NT) (here referred to as Latin America), v) the Palearctic (PA) (here referred to as Euro-Asia), and vi) Australia (AA) and Oceania (OC) combined.

Protected Areas We used the World Database on Protected Areas (WDPA) from February 2012 for all information on size, shape, location, and the IUCN management category of protected areas, which is a system of defining the management approach of the protected area – ranging from strict nature reserves with no human entry (IUCN 1a), through to protected areas that are sustainably used for human benefit (IUCN VI) (For details on the IUCN categories see Dudley 2008). All protected areas without information on date of establishment, as well as protected areas established after 1995, were excluded. Protected areas smaller than 50ha were also removed. The Netherlands and USA were, with the exception of a few protected areas, also excluded because information on their dates of establishment was not available in the WDPA. For overlapping protected areas with different IUCN categories, we did a stepwise erase; removing the highest (lowest level of management) IUCN category, always assigning the strictest IUCN management category (following from Jenkins & Joppa 2009; Joppa & Pfaff 2009, 2011). After this pre-processing we were left with a database of 27,795 protected areas for further analysis (see appendix II).

Data analyses The spatial extent of the layers was restricted to 75º to -58º longitude and -180º to 180º latitude, based on DMSP-OLS stable nightlights full extent. All non-terrestrial areas were removed using the Global Self-consistent Hierarchical High-resolution Geography (GSHHG) version 2.1 from NOAA (Wessel & Smith 1996). All layers were projected using Mollweide equal area projection.

ArcGIS Zonal analysis tool was used to estimate the difference in average THPI scores between ecoregions of the world as well as between protected areas of different IUCN management categories. An analysis of Covariance (ANCOVA) was used to test the relationship between THPI values of protected areas and IUCN management categories. Finally, we used a linear regression model to examine the correlation between the Human Influence Index (HII) (Sanderson et al. 2002) against changes in human pressure (THPI) for biomes within realms to test how our index compares to that developed using the earlier approach.

All spatial data management and analysis were conducted using ESRI ArcGIS 10 and all statistical analysis was conducted using R v. 2.14.1.

112 chapter V

Results

Changes within geopolitical units THPI scores illustrating changes in human pressure for terrestrial areas between 1995 and 2010 were mapped globally (Figure 1). Using our scoring approach, which ranges from -100 to +100 change over 15 years, 69% of the world’s terrestrial area experienced an increase in human pressure. Less than 1% of terrestrial area experienced negative changes greater than a THPI score = 95 and only ca. 5% was changed to a level greater than THPI = 5 (these breakpoints were chosen to illustrate the extreme ends of what was possible in our analysis).

The majority of the large changes were either around urban areas, or where oil and gas exploration is occurring. For example, large positive changes are seen in and around cites in North America (Figure 2a) and South-East Asia (Figure 2b). Moreover, areas experiencing intense increase in human pressure were, in some parts of the world, adjacent to areas of intense decrease in human pressure. Using high resolution satellite imagery for validation we could identify these to be point localities of resource extraction such as oil drillings in Nigeria (Figure 2c), or natural gas sites in the Russian tundra (Figure 2d).

Changes within natural units The Indo-Malay realm experienced the on average the greatest increases in THPI over the 15 years, while Australia-Oceania experienced the smallest negative increases with a ca. 27 fold difference between the two realms. The Afrotropic, Euro-Asia, and Latin America experienced intermediate negative increases in pressure ca. 3.5 fold less than the Indo-Malay (Figure S8).

Of the 14 global biomes, all except tundra experienced increases in pressure between 1995 and 2010. Mediterranean forest and woodland experienced the greatest increase in pressure, followed by temperate broadleaf forest and flooded grasslands. Taiga, deserts, and mountains experienced the smallest increases (Figure 3). When analyzing pressure changes in biomes within realms (n=62), only six experienced a decrease in human pressure between 1995 and 2010 (Table S3).

We also found a significant positive correlation (R2 = 0.42; p < 0.001) between the accumulated human pressure, the Human Influence Index, in biomes within realms, measured around 1995 (Sanderson et al. 2002, table 2), and the change in human pressure between 1995 and 2010 (Figure 4).

113 chapter V 114

Figure 1. Global distribution of the Temporal Human Pressure Index (THPI), where red indicates increased and green indicates decreased pressure between 1995 and 2010. Yellow areas are where pressure has not changed over time. Blue shows area of sea where the pressure metric has not been calculated. Change metrics are not calculated for land from 75º north and from -58º south, omitting the North and South Polar regions. chapter V

Figure 2. Changes in human pressure between 1995 and 2010 in different areas of the world. Red values indicate increases in human pressure and green values indicates decreases in human pressure. Larges increases can be observed around cities in North America (a), and South East Asia (b). Patterns of large increases adjacent to large decreases were observed in area of heavy resource extraction like oil 100 drillings in Nigeria (c) or natural 0 - 62 gas sites in the Russian tundra (d).

Changes within Protected Areas Differences between IUCN management categories were significant for average THPI values (F = 5.104; p < 0.001). Globally, protected areas in all IUCN management categories experienced an increase in human pressure from 1995 to 2010 (Figure 5).

WWF Biomes

Boreal forests/taiga Deserts and xeric shrublands Flooded grasslands and savannas Mangrove Mediterranean forests, woodlands, and scrub Montane grasslands and shrublands Temperate broadleaf and mixed forests Temperate coniferous forests

Temperate grasslands, savannas, and shrublands

Tropical and subtropical coniferous forests Tropical and subtropical dry broadleaf forests

Tropical and subtropical grasslands, savannas, and shrublands Tropical and subtropical moist broadleaf forests

Tundra

- 0.1 0 0.5 1.0 1.5 2.0 2.5 3.0 THPI Figure 3. Average changes in THPI across biomes for the world. Negative values indicates decreases in human pressure and positive values increases in human pressure between 1995 and 2010.

115 chapter V

Figure 4. Correlation between accumulated human pressure by circa 1995 (Human Influence Index) against changes in human pressure from 1995-2010 (THPI). Each point represents a biome within realm. The units are not directly comparable between HP and THPI.

Categories IV (habitat/species management area) and V (protected landscape) experienced the greatest pressure increase, while category Ib (wilderness areas) experienced the lowest. According to our results, aggregated protected areas in category VI (protected area with sustainable use of natural resources) experienced lower changes in pressure than protected areas in IUCN categories III (Natural Monument or Feature), IV and V.

Patterns of change in pressure within protected areas at the continental and realm scales did not mimic those globally. Across 5 of 6 global realms, the exception being North America, aggregated protected areas in different IUCN categories experienced increased human pressure (Figure 6). Protected areas in South East Asia experienced the greatest pressure increases, with little difference between management categories. Sub-Saharan protected areas experienced the greatest pressure increase in category III, while Australia and Oceania experienced greatest changes in category IV, and Latin America and Euro-Asia experienced them in category VI. No one category stood out as consistently best or worst.

116 chapter V

Figure 5. Average change in pressure (THPI) 1995-2010 in 11,608 protected areas of different IUCN categories globally. Positive THPI values indicate increased pressure over the 15 years. THPI scores are logged. Outliers have been removed. Discussion This study provides the first attempt to map changes in human pressure globally at a scale that is meaningful for conservation planning and can be used to evaluate the impact of conservation interventions such as protected areas. Our results show that the majority of the world has experienced increases in human pressure, with areas already under greater pressure experiencing the greatest increases. City fringes experienced high negative increases, corresponding well to the global urbanization patterns, where more than 50% of all humans now live in cities around the world (United Nations Population Fond 2007). Also African savannas, and European Mediterranean scrub- and flooded grassland (Figure 4) experienced considerably greater increases than what would have been expected from the accumulated pressure prior to the period of study. For African savannas, the changes we measure are supported by land cover changes (Riggio et al. 2012).

Previous studies have shown that habitat loss inside protected areas is lower in reserves which are more strictly managed using the IUCN management categories (Joppa et al. 2008; Scharlemann et al. 2010; Nelson & Chomitz 2011). However our results indicate that the relationship between change in pressure and IUCN management categories is more complex, and we find distinct patterns across realms, while our global average conforms better to our expectations of the impact of IUCN management categories.

Category VI protected areas together with Category V, are often considered to be of less value for biodiversity conservation (Locke & Dearden 2005), but our results show that Category VI protected areas had a lower increase in pressure than Category IV or V. A

117 chapter V similar pattern was observed using the Human Influence Index (Leroux et al. 2010). The observed differences between regions may not be surprising considering the IUCN categories have never been intended to be interpreted on an ordinal scale (Dudley et al. 2010), and our results highlight that IUCN categories are not necessarily a good proxy for protected area performance.

Figure 6. Average change in pressure (THPI) between 1995-2010 in protected areas within different geographical regions. Positive values indicate increased pressure over the 15 years, negative values indicate improvement. Geographical categories based on WWF realms: Australia and Oceania (AA and OC), Sub-Saharan Africa (AT), South East Asia (IM), North America (NA), Latin America (NT), and Euro-Asia (PA).

Challenges We identify six major challenges when interpreting our results. First; for both stable nightlights and human population, increases in values correspond with increases in pressure. However this relationship is not necessarily linear, and studies have suggested that per capita increase, for example of CO2 (Dietz & Rosa 1997), or material consumption (Bringezu et al. 2004), levels off or even declines with increasing

118 chapter V economic wealth and technological advances (Shafik 1994). Thus translating the THPI scores into actual impact measures is far from trivial. Second; the underlying driver of change may not always have the same impact on the ground. For example; besides electrical outlets, stable nightlights capture gas flares and other sites for resource extraction (Elvidge et al. 2009; Ghosh et al. 2010), as well as wild fires (Elvidge et al. 2001; Chand et al. 2007). While all components will have negative impact they might be very different, which can’t be extracted from the data. Third; when examining composite data layers, one major challenge is weighting the influence of one source compared to another. This is particularly challenging for data sources where weights can depend on idiosyncratic decisions and expert opinion (Malczewski 2006). Different Multicriteria Decision Analysis (MCDA) have been developed to evaluated multiple factor data, avoiding the most obvious pitfalls of comparing apples and oranges (Williams & Araújo 2002). The specific evaluation of these composite data sources all depends on the setting of targets that can be evaluated and compared (Csuti et al. 1997; Cowling et al. 1999; Araújo & Williams 2000). However, we do not select areas or evaluate decisions under a range of possible future scenarios but only to capture the change that has already occurred. This does not make the consideration of weighting between data sources less important, but does challenge the evaluation of which approach to take. Fourth; though stable nightlights and human populations are independent products and describe two distinct sources of pressure they will be inter- correlated. Areas of high population density will be predisposed to have higher values of stable nightlight, (Figure S7). This is not always true and areas with high population density, but relatively low stable nightlights, will generally be areas of smaller ecological footprint than areas scoring high in both (World Wildlife Fund 2012). Fifth; our map is at 5km2 grid scale. Yet some impacts originating from the two layers may have an effect on much larger scales making the individual pixel values a poor unit for comparison. Sixth; although we are able to present a temporally and spatially explicit map of change in human pressure, this does not include many instrumental drivers of biodiversity or habitat loss. By far the biggest challenge, however, is accessing datasets that are appropriate for the types of spatio-temporal analyses so desperately needed to address the biodiversity crisis.

A Paucity of Data The objective of this analysis was to map changes in anthropogenic pressure on land over time, using available and appropriate data of an acceptable quality and a sufficient spatial resolution. Our review of potential datasets that cover a range of pressures on biodiversity from humans (Salafsky et al. 2008; Baldwin 2010), illustrates the challenges in assessing comparable human pressures over time. Because few datasets met our criteria for inclusion, as they were either not sufficiently spatially resolved, had limited coverage, or were not comparable over time, our resulting pressure change metric is simple, only including stable nightlights and interpolated census estimations of human population. Despite that, we believe our efforts are relevant for mapping human

119 chapter V pressure on wild nature: both components we included have been shown to impact nature – and our map can provide the initial architecture of a much more comprehensive spatio-temporal dataset into the future.

We suggest that the current paucity of appropriate data to map spatial and temporal patterns of threat is to some extent a reflection of methods more than scientific questions driving data generation. We acknowledge the need for improvement in methods; however we also see a need for more data generation that allows for comparison over time. The advances in methods for data collection increases the availability of large-scale datasets as well as our knowledge about the world today, but it often neglects that conservation science is based on understanding ‘effects’ and thus changes over times of both pressures and responses (Dornelas 2010; Magurran & Dornelas 2010). For example: land-use change is a vitally important metric for conservation threat assessment and a major driver of biodiversity loss (Millinium Ecosystem Assessment 2005; United Nations Environmental Programme 2012) and has been called the most significant threat to nature (Vitousek et al. 1997). However no spatially and temporally comparable land-cover classification maps at fine scale and using a comparable classification exist for the globe over even a 10 year time period. This is despite the existence of several static products. We see a need for an increased focus on spatial data that can be compared over time, even if this comes at the expense of quality. Our analysis here is a step towards this goal, but also highlights the challenge in making an overall representation of human pressures and their changes around the world at scales that can be used for conservation decision-making.

Acknowledgements We thank the Danish National Research Foundation for financial support. We also thank the IUCN SSC/WCPA Joint Task-Force on Biodiversity and Protected Areas, UNEP-WCMC, WWF, and Microsoft research for financial and institutional support.

References Araújo, M. B., and P. H. Williams. 2000. Selecting areas for species persistence using occurrence data. Biological Conservation 96:331-345. Baldwin, R. F. 2010. Identifying Keystone Threats to Biological Diversity. Pages 17-32 in S. C. Trombulak, and R. F. Baldwin, editors. Landscape-scale Conservation Planning. Springer Netherlands. Balk, D., and G. Yetman. 2004. The Global Distribution of Population: Evaluating the gains in resolution refinement. Center for International Earth Science Information Network (CIESIN), Columbia University, Palisades, New York, USA. Balmford, A., P. Carey, V. Kapos, A. Manica, A. S. L. Rodrigues, J. P. W. Scharlemann, and R. E. Green. 2009. Capturing the Many Dimensions of Threat: Comment on Salafsky et al. Conservation Biology 23:482-487. Barnosky, A. D., et al. 2011. Has the Earth's sixth mass extinction already arrived? Nature 471:51-57.

120 chapter V

Bartholomé, E., and A. S. Belward. 2005. GLC2000: a new approach to global land cover mapping from Earth observation data. International Journal of Remote Sensing 26:1959-1977. Bontemps, S., P. Defourny, and E. Van Bogaert. 2010. CLOBECOVER 2009 - Product description and validation report. European Space Agency. Bringezu, S., H. Schutz, S. Steger, and J. Baudisch. 2004. International comparison of resource use and its relation to economic growth - The development of total material requirement, direct material inputs and hidden flows and the structure of TMR. Ecological Economics 51:97-124. Butchart, S. H. M., et al. 2010. Global Biodiversity: Indicators of Recent Declines. Science 328:1164-1168. Center for International Earth Science Information Network. 2005. Gridded Population of the World, Version 3 (GPWv3), Palisades, New York. Available from http://sedac.ciesin.columbia.edu/gpw Chand, T. R. K., K. V. S. Badarinath, M. S. R. Murthy, G. Rajshekhar, C. D. Elvidge, and B. T. Tuttle. 2007. Active forest fire monitoring in Uttaranchal State, India using multitemporal DMSP-OLS and MODIS data. International Journal of Remote Sensing 28:2123-2132. Convention on Biological Diversity. 2010. Strategic Plan for Biodiversity 2011-2020 - COP 10, decision X/2. Convention on Biological Diversity Available from http://www.cbd.int/decision/cop/?id=12268 Cowling, R. M., R. L. Pressey, A. T. Lombard, P. G. Desmet, and A. G. Ellis. 1999. From representation to persistence: Requirements for a sustainable system of conservation areas in the species-rich mediterranean-climate desert of southern Africa. Diversity and Distributions 5:51-71. Croft, T. A. 1978. Nighttime Images of Earth from Space. Scientific American 239:68-79. Csuti, B., et al. 1997. A comparison of reserve selection algorithms using data on terrestrial vertebrates in Oregon. Biological Conservation 80:83-97. De Poorter, M., and M. Browne. 2005. The Global Invasive Species Database (GISD) and international information exchange: using global expertise to help in the fight against invasive alien species. BCPC Symposium Proceedings 81:49-54. de Wit, C. A. 2002. An overview of brominated flame retardants in the environment. Chemosphere 46:583-624. Dietz, T., and E. A. Rosa. 1997. Effects of population and affluence on CO2 emissions. Proceedings of the National Academy of Sciences of the United States of America 94:175-179. Dornelas, M. 2010. Disturbance and change in biodiversity. Philosophical Transactions of the Royal Society B: Biological Sciences 365:3719-3727. Dudley, N. 2008. Guidelines for Applying Protected Area Management Categories. International Union for Conservation of Nature, Gland, Switzerland. Dudley, N., J. D. Parrish, K. H. Redford, and S. Stolton. 2010. The revised IUCN protected area management categories: the debate and ways forward. Oryx 44:485-490. Ellis, E. C., K. Goldweijk, K., S. Siebert, D. Lightman, and N. Ramankutty. 2010. Anthropogenic transformation of the biomes, 1700 to 2000. Global Ecology and Biogeography 19:589-606. Elvidge, C. D., K. E. Baugh, E. A. Kihn, H. W. Kroehl, and E. R. Davis. 1997a. Mapping City Lights With Nighttime Data from the DMSP Operational Linescan System. Photogrammetric Engineering & Remote Sensing 63:727-734.

121 chapter V

Elvidge, C. D., K. E. Baugh, E. A. Kihn, H. W. Kroehl, E. R. Davis, and C. W. Davis. 1997b. Relation between satellite observed visible-near infrared emissions, population, economic activity and electric power consumption. International Journal of Remote Sensing 18:1373-1379. Elvidge, C. D., I. Nelson, V. R. Hobson, J. Safran, and K. E. Baugh. 2001. Detection of fires at night using DMSP-OLS data. Pages 125-144 in F. J. Ahern, J. G. Goldammer, and C. O. Justice, editors. Global and Regional Vegetation Fire Monitoring from Space: Planning a Coordinated International Effort. SPB Academic Publishing, The Hauge, The Netherlands. Elvidge, C. D., D. Ziskin, K. Baugh, B. Tuttle, T. Ghosh, D. Pack, E. Erwin, and M. Zhizhin. 2009. A Fifteen Year Record of Global Natural Gas Flaring Derived from Satellite Data. Energies 2:595-622. European Space Agency. 2006. GlobCover Project led by MEDIAS-France. European Space Agency Available from http://ionia1.esrin.esa.int/news/_faq.asp?id=35 Food and Agricultural Organization of the United Nations. 2009. Global agriculture towards 2050. Page 4. High-Level Expert Forum. FOA, Rome, Italy. Fritz, S., and L. See. 2008. Identifying and quantifying uncertainty and spatial disagreement in the comparison of Global Land Cover for different applications. Global Change Biology 14:1057-1075. Ghosh, T., C. D. Elvidge, P. C. Sutton, K. E. Baugh, D. Ziskin, and B. T. Tuttle. 2010. Creating a Global Grid of Distributed Fossil Fuel CO2 Emissions from Nighttime Satellite Imagery. Energies 3:1895-1913. Goldewijk, K. K. 2001. Estimating global land use change over the past 300 years: The HYDE Database. Global Biogeochemical Cycles 15:417-433. Henderson, J. V., A. Storeygard, and D. N. Weil. 2012. Measuring Economic Growth from Outer Space. American Economic Review 102:994-1028. Hoffmann, A. A., and C. M. Sgro. 2011. Climate change and evolutionary adaptation. Nature 470:479-485. Hurlimann, E., et al. 2011. Toward an Open-Access Global Database for Mapping, Control, and Surveillance of Neglected Tropical Diseases. Plos Neglected Tropical Diseases 5. Intergovernmental Panel on Climate Change. 2007. Climate Change 2007 – Impacts, Adaptation and Vulnerability. the Intergovernmental Panel on Climate Change, Cambridge, UK. Jenkins, C. N., and L. N. Joppa. 2009. Expansion of the global terrestrial protected area system. Biological Conservation 142:2166-2174. Joppa, L. N., S. R. Loarie, and S. L. Pimm. 2008. On the protection of protected areas. Proceedings of the National Academy of Sciences 105:6673-6678. Joppa, L. N., and A. Pfaff. 2009. High and Far: Biases in the Location of Protected Areas. PLoS ONE 4:e8273. Joppa, L. N., and A. Pfaff. 2011. Global protected area impacts. Proceedings of the Royal Society B-Biological Sciences 278:1633-1638. Leroux, S. J., M. A. Krawchuk, F. Schmiegelow, S. G. Cumming, K. Lisgo, L. G. Anderson, and M. Petkova. 2010. Global protected areas and IUCN designations: Do the categories match the conditions? Biological Conservation 143:609-616. Locke, H., and P. Dearden. 2005. Rethinking protected area categories and the new paradigm. Environmental Conservation 32:1-10.

122 chapter V

Loveland, T. R., B. C. Reed, J. F. Brown, D. O. Ohlen, Z. Zhu, L. Yang, and J. W. Merchant. 2000. Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data. International Journal of Remote Sensing 21:1303-1330. Luck, G. W. 2007. A review of the relationships between human population density and biodiversity. Biological Reviews 82:607-645. Magurran, A. E., and M. Dornelas. 2010. Biological diversity in a changing world. Philosophical Transactions of the Royal Society B: Biological Sciences 365:3593-3597. Malczewski, J. 2006. GIS-based multicriteria descision analysis: a survey of the literature. International Journal of Geographical Information Science 20:703-726. McCloskey, J. M., and H. Spalding. 1989. A Reconnaissance-level Inventory of the Amount of Wilderness Remaining in the World. Ambio 18:221-227. Millinium Ecosystem Assessment. 2005. Millinium Ecosystem Assessment: Ecosystems and Human Well-being: Biodiversity Synthesis. World Resources Institute, Washington, DC. National Aeronautics & Space Administration. 2013. Pollution across Southwestern Asia. NASA Available from http://earthobservatory.nasa.gov/IOTD/view.php?id=80200 National Imagery and Mapping Agency. 1997. Vector Map Level 0 (VMAP0) ed. 003. National Imagery and Mapping Agency, Washington (DC). Available from http://egsc.usgs.gov/nimamaps/ Nelson, A., and K. M. Chomitz. 2011. Effectiveness of Strict vs. Multiple Use Protected Areas in Reducing Tropical Forest Fires: A Global Analysis Using Matching Methods. PLoS ONE 6:e22722. Newsome, S. D., J. S. Park, B. W. Henry, A. Holden, M. L. Fogel, J. Linthicum, V. Chu, and K. Hooper. 2010. Polybrominated Diphenyl Ether (PBDE) Levels in Peregrine Falcon (Falco peregrinus) Eggs from California Correlate with Diet and Human Population Density. Environmental Science & Technology 44:5248-5255. Olson, D. M., et al. 2001. Terrestrial Ecoregions of the World: A New Map of Life on Earth. Bioscience 51:933-938. Organisation for Economic Co-operation and Development. 1993. OECD core set of indicators for environmental performance reviews. OECD, Paris, France. Parmesan, C., and G. Yohe. 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37-42. Peres, C. A. 2000. Effects of subsistence hunting on vertebrate community structure in Amazonian forests. Conservation Biology 14:240-253. Riggio, J., et al. 2012. The size of savannah Africa: a lion’s (Panthera leo) view. Biodiversity and Conservation:1-19. Salafsky, N., et al. 2008. A Standard Lexicon for Biodiversity Conservation: Unified Classifications of Threats and Actions. Conservation Biology 22:897-911. Sanderson, E. W., M. Jaiteh, M. A. Levy, K. H. Redford, A. V. Wannebo, and G. Woolmer. 2002. The Human Footprint and the Last of the Wild. Bioscience 52:891-904. Scharlemann, J. P. W., V. Kapos, A. Campbell, I. Lysenko, N. D. Burgess, M. C. Hansen, H. K. Gibbs, B. Dickson, and L. Miles. 2010. Securing tropical forest carbon: the contribution of protected areas to REDD. Oryx 44:352-357. Shafik, N. 1994. Economic Development and Environmental Quality: An Econometric Analysis. Oxford Economic Papers-New Series 46:757-773. Shiklomanov, I. A. 2000. Appraisal and assessment of world water resources. Water International 25:11-32.

123 chapter V

Snow, R. W., C. A. Guerra, A. M. Noor, H. Y. Myint, and S. I. Hay. 2005. The global distribution of clinical episodes of Plasmodium falciparum malaria. Nature 434:214- 217. United Nations Children’s Fund. 2011. Levels & Trends in Child Mortality - Report 2011. United Nations Inter-agency Group for Child Mortality Estimation, New York, USA. United Nations Development Programme. 2011. Human Development Report 2011: Sustainability and Equity: A Better Future for All. UNDP, New York, USA. United Nations Environmental Programme. 2012. Global Environment Outcome 5: environment for the future we want. UNEP, Nairobi, Kenya. United Nations Population Fond. 2007. State of the World Population 2007: unleashing the potential of urban growth. UNPF, New York, USA. Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997. Human domination of Earth's ecosystems. Science 277:494-499. Wessel, P., and W. H. F. Smith. 1996. A global, self-consistent, hierarchical, high-resolution shoreline database. Journal of Geophysical Research-Solid Earth 101:8741-8743. Wilkie, D. S., and J. F. Carpenter. 1999. Bushmeat hunting in the Congo Basin: an assessment of impacts and options for mitigation. Biodiversity and Conservation 8:927-955. Williams, P., and M. Araújo. 2002. Apples, Oranges, and Probabilities: Integrating Multiple Factors into Biodiversity Conservation with Consistency. Environmental Modeling & Assessment 7:139-151. World Wildlife Fund. 2012. Living Planet Report 2012 - Biodiversity, biocapacity and better choices. WWF, Gland, Switzerland. Zalasiewicz, J., M. Williams, W. Steffen, and P. Crutzen. 2010. The New World of the Anthropocene. Environmental Science & Technology 44:2228-2231.

124

CHAPTER VI The effect of management in protected areas on species populations

Jonas Geldmann, Lauren Coad, Ben Collen, Megan Barnes, Ian Craigie, Thomas Brooks, Luke Harrison, Marc Hockings, Sarah Whitmee, Stephen Woodley, and Neil D Burgess

All supplementary material refers to appendix III

125

126 chapter VI

Management effectiveness of protected areas in 17 countries and the maintenance of species populations

Jonas Geldmanna, Lauren Coadb, Ben Collenc, Megan Barnesd,e, Ian Craigief, Thomas Brooksg , Luke Harrisonh, Marc Hockingsd, Sarah Whitmeei, Stephen Woodleyj, Neil D. Burgessa,k,l

a Center for Macroecology, Evolution and Climate, Natural History Museum of Denmark, University of Copenhagen, Denmark, b Environmental Change Institute, School of Geography, University of Oxford, Oxford, OX1 3QY, United Kingdom, c Centre for Biodiversity and Environmental Research, Department of Genetics, Evolution & Environment, University College London, United Kingdom, d School of Geography, Planning and Environmental Management, University of Queensland, Australia, e Environmental Decisions Group, Australia, f ARC Centre of Excellence for Coral Reef Studies, James Cook University, Australia, g International Union for the Conservation of Nature, Gland, Switzerland, h McGill University, Canada, i Zoological Society of London, j International Union for the Conservation of Nature, Chelsea, Canada, k United Nations Environmental Programme, World Conservation Monitoring Centre, Cambridge, United Kingdom, l WWF US Conservation Science Program, Washington DC, USA

Abstract Protected areas are amongst the most important conservation responses against the continued loss of biodiversity, and today cover more than 12.7% of the terrestrial surface. However we still know very little about the extent to which they are effective in protecting biodiversity, and in particular how management affects their performance. Here we analyze how the quality of management in protected areas globally, as measured by the Management Effectiveness Tracking Tool (METT) affects the trajectory of biodiversity using Living Planet Database (LPD) to track changes in animal populations. We further include socio-economic parameters to help explain the observed patterns. After spatially matching 4,366 population time-series and 3,023 METT assessments, our final sample covers 279 vertebrate populations across 53 protected areas from 17 countries with quantitative and independently sampled data on management and change in biodiversity. We show that the existence and implementation of management plans combined with increasing Human Development Index (HDI) improves changes in populations. We also show that population increased both before and after the period where METT was implemented, suggesting that our subset of protected areas is amongst the better managed ones. We conclude that management input and quality does have an impact on species outcomes. However, matched data on species trends and management effectiveness are few, reducing the sample size such that statistical generalization is problematic.

Keywords: Vertebrate population trends, Living Planet Index, World Database on protected areas, protected area management effectiveness

127 chapter VI

Introduction It is increasingly being recognized that the world is experiencing a biodiversity crisis resulting in habitat loss and declines in animal populations (Butchart et al. 2010). Setting aside land for the protection of nature has been proposed as a key strategy for reversing the present negative trajectory; which has resulted in an expanding network of more than 170,000 protected areas globally, covering ca. 12.7% of the terrestrial land surface (Bertzky et al. 2012). However, the extent to which protected areas are effective at safeguarding biodiversity is widely discussed (Joppa & Pfaff 2010; Barr et al. 2011), and it has been suggested that many of the worlds protected areas are present only as ‘paper parks’ lacking effective management, sometimes even boundaries and demarcation on the ground, making them likely to fail as a conservation tool (Dudley & Stolton 1999; McKinney 2002).

Past analyses of protected area effectiveness have focused on whether protected areas are located in the right places, by looking at overlaps between species ranges and protected areas (Rodrigues et al. 2004; Butchart et al. 2012). However, although location is an important prerequisite for successful reserves, such ‘gap analyses’ do not provide information on how well protected areas are performing in terms of conservation outcomes (Kapos et al. 2009). Ideally measuring the effectiveness of protected areas would involve tracking changed in a relevant biodiversity measure (e.g. change in habitat cover or species populations) over time in a controlled and unbiased manner (Ferraro 2009; Joppa & Pfaff 2009). This approach requires comparing the land within the protected area against comparable, but not protected, land or correlating actions and interventions in the protected areas against changes in biodiversity.

In an attempt to get at the question of protected area effectiveness, studies have used the protected area management categories developed by the International Union for the Conservation of Nature’s (IUCN) (Dudley 2008) as proxy values for level of conservation effort. These categories have been assigned to more than 92% of reserves globally (Bertzky et al. 2012). Stricter management categories have been shown to correlate with reduced rates of habitat loss (Scharlemann et al. 2010; Joppa & Pfaff 2011) as well as lower levels of human impact (Leroux et al. 2010). However, IUCN management categories have weaknesses in that they are not used in all countries, are interpreted differently around the world, are not assigned based on evaluation of on- ground activities, and depend on national legislation to provide context for how management will be undertaken. Alternatively studies have looked at how changes in red list status correlated with the quality of protected areas (Butchart et al. 2012) or whether governance structures changes the persistence of species inside protected areas (Tranquilli et al. 2012). However these studies do not include direct measures of conservation actions and interventions undertaken in the protected areas. We propose that changes in native and naturally occurring populations offer a good measure of conservation outcomes, and in combination with standardized scores of management

128 chapter VI input, can be used to assess the effectiveness of protected areas in delivering success or failure for biodiversity.

Protected Area Management Effectiveness (PAME) systems and evaluations have been developed and applied in more than 6,500 sites globally (Coad et al. 2013a), using more than 50 different methodologies (Leverington et al. 2010a). PAME tools often depend on local area managers who score management quality covering a range of input, activities and outputs of the protected area (Hockings 2003; Hockings et al. 2006). Protected areas conducting PAME are likely to be amongst the better functioning ones often receiving external funding. However, synthesis in 2010 showed that, more than 50% of the reserves where PAME tools have been completed had low scores, and thus suffered from major or significant deficiencies, indicating that conducting PAME evaluations does not always led to successful protected areas (Leverington et al. 2010b). Evidence of whether improved management is related to enhanced conservation outcomes in protected areas around the world is therefore still needed.

One of the most widely used methodologies is the Management Effectiveness Tracking Tool (METT) (Stolton et al. 2007), which has been applied in more than 1,600 sites globally (Coad et al. 2013b), and is a standard tool in all GEF-funded projects, and widely used by the World Bank, WWF, and other agencies (Coad et al. 2013a). METT reports on staff numbers, protected area budgets, threats as well as the adequacy and compliance of 30 specific questions ranging from implementation of management plans, staff training, management of budgets, outreach programs, local collaborations, and tourist facilities (Table S1) (Stolton et al. 2007). Ideally, METT thus produces a very comprehensive evaluation of the management performance of the protected area.

In this paper we investigate the correlation between individual METT scores (Leverington et al. 2010b) and changes in population time-series inside protected areas using the largest available global data-set for both Management Effectiveness (Coad et al. 2013b) and vertebrate population changes (Loh et al. 2005; Collen et al. 2009). We first calculate the greatest overlap between these two databases on input and outcomes respectively, and the World Database on Protected Areas (WDPA) (World Centre for Monitoring Conservation 2012), to obtain a global dataset for further analysis. We subsequently evaluate a suite of explanatory datasets which are also expected to influence changes in populations, covering landscape and socio-economic parameters. Finally we build models and test correlations between management inputs and species outcomes within protected areas, using the largest available dataset of independent data.

Methods Measuring biological outcomes Species population trends were obtained from the Living Planet Database (LPD), which contains more than 12,000 time-series across 2,500 vertebrate species (Collen et al. 2009). The LPD collates data from published scientific literature, on-line databases,

129 chapter VI large scale monitoring schemes (e.g. Pan-European Common Bird Monitoring Scheme) and grey literature (Loh et al. 2005). We only considered time-series within protected areas and added to the database before 29 November 2012. Only terrestrial mammals, birds and reptiles were included in the analysis. Species that are only partly dependent on land (i.e. Sea Turtles (Chelonioidea), Pinnipeds, or sea birds) were also included if a key life stage occurred inside a protected area. The LPD contains time-series measured at varies spatial scales and using a number of different change matrices (e.g. abundance, number of individuals, occupancy, breeding sites etc.). Following Collen et al. (2009), we considered all time-series as long as they had used a consistent methodology throughout the period of sampling.

For all population time-series we calculated the population slope (dependent variable), by fitting a generalized linear regression model (GLM) with a log-link function (Mountford 1982; Barnes et al. in prep). Values of zero in the beginning or end of the time-series were excluded, and slope values exceeding ± 0.5 on the log-scale, as well as time-series of less than five years, were removed. Slopes were calculated using two thresholds of at least two or three observations in a time-series. Slopes values were calculated using only data between 2000 and 2010 as this was considered to be a reasonable number of years for which the METT scores were valid.

For all species included in the analysis we extracted information on their body mass. For birds we used the CRC Handbook of Avian Body Masses (Payne 2009), for Ducks Geese and Swans (Kear 2005) and Handbook of the Birds of the World (del Hoyo et al. 1992), for mammals we used the PanTHERIA database (Jones et al. 2009) and Handbook of the Mammals of the World (Wilson & Mittermeier 2009). Body masses for green turtle (Chelonia mydas) were extracted from Hays et al. (2002), leatherback sea turtle (Dermochelys coriacea) from Wallace et al. (2005), and the Yellow-spotted Amazon River Turtle (Podocnemis unifilis) from Fachín-Terán and Vogt (2004). For all species where these data sources did not provide information we consulted experts within the IUCN SSC species specialist groups. All body mass data was log transformed for analysis. Information on threat status was extracted from the IUCN Red List (www.iucnredlist.org, (International Union for Conservation of Nature 2012).

Protected Area location and attribute data Data on protected areas were obtained from the World Database on Protected Areas (WDPA) managed by IUCN and UNEP-WCMC (World Centre for Monitoring Conservation 2012), which contains information on ca. 170,000 protected areas including their IUCN management categories, area, country, and location. Several protected areas have multiple designations and some population time-series were cataloged as having originated from multiple protected areas. For these, we reviewed the original sources for the time-series and selected only those for which we had certainty about their origin and information from METT.

130 chapter VI

Protected Area Management Data We constructed a database, which contains more than 3,000 individual records for sites evaluated using METT (Coad et al. 2013a). We used assessments conducted from 2011 and back (earliest 2001). For protected areas with multiple evaluations, the one nearest to 2005 was selected to provide only one METT evaluation per protected area. Data on dollar budgets and staff numbers, as well as information on threats and protected area objectives were excluded because of considerable missing data and inconsistent reporting. We therefore only considered the 30 specific questions for our analysis. For each of the 30 questions we first excluded all not directly linked to management activities or planning activities directed at biodiversity. Second we excluded all questions with more than 50 missing values. Finally we examined the distribution within each level (0-3) merging levels to the nearest adjacent, where the number of recorded observations were lower than 30. Based on this only: i) park regulation, ii) Park boundary iii) management plan, iv) research activities, v) staff numbers, and vi) current budget were kept (Table 1). Questions on ‘Protected area regulations’ and ‘current budget’ were transformed into two levels combining level 0 to1 and 2 to 3 respectively. ‘Park boundary’ was reduced to two levels combing categories 0-2 against category 3. ‘Management plan’, ‘research’, and ‘staff numbers’ were reduced to three levels combining 0 and 1) (Figure S1).

Table 1. Variables extracted from the METT scores included in the final analysis. Category Planning element Notes There are no mechanisms for controlling inappropriate land use and activities in the protected area to PA regulation Planning Mechanisms for controlling inappropriate land use and activities in the protected area exist and are being effectively implemented The boundaries of the protected area is not known to The Park boundary Process boundary of the protected area is known and is appropriately demarcated There is no management plan for the protected area to An Management plan Planning approved management plan exists and is being implemented There is no survey or research work taking place in the Research and protected area to There is a comprehensive, integrated Process monitoring programs programme of survey and research work, which is relevant to management needs There are no staff to Staff numbers are adequate for the Staff numbers Input management needs of the site There is no budget for the protected area to The available Current Budget Input budget is sufficient and meets the full management needs of the protected area

Pressures and contextual factors We included: i) Human Footprint (Sanderson et al. 2002) ii) Human Development Index (HDI) for 2005 (United Nations Development Programme 2011), and iii) Infant Mortality Rate (IMR) between 1990-2003 (median=2000) (Center for International Earth Science Information Network 2005) in our models. For protected areas with no

131 chapter VI appropriate sub-national data, the national mean was used. The median topological slope and mean elevation was calculated for all protected areas (Hijmans et al. 2005).

To test whether our measure of HDI from only one year could be used as a representation for the period over which trends were calculated (1990-2010) we examined the changes between 2005 and 2011, finding only very small changes (max change = 0.05) and a an R2 = 0.996 between the two years. Thus while HDI of different countries might have changed slightly over the years the ranking has not (Figure S2).

Statistical analysis All statistical analysis was done using R 2.14.1 (R Development Core Team 2011). For those population time-series where we could calculate reliable slopes for the period 2000-2010 we analyzed the data with a mixed effect model, using the lme4 (Bates et al. 2011) and lmerTest (Kuznetsova et al. 2013) packages. Mixed effect models can model nested data and thus able to control for effects at site or regional level when some protected areas had more than one population time-series (Zuur et al. 2009).

We first constructed a full model containing all independent parameters: i) park regulation, ii) Park boundary iii) management plan, iv) research activities, v) staff numbers, vi) current budget, vii) size of protected area, viii) species body size, ix) HDI 2005, x) mean elevation xi) median slope, and xii) species red list status as fixed effect.

Based on our full model, we subsequently constructed all possible sub-models containing all configurations of the included independent variable using a Dredge function (Barton 2013). For all models protected area within country was used as random effects. Finally, we selected the most parsimonious model amongst all sub- models based on Akaike AICc coefficients. This modeling procedure was repeated for slopes calculated using both two and three observations per time series.

To test the difference between using a threshold of either two or three observations to calculate the population slope, we used a paired t-test.

Results General characteristics of the datasets The initial dataset contained 4,366 population time-series from within protected areas and 3,023 METT evaluations. After cleaning, the final dataset contained population time-series from 72 vertebrate species across 17 countries in the tropical regions. Using a threshold of two observations per time-series the dataset contained 279 populations from 53 protected areas (Figure 1). Using a minimum of three observations to calculate the slope reduced the dataset to 184 populations from 30 protected areas Species included were generally large mammals with an mean weight of 698.6 kg (median = 161.9 kg) (Table S2).

132 chapter VI 133

birds

mammals

19 time-series

1 time-series Reptiles 0 3000 km

Figure 1. Location of the 53 protected areas included in the study. The size of the pie charts indicates the number of populations included from each protected area. Red colors are mammal time-series, blue are birds, and green are reptiles. chapter VI

Protected areas included had an mean size of 4030 ha. (median = 661 ha) and mean reported age of 43 years (median = 39 years) from WDPA date of establishment. They included two protected areas in IUCN management category Ia, 29 from category II, 10 from category IV, two from category V, two from category VI, and eight with no IUCN management category assigned. The mean year for the METT evaluation was 2006.4, S.D. = 2.9 years (Table S3).

The population time-series dataset with a threshold of two observation had a mean number of 4.05 observations (median = 3), while the dataset using a threshold of three observation had a mean number of 5.10 observations (median =5). A students t-test found no significant difference between the slopes (t= 1.01, p= 0.31, d.f. = 459) using two or three observations. We therefore concluded the remaining analyses using the dataset with two observations.

Species trends and impact of protected area management Across protected areas where METT evaluations had been conducted, animal population trends increased between 2000 and 2010; both for two (mean = 0.066, S.D.= 0.307, median = 0.0412) (Figure S3A) and three observations (mean = 0.043, S.D.= 0.190, median = 0.039) (Figure S3B).

Table 2. Model parameter estimates for the most parsimonious model. Parameter Estimate S.E. t-value p-values Intercept 0.799 0.309 2.585 0.010 * Management plan level 2 -1.089 0.405 -2.687 0.008 ** Management plan level 3 -0.875 0.334 -2.617 0.009 ** Human Development Index 2005 -0.929 0.508 -1.828 0.069 Management plan 2 : HDI 05 1.465 0.683 2.145 0.033 * Management plan 3 : HDI 05 1.119 0.565 1.980 0.049 * * significant at the 0.05 level., ** significant at 0.01 level. Management plan level 2 and 3 is the difference between no management plan and inadequate management plan or adequate management plan respectively.

In total, 17,152 model configurations were tested to find the best relationship between management species outcomes. A best fit model based on parsimony and Akaike AICc (AIC weight = 0.61) was selected from the full model containing i), HDI, ii) management plan, and iii) their interaction, with management plan and the interaction having a significant impact on population slopes (Table 2). Two other models also performed well containing either management plan alone or management plan and HDI without their interactions (accumulated AICc weight for all three models = 0.93).

These models show that, when not considering the interactions, population slopes were on average positive and significantly higher in protected areas with no management plan compared to protected area with well-implemented management plans or partly implemented management plans. For all levels of management plans (not existing to fully implemented) the mean population slope was positive (Figure 2). When the interaction between HDI and management plan was considered, population slopes were

134 chapter VI

Figure 2. The mean slopes of population for all populations divided by no management plan (No MP), inadequate management plans (inadequate MP), and adequate management plans (Adequate MP).

observed to decrease with increasing HDI in protected areas without a management plan. The opposite was observed for both poorly implemented and well implemented management plans, where population slopes increased in area with high HDI compared to areas with low HDI (Figure 3).

Figure 3. Partial effects plot from the final model showing the interaction between the level of management plan and HDI. Areas with no management plans had lower mean slopes in areas with high HDI (red). The opposite pattern was observed for populations in protected areas with inadequate (blue) and adequate management plans (green).

None of the most parsimonious models contained protected area regulation, boundaries, adequacy of staffing and budgets, or research and monitoring activities all expected to be important correlates of species outcomes. Neither where these variables found to have a significantly effect when tested alone against population slopes.

135 chapter VI

Discussion The influence of management and human development Our analyses show that animal population in protected areas increased over the period 2000-2010 where METT evaluations have been conducted. This positive result contradicts the generally negative trends in species population shown in tropical countries (Collen et al. 2009) and in tropical protected areas (Craigie et al. 2010). However, it should be noted that our results are confounded by the size of our dataset, and extending conclusions beyond the data included should be done with care. Still, the two datasets used in the analysis are the largest of their kinds, independently collected, and using quantitative measures for both input and outcome, making them ideal candidates for testing the link between the strength of management and changes in biodiversity.

Counter-intuitive to common expectations (Maiorano et al. 2007; Taylor et al. 2011; Laurance et al. 2012; Barnes et al. in prep), populations with the steepest positive slopes were found in protected areas with no management plan in countries with low HDI scores. Most of these populations were either from protected areas where populations of key species (Ethiopian wolf, Canis simensis) were decimated to the brink of extirpation by diseases in the 1990s (Marino et al. 2006) or from countries like Namibia, where community based game guard-schemes, running independently of protected area management, have helped see a resurrection of large mammals since the late 1980s (Child & Barnes 2010; Naidoo et al. 2011a). Previous studies have found evidence supporting a relation between successful community engagement and improved management in community reserves compared to traditional protected areas, fueled local economic benefits (Redford & Adams 2009; Naidoo et al. 2011b). The lack of information on governance regimes in our data does not allow us to conclude whether the same processes are working in the protected areas included in our analysis. However, our results suggest that governance structures and the role of local people in protected area effectiveness merit further attention, particularly as the evidence for community-based management ability to reduce deforestation is less conclusive (Bowler et al. 2010; Geldmann et al. 2013).

In protected areas with management plans, populations were observed to increase more in countries with higher HDI. This pattern has also been observed using an even larger sample of the LPD, however not including the direct management actions measured by the METT (Barnes et al. in prep). Even though differences were only small, populations in protected areas with only inadequate management plans increased more with increasing HDI than protected areas with fully implemented management plans. One explanation could be that parks with only partly implemented or inadequate management plans are responding to population declines in the past by implementing management procedures which are only starting to take effect, thus leading to populations increases. Contrary, adequate and well implemented management plans could be found in reserves with a long history of management and thus more likely

136 chapter VI stable populations close to carrying capacity (e.g. Carrillo et al. 2000; Metzger et al. 2010).

We found no significant effect of several management parameters often suggested to be important for conservation. Thus, in our study, neither adequacy of budgets nor staffing was associated with significantly improved trends in animal populations. This does not indicate that these factors are not important for protected area management and both have previously been associated with conservation success (Jachmann 2008; Geldmann et al. 2013). Our lack of clear results could suggest a more complicated relationship between budgets and resources for staffing on the one side and the changes in animal populations on the other. A new global study has tried to map global conservation expenditure at country level (Waldron et al. 2013), but our results suggest that the devil is in the details and that better data on conservation cost and staffing is needed at the protected area level understanding how resources are allocated and perceived within parks.

Utility of PAME assessments Given the attention that conservation donor organizations are investing in conducting PAME as part of their project evaluations, the lack of strong empirical evidence for improved management benefiting animal populations is worrying. This is particularly the case because similar results have been found between different PAME data and habitat outcomes (Nolte & Agrawal 2013; Nolte et al. 2013).

We have taken a conservative approach to METT, excluding a number of elements and questions, either because of quality, missing data, or irrelevance to changes in populations. Likewise we have not used accumulated scores across a range of questions covering very different elements of management or included questions that are not relevant for biodiversity outcomes. Hence, we have not conducted a comprehensive analysis of the entire suite of information collected by the METT, nor did we have a sufficient sample size to test all METT elements. Thus while our results suggest only weak links between METT and changes in population trends we do not suggest METT is not a valuable tool for management evaluations.

One issue which has been raised concerning METT pertains to the individuals completing the assessments who may be motivated to demonstrate improving scores for projects they have responsibility over and thus perhaps likely to report incorrect and inflated scores. Another challenge with correlating METT and population time-series is the temporal lack between interventions and responses in populations. Improved management plans may in time led to improved conditions for biodiversity, but whether this is expected right after its formulation and implementation is less certain.

Population trends as conservation outcome Often considered one of the gold standards for evaluating protected area performance, animal populations have shortcomings when used in large scale analysis (Mace et al.

137 chapter VI

2010). First, depending on the objectives of the protected area some species could be expected to increase while other decrease (Tambling & Toit 2005). In other cases mismatches between conservation objectives and the species being monitored can impede linking observed changes to management (Fellers & Drost 1993). Second, the delayed responses in vertebrate populations can limit their use in tracking changes of recent management interventions (Woinarski et al. 2010). Third, populations do not only respond to management, but are confounded by numerous other biotic and abiotic factors often of critical importance for the changes observed in populations (Scholte 2011). Fourth, positive slopes are not a direct measure of conservation success but only of populations growing. Many of the most successful parks could be expected to have neither populations increasing or decreasing but rather populations in equilibrium with no expected long-term directional change in population numbers (Packer et al. 2013). Thus while populations are a powerful indicator of changes in a conservation relevant biodiversity outcomes, understanding their interactions with human pressures and responses are complicated, making their application as measures of protected areas success challenging.

Limitations of analysis Though our study is global in scope it is still restricted by the availability of data to vertebrate species in tropical countries, often with considerable donor support. Patterns of change may therefore be very different for other less studies taxa, in protected areas in the developed countries, or not on the receiving end of outside investments. We a priori excluded all METT evaluations and population time-series from Europe, North America, and Australia because we evaluated the anthropogenic impact and governance history made them incompatible with the tropics. However even within the included data we also found a geographic bias towards Southern and Eastern Africa. METT evaluations in the database have relatively good coverage across the regions included in the analysis, but the same is not the case for population times-series which are skewed towards Africa ultimately skewing the final dataset. We ran our most parsimonious model including HDI, management plan and their interactions on the subset of location from Africa finding no significant effects of any parameters (Table S4). This could be either because of the further reduction in data or because these country level effects only occur over a larger spectrum of HDI. However given the large number of confounding factors between continents which are not included in our model this warrants further investigations.

We have used GLMs to estimate the slope of population tim-series between 2000 and 2010 as a measure of the average change (Mountford 1982). This approach is conservative, and with sensible threshold for populations increasing or decreasing too rapidly, eliminates most inappropriate population time-series (Braak et al. 1994). Other approaches exists, like Generalized Additive Models (GAM) which produce smoother and more accurate representations of the observed data thus accounting for more of the natural variation occurring in species populations (Fewster et al. 2000; Atkinson et al.

138 chapter VI

2006). However, models based on an overall linear trend are being widely applied for species population counts and GAMs are in many cases found to be inappropriate for population estimates (Soldaat et al. 2007). Based on the conservative nature of GLMs and the relative low number of mean observations in our time-series we have therefore persisted in using GLMs for estimating the mean slopes of animal populations.

Given the relatively low number of observation compared to the number of independent variables and the complex nature of mixed effects models, selecting the right model is no straight forward exercise. We have used a data-dredging which allows to run and compare all possible sub-models thus avoiding a stepwise reduction approach based on a complete model much larger than what is appropriate given the data (Coad et al. 2013c). However dredging is often considered a form of data mining and care should be taken when evaluating the best fit models based on this procedure (Burnham & Anderson 2002). In our case HDI and management plan were amongst the most influential parameters not only in the most parsimonious model but included in most of the top models.

Moving forward Our lack of strong correlations illuminates the complexity in estimating the effect of management using large scale data. Further, it illustrates the need for more data, when the number of confounding factors at site and population levels is high. However, small scale studies have similarly failed to present clear and prescriptive results of how management affects changes in biodiversity (Geldmann et al. 2013). While local case- studies have often excelled in complex explanation for the observed patterns (e.g. Stoner et al. 2007; Wegge et al. 2009), they have often used non-comparable experimental designs thus failing to make their results applicable across studies, regions and contexts. The concern about a lack of methodological rigor and consistency across studies of protected area effectiveness has already been raised (Ferraro 2009; Geldmann et al. 2013), and standardized reporting like the METT and population changes could be a way to ensure better standardized data. Particularly as many of the world most influential conservation organizations like UNDP and the GEF; investing billions of dollars in protected areas are already using METT. Similarly the LPD and other initiatives to collate large scale data on biodiversity continue to grow. Thus even with the challenges in combining these two large dataset we do see a need to approach the question of protected area effectiveness with correlative analyses using large collated datasets. At present most large scale analysis of protected area effectiveness have either used remote sensed data (Joppa & Pfaff 2011) or been based on qualitative information collected through questionnaires (Bruner et al. 2001; Laurance et al. 2012). The two datasets we present here are amongst the best candidates for investigating the effect of management. However this will still require investments in increasing the available data. Given the small overlap between the two datasets from our MET-database and the LPD targeting particular areas with either good population time-series or comprehensive management information is a cost effective approach to increase the available data.

139 chapter VI

Acknowledgement We thank F. Leverington, C. Nolte, M. Zimsky, J. Mee, N. Vansteelant, and many others who collected, compiled, and provided METT data. J. Loh and numerous staff and interns at the Zoological Society of London are thanked for their efforts in collating the Living Planet Database. We also thank the Danish National Research Foundation, the IUCN SSC/WCPA Joint Task-Force on Biodiversity and Protected Areas, UNEP- WCMC, the Zoological Society of London, WWF and for financial and institutional support. S. Grimstrup, P.S. Jørgensen, and E. Mousing helped with the statistical analysis.

References Atkinson, P. W., G. E. Austin, M. M. Rehfisch, H. Baker, P. Cranswick, M. Kershaw, J. Robinson, R. H. W. Langston, D. A. Stroud, C. V. Turnhout, and I. M. D. Maclean. (2006). Identifying declines in waterbirds: The effects of missing data, population variability and count period on the interpretation of long-term survey data. Biological Conservation 130:549-559. Barnes, M., I. D. Craigie, L. Harrison, J. Geldmann, T. Brooks, N. D. Burgess, B. Collen, M. Hockings, S. Whitmee, and S. Woodley. (in prep). Key correlates of population trends for birds and mammals in terrestrial protected areas. 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. Barton, K. (2013). R package MuMIn: Multi-model inference. Bates, D., M. Maechler, and B. Bolker. (2011). lme4: Linear mixed-effects models using S4 classes. Bertzky, B., C. Corrigan, J. Kemsey, S. Kenney, C. Ravilious, C. Besancon, and N. D. Burgess. (2012). Protected Planet report: Tracking progress towards global targets for protected areas. IUCN and UNEP-WCMC, Gland, Switzerland and Cambridge, UK. Bowler, D., L. Buyung-Ali, J. R. Healey, J. P. G. Jones, T. Knight, and A. S. Pullin. (2010). The Evidence Base for Community Forest Management as a Mechanism for Supplying Global Environmental Benefits and Improving Local Welfare. Environmental Evidence. Collaboration for Environmental Evidence. Braak, C. J. F. t., A. van Strien, R. Meijer, and T. Verstrael. (1994). Analysis of monitoring data with many missing values: which method? Pages 663–673 in W. Hagemeijer, and T. Verstrael, editors. Proceedings of the 12th International Conference of IBCC and EOAC. Statistics Netherlands, Voorburg/Heerlen & SOVON, Beek-Ubbergen, Noordwijkerhout, The Netherlands. Bruner, A. G., R. E. Gullison, R. E. Rice, and G. A. B. da Fonseca. (2001). Effectiveness of Parks in Protecting Tropical Biodiversity. Science 291:125-128. Burnham, K. P., and D. R. Anderson (2002). Model selection and multimodel inference - a practical information-theoretic approach Springer-Verlag, New York, USA. Butchart, S. H. M., J. P. W. Scharlemann, M. I. Evans, S. Quader, S. Aricò, J. Arinaitwe, M. Balman, L. A. Bennun, B. Bertzky, C. Besançon, T. M. Boucher, T. M. Brooks, I. J. Burfield, N. D. Burgess, S. Chan, R. P. Clay, M. J. Crosby, N. C. Davidson, N. De Silva, C. Devenish, G. C. L. Dutson, D. F. D. z. Fernández, L. D. C. Fishpool, C. Fitzgerald, M. Foster, M. F. Heath, M. Hockings, M. Hoffmann, D. Knox, F. W. Larsen, J. F. Lamoreux, C. Loucks, I. May, J. Millett, D. Molloy, P. Morling, M. Parr,

140 chapter VI

T. H. Ricketts, N. Seddon, B. Skolnik, S. N. Stuart, A. Upgren, and S. Woodley. (2012). Protecting Important Sites for Biodiversity Contributes to Meeting Global Conservation Targets. PLoS ONE 7:e32529. Butchart, S. H. M., M. Walpole, B. Collen, A. van Strien, J. P. W. Scharlemann, R. E. A. Almond, J. E. M. Baillie, B. Bomhard, C. Brown, J. Bruno, K. E. Carpenter, G. M. Carr, J. Chanson, A. M. Chenery, J. Csirke, N. C. Davidson, F. Dentener, M. Foster, A. Galli, J. N. Galloway, P. Genovesi, R. D. Gregory, M. Hockings, V. Kapos, J. F. Lamarque, F. Leverington, J. Loh, M. A. McGeoch, L. McRae, A. Minasyan, M. H. Morcillo, T. E. E. Oldfield, D. Pauly, S. Quader, C. Revenga, J. R. Sauer, B. Skolnik, D. Spear, D. Stanwell-Smith, S. N. Stuart, A. Symes, M. Tierney, T. D. Tyrrell, J. C. Vie, and R. Watson. (2010). Global Biodiversity: Indicators of Recent Declines. Science 328:1164-1168. Carrillo, E., G. Wong, and A. D. Cuaron. (2000). Monitoring mammal populations in Costa Rican protected areas under different hunting restrictions. Conservation Biology 14:1580-1591. Center for International Earth Science Information Network. (2005). Poverty Mapping Project: Global Subnational Infant Mortality Rates. Center for International Earth Science Information Network, Palisades, NY, USA. Available from http://sedac.ciesin.columbia.edu/data/set/povmap-global-subnational-infant-mortality- rates Child, B., and G. Barnes. (2010). The conceptual evolution and practice of community-based natural resource management in southern Africa: past, present and future. Environmental Conservation 37:283-295. Coad, L., F. leverington, N. D. burgess, I. C. Cuadros, J. Geldmann, T. R. Marthews, J. Mee, C. Nolte, S. Stoll-Kleemann, N. Vansteelant, C. Zamora, M. Zimsky, and M. Hockings. (2013a). Progress towards the CBD Protected Area Management Effectiveness Targets. Parks 19:13-24. Coad, L., F. Leverington, J. Geldmann, C. Nolte, and M. Hockings. (2013b). Management Effectiveness Tracking Tool, global database, University of Oxford and University of Queensland. Coad, L., J. Schleicher, E. J. Milner-Gulland, T. R. Marthews, M. Starkey, A. Manica, A. Balmford, W. Mbombe, T. R. Diop Bineni, and K. A. Abernethy. (2013c). Social and Ecological Change over a Decade in a Village Hunting System, Central Gabon. Conservation Biology 27:270-280. Collen, B., J. Loh, S. Whitmee, L. McRae, R. Amin, and J. E. M. Baillie. (2009). Monitoring Change in Vertebrate Abundance: the Living Planet Index. Conservation Biology 23:317-327. Craigie, I. D., J. E. M. Baillie, A. Balmford, C. Carbone, B. Collen, R. E. Green, and J. M. Hutton. (2010). Large mammal population declines in Africa's protected areas. Biological Conservation 143:2221-2228. del Hoyo, J., A. Elliott, and J. Sargatal (1992). Handbook of the birds of the World. Lynx Edicions, Barcelona, Spain. Dudley, N. (2008). Guidelines for Applying Protected Area Management Categories. International Union for Conservation of Nature, Gland, Switzerland. Dudley, N., and S. Stolton. (1999). Conversion of “Paper Parks” to Effective Management – Developing a Target. IUCN, WWF, WCPA. Fachín-Terán, A., and R. C. Vogt. (2004). Estrutura populacional, tamanho e razão sexual de Podocnemis unifilis (Testudines, Podocnemididae) no rio Guaporé (RO), norte do Brasil. Phyllomedusa 3:29-42.

141 chapter VI

Fellers, G. M., and C. A. Drost. (1993). Disappearance of the cascades frog Rana cascadae at the southern end of its range, California, USA. Biological Conservation 65:177-181. Ferraro, P. J. (2009). Counterfactual thinking and impact evaluation in environmental policy. New Directions for Evaluation 2009:75-84. Fewster, R. M., S. T. Buckland, G. M. Siriwardena, S. R. Baillie, and J. D. Wilson. (2000). Analysis of population trends for farmland birds using generalised additive models. Ecology 81:1970-1984. Geldmann, J., M. Barnes, L. Coad, I. D. Craigie, M. Hockings, and N. D. Burgess. (2013). Effectiveness of terrestrial protected areas in reducing habitat loss and population declines Biological Conservation 161:230-238. Hays, G. C., A. C. Broderick, F. Glen, and B. J. Godley. (2002). Change in body mass associated with long-term fasting in a marine reptile: the case of green turtles (Chelonia mydas) at Ascension Island. Canadian Journal of Zoology-Revue Canadienne De Zoologie 80:1299-1302. Hijmans, R. J., S. E. Cameron, J. L. Parra, P. G. Jones, and A. Jarvis. (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25:1965-1978. Hockings, M. (2003). Systems for assessing the effectiveness of management in protected areas. BioScience 53:823-832. Hockings, M., S. Stolton, F. Leverington, N. Dudley, J. Courrau, P. Valentine, and S. Editor. (2006). Evaluating Effectiveness: A framework for assessing management effectiveness of protected areas, Gland, Switzerland. International Union for Conservation of Nature. (2012). The IUCN Red List of Threatened Species. Version 2012.2. International Union for Conservation of Nature Available from http://www.iucnredlist.org Jachmann, H. (2008). Monitoring law-enforcement performance in nine protected areas in Ghana. Biological Conservation 141:89-99. Jones, K. E., J. Bielby, M. Cardillo, S. A. Fritz, J. O'Dell, C. D. L. Orme, K. Safi, W. Sechrest, E. H. Boakes, C. Carbone, C. Connolly, M. J. Cutts, J. K. Foster, R. Grenyer, M. Habib, C. A. Plaster, S. A. Price, E. A. Rigby, J. Rist, A. Teacher, O. R. P. Bininda-Emonds, J. L. Gittleman, G. M. Mace, A. Purvis, and W. K. Michener. (2009). PanTHERIA: a species-level database of life history, ecology, and geography of extant and recently extinct mammals. Ecology 90:2648-2648. Joppa, L. N., and A. Pfaff. (2009). High and Far: Biases in the Location of Protected Areas. PLoS ONE 4:e8273. Joppa, L. N., and A. Pfaff. (2010). Reassessing the forest impacts of protection The challenge of nonrandom location and a corrective method. Pages 135-149. Ecological Economics Reviews. Joppa, L. N., and A. Pfaff. (2011). Global protected area impacts. Proceedings of the Royal Society B-Biological Sciences 278:1633-1638. Kapos, V., A. Balmford, R. Aveling, P. Bubb, P. Carey, A. Entwistle, J. Hopkins, T. Mulliken, R. Safford, A. Stattersfield, M. Walpole, and A. Manica. (2009). Outcomes, not implementation, predict conservation success. Oryx 43:336-342. Kear, J. (2005). Ducks, Geese and Swans. Oxford University Press, NY, USA. Kuznetsova, A., P. B. Brockhoff, and R. H. B. Christensen. (2013). lmerTest: Tests for random and fixed effects for linear mixed effect models. Copenhagen, Denmark. Laurance, W. F., D. Carolina Useche, J. Rendeiro, M. Kalka, C. J. A. Bradshaw, S. P. Sloan, S. G. Laurance, M. Campbell, K. Abernethy, P. Alvarez, V. Arroyo-Rodriguez, P. Ashton, J. Benitez-Malvido, A. Blom, K. S. Bobo, C. H. Cannon, M. Cao, R. Carroll, C.

142 chapter VI

Chapman, R. Coates, M. Cords, F. Danielsen, B. De Dijn, E. Dinerstein, M. A. Donnelly, D. Edwards, F. Edwards, N. Farwig, P. Fashing, P.-M. Forget, M. Foster, G. Gale, D. Harris, R. Harrison, J. Hart, S. Karpanty, W. John Kress, J. Krishnaswamy, W. Logsdon, J. Lovett, W. Magnusson, F. Maisels, A. R. Marshall, D. McClearn, D. Mudappa, M. R. Nielsen, R. Pearson, N. Pitman, J. van der Ploeg, A. Plumptre, J. Poulsen, M. Quesada, H. Rainey, D. Robinson, C. Roetgers, F. Rovero, F. Scatena, C. Schulze, D. Sheil, T. Struhsaker, J. Terborgh, D. Thomas, R. Timm, J. Nicolas Urbina- Cardona, K. Vasudevan, S. Joseph Wright, J. Carlos Arias-G, L. Arroyo, M. Ashton, P. Auzel, D. Babaasa, F. Babweteera, P. Baker, O. Banki, M. Bass, I. Bila-Isia, S. Blake, W. Brockelman, N. Brokaw, C. A. Bruhl, S. Bunyavejchewin, J.-T. Chao, J. Chave, R. Chellam, C. J. Clark, J. Clavijo, R. Congdon, R. Corlett, H. S. Dattaraja, C. Dave, G. Davies, B. de Mello Beisiegel, R. d. Nazare Paes da Silva, A. Di Fiore, A. Diesmos, R. Dirzo, D. Doran-Sheehy, M. Eaton, L. Emmons, A. Estrada, C. Ewango, L. Fedigan, F. Feer, B. Fruth, J. Giacalone Willis, U. Goodale, S. Goodman, J. C. Guix, P. Guthiga, W. Haber, K. Hamer, I. Herbinger, J. Hill, Z. Huang, I. Fang Sun, K. Ickes, A. Itoh, N. Ivanauskas, B. Jackes, J. Janovec, D. Janzen, M. Jiangming, C. Jin, T. Jones, H. Justiniano, E. Kalko, A. Kasangaki, T. Killeen, H.-b. King, E. Klop, C. Knott, I. Kone, E. Kudavidanage, J. Lahoz da Silva Ribeiro, J. Lattke, R. Laval, R. Lawton, M. Leal, M. Leighton, M. Lentino, C. Leonel, J. Lindsell, L. Ling-Ling, K. Eduard Linsenmair, E. Losos, A. Lugo, J. Lwanga, A. L. Mack, M. Martins, W. Scott McGraw, R. McNab, L. Montag, J. Myers Thompson, J. Nabe-Nielsen, M. Nakagawa, S. Nepal, M. Norconk, V. Novotny, S. O'Donnell, M. Opiang, P. Ouboter, K. Parker, N. Parthasarathy, K. Pisciotta, D. Prawiradilaga, C. Pringle, S. Rajathurai, U. Reichard, G. Reinartz, K. Renton, G. Reynolds, V. Reynolds, E. Riley, M.-O. Rodel, J. Rothman, P. Round, S. Sakai, T. Sanaiotti, T. Savini, G. Schaab, J. Seidensticker, A. Siaka, M. R. Silman, T. B. Smith, S. S. d. Almeida, N. Sodhi, C. Stanford, K. Stewart, E. Stokes, K. E. Stoner, R. Sukumar, M. Surbeck, M. Tobler, T. Tscharntke, A. Turkalo, G. Umapathy, M. van Weerd, J. Vega Rivera, M. Venkataraman, L. Venn, C. Verea, C. Volkmer de Castilho, M. Waltert, B. Wang, D. Watts, W. Weber, P. West, D. Whitacre, K. Whitney, D. Wilkie, S. Williams, D. D. Wright, P. Wright, L. Xiankai, P. Yonzon, and F. Zamzani. (2012). Averting biodiversity collapse in tropical forest protected areas. Nature 489:290-294. Leroux, S. J., M. A. Krawchuk, F. Schmiegelow, S. G. Cumming, K. Lisgo, L. G. Anderson, and M. Petkova. (2010). Global protected areas and IUCN designations: Do the categories match the conditions? Biological Conservation 143:609-616. Leverington, F., K. L. Costa, J. Courrau, H. Pavese, C. Nolte, M. Marr, L. Coad, N. Burgess, B. Bomhard, and M. Hockings. (2010a). Management effectiveness evaluation in protected areas - a global study. Second edition 2010. The University of Queensland, Brisbane Australia. Leverington, F., K. L. Costa, H. Pavese, A. Lisle, and M. Hockings. (2010b). A Global Analysis of Protected Area Management Effectiveness. Environmental Management 46:685-698. Loh, J., R. E. Green, T. Ricketts, J. Lamoreux, M. Jenkins, V. Kapos, and J. Randers. (2005). The Living Planet Index: using species population time series to track trends in biodiversity. Philosophical Transactions of the Royal Society B: Biological Sciences 360:289-295. Mace, G. M., B. Collen, R. A. Fuller, and E. H. Boakes. (2010). Population and geographic range dynamics: implications for conservation planning. Philosophical Transactions of the Royal Society B-Biological Sciences 365:3743-3751. Maiorano, L., A. Falcucci, E. O. Garton, and L. Boitani. (2007). Contribution of the Natura 2000 network to biodiversity conservation in Italy. Conservation Biology 21:1433- 1444.

143 chapter VI

Marino, J., C. Sillero-Zubiri, and D. W. Macdonald. (2006). Trends, dynamics and resilience of an Ethiopian wolf population. Animal Conservation 9:49-58. McKinney, M. L. (2002). Effects of national conservation spending and amount of protected area on species threat rates. Conservation Biology 16:539-543. Metzger, K., A. Sinclair, R. Hilborn, J. Hopcraft, and S. Mduma. (2010). Evaluating the protection of wildlife in parks: the case of African buffalo in Serengeti. Biodiversity and Conservation 19:3431-3444. Mountford, M. D. (1982). Estimation of Population Fluctuations with Application to the Common Bird Census. Journal of the Royal Statistical Society. Series C (Applied Statistics) 31:135-143. Naidoo, R., L. C. Weaver, M. De Longcamp, and P. Du Plessis. (2011a). Namibia's community- based natural resource management programme: an unrecognized payments for ecosystem services scheme. Environmental Conservation 38:445-453. Naidoo, R., L. C. Weaver, G. Stuart-Hill, and J. Tagg. (2011b). Effect of biodiversity on economic benefits from communal lands in Namibia. Journal of Applied Ecology 48:310-316. Nolte, C., and A. Agrawal. (2013). Linking Management Effectiveness Indicators to Observed Effects of Protected Areas on Fire Occurrence in the Amazon Rainforest. Conservation Biology 27:155-165. Nolte, C., A. Agrawal, and P. Barreto. (2013). Setting priorities to avoid deforestation in Amazon protected areas: are we choosing the right indicators? Environmental Research Letters 8:015039. Packer, C., A. Loveridge, S. Canney, T. Caro, S. T. Garnett, M. Pfeifer, K. K. Zander, A. Swanson, D. MacNulty, G. Balme, H. Bauer, C. M. Begg, K. S. Begg, S. Bhalla, C. Bissett, T. Bodasing, H. Brink, A. Burger, A. C. Burton, B. Clegg, S. Dell, A. Delsink, T. Dickerson, S. M. Dloniak, D. Druce, L. Frank, P. Funston, N. Gichohi, R. Groom, C. Hanekom, B. Heath, L. Hunter, H. H. DeIongh, C. J. Joubert, S. M. Kasiki, B. Kissui, W. Knocker, B. Leathem, P. A. Lindsey, S. D. Maclennan, J. W. McNutt, S. M. Miller, S. Naylor, P. Nel, C. Ng'weno, K. Nicholls, J. O. Ogutu, E. Okot-Omoya, B. D. Patterson, A. Plumptre, J. Salerno, K. Skinner, R. Slotow, E. A. Sogbohossou, K. J. Stratford, C. Winterbach, H. Winterbach, and S. Polasky. (2013). Conserving large carnivores: dollars and fence. Ecology Letters 16:635-641. Payne, R. B. (2009). CRC Handbook of Avian Body Masses. Second Edition. The Wilson Journal of Ornithology 121:661-662. R Development Core Team. (2011). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Redford, K. H., and W. M. Adams. (2009). Payment for Ecosystem Services and the Challenge of Saving Nature. Conservation Biology 23:785-787. Rodrigues, A. S. L., S. J. Andelman, M. I. Bakarr, L. Boitani, T. M. Brooks, R. M. Cowling, L. D. C. Fishpool, G. A. B. da Fonseca, K. J. Gaston, M. Hoffmann, J. S. Long, P. A. Marquet, J. D. Pilgrim, R. L. Pressey, J. Schipper, W. Sechrest, S. N. Stuart, L. G. Underhill, R. W. Waller, M. E. J. Watts, and X. Yan. (2004). Effectiveness of the global protected area network in representing species diversity. Nature 428:640-643. Sanderson, E. W., M. Jaiteh, M. A. Levy, K. H. Redford, A. V. Wannebo, and G. Woolmer. (2002). The Human Footprint and the Last of the Wild. Bioscience 52:891-904. Scharlemann, J. P. W., V. Kapos, A. Campbell, I. Lysenko, N. D. Burgess, M. C. Hansen, H. K. Gibbs, B. Dickson, and L. Miles. (2010). Securing tropical forest carbon: the contribution of protected areas to REDD. Oryx 44:352-357.

144 chapter VI

Scholte, P. (2011). Towards understanding large mammal population declines in Africa’s protected areas: A West-Central African perspective. Tropical Conservation Science 4:11. Soldaat, L., H. Visser, M. Roomen, and A. Strien. (2007). Smoothing and trend detection in waterbird monitoring data using structural time-series analysis and the Kalman filter. Journal of Ornithology 148:351-357. Stolton, S., M. Hockings, N. Dudley, K. MacKinnon, T. Whitten, and L. F. (2007). Reporting Progress in Protected Areas A Site-Level Management Effectiveness Tracking Tool: second edition. World Bank/WWF Forest Alliance, Gland, Switzerland. Stoner, C., T. Caro, S. Mduma, C. Mlingwa, G. Sabuni, and M. Borner. (2007). Assessment of effectiveness of protection strategies in Tanzania based on a decade of survey data for large herbivores. Conservation Biology 21:635-646. Tambling, C. J., and J. T. D. Toit. (2005). Modelling Wildebeest Population Dynamics: Implications of Predation and Harvesting in a Closed System. Journal of Applied Ecology 42:431-441. Taylor, M. F. J., P. S. Sattler, M. Evans, R. A. Fuller, J. E. M. Watson, and H. P. Possingham. (2011). What works for threatened species recovery? An empirical evaluation for Australia. Biodiversity and Conservation 20:767-777. Tranquilli, S., M. Abedi-Lartey, F. Amsini, L. Arranz, A. Asamoah, O. Babafemi, N. Barakabuye, G. Campbell, R. Chancellor, T. R. B. Davenport, A. Dunn, J. Dupain, C. Ellis, G. Etoga, T. Furuichi, S. Gatti, A. Ghiurghi, E. Greengrass, C. Hashimoto, J. Hart, I. Herbinger, T. C. Hicks, L. H. Holbech, B. Huijbregts, I. Imong, N. Kumpel, F. Maisels, P. Marshall, S. Nixon, E. Normand, L. Nziguyimpa, Z. Nzooh-Dogmo, D. T. Okon, A. Plumptre, A. Rundus, J. Sunderland-Groves, A. Todd, Y. Warren, R. Mundry, C. Boesch, and H. Kuehl. (2012). Lack of conservation effort rapidly increases African great ape extinction risk. Conservation Letters 5:48-55. United Nations Development Programme. (2011). Human Development Report 2011: Sustainability and Equity: A Better Future for All. UNDP, New York, USA. Waldron, A., A. O. Mooers, D. C. Miller, N. Nibbelink, D. Redding, T. S. Kuhn, J. T. Roberts, and J. L. Gittleman. (2013). Targeting global conservation funding to limit immediate biodiversity declines. Proceedings of the National Academy of Sciences:PNAS early edition 1-5 pp. Wallace, B. P., C. L. Williams, F. V. Paladino, S. J. Morreale, R. T. Lindstrom, and J. R. Spotila. (2005). Bioenergetics and diving activity of internesting leatherback turtles Dermochelys coriacea at Parque Nacional Marino las Baulas, Costa Rica. Journal of Experimental Biology 208:3873-3884. Wegge, P., M. Odden, C. P. Pokharel, and T. Storaas. (2009). Predator-prey relationships and responses of ungulates and their predators to the establishment of protected areas: A case study of tigers, leopards and their prey in Bardia National Park, Nepal. Biological Conservation 142:189-202. Wilson, D. E., and R. A. Mittermeier (2009). Handbook of the Mammals of the World. Lynx Edicions, Barcelona, Spain. Woinarski, J. C. Z., M. Armstrong, K. Brennan, A. Fisher, A. D. Griffiths, B. Hill, D. J. Milne, C. Palmer, S. Ward, M. Watson, S. Winderlich, and S. Young. (2010). Monitoring indicates rapid and severe decline of native small mammals in Kakadu National Park, northern Australia. Wildlife Research 37:116-126. World Centre for Monitoring Conservation. (2012). World Database on Protected areas. UNEP World Conservation Monitoring Centre Available from http://protectedplanet.net/ Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. Smith (2009). Mixed Effects Models and Extentions in Ecology with R. Springer, New York, USA.

145

146

Appendices I-III

147

148 appendix

Appendix I (chapter I)

Evidence of protected area effectiveness – supporting information Table S1. List of search terms used in the systematic review to evaluate literature. Biodiversity Protected Area Management Output Biodiversity “Indigenous people” Monitor* Effect* Population* “Community conser- Management Effectiveness Species ved area$” Governance Outcome Threaten* Habitat$ Conserv* Success "Threatened species" "National park$" "Red list*" "Protected area$" Trend$ Reserve* Endanger* Increase* Decline* The table is printed in Geldmann et al. 2012 systematic review and has been subject to peer-review through the Collaboration for Environmental Evidence process.

Table S2. Sources used in the search Scientific sources Specialist sources Websites ISI Web of World Environment Library http://www.nzdl.org/fast- Google/Yahoo/chrome Knowledge cgi-bin/library?a=p&p=about&c=envl BIOSIS citation Forestscience.info http://www.forestscience.info/ Google scholar index Tropical forest conservation and development database Zoological records IUCN website http://forestry.lib.umn.edu/bib/trps.html ASFA Conservationevidence.org WWF website UN-REDD Web Platform SCRIS FAO website http://unfccc.int/methods_science/redd/items/4531.php Science Direct FAO online catalogue: http://www4.fao.org/faobib/ UNEP website Directory of Open World Environment Library http://www.nzdl.org/fast- CIFOR website Access Journals cgi-bin/library?a=p&p=about&c=envl Index to Theses Woods Hole research Forestscience.info http://www.forestscience.info/ Online centre Tropical forest conservation and development database Conservation CAB abstracts http://forestry.lib.umn.edu/bib/trps.html International COPAC Conservationevidence.org WCS University of World Bank Oxford Libraries SCOPUS UNDP ProQuest Dissertations and WCPA Theses The table is printed in Geldmann et al. 2012 systematic review and has been subject to peer-review through the Collaboration for Environmental Evidence process.

149 appendix

Table S3. Examples for each of the six categories used to group management actions and attributes for the population studies. Category Source Description Legislative and National regulations were tightened to protect Struhsaker et al., 2005 governmental regulations endangered species. Unspecified management Protected areas managed after management intervention (management Pettorelli et al., 2010 plans (MP) did better, but no details of the MP plans) was disclosed Specified management Anti-poaching efforts inside park and intervention (guards and Caro, 1999 increased guard presence anti-poaching) Specified management Gough and Kerley, Reserve boundaries was fenced to protect intervention (fencing) 2006 elephant populations Specified management Artificial nest sites were supplemented to Catrey et al., 2009 intervention (species) facilitate increased breading success Specified management Regulation of vegetation inside reserved with Schlicht et al. , 2009 intervention (Protected area) initiation of fire These examples corresponds to the columns in figure 2B.

150

Table S4. Summary of the 73 studies evaluating PA effectiveness for habitat extent. Intervention describes whether the PA did better (+), worse (-) or whether no difference could be detected (0). PA is the rate of change in the protected area while counterf. The rate in the scenario to which the PA is compared. Difference is the calculated difference between inside and counterfactual. Methods describes which method was used. Inter- Source Country Protected area Counterfactual PA Counterf Change measure Difference Method ventions Abbot and Homewood, 1999 Malawi Lake Malawi NP PA compared to buffer 0 -0.06 - Total - In-Out Alados et al. 2004 Spain Cabo de Gata-Nijar PA compared to buffer + - - - - Regression Alo and Pontius, 2008 Ghana Forest reserves PA compared to buffer - -0.014 -0.005 Total 0.36 In-Out PA compared to similar habitats + 0.111 0 Difference - Matching Andam et al. 2008 Costa Rica 150 protected areas outside Armenteras et al. 2006 Columbia Indigenous reserves Rerserve compared to buffer + 1.5 times Difference 1.5 Regression Armenteras et al. 2006 Columbia Guyana NP PA compared to buffer + -0.00071 -0.0028 Annual 3.94 Regression Armenteras et al. 2006 Columbia Guyana NP comparred to indigenous reserves + 5.8 times Difference 5.8 Regression Arroyo-Mora et al. 2005 Costa Rica Chorotega region PA compared to adjacent landscape + 0.6363 0.2934 Total 0.44 In-Out Bleher et al., 2006 Kenya Kakamega Compared to forest reserve + -3.5 -32.3 Trees harvest pr. ha. 9.23 Ground PA compared to community managed - -0.00043 -0.00024 Annual 0.56 In-Out + Reg Bray et al. 2008 Mexico 11 PAs area PA compared to community managed - -0.00356 -0.00243 Annual 0.68 In-Out + Reg Bray et al. 2008 Guatemala 11 PAs area appendix Brower et al. 2002 Mexico 4 reserves Region of the reserves - -0.02095 -0.01815 Annual 0.87 In-Out 151 Bruner et al. 2001 Global 93 protected areas Protected not protected + - - - - Interview Chatelain et al., 2010 Cote d'Ivoiry Tai NP PA compared to buffer + -0.0028 -0.0287 Annual 10.25 In-Out Chowdhury 2006 Mexico Calakmul BR PA compared to buffer + -0.1303 -0.6198 Percent convereted 4.76 Regression Probability of + -0.0026 -0.0043 1.39 Regression Cropper et al. 2001 Thailand Multible Wildlife sanctuaries compared to buffer clearing Probability of 0 -0.0031 -0.0043 1.39 Regression Cropper et al. 2001 Thailand Multible PA compared to buffer clearing Curran et al. 2004 Indonesia Gunung Palung NP PA compared to buffer + -0.56 -0.7 Total 1.25 In-Out Cushman and Wallin 2000 Russia Sikhote-alinskiy BR PA compared to buffer + -0.002 -0.007 Annual 3.5 In-Out DeFries et al. 2005 Global 198 protected areas PA compared to buffer + -0.0332 -0.0865 Total 2.61 In-Out PA compared to community managed - -0.00418 0.000003 Annual 0.001 Regression Ellis and Porter-Bolland 2008 Mexico Calakmul BR area Forrest et al. 2008 Bolivia Madidi NP, Tierras Origen Tacana PA compared to adjacent landscape + increas decrease Total - In-Out Gaveau et al. 2007 Indonesia Bukit Barisan Selatan NP PA compared to Wildlife Sanctuary + -0.005 -0.0256 Annual 5.12 Regression Gaveau et al., 2009 Indonesia Multible PA compared to buffer + -0.28 -0.45 Total 1.61 In-Out Gaveau et al., 2009 Indonesia Multible PA compared to region + -0.28 -0.58 Total 2.07 In-Out Hayes et al. 2002 Guatemala 5 NPs 4 BRs PA compared to buffer + -0.0016 -0.0075 Annual 4.69 Regression PA compared to similar habitats 0 -0.4055 -0.4051 Total 1.00 In-Out Ingram and Dawson, 2005 Madagascar All protected areas outside Joppa and Pfaff 2011 Global Global tropical forested PA's PA compared to matched outside + 0.07667 Effect of PA - Matching

Table S4. Continued Inter- Source Country Protected area Counterfactual PA Counterf Change measure Difference Method ventions Kinnaird et al. 2003 Indonesia Bukit Barisan Selatan NP PA compared to buffer + -0.02 Forest goneAnnual 4 Regression Protected compared to none-protected + - - - - Ground Kiragu Mwangi et al., 2010 Kenya 36 IBAs IBAs Linkie et al., 2004 Indonesia Kerinci Seblat NP PA compared to buffer + -0.0028 -0.0096 Annual 3.43 Regression Establishment of PA and compared to Ratio between inside - 1.15 0.29 0.25 In-Out Liu et al. 2001 China Wolong buffer and outside larger + decrease Total - In-Out Luque 2000 USA New Jersey Pinelands PA compared to buffer decrese PA compared to similar habitats - -0.0158 -0.0104 Annual 0 In-Out Mapaure and Campbell, 2002 Zimbabwe Sengwa outside Mas 2005 Mexico Calakmul BR PA compared to matched outside + -0.003 -0.006 Annual 2 Regression Mas 2005 Mexico Calakmul BR PA compared to buffer + -0.003 -0.013 Annual 4.33 Regression Mendoza and Dirzo 1999 Mexico Monte Azules BR PA compared to adjacent landscape + -0.0014 -0.0279 Annual 19.93 Regression Amboro NP, Noel Kempff, Mercado larger NP, BR, the Rios Blanco and Negro + decrease Total - Regression decrese Mertens et al. 2004 Bolivia WR PA compared to adjacent landscape Cuyabeno Wildlife Production + -0.0189 -0.2042 Total 10.80 In-Out Messina et al. 2006 Ecuador Reserve PA compared to buffer + -0.18 -0.82 Total 4.56 In-Out Mulley and Unruh, 2004 Uganda Kibale NP PA compared to buffer appendix PA compared to similar habitats + -0.137 -0.178 Total 1.30 In-Out 152 Nagendra et al. 2008 Nepal Chitwan NP and Parsa outside Nagendra et al. 2008 Nepal Chitwan NP and Parsa PA compared to buffer + -0.137 -0.244 Total 1.78 In-Out Difference in fire + 0.0115 - Matching Nelson and Chomitz, 2009 Africa IUCN I-IV PA compared to matched outside risk Difference in fire + 0.0185 - Matching Nelson and Chomitz, 2009 Asia IUCN I-IV PA compared to matched outside risk Difference in fire + 0.035 - Matching Nelson and Chomitz, 2009 Latin America IUCN I-IV PA compared to matched outside risk Difference in fire + 0.03 - Matching Nelson and Chomitz, 2009 Africa IUCN V-VI PA compared to matched outside risk Difference in fire + 0.051 - Matching Nelson and Chomitz, 2009 Asia IUCN V-VI PA compared to matched outside risk Difference in fire + 0.056 - Matching Nelson and Chomitz, 2009 Latin America IUCN V-VI PA compared to matched outside risk Same with protection as Probability of 0 - Regression Nelson et al. 2001 Panama Darién NP PA compared to matched outside without clearing Indigenous reserves compared to Lower with protection Probability of + - Regression Nelson et al. 2001 Panama Darién NP matched outside than without clearing Nepstad et al. 2006 Brazil 10 Extractive reserves Reserves compared to buffer + -0.0015 -0.0027 Annual 1.8 In-Out Nepstad et al. 2006 Brazil 121 indigenous reserves Indigenous reserves compared to buffer + -0.0018 -0.0146 Annual 8.11 In-Out Nepstad et al. 2006 Brazil 18 National forest National forests compared to buffer + -0.0008 -0.0079 Annual 9.88 In-Out Nepstad et al. 2006 Brazil 33 PAs PA compared to buffer + -0.0003 -0.0068 Annual 22.67 In-Out San Lorenzo, Soberani´a, Chagres, + - - - Interview Oestreicher et al. 2009 Panama Altos de Campana PA compared to buffer -

Table S4. Continued Inter- Source Country Protected area Counterfactual PA Counterf Change measure Difference Method ventions Oliveira et al. 2007 Peru all in the amazon region PA compared to buffer + -0.0115 -0.0476 Total 4.14 In-Out Probability of GCA compared to entire country - 1.53 clearing compared to 0.65 Regression Pelkey et al., 2000 Tanzania All GCA outside protection outside Probability of FR compared to entire country outside 0 0.91 clearing compared to 1.10 Regression Pelkey et al., 2000 Tanzania All FR protection outside Probability of PA compared to entire country outside + 0.62 clearing compared to 1.62 Regression Pelkey et al., 2000 Tanzania All NP protection outside Sader et al. 2001 Guatemala MBR PA compared to buffer + -0.0013 -0.0159 Annual 12.23 In-Out Sanchez-Azofeifa et al. 2002 Costa Rica Corcovado NP PA compared to buffer + 0 -0.0113 Annual - In-Out Sanchez-Azofeifa et al. 2003 Costa Rica 20 NPs PA compared to buffer + -0.0054 -0.0083 Annual 1.54 In-Out Sanchez-Azofeifa et al. 2003 Costa Rica 4 Biosphere reserves BR compared to buffer + -0.0029 -0.0083 Annual 2.86 In-Out Scharlemann et al. 2010 Global 5,787 tropical forested PA's 0 + - - - - In-Out Papua New PA compared to similar habitats + -0.089 Total 2.70 In-Out Shearman and Bryan 2011 Guinea 34 PAs outside -0.24 Smith 2003 Nicaragua Bosawas PA compared to buffer + 0 - Total - In-Out appendix Songer et al. 2009 Burma Chatthin PA compared to buffer + -0.0045 -0.0186 Annual 4.13 In-Out 153 Southworth et al. 2004 Honduras Celaque NP PA compared to buffer + -0.0387 -0.2512 Total 6.49 In-Out PA compared to similar habitats + -0.0001 Annual 8 In-Out Tabor et al., 2010 Kenya 75 PAs outside -0.008 Kenya PA compared to similar habitats + -0.002 Annual 8 In-Out Tabor et al., 2010 Tanzania 75 PAs outside -0.008 Kenya PA compared to similar habitats + -0.003 Annual 8 In-Out Tabor et al., 2010 Tanzania 30 KBAs outside -0.008 PA compared to similar habitats + -0.001 Annual 8 In-Out Tabor et al., 2010 Tanzania 2 AZEs outside -0.008 Tole 2002 Jamaca Hellshire Hills PA compared to buffer + -0.01 -0.15 Annual 15 In-Out

Table S5. Summary of the 73 studies evaluating PA effectiveness for habitat extent. Extension of table S4 including specific positive and negative drivers reported, methods used for data collection, habitat type and where reported IUCN categories for the PAs. IUC Source PA (short) Counterfactual Positive Drivers Negative Drivers Data type habitat N Abbot and Homewood, 1999 Lake Malawi NP II PA compared to buffer Fuel wood collection Aerial Forest Increased slope, elevation, reduced Alados et al. 2004 Forest Cabo de Gata-Nijar V PA compared to buffer soil quality human settlement Aerial Logging outside the reserve and Alo and Pontius, 2008 Forest Forest reserves - PA compared to buffer agricultural converstion outside. Remote sensing PA compared to similar habitats Andam et al. 2008 Forest 150 protected areas - outside Isolation, elevation, increased slope Human populations density Remote sensing Human population density, economic Armenteras et al. 2006 Forest Indigenous reserves - Rerserve compared to buffer Isolation, Size conditions, cattle, rivers and roads Remote sensing Human population density, economic Armenteras et al. 2006 conditions, cattle grazing, rivers and Forest Guyana NP - PA compared to buffer Isolation, Size roads Remote sensing Human population density, economic Armenteras et al. 2006 Forest Guyana NP - comparred to indigenous reserves Isolation, Size conditions, cattle, rivers and roads Remote sensing Governmental management, reduced Remote and Aerial Arroyo-Mora et al. 2005 Forest Chorotega region - PA compared to adjacent landscape cattle prices Cattle grazing, logging sensing Management for wildlife. National Bleher et al., 2006 Forest Kakamega -, II Compared to forest reserve reserve > forest reserve Logging Ground PA compared to community managed Human population density and Bray et al. 2008 Forest 11 PAs - area Remoteness distance to previous forest area remote sensing PA compared to community managed Human population density and Bray et al. 2008 Forest appendix 11 PAs - area Remoteness distance to previous forest area remote sensing

154 Logging, agricultural enchroachment Brower et al. 2002 Forest 4 reserves - Region of the reserves * before and after establishment Remote sensing Number og guards, level of deterrent, Bruner et al. 2001 Forest 93 protected areas - Protected not protected fencing and compensation programs Ground Human population density and Remote and Aerial Chatelain et al., 2010 Forest Tai NP II PA compared to buffer enchroachment sensing Management plan, community distance to roads, settlements and Chowdhury 2006 Multiple Calakmul BR VI PA compared to buffer involvement, Elevation previously forested areas Remote sensing Wildlife sanctuaries compared to Cropper et al. 2001 Forest Multible - buffer Human population density, roads Remote sensing Cropper et al. 2001 Multible - PA compared to buffer Human population density, roads Remote sensing Forest Curran et al. 2004 Gunung Palung NP II PA compared to buffer Logging by timer concessions Remote sensing Forest Cushman and Wallin 2000 Sikhote-alinskiy BR Ia PA compared to buffer Fires and human infrastructure Remote sensing Forest I and DeFries et al. 2005 Forest 198 protected areas II PA compared to buffer Encroachment Remote sensing Distance to roads, settlements *1988- Ellis and Porter-Bolland 2008 PA compared to community managed community managed > protected area, 2000/2000-05 (outside = community Forest Calakmul BR VI area external funding (GEF), Elevation managed) Remote sensing Madidi NP, Tierras II, -, Elevation, Natural ressource Forrest et al. 2008 Forest Origen Tacana - PA compared to adjacent landscape protection laws Human settlements, roads Remote sensing Bukit Barisan Gaveau et al. 2007 Forest Selatan NP II PA compared to Wildlife Sanctuary Increased slope, elevation Logging, roads, PA edge Remote sensing National Park >> Nature Reserve and Wildlife Sanctuary. Gaveau et al., 2009 Lawenforecement, Staffing, Forest antilogging campaigns and eviction of Multible - PA compared to buffer rural communities Human populations density Remote sensing

Table S5. Continued IUC Source PA (short) Counterfactual Positive Drivers Negative Drivers Data type habitat N National Park >> Nature Reserve and Wildlife Sanctuary. Gaveau et al., 2009 Lawenforecement, Staffing, Forest antilogging campaigns and eviction of Multible - PA compared to region rural communities Human populations density Remote sensing Elevation, roads and human Hayes et al. 2002 Forest 5 NPs 4 BRs - PA compared to buffer infrastructure Remote sensing PA compared to similar habitats Logging and fires (for agricultural Ingram and Dawson, 2005 Forest All protected areas - outside expantions) Remote sensing Isolation, elevation, increased slope Joppa and Pfaff 2011 Global tropical IUCN category I and II were effective Human population density, roads, Multiple forested PA's - PA compared to matched outside depending on method rivers remote sensing Bukit Barisan Kinnaird et al. 2003 Forest Selatan NP II PA compared to buffer Increased slope converstion to agriculture Remote sensing Management planning, Kiragu Mwangi et al., 2010 Protected compared to none-protected implementation of management Multiple 36 IBAs - IBAs actions species specific threat Ground Guads, Integrated Conservation and Logging concessions, road Linkie et al., 2004 Forest Kerinci Seblat NP II PA compared to buffer development project constructions Remote sensing Establishment of PA and compared to Liu et al. 2001 Forest Wolong V buffer Ground New Jersey Luque 2000 Multiple Pinelands V PA compared to buffer Management plan Urban encroachment Remote sensing appendix PA compared to similar habitats Regulation of Elephant populations

155 Mapaure and Campbell, 2002 Forest Sengwa - outside and fires Large elephant populations Aerial Human population density, roads Mas 2005 *higher number outside in none- Forest Calakmul BR VI PA compared to matched outside Elevation and increased slope matched Remote sensing Human population density, roads Mas 2005 *higher number outside in none- Forest Calakmul BR VI PA compared to buffer Elevation and increased slope matched Remote sensing Wildlife sanctuaries (-0.26%) > Mendoza and Dirzo 1999 Forest Monte Azules BR VI PA compared to adjacent landscape protected areas (-0.31%) Human population density Remote sensing Amboro NP, Noel Kempff, Mercado Mertens et al. 2004 NP, BR, the Rios Forest Blanco and Negro Human settlements, roads, favorable WR - PA compared to adjacent landscape Isolation agricultural conditions Remote sensing Cuyabeno Wildlife human population density, poverty, Messina et al. 2006 Forest Production Reserve VI PA compared to buffer urban expantion Remote sensing management plan, thea plantages Remote and Aerial Mulley and Unruh, 2004 Forest Kibale NP II PA compared to buffer outside PA Human encroachment sensing Chitwan NP and PA compared to similar habitats Nagendra et al. 2008 Forest Parsa II,IV outside Isolation Grazing and fuel wood extraction Remote sensing Chitwan NP and Nagendra et al. 2008 Forest Parsa II,IV PA compared to buffer Isolation Grazing and fuel wood extraction Remote sensing Indigenous land and multi use Nelson and Chomitz, 2009 Forest IUCN I-IV - PA compared to matched outside protected areas Remote sensing Indigenous land and multi use Nelson and Chomitz, 2009 Forest IUCN I-IV - PA compared to matched outside protected areas Remote sensing Indigenous land and multi use Nelson and Chomitz, 2009 Forest IUCN I-IV - PA compared to matched outside protected areas Remote sensing

Table S5. Continued IUC Source PA (short) Counterfactual Positive Drivers Negative Drivers Data type habitat N Nelson and Chomitz, 2009 IUCN V-VI - PA compared to matched outside Remote sensing Forest Indigenous land and multi use Nelson and Chomitz, 2009 Forest IUCN V-VI - PA compared to matched outside protected areas Remote sensing Nelson and Chomitz, 2009 IUCN V-VI - PA compared to matched outside Remote sensing Forest Remote and Aerial Nelson et al. 2001 Forest Darién NP II PA compared to matched outside Slope and isolation sensing Indigenous reserves compared to Remote and Aerial Nelson et al. 2001 Forest Darién NP - matched outside Slope and isolation sensing 10 Extractive Nepstad et al. 2006 Forest reserves - Reserves compared to buffer Remote sensing 121 indigenous Indigenous reserves compared to Nepstad et al. 2006 Land tenure to indigenous people. Forest reserves - buffer Stricter protection Fires * indigenous areas Remote sensing Nepstad et al. 2006 18 National forest - National forests compared to buffer Fires * indigenous areas Remote sensing Forest Nepstad et al. 2006 33 PAs - PA compared to buffer Remote sensing Forest San Lorenzo, Oestreicher et al. 2009 Soberani´a, Chagres, Guard numbers, funds and NGO Agricultural expantion and logging Remote sensing Forest Altos de Campana - PA compared to buffer involvement concessions and interviews all in the amazon Oliveira et al. 2007 Forest region - PA compared to buffer protected areas > Indigenous lands Remote sensing GCA compared to entire country Management under national Pelkey et al., 2000 Multiple All GCA - outside protection duristiction, guard patrols Sub-national management duristiction Remote sensing appendix FR compared to entire country outside Management under national Pelkey et al., 2000 Multiple

156 All FR - protection duristiction, guard patrols Sub-national management duristiction Remote sensing PA compared to entire country outside Management under national Pelkey et al., 2000 Multiple All NP - protection duristiction, guard patrols Sub-national management duristiction Remote sensing Sader et al. 2001 MBR - PA compared to buffer Isolation Human settlement, roads and rivers Remote sensing Forest Sanchez-Azofeifa et al. 2002 Corcovado NP II PA compared to buffer Remote sensing Forest Sanchez-Azofeifa et al. 2003 20 NPs - PA compared to buffer Isolation logging for agriculture Remote sensing Forest Sanchez-Azofeifa et al. 2003 4 Biosphere reserves - BR compared to buffer Isolation logging for agriculture Remote sensing Forest 5,787 tropical Stricter protection IUCN I-II > IUCN Scharlemann et al. 2010 Forest forested PA's - III-VI Remote sensing PA compared to similar habitats Shearman and Bryan 2011 Forest 34 PAs - outside Isolation, elevation, increased slope Human population density Remote sensing Smith 2003 Bosawas VI PA compared to buffer Buffers End of civil war. Remote sensing Multiple Songer et al. 2009 Chatthin III PA compared to buffer Staff and research program Logging Remote sensing Forest Agricultural expantion, increased Southworth et al. 2004 Forest Celaque NP II PA compared to buffer Increased slope, NGO intitiatives coffee prices Remote sensing PA compared to similar habitats Tabor et al., 2010 Forest 75 PAs - outside Remote sensing PA compared to similar habitats Tabor et al., 2010 Forest 75 PAs - outside Remote sensing PA compared to similar habitats Tabor et al., 2010 Forest 30 KBAs - outside Remote sensing PA compared to similar habitats Tabor et al., 2010 Forest 2 AZEs - outside Remote sensing Subsistence encroachment, human Tole 2002 Forest Hellshire Hills - PA compared to buffer settlements, Edge effect Ground

Table S6 Reporting of Outcome Survey type of Improvement Predator prey Source Protected area Counterfactual other factors Species list measure outcome ratio conflicts and biases Disease, prey Animals pr. availability and Gates of the Arctic national Implementatio Adams et al., 2008 area / Radiotelematry 1/1 Not reported migration of Wolf park and preserve n of regulation Abundance non-resident wolfs Intra-specific Animals pr. PA compared competition Balme et al., 2010 Phinda-Mkhuze Complex area / Radiotelematry 1/1 Not reported Leopard to buffer and prey Abundance availability Instroduction Population Bhattacharya, 1993 Kaziranga National Park ground count 1/1 Not addressed Not reported Indian Rhio of staffing estimate PA compared occupancy Population Blake et al., 2007 6 protected areas ground transect 1/1 Not addressed Forest elephant to buffer time density Transect Weather and Establishment population Brereton et al., 2008 Multiple counts on 1/1 Not addressed grazing Chalkhill blue butterfly of PA estimate ground pressure Elephant, Hippopotamus, Giraffe, Buffalo, Eland, Roan, Sable, Animals pr. PA compared Aerial census, Food Zebra, Waterbuck, Greater kudu, Hartebeest, Topi, Bushpig, Caro, 1999 Katavi national park area / 7/8 Not addressed to buffer ground counts availability Warthog, Reedbuck, Impala, Bushbuck, Lion, Spotted hyanea, Abundance small carnivores, mongoose, Baboon, Vervet

Elephant, Hippopotamus, Giraffe, Buffalo, Eland, Roan, Sable, appendix Animals pr. Game Aerial census, Food Zebra, Waterbuck, Greater kudu, Hartebeest, Topi, Bushpig, Caro, 1999 Katavi national park area / 15/? Not addressed

157 contolled area ground counts availability Warthog, Reedbuck, Impala, Bushbuck, Lion, Spotted hyanea, Abundance small carnivores, mongoose, Baboon, Vervet Elephant, Hippopotamus, Giraffe, Buffalo, Eland, Roan, Sable, Animals pr. Aerial census, Food Zebra, Waterbuck, Greater kudu, Hartebeest, Topi, Bushpig, Caro, 1999 Katavi national park Forest reserve area / 5/16 Not addressed ground counts availability Warthog, Reedbuck, Impala, Bushbuck, Lion, Spotted hyanea, Abundance small carnivores, mongoose, Baboon, Vervet Common opossum, Nine-banded armadillo, Lesser anteater, Mantled howler monkey, Geoffroy's spider monkey, White- Different Animals pr. Both predators Isolation, faced capuchin monkey, Central American squirrel monkey, Corcovado national park and Carrillo et al., 2000 levels of area / ground transect N/A and prey weather and White-nosed coati, Raccoon, Southern river otter, Ocelot, Golfo Dulce forest reserve protection Abundance increased landuse Margay, Jaguar, Puma, White-lipped peccary, Collared peccary, Red brocket deer, Central American tapir, Peca, and Central American agouti Intra-specific Castro Verde spetial Introduction of Population capture- competition, Catry et al., 2009 1/1 Not reported Lesser kestrel protection areas artifical nests estimate recapture nest location and predation National Density Devictor et al., 2007 All protected areas estimates compared to spot count 20/30 Not addressed Not reported 100 bird species, see orginal article outside PA trends Population Implementatio Population Ground and Wolf and elk Eberhardt et al., 2007 Yellowstone national park 2/2 structure and Elf and Wolf n of regulation estimate aerial both improved predators Invasive Lassen Volcanic National Establishment count on Fellers and Drost, 1993 Presence 0/1 Not addressed species and Cascades frog Park of management locations habitat loss Introduction of Population Population and Gough and Kerley, 2006 Addo Elephant national park ground count 1/1 Not addressed African elephant fence estimate weather

Table S6. Continued Reporting of Outcome Survey type of Improvement Predator prey Source Protected area Counterfactual other factors Species list measure outcome ratio conflicts and biases Predators, inter-specific Animals pr. competition, Closing of Harrington et al., 1999 Kruger National Park area / Aerial census 1/1 Not reported weather, Roan antelope waterpoints Abundance population structure, disease Animals pr. Herremans and Herremans- PA compared Multiple area / spot count 47/47 Not addressed Weather 47 raptor species Tonnoeyr , 2000 to buffer Abundance Implementatio Hilborn et al. 2006 Serengeti national park Not reported Not reported 3/3 Not addressed Not reported Buffalo, Elephant and Black rhino n of regulation Different zones Population Ma et al., 2009 Yancheng biosphere reserve ground count 1/1 Not addressed Habitat quality Red-crowned crane of PA estimate Population structure, Establishment Population Aerial and Mduma et al., 1999 Serengeti national park 1/1 Not reported weather, food Wildebeest of PA estimate ground census availability, and predators Translocation Population Meijaard and Nijman, 2000 Pulau Kraget nature reserve ground count 0/1 Not addressed Not reported Proboscis monkey of population estimate

Landscape appendix properties, Implementatio Population 158 Metzger et al., 2010 Serengeti national park Aerial 1/1 Not reported food Buffalo n of regulation estimate availability, and predators Buffalo, Eland, Elephant, Grant's gazelle, Thomson's gazelle, PA compared Population Vegetation Ottichilo et al, 2000 Masai Mara national reserve aerial 12* Not addressed Giraffe, Impala, Kongoni, Ostrich, Topi, Warthog, and to buffer estimate types Waterbug None- Yok Don, Cat Tien national protected land Area of ground count Weather, and Pedrono et al., 2009 parks, Ea So and Vinh Cuu 1/1 Not addressed Banteng within species occupancy and DNA disease nature reserves range Arusha NP, Kilimanjaro NP Aardwolf, African civet, African palm civet, Banded and FR, Mahale NP, Lake mongoose, Bat-eared fox, Black-backed jackal, Bushy-tailed Manyara NP, Minziro FR, Different mongoose, Clawless otter, Caracal, Common genet, Dwarf Ngorongoro Conservation Encounter Camera Landscape Pettorelli et al., 2010 levels of 23** Not addressed mogoose, Honey badger, Large spotted genet, Leopard, Lion, Area, Serengeti NP, Tanga rate trapping properties protection Marsh mongoose, Serval, Side-stripped jackal, Slender CF, Tarangire NP, mongoose, Spotted hyena, White-tailed mongoose, Wild cat, Biharamulo-Burigi-Kimisi and Zorilla GR, Zoraninge FR, Silver-bordered fritillary, Regal fritillary, Orange sukphur, Delaware skipper, Common rnglet, Great spangled fritillary, Areas not Animals pr. Landscape Nothern brown, Aphrodite fritillary, Long dash, Pearl crescent, Schlicht et al. , 2009 Multiple managed with area / ground transect - Not addressed properties Meadow fritillary, Melissa blue, Common wood-nymph, fire Abundance Clouded sulphur, Black Swallowtail, Dakota skipper, Poweshiek skipperling, and Monarch Nest location, Territories bird spotting Nest predator Populations inter-specific Sergio et al., 2005 Doñana national park pr. area / and nest 2/2 populations Black kite and Red kite outside PA competition, Abundance inventory increased density prey

Table S6. Continued Reporting of Outcome Survey type of Improvement Predator prey Source Protected area Counterfactual other factors Species list measure outcome ratio conflicts and biases Food, predation, Implementatio Population Lion populations Sinclair et al., 2007 Serengeti national park Aerial 2/2 habitat, Buffalo and Wildebeest n of regulation estimate increased disease, and weather Burigi-Biharamulo national PA compared Stoner et al., 2007 Aerial 20% Not addressed park to buffer PA compared Stoner et al., 2007 Greater Ruaha national park Aerial 25% Not addressed to buffer Species traits, Buffalo, Eland, Elephant, Giraffe, Grant's Gazelle, Greater Animals pr. PA compared human density, kudu, Hartebeest, Impala, Puku, Oryx, Reedbuck, Roan, Sable, Stoner et al., 2007 Tarangire national park area / Aerial 10% Not addressed to buffer feeding guilt, Thomson's gazelle, Topi, Warthog, Waterbuck, Wildebeest, Abundance PA compared and weather and Zebra Stoner et al., 2007 Selous-Mikumi national park Aerial 5% Not addressed to buffer PA compared Stoner et al., 2007 Ugalla national park Aerial 70% Not addressed to buffer Status of PA compared Isolation and Struhsaker et al., 2005 16 protected areas fauna and questionare N/A Not reported Fauna and flora to buffer species traits flora Study PA compared populations Las Amoladeras reserve and Nest Ground Dupont's lark, Black-bellied sandgrouse, Little bustard, and Suarez et al., 1993 to similar 0/5 possibly limited Predators Layna Paramos mortality observations Stone curlew habitat outside by predators appendix inside PA

159 Predators Population and Introduction of Population increased as habitat Tambling and Du Toit, 2005 Pilanesburg national park Aerial 1/2 Lion and Wildebeest fence estimate target species structure, and declined weather Population PA compared estimate and Population Theberge et al., 2006 Algonquin provincial park Radiotelematry 1/1 Not addressed Wolf to buffer annual loss structure to hunters Habitat heterogeneity, Animals pr. Both predators Establishment Camera inter-specific Tiger, Leopard, Chital deer, Muntjac, Hog deer, Wild boar, Wegge et al, 2009 Bardia national Park area / 5/8 and prey of PA trapping competition, Barasingha, and Nilgai Abundance increased and prey density PA compared Population Not Weather and Western, 2009 Tsavo national park ground count Not addressed to buffer estimate reported habitat PA compared Population Not Weather and Western, 2009 Mara national park ground count Not addressed Elephant, buffalo, Burchell's zebra, giraffe, Wildebeest, Eland, to buffer estimate reported habitat Waterbuck, Warthog, Grant's gazelle, Thomson's gazelle, PA compared Population Not Weather and Western, 2009 Amboseli national park ground count Not addressed Impala, Lesser kudu, Oryx, Black rhino, Topi, and Hartebeest to buffer estimate reported habitat PA compared Population Not Weather and Western, 2009 Meru national park ground count Not addressed to buffer estimate reported habitat Managed Animals pr. Eradication of Population and section Whitehead et al., 2008 Fiordland national park area / ground count 1/1 invasive habitat Whio duck compared to Abundance predators structure unmanaged

appendix

Appendix II (chapter V)

Protected Areas ability to reduce pressure – Supporting Information Spatial and temporal changes in human threats to wild nature and impacts on protected areas between 1995-2010

Jonas Geldmann, Lucas N. Joppa and Neil D. Burgess

Threat systems We used the threat categories included in the Salafsky et al. (2008) and Baldwin (2010) systems for categorizing threats. The two have been integrated in Table 1 of the main manuscript.

Salafsky et al. (2008):

i) Residential and commercial development, ii) Agriculture and aquaculture, iii) Energy production and mining, iv) iv) Transportation and service corridors, v) Biological resource use, vi) Human intrusions and disturbance, vii) Natural system modifications, viii) Invasive and other problematic species and genes, ix) Pollution, x) Geological events, and xi) Climate change and severe weather.

Baldwin (2010):

i) landuse and landcover change, ii) pollutants, iii) invasive species, iv) disease, and v) climate change

Balmford et al. (2009), which has also inspired our approach, does not have an explicit number of categories, but is organized based on i) mechanisms and ii) sources. Using this approach, “habitat destruction” (mechanism) could be on of multiple categories depending on the source responsible (e.g. whether “industry”, or “climate change” was the responsible factor). In their paper, Balmford et al. (2009) operates with three first order mechanisms: i) habitat destruction, ii) habitat degradation and fragmentation, and iii) direct reduction in fecundity of target and 12 categories of sources: i) agriculture,

160 appendix

plantations, and aquaculture, ii) housing and urban development, iii) industry, iv) recreation and tourism, v) energy production and mining, vi) transportation and service corridors, vii) war and civil unrest, viii) harvesting wild species, ix) climate change, x) natural disasters and geological events, xi) others, and xii) unknown.

Validation of included data-layers Stable nightlights

The Defense Meteorological Satellite Program Operational Line Scanner program has been running since the 1970s and the US National Oceanic and Atmospheric Administration (NOAA) has used these data to develop an annual average stable night light products from 1992 an onwards, with a spatial resolution of 2.8 km2. From 1992- 2010, six different satellites (F10, F12, F14, F15, F16 and F18) have measured the night lights with some overlap. Stable nightlight is graded between 0-63 based on light intensity regardless of the light source, and the underlying drivers of the score may vary, complicating comparison of change across the world. Intensified light from cities is expected to be predominantly from an increase in electric lights and thus be directly related to human influence (Elvidge et al., 1997). However, light measured in remote areas may originate from natural or human intensified phenomena, such as forest or savannah fire, which can be highly spatially variable over time (Chand et al., 2007). Evaluating the specific impact of the changes in night light can therefore be challenging and need a posteriori consideration of the context in which changes are measured.

Data from satellites F12 and F18 were used for 1995 and 2010 respectively to estimate the light intensity. Earlier studies have shown different satellites in overlapping years to produce slightly different results, though this difference was found to be constant across year and not cumulative (Doll, 2008). The impact of the inaccuracy in the satellite’s light sensor is thus expected to decrease with the number of years over which real differences are compared. Data were spatially aggregated to a resolution of 2.5 arc minutes (approximately 5km2 at Equator) from their original resolution to reduce differences observed in time originating from possible miss-alignment of data. Human population

Human population estimates are generated by Center for International Earth Science Information Network (CIESIN) based on a combination of available census data and modeling (Balk et al., 2010). We used the GPWv3 which is the newest version that uses census data and population growth algorithms that do not include any secondary data sources for smoothing (CIESIN, 2005). The quality of the human population data is largely dependent on the underlying census data used for the estimation, which is of variable quality and spatial resolution.

161 appendix

The impact of changes in human population is not expected to be linear, with per capita impact decreasing with increasing population density (Shafik, 1994). To account for the non-linear relationship we therefore tested three different data-transformations as well as no transformation: i) log transformed, ii) square-root transformed, and iii) sigmoid transformed (Figure S1). Sigmoid transformed data was abandoned as this effectively gave a bimodal distribution of either no impact of a very high impact (about 10 times higher than stable nightlights). Log transforming human population data increased the relative weight of areas with very low population density, while changing the distribution of data considerable compared to the non-transformed (Figure S2a,c). Square root transformed data, like log transformed, decreases the impact of changes in areas already densely populated, as desired. At the same time this transformation yields a range distribution similar to the original data also visible when comparing the frequency distributions (Figure S2b,c). When comparing the influence of stable nightlights and human population density respectively, square root transformation also comes much closer to the desired equal weighting (Figure S3b).

Critical appraisal of the THPI Impact and drivers

Specific sources of pressure may have very different impacts on ground depending on the habitat and the condition under with the pressure works. For example; fires are considered damaging in many tropical forested habitats (Nelson & Chomitz, 2011), while at the same time essential in preserving many grass land systems (Newton et al., 2009). Legislative frameworks and governance structures also affects the impact from the same pressure (Bajracharya et al., 2005, Camargo et al., 2009), so that the even globally standardized indices are not easily compared between areas. At a species level, pressures may also have very positive impacts on some while being damaging to other which can result in changes in ecosystem-functioning not necessarily “bad” or “good” (For example: Fellers & Drost, 1993, Faubert et al., 2012), so that the result of the pressure dependents on what is being measured. For this reason, we have focused on data sources that measure pressure-input. Thus, like the “empty forest” hypothesis (Redford, 1992) our measure of change in pressure may hide pressure-impact that could be revealed with the inclusion of more information. We identify seven generic categories of impact modifiers that all affect how the same measure of input may have different impact on the ground: i) human behaviors and compliance, ii) ecosystem or organism specific responses, iii) Inter-correlations between inputs, iv) function/mode of impacts, v) historic conditions, vi) ecosystem complexity and displacement of effect, and vii) interactions with responses. However while these are important in interpreting the consequences of increased or decreased human pressure, they vary with local condition as well as scale and has not been considered in our map. Weighting of data

162 appendix

When examining composite data layers, a major challenge is weighting the influence of one source compared to another. Two factors needs to be considered when combining data sources: i) type of data (e.g. continues, discrete or grouped variables) and ii) shape of the expected response function (e.g. linear, log, sigmoid etc.) (Figure S1). In some cases empirical evidence can suggest an a priori approach, while expert opinion and a posteriori validation is needed in others. This is particularly challenging for data sources where weights depend on idiosyncratic decisions and expert opinion (Malczewski, 2006). The challenge and assumptions of weighting has inspired systematic conservation planning (Pressey et al., 1993, Margules & Pressey, 2000) and has led to the development of several software packages for reserve selection and landscape prioritization in general (Moilanen et al., 2009). These methods depend on an a posteriori evaluation of the results to interpret whether the used weighting is appropriate. However, because we do not select areas or evaluate decisions under a range of possible future scenarios, but only capture the change which has already occurred, we cannot in the same way evaluate the weighting of our difference source layers. We have therefore used the same methods as the original human footprint (Sanderson et al., 2002) which used an equal weighting of each data source, so that the maximum contribution of all layers were the same.

Data colliniarity When evaluating the influence of human population or stable nightlights separately, human population changes were the strongest elements of change in 58.4% of the world. No clear geographical pattern was observed in this distribution (Figure S6). Our map is therefore not only a map of either human population changes or changes in stable nightlight, but a combination of both. The correlation between scores of stable nightlight and human population density in 2010 was plotted, showing no signs of collinarity between the two data layers (Figure S7).

References Bajracharya SB, Furley PA, Newton AC (2005) Effectiveness of community involvement in delivering conservation benefits to the Annapurna Conservation Area, Nepal. Environmental Conservation, 32, 239-247.

Baldwin RF (2010) Identifying Keystone Threats to Biological Diversity. In: Landscape-scale Conservation Planning. (eds Trombulak SC, Baldwin RF) pp 17-32. Springer Netherlands.

Balk D, Yetman G, De Sherbiniin A (2010) Construction of Gridded Population and Poverty Data Sets from Different Data Sources. In: European Forum for Geostatistics Conference. pp 1-10, Tallinn, Estonia.

Balmford A, Carey P, Kapos V, Manica A, Rodrigues ASL, Scharlemann JPW, Green RE (2009) Capturing the Many Dimensions of Threat: Comment on Salafsky et al. Conservation Biology, 23, 482-487.

Bontemps S, Defourny P, Van Bogaert E (2010) CLOBECOVER 2009 - Product description and validation report. pp 30, European Space Agency.

163 appendix

Camargo C, Maldonado JH, Alvarado E et al. (2009) Community involvement in management for maintaining coral reef resilience and biodiversity in southern Caribbean marine protected areas. Biodiversity and Conservation, 18, 935-956.

Chand TRK, Badarinath KVS, Murthy MSR, Rajshekhar G, Elvidge CD, Tuttle BT (2007) Active forest fire monitoring in Uttaranchal State, India using multi-temporal DMSP- OLS and MODIS data. International Journal of Remote Sensing, 28, 2123-2132.

Ciesin (2005) Gridded Population of the World, Version 3 (GPWv3). Palisades, New York, Center for International Earth Science Information Network (CIESIN), Columbia University, Centro Internacional de Agricultura Tropical (CIAT). .

Doll CNH (2008) CIESIN Thematic Guide to Night-time Light Remote Sensing and its Applications. In: CIESIN Thematic Guide. pp 1-41, Columbia University Palisades, NY, USA, Center for International Earth Science Information Network (CIESIN).

Elvidge CD, Baugh KE, Kihn EA, Kroehl HW, Davis ER (1997) Mapping City Lights With Nighttime Data from the DMSP Operational Linescan System. Photogrammetric Engineering & Remote Sensing, 63, 727-734.

Esa (2006) GlobCover Project led by MEDIAS-France. Avaliable at: http://ionia1.esrin.esa.int/news/_faq.asp?id=35.

Faubert P, Tiiva P, Michelsen A, Rinnan Å, Ro-Poulsen H, Rinnan R (2012) The shift in plant species composition in a subarctic mountain birch forest floor due to climate change would modify the biogenic volatile organic compound emission profile. Plant and Soil, 352, 199-215.

Fellers GM, Drost CA (1993) Disappearance of the cascades frog Rana cascadae at the southern end of its range, California, USA. Biological Conservation, 65, 177-181.

Fritz S, See L (2008) Identifying and quantifying uncertainty and spatial disagreement in the comparison of Global Land Cover for different applications. Global Change Biology, 14, 1057-1075.

Malczewski J (2006) GIS-based multicriteria descision analysis: a survey of the literature. International Journal of Geographical Information Science, 20, 703-726.

Margules CR, Pressey RL (2000) Systematic conservation planning. Nature, 405, 243-253.

Moilanen A, Wilson K, Possingham H (2009) Spatial Conservation Prioritization Quantitative Methods and Computational Tools, Oxford, UK, Oxford University Press.

Nelson A, Chomitz KM (2011) Effectiveness of Strict vs. Multiple Use Protected Areas in Reducing Tropical Forest Fires: A Global Analysis Using Matching Methods. PLoS ONE, 6, e22722.

Newton A, Stewart G, Myers G, Lake S, Bullock J, Pullin AS (2009) How Does the Impact of Grazing on Heathland Compare with the Impact of Burning, Cutting or No Management? In: CEE review 05-008 (SR14). Collaboration for Environmental Evidence.

Pressey RL, Humphries CJ, Margules CR, Vane-Wright RI, Williams PH (1993) Beyond opportunism: Key principles for systematic reserve selection. Trends in Ecology & Evolution, 8, 124-128.

Redford KH (1992) The Empty Forest. Bioscience, 42, 412-422.

164 appendix

Salafsky N, Salzer D, Stattersfield AJ et al. (2008) A Standard Lexicon for Biodiversity Conservation: Unified Classifications of Threats and Actions. Conservation Biology, 22, 897-911.

Sanderson EW, Jaiteh M, Levy MA, Redford KH, Wannebo AV, Woolmer G (2002) The Human Footprint and the Last of the Wild. Bioscience, 52, 891-904.

Shafik N (1994) Economic Development and Environmental Quality: An Econometric Analysis. Oxford Economic Papers-New Series, 46, 757-773.

Supplementary Figures

Figure S1.Types of responses to changes in an independent variable ranging between 0-100 equivalent to the score of the THPI.

165

appendix 164

Figure S2. Frequncy distribution of THPI scores globally based on the tranformation of data. Top: log-transformation. Middle: squareroot transformation and buttom: no transformation.

appendix

Figure S3. Correlation between frequency scores of transformed values (y-axis) and original values (x- axis) for human population density. Left figure non-transformed against log transformed values. Right figure. Square root transformed values against non-transformed values.

Figure S4. Histogram of values for change in ClobeCover300 between 2004 and 2009. More than 99% of the terrestrial pixels have a values of zero (0) change using the Scaling suggested by Sanderson et al. 2002.

167 appendix

Figure S5. Hitogram of the residuals (the difference in THPI scores with or without inclusion of GlobeCover300 values). Inclussion of landcover change introduces very few and small differences.

Figure S6. Comparison between stable nightlights and human population. Yellow indicates areas where the human population score is greater than the stable nightlight scores. Black indicates areas where the stable nightlights are greater than human population. Blue indicates areas where both datasets are zero.

168 appendix

Figure S7. Corelation between Stable Nightlight scores and human population density in 2010. Values are log transformed. The figure shows not clear signs of intercorrelation.

3.5

2.5

1.5

0.5

Figure S8. Average THPI scores for the included realms between 1995 and 2010.

169

Supplementary Tables Table S1. Original categories used in GlobeCover300 and the values scores used in the production of the Human footprint (HII). category Description HII value 11 Post-flooding or irrigated croplands (or aquatic) 8 14 Rainfed croplands 3 20 Mosaic cropland (50-70%) / vegetation (grassland/shrubland/forest) (20-50%) 0 30 Mosaic vegetation (grassland/shrubland/forest) (50-70%) / cropland (20-50%) 0 40 Closed to open (>15%) broadleaved evergreen or semi-deciduous forest (>5m) 0 50 Closed (>40%) broadleaved deciduous forest (>5m) 0 60 Open (15-40%) broadleaved deciduous forest/woodland (>5m) 0

70 Closed (>40%) needleleaved evergreen forest (>5m) 0 appendix 90 Open (15-40%) needleleaved deciduous or evergreen forest (>5m) 0 170 100 Closed to open (>15%) mixed broadleaved and needleleaved forest (>5m) 0 110 Mosaic forest or shrubland (50-70%) / grassland (20-50%) 0 120 Mosaic grassland (50-70%) / forest or shrubland (20-50%) 0 130 Closed to open (>15%) (broadleaved or needleleaved, evergreen or deciduous) shrubland (<5m) 0 140 Closed to open (>15%) herbaceous vegetation (grassland, savannas or lichens/mosses) 0 150 Sparse (<15%) vegetation 0 160 Closed to open (>15%) broadleaved forest regularly flooded (semi-permanently or temporarily) - Fresh or brackish water 0 170 Closed (>40%) broadleaved forest or shrubland permanently flooded - Saline or brackish water 0 180 Closed to open (>15%) grassland or woody vegetation on regularly flooded or waterlogged soil - Fresh, brackish or saline water 0 190 Artificial surfaces and associated areas (Urban areas >50%) 10 200 Bare areas 0 210 Water bodies No Data 220 Permanent snow and ice 0 230 No data (burnt areas, clouds,…) 0

Table S2. Frequency scores for the change values in land cover change, following the impact-scheme used by Sanderson et al. (2002) All terrestrial areas Protected Areas Change Pixels Percent Pixels Percent Pixels Percent Pixels Percent category 300m 300m 5km 5km 300m 300m 5km 5km 10 268148 0.01 8 <0.01 10283 0.06 0 0.00 9 9 <0.01 1 <0.01 8 950768 0.05 78 <0.01 37855 0.22 6 <0.01 7 65791 <0.01 243 <0.01 2957 0.02 61 <0.01 6 601 0.01 165 0.01 5 1793584 0.09 1973 0.03 27432 0.16 285 0.01 4 3379 0.05 496 0.02 appendix 3 8190637 0.42 8562 0.12 178193 <0.01 1140 0.05 171 2 18710 <0.01 22278 0.33 2822 0.11 1 64333 0.94 468 <0.01 9058 0.37 0 1928448644 99.02 6670861 97.39 16636785 96.34 2440442 98.66 -1 58882 0.86 28 <0.01 14547 0.59 -2 6320 <0.01 13893 0.20 3271 0.13 -3 7300341 0.37 3169 0.05 353086 2.04 775 0.03 -4 619 0.01 137 0.01 -5 2 <0.01 388 0.01 2 <0.01 100 <0.01 -6 207 0.00 67 <0.01 -7 19553 <0.01 95 <0.01 584 <0.01 49 <0.01 -8 64 <0.01 15 <0.01 1 <0.01 30 <0.01 -9 53 <0.01 15 <0.01 -10 545276 0.03 33 <0.01 20498 0.12 12 <0.01 Values are the impact scores, so that 10 and -10 are the maximum possible increases or decreases in pressure and 0 is areas experiencing no changes between 2004 and 2009.

appendix Table S3. Summary of pressure changes across reams and biomes globally Minimum Maximum Meanpressure Standard Realm Biome Area (km2) pressure pressure score*† diviation* score* score* Afrotropic Tropical and Subtropical Moist Broadleaf Forests 3,482,220 -51.590 66.975 1.120 2.210 Afrotropic Tropical and Subtropical Dry Broadleaf Forests 193,167 -3.715 25.232 0.651 0.937 Afrotropic Tropical and Subtropical Grasslands, Savannas, and Shrublands 14,006,100 -22.480 68.511 0.979 1.585 Afrotropic Temperate Grasslands, Savannas, and Shrublands 25,758 -1.263 59.659 5.752 8.810 Afrotropic Flooded Grasslands and Savannas 459,038 -5.618 37.225 0.707 1.086 Afrotropic Montane Grasslands and Shrublands 863,919 -27.910 67.737 1.407 3.047 Afrotropic Mediterranean Forests, Woodlands, and Scrub 93,932 -23.243 37.369 0.839 3.043 Afrotropic Deserts and Xeric Shrublands 2,396,580 -34.614 80.150 0.976 3.551 Afrotropic Mangroves 70,199 -61.949 55.340 0.517 9.058 Australia Tropical and Subtropical Moist Broadleaf Forests 1,131,490 -15.674 37.472 0.473 0.984 Australia Tropical and Subtropical Dry Broadleaf Forests 83,386 -3.478 25.309 0.537 1.592 Australia Temperate Broadleaf and Mixed Forests 721,362 -28.807 43.105 0.458 2.087 Australia Tropical and Subtropical Grasslands, Savannas, and Shrublands 2,163,680 -22.002 28.975 -0.009 0.796 Australia Temperate Grasslands, Savannas, and Shrublands 627,507 -11.092 19.935 0.045 0.574 Australia Montane Grasslands and Shrublands 67,634 -3.943 14.650 0.177 0.634 Australia Mediterranean Forests, Woodlands, and Scrub 800,256 -16.426 41.146 0.123 1.199 Australia Deserts and Xeric Shrublands 3,578,720 -32.095 42.123 -0.012 0.480 Australia Mangroves 23,078 -11.869 13.812 0.255 0.737 Indo-Malay Tropical and Subtropical Moist Broadleaf Forests 5,375,180 -46.702 65.161 2.569 5.109 Indo-Malay Tropical and Subtropical Dry Broadleaf Forests 1,529,540 -27.215 60.660 3.439 4.987 Indo-Malay Tropical and Subtropical Coniferous Forests 96,053 -3.176 47.714 2.964 3.684 Indo-Malay Temperate Broadleaf and Mixed Forests 150,153 -22.767 52.535 1.490 2.822 Indo-Malay Temperate Coniferous Forests 67,499 -10.708 52.733 2.350 4.238 Indo-Malay Tropical and Subtropical Grasslands, Savannas, and Shrublands 34,588 -2.715 47.111 4.357 3.662 Indo-Malay Flooded Grasslands and Savannas 27,166 -0.751 8.371 0.291 0.589 Indo-Malay Montane Grasslands and Shrublands 4,396 -9.272 12.600 1.510 1.589 Indo-Malay Deserts and Xeric Shrublands 1,087,120 -24.974 60.959 4.295 4.979 Indo-Malay Mangroves 105,173 -28.094 61.367 3.562 6.744 Neotropical Tropical and Subtropical Moist Broadleaf Forests 9,263,590 -42.413 54.713 0.682 2.435 Neotropical Tropical and Subtropical Dry Broadleaf Forests 1,139,060 -29.011 52.070 1.231 3.257 Neotropical Tropical and Subtropical Coniferous Forests 324,734 -6.393 42.893 2.028 3.851 Neotropical Tropical and Subtropical Grasslands, Savannas, and Shrublands 3,991,000 -41.618 57.014 0.612 2.486 Neotropical Temperate Grasslands, Savannas, and Shrublands 1,624,150 -54.587 41.696 0.458 2.553 Neotropical Flooded Grasslands and Savannas 270,865 -6.256 36.883 0.582 2.416 Neotropical Montane Grasslands and Shrublands 874,831 -25.977 56.862 0.365 1.611 Neotropical Mediterranean Forests, Woodlands, and Scrub 147,878 -5.317 40.074 1.504 3.806 Neotropical Deserts and Xeric Shrublands 1,171,050 -46.368 49.778 1.048 3.052 Neotropical Mangroves 107,583 -22.831 49.610 1.857 4.715 North America Tropical and Subtropical Dry Broadleaf Forests 50,494 -5.130 20.062 0.722 2.015 North America Tropical and Subtropical Coniferous Forests 289,952 -13.521 32.609 0.114 1.310 North America Temperate Broadleaf and Mixed Forests 2,832,180 -27.058 54.388 0.735 4.303 North America Temperate Coniferous Forests 2,268,760 -39.612 53.583 0.705 3.022 North America Boreal Forests/Taiga 5,077,860 -37.140 61.199 -0.015 1.099 North America Tropical and Subtropical Grasslands, Savannas, and Shrublands 75,906 -17.407 52.889 2.182 4.901 North America Temperate Grasslands, Savannas, and Shrublands 3,097,390 -33.846 57.451 0.717 3.348 North America Tundra 3,682,180 -23.261 36.222 -0.017 0.631 North America Mediterranean Forests, Woodlands, and Scrub 120,250 -16.020 48.116 1.835 4.107 North America Deserts and Xeric Shrublands 2,321,550 -32.440 60.885 0.755 3.058 Oceania Tropical and Subtropical Moist Broadleaf Forests 23,271 -28.833 17.927 0.649 1.983 Oceania Tropical and Subtropical Dry Broadleaf Forests 13,940 -13.990 17.441 0.832 1.926 Oceania Tropical and Subtropical Grasslands, Savannas, and Shrublands 3,027 -1.811 8.079 0.469 1.198 Palearctic Tropical and Subtropical Moist Broadleaf Forests 510,786 -30.385 56.871 1.752 4.595 Palearctic Temperate Broadleaf and Mixed Forests 8,636,390 -53.503 100.000 2.499 6.680 Palearctic Temperate Coniferous Forests 1,707,670 -32.618 70.564 0.810 3.310 Palearctic Boreal Forests/Taiga 9,944,970 -59.864 66.053 0.095 2.531 Palearctic Temperate Grasslands, Savannas, and Shrublands 4,716,920 -41.651 61.209 0.534 3.010 Palearctic Flooded Grasslands and Savannas 338,384 -17.988 67.735 4.716 9.576 Palearctic Montane Grasslands and Shrublands 3,391,470 -21.037 64.985 0.390 1.981 Palearctic Tundra 3,725,580 -50.002 66.018 -0.056 1.978 Palearctic Mediterranean Forests, Woodlands, and Scrub 2,027,290 -38.517 56.638 3.793 6.205 Palearctic Deserts and Xeric Shrublands 17,395,900 -56.001 86.258 0.559 2.708 * units are Temporal Human Pressure Index (THPI). † Values used in figure 4 compared with data from Sanderson et al. (2002).

172

appendix Appendix III (chapter VI)

Table S1. Variables collected by the Management Effectiveness Tracking Tool Category Notes Budget Annual budget of the protected area. Often a combination of core funding and project funding (Input) Staff The number of people working in the protected area (Input) Primary threats Divided into 43 categories ranked from not relevant to severe (0-3) (Context) Primary objectives The primary objectives of the protected area (Context) Critical Management activities Most important activities in the protected area to achieve the primary objectives. (Planning/process) 0= The protected area is not gazetted 1 Legal status 1= There is agreement that the protected area should be gazetted 2= The protected area is in the process of being azetted (Context) 3= The protected area has been formally gazetted/covenanted 0 = There are no regulations 2 Protected area regulations 1 = Regulations with major weaknesses 2 = Regulations with some weaknesses or gaps (Planning) 3 = Regulations provide an excellent basis for management 0 = No effective capacity/resources 3 Law enforcement 1 = There are major deficiencies in staff capacity/resources 2 = The staff have acceptable capacity/resources (Input) 3 = The staff have excellent capacity/resources 0 = No firm objectives have been agreed for the protected area 4 Protected area objectives 1 = Objectives exist, but not managed according to these 2 = Objectives exist, but is only partially managed according to these (Planning) 3 = Objectives exist, and is managed to meet these 0 = Inadequacies in protected area design mean achieving the major objectives of the protected area is very difficult 1 = Inadequacies in protected area design mean that achievement of major objectives is difficult but some mitigating actions are 5 Protected area design being taken 2 = Protected area design is not significantly constraining achievement of objectives, but could be improved (Planning) 3 = Protected area design helps achievement of objectives; it is appropriate for species and habitat conservation 0 = The boundary of the protected area is not known 6 Protected area boundary 1 = The boundary of the protected area is known by the management authority but is not known by local residents 2 = The boundary of the protected area is known but is not demarcated (Process) 3 = The boundary of the protected area is known and is appropriately demarcated 0 = There is no management plan 7 Management plan 1 = Management plan is not being implemented 2 = Management plans is partially implemented (Planning) 3 = A management plan exists and is being implemented 0 = No regular work plan exists 8 Regular work plan 1 = Exists but few of the activities are implemented 2 = Exists and many activities are implemented (Planning/output) 3 = Exists and all activities are implemented 0 = There is little or no information available on the critical habitats, species and cultural values of the protected area 9 Resource inventory 1 = Information is not sufficient to support planning and decision making 2 = Information is sufficient for most key areas (Input) 3 = Information is sufficient to support all areas 0 = There is no survey or research work taking place 10 Research 1 = There is a small amount of survey and research work 2 = There is considerable survey and research work (Process) 3 = There is a comprehensive, integrated research programme 0 = Active resource management is not being undertaken 11 Resource management 1 = Very few of the requirements for active management are being implemented 2 = Many of the requirements for active management are being implemented (Process) 3 = Requirements are being substantially or fully implemented 0 = There are no staff 12 Staff numbers 1 = Staff numbers are inadequate 2 = Staff numbers are below optimum (Input) 3 = Staff numbers are adequate 0 = Problems with personnel management constrain the achievement of major management objectives 13 Personal management 1 = Problems with personnel management partially constrain the achievement of major management objectives 2 = Personnel management is adequate to the achievement of major management objectives but could be improved (Input/process) 3 = Personnel management is excellent and aids the achievement major management objectives 0 = Staff lack the skills needed for protected area management 14 Staff training 1 = Staff training and skills are low relative to the needs 2 = Staff training and skills are adequate, but could be further improved to fully achieve the objectives of management (Input/process) 3 = Staff training and skills are aligned with the management needs

173

appendix Category Notes 0 = There is no budget 15 Current budget 1 = The available budget is inadequate for basic management needs 2 = The available budget is acceptable but could be further improved (Input) 3 =The available budget is sufficient 0 = Wholly reliant on outside or highly variable funding 16 Security of budget 1 = There is very little secure budget 2 = There is a reasonably secure core budget (Input) 3 = There is a secure budget 0 = Budget management is very poor and significantly undermines effectiveness 17 Management of budget 1 = Budget management is poor and constrains effectiveness 2 = Budget management is adequate but could be improved (Process) 3 =Budget management is excellent and meets management needs 0 = There are little or no equipment and facilities 18 Equipment 1 = There are some equipment and facilities but these are inadequate 2 = There are equipment and facilities, but still some gaps (Input) 3 = There are adequate equipment and facilities 0 = There is little or no maintenance of equipment and facilities 19 Maintenance of equipment 1 = There is some ad hoc maintenance of equipment and facilities

2 = There is basic maintenance of equipment and facilities (Process) 3 = Equipment and facilities are well maintained 0 = There is no education and awareness programme 20 Education program 1 = There is a limited and ad hoc education and awareness programme 2 = There is an education and awareness programme but it only partly meets needs (Process) 3 =There is an appropriate and implemented education and awareness programme 0 = There is no contact between managers and neighbouring official 21 State and comm. Neighbors 1 = There is contact between managers and neighbouring official but little or no cooperation

2 = There is contact between managers and neighbouring official but only some co-operation (Process) 3 = There is regular contact between managers and neighbouring official 0 = Indigenous and traditional peoples have no input into decisions 22 Indigenous people 1 = Indigenous and traditional peoples have some input into discussions 2 = Indigenous and traditional peoples directly contribute to some decisions (Process) 3 = Indigenous and traditional peoples directly participate in all relevant decisions 0 = Local communities have no input into decisions 23 Local communities 1 = Local communities have some input into discussions 2 = Local communities directly contribute to some relevant decisions (Process) 3 = Local communities directly participate in all relevant decisions 0 = There are no visitor facilities and services despite an identified need 24 Visitor facilities 1 = Visitor facilities and services are inappropriate for current levels of visitation 2 = Visitor facilities and services are adequate for current levels of visitation but could be improved (Outputs) 3 = Visitor facilities and services are excellent for current levels of visitation 0 = There is little or no contact between managers and tourism operators 1 = There is contact between managers and tourism operators but this is largely confined to administrative or regulatory matters 25 Commercial tourism 2 = There is limited co-operation between managers and tourism operators to enhance visitor experiences and maintain protected area values (Process) 3 = There is good co-operation between managers and tourism operators to enhance visitor experiences, and maintain protected area values 0 = Although fees are theoretically applied, they are not collected 1 = Fees are collected, but make no contribution to the protected area 26 Fees 2 = Fees are collected, and make some contribution to the protected area 3 = Fees are collected and make a substantial contribution to the protected area 0 = Many important biodiversity, ecological or cultural values are being severely degraded

1 = Some biodiversity, ecological or cultural values are being severely degraded 27 Condition assessment 2 = Some biodiversity, ecological and cultural values are being partially degraded but the most important values have not been

significantly impacted (Outcome) 3 = Biodiversity, ecological and cultural values are predominantly intact 0 = Protection systems are ineffective in controlling access or use of the reserve in accordance with designated objectives 1 = Protection systems are only partially effective in controlling access or use of the reserve in accordance with designated

objectives 28 Access assessment 2 = Protection systems are moderately effective in controlling access or use of the reserve in accordance with designated

objectives (Output) 3 = Protection systems are largely or wholly effective in controlling access or use of the reserve in accordance with designated objectives 0 = The protected area does not deliver economic benefits to local communities 29 Economic benefit assessment 1 = Potential economic benefits are recognized. Plans are being developed (Outcome) 2 = There is some flow of economic benefits to local communities 3 = There is a major flow of economic benefits to local communities 0 = There is no monitoring and evaluation 30 Monitoring and evaluation 1 = There is some ad hoc monitoring and evaluation, but no overall strategy 2 = There is an agreed and implemented monitoring and evaluation system but results do not feed back into management (Panning/ Process) 3 = A good monitoring and evaluation system exists, and is well implemented Based on the second version of the METT evaluation. A never version has been launched but this has only been completed for a very small subset of the sites included in the PhD thesis. And where the METT evaluations using the third versions has been included, they have been modified to fit the format of the second version.

174

appendix Table S2. Species included in the analysis Species Common name Weight (g) Red list status Aepyceros melampus Impala 52591.69 LC Alcelaphus buselaphus Hartebeest 160937.86 LC Antidorcas marsupialis Springbok 33571.24 LC Bos frontalis Gaur 800143.05 VU Canis simensis Ethiopian wolf 14361.86 EN Cephalophus natalensis Duiker, Harvey's 12724.51 LC Ceratotherium simum Southern white rhinoceros 2285939.43 NT Chelonia mydas Green turtle 166300 EN Connochaetes gnou Wildebeest, Black 156547.53 LC Connochaetes taurinus Wildebeest, Blue 198619.68 LC Copsychus sechellarum Seychelles magpie robin 77 EN Crocodylus niloticus Crocodile 362500 LC Crocuta crocuta Spotted hyaena 63369.1 LC Damaliscus korrigum Korrigum 127194.87 LC Damaliscus lunatus Topi 136000.33 LC Damaliscus pygargus Blesbok 77784.55 LC Dermochelys coriacea Leatherback turtle 268000 CR Diceros bicornis Black rhinoceros 995940.54 CR Elephas maximus Asian elephant 3269794.34 EN Equus burchellii Zebra, Burchells 279160.65 LC Equus zebra Cape Mountain Zebra 282462.13 VU Eudorcas rufifrons Red fronted gazelle 26999.77 VU Giraffa camelopardalis Giraffe 964654.73 LC Gyps bengalensis White-rumped vulture 4385 CR Gyps indicus Long-billed vulture 5515 CR Hippopotamus amphibius Hippopotamus 1536310.4 VU Hippotragus equinus Roan antelope 264173.96 LC Hippotragus niger Sable antelope 236405.9 LC Hyaena brunnea Brown Hyaena 40000 NT Kobus ellipsiprymnus Waterbuck 204393.48 LC Kobus kob Kob 80035.21 LC Kobus leche Red Lechwe 88645.04 LC Kobus vardonii Puku 71462.72 NT Larus dominicanus Kelp gull 1118 LC Leptailurus serval Serval 10250 LC Loxodonta africana African Elephant 3824539.93 NT Mastomys coucha Southern Multimammate mouse 53.81 LC Micaelamys namaquensis Namaqua Micaelamys,Namaqua Rock Rat 57.1 LC Mus minutoides pygmy mouse 6.43 LC Myotomys unisulcatus Bush Karroo Rat 102.5 LC Octodon degus Degu 203.275 LC Oreotragus oreotragus Klipspringer 13486.55 LC Oryx gazella Gemsbok 188404.45 LC Otomys irroratus Vlei Rat 114.45 LC Otus ireneae Sokoke Scops Owl 50.3 EN Ozotoceros bezoarticus Pampas Deer 34620.4 NT Pan troglodytes Chimpanzee 45000 EN Panthera pardus Leopard 52399.99 NT Panthera tigris Bengal tiger 161914.6825 EN Pelea capreolus Grey Rhebok 22731.33 NT Pelecanus onocrotalus Great White Pelican 533 LC Phacochoerus africanus Warthog 82499.99 LC Phalacrocorax capensis Cape cormorant 1330 NT Platanista gangetica Ganges river dolphin 93481.965 EN Podocnemis unifilis Yellow-spotted river turtle 2254 VU Potamochoerus larvatus Bush-pig 69063.79 LC Raphicerus campestris Steenbok 11661.53 LC Redunca arundinum Reedbuck 58059.24 LC Redunca fulvorufula Mountain Reedbuck 29352.66 LC Redunca redunca Bohar reedbuck 43288.95 LC Rhabdomys dilectus Mesic four-striped grass rat 35 LC Rusa unicolor Sambar 177522.9 VU Struthio camelus Ostrich 111000 LC Sus scrofa Wild Boar 84471.54 LC Sylvicapra grimmia Duiker, Common 15639.15 LC Syncerus caffer Buffalo 592665.98 LC Tapirus indicus Asian Tapir 311209.19 EN Taurotragus oryx Eland 562592.69 LC Tragelaphus angasii Nyala 87616.76 LC Tragelaphus oryx Eland 560 LC Tragelaphus scriptus Bushbuck 43250.39 LC Tragelaphus strepsiceros Greater kudu 206056.41 LC

175

appendix Table S3. List of protected area included in the analysis

WDPA-id Country IUCN cat. Established Size (ha) 24 BHS II 1965 908.6 101 CHL II 1989 89.6 167 CRI II 1970 797.6 608 CMR II 1968 1415.8 691 IND II 1977 719.1 694 IND II 1975 904.4 805 NPL II 1973 1184.3 873 ZAF NR 1926 19093.5 875 ZAF II 1931 1410.2 876 ZAF II 1979 770.1 877 ZAF II 1937 246.6 884 NAM II 1975 23063.5 887 NAM II 1990 397.8 917 TZA II 1964 10894.3 1085 ZMB II 1971 22314.3 1086 ZMB II 1971 8985.5 1099 ZMB II 1971 437.0 1308 NPL II 1976 912.5 1309 NPL IV 1976 370.8 1350 ZAF II 1972 294.7 1361 NAM IV 1969 9.9 1362 NAM II 1968 40.2 1399 TZA IV 1905 47971.6 1405 THA Ia 1974 3716.0 1438 UGA III 1968 476.3 1792 IND IV 1950 1019.4 1808 IND II 1980 555.1 2281 ETH II 1969 1579.1 2499 ECU VI 1979 5898.0 4029 ZAF II 1956 322.4 4081 ZMB VI 1971 32421.0 6939 SYC Ia 1975 1.7 7422 KEN NR 1943 370.8 7434 TZA NR 1969 1182.2 7962 ZMB II 1983 4153.7 9061 ZAF II 1985 19.3 9129 ZAF IV 1979 470.2 12414 IND IV 1987 426.9 17368 ZAF II 1985 242.5 17999 NAM II 2008 3861.2 18004 NAM V 1972 41.4 18005 NAM V 1968 244.2 19604 ARG NR 1982 32883 ZAF IV 1990 236.81 39758 ZAF IV 1983 301.070 39832 ZAF IV 1992 598.02 102310 CRI II 1991 273.20 116189 ZAF II 1895 903.47 116257 ZAF II 1987 661.69 198302 ZAF NA 1999 2331.29 300597 ZAF IV NA 187.38 351272 ZAF NR NA 214.06 351523 ZAF NR NA 53.50 NR = not reported, NA = not applicable. ISO 3 codes are given for country. Name and location of the individual protected areas can be found on: www.protectedplnaet.net/sites/wdpa_id.

Table S4. Model parameter estimates for the most parsimonious model using only time-series from African countries. Parameter Estimate S.E. t-value p-values Intercept 0.5455 0.4087 1.335 0.184 Management plan level 2 -0.5682 0.5154 -1.103 0.273 Management plan level 3 -0.6519 0.4462 -1.461 0.147 Human Development Index 2005 -0.4622 0.7099 -0.651 0.516 Management plan 2* HDI 05 0.5480 0.9016 0.608 0.545 Management plan 3* HDI 05 0.7319 0.7937 0.922 0.359

176

A B appendix 177

Figure S1. Frequency plot of the original data distribution of all management categories included from METT score cards (A) before transformation and (B) after transformation.

appendix

Figure S2. Correlation between country Human Development Index in 2005 and 2011. There is a highly significant correlations (t = 250.97 p < 0.0001) between the two measures with the biggest difference between 2010 and 2005 being 0.05.

178

appendix 179

Figure S3. Histograms of slopes for the 279 slopes for mammals, birds and reptiles included in the analysis with 2 observations (A), and the 184 populations included with 3 observations (B).