Species Distribution Ranges and Conservation: Spatial Structure, Contraction Patterns, Global Change Impacts, and Bias in Species Data

A aquellos que aman la naturaleza

RECOMMENDED CITATION: Lucas, P.M. (2016) Species distribution ranges and conservation: Spatial structure, contraction patterns, global change impacts, and bias in species data. PhD Thesis. Pablo de Olavide University, Sevilla, Spain.

DESIGN & LAYOUT: Pablo Miguel Lucas. Show the distribution range of the Brown bear (Ursus arctos), one of the most originally widely distributed terrestrial mammals in the world but which have suffered an important range contraction (Red color) due to persecution and habitat destruction by humans. Its current distribution (Blue color) shows continuous big areas at high latitudes and fragmented populations in areas of Europe.

Species distribution ranges and conservation: Spatial structure, contraction patterns, global change impacts, and bias in species data

— Distribuciones de especies y conservación Estructura espacial, patrones de contracción, impactos de cambio global y sesgos en los datos de especies

Pablo Miguel Lucas

PhD Thesis

Sevilla, 2016

Estación Biológica de Doñana, CSIC Departamento de Biología de la Conservación

Universidad Pablo de Olavide Facultad de Ciencias Experimentales Departamento de Biología Molecular e Ingeniería Bioquímica Doctorado en Estudios Medioambientales

"Species distribution ranges and conservation: Spatial structure, contraction patterns, global change impacts, and bias in species data"

Memoria presentada por el Licenciado en Ciencias Ambientales Pablo Miguel Lucas Ibáñez para optar al título de Doctor por la Universidad Pablo de Olavide

Fdo. Pablo Miguel Lucas Ibáñez Dr. Eloy Revilla Sánchez, Estación Biológica de Doñana - CSIC, Sevilla (España) y

Dra. Manuela González Suárez, University of Reading, Reading (Reino Unido)

CERTIFICAN

Que los trabajos de investigación desarrollados en la Memoria de Tesis Doctoral "Species distribution ranges and conservation: Spatial structure, contraction patterns, global change impacts, and bias in species data" son aptos para ser presentados por el Licenciado Pablo Miguel Lucas Ibáñez ante el Tribunal que en su día se designe, para aspirar al Grado de Doctor por la Universidad Pablo de Olavide.

Y para que así conste, y en cumplimiento de las disposiciones legales vigentes, extendemos el presente certificado a 7 de Octubre de 2016.

Directores:

Fdo. Eloy Revilla Sánchez Fdo. Manuela González Suárez

Tutor:

Fdo. Eduardo Narbona Fernández Universidad Pablo de Olavide

Look deep into nature, and then you will understand everything better.

Albert Einstein

Index

Summary ...... 12 Resumen ...... 14 Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities ...... 16 Introduction ...... 18 Range structure and dynamics ...... 18 Ranges in the antropocene ...... 25 Gaps and opportunities ...... 33 Chapter 1: Toward multifactorial null models of range contraction in terrestrial vertebrates . 37 Abstract...... 39 Introduction ...... 40 Methods ...... 42 Results ...... 47 Discussion ...... 52 Acknowledgments ...... 55 Supplementary material ...... 56 Chapter 2: Size matters, and so does spatial configuration: predicting vulnerability to extinction in vertebrates...... 115 Abstract...... 117 Introduction ...... 118 Material and methods ...... 121 Results ...... 124 Discussion ...... 132 Acknowledgements ...... 135 Supplementary material ...... 137 Chapter 3: The roles of climate change and land use in recent terrestrial vertebrate range contractions ...... 139 Abstract...... 141 Introduction ...... 141 Material and methods ...... 144 Results ...... 148 Discussion ...... 153 Acknowledgments ...... 158 Supplementary material ...... 159 Chapter 4: Biases in comparative analyses of extinction risk: mind the gap...... 266 Abstract...... 268 Introduction ...... 269 Materials and Methods ...... 271 Results ...... 277 Discussion ...... 285 Acknowledgments ...... 289 Supporting material ...... 291 Conclusions ...... 298 Bibliography ...... 302 Agradecimientos ...... 319

Summary

The impact of human activities is causing an accelerated loss of biodiversity. The two main human impacts, habitat loss through land use modifications and climate change are predicted to be widespread and more intense in the future, increasing the risk of extinction of many species. Therefore, there is an urgent necessity to improve our understanding of how extinctions occur in order to improve predictions and to maximize the effectiveness of conservation policies. The complete extinction of a species is usually preceded by a process of extinction of its local populations which causes a reduction of the geographic extension of its distribution range, in a process called range contraction. Distribution ranges are naturally dynamic, but nowadays, the impact of human activities is the main factor driving colonizations and local extinctions. The spatial structure and dynamics of species’ ranges are related to their ecology, their coexistence with other species and other historical and evolutionary factors, affecting the overall probability of extinction and recolonization. These factors and processes show variability among taxa due to differences in the spatial characteristics of the range including the prevalence of human impacts. Studying the structure and dynamics of geographic ranges including how they contract offers the opportunity to assess the vulnerability of populations to human impacts and understand the processes leading to global extinction. In the present thesis I explore several aspects of range dynamics including whether we can define a null model of range contraction (Chapter 1), how the spatial structure of the range determines the vulnerability of species (Chapter 2), the distinct and interacting effects of climate change and land use in driving distribution range contractions (Chapter 3), and the existing bias in data availability and how these biases influence our ability to assess vulnerability to extinction (Chapter 4). In the first chapter, I show that traditional null models of range contraction fail to explain range contractions in terrestrial vertebrates and propose new multifactorial models that combine land use and spatial configuration to generate baseline expectations of range contraction. In the second chapter, I identify several spatial configuration descriptors, in addition to the widely use total range area, which are associated with vulnerability to extinction. In particular, fragment size heterogeneity and shape are the most important predictors, but there are also multiple interactions that illustrate the complex ways in which spatial configuration influences extinction risk. In the third chapter, I show the importance of

Summary both land use and climate change in currently observed biodiversity loses. The combination of both threats leads to severe range contraction, with species remaining in refuge areas which may not be representative of their original niche breadth. In the fourth chapter, I explore possible bias in species data availability and their effect on the findings of macroecological studies of extinction risk.

Resumen Resumen

El impacto de las actividades humanas está causando una acelerada pérdida de biodiversidad. Los dos principales impactos humanos, pérdida de hábitat a través de modificaciones del uso del suelo y cambio climático están predichos a estar más extendidos y ser más intensos en el futuro, incrementando el riesgo de extinción de muchas especies. Por lo tanto, hay una necesidad urgente de mejorar nuestro entendimiento de cómo las extinciones ocurren para mejorar las predicciones y maximizar la eficacia de las políticas de conservación. La completa extinción de una especie es normalmente precedida por la reducción de la extensión geográfica de su distribución, en un proceso denominado contracción de la distribución. Las distribuciones son dinámicas por naturaleza, pero normalmente, el impacto de las actividades humanas es el principal factor que dirige las colonizaciones y extinciones locales. La estructura espacial y la dinámica de las distribuciones de las especies están relacionadas con su ecología, su coexistencia con otras especies y otros factores históricos y evolutivos, afectando todos ellos a la probabilidad de extinción y recolonización. Estos factores y procesos muestran variabilidad entre taxones debido a diferencias en las características espaciales de la distribución incluyendo el predominio de los impactos humanos. Estudiando la estructura y la dinámica de las distribuciones geográficas incluyendo como estas se contraen ofrece la oportunidad de analizar la vulnerabilidad de poblaciones a impactos humanos y entender los procesos que llevan a la extinción global. En la presente tesis exploro diferentes aspectos de la dinámica de las distribuciones incluyendo si nosotros podemos definir un modelo nulo de dinámica de la contracción (Capítulo 1), como la estructura espacial de la distribución determina la vulnerabilidad de las especies (Capítulo 2), los diferentes efectos y sus interacciones del cambio climático y uso del suelo sobre las contracciones de las distribuciones (Capítulo 3), y los sesgos existentes en la disponibilidad de datos y cómo estos sesgos influyen en nuestra habilidad para analizar vulnerabilidad a la extinción (Capítulo 4). En el primer capítulo, yo muestro que los tradicionales modelos nulos de contracción de las distribuciones fallan al explicar las contracciones en vertebrados terrestres y propongo nuevos modelos multifactoriales que combinan uso del suelo con configuración espacial para generar un punto de partida de lo que se espera en las contracciones de las distribuciones. En el segundo capítulo, identifico diferentes descriptores de configuración espacial, además del ampliamente usado área total de la distribución, que son asociados con vulnerabilidad a la extinción. En particular, heterogeneidad del tamaño del fragmento y forma son los más importantes predictores, pero hay además múltiples interacciones que ilustran las complejas maneras en las que la configuración espacial influencia el riesgo a la extinción. En el tercer capítulo, muestro la importancia de ambos, uso del suelo y cambio climático en observaciones actuales de pérdidas de biodiversidad. La combinación de ambas amenazas conduce a severas contracciones de la distribución, con especies manteniéndose en áreas de refugio las cuales parece que no son representativas de la variedad de su nicho original. En el cuarto capítulo, exploro los posibles sesgos en la disponibilidad de datos de las especies y sus efectos en los resultados de estudios macroecológicos de riesgo de extinción. 14

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

Lucas, P.M; González-Suárez, M; Revilla, E. (in prep) Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities Introduction

The distribution range of a species is a concept used to describe the geographical extent where a species can be found. Species distribution ranges are often complex, varying in the number of fragments/populations, local abundances and total area. Additionally, ranges change over time due to colonization and local extinction events which can occur at different time scales for different species (Evans et al. 2005, Gaston 2003, Gaston 2008, Gaston and Curnutt 1998, Webb and Gaston 2000). Colonizations and extinctions are affected by species- specific life history, ecological and biological traits (Kier et al. 2009). In addition, human impacts also shape species ranges favoring the expansion of some species and contributing to the extinction of others (Crees et al. 2016, Turvey et al. 2015, Vilà et al. 2011). Understanding the spatial configuration and dynamics of species distribution ranges has been important for the fields of biogeography, evolution and ecology (Cunningham et al. 2016, Gaston 2003, Gaston and Fuller 2009, Holt and Barfield 2009). In the face of the current biodiversity crisis, this understanding has also become critical in conservation biology. Vulnerability to extinction is strongly influenced by distribution range size and spatial configuration. Here I review our current understanding of distribution ranges dynamics exploring how range spatial configuration is associated with vulnerability and human impacts. I conclude with a discussion of knowledge gaps and future conservation issues.

Range structure and dynamics

Range area Distribution range areas can be described using two different approaches: the extent of occurrence or the area of occupancy (Gaston 1991). The extent of occurrence is the area which lies within the outermost geographic limits of the occurrences of a species. The area of occupancy is the area within those outer most limits over which it actually occurs. Regardless of the approach used the final delineation of a range area is determined by the working scale, with areas potentially included or excluded at different scales and resolutions (Cowley et al. 2001, Wilson et al. 2004). Species range areas are considered fundamental units in

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities biogeography and are a well-used descriptor in other fields as well because they are relatively easy to obtain. For example, the majority of ranges for amphibians, birds and mammals are available in the IUCN website (International Union for Conservation of Nature 2010c). As discussed above, range size and configuration are also important to evaluate extinction risk and management actions (Cardillo 2003, Cardillo et al. 2008, Cardillo et al. 2005, Cardillo et al. 2004, González-Suárez et al. 2013, González-Suárez et al. 2012). Distribution range sizes can vary orders of magnitude among species (Brown et al. 1996, Gaston 1996). There are species with small ranges which live in small islands with no more than a few square kilometers, while others occupy a large fraction of the planet, e.g., black rats or humans. The frequency distribution of range sizes is strongly right skewed, many species have small ranges but only a few have very large ranges (Gaston 1996, Webb and Gaston 2000). Three factors have been proposed to explain this variation in range area: (1) niche breath, (2) evolutionary history, and (3) species traits such as dispersal capacity (Olalla-Tárraga et al. 2011, Thompson et al. 1999). Niche breadth is one of the most commonly cited factors to explain range area. Species with broad niches can occupy and maintain populations in a greater number of habitats, thus potentially occupying a larger range area than species with narrower niche requirements which, as specialists, should tend to have smaller ranges (Brown 1984, Gaston 1994b, Kunin and Gaston 1993). This relationship between niche breadth and range size is important for conservation because specialization influences the responses of species to current and future environmental conditions (Heim and Peters 2011, Slatyer et al. 2013). When studying niche breadth at range scale the recommended niche characterization is the realized Grinellian niche that characterises the abiotic conditions (Devictor et al. 2010, Grinnell 1917, Hirzel and Le Lay 2008, Olalla-Tárraga et al. 2011, Soberon 2007, Wiens et al. 2010). Eltonian niches characterize the biotic conditions (diet, species competition) and are less well studied and more relevant at population scales (Botts et al. 2013). The influence of niche specialization on range size has been studied in different taxa and geographical regions with a special focus on the climatic (Grinellian) niche. These studies have, for example, identified positive associations between range size and climate breadth in North American trees (Morin and Lechowicz 2013) and in alpine and subalpine plants species (Yu et al. 2016). Although assumed as a general pattern, some studies have found the opposite, with geographically rare species being diet generalists and widespread species diet specialists 19

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities suggesting interaction with other factors (Williams et al. 2006). Slatyer et al (2013) reviewed the evidence of the relationship between niche breadth and range area based on 64 different studies. They found high among-species variability in response, but in general, there was a positive relationship between habitat tolerance and diet breadth. Several studies suggest that the positive relationship between niche breadth and range size may indicate that habitat loss and climate change will have a stronger impact on more specialized species due to synergistic effects between a narrower niche and smaller range size (Chen et al. 2011, Lenoir and Svenning 2015, Ohlemüller et al. 2008, Slatyer et al. 2013). Another factor explaining variation in range size is evolutionary history. For example, Morueta-Holme et al (2013) showed that plant species with broader range sizes had an evolutionary history associated with climate instability whereas species with small ranges concentrated more in areas where climate have been stable over longer periods of time. The final factor that determines range size is species dispersal ability (Brown et al. 1996, Gaston 1996, 2003). Lester et al (2007) indicated that mechanisms which explain this association can be categorized in three groups: (1) the site colonization hypothesis, which is based on theoretical metapopulation models (Hanski 1998, Hanski 1999), states that a low ability to disperse will generate low occupation of farther areas even if suitable habitat is available, thus ranges will be small (Thompson et al. 1999); (2) the speciation rate hypothesis which states that species with low dispersal ability will have greater isolation and lower gene flow among populations leading to higher speciation rates that in the long run will divide the original species ranges into multiple fragments occupied by new species (Kozak and Wiens 2016, Scholl and Wiens 2016); and (3) the selection hypothesis, which explain that it is possible that range size could be a cause of dispersal ability instead of an effect,as occurs in situations leading to selection for reduced dispersal in species with small ranges (e.g. energy conservation contributes to the loss of flight ability in island birds, (McNab 1994) Lester et al (2007) concluded that dispersal ability is likely not the main drivers of range size but can play a role in particular cases. In addition to these three most commonly evaluated predictors or correlates of range size, there are other factors that can be important including body size, population abundance, latitude, and environmental variability (Gaston 2003, Thompson et al. 1999). Often many of these factors co-occur, e.g. species with bigger niche breadth usually have high ability to disperse and are often large; therefore, it is difficult to identify single factors and often 20

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities different processes and characteristics may interact to ultimately shape the size and spatial configuration of a range (Lester et al. 2007). Finally, humnan impacts are behind many reductions and expansions of range area, probably being one of the most important determinants of current range size for many species.

Range boundaries Gaston (1990) defines the range limit as “the point where the sum of immigration and birth declines below emigration and death”. We should add to this definition a temporal extent on which it applies. Defined geographical range limits are essentially a simplification of reality because no species is found in all areas of its range nor it is always completely absent from areas outside (Gaston 2003). Instead, range limits should be understood as fuzzy boundaries that can help visualize and represent the distribution of species. Ranges borders are spatially complex with complexity increasing at smaller scales where more discontinuities and gaps become apparent (Wilson et al. 2004). There are two general types of limits in a range: exterior limits refer to the outermost external occurrences e.g. polewards range limits of the treeline in the taiga; and interior limits which represent the discontinuities and gaps within the external limits e.g. the treeline elevational limit. Both of these limits are dynamic over time, e.g. changes in weather such as an extraordinary cold wave could shift the treeline towards the Equator (shift in external limits) and/or to lower altitudes (internal limit shift). In summary, a working definition of range limits could be: imaginary lines which separate areas of very low abundance (as a proxy to complete absence, which is very difficult to quantify) at given spatial resolution and time interval. Gaston (Gaston 2003) proposes a global explanation to understand what influences range limits within a species based on three components: (1) abiotic and biotic interactions; (2) population dynamics, and (3) evolutionary constraints. In terms of the abiotic and biotic factors that determine range limits, there are multiple studies relating climate, dispersal barriers or interactions with other species (Gaston 2009, Parmesan et al. 2005, Sexton et al. 2009). Physical barriers are among the more obvious factors limiting dispersal and examples include oceans, high mountain ranges, rivers and deserts (Moussalli et al. 2009, Smissen et al. 2013). These physical barriers can affect a great number of species, being particularly important for insular species (Weigelt et al. 2015), and often delimits terrestrial biomes. Another very well studied abiotic factor delimiting species ranges at broad scales is climate 21

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

(Buckley and Jetz 2007, Cunningham et al. 2016, Sunday et al. 2011, 2012). For example Root (Root 1988) showed that the northern range limits of North American birds coincided with the minimum January temperature and the mean length of frost-free period. Similarly, the altitudinal and latitudinal limit of many forests is usually associated with minimum temperatures (Morueta-Holme et al. 2015). In koalas their western range limit is linked to the prevalence of extreme drought events (Seabrook et al. 2014). These range limits change with climate and weather oscillations, including changes associated to annual variation (species moving up and down a mountain during the year) to inter-annual or decadal changes (Seabrook et al. 2014). The influence of climate is also modulated by habitat conditions; microhabitat variation can generate refuges within largely unsuitable zones. Refuge areas have been relevant for species persistence during past climatic changes maintaining populations for subsequent colonization and expansion (Bennett and Provan 2008, Tzedakis et al. 2002), (Jackson and Betancourt 2009). Climatic variables directly delimit the distribution of the species and also have an indirect influence in other relevant factors, i.e., climate conditions influence soil conditions which in turn affect the distribution of many plant species (Cunningham et al. 2016). Biotic interactions, including competition, predation, and parasitism, are also important to explain species range limits with direct and indirect effects, often also mediated by abiotic factors (Ockendon et al. 2014). Araújo and Rozenfeld (2014) suggested that biotic interactions are the most relevant drivers at local scales, while climate is the key driver at large scales. Theoretical studies have highlighted the importance of species interactions in delimiting range boundaries in small communities (Gilman et al. 2010). In an empirical study, Gross and Price (2000) also showed that the northern range limit of Phylloscopus humei, the Hume’s yellow-browed leaf warbler, is driven by the distribution of its main food source which is in turn limited by climatic conditions. On the other hand, the warbler’s southern range limit is best explained by competition with P. trochiloides. Overall, there are fewer studies exploring biotic factors than abiotic factors, likely because these interactions are more difficult to quantify (Lavergne et al. 2010). Nevertheless, as the importance of these biotic interactions becomes clear, new species distribution models that incorporate these biotic factors are being proposed (Cazelles et al. 2015). The second component proposed by Gaston (Gaston 2003) as important to understand range limits are species’ population dynamics. Biotic and abiotic factors 22

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities determine the number of individuals of a population and its rates of death/births and migrations (Brown 1971, David et al. 2003, Jones and Diamond 1976, Pimm et al. 1988). The population dynamics will shape the distribtuon of a species (David et al. 2003, MacArthur and Wilson 1963, MacArthur and Wilson 1967). These parameters (number of individuals, rates of death/births and migration) depend of the quality of the habitat (Bascompte and Solé 1995, Forman 1995, Richard T. T. Forman and Godron 1986), the spatial structure (Bascompte and Solé 1998, David Tilman and Kareiva 1997, Hanski 1999, Levins 1969, MacArthur and Wilson 1967), the connectivity in the landscape (Gyllenberg and Hanski 1992, Hanski and Gyllenberg 1997, Hanski and Gyllenberg 1993), and any threatening factors (Bascompte and Solé 1995, Hanski and Ovaskainen 2000, Quinn and Hastings 1987). Populations located near range limits often have less suitable, lower quality and more fragmented, habitat than populations in the central areas of the range, thus often have lower growth rates (fewer births and more deaths) and smaller abundances. Lower rates make these populations more susceptible to extinction and because they are also less well-connected to other populations, when extinction occurs these areas are less likely to be recolonized. The third component that explains range limits are evolutionary constraints, the factors that prevent species from evolving traits to adapt to new habitats and disperse farther (Mayr 1963). Evolutionary constraints may be associated with low genetic variation and heritability at the range borders, trade-offs among traits that restrain changes, and differences in gene flow from central to peripheral populations (Gaston 2003). While potentially important this third component has not been well-studied and the role of evolutionary processes in range delimitation is yet to be clearly understood.

Patterns of abundance across a range As explained previously, local abundance varies across a species’ range influencing species range boundaries, population dynamic and gene flow. This variation is important to understand macroecological patterns, population dynamics, ecological interactions, and for conservation (Sagarin et al. 2006). In general, ranges are characterized by few areas with a high abundance of individuals and many with relatively low abundances, so histograms of frequency of abundance are often positively skewed (Gaston 1990, 1994b, Gaston and Curnutt 1998).

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

A general rule of thumb in biogeography, called “abundant-center hypothesis”, states that the center of a species’ range represents the area with the most favorable conditions for the species and thus should have the highest population density (Hengeveld and Haeck 1982, Whittaker 1956). This hypothesis, based on theoretical ecology, has been frequently applied to generate other hypotheses or predict patterns of evolutionary changes and vulnerability to extinction. For instance, it is often assumed that border areas are more prone to extinction (Sagarin and Gaines 2002, Sagarin et al. 2006). However, numerous studies (Sagarin and Gaines (2003, 2002) that have tested the abundant-center hypothesis has found its predictions are often not supported. The abundant-center hypothesis is very simplistic compared to observed patterns of abundance (Gaston 2003, Sagarin and Gaines 2002) and some of its assumptions fail to represent reality. For instance, the best habitat is not always found at the center of ranges, and changes in quality are rarely gradual as this hypothesis assumes. Because changes in quality are not gradual ranges are often not defined as one continuous area but rather formed by different fragments. As a result it also becomes difficult to identify a unique center, or if identified using common methods, the “central” area may even be located outside the occupied fragments. As the generality of the abundant-center hypothesis continues to be questioned, it is becoming apparent that many factors determine abundance within a range, and that these factors can interact in complicated ways which render simple rules of thumb inadequate (Lucas et al. 2016, Sagarin et al. 2006). Sagarin and Gaines (2006) suggested we need to consider population demography, biophysical variables and genetic population structure to properly understand range limits and abundance patterns (Lucas et al. 2016).

Range dynamics Range dynamics reflect the changes in abundance and spatial configuration that occur within a species distribution range over a given time period. Dynamics can be driven by changes in biotic and abiotic factors and by evolutionary processes that change species requirements. Pease (1989) used theoretical models to explore range dynamics and suggested that species ranges are dynamic mostly due to changes in the environmental variables with a less important role of adaptation. Davis and Shaw (Davis and Shaw 2001) also propose that range dynamics are the result of environmental changes followed by niche tracking, rather than adaptation. In biogeography, niche conservatism is often assumed, under the expectation that 24

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities species’ niche will change little over time (Huntley et al. 1989, Peterson et al. 1999). Pearman et al (2008) explored this assumption and showed that conservatism is not a general pattern, with several studies showing evidence of rapid niche shifts (Ackerly 2003, Dietz and Edwards 2006, Olalla‐Tárraga et al. 2016, Orr and Smith 1998). These findings are important to understand the consequences and responses that different species may have to anthropogenic-driven climate change and habitat loss/change. If species are more likely to track their niche, species confined to areas where niche tracking is not possible would be likely to become extinct. If, instead, niche shifts are possible, species may persist although their characteristics may also be altered, with potential consequences for the communities they belong to.

Ranges in the antropocene

Many large scale patterns, including species diversity, macroecological and biogeographical patterns, have been intensely modified by humans (Murray and Dickman 2000, Murray et al. 2011, Steffen et al. 2015a, Watson et al. 2016). Many species have reduced their geographic ranges, changed its spatial configuration and/or its patterns of abundance or have even become globally extinct (Ceballos and Ehrlich 2002, Channell and Lomolino 2000a, b, Rodrıguez and Delibes 2003, Rodríguez and Delibes 2002, Rodriguez 2002). However, other species have been positively affected by human changes and have extended their historical/natural ranges (Vilà et al. 2011, Vilà and Ibáñez 2011). Human induced changes started 100 kyr ago, with the megafauna extinctions of the Holocene in which natural climatic events combined with overexploitation by humans likely caused the extinction of several large terrestrial species (Bartlett et al. 2016). Nevertheless, it has been during the postindustrial era when unprecedented shifts in the Earth System indicators driven by human activities have occurred (Steffen et al. 2015a). Humans have led to a new period of biodiversity loss, commonly referred to as “the Sixth Mass Extinction”, with rates of extinction three orders of magnitude greater than background rates (Pimm et al. 2014). During this time land use, climate change, invasive species, pollution, and overexploitation have been the main anthropogenic drivers of extinctions (Maxwell et al. 2016, Pimm et al. 2006). Given this situation there is an urgent need to understand biodiversity loss, exploring

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities how different human impacts influence species’ vulnerability to extinction and their distribution within the planet. This exploration needs to account for how different species respond based on their distinct traits, including those that affect their distribution ranges (Urban et al. 2016).

Range expansions in the Anthropocene: invasive species Invasive species are one of the main threats to biodiversity (Clavero and García-Berthou 2005, McGeoch et al. 2010, Sala et al. 2000) and an increasing global problem (Hulme 2009). Invasive species have direct impacts reducing fitness, growth, abundance, modifying behavior of other species, as well as indirect impacts via induced changes in soil composition or nutrient cycles (Reinhart et al. 2003, Vilà et al. 2011). These impacts can reduce the abundance or even lead to extirpation of native species, changing species richness at local and regional scales (Levine et al. 2003, Pino et al. 2009, Vilà et al. 2011) with potentially enormous economic costs (Pimentel et al. 2000). The same invasive species can also have different impacts over distinct ecosystems. Insular species are most threatened by invasive species because they usually have small ranges population sizes (making them intrinsically more vulnerable) and because island endemics usually have fewer behavioral, morphological, and life-history defenses (Jones et al. 2008, Kier et al. 2009, Medina et al. 2011). Based on the definition of Pyšek et al. (2004), an invasive species is an organism present in an area due to human action (intended or accidental) that produces a high number of new individuals and has high dispersal capacity. This definition highlights two important conditions: (1) the natural range borders have been extended by human action, and (2) once introduced, species traits like reproductive speed, dispersal ability, and niche breadth play a determinant role allowing invasive species to persist and expand (Vilà et al. 2011). Concerned by future impacts of invasive species, several studies have aimed to detect if invasive and non-invasive species have distinct characteristics (Richardson and Rejmánek 2011). Van Kleunen et al. (2010) found that invasive species had higher values of leaf-area allocation, shoot allocation, rapid growth, and fitness, thus, it could be possible to predict what plant species could become invasive from their characteristics. In addition to understanding which species may become invasive, we may also wonder what may be the potential new limits to an invasive species’ range and how fast it is likely to expand to fill these limits. These two questions are fundamental to best allocate 26

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities efforts to control the expansion and the impact of these species (Morrison et al. 2004). To predict potential range limits, Morrison et al. (2004), for example, used an ecophysiological model that relies on temperature and precipitation data. But usually these models assumes that niche preferences are conserved (Gallien et al. 2010, Jiménez-Valverde et al. 2011, Václavík and Meentemeyer 2012, Venette et al. 2010). In reality, invasive species show high levels of plasticity (Davidson et al. 2011) and they are able to shift/expand their niche. For example, Argentine ants can change their diet in new areas (Rödder and Lötters 2009). As the biotic/abiotic conditions in the invasive range areas are often quite different from the native ranges, it can be difficult to predict how a species may behave in its new environments. Jakobs et al (2004) compared population structure and abundance of Solidago gigantea (Asteraceae) in native and introduced areas. They found that invasive populations were more abundant and had greater biomass than populations in native areas. Finally, to determine where and how an invasive species can expand it is also important to consider interactions among different factors. For example, climate change has been reported to favor the expansion of some invasive species (Muhlfeld et al. 2014, Zeidberg and Robison 2007).

Range contractions in the Anthropocene: endangered species Species extinction is usually a process which starts with the disappearance of local populations leading to a reduction in range area (Gilpin and Soulé 1986, Yackulic et al. 2011), but also causing changes in the spatial configuration and abundance patterns across the range (Di Fonzo et al. 2016, Rodrıguez and Delibes 2003, Rodríguez and Delibes 2002, Rodriguez 2002, Turvey et al. 2015, Turvey et al. 2016). Extirpation of local populations is often considered as a better measure of biodiversity loss than global species extinctions (Ceballos and Ehrlich 2002) because: (1) local extirpations are more likely to occur (easier to detect and quantify) than global extinctions (Hanski 1998, Hanski 1999, Hanski and Ovaskainen 2000); (2) biotic interactions/ecosystems services can be altered/lost when a local extirpation occurs even if the species is not globally extinct (Hollings et al. 2016, Tylianakis et al. 2008); and (3) local extirpations often reduce genetic diversity (Janecka et al. 2014). Studying range contractions also offers the possibility to understand the cause and consequences of biodiversity loss before extinction occurs; thus, providing means for more effective management and proactive conservation actions (Channell and Lomolino 2000a, b, Safi and Pettorelli 2010). As we previously explained, range dynamics are naturally complex 27

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

(Gaston 2003, Gaston 2009), thus understanding the consequences of the direct and indirect effects of human impacts is not easy (Yackulic et al. 2011).

Null models of range contraction The causes which lead to extinction can be numerous, change over time, and be influenced by intrinsic species’ traits and among-species interactions. Therefore, we may a priori expect a variety of patterns of range contractions. However, we may also expect commonalities and general patterns, which ecologists and conservation biologists have verbalized as null models or simple hypotheses making predictions on how ranges should contract. One of these null models, the demographic hypothesis, assumes that the center of the range is more suitable for the species, thus populations located in the borders are less abundant and have lower genetic variability and poorer connectivity with other populations compared to populations located in the center of a geographic range (Brown 1995, Lawton 1993). Metapopulation/landscape ecology (Bascompte and Solé 1995, Bascompte and Solé 1996, Bascompte and Solé 1998, Hanski 1998, Hanski 1999, Hanski and Ovaskainen 2000, MacArthur and Wilson 1967) and population ecology theory (Brown 1971, David et al. 2003, Jones and Diamond 1976, Pimm et al. 1988) tells us that vulnerability to extinction correlates inversely with the number of individuals, the connectivity within populations, and with genetic variability. Therefore, the demographic hypothesis predicts that ranges should contract inwards as border populations are more likely to become extinct (Brown 1995, Lawton 1993). This basic prediction from the demographic hypothesis has been often assumed to be true for conservation planning of endangered species, focusing efforts on core areas within the geographic range (Lawton 1993). However, Channell and Lomolino (2000a) showed that this prediction is not often supported by observed historical and current distribution data. They analyzed data from multiple species across a wide range of taxonomic groups and geographical regions and found that remaining populations were often located near the borders of historical ranges. In their study, they also showed that larger fragments are more likely to persist, which confirms at range scale the predictions of theoretical/population ecology; and that for species with mainland and insular historical ranges, contrary to theoretical predictions, insular areas, even being smaller range fragments, had lower probability of extinction. Channell and Lomolino (2000a) suggested that these findings can be explained by a new null model of range contraction, the contagion hypothesis, which 28

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities postulates that contraction is driven by human impacts that spread like a contagion across the landscape and render density patterns largely irrelevant. In a subsequent study, these same authors (2000b) tested their proposed contagion hypothesis reporting the expected patterns: an initial loss of populations near a border, followed by extirpation of populations located in central areas, and the last populations remaining in the opposite border from where extirpations started. A third proposed null model of range contraction, called here the refuge hypothesis, postulates that extinction probability is largely associated to human use and not directly predictable from the position of a population within a range (Ceballos and Ehrlich 2002, Fisher 2011, Hoffmann et al. 2010, Laliberte and Ripple 2004, Li et al. 2015, Maxwell et al. 2016, Pomara et al. 2014, Schipper et al. 2008, Watson et al. 2016, Yackulic et al. 2011). This hypothesis predicts that populations are first extirpated in that areas more modified/used by humans and that remaining populations should occur in less used areas that act as refuges. Lucas et al. (Chapter 1) compared these three null models: demographic, contagion, and refuge, using global data from terrestrial vertebrates and concluded that neither model was sufficiently general to predict the observed patterns. Then, assuming that range dynamics usually depend on several biotic and abiotic factors/processes, including human impacts, these authors proposed and tested new null multifactorial models of range contractions that simutaneously incorporate several of the processes assumed by the null models. These new combined models were found to offer a better fit to the available data. Nevertheless, different multifactorial models were identified as best depending on the scale, suggesting that different and potentially complex mechanisms drive range contraction in vertebrates. Simple null models are too simplistic and overall, multifactorial null models provide a better baseline to understand and predict future changes.

Human impacts and patterns of contraction Human impacts are currently the main cause of range contraction, with climate change and land use recognized as the most important drivers (Hoffmann et al. 2010, Parmesan and Yohe 2003, Pounds et al. 2006, Sekercioglu et al. 2008, Steffen et al. 2015b, Thomas et al. 2004, Thuiller et al. 2005a, Thuiller et al. 2005b). As previously described, climate is one of the main factors determining geographic range limits. This effect can be direct, e.g. physiological limits, or indirect via its effects on other species or abiotic factors, e.g. favoring a competitive 29

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities species or changing oxygen concentration in water. Climate change entails changes in mean values, e.g., increased mean temperatures, but also changes in variance and the frequency of extreme climatic events (IPCC 2015). These changes have already affected distribution range dynamics for many species and are predicted to affect many more in the future. Although climate change is of great concern, human land use has been and continues to be the most damaging human impact (Maxwell et al. 2016, Sala et al. 2000) Chapter 3). Land use changes directly lead to habitat loss and fragmentation, but also have indirect effects via altered nitrogen deposition, induced climate change, increased overexploitation, and introduction and expansion of invasive species (Newbold et al. 2015). Worryingly impacts from both land use and climate change are expected to increase in extension and magnitude during this century, often acting jointly on the same populations and species (Sala et al. 2000). Reliable projections of the effects of human impacts are a priority for species conservation, but their validation is a challenge. One alternative is to use past species range shifts to understand how different impacts affect species distributions. These analyses may allow us to develop more mechanistic models that can generate improved projections (Araújo et al. 2005, Urban et al. 2016). Even though species are generally affected by several impacts simultaneously, most assessments of range shifts have focused on single human impacts, particularly on climate change (Oliver and Morecroft 2014). As a consequence, the potential effects of simultaneous factors may be misrepresented or their effects wrongly attributed (Oliver and Morecroft 2014). Some recent studies have addressed this issue exploring several impacts concurrently. For example, Kerr et al (Kerr et al. 2015) evaluated the effects of climate, land use, and pesticide use in observed range shifts in bumblebees. For terrestrial vertebrates, a recent study (Chapter 3) evaluated climate change and land use effects on terrestrial vertebrates and found that although land use is the most important factor explaining range shifts in this group of species, climate change has also played a role. The authors concluded that considering only one of these impacts considerably reduces our capacity to explain range shifts. Indeed human impacts have been shown to interact in potentially complex and often worrisome ways (Oliver et al. 2015, Oliver and Morecroft 2014, Urban et al. 2016). Late quaternary extinctions can provide interesting information to advance our knowledge of natural and human impacts on range dynamics. Nevertheless, data traditionally used in paleontological studies come from archeological sites and often have a more limited 30

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities spatial scale and reflect a wider time frame compared to historical distribution datasets. Therefore, inferences need to be made with caution. The Holocene was characterized by climatic changes and the first apparent human impacts (which were, nevertheless, several orders of magnitude weaker than current human impacts). The causes of observed extinction over this period have generated an intense debate (Alroy 2001, Koch and Barnosky 2006, Lorenzen et al. 2011), with many studies suggestting that both types of impacts, climate change and humans, have played a role in past extinctions, and also often reporting differences in impact importance among populations and species (Bartlett et al. 2016, Lorenzen et al. 2011, Villavicencio et al. 2016). Some of these taxonomic differences have been associated to species traits such as trophic position (Crees et al. 2016, Villavicencio et al. 2016). Past and current climatic changes have been associated with shifts in environmental conditions, including changes in latitude and elevation (Davis and Shaw 2001). In order to remain within their climatic niche as warming occurs due to climate change, species tend to shift their ranges to higher latitudes (Hill et al. 1999, Parmesan et al. 1999), higher elevations (Menéndez and González‐Megías 2014, Merrill et al. 2008, Morueta-Holme et al. 2015, Sekercioglu et al. 2008, Shoo et al. 2005, Tingley et al. 2009), and steeper areas which create microrefugia (Hannah et al. 2014). However, land use changes can also cause range shifts in elevation or latitude due to direct competition for space with humans who prefer flatter temperate or warm lowlands (Keppel et al. 2012). Therefore, it can be difficult to attribute observed patterns to specific impacts. Recent work evaluating both climate change and human land use shows slightly different patterns associated with each impact and with their combined effects: (1) species associated to climate change, showed contraction to higher latitudes mainly at intermediate-higher latitudes; contraction to more elevated areas, mainly at lower latitudes; (2) species associated to land use, showed contraction to lower latitudes, no clear change for elevation and a tendency to contract to steeper areas at lower latitudes; and (3) species associated to both impacts showed bigger rate of range contraction and clearer environmental patterns, contraction towards higher latitudes, higher altitudes and steeper areas (Chapter 3). Understanding this complexity is key to identify areas which may act as refuges in future years (Gómez and Lunt 2007).

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

Projections Projections indicate that effects of the main drivers of global change, including land use change and climate change, will continue and even accelerate during this century (Lawton and May 1995). Species are disappearing fast and will continue disappearing, thus, it is urgent to improve our ability to predict how, where, and who is most vulnerable to extinction. Several studies have predicted the future biodiversity consequences of climate change (IPCC 2007). However, during the next century the most important driver is expected to still be land use changes, followed by climate change, nitrogen deposition, biotic exchange and atmospheric CO2 (Sala et al. 2000, Thuiller et al. 2005b). The different effects of these impacts across biomes will be likely variable, with Artic regions highly impacted by climate change and Tropical areas particularly affected by land use (Jetz et al. 2007). However, these projections are fairly simplistic and rarely account for impact interactions (Brook et al. 2008, Oliver et al. 2015, Oliver and Morecroft 2014) or how the different biotic and abiotic conditions influence the magnitude of the impacts (Tylianakis et al. 2008).

The role of species traits Although impacts are widespread, not all species respond equally, some have experienced important declines others are expanding or remain stable (Gaston 1994b, Gaston and Blackburn 1995, Laurance 1991) and for the majority we have no information. These differences are often associated to the life-history, behavioral, and ecological traits of the different species. The value of these intrinsic traits as predictors of vulnerability has been explored in many studies (Purvis et al. 2000). In context of the current global extinction crisis understanding which species are more prone to extinction is of crucial importance to conservation biology (Safi and Pettorelli 2010). However, necessary demographic data are not available for most taxa. Nevertheless, traits may be useful proxies. There are robust ecological basis to propose that several characteristics such as a small population size, endemic status, or large body size could determine vulnerability (McKinney 1997, Simberloff 1998). Based on these expectations, Purvis et al. (2000) presented the first wide and systematic study using species’ traits to predict vulnerability in the mammalian families Carnivora and Primates. They found that high trophic position, low abundance, slow growth, and small geographical range were associated with increased vulnerability. However, traits alone do not tell the whole history, which traits make species more vulnerable partly depends 32

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities on what the main threats are (Cardillo et al. 2005, Cardillo et al. 2004, González-Suárez et al. 2013). Comparative studies of species traits have offered interesting insights but also may present some limitations. These analyses rely on available information, in some cases compiled in useful and public databases like PANTHERIA (Jones et al. 2009); or often requiring an intense search and compilation effort. Unfortunately, some taxa, traits, and areas are better studied that others, which causes biases in the available information. These biases can unfortunately influence our ability to detect how species’ traits influence vulnerability to extinction and lead to erroneous conclusions if results are not interpreted with caution (Chapter 4). Understanding the limitations of these data, including geographic range area, is important to ensure conservation measures are adequately designed and implemented. Geographic range area is often considered a species trait (although as we have discussed above it is influenced by both species and environmental characteristics) and has been found to be one of the best predictors of vulnerability (Bielby et al. 2008, Davidson et al. 2009, Giam et al. 2011, Harris and Pimm 2008). Range area is generally correlated with population size, and larger populations are overall less likely to go extinct (David et al. 2003, Jones and Diamond 1976). Additionally, the spatial structure of populations is associated with extinction risk (Bascompte and Solé 1998, David Tilman and Kareiva 1997, Hanski 1999, Levins 1969, MacArthur and Wilson 1967). At the range scale its structure is also associated with vulnerability, but causality is poorly understood and it is possible that interactions exist with other species-specific traits (Chapter 2). The process of extinction is driven by many factors and it is not surprising that vulnerability predictors differ among taxa, biogeographical regions and threatening factors (Cardillo et al. 2008, Fisher and Owens 2004, Pinsky et al. 2011). Combining information about all these components, albeit complicated, is important to better understand and more importantly, predict extinction risk.

Gaps and opportunities

The complexity in the configuration and dynamics of geographic ranges represent a big challenge, but also a great opportunity to address key questions in ecology, evolution, biogeography, and conservation biology. These questions can generally be grouped into three

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities main themes: (1) exploring the role of natural processes; (2) evaluating the effect of human impacts; and (3) quantifying/predicting the effects/consequences for species vulnerability to extinction. Many studies (see sections above) have focused on studying total range size, but there are additional descriptors of spatial structure, including number and shape of fragments, and their connectivity that remain poorly understood. We need to better understand how natural processes and human impacts influence spatial configuration, and how this configuration in turn determines extinction risk. Similarly, we need to gain a better grasp of how natural processes affect range dynamics. It is often assumed that observed ranges represent an equilibrium, but ranges are naturally dynamic and our ability to attribute changes to human impacts is greatly affected by our inability to describe how natural changes occur. Unfortunately, understanding dynamics requires studying long and detailed time series ideally representing different taxa and biomes, yet these datasets are rarely available. Understanding dynamics will also require identifying and describing mechanisms, which can be challenging but also is essential to gain improved predictive capacity. In addition, we need to understand more broadly the influence of diverse human impacts. Much attention has focused on land use and climate change, but future work should consider the simultaneous role of additional threats and importantly how threats interact with each other and with natural processes (and species characteristics). Finally, an important gap in our understanding is the importance of scale. Some factors and processes may be important at large scales but others may prevail within local areas. Although range data present great potential to address ecological and conservation questions, there are also concerns regarding how information is presented, particularly regarding the spatial resolution used. Most range maps are drawn from presence data, but these data may be available at different spatial scales for different areas and/or species. This variation may reflect true differences among areas/species and/or different methodologies and efforts. Nevertheless, general databases (like the IUCN) do not provide this information to the potential user. Ideally, maps should be accompanied by the original presence data with details on methodology so that analyses can take any differences into account. Creative use of available historical sources also needs to be promoted, even if caution is necessary to interpret this information (Clavero and Revilla 2014).

Synthesis and review: Species ranges, natural processes, human interactions and gaps/opportunities

To conclude we would like to propose that to move the field forward we need to: (1) develop and test new theoretical models; (2) conduct controlled experiments in laboratory or semi-natural conditions; (3) ensure that spatial range data are accompanied by informative metadata; and (4) develop new methodological and conceptual approaches to compare data across temporal and spatial scales.

Chapter 1: Toward multifactorial null models of range contraction in terrestrial vertebrates

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Lucas, P.M; González-Suárez, M; Revilla, E. (2016) Toward multifactorial null models of range contraction in terrestrial vertebrates. Ecography. doi: 10.1111/ecog.01819.

1. Toward multifactorial null models of range contraction in terrestrial vertebrates Abstract

The contraction of a species’ distribution range, which results from the extirpation of local populations, generally precedes its extinction. Therefore, understanding drivers of range contraction is important for conservation and management. Although there are many processes that can potentially lead to local extirpation and range contraction, three main null models have been proposed: demographic, contagion, and refuge. The first two models postulate that the probability of local extirpation for a given area depends on its relative position within the range; but these models generate distinct spatial predictions because they assume either a ubiquitous (demographic) or a clinal (contagion) distribution of threats. The third model (refuge) postulates that extirpations are determined by the intensity of human impacts, leading to heterogeneous spatial predictions potentially compatible with those made by the other two null models. A few previous studies have explored the generality of some of these null models, but we present here the first comprehensive evaluation of all three models. Using descriptive indices and regression analyses we contrast the predictions made by each of the null models using empirical spatial data describing range contraction in 386 terrestrial vertebrates (mammals, birds, amphibians, and reptiles) distributed across the World. Observed contraction patterns do not consistently conform to the predictions of any of the three models, suggesting that these may not be adequate null models to evaluate range contraction dynamics among terrestrial vertebrates. Instead, our results support alternative null models that account for both relative position and intensity of human impacts. These new models provide a better multifactorial baseline to describe range contraction patterns in vertebrates. This general baseline can be used to explore how additional factors influence contraction, and ultimately extinction for particular areas or species as well as to predict future changes in light of current and new threats.

Keywords: Border, extinction, habitat loss, historical range, human, land use, range dynamics.

1. Toward multifactorial null models of range contraction in terrestrial vertebrates Introduction

Species extinctions generally start with the vanishing of particular populations that continue until no populations remain (Yackulic et al. 2011). In other words, complete extinction is usually preceded by a contraction of the distribution range that results from the extirpation of local populations. Local extirpations and contractions are considered good descriptors of biological capital loss, possibly even preferable to quantifying extinction itself (Ceballos and Ehrlich 2002, Davis et al. 1998). Therefore, understanding the general dynamics of range contraction is key for effective conservation (Safi and Pettorelli 2010). The list of proximate and ultimate causes of local extinction is long, and taxon-dependent (Cahill et al. 2012, González-Suárez and Revilla 2014); thus, we may expect a wide variety of range contraction patterns. Nevertheless, ecologists and conservation biologists have used null models or simple hypotheses to describe the expected spatial patterns of local extinction and range contraction, especially when detailed information is not available. Null models are representations based on the simplest and most general mechanisms, and deliberately focus on a few key factors or processes to provide a baseline for comparison with empirical observations or with more complex models (Gotelli 2001). The simplicity of null models can be useful for species for which little information exists, as well as in theoretical studies (Hanski 1998, Hanski and Ovaskainen 2000). Generalized patterns of distribution range contraction have been described in the literature using three different null models: demographic, contagion, and refuge. These models describe contraction based on distinct mechanisms derived from theoretical principles in ecology, biogeography, and conservation biology (Hanski 1998, Hemerik et al. 2006); and have been used in empirical studies as baselines to determine the role of additional factors or to broadly describe observed contraction patterns (Franco et al. 2006, Parmesan 1996, Pomara et al. 2014, Thomas et al. 2004, Turvey et al. 2015, Yackulic et al. 2011). The demographic null model derives from basic population dynamic principles, and from the ecological assumption which postulate that environmental conditions and resources at the center of a distribution range are more suitable than at the border, resulting in higher population growth rates and thus, higher abundance in central areas (Brown 1995, Lawton 1993). Because extinction is directly determined by population abundance (Brown 1971, David et al. 2003, Jones and Diamond 1976, Pimm et al. 1988), when the drivers of

1. Toward multifactorial null models of range contraction in terrestrial vertebrates extinction (threats) are ubiquitous, central areas would have lower extinction/extirpation risk (MacArthur and Wilson 1967). Assuming threats are indeed ubiquitous, this null model then predicts that populations would be first extirpated along the historical border (where density is lower) and would continue toward the center, where the last (most dense) population would be found (Fig. 1). The contagion null model, on the other hand, assumes that the treats have clinal distribution, with threats spreading across the landscape with distinct directionality, like a contagious disease (Channell and Lomolino 2000a, b, Lawton 1993). Based on this clinal threat pattern, the contagion null model predicts that populations would be first extirpated in the historical border closest to the extinction driver’s origin, and then as the threat spreads across the range, the central areas would become extirpated until only the historical border located farthest from the initial point remains (Fig. 1). Finally, the refuge model assumes that more humanized land uses are associated with higher risk of extinction (Ceballos and Ehrlich 2002, Fisher 2011, Hoffmann et al. 2010, Laliberte and Ripple 2004, Li et al. 2015, Pomara et al. 2014, Schipper et al. 2008, Yackulic et al. 2011), and predicts that populations would be first extirpated in areas that are more modified and heavily used by humans. According to this model, the last population will be located in the least used area, which represents a final refuge for the species (Fig. 1). Some of the assumptions and the predictions of primarily the demographic and contagion models have been tested by previous studies, which collectively suggest these models may not be broadly applicable (Ceballos and Ehrlich 2002, Fisher 2011, Hemerik et al. 2006, Laliberte and Ripple 2004, Sagarin and Gaines 2002, Thomas et al. 2008, Yackulic et al. 2011). However, there has been no comprehensive evaluation of all three null models; partly because spatial data quantifying range contraction at the global scale are limited, but also because there are important methodological challenges including the difficulties in defining a unique center and a relative position within a species range. In this study we overcome these challenges to simultaneously evaluate these three null models using a global dataset for 386 terrestrial vertebrates (mammals, birds, amphibians and reptiles). We first identify the key predictions derived from each null model and then, using descriptive indices and regression analyses, we evaluate if empirical range contraction data conform to the models predictions. Our goals are: 1) to determine which, if any, of the proposed null models represents the most adequate general baseline to explain range contractions; 2) if necessary, to propose and evaluate alternative multifactorial null models; and 3) to provide a more 41

1. Toward multifactorial null models of range contraction in terrestrial vertebrates consistent framework regarding the general underlying causes of range contraction dynamics among terrestrial vertebrates.

Methods

Spatial distribution data We used global distribution data of 386 terrestrial vertebrates (International Union for Conservation of Nature 2010a) with known range contraction (i.e., a distribution with extirpated areas, where the species was present in the past but is no longer found, and current areas, where the species is currently present, and following the notation of the International Union for Conservation of Nature 2010; detailed information is provided in Appendix A1.1 in Supplementary Material). Since most species distributions are fragmented and have complex shapes, our analyses were conducted at two different scales. At the range scale, we used data from the complete historical distribution range of each species (N=374), which often included multiple fragments separated by unoccupied areas. At the fragment scale, we used data from all individual fragments with observed contraction (N=273. See Appendix 1.2 for additional information in data preparation). Supplementary material Appendix A1.2, Table A1.3 and A1.4, and Fig. A1.1 provide descriptive summaries of these data including total area in km2 and percentage of contraction (calculated as the percentage of the historical range area classified as extirpated) for complete ranges and individuals fragments. For complete ranges we also summarize the number of fragments present in the historical, extirpated, and current ranges, as well as the percentage of extirpated fragments (percentage of historical fragments classified as extirpated). Spatial data were projected into an equal area projection (Cylindrical Equal Area) and rasterized.

Analyses We followed a two-step approach to evaluate the key predictions of each null model (Fig. 1.1). First, we defined three indices to visually explore the support of model predictions by the empirical data. Second, we defined and compared three regression models that estimate the probability of extirpation based on the key model predictions, thus providing a quantitative test of support for each null model.

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Figure 1.1. Assumptions and predicted range contraction patterns for each of the three null models. The demographic null model assumes higher density in the center of the range and a ubiquitous threat pattern. As a result, contractions are predicted to occur toward the core in multiple directions. The contagion null model assumes that threats are distributed in a cline resulting in a directional contraction along this cline. The refuge null model assumes that the extirpation is determined by human land use and predicts a heterogeneous range contraction pattern with less used areas being less likely to become extirpated.

Indexes The demographic and contagion null models both associate the probability of extirpation with an area’s relative position within a range (Fig. 1.1). Therefore, we defined a position index based on relative distance to the border. We use the border instead of the center because identifying meaningful centers is complicated in complexly shaped and fragmented distributions (Sagarin et al. 2006). For each distribution range and fragment analyzed, we first estimated the geodetic distance from each grid cell to the closest historical border cell (Fig. 1.2, and Appendix A1.2). A geodetic distance is the distance between two unprojected points

1. Toward multifactorial null models of range contraction in terrestrial vertebrates on the spheroid of the Earth (using the spheroid World Geodetic System 1984, WGS84). Distances were standardized dividing species’ values by the

Figure 1.2. Examples of the three variables defined to represent the key predictors of the three null models: Distance to the Border (Border): average distance to border from each cell; Angle of contraction: geodetic angle of contraction (from each extirpated cell to the closest current cell), and Human Use (Land use): proportion of human use in the cell. Examples represent the Saint Lucia amazon (Amazona versicolor) which illustrates the pattern of contraction predicted by the demographic null model (also partly congruent with the refuge null model); the La Palma giant lizard (Gallotia auaritae) illustrates contraction from a border to the opposite border in a unique direction as predicted by the contagion null model (and is also partly congruent with the refuge null model); and the blue duck (Hymenolaimus malacorhynchos) which adjusts to the refuge null model prediction. maximum distance observed for the range (at range scale) or fragment (at fragment scale) to facilitate comparison among species with different distribution ranges. Using these distance values from each cell to the nearest border, we then calculated the variable Border as the arithmetic mean distance to the border from all cells within one area, with Border_ext representing extirpated areas and Border_curr current areas. Using these values we defined

1. Toward multifactorial null models of range contraction in terrestrial vertebrates the Centrality Index = Border_ext/Border_curr for each range and fragment. The demographic null model predicts Centrality Index < 1 (extirpated areas are closer to the border), whereas the contagion model predicts Centrality Index<1 only for initial initial stages of contraction (approximately <50% of the historical range extirpated), and Centrality Index > 1 for contractions >50%. Therefore, both the contagion and demographic null models predict the same values of Centrality Index in early stages of contraction but different values in later stages. The refuge null model makes no general prediction for the Centrality Index (Fig. 1.1). The second prediction made by the demographic and contagion null models relates to the directionality in contraction. The demographic null model predicts that contraction occurs in multiple directions, while the contagion null model states that contraction occurs along a unique general direction that can be detected as a predominant contraction angle (Fig. 1.1). We calculated the geodetic angle of contraction for each extirpated cell as the azimuth of the direction defined by the vector joining each extirpated grid cell with its closest current cell (Fig. 1.2 and Appendix A1.2). Using all angles of contraction for each distribution (complete range or individual fragment) we calculated the Directionality Index as the angular concentration. Directionality Index ranges from 0 to 1 and is the inverse of the dispersion of the angles (Zar 1999). The demographic null model predicts Directionality Index values close to 0 (high angle dispersion) and the contagion null model predicts values close to 1 (a low angle dispersion). The refuge model makes no prediction for the Directionality Index (Fig. 1.1). The last index we defined captures the predictions of the refuge model (Fig. 1.1). Although human land use has changed over time and past uses likely influenced observed contraction, data are not available at a global scale to describe past land use. Therefore, we defined land use based on the 1-km resolution MODIS (MCD12Q1) Land Cover Product (Oak Ridge National Laboratory Distributed Active Archive Center 2010). We determined the extent of land classified as covered/used (henceforward used) by humans for each range or fragment (Appendix A1.2 and Table A1.5). From these cell values we then calculated the variables Land use_ext as the proportion of cells used by humans in the extirpated area, and Land use_curr as the proportion of cells used by humans in the current area. Using these variables, we defined a Land use Index which is calculated as Land use_ext/ Land use_curr. If extirpated areas have a greater proportion of human use, then Land use Index > 1 as 45

1. Toward multifactorial null models of range contraction in terrestrial vertebrates predicted by the refuge null model. The contagion and demographic null models make no specific predictions regarding the Land use Index. We calculated and investigated the distribution of these three indices for terrestrial vertebrates. Prior to visualizing the empirical data the behavior of the Centrality and Directionality indexes was evaluated using simulated scenarios. We sketched three example distribution range areas (Fig. A1.3) for which we simulated two patterns: range contraction towards the center (demographic model), and clinal range contraction (contagion model). For irregularly shaped distributions we explored two different directions of contraction because distinct clines could influence results. The indexes were then validated exploring the behavior of values calculated at seven stages along the contraction process in these simulated scenarios (Fig. A1.3).

Regression analyses We defined regression models to estimate the probability of extirpation of an area based on two of the previously defined variables (Border and Land use) and the percentage of contraction (Contraction). For this approach we excluded distributions (ranges and fragments) with <10% or >90% contraction (Appendix A1.1, Tables A1.1 and A1.2) because at early and late stages of contraction stochastic noise may confound existing patterns (Yackulic et al. 2011). Under the demographic model, the probability of extirpation should continuously decrease with the distance to the border independently of the percentage of contraction. Thus, the probability of extirpation of an area could be simply defined by the variable Border (Mod_Demographic, Table 1.1). A key prediction of the contagion null model is that there is directionality in contraction, but the angle of contraction is a relative concept that compares extirpated and current areas and thus, cannot be estimated for completely extirpated or current areas. Instead, we evaluated another prediction of this null model, namely that the effect of distance to the border on the probability of extirpation depends on the percentage of contraction. We modeled this prediction using an interaction term between the variables Border and Contraction (Mod_Contagion, Table 1.1). Finally, under the refuge null model, the probability of extirpation should simply depend on the human land use intensity, which is represented by the variable Land use (Mod_Refuge, Table 1.1). For each of the analysis scales (range and fragment) we fitted generalized linear mixed regression models (GLMM) with family binomial and a logit link using the function glmer 46

1. Toward multifactorial null models of range contraction in terrestrial vertebrates from the lme4 package in R (R Development Core Team 2013). All models included taxonomic class, order, family, and genus as random factors to control for evolutionary non- independence of the observations. We compared models using an information theoretic approach based on Akaike Information Criterion, AIC (Burnham and Anderson 2002). Finally, we explored the possibility that the multiple processes postulated by these null models may occur simultaneously. We fitted two additional models that combine predictions from compatible null models. Combined_1 modelled the probability of extirpation considering both Land use and Border, Combined_2 included Land use and allowed for the interaction of Border with Contraction (Table 1.1).

Results

We analyzed spatial data for 386 species (374 species at range scale and 213 at fragment scale) which represent ~1.6% of the terrestrial vertebrates listed by the IUCN. The studied distribution ranges and fragments have widely variable areas, with an observed mean percentage of contraction of 41% for complete ranges and 51% for fragments (Appendix A1.2, Tables A1.3 and A1.4 and Fig. A1.1). Distribution ranges are often fragmented with a mean of 6.7 fragments per historical range. Validation of the indexes showed that as expected, when contraction was simulated following the demographic model, Centrality Index values decreased and Directionality Index values were generally close to 0 (although for irregular shapes values showed a small increase at high contraction stages). When contraction was simulated following a cline (as proposed by the contagion model), we detected the predicted shift in the Centrality Index and values for the Directionality Index generally close to 1. Empirical estimates of the three indices did not identify a single best-supported null model at the range or fragment scale (Fig. 1.3). Centrality Index values show a tendency to change with the percentage of contraction as predicted by the contagion null model. However, Directionality Index values show no support for either the contagion or demographic models. The Land use Index suggests extirpation has been more likely in humanized areas as predicted by the refuge null model (median values are consistently above

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Figure 1.3. The distribution of three indices at the range (a, c, e) and fragment scale (b, d, f). For initial stages of contraction (< 50% contraction) both demographic and contagion null model predict Centrality Index < 1. For higher stages of contraction (> 50% contraction) Centrality Index < 1 supports the demographic null model while Centrality Index > 1 supports the contagion null model (a, b). Directionality Index close to 0 is predicted by the demographic null model, whereas values close to 1 support the contagion null model (c, d). Land use Index > 1 is predicted by the refuge null model (e, f). Ends of the whiskers represent the lowest datum still within the 1.5 interquartile range (IQR) of the lower quartile, and the highest datum still within the 1.5 IQR of the upper quartile (Tukey boxplot).

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

1; Fig. 1.3). However, in many cases current areas are more humanized than those extirpated. Results were broadly consistent among taxonomic classes (Appendix A1.2, Fig. A1.4). Results from the regression analyses at both scales also failed to clearly identify a single best null model. At the range scale, both the refuge (Mod_Refuge) and the contagion (Mod_Contagion) null models received support; whereas at the fragment scale the only supported model was Mod_Refuge (Table 1.1). Although overall the refuge null model received greater support compared to other null models, results at both range and fragment scales revealed that either of the combined models represents a great improvement (based on AIC) over models based on the unifactorial null models (Table 1.1). At least for the available data, multiple processes appear to best explain the general patterns of contraction among terrestrial vertebrates. At the range scale Combined_2 was the only supported model (Table 1.1), which describes the probability of extirpation as positively correlated with human use (Land use) and identifies a contraction-dependent effect of distance to the border. In particular, at early stages of contraction (up to ~60% contraction, obtained when the ∂Probability of Extirpation/∂Border is equal to zero) areas near the border are more likely to be extirpated whereas at later stages the pattern is reversed (Fig. 1.4a). At the fragment scale, both combined models were supported (being within 2 AIC units of each other, Table 1.1) and show a positive association between the probability of extirpation and Land use, with the best supported model, Combined_2, additionally supports an interaction between Border with Contraction with extirpation being generally more likely near the border, but with a weakening effect as contraction advances. In this model, extirpation only becomes more likely near the center outside the range of data values used to fit the model (approximately >98% contraction, obtained when the ∂Probability of Extirpation/∂Border is equal to zero. Data used to fit the models exclude fragments with <10% or >90% contraction). The simpler supported model (Combined_1) does not include an interaction term and suggests that extirpation is consistently more likely near the border (Figs. 1.4b and 1.4c). Thus, at the fragment scale, and considering both supported models we interpret the results as that in the early stages of contraction areas close to the border have higher probability of extirpation than central areas. However, this difference between border and central areas may weaken as contraction

2. Toward multifactorial null models of range contraction in terrestrial vertebrates

1 Table 1.1. Results from the regression analyses based on regression models (GLMM) to evaluate the three main null models of range contraction 2 (demographic, contagion and refuge) and two combined models that incorporate multiple processes. Combined_1 proposes that the probability of 3 extirpation of an area is determined by the proportion of human use in the area (variable Land use) and the distance to the historical border 4 (variable Border). Combined_2 proposes that the probability of extirpation depends on Land use and the interaction of Border and Contraction 5 (reflecting the expectation that as range contraction progresses the risk associated with being near the border changes). All models were fitted at 6 two scales: complete historical range and historical fragment. We report model coefficients (best estimates and their SE), AIC, ΔAIC (difference in

7 AIC with the best model comparing all five models), and ΔAICsm (difference in AIC comparing only the three models derived from the main 8 proposed null models). Dashes indicate variables not included in the model.

Model Coefficients Model comparison

Land use Border Contraction Border*Contraction AIC ΔAIC ΔAICsm

Range scale (N=457, 229 species) Combined_2 2.13 (0.466)* -9.74 (2.145)* -2.66 (0.688)* 15.86 (3.699)* 605.21 0.00 Combined_1 2.03 (0.443)* -1.78 (0.919) † - - 621.33 16.13 Mod_Refuge 2.02 (0.441)* - - - 623.15 17.94 0.00 Mod_Contagion - -9.81 (2.110)* -2.23 (0.664)* 15.74 (3.650)* 625.49 20.28 2.34 Mod_Demographic - -1.74 (0.887) † - - 641.64 36.43 18.49 Fragment scale (N=362, 142 species) Combined_2 2.73 (0.541)* -9.15 (2.497)* -2.03 (0.977)* 9.35 (4.131)* 468.09 0.00 Combined_1 2.62 (0.527)* -4.16 (1.008)* - - 469.35 1.26 Mod_Refuge 2.57 (0.514)* - - - 486.24 18.14 0.00 Mod_Contagion - -8.30 (2.430)* -1.22 (0.927) 7.65 (3.975) † 494.30 26.21 8.06 Mod_Demographic - -3.98 (0.952)* - - 494.72 26.62 8.48 9 *P < 0.05; †P < 0.10 10 11 12 13 50

1. Toward multifactorial null models of range contraction in terrestrial vertebrates progresses. Separate analyses for data rasterized at different resolutions offered results consistent with these analyses (Appendix A1.2, Table A1.9)

Figure 1.4. Predictions of the supported regression models explaining probability of extirpation of an area as a function of its distance to the historical border (Border) and its human land use (Land use) with a possible interaction of Land use and the percentage of contraction (Contraction). At the range scale, panel (a), Model Combined_2 (including the interaction) was the single supported model (Table 1). At the fragment scale both Model Combined_2 (b) and Model Combined_1 (c, no interaction) were supported. To visualize the effect of the interaction between Border and Contraction (a, b), we represent predictions at three levels of contraction: 20% in darker grey, 50% in medium dark grey, and 80% in light grey. 51

1. Toward multifactorial null models of range contraction in terrestrial vertebrates Discussion

The three main null models of range contraction proposed to date make diverse predictions derived from their theoretical underpinnings. Our evaluation using global spatial data for terrestrial vertebrates reveals that none of these null models is sufficiently general to describe contraction range patterns. Even though in the majority of species extirpated areas are more likely to be heavily humanized, as predicted by the refuge null model, we also find support for models that incorporate two distinct mechanisms that likely act together. In addition, the relative position within a range also appears to influence extirpation probability (independently of human use). For many of the studied species, extirpation is more likely near the border during early stages of contraction but during the final stages of contraction extirpation becomes more likely in central areas, as proposed by the contagion null model. Yet, we also find support for the demographic model which postulates that the probability of extirpation is always higher near the border. Future research focused on the final stages of contraction would be necessary to disentangle these patterns. Nevertheless, our results show that contraction is better described by multi-process models that consider both human impacts and relative position, than by the three originally-proposed null models.

Contraction and human land use We find that human use is probably the best single predictor of extirpation probability, as previously suggested by Yackulic et al. (2011). The key role of human land use changes in species extinction has been proposed by previous studies that identified habitat loss due to human land use as the main threat for diverse vertebrate groups (González-Suárez and Revilla 2014, Hayward 2011, Pekin and Pijanowski 2012, Schipper et al. 2008). In our study, we find that indeed greater extirpation risk is generally associated with more humanized areas. However, a correlation between human use and extirpation does not imply a direct causal relationship. Other factors, such as the presence of invasive species or climate change, could be spatially correlated with human uses leading to similar patterns of contraction (Franco et al. 2006, Thomas et al. 2006). The potential role of these other factors could be explored considering our new proposed baseline that accounts for relative position and human impacts.

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Although extirpations are generally more common in humanized areas, some species persist within these regions. Distinct patterns may be due to intrinsic responses; some species are less sensitive to human impacts than others (Maklakov et al. 2011), and some even benefit from humanized conditions (Maclean et al. 2011). Additionally, extirpation may be determined by other drivers of extinction with different spatial configurations (Clavero et al. 2009, González-Suárez et al. 2013, González-Suárez and Revilla 2014, Thomas et al. 2006). A caveat of our approach is that our data reflect only current human land uses, which may not correspond to the past uses potentially responsible for observed extirpations (Carvalheiro et al. 2013, Plieninger et al. 2006). It is not clear to us, however, how this could bias our results since we analyzed a large number of species at a global scale, and the progress of land use changes has been heterogeneous across the world. While land uses often intensify with time, the rates of intensification vary by area, and may affect species differently (Bregman et al. 2014, Gilroy et al. 2014). For example, in some areas of Europe and North America there has been a reversal toward more natural uses as agricultural land has been abandoned, but this reversal has not occurred in other areas (Gellrich et al. 2007, MacDonald et al. 2000, Mottet et al. 2006, Strijker 2005). Future studies would be necessary to address the temporal aspect of land use changes; however, human activities and land use are still likely to be key factors driving range contraction. In fact, they may well play an even more important and complex role than identified here, e.g., areas with intense agricultural uses have a greater impact that agri-environmental management areas (Carvalheiro et al. 2013, Franco et al. 2006).

Contraction and relative position within the range: different patterns at different scales In addition to the importance of human land use, our analyses show that the relative position of an area also influences its probability of extirpation (Brown 1995, Channell and Lomolino 2000a, b, Lawton 1993). At the range scale our results indicate that the probability of extirpation near the border (or the center) depends on the contraction stage. This pattern can be caused by directional threats as proposed by Channell & Lomolino (2000a, b). For example, climate change can create latitudinal and altitudinal clines (Parmesan 1996, Parmesan and Yohe 2003). However, there are alternative mechanisms that can also lead to this observed pattern. Climatic and biotic factors generally define range limits (Araújo and Rozenfeld 2014), but some boundaries are due to abrupt ecosystem changes or physical barriers, such as mountain chains or the transition from land to ocean. In these cases, border 53

1. Toward multifactorial null models of range contraction in terrestrial vertebrates areas may actually represent optimal habitat and thus, be the most populated (Caughley et al. 1988, Gaston 2003, Sagarin and Gaines 2002). When optimal habitat occurs in a range border, a directional pattern of contraction could simply occur due to intrinsic population dynamics, as less dense populations are more likely to go extinct. At the fragment scale we found support for two apparently contrasting models. The simplest model predicts that the probability of extirpation is always higher near the border, while the best model suggests that the probability of extirpation near the border depends on the contraction level. However, the predicted shift from higher extirpation risk near the border to higher near the center occurs at the very final stages of contraction (which lay beyond the range of values analyzed, >90% contraction). In comparison, at the range scale this shift is predicted at ~60% contraction. Therefore, we interpret these results as supporting a higher probability of extirpation near fragment borders in early stages with a potential weakening of this effect as contraction progresses. There are various possible reasons that could explain the discrepancy in the results between range and fragment scales. First, different factors and process influence dynamics at different scales, e.g., climate acts at broader scale while biotic interactions are more relevant locally (Araújo and Rozenfeld 2014, Pearson and Dawson 2003, Whittaker et al. 2001). Second, the meaning and identification of relative positions in complexly shaped distributions is complicated and this may confound results. For example, the border area in a fragment located near other fragments has a greater probability of receiving migrants than a “true border”, and thus, could have a lower probability of extirpation. Null models are commonly defined based on idealized distributions that largely fail to represent reality. Most species distributions are complex, often formed by multiple fragments with different shapes that change over time (Gaston 2003, Wilson et al. 2004). To study range dynamics we need to embrace this complexity, considering all types of ranges and not only those that conform to some theoretical or idealized depictions. Importantly, as shown here, we must evaluate predictions at different scales because results and inferences may differ (Thomas et. al. 2008).

A new baseline to understand range contraction: multifactorial null models Earlier null models of range contraction have focused on single processes –basic population rules and simple threat dynamics (Brown 1995, Brown and Kodricbrown 1977, Channell and 54

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Lomolino 2000a, b, Lawton 1993). Here we show that these null models are not adequate baselines, at least for terrestrial vertebrates. Species persistence may be influenced by multiple external threats and intrinsic processes (González-Suárez et al. 2013, Yackulic et al. 2011). To partly account for this complexity, Yackulic et al. (2011) proposed multifactorial models (including biome, human impacts, and relative position) to explain range contraction in large mammals. Here, we generalized the importance of multifactorial models for a wide range of terrestrial vertebrates. Understanding range contraction is important for conservation and management, particularly if we hope to accurately predict future range changes and assess the effects of new threats (Newbold et al. 2014, Peters et al. 2014, Selwood et al. 2014, Stanton et al. 2014, Thomas et al. 2004, Thomas et al. 2011). Our global study based on data from four different groups of vertebrates reveals the need to develop more realistic null models to use as baselines. Without departing from the objective of simplicity, we propose to combine simple key elements already identified as relevant to define new multi-process null models of range contraction. We realize that data at this scale could have their own limitations, but we feel that these models can offer a more realistic baseline to evaluate the role of additional factors, such as the effect of different types of range borders, the role of environmental conditions, additional human and natural threats, as well as how intrinsic species’ traits influence contraction range dynamics.

Acknowledgments

We are sincerely grateful to Kevin Gaston, Thomas Wilson, Miguel Delibes, Miguel Ángel Olalla, the members of the Spatial Ecology Lab (University of Queensland), the members of the Department of Conservation Biology (EBD-CSIC) and the personal from the Geographic Information System Laboratory (LAST-EBD-CSIC) for their help with technical aspects and helpful suggestions about the manuscript. We are also in debt to Ángel Lucas for the artwork in Figure 2. We also thank two anonymous reviewers for helpful comments that improved earlier drafts of the article. This work was funded by the Spanish Ministry of Economy and Competitiveness (CGL2009-07301⁄BOS and CGL2012-35931/BOS co-funded by FEDER, and the FPI grant BES-2010-034151), by the European Community’s Seventh Framework Programme (FP7⁄2007-2013) under grant agreement no 235897, and by a Juan de la Cierva post-doctoral fellowship (JCI-2011-09158). We also acknowledge funding from the Spanish Severo Ochoa Program (SEV-2012-0262). 55

1. Toward multifactorial null models of range contraction in terrestrial vertebrates Supplementary material

Appendix 1.1

Table A1.1 Description of data for all terrestrial vertebrates species with range contraction (N=628) obtained from the IUCN including: scientific name (Species), taxonomic class (Class), the group in which it was rasterized (Raster group), the area of its historical range (Historical range km2), the number of historical cells (Historical cells), the number of current cells (Current cells), the percentage of contraction of the range (Contraction), if the species was included for analysis (Included), and the Reason for Exclusion. Exclusion was based on number of historical cells (Historical), number of current cells (Current), number of extirpated cells (Extirpated) or on topological errors defined as the overlap between current and extirpated areas (Topology). Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Atelopus chiriquiensis Amphibian R5km 4450 178 63 115 64.61 Yes - Atelopus varius Amphibian R10km 21600 216 61 155 71.00 Yes - Bolitoglossa robusta Amphibian R10km 6800 68 60 8 11.76 Yes - Craugastor ranoides Amphibian R10km 40200 402 127 275 68.41 Yes - Craugastor rhyacobatrachus Amphibian R5km 1325 53 14 39 73.58 Yes - Duellmanohyla uranochroa Amphibian R10km 14800 148 7 141 95.27 Yes - Eleutherodactylus martinicensis Amphibian R5km 4600 184 158 26 14.13 Yes - Eleutherodactylus schwartzi Amphibian R1km 146 146 89 57 39.04 Yes - Hyloscirtus colymba Amphibian R10km 16500 165 146 19 11.52 Yes - Isthmohyla angustilineata Amphibian R5km 1500 60 38 22 36.67 Yes - Isthmohyla calypsa Amphibian R1km 318 318 291 27 8.49 Yes - Lithobates tarahumarae Amphibian R10km 97000 970 952 18 1.86 Yes - Lithobates vibicarius Amphibian R5km 2650 106 43 63 59.43 Yes - Neurergus microspilotus Amphibian R1km 576 576 394 182 31.60 Yes - Ommatotriton vittatus Amphibian R10km 53100 531 524 7 1.32 Yes - Pristimantis caryophyllaceus Amphibian R10km 51600 516 448 68 13.18 Yes - 56

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Rana muscosa Amphibian R10km 16000 160 10 150 93.75 Yes - Rana sierrae Amphibian R10km 29000 290 17 273 94.14 Yes - Strabomantis bufoniformis Amphibian R10km 58000 580 569 11 1.90 Yes - Accipiter butleri Bird R1km 279 279 158 121 43.37 Yes - Accipiter gundlachi Bird R10km 111800 1118 128 990 88.55 Yes - Aceros waldeni Bird R10km 23400 234 110 124 52.99 Yes - Acrocephalus aequinoctialis Bird R1km 508 508 465 43 8.00 Yes - Acrocephalus caffer Bird R1km 655 655 270 385 58.78 Yes - Acrocephalus luscinius Bird R1km 706 706 143 563 79.75 Yes - Aethopyga duyvenbodei Bird R1km 673 673 557 116 17.00 Yes - Agapornis nigrigenis Bird R10km 19300 193 164 29 15.03 Yes - Agriornis albicauda Bird R50km 675000 270 244 26 9.00 Yes - Alectrurus risora Bird R50km 805000 322 38 284 88.20 Yes - Alectrurus tricolor Bird R50km 995000 398 251 147 36.93 Yes - Amazona agilis Bird R10km 8500 85 26 59 69.41 Yes - Amazona arausiaca Bird R1km 771 771 74 697 90.40 Yes - Amazona imperialis Bird R1km 220 220 53 167 75.91 Yes - Amazona leucocephala Bird R10km 125800 1258 173 1085 86.25 Yes - Amazona oratrix Bird R50km 167500 67 19 48 71.64 Yes - Amazona pretrei Bird R10km 26100 261 104 157 60.15 Yes - Amazona rhodocorytha Bird R10km 27200 272 25 247 90.81 Yes - Amazona ventralis Bird R10km 8700 87 75 12 13.79 Yes - Amazona versicolor Bird R1km 615 615 138 477 77.56 Yes - Amazona vinacea Bird R50km 157500 63 44 19 30.16 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Amazona viridigenalis Bird R10km 68200 682 77 605 88.71 Yes - Anairetes alpinus Bird R10km 15200 152 122 30 19.00 Yes - Anas strepera Bird R50km 34190000 13676 13627 49 0.36 Yes - Anodorhynchus hyacinthinus Bird R50km 672500 269 215 54 20.07 Yes - Anthocephala floriceps Bird R10km 10600 106 100 6 5.66 Yes - Anthracoceros montani Bird R5km 1475 59 21 38 64.41 Yes - Anthus nattereri Bird R10km 94500 945 707 238 25.19 Yes - Anthus spragueii Bird R50km 4027500 1611 1529 82 5.09 Yes - Apalharpactes reinwardtii Bird R10km 6300 63 11 52 82.54 Yes - Aphelocoma coerulescens Bird R10km 64600 646 64 582 90.09 Yes - Apteryx haastii Bird R10km 39900 399 90 309 77.44 Yes - Ara ambiguus Bird R10km 103400 1034 991 43 4.16 Yes - Ara macao Bird R50km 7022500 2809 2686 123 4.38 Yes - Ara militaris Bird R50km 365000 146 110 36 24.66 Yes - Aramides wolfi Bird R1km 143 143 137 6 4.20 Yes - Aramus guarauna Bird R50km 12207500 4883 4872 11 0.23 Yes - Aratinga chloroptera Bird R10km 67000 670 190 480 71.64 Yes - Aratinga euops Bird R10km 110900 1109 118 991 89.36 Yes - Ardeotis nigriceps Bird R50km 1117500 447 189 258 57.72 Yes - Artamus mentalis Bird R10km 16000 160 150 10 6.25 Yes - Asthenes heterura Bird R10km 43600 436 402 34 7.80 Yes - Atlapetes flaviceps Bird R1km 302 302 228 74 24.50 Yes - Atlapetes fuscoolivaceus Bird R5km 2825 113 44 69 61.06 Yes - Atrichornis rufescens Bird R10km 61300 613 171 442 72.00 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Aythya innotata Bird R5km 1825 73 26 47 64.38 Yes - Bangsia aureocincta Bird R1km 911 911 664 247 27.11 Yes - Bangsia melanochlamys Bird R10km 8900 89 42 47 52.81 Yes - Basileuterus basilicus Bird R1km 875 875 699 176 20.11 Yes - Basileuterus conspicillatus Bird R10km 5700 57 32 25 43.86 Yes - Biatas nigropectus Bird R10km 28600 286 189 97 33.92 Yes - Bolborhynchus ferrugineifrons Bird R5km 3500 140 125 15 10.71 Yes - Branta sandvicensis Bird R10km 7400 74 23 51 68.92 Yes - Brotogeris pyrrhoptera Bird R10km 13200 132 93 39 29.55 Yes - Bucco noanamae Bird R10km 31300 313 296 17 5.43 Yes - Cacatua haematuropygia Bird R50km 217500 87 9 78 89.66 Yes - Callaeas cinereus Bird R10km 99700 997 61 936 93.88 Yes - Calyptophilus frugivorus Bird R10km 14400 144 72 72 50.00 Yes - Capito hypoleucus Bird R5km 4250 170 150 20 11.76 Yes - Carduelis cucullata Bird R10km 103200 1032 117 915 88.66 Yes - Carduelis siemiradzkii Bird R10km 24500 245 183 62 25.31 Yes - Carduelis yarrellii Bird R10km 40200 402 9 393 97.76 Yes - Carpornis melanocephala Bird R10km 68300 683 190 493 72.00 Yes - Chaetocercus berlepschi Bird R1km 1113 1113 1005 108 9.00 Yes - Chaetornis striata Bird R50km 1132500 453 240 213 47.02 Yes - Charitospiza eucosma Bird R10km 57000 570 520 50 8.77 Yes - Chasiempis ibidis Bird R5km 1575 63 7 56 88.89 Yes - Chasiempis sandwichensis Bird R10km 9500 95 58 37 38.00 Yes - Chasiempis sclateri Bird R5km 1325 53 15 38 71.70 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Chondrohierax wilsonii Bird R10km 107800 1078 38 1040 96.47 Yes - Cinclodes aricomae Bird R10km 54300 543 28 515 94.84 Yes - Cinclus schulzi Bird R10km 25400 254 235 19 7.48 Yes - Cisticola eximius Bird R50km 687500 275 268 7 2.00 Yes - Cistothorus apolinari Bird R5km 2900 116 62 54 46.55 Yes - Claravis godefrida Bird R10km 60100 601 48 553 92.01 Yes - Clytorhynchus vitiensis Bird R10km 16900 169 163 6 3.55 Yes - Coccyzus rufigularis Bird R10km 17000 170 7 163 95.88 Yes - Coeligena prunellei Bird R10km 8300 83 68 15 18.07 Yes - Colaptes fernandinae Bird R10km 106700 1067 74 993 93.06 Yes - Collocalia bartschi Bird R1km 938 938 740 198 21.11 Yes - Collocalia leucophaea Bird R1km 219 219 109 110 50.23 Yes - Columba thomensis Bird R1km 349 349 323 26 7.45 Yes - Columbina cyanopis Bird R5km 2500 100 47 53 53.00 Yes - Conothraupis mesoleuca Bird R1km 1032 1032 721 311 30.14 Yes - Coracias garrulus Bird R50km 35457500 14183 14144 39 0.27 Yes - Coracina newtoni Bird R1km 534 534 29 505 94.57 Yes - Coracina ostenta Bird R10km 24000 240 234 6 2.50 Yes - Corvus leucognaphalus Bird R10km 84300 843 79 764 90.63 Yes - Coryphaspiza melanotis Bird R50km 237500 95 69 26 27.37 Yes - Cotinga maculata Bird R10km 82900 829 8 821 99.03 Yes - Cotinga ridgwayi Bird R10km 11800 118 86 32 27.12 Yes - Crax alberti Bird R10km 43500 435 21 414 95.17 Yes - Crax globulosa Bird R10km 58000 580 258 322 55.52 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Crypturellus erythropus Bird R10km 68000 680 123 557 81.91 Yes - Cyanoramphus forbesi Bird R1km 377 377 314 63 16.71 Yes - Dacnis hartlaubi Bird R1km 289 289 135 154 53.29 Yes - Dacnis nigripes Bird R10km 29200 292 273 19 6.51 Yes - Dasyornis brachypterus Bird R10km 28900 289 20 269 93.08 Yes - Dicaeum haematostictum Bird R10km 24300 243 237 6 2.00 Yes - Didunculus strigirostris Bird R5km 2750 110 48 62 56.36 Yes - Ducula aurorae Bird R1km 298 298 26 272 91.28 Yes - Dysithamnus occidentalis Bird R10km 11300 113 79 34 30.09 Yes - Dysithamnus plumbeus Bird R10km 16500 165 111 54 32.00 Yes - Electron carinatum Bird R10km 42600 426 356 70 16.43 Yes - Eleothreptus candicans Bird R10km 8600 86 29 57 66.28 Yes - Empidonax fulvifrons Bird R50km 700000 280 231 49 17.50 Yes - Eos histrio Bird R5km 1825 73 42 31 42.47 Yes - Eremomela turneri Bird R5km 2950 118 108 10 8.00 Yes - Eriocnemis godini Bird R1km 408 408 96 312 76.47 Yes - Eriocnemis nigrivestis Bird R1km 947 947 79 868 91.66 Yes - Eunymphicus cornutus Bird R1km 770 770 691 79 10.26 Yes - Eutrichomyias rowleyi Bird R1km 557 557 7 550 98.74 Yes - Falco femoralis Bird R50km 12685000 5074 4666 408 8.00 Yes - Falco punctatus Bird R1km 775 775 159 616 79.48 Yes - Francolinus gularis Bird R50km 280000 112 48 64 57.14 Yes - Francolinus nahani Bird R5km 1975 79 69 10 12.66 Yes - Fulica cornuta Bird R50km 377500 151 77 74 49.01 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Galbula pastazae Bird R10km 22500 225 218 7 3.11 Yes - Gallicolumba rubescens Bird R1km 352 352 11 341 96.88 Yes - Garrulus lidthi Bird R1km 1068 1068 818 250 23.41 Yes - Geospiza difficilis Bird R5km 2475 99 53 46 46.46 Yes - Geospiza magnirostris Bird R10km 8100 81 74 7 8.64 Yes - Geothlypis speciosa Bird R1km 641 641 501 140 21.84 Yes - Geotrygon caniceps Bird R10km 114900 1149 281 868 75.00 Yes - Grallaria gigantea Bird R10km 8700 87 24 63 72.41 Yes - Grallaria kaestneri Bird R1km 1022 1022 434 588 57.53 Yes - Grallaria milleri Bird R1km 705 705 663 42 5.96 Yes - Grallaria rufocinerea Bird R10km 7600 76 54 22 28.95 Yes - Grallaricula cucullata Bird R5km 2400 96 59 37 38.54 Yes - Grus leucogeranus Bird R50km 350000 140 55 85 60.71 Yes - Guaruba guarouba Bird R50km 382500 153 42 111 72.55 Yes - Gubernatrix cristata Bird R50km 770000 308 297 11 3.57 Yes - Gyalophylax hellmayri Bird R10km 20100 201 126 75 37.31 Yes - Gymnogyps californianus Bird R50km 405000 162 38 124 76.00 Yes - Gymnomyza aubryana Bird R1km 1078 1078 1051 27 2.50 Yes - Gyps bengalensis Bird R50km 6035000 2414 1906 508 21.04 Yes - Gyps coprotheres Bird R50km 1390000 556 439 117 21.04 Yes - Gyps tenuirostris Bird R50km 1807500 723 269 454 62.79 Yes - Habia atrimaxillaris Bird R5km 2150 86 38 48 55.81 Yes - Hapalopsittaca fuertesi Bird R1km 840 840 53 787 93.69 Yes - Harpyhaliaetus coronatus Bird R50km 3907500 1563 1272 291 18.62 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Hemignathus flavus Bird R5km 1575 63 14 49 77.78 Yes - Hemignathus kauaiensis Bird R5km 1325 53 9 44 83.02 Yes - Hemignathus parvus Bird R5km 1400 56 11 45 80.36 Yes - Hemitriccus furcatus Bird R10km 10900 109 71 38 34.86 Yes - Hemitriccus mirandae Bird R5km 4350 174 74 100 57.47 Yes - Herpsilochmus pectoralis Bird R10km 16000 160 88 72 45.00 Yes - Houbaropsis bengalensis Bird R50km 287500 115 45 70 60.87 Yes - Hymenolaimus malacorhynchos Bird R50km 200000 80 21 59 73.75 Yes - Hypopyrrhus pyrohypogaster Bird R10km 8700 87 37 50 57.47 Yes - Hypsipetes olivaceus Bird R5km 1850 74 6 68 91.89 Yes - Ibycter americanus Bird R50km 8452500 3381 3234 147 4.00 Yes - Iodopleura pipra Bird R50km 160000 64 54 10 15.63 Yes - Ixos siquijorensis Bird R10km 5400 54 6 48 88.89 Yes - Jacamaralcyon tridactyla Bird R50km 412500 165 27 138 83.64 Yes - Laniisoma elegans Bird R50km 200000 80 69 11 13.75 Yes - Lepidopyga lilliae Bird R1km 242 242 160 82 33.00 Yes - Leptoptilos dubius Bird R50km 2255000 902 107 795 88.14 Yes - Leptotila ochraceiventris Bird R10km 41700 417 62 355 85.13 Yes - Leptotila wellsi Bird R1km 65 65 9 56 86.15 Yes - Lipaugus weberi Bird R1km 936 936 42 894 95.51 Yes - Macgregoria pulchra Bird R10km 5500 55 42 13 23.64 Yes - Macronyx sharpei Bird R5km 2725 109 98 11 10.09 Yes - Margarops fuscus Bird R10km 6100 61 55 6 9.84 Yes - Megapodius nicobariensis Bird R5km 2375 95 65 30 31.58 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Meleagris ocellata Bird R50km 167500 67 59 8 11.94 Yes - Mergus octosetaceus Bird R50km 1297500 519 13 506 97.50 Yes - Merulaxis stresemanni Bird R1km 92 92 24 68 73.91 Yes - Micrastur plumbeus Bird R10km 64400 644 625 19 2.95 Yes - Mohoua ochrocephala Bird R10km 29200 292 177 115 39.00 Yes - Myadestes obscurus Bird R10km 6800 68 26 42 61.76 Yes - Myadestes palmeri Bird R1km 177 177 31 146 82.49 Yes - Myiarchus semirufus Bird R10km 32100 321 24 297 92.52 Yes - Myrmoborus melanurus Bird R10km 19600 196 184 12 6.12 Yes - Myrmotherula fluminensis Bird R1km 78 78 10 68 87.18 Yes - Myrmotherula minor Bird R10km 11700 117 51 66 56.41 Yes - Myrmotherula unicolor Bird R10km 54900 549 347 202 36.79 Yes - Myrmotherula urosticta Bird R10km 14900 149 32 117 78.52 Yes - Myzomela rubratra Bird R5km 2150 86 65 21 24.42 Yes - Nemosia rourei Bird R1km 431 431 30 401 93.04 Yes - Neopelma aurifrons Bird R10km 8200 82 45 37 45.12 Yes - Nestor meridionalis Bird R10km 89700 897 717 180 20.07 Yes - Nothoprocta taczanowskii Bird R10km 22000 220 169 51 23.18 Yes - Nothura minor Bird R10km 39900 399 29 370 92.73 Yes - Odontophorus atrifrons Bird R5km 2425 97 89 8 8.25 Yes - Odontophorus hyperythrus Bird R10km 42100 421 13 408 96.91 Yes - Onychorhynchus occidentalis Bird R10km 31700 317 74 243 76.66 Yes - Oreomystis mana Bird R1km 750 750 691 59 7.87 Yes - Oreothraupis arremonops Bird R10km 16100 161 138 23 14.29 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Ortalis erythroptera Bird R10km 22200 222 192 30 13.51 Yes - Otus brookii Bird R10km 38800 388 296 92 23.71 Yes - Oxyura jamaicensis Bird R50km 14740000 5896 5888 8 0.14 Yes - Oxyura leucocephala Bird R50km 10917500 4367 4361 6 0.14 Yes - Pachycephala rufogularis Bird R10km 105800 1058 1013 45 4.25 Yes - Padda oryzivora Bird R50km 142500 57 43 14 24.56 Yes - Palmeria dolei Bird R1km 1065 1065 153 912 85.63 Yes - Paradisaea rudolphi Bird R10km 27200 272 221 51 18.75 Yes - Paroreomyza montana Bird R5km 1500 60 8 52 86.67 Yes - Parus nuchalis Bird R10km 66400 664 624 40 6.02 Yes - Patagioenas inornata Bird R50km 205000 82 19 63 76.83 Yes - Patagioenas oenops Bird R10km 8200 82 58 24 29.27 Yes - Pauxi pauxi Bird R10km 35700 357 339 18 5.04 Yes - Pauxi unicornis Bird R10km 59600 596 154 442 74.16 Yes - Pavo muticus Bird R50km 1910000 764 398 366 47.91 Yes - Penelope ochrogaster Bird R10km 74200 742 310 432 58.22 Yes - Penelope ortoni Bird R10km 77100 771 517 254 32.94 Yes - Penelope perspicax Bird R1km 228 228 110 118 51.75 Yes - Penelope pileata Bird R50km 980000 392 386 6 1.53 Yes - Penelopides panini Bird R10km 27700 277 267 10 3.61 Yes - Phapitreron cinereiceps Bird R1km 656 656 608 48 7.32 Yes - Philesturnus carunculatus Bird R1km 108 108 99 9 8.33 Yes - Phytotoma raimondii Bird R10km 16900 169 27 142 84.02 Yes - Picoides borealis Bird R50km 905000 362 245 117 32.32 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Picumnus fulvescens Bird R10km 77700 777 767 10 1.29 Yes - Picus squamatus Bird R50km 777500 311 282 29 9.32 Yes - Pipile jacutinga Bird R50km 1087500 435 15 420 96.55 Yes - Pipile pipile Bird R1km 339 339 254 85 25.07 Yes - Piprites pileata Bird R10km 45900 459 445 14 3.05 Yes - Podiceps taczanowskii Bird R1km 145 145 130 15 10.34 Yes - Poephila cincta Bird R50km 832500 333 201 132 39.64 Yes - Poliocephalus rufopectus Bird R50km 145000 58 34 24 41.38 Yes - Pomarea iphis Bird R1km 125 125 79 46 36.00 Yes - Pomarea mendozae Bird R1km 854 854 14 840 98.36 Yes - Poospiza cinerea Bird R10km 53200 532 171 361 67.86 Yes - Poospiza rubecula Bird R5km 1850 74 46 28 37.00 Yes - Porphyrio hochstetteri Bird R10km 5700 57 6 51 89.47 Yes - Primolius maracana Bird R50km 3852500 1541 1455 86 5.58 Yes - Prioniturus verticalis Bird R1km 756 756 608 148 19.00 Yes - Procnias tricarunculatus Bird R10km 123300 1233 1098 135 10.95 Yes - Psephotus chrysopterygius Bird R5km 3200 128 114 14 10.94 Yes - Pseudibis davisoni Bird R10km 61200 612 477 135 22.06 Yes - Pseudonestor xanthophrys Bird R1km 142 142 121 21 14.79 Yes - Psittacula eques Bird R1km 1013 1013 57 956 94.37 Yes - Psophodes nigrogularis Bird R50km 217500 87 29 58 66.67 Yes - Ptilinopus jambu Bird R10km 72700 727 457 270 37.14 Yes - Ptilinopus rarotongensis Bird R1km 76 76 59 17 22.37 Yes - Pyrilia pyrilia Bird R50km 185000 74 52 22 29.73 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Pyrrhura calliptera Bird R10km 5400 54 11 43 79.63 Yes - Pyrrhura cruentata Bird R10km 54200 542 104 438 80.81 Yes - Pyrrhura griseipectus Bird R5km 2475 99 26 73 73.00 Yes - Rallus antarcticus Bird R10km 110100 1101 1092 9 0.82 Yes - Rallus semiplumbeus Bird R5km 1525 61 31 30 49.18 Yes - Rhinomyias albigularis Bird R10km 24000 240 234 6 2.50 Yes - Rhopornis ardesiacus Bird R5km 2625 105 93 12 11.43 Yes - Saltator rufiventris Bird R10km 27500 275 242 33 12.00 Yes - Sarcoramphus papa Bird R50km 14415000 5766 5714 52 0.90 Yes - Saxicola macrorhynchus Bird R50km 722500 289 99 190 65.00 Yes - Sephanoides fernandensis Bird R1km 148 148 11 137 92.57 Yes - Serinus xantholaemus Bird R5km 1425 57 10 47 82.46 Yes - Simoxenops striatus Bird R10km 20400 204 196 8 3.92 Yes - Sporophila caerulescens Bird R50km 6275000 2510 1324 1186 47.25 Yes - Sporophila falcirostris Bird R10km 44000 440 411 29 6.59 Yes - Sporophila frontalis Bird R10km 82500 825 758 67 8.00 Yes - Starnoenas cyanocephala Bird R10km 110900 1109 282 827 74.00 Yes - Sterna acuticauda Bird R50km 3390000 1356 989 367 27.06 Yes - Sturnella defilippii Bird R10km 7400 74 49 25 33.78 Yes - Synallaxis infuscata Bird R1km 793 793 126 667 84.11 Yes - Synallaxis tithys Bird R10km 7500 75 51 24 32.00 Yes - Sypheotides indicus Bird R50km 2445000 978 160 818 83.64 Yes - Tachycineta euchrysea Bird R10km 11300 113 76 37 32.74 Yes - Tangara peruviana Bird R10km 100400 1004 258 746 74.30 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Taoniscus nanus Bird R10km 63300 633 578 55 8.69 Yes - Terenura sharpei Bird R5km 2775 111 101 10 9.01 Yes - Terpsiphone corvina Bird R1km 56 56 14 42 75.00 Yes - Thripophaga berlepschi Bird R5km 1450 58 41 17 29.31 Yes - Todiramphus gambieri Bird R1km 54 54 24 30 55.56 Yes - Touit melanonotus Bird R10km 7800 78 71 7 8.00 Yes - Touit stictopterus Bird R10km 34700 347 322 25 7.20 Yes - Touit surdus Bird R10km 63800 638 326 312 48.90 Yes - Triclaria malachitacea Bird R10km 49500 495 300 195 39.39 Yes - Turdoides bicolor Bird R50km 1147500 459 447 12 2.61 Yes - Turdoides hindei Bird R10km 24200 242 140 102 42.15 Yes - Turdus lherminieri Bird R5km 1975 79 55 24 30.38 Yes - Turdus swalesi Bird R5km 4575 183 156 27 14.75 Yes - Turdus xanthorhynchus Bird R1km 145 145 83 62 42.76 Yes - Tyrannus cubensis Bird R10km 110800 1108 32 1076 97.11 Yes - Vini australis Bird R5km 4150 166 146 20 12.05 Yes - Vini kuhlii Bird R1km 663 663 516 147 22.17 Yes - Vini ultramarina Bird R1km 335 335 163 172 51.34 Yes - Vireo atricapilla Bird R50km 627500 251 177 74 29.48 Yes - Vireo masteri Bird R5km 4600 184 132 52 28.26 Yes - Xanthopsar flavus Bird R50km 650000 260 78 182 70.00 Yes - Xenospingus concolor Bird R10km 31100 311 303 8 2.57 Yes - Xenospiza baileyi Bird R1km 342 342 69 273 79.00 Yes - Xiphocolaptes falcirostris Bird R10km 100500 1005 858 147 14.63 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Xipholena atropurpurea Bird R10km 43200 432 21 411 95.14 Yes - Zaratornis stresemanni Bird R10km 74000 740 733 7 0.95 Yes - Acinonyx jubatus Mammal R50km 2797500 1119 1110 9 0.80 Yes - Aegialomys galapagoensis Mammal R1km 568 568 23 545 95.95 Yes - Alcelaphus buselaphus Mammal R50km 6827500 2731 2690 41 1.50 Yes - Ammospermophilus nelsoni Mammal R10km 8600 86 25 61 70.00 Yes - Axis porcinus Mammal R50km 350000 140 106 34 24.29 Yes - Babyrousa babyrussa Mammal R10km 13300 133 127 6 4.51 Yes - Babyrousa celebensis Mammal R50km 135000 54 48 6 11.11 Yes - Barbastella barbastellus Mammal R50km 3767500 1507 1488 19 1.26 Yes - Bettongia tropica Mammal R1km 398 398 318 80 20.00 Yes - Canis aureus Mammal R50km 27145000 10858 10832 26 0.24 Yes - Canis lupus Mammal R50km 51455000 20582 20434 148 0.72 Yes - Canis simensis Mammal R10km 7100 71 63 8 11.27 Yes - Capreolus capreolus Mammal R50km 7227500 2891 2876 15 0.52 Yes - Choeropsis liberiensis Mammal R50km 137500 55 49 6 10.91 Yes - Coleura seychellensis Mammal R1km 227 227 175 52 22.91 Yes - Cricetulus migratorius Mammal R50km 7637500 3055 3048 7 0.23 Yes - Dendrolagus pulcherrimus Mammal R5km 2950 118 8 110 93.22 Yes - Diceros bicornis Mammal R50km 5162500 2065 1569 496 24.02 Yes - Emballonura semicaudata Mammal R10km 15900 159 18 141 88.68 Yes - Equus africanus Mammal R10km 100300 1003 882 121 12.06 Yes - Haeromys pusillus Mammal R10km 51600 516 404 112 21.71 Yes - Hapalomys longicaudatus Mammal R5km 2000 80 19 61 76.25 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Helarctos malayanus Mammal R50km 3225000 1290 596 694 53.80 Yes - Isoodon macrourus Mammal R50km 1215000 486 432 54 11.11 Yes - Kobus kob Mammal R50km 2990000 1196 1189 7 0.59 Yes - Komodomys rintjanus Mammal R10km 16200 162 22 140 86.42 Yes - Leptailurus serval Mammal R50km 12767500 5107 4990 117 2.29 Yes - Lynx pardinus Mammal R10km 26400 264 10 254 96.21 Yes - Macaca arctoides Mammal R50km 1440000 576 568 8 1.39 Yes - Mandrillus leucophaeus Mammal R10km 43900 439 429 10 2.00 Yes - Mastacomys fuscus Mammal R10km 65500 655 631 24 3.66 Yes - Melursus ursinus Mammal R50km 2775000 1110 602 508 45.77 Yes - Mesocapromys nanus Mammal R5km 3800 152 63 89 58.55 Yes - Microtus cabrerae Mammal R10km 119100 1191 1154 37 3.11 Yes - Microtus ochrogaster Mammal R50km 3247500 1299 1286 13 1.00 Yes - Myrmecophaga tridactyla Mammal R50km 13227500 5291 5169 122 2.31 Yes - Nomascus leucogenys Mammal R10km 51700 517 264 253 48.94 Yes - Nyctimene cephalotes Mammal R50km 222500 89 77 12 13.48 Yes - Oryx beisa Mammal R50km 1085000 434 418 16 3.69 Yes - Pecari tajacu Mammal R50km 15952500 6381 6150 231 3.62 Yes - Perognathus alticolus Mammal R1km 1038 1038 1011 27 2.60 Yes - Peromyscus madrensis Mammal R1km 414 414 288 126 30.43 Yes - Procolobus badius Mammal R50km 440000 176 141 35 19.89 Yes - Procolobus kirkii Mammal R5km 1525 61 29 32 52.46 Yes - Pseudantechinus mimulus Mammal R5km 4925 197 190 7 3.55 Yes - Pseudomys desertor Mammal R50km 5175000 2070 1408 662 31.98 Yes -

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Pteralopex flanneryi Mammal R10km 14100 141 107 34 24.11 Yes - Pteralopex taki Mammal R5km 3375 135 106 29 21.48 Yes - Pteropus niger Mammal R5km 4375 175 74 101 57.71 Yes - Reithrodontomys raviventris Mammal R1km 988 988 916 72 7.29 Yes - Rusa marianna Mammal R10km 108300 1083 1037 46 4.25 Yes - Sus barbatus Mammal R50km 585000 234 225 9 3.85 Yes - Sus philippensis Mammal R10km 64200 642 633 9 1.40 Yes - Tapirus bairdii Mammal R50km 835000 334 300 34 10.18 Yes - Tayassu pecari Mammal R50km 13777500 5511 4540 971 17.62 Yes - Thylogale brunii Mammal R10km 108600 1086 987 99 9.00 Yes - Thylogale calabyi Mammal R1km 1005 1005 644 361 35.92 Yes - Trachypithecus geei Mammal R10km 7300 73 55 18 24.66 Yes - Ursus arctos Mammal R50km 29747500 11899 10967 932 7.83 Yes - Ursus thibetanus Mammal R50km 7065000 2826 1321 1505 53.26 Yes - Wilfredomys oenax Mammal R10km 119400 1194 1156 38 3.18 Yes - Zaglossus bruijnii Mammal R10km 91700 917 899 18 1.96 Yes - Acanthodactylus schreiberi Reptile R10km 10800 108 85 23 21.30 Yes - Gallotia auaritae Reptile R1km 409 409 149 260 63.57 Yes - Iguana delicatissima Reptile R5km 4475 179 117 62 34.64 Yes - Phrynosoma cornutum Reptile R50km 1795000 718 710 8 1.00 Yes - Pituophis ruthveni Reptile R10km 74400 744 14 730 98.12 Yes - Psammodromus microdactylus Reptile R10km 7900 79 39 40 50.63 Yes - Agalychnis annae Amphibian R5km 3875 155 2 153 98.71 No Current Craugastor taurus Amphibian R10km 5900 59 1 58 98.31 No Current

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Lithobates sevosus Amphibian R10km 21600 216 1 215 99.54 No Current Acrocephalus rodericanus Bird R1km 114 114 4 110 96.49 No Current Agelaius xanthomus Bird R10km 9100 91 4 87 95.60 No Current Amazona vittata Bird R10km 9100 91 1 90 98.90 No Current Buteo ridgwayi Bird R10km 35400 354 2 352 99.44 No Current Charmosyna diadema Bird R1km 64 64 1 63 98.44 No Current Copsychus sechellarum Bird R1km 210 210 1 209 99.52 No Current Coracina typica Bird R5km 1850 74 4 70 94.59 No Current Crax blumenbachii Bird R50km 142500 57 1 56 98.25 No Current Curaeus forbesi Bird R10km 18600 186 1 185 99.46 No Current Cyanoramphus malherbi Bird R10km 44600 446 3 443 99.33 No Current Cyanoramphus novaezelandiae Bird R50km 272500 109 2 107 98.17 No Current Foudia flavicans Bird R1km 114 114 5 109 95.61 No Current Foudia sechellarum Bird R1km 54 54 1 53 98.15 No Current Glaucis dohrnii Bird R10km 109300 1093 4 1089 99.63 No Current Hemignathus munroi Bird R10km 7800 78 1 77 98.72 No Current Leucopeza semperi Bird R1km 77 77 1 76 98.70 No Current Loxops coccineus Bird R10km 10500 105 4 101 96.19 No Current Macroagelaius subalaris Bird R10km 8000 80 5 75 93.75 No Current Mimus trifasciatus Bird R1km 172 172 1 171 99.42 No Current Myadestes lanaiensis Bird R5km 1450 58 2 56 96.55 No Current Myrmeciza ruficauda Bird R10km 46100 461 4 457 99.13 No Current Nesoenas mayeri Bird R5km 1850 74 3 71 95.95 No Current Oreomystis bairdi Bird R5km 1275 51 4 47 92.16 No Current

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Pardalotus quadragintus Bird R10km 6100 61 1 60 98.36 No Current Paroreomyza maculata Bird R5km 1500 60 4 56 93.33 No Current Pomarea dimidiata Bird R1km 69 69 3 66 95.65 No Current Psittirostra psittacea Bird R10km 16100 161 3 158 98.14 No Current Rhipidura semirubra Bird R5km 2000 80 5 75 93.75 No Current Sporophila melanops Bird R1km 706 706 2 704 99.72 No Current Terenura sicki Bird R10km 13800 138 1 137 99.28 No Current Thripophaga macroura Bird R50km 145000 58 2 56 96.55 No Current Zosterops modestus Bird R1km 152 152 4 148 97.37 No Current Conilurus penicillatus Mammal R50km 240000 96 5 91 94.79 No Current Sminthopsis aitkeni Mammal R5km 1425 57 3 54 94.74 No Current Alytes obstetricans Amphibian R50km 970000 388 386 2 0.52 No Extirpated Gastrotheca christiani Amphibian R5km 1475 59 58 1 1.69 No Extirpated Liuixalus romeri Amphibian R1km 82 82 80 2 2.44 No Extirpated Pelobates syriacus Amphibian R50km 535000 214 213 1 0.47 No Extirpated Pleurodeles waltl Amphibian R50km 377500 151 150 1 0.66 No Extirpated Rana latastei Amphibian R10km 33200 332 330 2 0.60 No Extirpated Salamandra algira Amphibian R10km 21200 212 210 2 0.94 No Extirpated Acrocephalus brevipennis Bird R5km 1425 57 54 3 5.26 No Extirpated Agapornis pullarius Bird R50km 2662500 1065 1064 1 0.09 No Extirpated Amaurolimnas concolor Bird R50km 1595000 638 634 4 0.63 No Extirpated Apus sladeniae Bird R10km 12700 127 126 1 0.79 No Extirpated Ara ararauna Bird R50km 7750000 3100 3097 3 0.10 No Extirpated Ara rubrogenys Bird R10km 9900 99 97 2 2.02 No Extirpated

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Asthenes anthoides Bird R50km 335000 134 133 1 0.75 No Extirpated Botaurus lentiginosus Bird R50km 14207500 5683 5682 1 0.02 No Extirpated Bubulcus ibis Bird R50km 63440000 25376 25375 1 0.00 No Extirpated Cacatua moluccensis Bird R10km 18300 183 179 4 2.19 No Extirpated Campylopterus phainopeplus Bird R5km 2850 114 111 3 2.63 No Extirpated Charmosyna palmarum Bird R10km 12000 120 117 3 2.50 No Extirpated Chlorochrysa nitidissima Bird R10km 18200 182 177 5 2.75 No Extirpated Chloropeta gracilirostris Bird R10km 75800 758 753 5 0.66 No Extirpated Chlorospingus flavovirens Bird R10km 5600 56 54 2 3.57 No Extirpated Colaptes auratus Bird R50km 15722500 6289 6288 1 0.02 No Extirpated Collocalia elaphra Bird R1km 207 207 204 3 1.45 No Extirpated Collocalia orientalis Bird R10km 6800 68 63 5 7.35 No Extirpated Compsospiza garleppi Bird R5km 3775 151 150 1 0.66 No Extirpated Coturnicops noveboracensis Bird R50km 4145000 1658 1657 1 0.06 No Extirpated Dendrocygna bicolor Bird R50km 18717500 7487 7483 4 0.05 No Extirpated Dendrocygna viduata Bird R50km 32687500 13075 13074 1 0.01 No Extirpated Ducula brenchleyi Bird R10km 10600 106 105 1 0.94 No Extirpated Euscarthmus rufomarginatus Bird R50km 1915000 766 761 5 0.65 No Extirpated Falco araea Bird R1km 239 239 235 4 1.67 No Extirpated Formicivora iheringi Bird R10km 21200 212 208 4 1.89 No Extirpated Gallicolumba sanctaecrucis Bird R5km 1850 74 72 2 2.70 No Extirpated Gallinago stricklandii Bird R50km 355000 142 137 5 3.52 No Extirpated Gallus gallus Bird R50km 5097500 2039 2035 4 0.20 No Extirpated Garrulax courtoisi Bird R1km 324 324 323 1 0.31 No Extirpated

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Grallaria alleni Bird R10km 7800 78 77 1 1.28 No Extirpated Hemignathus virens Bird R10km 6400 64 63 1 1.56 No Extirpated Hypothymis coelestis Bird R50km 182500 73 68 5 6.85 No Extirpated Icterus leucopteryx Bird R10km 11600 116 114 2 1.72 No Extirpated Jacana spinosa Bird R50km 1037500 415 413 2 0.48 No Extirpated Leptasthenura xenothorax Bird R5km 2625 105 103 2 1.90 No Extirpated Lipaugus lanioides Bird R50km 205000 82 81 1 1.22 No Extirpated Loxigilla portoricensis Bird R10km 9200 92 90 2 2.17 No Extirpated Megapodius layardi Bird R10km 6800 68 67 1 1.47 No Extirpated Megascops nudipes Bird R10km 9600 96 94 2 2.08 No Extirpated Megaxenops parnaguae Bird R50km 667500 267 264 3 1.00 No Extirpated Merops breweri Bird R50km 1617500 647 645 2 0.31 No Extirpated Microhierax erythrogenys Bird R50km 250000 100 99 1 1.00 No Extirpated Myrmeciza griseiceps Bird R10km 17800 178 175 3 1.69 No Extirpated Notiomystis cincta Bird R1km 58 58 54 4 6.90 No Extirpated Oreophasis derbianus Bird R10km 7600 76 75 1 1.32 No Extirpated Penelope barbata Bird R10km 16300 163 162 1 0.61 No Extirpated Pipilo erythrophthalmus Bird R50km 8057500 3223 3222 1 0.03 No Extirpated Pipilo maculatus Bird R50km 4592500 1837 1836 1 0.05 No Extirpated Platyrinchus leucoryphus Bird R50km 437500 175 172 3 1.71 No Extirpated Prosopeia personata Bird R10km 10500 105 104 1 0.95 No Extirpated Ptilinopus perousii Bird R10km 19700 197 192 5 2.54 No Extirpated Regulus calendula Bird R50km 14227500 5691 5690 1 0.02 No Extirpated Rhaphidura sabini Bird R50km 1742500 697 695 2 0.29 No Extirpated

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Sitta canadensis Bird R50km 12255000 4902 4901 1 0.02 No Extirpated Strix occidentalis Bird R50km 1247500 499 494 5 1.00 No Extirpated Syndactyla ruficollis Bird R10km 18500 185 183 2 1.08 No Extirpated Taphrolesbia griseiventris Bird R5km 4325 173 172 1 0.58 No Extirpated Troglodytes aedon Bird R50km 29015000 11606 11604 2 0.02 No Extirpated Turnix varius Bird R50km 2305000 922 917 5 0.54 No Extirpated Tyto longimembris Bird R50km 4292500 1717 1712 5 0.29 No Extirpated Vestiaria coccinea Bird R10km 16200 162 160 2 1.23 No Extirpated Vini peruviana Bird R5km 1850 74 73 1 1.35 No Extirpated Xenornis setifrons Bird R10km 6200 62 60 2 3.23 No Extirpated Zimmerius villarejoi Bird R5km 1750 70 68 2 2.86 No Extirpated Acerodon jubatus Mammal R50km 150000 60 56 4 6.67 No Extirpated Bos javanicus Mammal R50km 180000 72 70 2 2.78 No Extirpated Capricornis milneedwardsii Mammal R50km 2917500 1167 1166 1 0.09 No Extirpated Cephalophus adersi Mammal R5km 2575 103 102 1 0.97 No Extirpated Cervus nippon Mammal R50km 352500 141 140 1 0.71 No Extirpated Crocidura canariensis Mammal R5km 2475 99 97 2 2.02 No Extirpated Crocidura sicula Mammal R10km 25900 259 257 2 0.77 No Extirpated Dicerorhinus sumatrensis Mammal R10km 8000 80 78 2 2.50 No Extirpated Eptesicus fuscus Mammal R50km 13152500 5261 5260 1 0.02 No Extirpated Lasiurus borealis Mammal R50km 4870000 1948 1947 1 0.05 No Extirpated Macaca nigra Mammal R10km 5800 58 56 2 3.45 No Extirpated Martes zibellina Mammal R50km 6725000 2690 2685 5 0.19 No Extirpated Oryzomys couesi Mammal R50km 1475000 590 586 4 0.68 No Extirpated

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Phascogale pirata Mammal R10km 40500 405 404 1 0.25 No Extirpated Pseudomys shortridgei Mammal R10km 31900 319 315 4 1.25 No Extirpated Pteropus melanotus Mammal R10km 8000 80 75 5 6.25 No Extirpated Rhinolophus canuti Mammal R10km 5700 57 54 3 5.26 No Extirpated Rucervus eldii Mammal R10km 84700 847 843 4 0.47 No Extirpated Sus cebifrons Mammal R5km 2375 95 92 3 3.16 No Extirpated Chalcides mauritanicus Reptile R5km 4250 170 169 1 0.59 No Extirpated Coronella girondica Reptile R50km 1212500 485 482 3 0.62 No Extirpated Crotalus horridus Reptile R50km 2050000 820 818 2 0.24 No Extirpated Cyrtopodion russowii Reptile R50km 1352500 541 538 3 0.55 No Extirpated Phrynosoma douglasii Reptile R50km 372500 149 148 1 0.67 No Extirpated Trapelus savignii Reptile R10km 36500 365 364 1 0.27 No Extirpated Vipera ursinii Reptile R10km 67900 679 677 2 0.29 No Extirpated Pseudoeurycea robertsi Amphibian R1km 15 15 8 7 46.67 No Historical Acrocephalus familiaris Bird R1km 6 6 1 5 83.33 No Historical Attila torridus Bird R10km 4800 48 47 1 2.08 No Historical Gallirallus sylvestris Bird R1km 13 13 1 12 92.31 No Historical Icterus oberi Bird R1km 26 26 9 17 65.38 No Historical Melamprosops phaeosoma Bird R1km 11 11 4 7 63.64 No Historical Otus insularis Bird R1km 47 47 38 9 19.00 No Historical Pyriglena atra Bird R10km 5000 50 48 2 4.00 No Historical Spizella wortheni Bird R1km 26 26 24 2 7.69 No Historical Toxostoma guttatum Bird R1km 36 36 24 12 33.33 No Historical Zosterops albogularis Bird R1km 36 36 6 30 83.33 No Historical

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Bombina variegata Amphibian R50km 1080000 432 432 0 0.00 No Topology Craugastor punctariolus Amphibian R5km 3775 151 151 0 0.00 No Topology Hylomantis lemur Amphibian R10km 16200 162 91 72 44.44 No Topology Lissotriton helveticus Amphibian R50km 1067500 427 427 0 0.00 No Topology Lithobates warszewitschii Amphibian R50km 162500 65 65 0 0.00 No Topology Acrocephalus sechellensis Bird R1km 1 1 0 1 100.00 No Topology Amazilia castaneiventris Bird R5km 3200 128 128 18 14.06 No Topology Amazona barbadensis Bird R10km 14600 146 111 40 27.40 No Topology Athene cunicularia Bird R50km 15145000 6058 6058 0 0.00 No Topology Atlapetes pallidiceps Bird R5km 1325 53 0 53 100.00 No Topology Bradypterus grandis Bird R10km 12500 125 125 0 0.00 No Topology Calyptura cristata Bird R5km 2675 107 0 107 100.00 No Topology Camarhynchus psittacula Bird R10km 7300 73 73 0 0.00 No Topology Campephilus imperialis Bird R50km 355000 142 0 142 100.00 No Topology Campylopterus ensipennis Bird R5km 2400 96 96 0 0.00 No Topology Carpodectes antoniae Bird R5km 1475 59 59 0 0.00 No Topology Cephalopterus penduliger Bird R10km 70400 704 699 7 0.99 No Topology Cinclodes palliatus Bird R5km 3200 128 128 0 0.00 No Topology Coccyzus americanus Bird R50km 21085000 8434 8434 0 0.00 No Topology Compsospiza baeri Bird R10km 24600 246 246 0 0.00 No Topology Corvus palmarum Bird R10km 116100 1161 1100 73 6.29 No Topology Crotophaga ani Bird R50km 14150000 5660 5660 0 0.00 No Topology Crypturellus kerriae Bird R10km 5900 59 59 0 0.00 No Topology Cyanolanius madagascarinus Bird R50km 452500 181 181 0 0.00 No Topology

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Cyanopsitta spixii Bird R50km 125000 50 0 50 100.00 No Topology Dendrocopos canicapillus Bird R50km 8007500 3203 3203 0 0.00 No Topology Ducula latrans Bird R10km 15900 159 159 0 0.00 No Topology Ducula mindorensis Bird R5km 2400 96 50 93 96.88 No Topology Euphonia musica Bird R10km 88800 888 888 0 0.00 No Topology Eurochelidon sirintarae Bird R5km 2125 85 0 85 100.00 No Topology Foudia rubra Bird R5km 1850 74 0 74 100.00 No Topology Gallicolumba erythroptera Bird R5km 1425 57 0 57 100.00 No Topology Hapalopsittaca amazonina Bird R10km 29800 298 256 45 15.10 No Topology Hapalopsittaca pyrrhops Bird R10km 10300 103 103 0 0.00 No Topology Heliodoxa gularis Bird R10km 25500 255 230 55 21.57 No Topology Hemignathus lucidus Bird R5km 3875 155 0 155 100.00 No Topology Hemitriccus kaempferi Bird R10km 7700 77 77 1 1.30 No Topology Himatione sanguinea Bird R10km 16300 163 163 0 0.00 No Topology Hylocryptus erythrocephalus Bird R10km 15400 154 154 0 0.00 No Topology Junco hyemalis Bird R50km 16202500 6481 6481 1 0.02 No Topology Lathrotriccus euleri Bird R50km 10297500 4119 4119 0 0.00 No Topology Lathrotriccus griseipectus Bird R10km 65500 655 136 526 80.31 No Topology Loxioides bailleui Bird R5km 1325 53 7 47 88.68 No Topology Loxops caeruleirostris Bird R1km 5 5 5 0 0.00 No Topology Myrmotherula grisea Bird R10km 32500 325 325 0 0.00 No Topology Neochmia ruficauda Bird R50km 1480000 592 531 100 16.89 No Topology Numenius borealis Bird R50km 2252500 901 0 901 100.00 No Topology Odontophorus strophium Bird R5km 2750 110 48 69 62.73 No Topology

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Ognorhynchus icterotis Bird R50km 145000 58 0 58 100.00 No Topology Onychorhynchus coronatus Bird R50km 6942500 2777 2777 9 0.32 No Topology Oreopholus ruficollis Bird R50km 3120000 1248 1248 0 0.00 No Topology Passer hispaniolensis Bird R50km 8875000 3550 3550 0 0.00 No Topology Penelope albipennis Bird R5km 3075 123 123 0 0.00 No Topology Penelopides mindorensis Bird R10km 8400 84 7 84 100.00 No Topology Petroica traversi Bird R1km 3 3 3 0 0.00 No Topology Phylloscartes paulista Bird R50km 490000 196 195 12 6.12 No Topology Phylloscartes roquettei Bird R50km 180000 72 72 0 0.00 No Topology Picathartes oreas Bird R50km 375000 150 150 0 0.00 No Topology Picumnus steindachneri Bird R10km 7000 70 70 0 0.00 No Topology Platyspiza crassirostris Bird R10km 7300 73 73 0 0.00 No Topology Pseudocolopteryx dinelliana Bird R50km 727500 291 291 0 0.00 No Topology Quelea erythrops Bird R50km 2565000 1026 1026 1 0.10 No Topology Ramphocinclus brachyurus Bird R5km 1375 55 0 55 100.00 No Topology Salpinctes obsoletus Bird R50km 5550000 2220 2220 0 0.00 No Topology Siphonorhis americana Bird R1km 108 108 0 108 100.00 No Topology Siptornopsis hypochondriaca Bird R10km 7300 73 73 0 0.00 No Topology Sporophila nigrorufa Bird R50km 192500 77 77 0 0.00 No Topology Strigops habroptila Bird R10km 18000 180 0 180 100.00 No Topology Sylvia atricapilla Bird R50km 22410000 8964 8964 0 0.00 No Topology Sylvia melanocephala Bird R50km 4360000 1744 1744 0 0.00 No Topology Tangara fastuosa Bird R10km 15600 156 156 0 0.00 No Topology Thinocorus rumicivorus Bird R50km 2240000 896 896 0 0.00 No Topology

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Tympanuchus cupido Bird R50km 3280000 1312 153 1179 89.86 No Topology Tympanuchus pallidicinctus Bird R50km 455000 182 27 182 100.00 No Topology Tyrannus caudifasciatus Bird R50km 200000 80 80 0 0.00 No Topology Vanellus gregarius Bird R50km 11862500 4745 3333 1565 32.00 No Topology Vanellus macropterus Bird R10km 7700 77 0 77 100.00 No Topology Vermivora bachmanii Bird R50km 917500 367 0 367 100.00 No Topology Zenaida aurita Bird R50km 235000 94 94 0 0.00 No Topology Zosterops chloronothus Bird R5km 1850 74 3 72 97.30 No Topology Blastocerus dichotomus Mammal R50km 4082500 1633 128 1508 92.35 No Topology Cephalophus harveyi Mammal R50km 347500 139 139 0 0.00 No Topology Chaetodipus rudinoris Mammal R10km 118300 1183 1183 0 0.00 No Topology Dipodomys stephensi Mammal R5km 2875 115 115 0 0.00 No Topology Gazella dorcas Mammal R50km 9912500 3965 3965 0 0.00 No Topology Hipposideros cervinus Mammal R50km 1972500 789 789 0 0.00 No Topology Hystrix cristata Mammal R50km 5347500 2139 2139 0 0.00 No Topology Macaca fuscata Mammal R50km 177500 71 71 0 0.00 No Topology Muntiacus muntjak Mammal R50km 1500000 600 600 0 0.00 No Topology Mustela lutreola Mammal R50km 2297500 919 176 745 81.00 No Topology Myotis nattereri Mammal R50km 5827500 2331 2331 0 0.00 No Topology Panthera pardus Mammal R50km 22205000 8882 8859 24 0.27 No Topology Procolobus rufomitratus Mammal R50km 1342500 537 537 0 0.00 No Topology Pteralopex atrata Mammal R10km 4700 47 47 0 0.00 No Topology Rhinolophus pusillus Mammal R50km 3927500 1571 1571 0 0.00 No Topology Spermophilus mohavensis Mammal R10km 20500 205 205 0 0.00 No Topology

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historial Current Extirpated Contraction Included Reason for group range (km2) cells cells cells (%) Exclusion Sus celebensis Mammal R50km 172500 69 69 0 0.00 No Topology Sus scrofa Mammal R50km 27907500 11163 11163 0 0.00 No Topology Tapirus terrestris Mammal R50km 13020000 5208 4505 711 13.65 No Topology Thylogale browni Mammal R50km 227500 91 91 0 0.00 No Topology Tragulus napu Mammal R50km 1400000 560 560 0 0.00 No Topology Lacerta bilineata Reptile R50km 797500 319 319 0 0.00 No Topology Lacerta media Reptile R50km 855000 342 342 0 0.00 No Topology Lacerta schreiberi Reptile R50km 140000 56 56 0 0.00 No Topology Phrynocephalus persicus Reptile R50km 127500 51 51 0 0.00 No Topology Regina septemvittata Reptile R50km 1012500 405 405 0 0.00 No Topology Timon lepidus Reptile R50km 632500 253 253 0 0.00 No Topology Timon tangitanus Reptile R50km 130000 52 52 0 0.00 No Topology

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Table A1.2. Description of data for all terrestrial vertebrates species with fragments with contraction (n=430) obtained from the IUCN including: scientific name (Species), class (Class), the group in which it was rasterized (Raster group), the historical area of the fragment (Historical fragment km2), the number of historical cells (Historical cells), the number of current cells (Current cells), the percentage of contraction of the fragment (Contraction), if the fragment was included for analysis (Included), and the Reason for Exclusion. Exclusion was based on number of historical cells (Historical), number of current cells (Current), number of extirpated cells (Extirpated) or on topological errors defined as the overlap between current and extirpated areas (Topology). Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Atelopus chiriquiensis Amphibian R5km 4450 178 63 115 64.00 Yes - Atelopus varius Amphibian R10km 21600 216 61 155 71.00 Yes - Bolitoglossa robusta Amphibian R5km 4650 186 150 36 19.00 Yes - Craugastor ranoides Amphibian R10km 11000 110 13 97 88.00 Yes - Craugastor ranoides Amphibian R10km 29200 292 114 178 60.96 Yes - Craugastor rhyacobatrachus Amphibian R5km 1325 53 14 39 73.00 Yes - Duellmanohyla uranochroa Amphibian R10km 14800 148 7 141 95.00 Yes - Hyloscirtus colymba Amphibian R5km 3750 150 80 70 46.00 Yes - Isthmohyla angustilineata Amphibian R5km 1325 53 31 22 41.00 Yes - Isthmohyla calypsa Amphibian R1km 318 318 291 27 8.00 Yes - Neurergus microspilotus Amphibian R1km 576 576 394 182 31.00 Yes - Pleurodeles waltl Amphibian R10km 56300 563 540 23 4.00 Yes - Pristimantis caryophyllaceus Amphibian R10km 32800 328 260 68 20.00 Yes - Rana muscosa Amphibian R10km 8200 82 7 75 91.00 Yes - Rana muscosa Amphibian R5km 3250 130 13 117 90.00 Yes - Rana sierrae Amphibian R10km 29000 290 17 273 94.00 Yes - Strabomantis bufoniformis Amphibian R10km 57900 579 568 11 1.00 Yes - Accipiter gundlachi Bird R10km 106700 1067 128 939 88.00 Yes - Agapornis nigrigenis Bird R10km 19300 193 164 29 15.00 Yes - 83

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Agriornis albicauda Bird R50km 625000 250 224 26 10.40 Yes - Alectrurus tricolor Bird R50km 722500 289 178 111 38.00 Yes - Amazona agilis Bird R10km 8500 85 26 59 69.00 Yes - Amazona arausiaca Bird R1km 771 771 74 697 90.00 Yes - Amazona imperialis Bird R1km 165 165 53 112 67.00 Yes - Amazona leucocephala Bird R10km 106700 1067 122 945 88.00 Yes - Amazona oratrix Bird R10km 50200 502 31 471 93.00 Yes - Amazona oratrix Bird R10km 26100 261 35 226 86.00 Yes - Amazona oratrix Bird R10km 75500 755 168 587 77.00 Yes - Amazona versicolor Bird R1km 615 615 138 477 77.00 Yes - Amazona viridigenalis Bird R10km 68200 682 77 605 88.00 Yes - Anairetes alpinus Bird R5km 1775 71 56 15 21.00 Yes - Anodorhynchus hyacinthinus Bird R50km 147500 59 52 7 11.86 Yes - Anthocephala floriceps Bird R5km 3075 123 110 13 10.00 Yes - Anthus spragueii Bird R50km 1217500 487 405 82 16.00 Yes - Aphelocoma coerulescens Bird R10km 64600 646 64 582 90.00 Yes - Apteryx haastii Bird R10km 39900 399 90 309 77.00 Yes - Ara ambiguus Bird R1km 1097 1097 183 914 83.00 Yes - Ara ambiguus Bird R10km 66200 662 628 34 5.00 Yes - Ara macao Bird R50km 340000 136 13 123 90.00 Yes - Aratinga chloroptera Bird R10km 66300 663 190 473 71.00 Yes - Aratinga euops Bird R10km 106700 1067 118 949 88.00 Yes - Ardeotis nigriceps Bird R50km 1070000 428 189 239 55.00 Yes - 84

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Asthenes heterura Bird R10km 14500 145 111 34 23.45 Yes - Atlapetes fuscoolivaceus Bird R5km 2825 113 44 69 61.00 Yes - Atrichornis rufescens Bird R10km 61300 613 171 442 72.00 Yes - Bangsia aureocincta Bird R1km 805 805 636 169 20.00 Yes - Bangsia melanochlamys Bird R1km 1244 1244 165 1079 86.00 Yes - Bangsia melanochlamys Bird R10km 7600 76 40 36 47.00 Yes - Basileuterus basilicus Bird R1km 576 576 436 140 24.00 Yes - Basileuterus basilicus Bird R1km 228 228 212 16 7.00 Yes - Basileuterus conspicillatus Bird R10km 5500 55 32 23 41.00 Yes - Bolborhynchus ferrugineifrons Bird R1km 75 75 34 41 54.00 Yes - Bolborhynchus ferrugineifrons Bird R1km 512 512 346 166 32.42 Yes - Bolborhynchus ferrugineifrons Bird R1km 76 76 57 19 25.00 Yes - Bolborhynchus ferrugineifrons Bird R1km 72 72 58 14 19.00 Yes - Bolborhynchus ferrugineifrons Bird R1km 124 124 116 8 6.00 Yes - Bolborhynchus ferrugineifrons Bird R1km 389 389 376 13 3.00 Yes - Branta sandvicensis Bird R10km 6000 60 9 51 85.00 Yes - Brotogeris pyrrhoptera Bird R10km 9000 90 81 9 10.00 Yes - Bucco noanamae Bird R10km 31300 313 296 17 5.00 Yes - Buteo ridgwayi Bird R5km 3825 153 8 145 94.00 Yes - Callaeas cinereus Bird R10km 34500 345 14 331 95.00 Yes - Callaeas cinereus Bird R10km 21600 216 41 175 81.00 Yes - Capito hypoleucus Bird R1km 1148 1148 1001 147 12.00 Yes - Carduelis cucullata Bird R10km 82700 827 16 811 98.00 Yes - 85

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Carduelis siemiradzkii Bird R10km 23500 235 173 62 26.00 Yes - Carpornis melanocephala Bird R10km 11500 115 6 109 94.00 Yes - Carpornis melanocephala Bird R10km 28500 285 180 105 36.00 Yes - Chaetornis striata Bird R50km 227500 91 41 50 54.00 Yes - Chasiempis ibidis Bird R5km 1575 63 7 56 88.00 Yes - Chasiempis sandwichensis Bird R10km 9500 95 58 37 38.00 Yes - Chasiempis sclateri Bird R5km 1325 53 15 38 71.00 Yes - Chondrohierax wilsonii Bird R10km 106700 1067 38 1029 96.00 Yes - Cinclodes aricomae Bird R10km 51500 515 11 504 97.00 Yes - Cinclus schulzi Bird R1km 702 702 46 656 93.00 Yes - Cinclus schulzi Bird R1km 748 748 97 651 87.00 Yes - Cinclus schulzi Bird R5km 2400 96 82 14 14.00 Yes - Coccyzus americanus Bird R1km 1130 1130 405 725 64.00 Yes - Coccyzus americanus Bird R5km 1700 68 33 35 51.00 Yes - Coccyzus rufigularis Bird R5km 1575 63 36 27 42.00 Yes - Coeligena prunellei Bird R10km 6400 64 53 11 17.00 Yes - Colaptes fernandinae Bird R10km 106700 1067 74 993 93.00 Yes - Collocalia orientalis Bird R5km 1375 55 35 20 36.00 Yes - Columba thomensis Bird R1km 348 348 323 25 7.00 Yes - Coracina newtoni Bird R1km 534 534 29 505 94.00 Yes - Corvus leucognaphalus Bird R10km 74100 741 78 663 89.00 Yes - Coryphaspiza melanotis Bird R10km 47500 475 411 64 13.00 Yes - Cotinga ridgwayi Bird R5km 2325 93 47 46 49.00 Yes - 86

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Cotinga ridgwayi Bird R10km 8700 87 68 19 21.00 Yes - Crax alberti Bird R10km 27300 273 15 258 94.00 Yes - Crax alberti Bird R1km 882 882 63 819 92.00 Yes - Crypturellus erythropus Bird R10km 67900 679 123 556 81.00 Yes - Dacnis nigripes Bird R10km 11400 114 101 13 11.00 Yes - Dasyornis brachypterus Bird R10km 27800 278 9 269 96.00 Yes - Didunculus strigirostris Bird R1km 1130 1130 404 726 64.00 Yes - Didunculus strigirostris Bird R5km 1700 68 33 35 51.00 Yes - Dysithamnus occidentalis Bird R5km 3000 120 88 32 26.00 Yes - Dysithamnus plumbeus Bird R10km 7600 76 58 18 23.00 Yes - Empidonax fulvifrons Bird R50km 610000 244 195 49 20.00 Yes - Eremomela turneri Bird R1km 377 377 149 228 60.00 Yes - Eriocnemis nigrivestis Bird R1km 884 884 58 826 93.00 Yes - Eutrichomyias rowleyi Bird R1km 557 557 7 550 98.00 Yes - Falco femoralis Bird R50km 1145000 458 65 393 85.00 Yes - Falco punctatus Bird R1km 775 775 159 616 79.00 Yes - Francolinus gularis Bird R50km 280000 112 48 64 57.00 Yes - Fulica cornuta Bird R50km 375000 150 76 74 49.00 Yes - Galbula pastazae Bird R10km 22500 225 218 7 3.00 Yes - Geotrygon caniceps Bird R10km 106700 1067 233 834 78.00 Yes - Geotrygon caniceps Bird R5km 2025 81 30 51 62.00 Yes - Grallaria alleni Bird R1km 88 88 41 47 53.00 Yes - Grallaria gigantea Bird R5km 4075 163 73 90 55.00 Yes - 87

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Grallaria kaestneri Bird R1km 1022 1022 434 588 57.00 Yes - Grallaria milleri Bird R1km 705 705 663 42 5.00 Yes - Guaruba guarouba Bird R50km 382500 153 42 111 72.00 Yes - Gubernatrix cristata Bird R50km 770000 308 297 11 3.00 Yes - Gymnogyps californianus Bird R50km 367500 147 26 121 82.00 Yes - Gyps bengalensis Bird R50km 5620000 2248 1740 508 22.00 Yes - Gyps tenuirostris Bird R50km 1382500 553 109 444 80.00 Yes - Gyps tenuirostris Bird R50km 425000 170 160 10 5.00 Yes - Habia atrimaxillaris Bird R5km 2150 86 38 48 55.00 Yes - Harpyhaliaetus coronatus Bird R50km 3900000 1560 1272 288 18.00 Yes - Hemignathus flavus Bird R5km 1575 63 14 49 77.00 Yes - Hemignathus kauaiensis Bird R5km 1325 53 9 44 83.00 Yes - Hemignathus lucidus Bird R1km 629 629 14 615 97.00 Yes - Hemignathus parvus Bird R5km 1400 56 11 45 80.00 Yes - Houbaropsis bengalensis Bird R50km 250000 100 30 70 70.00 Yes - Hymenolaimus malacorhynchos Bird R10km 94500 945 222 723 76.00 Yes - Hymenolaimus malacorhynchos Bird R10km 106100 1061 325 736 69.00 Yes - Hypopyrrhus pyrohypogaster Bird R1km 431 431 45 386 89.00 Yes - Hypopyrrhus pyrohypogaster Bird R5km 1750 70 39 31 44.00 Yes - Hypopyrrhus pyrohypogaster Bird R5km 1900 76 46 30 39.00 Yes - Hypsipetes olivaceus Bird R5km 1850 74 6 68 91.00 Yes - Ibycter americanus Bird R50km 8027500 3211 3105 106 3.00 Yes - Jacamaralcyon tridactyla Bird R50km 410000 164 27 137 83.00 Yes - 88

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Laniisoma elegans Bird R50km 195000 78 68 10 12.00 Yes - Lathrotriccus griseipectus Bird R1km 72 72 46 26 36.00 Yes - Leptoptilos dubius Bird R50km 1830000 732 69 663 90.00 Yes - Leptotila ochraceiventris Bird R10km 36600 366 11 355 96.00 Yes - Leptotila wellsi Bird R1km 59 59 9 50 84.00 Yes - Lipaugus weberi Bird R1km 566 566 8 558 98.00 Yes - Lipaugus weberi Bird R1km 356 356 34 322 90.00 Yes - Macgregoria pulchra Bird R1km 950 950 444 506 53.00 Yes - Macgregoria pulchra Bird R1km 1102 1102 614 488 44.00 Yes - Meleagris ocellata Bird R50km 167500 67 59 8 11.00 Yes - Micrastur plumbeus Bird R10km 64400 644 625 19 2.00 Yes - Mohoua ochrocephala Bird R1km 835 835 297 538 64.00 Yes - Mohoua ochrocephala Bird R1km 221 221 106 115 52.00 Yes - Mohoua ochrocephala Bird R1km 410 410 208 202 49.00 Yes - Mohoua ochrocephala Bird R10km 15500 155 143 12 7.00 Yes - Mohoua ochrocephala Bird R1km 1147 1147 1071 76 6.00 Yes - Myadestes lanaiensis Bird R1km 309 309 18 291 94.00 Yes - Myadestes obscurus Bird R10km 6800 68 26 42 61.00 Yes - Myadestes palmeri Bird R1km 177 177 31 146 82.00 Yes - Myiarchus semirufus Bird R10km 29400 294 18 276 93.00 Yes - Myiarchus semirufus Bird R5km 2625 105 25 80 76.00 Yes - Myrmoborus melanurus Bird R10km 19400 194 182 12 6.00 Yes - Myrmotherula minor Bird R5km 4675 187 179 8 4.00 Yes - 89

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Myrmotherula unicolor Bird R10km 5500 55 39 16 29.00 Yes - Myrmotherula urosticta Bird R5km 3800 152 57 95 62.50 Yes - Nestor meridionalis Bird R10km 54500 545 445 100 18.00 Yes - Odontophorus hyperythrus Bird R10km 40300 403 13 390 96.00 Yes - Onychorhynchus occidentalis Bird R10km 27100 271 35 236 87.00 Yes - Oreomystis mana Bird R1km 441 441 385 56 12.00 Yes - Oreothraupis arremonops Bird R10km 5900 59 50 9 15.00 Yes - Pachycephala rufogularis Bird R10km 102500 1025 1013 12 1.00 Yes - Padda oryzivora Bird R50km 127500 51 37 14 27.00 Yes - Palmeria dolei Bird R1km 534 534 153 381 71.00 Yes - Paradisaea rudolphi Bird R5km 3225 129 91 38 29.00 Yes - Pardalotus quadragintus Bird R1km 992 992 72 920 92.00 Yes - Paroreomyza montana Bird R1km 1041 1041 253 788 75.00 Yes - Patagioenas inornata Bird R10km 106700 1067 89 978 91.00 Yes - Patagioenas inornata Bird R10km 74100 741 205 536 72.00 Yes - Patagioenas oenops Bird R5km 2675 107 61 46 42.00 Yes - Pauxi pauxi Bird R5km 1400 56 43 13 23.00 Yes - Pauxi pauxi Bird R10km 24000 240 226 14 5.00 Yes - Pauxi unicornis Bird R10km 51700 517 97 420 81.00 Yes - Pavo muticus Bird R50km 1782500 713 347 366 51.00 Yes - Penelope ortoni Bird R10km 77100 771 517 254 32.00 Yes - Penelope pileata Bird R50km 980000 392 386 6 1.00 Yes - Penelopides panini Bird R1km 333 333 10 323 96.00 Yes - 90

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Picoides borealis Bird R50km 270000 108 64 44 40.00 Yes - Picoides borealis Bird R50km 635000 254 181 73 28.00 Yes - Picumnus fulvescens Bird R10km 77400 774 767 7 0.90 Yes - Picus squamatus Bird R50km 300000 120 91 29 24.00 Yes - Pipile jacutinga Bird R50km 1087500 435 15 420 96.00 Yes - Pipile pipile Bird R1km 289 289 254 35 12.00 Yes - Podiceps taczanowskii Bird R1km 145 145 130 15 10.00 Yes - Poephila cincta Bird R50km 815000 326 194 132 40.00 Yes - Poospiza cinerea Bird R10km 19300 193 155 38 19.00 Yes - Porphyrio hochstetteri Bird R10km 5700 57 6 51 89.00 Yes - Primolius maracana Bird R50km 3815000 1526 1440 86 5.00 Yes - Procnias tricarunculatus Bird R10km 83800 838 704 134 15.00 Yes - Pseudonestor xanthophrys Bird R1km 142 142 121 21 14.00 Yes - Psittacula eques Bird R1km 1013 1013 57 956 94.00 Yes - Psophodes nigrogularis Bird R50km 195000 78 20 58 74.00 Yes - Pyrilia pyrilia Bird R50km 185000 74 52 22 29.00 Yes - Pyrrhura calliptera Bird R5km 2400 96 40 56 58.00 Yes - Rallus semiplumbeus Bird R1km 1205 1205 672 533 44.00 Yes - Rhaphidura sabini Bird R10km 22800 228 187 41 17.00 Yes - Saltator rufiventris Bird R10km 15100 151 134 17 11.00 Yes - Sarcoramphus papa Bird R50km 14222500 5689 5651 38 0.67 Yes - Saxicola macrorhynchus Bird R50km 722500 289 99 190 65.00 Yes - Sephanoides fernandensis Bird R1km 94 94 11 83 88.00 Yes - 91

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Sporophila caerulescens Bird R50km 6275000 2510 1324 1186 47.00 Yes - Sporophila frontalis Bird R10km 16500 165 156 9 5.00 Yes - Sporophila frontalis Bird R10km 60900 609 598 11 1.00 Yes - Starnoenas cyanocephala Bird R10km 106700 1067 282 785 73.00 Yes - Synallaxis tithys Bird R5km 2275 91 58 33 36.00 Yes - Synallaxis tithys Bird R5km 4700 188 135 53 28.00 Yes - Sypheotides indicus Bird R50km 2445000 978 160 818 83.00 Yes - Tachycineta euchrysea Bird R1km 945 945 54 891 94.00 Yes - Tangara peruviana Bird R10km 99700 997 255 742 74.00 Yes - Turdoides hindei Bird R10km 24200 242 140 102 42.00 Yes - Turdus lherminieri Bird R1km 100 100 27 73 73.00 Yes - Turdus lherminieri Bird R1km 689 689 423 266 38.00 Yes - Turdus swalesi Bird R1km 1196 1196 512 684 57.00 Yes - Turdus xanthorhynchus Bird R1km 145 145 83 62 42.00 Yes - Tyrannus cubensis Bird R10km 106700 1067 32 1035 97.00 Yes - Vireo atricapilla Bird R50km 452500 181 107 74 40.00 Yes - Vireo masteri Bird R1km 1018 1018 551 467 45.00 Yes - Vireo masteri Bird R5km 2000 80 52 28 35.00 Yes - Vireo masteri Bird R1km 307 307 223 84 27.00 Yes - Vireo masteri Bird R1km 1123 1123 1104 19 1.00 Yes - Xanthopsar flavus Bird R50km 557500 223 45 178 79.00 Yes - Xenospingus concolor Bird R1km 1000 1000 696 304 30.40 Yes - Xenospingus concolor Bird R1km 770 770 580 190 24.00 Yes - 92

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Zaratornis stresemanni Bird R10km 74000 740 733 7 0.95 Yes - Acerodon jubatus Mammal R10km 11300 113 7 106 93.00 Yes - Ammospermophilus nelsoni Mammal R5km 1525 61 14 47 77.00 Yes - Ammospermophilus nelsoni Mammal R10km 7000 70 20 50 71.00 Yes - Axis porcinus Mammal R10km 89000 890 107 783 87.00 Yes - Barbastella barbastellus Mammal R50km 3217500 1287 1268 19 1.00 Yes - Canis aureus Mammal R50km 26542500 10617 10591 26 0.24 Yes - Diceros bicornis Mammal R50km 4702500 1881 1569 312 16.00 Yes - Equus africanus Mammal R10km 79700 797 676 121 15.00 Yes - Helarctos malayanus Mammal R1km 1174 1174 68 1106 94.00 Yes - Helarctos malayanus Mammal R10km 11600 116 18 98 84.00 Yes - Helarctos malayanus Mammal R1km 301 301 67 234 77.00 Yes - Helarctos malayanus Mammal R1km 404 404 104 300 74.00 Yes - Helarctos malayanus Mammal R5km 2025 81 21 60 74.00 Yes - Helarctos malayanus Mammal R1km 940 940 328 612 65.00 Yes - Helarctos malayanus Mammal R50km 2035000 814 307 507 62.00 Yes - Helarctos malayanus Mammal R50km 435000 174 73 101 58.00 Yes - Helarctos malayanus Mammal R5km 1500 60 37 23 38.00 Yes - Helarctos malayanus Mammal R50km 730000 292 213 79 27.00 Yes - Helarctos malayanus Mammal R1km 764 764 608 156 20.00 Yes - Helarctos malayanus Mammal R5km 1600 64 52 12 18.75 Yes - Isoodon macrourus Mammal R50km 817500 327 273 54 16.00 Yes - Leptailurus serval Mammal R50km 12752500 5101 4984 117 2.00 Yes - 93

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Macaca arctoides Mammal R50km 1122500 449 441 8 1.00 Yes - Martes zibellina Mammal R10km 78100 781 638 143 18.00 Yes - Melursus ursinus Mammal R10km 42400 424 226 198 46.00 Yes - Melursus ursinus Mammal R50km 2727500 1091 592 499 45.00 Yes - Microtus cabrerae Mammal R10km 5600 56 19 37 66.00 Yes - Myrmecophaga tridactyla Mammal R50km 12920000 5168 5046 122 2.00 Yes - Nomascus leucogenys Mammal R10km 26500 265 20 245 92.00 Yes - Pecari tajacu Mammal R50km 15947500 6379 6148 231 3.00 Yes - Procolobus badius Mammal R50km 377500 151 116 35 23.00 Yes - Procolobus kirkii Mammal R5km 1525 61 29 32 52.00 Yes - Pseudomys desertor Mammal R50km 5172500 2069 1407 662 31.00 Yes - Tapirus bairdii Mammal R50km 835000 334 300 34 10.00 Yes - Tayassu pecari Mammal R50km 13572500 5429 4458 971 17.00 Yes - Trachypithecus geei Mammal R10km 7300 73 55 18 24.00 Yes - Ursus arctos Mammal R10km 78100 781 580 201 25.00 Yes - Ursus arctos Mammal R50km 21367500 8547 7700 847 9.00 Yes - Ursus thibetanus Mammal R10km 17600 176 7 169 96.00 Yes - Ursus thibetanus Mammal R10km 33800 338 70 268 79.00 Yes - Ursus thibetanus Mammal R10km 27000 270 65 205 75.00 Yes - Ursus thibetanus Mammal R50km 6750000 2700 1260 1440 53.00 Yes - Ursus thibetanus Mammal R50km 230000 92 52 40 43.00 Yes - Gallotia auaritae Reptile R1km 409 409 149 260 63.00 Yes - Phrynosoma cornutum Reptile R50km 1795000 718 710 8 1.00 Yes - 94

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Pituophis ruthveni Reptile R10km 74400 744 14 730 98.00 Yes - Agalychnis annae Amphibian R5km 3150 126 2 124 98.00 No Current Craugastor taurus Amphibian R10km 5900 59 1 58 98.00 No Current Lithobates sevosus Amphibian R10km 21600 216 1 215 99.00 No Current Acrocephalus rodericanus Bird R1km 114 114 4 110 96.00 No Current Agelaius xanthomus Bird R10km 9000 90 3 87 96.00 No Current Amazona vittata Bird R10km 8900 89 1 88 98.00 No Current Atlapetes flaviceps Bird R1km 77 77 3 74 96.00 No Current Atlapetes pallidiceps Bird R1km 1151 1151 1 1150 99.00 No Current Cacatua haematuropygia Bird R10km 94100 941 2 939 99.00 No Current Cacatua haematuropygia Bird R10km 25400 254 1 253 99.00 No Current Callaeas cinereus Bird R10km 5200 52 5 47 90.00 No Current Calyptophilus frugivorus Bird R5km 4475 179 1 178 99.00 No Current Carduelis cucullata Bird R10km 7500 75 4 71 94.00 No Current Cistothorus apolinari Bird R1km 310 310 4 306 98.00 No Current Conothraupis mesoleuca Bird R1km 312 312 1 311 99.00 No Current Coracina typica Bird R5km 1850 74 4 70 94.00 No Current Crax alberti Bird R10km 6500 65 5 60 92.00 No Current Crax blumenbachii Bird R50km 142500 57 1 56 98.00 No Current Curaeus forbesi Bird R10km 18300 183 1 182 99.00 No Current Cyanoramphus malherbi Bird R10km 17000 170 3 167 98.00 No Current Foudia flavicans Bird R1km 114 114 5 109 95.00 No Current Glaucis dohrnii Bird R10km 109300 1093 4 1089 99.00 No Current 95

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Grus leucogeranus Bird R50km 220000 88 3 85 96.00 No Current Gyps coprotheres Bird R50km 285000 114 3 111 97.00 No Current Hemignathus munroi Bird R10km 7800 78 1 77 98.00 No Current Herpsilochmus pectoralis Bird R5km 1325 53 1 52 98.00 No Current Loxops coccineus Bird R10km 8000 80 4 76 95.00 No Current Mergus octosetaceus Bird R50km 1275000 510 4 506 99.00 No Current Mohoua ochrocephala Bird R10km 6100 61 2 59 96.00 No Current Myrmeciza ruficauda Bird R10km 23200 232 1 231 99.00 No Current Nesoenas mayeri Bird R5km 1850 74 3 71 95.00 No Current Oreomystis bairdi Bird R5km 1275 51 4 47 92.00 No Current Paroreomyza maculata Bird R5km 1500 60 4 56 93.00 No Current Pomarea dimidiata Bird R1km 69 69 3 66 95.00 No Current Psittirostra psittacea Bird R10km 10400 104 3 101 97.00 No Current Ramphocinclus brachyurus Bird R1km 239 239 5 234 97.00 No Current Sporophila melanops Bird R1km 706 706 2 704 99.00 No Current Tachycineta euchrysea Bird R5km 3075 123 2 121 98.00 No Current Terenura sicki Bird R10km 13800 138 1 137 99.00 No Current Thripophaga macroura Bird R50km 145000 58 2 56 96.00 No Current Zosterops modestus Bird R1km 152 152 4 148 97.00 No Current Dendrolagus pulcherrimus Mammal R5km 2800 112 2 110 98.00 No Current Sminthopsis aitkeni Mammal R5km 1425 57 3 54 94.00 No Current Liuixalus romeri Amphibian R1km 75 75 73 2 2.00 No Extirpated Bolborhynchus ferrugineifrons Bird R5km 1575 63 62 1 1.00 No Extirpated 96

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Campylopterus phainopeplus Bird R5km 2825 113 111 2 1.00 No Extirpated Capito hypoleucus Bird R5km 2675 107 106 1 0.93 No Extirpated Chlorochrysa nitidissima Bird R10km 9300 93 90 3 3.00 No Extirpated Leptasthenura xenothorax Bird R5km 1525 61 59 2 3.00 No Extirpated Megaxenops parnaguae Bird R50km 667500 267 264 3 1.00 No Extirpated Nothoprocta taczanowskii Bird R10km 16300 163 160 3 1.00 No Extirpated Oreomystis mana Bird R1km 208 208 205 3 1.00 No Extirpated Oreophasis derbianus Bird R10km 6200 62 61 1 1.00 No Extirpated Platyrinchus leucoryphus Bird R50km 435000 174 172 2 1.00 No Extirpated Simoxenops striatus Bird R10km 9000 90 86 4 4.00 No Extirpated Simoxenops striatus Bird R10km 8000 80 76 4 5.00 No Extirpated Strix occidentalis Bird R50km 390000 156 151 5 3.21 No Extirpated Tangara fastuosa Bird R1km 128 128 127 1 0.78 No Extirpated Babyrousa celebensis Mammal R50km 127500 51 48 3 5.00 No Extirpated Ursus arctos Mammal R50km 2305000 922 920 2 0.22 No Extirpated Basileuterus basilicus Bird R1km 39 39 21 18 46.00 No Historical Basileuterus basilicus Bird R1km 11 11 10 1 9.00 No Historical Basileuterus basilicus Bird R1km 4 4 3 1 25.00 No Historical Bolborhynchus ferrugineifrons Bird R1km 43 43 26 17 39.00 No Historical Campylopterus ensipennis Bird R1km 39 39 28 11 28.00 No Historical Crypturellus erythropus Bird R1km 48 48 44 4 8.00 No Historical Crypturellus erythropus Bird R1km 8 8 5 3 37.50 No Historical Gallirallus sylvestris Bird R1km 13 13 1 12 92.00 No Historical 97

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Icterus oberi Bird R1km 26 26 9 17 65.00 No Historical Melamprosops phaeosoma Bird R1km 7 7 4 3 42.00 No Historical Otus insularis Bird R1km 46 46 38 8 17.00 No Historical Penelope barbata Bird R10km 5000 50 49 1 2.00 No Historical Toxostoma guttatum Bird R1km 27 27 15 12 44.00 No Historical Xenospiza baileyi Bird R1km 21 21 5 16 76.00 No Historical Xenospiza baileyi Bird R1km 5 5 1 4 80.00 No Historical Xenospiza baileyi Bird R1km 3 3 2 1 33.00 No Historical Zosterops albogularis Bird R1km 36 36 6 30 83.00 No Historical Craugastor punctariolus Amphibian R5km 3725 149 149 0 0.00 No Topology Hylomantis lemur Amphibian R10km 14600 146 75 72 49.00 No Topology Lithobates warszewitschii Amphibian R50km 162500 65 65 0 0.00 No Topology Aceros waldeni Bird R10km 12400 124 0 124 100.00 No Topology Amazilia castaneiventris Bird R5km 1950 78 78 9 11.00 No Topology Amazilia castaneiventris Bird R1km 859 859 859 73 8.00 No Topology Amazilia castaneiventris Bird R1km 154 154 154 154 100.00 No Topology Amazona barbadensis Bird R1km 442 442 442 442 100.00 No Topology Attila torridus Bird R5km 3075 123 123 0 0.00 No Topology Buteo ridgwayi Bird R1km 636 636 18 636 100.00 No Topology Cacatua haematuropygia Bird R10km 13000 130 0 130 100.00 No Topology Cacatua haematuropygia Bird R10km 12600 126 0 126 100.00 No Topology Cacatua haematuropygia Bird R10km 7300 73 0 73 100.00 No Topology Campephilus imperialis Bird R50km 355000 142 0 142 100.00 No Topology 98

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Carpornis melanocephala Bird R10km 7700 77 0 77 100.00 No Topology Carpornis melanocephala Bird R1km 27 27 0 27 100.00 No Topology Cephalopterus penduliger Bird R10km 69600 696 691 7 1.00 No Topology Compsospiza baeri Bird R10km 20100 201 201 0 0.00 No Topology Corvus palmarum Bird R10km 30400 304 243 73 24.00 No Topology Crypturellus erythropus Bird R1km 76 76 1 76 100.00 No Topology Crypturellus kerriae Bird R10km 5900 59 59 0 0.00 No Topology Cyanopsitta spixii Bird R50km 125000 50 0 50 100.00 No Topology Ducula mindorensis Bird R5km 2150 86 48 84 97.00 No Topology Ducula mindorensis Bird R1km 112 112 34 111 99.00 No Topology Eurochelidon sirintarae Bird R5km 2125 85 0 85 100.00 No Topology Euscarthmus rufomarginatus Bird R5km 5000 200 0 200 100.00 No Topology Foudia rubra Bird R5km 1850 74 0 74 100.00 No Topology Gyps coprotheres Bird R10km 19700 197 70 136 69.00 No Topology Hapalopsittaca amazonina Bird R10km 5200 52 10 45 86.00 No Topology Heliodoxa gularis Bird R10km 23600 236 230 36 15.00 No Topology Hemignathus lucidus Bird R5km 1325 53 0 53 100.00 No Topology Hemitriccus kaempferi Bird R10km 7700 77 77 1 1.30 No Topology Hylocryptus erythrocephalus Bird R10km 13900 139 139 0 0.00 No Topology Junco hyemalis Bird R1km 294 294 121 224 76.00 No Topology Lathrotriccus griseipectus Bird R10km 64600 646 135 518 80.00 No Topology Loxioides bailleui Bird R5km 1325 53 7 47 88.00 No Topology Macroagelaius subalaris Bird R5km 2875 115 1 115 100.00 No Topology 99

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Myrmotherula grisea Bird R10km 22800 228 228 0 0.00 No Topology Myrmotherula unicolor Bird R10km 16600 166 0 166 100.00 No Topology Neochmia ruficauda Bird R50km 515000 206 148 96 46.00 No Topology Neochmia ruficauda Bird R10km 9400 94 75 48 51.00 No Topology Numenius borealis Bird R50km 510000 204 0 204 100.00 No Topology Odontophorus strophium Bird R5km 1650 66 17 56 84.00 No Topology Onychorhynchus coronatus Bird R10km 48200 482 462 243 50.00 No Topology Oreomystis mana Bird R1km 92 92 92 0 0.00 No Topology Oreomystis mana Bird R1km 9 9 9 0 0.00 No Topology Penelopides mindorensis Bird R10km 8400 84 7 84 100.00 No Topology Phylloscartes paulista Bird R50km 490000 196 195 12 6.00 No Topology Phylloscartes roquettei Bird R50km 180000 72 72 0 0.00 No Topology Phytotoma raimondii Bird R10km 6900 69 0 69 100.00 No Topology Phytotoma raimondii Bird R5km 1650 66 0 66 100.00 No Topology Psittirostra psittacea Bird R5km 1400 56 0 56 100.00 No Topology Quelea erythrops Bird R1km 145 145 145 145 100.00 No Topology Ramphocinclus brachyurus Bird R1km 1136 1136 19 1136 100.00 No Topology Sarcoramphus papa Bird R10km 33800 338 0 338 100.00 No Topology Siphonorhis americana Bird R1km 30 30 0 30 100.00 No Topology Siptornopsis hypochondriaca Bird R10km 7300 73 73 0 0.00 No Topology Sporophila nigrorufa Bird R50km 192500 77 77 0 0.00 No Topology Sylvia melanocephala Bird R50km 997500 399 399 0 0.00 No Topology Syndactyla ruficollis Bird R10km 6500 65 65 0 0.00 No Topology 100

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Species Class Raster Historical Historical Current Extirpated Contraction Included Reason group fragment cells cells cells (%) for (km2) Exclusion

Tangara fastuosa Bird R10km 14900 149 149 0 0.00 No Topology Taphrolesbia griseiventris Bird R5km 4200 168 168 0 0.00 No Topology Turdoides bicolor Bird R10km 9600 96 96 0 0.00 No Topology Turdus swalesi Bird R1km 3 3 3 3 100.00 No Topology Tympanuchus cupido Bird R50km 3150000 1260 153 1127 89.00 No Topology Tympanuchus pallidicinctus Bird R50km 455000 182 27 182 100.00 No Topology Vanellus gregarius Bird R50km 9947500 3979 3286 846 21.00 No Topology Vermivora bachmanii Bird R50km 797500 319 0 319 100.00 No Topology Xipholena atropurpurea Bird R10km 14800 148 0 148 100.00 No Topology Zosterops chloronothus Bird R5km 1850 74 3 72 97.00 No Topology Blastocerus dichotomus Mammal R50km 4082500 1633 128 1508 92.00 No Topology Lynx pardinus Mammal R5km 4475 179 0 179 100.00 No Topology Lynx pardinus Mammal R1km 156 156 52 113 72.00 No Topology Mustela lutreola Mammal R50km 2207500 883 150 735 83.00 No Topology Myotis nattereri Mammal R50km 152500 61 61 0 0.00 No Topology Panthera pardus Mammal R5km 1525 61 61 61 100.00 No Topology Tapirus terrestris Mammal R50km 13020000 5208 4505 711 13.00 No Topology Tapirus terrestris Mammal R1km 11 11 11 10 90.00 No Topology Tapirus terrestris Mammal R1km 10 10 10 7 70.00 No Topology Tapirus terrestris Mammal R1km 7 7 7 6 85.00 No Topology

101

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Appendix 1.2. Supplementary methods

Distributions with contraction Distribution maps of terrestrial vertebrates with observed range contraction were obtained from the International Union for Conservation of Nature (International Union for Conservation of Nature 2010a). From all terrestrial vertebrate species with available spatial data (N=23,743), we selected areas defined by the IUCN as native in origin and in the presence categories of extant, probably extant, possibly extinct, or extinct (Based in IUCN, maps of the historical range of a species are a combination of polygons coded as Extant, Probably Extant, Possibly Extinct, and Extint, IUCN 2010b). Original presence categories of the IUCN were reclassified into current (extant and probably extant) and extirpated (extinct and probably extinct), defining the historical range as the sum of both, and defining a species with range contraction as a species with categories of current and extirpated (N=628, Table A1.1). Since most studied distributions are fragmented and complexly shaped we considered two different scales to describe the empirical data. In the range scale we characterized the complete historical distribution range of each species; these ranges often include multiple fragments separated by unoccupied matrix. In the fragment scale we identified all individual fragments (i.e., continuous areas) of the historical range with observed contraction (N=430, Table A1.2). These historical fragments were originally continuous but as contraction progresses these continuous areas could have become divided into multiple current and/or extirpated fragments. We used an equal area projection (Cylindrical Equal Area projection) in ArcMap 9.3 (ESRI 2008), and rasterized, for each distribution, the current areas and the extirpated areas, using Idrisi Taiga (Clark Labs 2009). The use of raster format for range distributions is common in biodiversity studies at range scale, including others that analyzed range contraction (Ceballos and Ehrlich 2002, Yackulic et al. 2011), as it facilitates the management of spatial calculations in large datasets. To minimize the loss of information during the rasterization process, we applied four different grid cell resolutions (e.g., 1, 5, 10, 50 km cell side) defined based on the total area of the historical distribution of the range/fragment in the vector format (Table A1.6). From the 628 rasterized ranges and 430 rasterized fragments with contraction we eliminated ranges/fragments with detected topological errors (i.e., overlapping of mutually exclusive classes of presence: current and extirpated) and distributions with: <51 cells of historical distribution, <6 cells classified as current, and/or <6 cells classified as extirpated. The final dataset included 374 ranges and 273 fragments with contraction (Tables A1.1 and A1.2 and Fig. A1.2) representing 386 different species. These rasterized distributions had a minimum Nrange grids cells =53, minimum Nfragment grids cells =51 (Tables A1.7 and A1.8). Additional descriptive information for these data is provided in Table A1.3 and A1.4 and Fig. A1.1.

Calculation of geodetic distances The best distance-preserving projections available at a global scale have a root-mean- square logarithmic distance error of σ ≈ 0.340 (Gott et al. 2007). These errors are biased with 102

1. Toward multifactorial null models of range contraction in terrestrial vertebrates latitude and/or longitude (Mena Berrios 2008) and could affect our estimates of distances. Therefore, we calculated the spatial position of each grid cell center over the spheroid of the Earth (using the spheroid World Geodetic System 1984, WGS84), and then calculated the distance between each pair of grid cell centers as the geodetic distance, that is the distance between two unprojected points on the Earth's surface. Geodetic distances were calculated in Visual Fox Pro applying the Vicenty’s algorithm (Vincenty 1975) over the spheroid WGS84. Estimates generated using this algorithm agreed to within 0.115 mm for distances between 10 to 18,000 km (Thomas and Featherstone 2005) and thus, provide accurate estimates of varying distances at global scales.

Calculation of geodetic angles Similar to distances, angles are also distorted in projected maps resulting in errors that could bias our results and conclusions. To avoid this problem, we first calculated the linear vector between each extirpated grid cell and the closest current cell (vector d) following the method for calculation of geodetic distances described above. Second, from the spatial position over the spheroid of the Earth (using the spheroid World Geodetic System 1984, WGS84) of each extirpated grid cell center, we calculated the angle of contraction as the geodetic angle, the azimuth of the direction of vector d. The azimuth was calculated in Visual Fox Pro applying the Vicenty’s algorithm (Vincenty 1975) over the spheroid WGS84.

Calculation of land use We reclassified the original categories of the MODIS (MCD12Q1) Land Cover Product (Oak Ridge National Laboratory Distributed Active Archive Center 2010) to define the variable Land use used in our analyses (Table A1.5). Original data at 1km grid cell resolution were projected into Cylindrical Equal Area with ArcMap 9.3 (ESRI 2008) using the same registration point as distributions rasterized at 1km (x and y coordinates, in the output space, used for pixel alignment) to obtain the alignment of pixels between the land use map and the distributions. Rasterized land use values at a resolution of 1km2 were converted into the resolutions used in our analyses (Table A1.5) estimating for each pixel the percentage of land use from the terrestrial surface of each pixel.

103

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Table A1.3. Descriptive statistics for the distributions used at range scale. We report the mean, standard deviation (SD), median and range (minimum and maximum) of the total area (in km2) and the number of fragments for the historical, current and extirpated ranges, as well as the percentage of contraction (% Contraction) and the percentage of fragments that are defined as extirpated (% Extirpated fragments). Estimates Historical Current Extirpated % Contraction % Extirpated fragments Area Fragments Area Fragments Area Fragments Mean 1,194,127 6.7 1,075,282 5.5 118,845 3.7 40.80 26.0 SD 4,751,117 8.1 4,640,053 7.3 397,116 5.0 32.75 25.9 Median 24,000 4 7,900 3 4,100 2 32.16 18.4 Range 54-51,455,000 1-96 7-51,085,000 1-96 6-3,762,500 1-42 0.14-99.03 0-95

Table A1.4. Descriptive statistics for the distributions used at fragment scale. For this scale the units of analysis were historical fragments, continuous areas of the historical range, which may become fragmented as extirpation progressed leading to multiple current and/or extirpated fragments within the original continuous historical fragments. We report the percentage of contraction (% Contraction) and the mean, standard deviation (SD), median and range (minimum and maximum) of the historical, current, and extirpated areas (in km2), and the resulting number of current and extirpated fragments observed within historical fragments. Estimates Historical Current Extirpated % Contraction Area Area Fragments Area Fragments Mean 45,614 30,775 2.5 14,839 2.4 50.67 SD 115,810.58 108,641 3.1 26,304 3.4 32.52 Median 9,000 2,600 1 1,800 1 51.47 Range 59-1,061,700 7-1,059,100 1-29 8-144,000 1-24 0.24-98.74

104

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Table A1.5. Reclassification of the original categories of the MODIS (MCD12Q1) Land Cover Product (Oak Ridge National Laboratory Distributed Active Archive Center 2010) used to define the land use (1 for land use, 0 for non land use) variable used in our analyses. Original categories Original code Land use Water 0 No data Evergreen needleleaf forest 1 0 Evergreen broadleaf forest 2 0 Deciduous needleleaf forest 3 0 Deciduous broadleaf forest 4 0 Mixed forest 5 0 Closed shrublands 6 0 Open shrublands 7 0 Woody savannas 8 0 Savannas 9 0 Grasslands 10 0 Permanent wetlands 11 0 Croplands 12 1 Urban and built-up lands 13 1 Cropland/natural vegetation mosaics 14 1 Snow and ice 15 0 Barren 16 0

Table A1.6. Description of the groups used for rasterization of distributions (used for both range and fragment scales). Group Historical distribution area (km2) Cell area (km2) Min Max R1km 54 1250 1 R5km 1,250 5,000 25 R10km 5,000 125,000 100 R50km 125,000 51,455,000 2,500

105

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Table A1.7. Descriptive statistics of grid cells for the distributions at range scale for each rasterized group. We report report the mean, standard deviation (SD), median and range (minimum and maximum) of the total area (in number of grids cells) and the number of fragments for the historical, current and extirpated ranges, as well as the percentage of contraction (% Contraction) and the percentage of fragments that have been extirpated (% Ext. fragments). Group Historical Current Extirpated % Contraction % Ext. fragments Cells Fragments Cells Fragments Cells Fragments

R1km Mean 526.0 4.375 279.8 3.4 246.2 2.8 48.0 30.1 SD 329.98 4.03 290.85 4.08 261.5 4.50 33.2 25.53 Median 469.5 3 146 2 138.5 1 40.9 33.33 Range 54-1,113 1-19 7-1,051 1-21 6-956 1-35 2.5-98.7 0-80 R5km Mean 105.8 5.3 63.3 4.2 42.6 2.5 45.5 25.8 SD 43.93 4.67 47.17 4.15 27.13 2.39 27.14 23.72 Median 99 4 53 3 38 2 46.6 25 Range 53-197 1-19 6-190 1-19 7-115 1-15 3.55-93.22 0-98.4 R10km Mean 434.4 7.9 242.8 5.6 191.5 4.4 42.8 30.0 SD 345.33 7.29 284.7 4.87 267.72 5.33 34.58 27.52 Median 311 6 126 4 57 3 33.8 25.0 Range 54-1,258 1-44 6-1,156 1-30 6-1,085 1-42 0.82-99.03 0-95.4 R50km Mean 1,953.1 7.3 1,769.5 7.6 183.5 3.81 28.8 15.8 SD 3,498.18 12.05 3,471 12.23 284.38 5.65 29.11 21.55 Median 450 4 275.5 4 61 2 19.3 Range 54-20,582 1-101 9-20,434 1-106 6-1,505 1-31 0.14-97.5 0-95 All scales Mean 772.7 592.8 179.8 SD 1,857.46 1,835.78 258.15 Median 295 126 61.5 Range 53-20,582 6-20,434 6-1,505

106

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Table A1.8. Descriptive statistics in grid cells for the distributions at fragment scale for each rasterized group. We report report the mean, standard deviation (SD), median and range (minimum and maximum) of the total area (in number of grids cells) and the number of fragments for the historical, current and extirpated ranges, as well as the percentage of contraction (% Contraction). Group Historical Current Extirpated % Contraction

Cells Cells Fragments Cells Fragments

R1km Mean 603.5 251.2 1.9 352.3 2.3 55.5 SD 371.37 261.71 2.15 311.20 3.47 31.89 Median 571 149 1 263 1 59 Range 59-1,244 7-1,104 1-13 8.0-1,106 1-24 1.7-98.7 R5km Mean 96.0 49.0 1.8 47.1 1.5 51.27 SD 42.06 38.73 1.47 30.05 1.30 24.04 Median 81 39 1 39 1 51.47 Range 53-188 6-179 1-8 8-145 1-9 4.28-94.77 R10km Mean 446.4 176.3 2.8 270.1 2.3 55.87 SD 337.72 216.35 2.85 307.27 2.55 35.70 Median 313 90 2 134 1 71.43 Range 55-1,067 6-1,013 1-17 7-1,035 1-15 0.90-98.12 R50km Mean 1,307.6 1,085.4 3.0 222.1 3.3 36.34 SD 2,187.11 2,111.51 4.69 307.42 4.69 29.42 Median 327 189 1 86 1 28.74 Range 51-10,617 13-10,591 1-29 6-1,440 1-22 0.24-96.55 All scales Mean 625.1 379.1 245.9 SD 1,136.63 1,079.92 298.61 Median 301 91 101 Range 51-10,617 6-10,591 6-1,440

107

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Table A1.9. Results from separate regression analyses for each rasterization group (R1km, R5km, R10, and R50km). Estimates are from GLMM used to evaluate the three main null models of range contraction (demographic, contagion and refuge) and two combined models that incorporate multiple processes. We report model AIC, ΔAIC (difference in AIC with the best model comparing all five models). Analysis at range scale Analysis at fragment scale Model Model comparison Model Model comparison AIC ΔAIC AIC ΔAIC R1km (n=91, 46 species) R1km (n=82, 31 species) Combined_2 131.39 0.00 Combined_1 109.42 0.00 Mod_Contagion 132.58 1.19 Combined_2 112.84 3.42 Mod_Refuge 133.25 1.85 Mod_Demographic 115.03 5.61 Combined_1 134.12 2.72 Mod_Refuge 116.19 6.77 Mod_Demographic 134.78 3.39 Mod_Contagion 118.76 9.34 R5km (n=86, 43 species) R5km (n=76, 34 species) Combined_2 116.02 0.00 Combined_1 93.17 0.00 Combined_1 116.84 0.81 Mod_Refuge 95.70 2.53 Mod_Refuge 118.99 2.96 Combined_2 96.92 3.75 Mod_Contagion 120.23 4.21 Mod_Demographic 108.61 15.44 Mod_Demographic 122.44 6.42 Mod_Contagion 109.84 16.67 R10km (n=176, 88 species) R10km (n=116, 54 species) Combined_2 234.90 0.00 Combined_1 157.96 0.00 Mod_Refuge 237.88 2.98 Mod_Refuge 159.01 1.04 Combined_1 239.55 4.65 Combined_2 159.12 1.15 Mod_Contagion 250.87 15.97 Mod_Demographic 169.05 11.09 Mod_Demographic 253.99 19.09 Mod_Contagion 171.41 13.44 R50km (n=104, 52 species) R50km (n=88, 40 species) Mod_Contagion 149.39 0.00 Mod_Contagion 120.71 0.00 Combined_2 150.92 1.53 Combined_2 122.39 1.68 Mod_Refuge 152.77 3.38 Mod_Demographic 123.82 3.11 Mod_Demographic 152.94 3.55 Combined_1 125.23 4.52 Combined_1 153.73 4.34 Mod_Refuge 130.53 9.83

108

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Figure A1.1. Distribution frequencies of the area of historical distributions (a, b), the area of extirpated distributions (c, d), the area of current distributions (e, f) and the percentage of contraction (g, h). Left panels represent data at the range scale, and right panels at the fragment scale.

109

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Fig. A1.2. Diagram describing the steps used to process the IUCN spatial data for our analyses 110

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

111

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Fig. A1.3. Representative distributions with simulated range contraction patterns to validate the Centrality Index and Directionallity Index. Panels a, c, and f reflects patterns following the demographic null model; b, d, e, g, and h for the contagion null model. Color gradient indicates the pattern of contraction with light areas representing initial stages of contraction (areas were initial extirpations occur) and dark areas final stages. 112

1. Toward multifactorial null models of range contraction in terrestrial vertebrates

Fig. A1.4. Empirical indices values as as function of the percentage of contraction at the range (a, c, e) and fragment scale (b, d, f). Blue circles representing estimates from birds, red triangles represent mammals, green vertical crosses (+) represent amphibians, and green diagonal crosses (x) represent reptiles. Horizontal lines represent reference indices value

113

114

Chapter 2: Size matters, and so does spatial configuration: predicting vulnerability to extinction in vertebrates

115

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Lucas, P.M; González-Suárez, M; Revilla, E. Size matters, and so does spatial configuration: predicting vulnerability to extinction in vertebrates (In prep)

116

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates Abstract

Species extinctions caused by anthropogenic activities are widespread and there are global efforts aimed to halt this loss of biodiversity. Preventing future losses with effective conservation policies requires a good understanding of what factors influence species’ extinction risk. The degree of vulnerability to extinction varies among taxa and is influenced by intrinsic species’ traits and extrinsic factors including anthropogenic impacts. The most consistently recognized and widely available predictor of a species’ extinction risk is the area of its distribution range. However, distribution ranges of equal area can have diverse spatial configurations that reflect varying spatial patterns in threatening processes and in species’ characteristics (e.g., dispersal) which likely influence extinction risk. The aim of this study is to assess if and how the spatial configuration of the distribution range influences vulnerability to extinction evaluating the relative importance of descriptors of shape and fragmentation (fragment number and size heterogeneity). We analyzed global distribution maps and conservation status data obtained from the IUCN for 11,160 species of non-marine amphibians, birds and mammals. Our results confirm that the most important predictor of vulnerability to extinction is range area, but also identify fragment heterogeneity and shape as valuable predictors. We detect complex relationships, revealed by multiple interaction terms, which demonstrate that simple rules of thumb to associate vulnerability to range spatial configuration are likely to be inadequate. Our study shows the value of explicitly considering descriptors of spatial configuration beyond area as additional indicators of threating processes and vulnerability among vertebrates. These metrics are relatively easy to define (although values are sensitive to data quality), and unlikely other correlates of extinction risk are readily available for many species (all of those with distribution range maps). Considering these additional spatial configuration predictors improves our capacity to predict vulnerability and thus, could promote more effective conservation.

Keywords: border, colonization, distribution, expansion, fragmentation, habitat lost, heterogeneity, range

117

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates Introduction

Anthropogenic activities are causing the loss of many populations and species with an important reduction in natural, economic and social capital (CBD 2010). Estimations suggest that current rates of extinction are 3-4 orders of magnitude higher than natural rates (Barnosky et al. 2011). Approximately 20% of extant vertebrate species are classified as Threatened by the International Union for the Conservation of Nature (IUCN), but there are taxonomic differences with amphibians being the most affected (41%), birds the least (13%), and mammals and reptiles having intermediate values (25% and 22% respectively, Hoffmann et al. 2010). Current trends predict more prevalent and intense human activities in the future which would likely lead to further extinctions (Hurtt et al. 2011, Moss et al. 2010, Pereira et al. 2010). As a result, there is growing concern regarding how to achieve a significant reduction in future biodiversity loss (CBD 2010, Sala et al. 2000). Understanding what makes some taxa or species more prone to extinction is key to facilitate better allocation of limited resources and to develop more effective conservation management actions (Safi and Pettorelli 2010). The best predictor of extinction risk is probably the total number of (reproducing) individuals in a population (Jones and Diamond 1976). Everything else being equal, the more individuals in a population, the lower its risk of extinction (David et al. 2003). However, obtaining good estimates of abundance can be complicated. As a result, estimates of total population size are available for relatively few species and areas. To overcome this limitation, many studies have searched for correlates of extinction risk, including morphological, ecological, and behavioral species’ traits (Cardillo et al. 2008, Davidson et al. 2009, Fritz et al. 2009, González-Suárez et al. 2013, González-Suárez and Revilla 2013, Purvis et al. 2000). Among these correlates, probably the most widely accepted is the size of the geographic distribution range (henceforward range). Everything else being equal, wider ranges can host more individuals, and thus are associated with lower risk of extinction (Cardillo et al. 2008, Cardillo et al. 2005, Gaston 1994a, Gaston and Fuller 2009, Orzechowski et al. 2015, Runge et al. 2015). Although, range size can also be difficult to estimate accurately (Gaston 2003), distribution range maps (IUCN 2010) from which

118

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates approximate range sizes can be calculated are currently available for many more species that abundance estimates. While total range area has been identified as a key correlate of extinction risk, at the local/population scale, theoretical and empirical research has shown that in addition to the amount of area (habitat) available, the spatial configuration of the habitat also influences vulnerability (Bascompte and Solé 1998, David Tilman and Kareiva 1997, Hanski 1999, Levins 1969, MacArthur and Wilson 1967). Landscapes are heterogeneous spaces with varying degrees of habitat suitability (Forman 1995, Richard T. T. Forman and Godron 1986). Habitat suitability also varies within an occupied fragment between the border, where it is usually lower, and the core areas, where tends to be higher (Bascompte and Solé 1995). Therefore, for a given area the shape of the fragment is important. Fragments with a greater ratio of border to core have lower habitat suitability and less carrying capacity, than more compact or circular fragments. The viability of spatially structured populations is also influenced by the degree of fragmentation, size and number of fragments, of the available habitat. The probability of (re)colonization of a suitable but unoccupied fragment depends on the number of immigrants arriving from neighboring occupied areas (Hanski and Gyllenberg 1997), which, in turn, depends on the population size of these potential donors (which is positively correlated with their area) and the distance between donor and recipient fragments (the further away, the lower the chance for dispersers to arrive. (Gyllenberg and Hanski 1992, Hanski and Gyllenberg 1993). In general, there are three main characteristics that determine the direct effect of fragmentation on extinction probability: the number of fragments, the connectivity between fragments, and the causes of local extinctions (Bascompte and Solé 1995, Hanski and Ovaskainen 2000, Quinn and Hastings 1987). Because there are many possible interactions among these characteristics, the question of what is the best spatial configuration to reduce extinction risk has long been a central debate in conservation biology, often referred to as the Single Large or Several Small (SLOSS) problem (Diamond 1975, Ovaskainen 2002, Simberloff and Abele 1976, Wright and Hubbell 1983). The SLOSS problem is relevant in the design of protected areas when aiming to maximize the number of species, their metapopulation capacity and genetic variation (Ovaskainen 2002). If extinction risk is driven by demographic stochasticity a higher number

119

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates of individuals will reduce risk (Narendra S. Goel and Richter-Dyn 1974), thus a larger fragment would be preferable to several small areas (Quinn and Hastings 1987). However, if the primarily causes of extinction are environmental stochastic processes, even large populations are vulnerable to extinction, thus multiple fragments (with independent environmental risks) would reduce the overall risk (Quinn and Hastings 1987). Connectivity among fragments allows for rescue effects and also needs to be consider; if fragments are not connected a single large area is preferable to several small (Pelletier 2000). Finally, heterogeneity in fragment size may also influence vulnerability. When heterogeneity is large, with one fragment much larger than the rest, vulnerability is mostly determined by the probability of extinction of this largest fragment, and larger fragments are less likely to become extinct (Hanski et al. 1996). However, if threatening impacts concentrate on that larger patch the risk could be greater with high heterogeneity than if similarly sized fragments (exposed to different risks) existed. Complete distribution ranges also show diverse spatial configurations, often including multiple fragments of varying sizes, located at different distances, and with diverse shapes that differ in their border to area ratios (Brown 1995, Channell and Lomolino 2000a, b, Gaston 1990, Gaston 1994a, Gaston 2003, Gaston 2008, 2009, Lawton 1993). Some of this variation reflects differences in geographic conditions and species’ traits (dispersal abilities or habitat specialization). Additionally, variation in spatial configuration can reflect effects of human impacts, such as changes in land use or climate change, which can cause local extinctions leading to area loss, changes in shape and fragmentation, and altered patterns of dispersion and colonization (e.g., invasive species). The relevance of some aspects of spatial configuration is already recognized by conservation organizations like the IUCN which considers as criteria for classification as threatened not only the total range area but also the number of populations and their level of isolation (IUCN 2012). However, comparative studies of extinction risk have focused on the role of range area as predictor of vulnerability (Cardillo et al. 2008, Cardillo et al. 2005, Fritz et al. 2009, Purvis et al. 2000) and the relative importance of other descriptors of spatial configuration at a range scale remains largely unexplored. Here, we address this knowledge gap by exploring how distinct spatial aspects of a species’ range affect its vulnerability to extinction. In particular, we evaluate the influence of

120

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates the shape and fragmentation (defined by fragment number and size heterogeneity) for non- marine vertebrate species (amphibians, birds, and mammals). Although potentially important, we could not consider fragment isolation because it is largely driven by dispersal abilities which are not well-described yet likely distinct for the studied taxa (so generalizations are likely to be inaccurate). We expect that, for a given area, vulnerability will be higher in ranges with more fragments, higher border to area ratios (irregular shapes), and with more homogeneous (equally sized) fragments (Fig. 2.1). We also expect these effects of spatial configuration to be particularly relevant for species with small ranges because they presumably have smaller population sizes which are more susceptible to extinction (Hanski 1999).

Figure 2.1. Description of the studied spatial configuration variables with illustrative examples of values, predicted association with increased vulnerability to extinction, and their definition. *Note that threats acting on borders can increase circularity.

Material and methods

Data Spatial distribution range data were download from the International Union for Conservation of Nature (IUCN 2010) for all available terrestrial amphibians, birds, and mammals. We used areas of the range defined by the IUCN as native or reintroduced in origin, extant or probably extant presence, and resident, breeding season, and non-breeding season in seasonality. We 121

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates projected the distributions into Winkel triplel projection, which aims to minimize the three kinds of distortions: area, direction and distance). Polygons of less than 0.1 km2 were excluded to eliminate potential mapping errors assuming these very small areas are unlikely to represent sustainable isolated fragments (populations). From each range map we calculated three variables: fragment shape (Circularity), number of fragments (N_frag), and fragment size heterogeneity (Heterogeneity; Fig. 2.1). We also calculated the total range size (Area). To better evaluate the role of fragmentation we limited our analyses to ranges with >1 distinct fragments (minimum required to estimate Heterogeneity), but results were generally consistent considering all species (results not shown). To define vulnerability to extinction we used two different metrics obtained from the IUCN (IUCN 2015). First, we used the Red List status categories, an ordinal variable describing vulnerability to extinction with levels Least Concern, Near Threatened, Vulnerable, Endangered and Critically Endangered. Because we used species with current distributions only, no species in our data were classified as Extinct in the Wild or Extinct. Second, we used the Population Trend category, which is also an ordinal variable with decreasing, stable, or increasing levels. Species classified as Data Deficient status or Unknown population trend were not included in our analyses.

Analyses We defined regression models that aimed to predict vulnerability to extinction (Red List Status or Population Trend) as a function of area, shape, number of fragments, and fragment size heterogeneity. Because our objective was to asses if additional descriptors of spatial configuration improve predictability, we used as our null model a regression including Area as the single predictor. This null model was compared with increasingly complex models that incorporated the other variables describing shape and fragmentation (Table 2.1) using an information theoretic approach based on AIC (Burnham and Anderson 2002).Because we hypothesized that spatial configuration may have different effects depending on range size we also defined models including interaction terms of shape (Circularity) and/or fragmentation (N_frag and Heterogeneity) with Area. We fitted separate models for each taxonomic class because of their distinct characteristics in dispersal and life-

122

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Table 2.1. Results of the GLMM analyses aimed to predict vulnerability to extinction in a species as a function of distinct descriptors of the spatial configuration of its distribution range. We report AIC, ΔAIC (difference in AIC with the null model Area. More negative ΔAIC represent stronger support). Models in bold are the best supported within each category (the best supported were the simplest model within two AIC of the model with lowest AIC). Models in bold and underlined indicate the best overall model. Variable Heterogeneity is indicated as Het. Model ΔAIC (AIC) Red List status Population trend Amphibians Birds Mammals Amphibians Birds Mammals (n=1498) (n=7215) (n=2447) (n= 1665) (n=6968) (n=1835) Size Area (Null model) 0.00 0.00 0.00 0.00 0.00 0.00 (1634.15) (6990.79) (2862.92) (1939.49) (11473.40) (2381.48) Size and Shape (Circularity) Area+Circularity -16.99 0.59 0.40 -19.08 -9.18 -7.79 Area*Circularity -20.20 -28.24 2.40 -31.78 -67.98 -11.01 Size and Fragmentation Area+N_frag -3.59 1.92 1.67 -9.30 2.00 -0.02 Area*N_frag -3.04 3.00 2.71 -7.31 -5.91 -0.23 Area+Het 1.29 -9.68 -3.13 1.40 -32.18 -0.84 Area*Het -10.76 -27.67 -11.96 0.64 -38.94 -5.60 Area*Het+ Area*N_frag -19.43 -24.99 -12.66 -12.28 -43.54 -9.84 Area*Het+ Area*N_frag+ Het*N_frag -19.03 -29.59 -11.18 -11.13 -49.60 -10.35 Size, Fragmentation and Shape Area* Circularity+Area*Het+Area*N_frag -33.06 ─ ─ -33.40 ─ -15.97 Area*Circularity+Area*Het+Area*N_frag+N_frag*Het ─ -49.15 ─ ─ -100.74 ─

123

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates history. To avoid circularity in our analyses, we excluded all species classified as threatened by the IUCN based on criterion B (small geographic range or area of occupancy, possibly fragmented). Models were fitted as multivariate GLMM with cumulative logits for ordered multinomial data and random intercepts using the function clmm package in R (R Development Core Team 2013). Models included taxonomic information (order, family, and genus) as random factors to control for evolutionary non-independence of the observations following González-Suárez and Revilla (2013).

Results

The final database (excluding single-fragment distributions and species classified based on B- criteria) included data for 11,160 species with valid Red List Status (55.58% of the studied taxa) and 10,468 species with valid Population Trend categories (47.30%. Tables S3 and S4). Initial descriptive analyses of these data show that species with more vulnerable Red List status and decreasing population trends generally have smaller ranges, with more circular shapes and possibly fewer, more evenly-sized fragments (Figs. 2.2, 2.3).

Spatial Configuration and Red List Status Models that included descriptors of shape and/or fragmentation were identified as improvements over the null (Area only models) based on AIC for all taxa, although the particular descriptors included in the best model varied among groups (Fig. 2.4, Table 2.1). For example, in mammals, the best model did not include any descriptors of shape, but revealed an effect of fragmentation. In particular, for mammals having distinctly-sized fragments reduced the risk in larger ranges, yet for small ranges having similarly-sized fragments was slightly better (Fig. 2.4, Table 2.2). For birds and amphibians both shape and fragmentation were revealed as important. Among amphibians, more threatened species generally had ranges with more circular shapes and fewer fragments. Also as found for mammals, having distinctly-sized fragments reduced the risk in larger ranges but slightly increased it in smaller ones. Among birds, more circular shapes, particularly for larger

124

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates ranges, were also associated with slightly higher risk, and when many fragments existed distinctly-sized fragments generally reduced risk. As expected, in all taxa generally larger ranges were associated with lower risk (Fig. 2.4, Table 2.2).

Spatial Configuration and Population trend Analyses based on Population Trend also support the importance of additional spatial configuration descriptors (Table 2.1). Descriptors of both shape and fragmentation were included in the best models for all taxa and were largely consistent with those based on status. Generally, greater heterogeneity in more fragmented areas and more irregular shapes reduced risk (Fig. 2.5, Table 2.2). In contrast to results based on status effects were generally more noticeable for larger ranges especially among mammals.

125

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Figure 2.2. Frequency distributions for the variables Area, Circularity, N_frag and Heterogeneity for multi-fragment distributions of amphibians, mammals and birds (Tables A2.3 and A2.4). Light grey show non-threatened species (Least Concern and Near Threatened status), medium grey indicates threatened (Vulnerable, Threatened and Critically Endangered) species classified based on criterion B (not included in our regression analyses) and dark grey all other threatened species (analyzed in this study).

126

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Figure 2.3. Frequency distributions for the variables Area, Circularity, N_frag and Heterogeneity for multi-fragment distributions of amphibians, mammals and birds (Tables A2.3 and A2.4). Light grey show species with an increasing population trend (generally few species and thus, hardly visible), medium grey indicates species with stable population trend and dark grey represents species with decreasing population trend.

127

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Table 2.2. Coefficient estimates and SE of the best GLMM predicting vulnerability to extinction in a species as a function of distinct descriptors of the spatial configuration of its distribution range. Model selection results are shown in table 1. We modeled the probability of increase in the Red List Status and Population Trend. A dash (-) indicates variables not included in the best model. Sample sizes (n) indicate the number of species included in each model. Variables β (SE) Red List category β (SE) Population trend Amphibians Birds Mammals Amphibians Birds Mammals (n= 1498) (n=7215) (n=2447) (n= 1665) (n=6968) (n=1835) Area -0.60 (0.617) -1.71 (0.232) -1.17 (0.188) -1.32 (0.450) -1.02 (0.174) -0.95 (0.398) Circularity 0.37 (1.914) -4.71 (1.132) - -2.53 (1.467) -5.20 (0.854) -2.19 (1.739) Heterogeneity 4.47 (1.925) 1.76 (0.784) 2.73 (1.073) -0.28 (1.132) 0.06 (0.669) 1.46 (1.643) Nfrag -3.39 (1.509) 0.01 (0.330) - -2.33 (1.017) -5.20 (0.854) -2.04 (0.881) Area* Circularity 0.38 (0.437) 1.06 (0.229) - 0.96 (0.325) 1.11 (0.160) 0.61 (0.319) Area* Nfrag 0.52 (0.304) 0.09 (0.056) - 0.33 (0.187) 0.04 (0.040) 0.29 (0.140) Area* Heterogeneity -1.22 (0.435) -0.29 (0.164) -0.65 (0.213) -0.06 (0.319) 0.02 (0.131) -0.41 (0.292) Nfrag* Heterogeneity - -0.68 (0.253) - - -0.54 (0.170) -

128

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Figure 2.4. Predicted marginal probabilities for each Red List status, with dark green for Least Concern (LC), light green for Near Threatened (NT), yellow for Vulnerable (VU), orange for Endangered (EN) and red for Critically Endangered species (CR), based on the best models with descriptors of spatial configuration for each class (coefficients in table 2). Note that threat categories are underestimated for some plots, or even not represented, which is due to the relatively small number of species in these categories. We used combinations of values along the 129

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates range of the variables to show the pattern of the interactions. For amphibians, small area refers to ranges of 1,995 km2 (Percentile 10), big areas refers to ranges of 706,651 km2 (Percentile 80); few fragments refers to ranges of 2 fragments (Percentile 20) and many fragments refers to ranges of 8 fragments (Percentile 80); irregular shapes refers to ranges of 0.257 of Circularity (Percentile 10) and regular to shapes to ranges of 0.828 (Percentile 90). For birds, small area refers to ranges of 17,568 km2 (Percentile 10), big areas refers to ranges of 4,099,218 km2 (Percentile 80); few fragments refers to ranges of 4 fragments (Percentile 20) and many fragments refers to ranges of 155 fragments (Percentile 80); irregular shapes refers to ranges of 0.264 of Circularity (Percentile 10) and regular to shapes to ranges of 0.637 (Percentile 90). For mammals, small area refers to ranges of 7051 km2 (Percentile 5), big areas refers to ranges of 3,593,317 km2 (Percentile 80).

130

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Figure 2.5. Predicted marginal probabilities for each category of population trend, with dark green for Increasing Population (IP), light green for Stable Population (SP) and red for Decreasing Population (DP), based on the best models with descriptors of spatial configuration for each class (coefficients in table 2). Note that category IP is usually not represented in the plots, which is due to the small number of species in that category. We used combinations of values along the range

131

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates of the variables to show the pattern of the interactions. For amphibians, small area refers to ranges of 1,995 km2 (Percentile 10), big areas refers to ranges of 706,651 km2 (Percentile 80); few fragments refers to ranges of 2 fragments (Percentile 20) and many fragments refers to ranges of 8 fragments (Percentile 80); irregular shapes refers to ranges of 0.257 of Circularity (Percentile 10) and regular to shapes to ranges of 0.828 (Percentile 90). For birds, small area refers to ranges of 17,568 km2 (Percentile 10), big areas refers to ranges of 4,099,218 km2 (Percentile 80); few fragments refers to ranges of 4 fragments (Percentile 20) and many fragments refers to ranges of 155 fragments (Percentile 80); irregular shapes refers to ranges of 0.264 of Circularity (Percentile 10) and regular to shapes to ranges of 0.637 (Percentile 90). For mammals, small area refers to ranges of 7051 km2 (Percentile 5), big areas refers to ranges of 3,593,317 km2 (Percentile 80).

Discussion

Our results showed that range area is negatively associated with vulnerability to extinction but that considering different descriptors of the spatial configuration of distribution range, namely shape, number of fragments, and fragment size heterogeneity results in improved models predicting vulnerability. In summary we found that distributions with high fragment size heterogeneity and fewer irregularly shaped fragments are less vulnerable.

Capturing spatial complexity in distribution range data Our results show that understanding the spatial complexity in distribution ranges can be useful to evaluate vulnerability to extinction. Previous comparative analyses have largely described distribution ranges considering only the area and thus, have assumed that there is limited or no variation in other aspects of range spatial configuration, or that if variation exists that it is largely irrelevant to understand vulnerability to extinction. However, these assumptions greatly simplify distribution range dynamics and the processes that influence spatial configuration and extinction. The processes that define range boundaries are naturally complex, varying temporally and spatially and can cause heterogeneity that results in fragments of different sizes and shapes for which the use of a single descriptor may be inadequate (Gaston 2003, Lucas et al. 2016). In addition, human impacts have altered the spatial configuration of distributions (Crees et al. 2016, Turvey et al. 2015, Turvey et al. 2016), so the inclusion of diverse spatial descriptors can be helpful to understand processes and patterns related to range dynamic including extinction risk. 132

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

Range area and vulnerability Our results showed that the area of the range is negatively associated with vulnerability to extinction as previously reported by other studies (Bielby et al. 2008, Davidson et al. 2009, Giam et al. 2011, Harris and Pimm 2008). Range area is generally positively associated with abundance (Gaston 1994a, Gaston 2003) which in turn, reduces extinction risk. Small ranges may increase vulnerability directly via lower abundance but can also signal the existence of ongoing threats which can be more worrisome than low abundance alone (Channell and Lomolino 2000a, b, Lomolino and Channell 1995, 1998, Lucas et al. 2016).

Size heterogeneity The manner in which the total range area was distributed among the different fragments (Heterogeneity) was a relevant predictor for all classes and both measures of vulnerability with an effect that often depended on total range size. As predicted, high size heterogeneity was generally associated with lower vulnerability, with weaker effects for small ranges perhaps because size heterogeneity is more limited when ranges are smaller (very large fragments are not possible). Additionally in birds, the association of high heterogeneity with lower vulnerability was more important for distributions with many fragments (for which the potential for higher heterogeneity may be greater) and relatively weak for distributions with few fragments. As mentioned in the introduction size heterogeneity in which one large fragment concentrates most of the existing area can reduce vulnerability because large fragments are less likely to become extinct (assuming risk is not concentrated on that area). Additionally, increased vulnerability in ranges with more homogeneously sized fragments may be explained as a consequence of range contraction and expansion dynamics. Processes of range contraction, associated to vulnerability, usually lead to range fragmentation (Rodriguez and Delibes 1992, Rodrıguez and Delibes 2003, Rodríguez and Delibes 2002) and fragmentation causes homogenization in fragment size because there is a minimum viable fragment size that limits the range of possible sizes. During contraction, fragments split into smaller fragments, thus reducing maximum fragment size. However, minimum fragment size is constrained because very small fragments cannot support viable populations, thus contraction, associated with vulnerability, may lead to more homogeneously-sized areas (Rodrıguez and Delibes 2003). On the other hand, for a species expanding its range (lower

133

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates vulnerability) small fragments can be maintained through immigration from the bigger fragments and initially separated fragments may be connected by newly established populations generating greater size disparity (Lieury et al. 2016). Nevertheless, at range scale, we should not expect fragments to be connected by dispersal (that process should occur more at population scales).

The shape of the range and vulnerability Threatened amphibians, birds and mammals are characterized by more circular shapes and the effect is more important for species with big ranges. This contradicts our expectations based on metapopulation and island biogeography theory, which postulates that higher border to core ratio should increase vulnerability. This result may be explained if circular shapes are actually the result of higher vulnerability rather than a cause. Through the process of range contraction, local extinctions produce several changes in the spatial configuration of ranges, e.g. fragmentation and/or shape changes. The resultant spatial configuration is context specific, being determined by the distribution of impacts and species abundance, but in general border areas are more prone to be extirpated (Brown 1995, Lawton 1993). Thus, in initially irregularly shaped ranges, contraction may lead to increased circularity in shape by removing the most irregular fragments and increasing the circularity of those that remain as border areas become extirpated (Mehlman 1997, Smale and Wernberg 2013). This intriguing hypothesis cannot, unfortunately, be tested with our data which represent static distribution maps.

The number of fragments of the range and vulnerability The association between the number of fragments and vulnerability was important for the three taxonomic classes, being more relevant for models based in population trend. In general and contrary to our predictions, less vulnerable species had more fragmented ranges, with a more marked effect for those with small ranges. As discussed regarding size heterogeneity above, it is possible that the number of fragments may reflect a consequence of contraction rather than a driver. Range fragmentation is common among species suffering range contraction (Turvey et al. 2015), but contraction also leads to the extirpation of small fragments so that the total number of fragments may not actually increase significally. For

134

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates example, Rodriguez and Delibes (Rodríguez and Delibes 2002) showed that the Iberian lynx distribution suffered an important range contraction in which the largest fragments were fragmented but also small populations were lost such that at the end the total number of populations/fragments only increased slightly. At the other extreme, species showing an increase in its populations (or non-threatened species) may have a higher rate of inmigration which will result in the colonization of new areas and in a more fragmented distribution,even small fragments may be supported by migration.

Conclusion Most species distribution ranges are spatially complex, often formed by multiple fragments with different shapes which also change over time (Gaston 2003, Wilson et al. 2004). Using additional range descriptors beyond size provides more informative predictive models of vulnerability and reveals interesting hypothesis regarding the spatial consequences of range expansion and contraction. Whereas in population biology it is widely accepted that spatial complexity affects extinction probability, to our knowledge this is the first time that these relationships have been quantified at biogeographical scales. Our selected variables have a clear ecological basis, are simple to calculate, can be used for studies at different scales, and only require range distribution maps. In spite of the uncertainty associated with definition of species ranges and their heterogeneous quality (Hurlbert and Jetz 2007), their availability for most terrestrial vertebrates makes them a suitable source of information to be used in vulnerability assessments. Future work considering how species’ traits, distinct threatening processes, and local environmental conditions may affect these patterns would be necessary but our study shows we have much to gain from considering more than just size.

Acknowledgements

This work was funded by the Spanish Ministry of Economy and Competitiveness (CGL2009- 07301⁄BOS and CGL2012-35931/BOS co-funded by FEDER, and the FPI grant BES-2010- 034151), by the European Community’s Seventh Framework Programme (FP7⁄2007-2013) under grant agreement no 235897, and by a Juan de la Cierva post-doctoral fellowship (JCI-

135

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates

2011-09158). We also acknowledge funding from the Spanish Severo Ochoa Program (SEV- 2012-0262).

136

2. Size matters, and so does spatial configuration predicting vulnerability to extinction in vertebrates Supplementary material

Table A2.1. Number of species after each filter by class. Described in the IUCN database refers to the number of species that include the IUCN database as species; IUCN spatial database refers to the number of species which are included in the spatial database of the IUCN (Include Extinct species); Presence and seasonality refers to the filters applied to these categories, for Presence we selected Extant and Probably Extant and for Seasonality we selected all categories in Amphibians and Mammals and “Resident”, “Breeding Season”, and “Non-Breeding Season” for Birds. Systems refers to the number of species selected after select species living only in terrestrial systems for Birds, for Mammals we directly selected the spatial information which include only terrestrial species as defined by the IUCN. Size of the fragments refers to the selected species after exclude fragments of less than 0,1 km2. Multifragment refers to the number of species after select only species with a minimum of two fragment in its distribution. And not categorized by B criteria refers to the number of listed species in the IUCN Red List excluding species categorized by B criteria.

Described Spatial Categories Systems Min. Multifrag. Excl. by B data frag. criteria; Pop. trend Amphibians 6414 6312 6277 - 6276 3207 1498; 1665 Birds 10425 10424 10249 9400 9347 7516 7215; 6968 Mammals 5408 5292 5269 - 5268 3903 2447; 1835

Table A2.2. Number of threatened (indicating those listed based on B-criteria) and non-threatened species with multi-fragment distributions (Multifragment column in Table S3).

Red List criteria Population trend Threat Not Decreasing Stable Increasing threat B Excluding B criteria criteria Amphibians 578 100 1398 923 723 19 Birds 301 502 6713 3187 3311 470 Mammals 262 340 2107 960 826 49

137

138

Chapter 3: The roles of climate change and land use in recent terrestrial vertebrate range contractions

139

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Lucas, P.M; González-Suárez, M; Revilla, E. The roles of climate change and land use in recent terrestrial vertebrate range contractions (In prep) 140

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions Abstract

Climate change and land use are the main drivers threatening biodiversity. Many studies have studied these aspects separately, but little is known about the combined effects of both impacts even though most species are likely affected by both. This limited understanding of the mechanisms leading to range contraction reduces our ability to generate accurate predictions and thus, make effective management decision. To address this knowledge gap we simultaneously study the roles of climate change and land use in observed range contractions for 335 terrestrial vertebrate species, disentangling the relative importance of each of these drivers and exploring if different drivers lead to different detectable changes in the environmental characteristics of the remaining range areas. Our regression analyses show that observed biodiversity loss is best explained by considering both land use and climate change, although some species seem to be primarily affected by a single driver, and for many contractions were not clearly explained by either threat. Species apparently affected by both impacts show the most severe contractions. The environmental characteristics of the remaining range differ for species affected by land use and climate change. Current areas are located in poleward, steeper, higher altitude areas for species affected by climate changes and a combination of both impacts. For species affected by land use changes current areas tend to be located closer to the equator and be slightly steeper. Our results highlight the importance of considering diverse impacts to evaluate and predict biodiversity loss and highlight how humans have changed the environmental characteristics of many species ranges, which influences our ability to define niches and habitat preferences.

Keywords: climate change, land use, extinction, conservation, impacts, global

Introduction

Climate change and human land use are the two main drivers of the current biodiversity crisis (Hoffmann et al. 2010, Parmesan and Yohe 2003, Pounds et al. 2006, Sekercioglu et al. 2008, Steffen et al. 2015b, Thomas et al. 2004, Thuiller et al. 2005a, Thuiller et al. 2005b). Climate change comprises changes in the mean and variability of weather patterns, including a

141

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions general increase in temperature, and changes in the variability of climatic events, such as more frequent and longer droughts (IPCC 2015). Species’ responses to these climatic challenges include changes in phenology (Butt et al. 2015, Hurlbert and Liang 2012, Lane et al. 2012), distribution range shifts (Devictor et al. 2008, Kerr et al. 2015, Tingley et al. 2009), altered migration patterns (Pilfold et al. 2016), and changes in abundance (Hinder et al. 2012, Johnston et al. 2013). Although currently climate change may appear as a key threat to relatively few species (Ameztegui et al. 2016, Maxwell et al. 2016, Parmesan and Yohe 2003,

Root et al. 2003), the first mammal extinction due climate change was recently reported (Gynther et al. 2016) and future impacts are predicted to be very large (Parmesan and Yohe 2003). On the other hand, land use is currently the most important impact on the biodiversity (Hoffmann et al. 2010, Maxwell et al. 2016) due to direct habitat loss and/or fragmentation, but also due to its association with other human impacts such as species overexploitation, pollution and invasive species (Newbold et al. 2015). Impacts related to land use can reduce species richness and abundance at local scale (Newbold et al. 2015) leading to changes in the current distribution of species (Kehoe et al. 2015, Marco and Santini 2015) and increasing extinction risk (Pekin and Pijanowski 2012). Although climate change and land use are both globally important threats, there are differences in the spatial distribution of their impacts, generating observable patterns in the loss of biodiversity. Observed climate changes have been spatially heterogeneous, with higher temperature increases recorded in the northern hemisphere (Armour et al. 2016, IPCC 2014). Similarly, climate change affects organisms differently within an area. Populations near their maximum thermal limit are very vulnerable to temperature increases, while those close to their minimum thermal limits can even be positively affected by climate change (Gray et al. 2015). Climate change has been linked to local extinction of populations in the areas of their distribution ranges that are closer to the Equator (Hill et al. 1999, Parmesan et al. 1999) and at lower altitudes (Menéndez and González‐Megías 2014, Merrill et al. 2008, Morueta-Holme et al. 2015, Sekercioglu et al. 2008, Shoo et al. 2005, Tingley et al. 2009). Human-induced land use changes have already occurred in more than 75% of Earth’s ice-free surface (Ellis and Ramankutty 2008, Ramankutty et al. 2008, Watson et al. 2016) but the magnitude of these changes also shows important spatial heterogeneity. Some areas of the world have been intensively used by humans for millennia (e.g., West Europe or the Indian subcontinent), whereas human impacts are more recent and localized in others (e.g., 142

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Amazonian basin). Human land-use has been associated with the extinction of populations in flatter, lower elevation areas (Keppel et al. 2012) (Ellis and Ramankutty 2008, Parmesan and Yohe 2003, Ramankutty et al. 2008). Humans tend to occupy subtropical and temperate lowland areas with relatively low roughness, which a priori are also the most affected by climate change (Ellis and Ramankutty 2008, Parmesan and Yohe 2003, Ramankutty et al. 2008). For an effective conservation policy it is mandatory to correctly attitbute range shifts to the relevant driver, either climate change, land use or their simultaneous impact. To do so it is neccesary to include proxies of both impacts in analyses (Oliver and Morecroft 2014). However, up to date most studies have investigated climate change and land use impacts in isolation (Randin et al. 2009) or have focused on specific taxa and/or regions (Kerr et al. 2015, Oliver et al. 2015, Oliver and Morecroft 2014), and the doubt remains if we are correctly attributing range shifts to the relevant environmental drivers and on how prevalent is the simultaneous effect of both climate change and land use. The current combination of rapid climate change and intense land use modifications may be contributing to an amplification of their impact on biodiversity (Root et al. 2003, Sala et al. 2000, Scharlemann et al. 2004). Species may shift their ranges to track climate changes, but rapid changes combined with impacts caused by human land use may limit and even impede this tracking, leading to stronger range contractions (Brook et al. 2008). The spatial configuration of these drivers may suggest that there are areas which are acting as refuges providing protection from disturbances, even if they are environmentally suboptimal (Maclean et al. 2016, Turvey et al. 2015). Refuge areas have been relevant for species persistence during past climatic changes (Bennett and Provan 2008, Tzedakis et al. 2002), maintaining populations that acted as sources in subsequent expansions (Jackson and Betancourt 2009). Identifying the existence and the mechanisms associated to the maintenance of current refuge area is critical to understand the ongoing loss of biodiversity. Here we present the first, in our knowledge, global analysis of a broad taxonomic group that simultaneously evaluates the importance of land use and climate change in recent range contractions. Our goals are to understand the relative and combined importance of each driver and whether unique identifiable patterns of range contraction can be associated with each impact at global scales. To address these goals we analyze range contraction data for 335 terrestrial vertebrate species distributed worldwide. Because climate and land use changes are 143

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions key factors behind biodiversity loss, we hypothesize that patterns of range contraction should be generally better explained by their combined impact. However, because human-induced climate change is a more recent threat, we expect that land use changes should be a stronger predictor of observed contraction. Finally, we expect that the spatiotemporal heterogeneity of both threats would lead to the existence of identifiable areas acting as refuges for species showing range contraction.

Material and methods

Spatial distribution data We initially downloaded available global distribution data for 628 terrestrial vertebrates (International Union for Conservation of Nature 2010a) with quantified range contraction. These species had distributions that included extirpated areas where the species had been present in the past but is no longer found, and current areas where the species is currently present (following the notation of the International Union for Conservation of Nature 2010; detailed information is provided in Appendix A3.1). All spatial data were projected into an equal area projection (Cylindrical Equal Area) and rasterized using a cell size of 10 x 10 km. For analyses, we selected species with no topological errors in their distribution ranges (overlapping of mutually exclusive extirpated and current areas) and with a minimum of five cells for both extirpated and current areas (n = 344). We excluded a few additional species for which data on some of the variables describing drivers’ impacts were not available resulting in a final dataset with 335 distribution ranges. See Appendix A3.2 for additional information on data preparation. Tables A3.2, and A3.6- A3.9 and Figures A3.1 and A3.2 provide descriptive summaries of these data including total area in km2, the percentage of contraction (calculated as the percentage of the historical range area classified as extirpated) and the distribution of historical, current and extirpated ranges of species across continents and biomes.

Climate change impact Although climate change shows general trends, like a global increase between 1901-2008 of 0.07ºC/decade in land surface temperatures, climatic changes are spatially heterogeneous

144

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

(IPCC 2014). For example, during that period Alaska experienced an increase of 0.13ºC/decade but in the Amazon region the change was 0.04ºC/decade (Harris et al. 2014, IPCC 2014). To account for these spatial variability in climate change we used the Time- Series (TS) Version 3.21 provided by the Climate Research Unit (CRU TS v. 3.21 downloaded from https://crudata.uea.ac.uk, Harris et al. 2014) at a resolution of 30 arcminutes. The data describe month-by-month variations in climate during the last century. We focused on temperature as main descriptor of climate change because it is one of the main determinants of species’ niches and is the climate variable with the clearer observed tendency. We downscaled the data to 5 arcminutes (10 km) applying the dry adiabatic lapse rate (9.8 K km-1) where the original temperature from the coarse cell is corrected by the elevation of the downscaled cells using a digital elevation model form (downloaded from http://www.earthenv.org/DEM.html, Robinson et al. 2014). The IPCC defines climate as the mean and variability of weather variables over a period of time, which can range from months to millions of years (IPCC 2014). We used the classical period of 30 years adopted from the World Meteorological Organization focusing on two periods: Baseline Climate from 1901 to 1931 and Current Climate from 1982 to 2012 (the time in which the strongest shifts in climate have been reported, (IPCC 2014). As mentioned above climate change can affect populations and species differently. Species have different thermal niches representing the range of temperatures where they can live, reproduce and maintain required biological interactions (Willis and Bhagwat 2009). Areas of a distribution range near the thermal niche should be more prone to extinction as small changes in temperature can lead to unsuitable conditions. To understand how different species in different areas were affected by detected climate changes we conservatively defined thermal niches for each species as follows. First, for each grid cell of the historical range we calculated the average baseline temperature for each month using data from the Baseline Climate period as well as the current monthly temperature based on the Current Climate. We then defined proxies of upper thermal limits based on five quantiles per month (0.75, 0.80, 0.85, 0.90, 0.95) obtained from the distribution of baseline temperatures generating five separate threshold variables. For each grid cell of the historical range we then determined for each month if the current temperature exceeded the upper thermal limit based on the different quantiles. We then calculated the average (over all months) proportion of cells that exceeded the limits in the extirpated and current range for each quantile (used in the 145

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions global analysis described below). These values range from 0 when no grid in any month exceeded the thermal limit, to 1 when all grids exceeded the thermal limit every month. For the species models (see below) we calculated the average number of months in which thermal limits were exceeding for each grid cell. Values range from 0 when the thermal limit was not exceeded in any month, to 1, when the limit was exceeded in all 12 months. We expect that if local extinction is associated with climate change, extirpated areas should have a greater proportion of cells in which current temperatures exceeded the upper thermal limits.

Land use impact We used two proxies of human land use impact, one based on the current land cover types linked with human activity and another based on human density. We defined land cover based on the 1-km resolution MODIS (MCD12Q1) Land Cover Product (Oak Ridge National Laboratory Distributed Active Archive Center 2010). We aggregated the information over cells of 10-km resolution defining the proportion of human land cover per cell (for species models). For the global analysis we calculated the proportion of cells which correspond to human land uses in current and extirpated areas (additional details in Appendix A3.2 and Table A3.3). To define human land use impact based on human density we used the gridded population of the world (Center for International Earth Science Information Network - CIESIN - Columbia University 2005) and calculated human density over cells of 10-km resolution, as the number of humans/km2 (for species models) and the human density (humans/km2) in extirpated and current areas for global analysis. If human land uses influence local extinction, we expect that extirpated areas should have a greater proportion of human use and/or higher human density.

Statistical analyses Global models First, to identify the best descriptors for each impact, we fitted single-impact regression models describing the probability of extirpation of an area based on descriptors of climate change and land use impacts separately. We selected the best descriptor for each impact based on Akaike Information Criterion, AIC values (Burnham and Anderson 2002). Using the selected descriptors (those in the model with lowest AIC) we defined additive-effect and multiplicative-effect multi-impact models. We used generalized linear mixed regression 146

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions models (GLMM) with family binomial and a logit link using the glmer function from the lme4 package in R (R Development Core Team 2013). All models included taxonomic class, order, and family as random factors to control for evolutionary non-independence of the observations (González-Suárez and Revilla 2013). We compared single-impact and multi- impact models using an information theoretic approach based on AIC.

Species models Results from the global model could be difficult to interpret if different species are affected by different impacts or affected differently by the same impact. Therefore, we fitted individual species models predicting the probability of extirpation in each cell for species with at least 40 historical range cells (Table A3.1). We defined four models for each species: single-impact (using the best indicators for land use and climate change identified in the global analysis), additive-effect multi-impact, and a null model (only including random effects). We did not test the more complex multiplicative-effect multi-impact model based on results from the global analysis (see below) and to avoid overfitting in species with small ranges (few data points). We used GLMM with family binomial and a logit link and included a random variable describing the aggregation of 2x2 neighboring cells to account for spatial autocorrelation (Yackulic et al. 2011; Appendix A3.2, Tables A3.4, A3.5 and Figure A3.3) . We classified species as being affected by climate change when (1) the single-impact climate model was identified as best (lowest AIC) and the estimated coefficient indicated higher extirpation probability in cells that had more often exceeded the upper thermal thresholds; or (2) when the additive-effect multi-impact was the best model and the climate estimated coefficient indicated higher extirpation associated to exceeding the upper thermal thresholds but the land use coefficient did not support higher extirpation in cells with more human use or higher density. Similarly, species were classified as affected by land use when (1) the single- impact land use model was identified as best and the coefficients indicated higher extirpation probability in cells with more human use or higher density; or (2) the additive-effect multi- impact model was identified as best and the land use coefficient indicated higher extirpation probability in cells with more human use or higher density but the climate coefficient did not indicate higher extirpation in cells that had more often exceeded the upper thermal thresholds. We classified species affected by both drivers when the best model was the multi-impact

147

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions additive-effect model and coefficients supported higher extirpation probability associated with exceeding upper thermal thresholds and more human uses/density.

Environmental characteristics of current areas To characterize refuges or areas where species persist we calculated the change in distance to the Equator, elevation, and slope comparing current and extirpated areas. For each rasterized distribution range, we first estimated the geodetic distance to the Equator for the center of each grid cell (additional details in Appendix A3.2). We then calculated the arithmetic mean distance within each area (Dis_Equator_ext representing extirpated areas and Dis_Equator_curr current areas). We defined the ΔDis_Equator = Dis_Equator_curr - Dis_Equator_ext for each species range; ΔDis_Equator > 0 if species contracts polewards (extinction more likely near the equator). We used a Digital Elevation Model of the World at a scale of 100 meters (downloaded from http://www.earthenv.org/DEM.html, Robinson et al. 2014) to calculate the arithmetic mean elevation within extirpated and current areas (Elevation_ext and Elevation_curr, respectively). For each range, we defined the ΔElevation = Elevation_curr - Elevation_ext. ΔElevation > 0 occur when the species remains in more elevated areas. We used the same Digital Elevation Model of the World, first we calculated the slope over the original map at 100x100m resolution and then we used it to calculate the arithmetic mean slope within extirpated and current areas (Slope_ext and Slope_curr, respectively). Using these values we defined the ΔSlope = Slope_curr - Slope_ext for each range. ΔSlope >0 if a species remains in the areas with higher slope. We described changes in environmental variables ΔDis_Equator, ΔElevation, and ΔSlope for species classified as affected by climate change, land use and both drivers using representative plots and linear regression models fitted with the lm function from stats package in R.

Results

Our analyses included 335 species (~ 1.4% of the terrestrial vertebrates listed in the IUCN) distributed across all biomes and all continents except Antarctica (Tables A3.6-A3.9). All

148

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions taxa were included but birds were overrepresented: 234 birds, 70 mammals, 19 amphibians and 12 reptiles. Distribution ranges vary in size (1.400 to 62.356.700 km2) and exhibited a mean contraction of 34% of their historical range.

Global models Results from the single-impact regression analyses identified the 0.75 upper thermal threshold as the best descriptor of climate change, although 0.80 and 0.85 were also supported thresholds (Table A3.13) and land cover as the best descriptor of land use (Table A3.14). Both multi-impact models received more supported than the single-impact models with the simpler additive-effects model identified as best (Table 3.1). As expected, extirpation was positively associated with more human land uses and areas where thermal limits had been exceeded more often, i.e., more impacted by climate change.

Species models From the 277 species with sufficient data to fit individual models, 15 failed to show model convergence. From the remaining 262 species (Tables 3.2, A3.1, A3.5 and A3.10) the null model was identified as the best for nearly half (48.09%), while the two single and additive models were supported in similar proportions (Tables 3.2, A3.14). Support for a given model did not necessarily indicated the predicted effect was observed, but species for which the land use model was supported were more likely to model coefficients suggesting the expected effect than those species for which climate change was the best model (Table 3.2). Overall, we found that range contraction was positively associated with climate change for 31 (11.83 %) species, with human land use for 49 (18.70%), and with both impacts for 24 (9.16%). For the majority of species (60.31%) we found no clear predictors of contraction patterns (supporting a null model) or patterns than contradicted the expected (higher persistence in areas with more human land use or that more often exceeded thermal thresholds). These results were not qualitatively different when we used more restrictive criteria to define support from model coefficients (i.e., accounting for parameter uncertainty, Table A3.10). Species classified in the different impact groups have similar taxonomic representation and historical range areas, but differ in Red List status (IUCN 2015) and observed percentage of range contraction (Figures 3.1, A3.4). Species for which the best model was the null generally showed a low percentage of range contraction, those associated 149

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions to single impacts showed intermediate levels of range contraction, while species associated with both drivers showed a high percentage of contraction and were more likely to be in higher risk Red List categories (Figure 3.1).

150

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table 3.1. Results from the global analyses based on regression models (GLMM) to evaluate the effect of each impact from single-impact models and the effect of multi-impact models (with additive effect and with interaction effect between impacts) over range contraction of 335 terrestrial vertebrates. We also include a null model (a GLMM with only the intercept and the random effects) for compare with models based in impacts. We report model coefficients (best estimates and their SE), AIC, ΔAIC (difference in AIC with the best of all five models, indicated in bold), and ΔAICum (difference in AIC comparing with the best single-impact model for each impact). Dashes indicate variables not included in the model. Model Coefficients Model comparison

Land use Climate Change Land use* Climate Change AIC ΔAIC ΔAICsm

Mod_ Additive 1.67 (0.377)* 1.47 (0.368)* - 892.05 0.00 Mod_Interaction 1.37 (0.744)† 1.27 (0.570)* 0.71 (1.556) 893.84 1.79 Mod_Use 2.05 (0.363)* - - 906.63 14.58 0.00 Mod_Clim - 1.89 (0.356)* - 910.25 18.21 3.62 Mod_Null - - - 938.82 46.77 14.68 *P < 0.05; †P < 0.10; P < 1

Table 3.2. Summary of the results from the species models based on regression models (GLMM, Table A2) to evaluate the importance within species of each descriptor of impact and the effect of combined impacts. We report the number of species (and percentage it represent of total) for which convergence was achieved for each tested model. For each model we then list the number (and percentage) of species for which estimated coefficients supported the prediction. We also include the total number of species (and percentage) classified by impact. Dashes indicate prediction not supported in the model. Model; N=262 Species better fit to each model Classification (supporting the prediction) Climate Land use Additive Null Mod_Clim 43 (16.41%) 24 (9.16%) - - - Mod_Use 44 (16.79%) - 34 (12.98%) - - Mod_ Additive 49 (18.70%) 7 (2.67%) 15 (5.76%) 24 (9.16%) - Mod_Null 126 (48.09%) - - - 126 (48.09%) TOTAL 262 (100.00%) 31 (11.83%) 49 (18.70%) 24 (9.16%) 126 (48.09%)

151

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Figure 3.1. The distribution for the percentage of contraction (a), the IUCN red list category (b) and the taxonomy (c) for each group of associated impact based in species analysis. Ends of the whiskers represent the lowest datum still within the 1.5 interquartile range (IQR) of the lower quartile, and the highest datum still within the 1.5 IQR of the upper quartile (Tukey boxplot).

Environmental characteristics of current areas Species with contraction patterns associated to climate change showed: (1) range contraction towards higher latitudes particularly marked at intermediate-higher latitudes (mean ± SD ΔDis_Equator = 370 ± 652.32 km. Figure 3.2 and Table A3.11); (2) contraction towards higher altitude mainly at low latitudes (ΔElevation = 159 ± 278.95 m); and (3) weak tendency for contraction towards steeper areas at low latitudes (ΔSlope = 0.68 ± 6.5764%). Species with contraction patterns associated to land use showed: (1) consistent range contraction towards lower latitude (ΔDis_Equator = -240 ± 423.10 km); (2) no clear tendency in elevation changes (ΔElevation = 19 ± 209.22 m); and (3) a tendency towards remaining in steeper areas at low latitudes (ΔSlope = 2.14 ± 6.1056 %. Figure 3.2). Relative few species had contraction patterns best explained by both drivers (additive model) but these species showed clearer environmental patterns with range contraction towards higher latitudes particularly at intermediate-higher latitudes (ΔDis_Equator = 177 ± 465.84 km ), and consistent changes towards higher elevations and steeper areas (ΔElevation = 272 ± 278.34 m; ΔSlope = 7.08 ± 6.7591%. Figure 3.2).

152

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Figure 3.2. Changes in environmental variables (Latitude, elevation and slope, from left to right column) relative to the mean latitude of its historical range (Distance to Equator) for the group associated to climate change (a, b, c), associated to land use (d, e, f) and for the additive group (g, h, i). No change expectations (black lines) are for no temporal change in the values of the environmental variables. The confidence bands (95%) of linear models are shown.

Discussion

The importance and consequences of climate change and land use modifications on species’ distributions have usually been studied in isolation, with more limited attention to the possible combined effects of both impacts. Based in our evaluation of terrestrial vertebrates at a global scale, both climate change and land use appear to have played a role in recent range contractions, but differences among species exist. Contraction patterns were best predicted by land use for some species, but for others it was climate change, the combined effects of both impacts, or could not be predicted by the tested variables. Species affected by both impacts

153

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions had higher risk of extinction and exhibited greater range contraction. Although our global analyses supported a simpler additive model, the possibility of synergetic effects (represented by the multiplicative model) cannot be ruled out and should be further explored. Range contractions associated to each impact resulted in some similar as well as distinct environmental patterns which could be helpful to predict responses and identify existing impacts.

Climate change or land use impact? Our results suggest that no single impact is clearly responsible for observed biodiversity losses among the studied species. We found support for the importance of both impacts (individually and combined), yet for many species others threats or processes appear to be more relevant as none of the proposed predictors were relevant (alternatively, patterns may not be yet detectable as these species tended to be in earlier stages of contraction). We cannot obviate that species ranges are naturally dynamic (Gaston 2003), and that while there is strong evidence that human impacts are causing “the sixth mass extinction”, natural environmental changes (which were not modelled here) can lead to local extinction and colonization (Gaston 2003, Pechmann and Wilbur 1994, Skelly et al. 1999). In addition, we only characterized some of the human impacts that shape species distribution yet potentially other activities, like direct exploitation, fires, deforestation and interactions with invasive species, may be important (Guisan and Thuiller 2005, Laurance et al. 2009, Lindenmayer et al. 2014, Marco and Santini 2015, Maxwell et al. 2016, Watson et al. 2016). While our results indicate both impacts are relevant, we found more evidence of contraction associated to land use than climate change. This may be due to the longer history of land use changes with increasing intensity affecting biodiversity (González-Suárez & Revilla 2014). Humans have been modifying land for thousands of years with some areas intensively used since the Roman Ages (Turvey et al. 2015). On the other hand, human- induced climate changes became noticeable around mid-20th century (IPCC 2014). Therefore, species have been exposed to land use changes for much longer which could make effects more noticeable. Instead, human-induced climate change is a more recent threat, and thus, more difficult to detect in recent data (Franco et al. 2006, Wilson et al. 2005). Worryingly, climate change effects may soon become prevalent(Urban 2015, Urban et al. 2016).

154

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Spatial patterns in climate change, land use and species/richness could also explain the greater importance of land use in recent contraction patterns. Land use changes have been traditionally more intense in temperate areas and with recent development in the tropic where vertebrate diversity is higher (Butchart et al. 2010, Kehoe et al. 2015, Watson et al. 2016). Instead, climate change has been particularly noticeable at high latitudes (IPCC 2007, 2014, Parmesan and Yohe 2003) were fewer terrestrial vertebrate species occur. As climate changes become more widespread across the globe more species may suffer from this threat.

A deadly cocktail, the joined impact of climate change and land use Land use and climate change independently represent important biodiversity threats but when acting together further losses can be expected (González-Suárez and Revilla (2014). Although we found relatively few species for which the multi-impact additive-effect model was supported, predicted changes in land use and climatic conditions are likely to increase that number rapidly in the future. Furthermore, synergistic, multiplicative effects may become apparent in the future (Frishkoff et al. 2016). The more complex multiplicative model was not clearly supported (or rejected) by our analyses but there are various potential synergetic mechanisms. For example, land uses modify habitat structure leading to habitat loss and deterioration of the connectivity among populations, thus reducing the capacity of species to track their climatic niche (Hill et al. 2001, Travis 2003). Deforestation also generates changes in carbon emissions leading to increased global, and localized, climate changes (Longobardi et al. 2016). Similarly, as humans are also affected by climate change, we may start using new areas further limiting options for wildlife (Nelson et al. 2014, Thornton et al. 2014).

Distinct and identifiable refuges? We found expected changes in environmental conditions associated to contraction due to climate change including increases in altitude and latitude. The most vulnerable populations to climate change are those living close to their maximum thermal limit which often occurs at low latitudes and/or low elevations (Burrows et al. 2014, Tingley et al. 2009). Changes in the abruptness (slope) of the terrain were weaker than expected and primarily detectable at low latitudes. Microclimatic conditions are influenced by the geometry of the terrain which determines solar radiation. Steeper areas exhibit higher local-scale variation in solar 155

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions radiation, and thus, can potentially generate more microclimates that can buffer organisms against large scale changes (Tian et al. 2001). The environmental changes in species affected by land use were weaker than those for species affected by climate change and in some cases showed opposite patterns. In particular, contraction occurred towards the Equator in species classified as affected by land use, possibly because template areas are more affected by human land use, whereas species affected by climate change moved polewards (to cooler areas). Species affected by land use alone do not show consistent change in elevation, except for a possible tendency towards higher elevations at higher latitudes. The most extended human land use is agriculture and at low latitudes the altitudinal limit for agriculture is situated around 3,800 meters above sea level (Morueta-Holme et al. 2015). Therefore, at low latitudes few areas are unavailable for human uses (i.e., higher elevations may not represent clear refuges). On the other hand, at high latitudes, agriculture is more restricted by elevation. Additional data for temperate species is necessary to evaluate this hypothesis. Finally, humans also tend to occupy flatter areas which could explain the trend towards steeper zones. However, there was considerable variation among species and a latitude gradient. When contraction was best explained by a combination of both impacts, species more clearly associated with higher and steeper areas. However, changes in latitude were less clear perhaps due to the contrasting trends observed for each impact. More clear changes in elevation and abruptness may be explained by the consistent additive effects observed by each impact and because these species were in later stages of range contraction when environmental changes may be easier to detect and attribute. Understanding the characteristics of the areas in which species persist can be important for conservation. If adequately protected these areas may act as refuges as has occurred during past climatic changes (Gómez and Lunt 2007). This knowledge is also valuable for describing habitat preferences and characterizing species’ niches. Based on currently occupied areas of species affected by climate change and land use we could falsely document preferences for abrupt and higher altitude areas that in reality may represent suboptimal habitat (Buuveibaatar et al. 2016, Lea et al. 2016, Li et al. 2016). Finally, our results show that observed environmental changes are not clearly linked to unique impacts; thus, linking observed patterns to causal processes needs to be done with caution and ideally

156

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions based on observed temporal data that characterize changes in species distributions, impacts, and environmental conditions.

Limitations and future steps Conservation biology needs a better understanding of how different impacts determine biodiversity changes (Urban et al. 2016). Because extinctions are relatively rare events and difficult to document, exploring range dynamics can be a useful approach to understand these patterns and mechanisms (Bonebrake et al. 2010, Clavero and Hermoso 2015, Clavero and Revilla 2014, Li et al. 2015). Many studies have predicted the consequences that different impacts can have on existing biodiversity, but all approaches have limitations and make assumptions that need to be carefully acknowledged and considered (Urban et al. 2016). In our case, available contraction data are admittedly coarse and our evaluation of impacts is a simplification of the complexity of threats. There is a need to further explore the role that different processes (physiological, competition, predation, habitat selection, geographic) play on range dynamics and population extinction. In general there are multiple processes and impacts acting simultaneously and their combined effects can lead to unexpected consequences(Brook et al. 2008, Gaston 2008, 2009, Oliver et al. 2015, Oliver and Morecroft 2014). Here we evaluate the role of two important human impacts to show that both are relevant but also that we still have much to learn, including the impact of different threats, the potential for synergetic effects, and how impacts affect local abundance. Complementary information to better understand patterns and mechanisms associated with climate and land use impacts could be obtained from time-series of population dynamics, which unfortunately, are even scarcer than historical distribution information. Additionally, future work would be necessary to integrate additional threats like overexploitation, invasive species, and pollution, as consideration of intrinsic species traits which can influence responses to threats (González- Suárez et al 2013). We are conscientious that historical data are not such rigorous/precise than current data and that out database used may have variability in the quality of the data between species and geographical areas (González-Suárez et al. 2012), but instead they offer a global visualization across near all continents and biomes, include a big number of different taxons and a have longer temporal bin, which should reduce its possible bias. Despite these data limitations, our study helped us identify some general environmental characteristics of sites 157

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions that may represent refuge areas for species showing range contraction. Worryingly, these higher and steeper areas may not be suitable or accessible to all species. Some species may exist already in the highest elevations or biogeographical barriers may limit their ability to reach these locations. Understanding how local conditions and species’ traits influence vulnerability in light of ongoing human threats has become increasingly relevant (Pöyry et al. 2009, Warren et al. 2001).

Acknowledgments

158

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions Supplementary material

Appendix S3.1

Table A3.1. Description of data for all terrestrial vertebrates species with range contraction (N=628) obtained from the IUCN including: scientific name, historical range area (Hist range km2), extirpated range area (Ext range km2), current range area (Cur range km2), percentage of contraction (% contract), and if the species was included for analysis (Included) or the criteria for exclusion for the global and species models. Exclusion was based on number of current cells (Current), number of extirpated cells (Extirpated), on topological errors defined as the overlap between current and extirpated areas (Topology), on absent of data of some of the human ipact (NA data), or on that not all models were available due not convergence (Not all models, this criterion only applied for species analysis).

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Amphibia Agalychnis annae 3900 3900 0 100.00 Current Current Alytes obstetricans 992200 2300 989900 0.23 Included Not all models Atelopus chiriquiensis 4300 2700 1600 62.79 Included Included Atelopus varius 21600 15500 6100 71.76 Included Included Bolitoglossa robusta 6800 800 6000 11.76 Included Included Bombina variegata 1096400 300 1096100 0.03 Extirpated Extirpated Craugastor punctariolus 3800 100 3800 2.63 Extirpated Extirpated Craugastor ranoides 40200 27500 12700 68.41 Included Included Craugastor rhyacobatrachus 1500 1100 400 73.33 Current Current Craugastor taurus 5900 5800 100 98.31 Current Current Duellmanohyla uranochroa 14800 14100 700 95.27 Included Included Eleutherodactylus martinicensis 4800 600 4200 12.50 NA data NA data Eleutherodactylus schwartzi 200 100 100 50.00 Extirpated Extirpated

159

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Gastrotheca christiani 1400 0 1400 0.00 Extirpated Extirpated Hylomantis lemur 16200 7200 9100 44.44 Topology Topology Hyloscirtus colymba 16500 1900 14600 11.52 Included Not all models Isthmohyla angustilineata 1400 600 800 42.86 Included Historical Isthmohyla calypsa 300 0 300 0.00 Extirpated Extirpated Lissotriton helveticus 1086600 1200 1085400 0.11 Included Included Lithobates sevosus 21600 21500 100 99.54 Current Current Lithobates tarahumarae 97000 1800 95200 1.86 Included Included Lithobates vibicarius 2800 1800 1000 64.29 Included Historical Lithobates warszewitschii 156800 1200 155600 0.77 Included Included Liuixalus romeri 100 0 100 0.00 Extirpated Extirpated Neurergus microspilotus 700 300 400 42.86 Extirpated Extirpated Ommatotriton vittatus 53100 700 52400 1.32 Included Not all models Pelobates syriacus 520500 1700 518800 0.33 Included Included Pleurodeles waltl 379900 2300 377600 0.61 Included Included Pristimantis caryophyllaceus 51600 6800 44800 13.18 Included Included Pseudoeurycea robertsi 0 0 0 NA Extirpated Extirpated Rana latastei 33200 200 33000 0.60 Extirpated Extirpated Rana muscosa 16000 15000 1000 93.75 Included Included Rana sierrae 29000 27300 1700 94.14 Included Included Salamandra algira 21200 200 21000 0.94 Extirpated Extirpated Strabomantis bufoniformis 58000 1100 56900 1.90 Included Included Aves Accipiter butleri 200 100 100 50.00 Extirpated Extirpated 160

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Accipiter gundlachi 111800 99000 12800 88.55 Included Included Aceros waldeni 23400 12400 11000 52.99 Included Included Acrocephalus aequinoctialis 300 0 300 0.00 Extirpated Extirpated Acrocephalus brevipennis 1700 100 1600 5.88 Extirpated Extirpated Acrocephalus caffer 600 300 300 50.00 Extirpated Extirpated Acrocephalus familiaris 0 0 0 NA Extirpated Extirpated Acrocephalus luscinius 800 600 200 75.00 Current Current Acrocephalus rodericanus 100 100 0 100.00 Extirpated Extirpated Acrocephalus sechellensis 0 0 0 NA Extirpated Extirpated Aethopyga duyvenbodei 700 100 600 14.29 Extirpated Extirpated Agapornis nigrigenis 19300 2900 16400 15.03 Included Included Agapornis pullarius 2651500 200 2651300 0.01 Extirpated Extirpated Agelaius xanthomus 9100 8700 400 95.60 Current Current Agriornis albicauda 651800 72100 579700 11.06 Included Included Alectrurus risora 797900 699700 98200 87.69 Included Included Alectrurus tricolor 996800 371900 624900 37.31 Included Included Amaurolimnas concolor 1595000 11200 1583800 0.70 Included Included Amazilia castaneiventris 2900 500 2900 17.24 Topology Topology Amazona agilis 8500 5900 2600 69.41 Included Included Amazona arausiaca 300 200 100 66.67 Extirpated Extirpated Amazona barbadensis 14600 4000 11100 27.40 Topology Topology Amazona imperialis 300 200 100 66.67 Extirpated Extirpated Amazona leucocephala 125800 108500 17300 86.25 Included Included Amazona oratrix 173700 128500 45200 73.98 Included Included 161

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Amazona pretrei 26100 15700 10400 60.15 Included Included Amazona rhodocorytha 27200 24700 2500 90.81 Included Included Amazona ventralis 8700 1200 7500 13.79 Included Not all models Amazona versicolor 600 500 100 83.33 Current Current Amazona vinacea 156000 50200 105800 32.18 Included Included Amazona viridigenalis 68200 60500 7700 88.71 Included Included Amazona vittata 9100 9000 100 98.90 Current Current Anairetes alpinus 15200 3000 12200 19.74 Included Included Anas strepera 18159300 100 18159200 0.00 Extirpated Extirpated Anodorhynchus hyacinthinus 668800 131800 537000 19.71 Included Included Anthocephala floriceps 10600 600 10000 5.66 Included Included Anthracoceros montani 1400 700 700 50.00 Included Historical Anthus nattereri 94500 23800 70700 25.19 Included Included Anthus spragueii 1237100 199900 1037200 16.16 Included Included Apalharpactes reinwardtii 6300 5200 1100 82.54 Included Included Aphelocoma coerulescens 64600 58200 6400 90.09 Included Included Apteryx haastii 39900 30900 9000 77.44 Included Included Apus sladeniae 12700 100 12600 0.79 Extirpated Extirpated Ara ambiguus 103400 4300 99100 4.16 Included Included Ara ararauna 7739600 4800 7734800 0.06 Included Included Ara macao 7031900 318000 6713900 4.52 Included Included Ara militaris 378300 97900 280400 25.88 Included Included Ara rubrogenys 9900 200 9700 2.02 Extirpated Extirpated Aramides wolfi 200 0 200 0.00 Extirpated Extirpated 162

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Aramus guarauna 12205700 23600 12182100 0.19 Included Included Aratinga chloroptera 67000 48000 19000 71.64 Included Included Aratinga euops 110900 99100 11800 89.36 Included Included Ardeotis nigriceps 1112100 649200 462900 58.38 Included Included Artamus mentalis 17600 1100 16500 6.25 Included Not all models Asthenes anthoides 196600 2500 194100 1.27 Included Not all models Asthenes heterura 43600 3400 40200 7.80 Included Included Athene cunicularia 14309300 800 14308500 0.01 NA data NA data Atlapetes flaviceps 300 100 200 33.33 Extirpated Extirpated Atlapetes fuscoolivaceus 2600 1600 1000 61.54 Included Historical Atlapetes pallidiceps 1300 1300 0 100.00 Current Current Atrichornis rufescens 61300 44200 17100 72.10 Included Included Attila torridus 4800 100 4700 2.08 Extirpated Extirpated Aythya innotata 1800 1100 700 61.11 Included Historical Bangsia aureocincta 1000 200 800 20.00 Extirpated Extirpated Bangsia melanochlamys 8900 4700 4200 52.81 Included Included Basileuterus basilicus 800 300 500 37.50 Extirpated Extirpated Basileuterus conspicillatus 5700 2500 3200 43.86 Included Included Biatas nigropectus 28600 9700 18900 33.92 Included Included Bolborhynchus ferrugineifrons 3200 400 2800 12.50 Extirpated Extirpated Botaurus lentiginosus 8779300 0 8779300 0.00 Extirpated Extirpated Bradypterus grandis 12500 0 12500 0.00 Extirpated Extirpated Branta sandvicensis 7400 5100 2300 68.92 Included Included Brotogeris pyrrhoptera 13200 3900 9300 29.55 Included Included 163

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Bubulcus ibis 62356700 1800 62354900 0.00 Included Not all models Bucco noanamae 31300 1700 29600 5.43 Included Included Buteo ridgwayi 35400 35200 200 99.44 Current Current Cacatua haematuropygia 215700 193100 22600 89.52 Included Included Cacatua moluccensis 18300 400 17900 2.19 Extirpated Extirpated Callaeas cinereus 99700 93600 6100 93.88 Included Included Calyptophilus frugivorus 14400 7200 7200 50.00 Included Included Calyptura cristata 2500 2500 0 100.00 Current Current Camarhynchus psittacula 7300 0 7300 0.00 Extirpated Extirpated Campephilus imperialis 365800 365800 0 100.00 Current Current Campylopterus ensipennis 2600 0 2600 0.00 Extirpated Extirpated Campylopterus phainopeplus 1300 0 1300 0.00 Extirpated Extirpated Capito hypoleucus 4500 700 3800 15.56 Included Included Carduelis cucullata 103200 91500 11700 88.66 Included Included Carduelis siemiradzkii 24500 6200 18300 25.31 Included Included Carduelis yarrellii 40200 39300 900 97.76 Included Not all models Carpodectes antoniae 1600 0 1600 0.00 Extirpated Extirpated Carpornis melanocephala 68300 49300 19000 72.18 Included Included Cephalopterus penduliger 70400 700 69900 0.99 Topology Topology Chaetocercus berlepschi 1300 200 1100 15.38 Extirpated Extirpated Chaetornis striata 1119300 527600 591700 47.14 Included Included Charitospiza eucosma 57000 5000 52000 8.77 Included Included Charmosyna diadema 100 100 0 100.00 Extirpated Extirpated Charmosyna palmarum 12000 300 11700 2.50 Extirpated Extirpated 164

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Chasiempis ibidis 1400 1200 200 85.71 Current Current Chasiempis sandwichensis 9500 3700 5800 38.95 Included Included Chasiempis sclateri 1400 800 600 57.14 Included Historical Chlorochrysa nitidissima 18200 500 17700 2.75 Included Included Chloropeta gracilirostris 75800 500 75300 0.66 Included Not all models Chlorospingus flavovirens 5600 200 5400 3.57 Extirpated Extirpated Chondrohierax wilsonii 107800 104000 3800 96.47 Included Included Cinclodes aricomae 54300 51500 2800 94.84 Included Included Cinclodes palliatus 3100 0 3100 0.00 Extirpated Extirpated Cinclus schulzi 25400 1900 23500 7.48 Included Included Cisticola eximius 674800 20300 654500 3.01 Included Not all models Cistothorus apolinari 2800 1200 1600 42.86 Included Historical Claravis godefrida 60100 55300 4800 92.01 Included Included Clytorhynchus vitiensis 18500 600 17900 3.24 Included Not all models Coccyzus americanus 5273100 1700 5271400 0.03 Included Included Coccyzus rufigularis 17000 16300 700 95.88 Included Included Coeligena prunellei 8300 1500 6800 18.07 Included Included Colaptes auratus 14958000 300 14957700 0.00 Extirpated Extirpated Colaptes fernandinae 106700 99300 7400 93.06 Included Included Collocalia bartschi 1000 100 900 10.00 Extirpated Extirpated Collocalia elaphra 200 0 200 0.00 Extirpated Extirpated Collocalia leucophaea 100 0 100 0.00 Extirpated Extirpated Collocalia orientalis 6800 500 6300 7.35 Included Included Columba thomensis 400 0 400 0.00 Extirpated Extirpated 165

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Columbina cyanopis 2700 1500 1200 55.56 Included Historical Compsospiza baeri 11600 0 11600 0.00 Extirpated Extirpated Compsospiza garleppi 3400 0 3400 0.00 Extirpated Extirpated Conothraupis mesoleuca 1000 400 600 40.00 Extirpated Extirpated Copsychus sechellarum 200 200 0 100.00 Extirpated Extirpated Coracias garrulus 10473900 103100 10370800 0.98 Included Not all models Coracina newtoni 400 400 0 100.00 Extirpated Extirpated Coracina ostenta 24000 600 23400 2.50 Included Included Coracina typica 1700 1400 300 82.35 Current Current Corvus leucognaphalus 84300 76400 7900 90.63 Included Included Corvus palmarum 116100 7300 110000 6.29 Topology Topology Coryphaspiza melanotis 236300 67600 168700 28.61 Included Included Cotinga maculata 82900 82100 800 99.03 Included Included Cotinga ridgwayi 11800 3200 8600 27.12 Included Included Coturnicops noveboracensis 3666300 1400 3664900 0.04 Included Included Crax alberti 43500 41400 2100 95.17 Included Included Crax blumenbachii 138000 137000 1000 99.28 Included Included Crax globulosa 58000 32200 25800 55.52 Included Included Crotophaga ani 14120900 300 14120600 0.00 Extirpated Extirpated Crypturellus erythropus 68000 55700 12300 81.91 Included Included Crypturellus kerriae 5900 0 5900 0.00 Extirpated Extirpated Curaeus forbesi 18600 18500 100 99.46 Current Current Cyanolanius madagascarinus 446300 600 445700 0.13 Included Included Cyanopsitta spixii 130800 130700 100 99.92 Current Current 166

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Cyanoramphus forbesi 0 0 0 NA Extirpated Extirpated Cyanoramphus malherbi 44600 44300 300 99.33 Current Current Cyanoramphus novaezelandiae 269600 266100 3500 98.70 Included Not all models Dacnis hartlaubi 200 200 0 100.00 Extirpated Extirpated Dacnis nigripes 29200 1900 27300 6.51 Included Included Dasyornis brachypterus 28900 26900 2000 93.08 Included Included Dendrocopos canicapillus 8007500 600 8006900 0.01 Included Not all models Dendrocygna bicolor 18670400 0 18670400 0.00 Extirpated Extirpated Dendrocygna viduata 32639100 800 32638300 0.00 Included Included Dicaeum haematostictum 24300 600 23700 2.47 Included Included Didunculus strigirostris 2800 1700 1100 60.71 Included Historical Ducula aurorae 300 300 0 100.00 Extirpated Extirpated Ducula brenchleyi 10600 100 10500 0.94 Extirpated Extirpated Ducula latrans 17500 0 17500 0.00 Extirpated Extirpated Ducula mindorensis 2700 2600 1300 96.30 Topology Topology Dysithamnus occidentalis 11300 3400 7900 30.09 Included Included Dysithamnus plumbeus 16500 5400 11100 32.73 Included Included Electron carinatum 42600 7000 35600 16.43 Included Included Eleothreptus candicans 8600 5700 2900 66.28 Included Included Empidonax fulvifrons 644400 132100 512300 20.50 Included Included Eos histrio 1900 700 1200 36.84 NA data NA data Eremomela turneri 2500 100 2400 4.00 Extirpated Extirpated Eriocnemis godini 400 300 100 75.00 Extirpated Extirpated Eriocnemis nigrivestis 1200 1200 0 100.00 Current Current 167

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Eunymphicus cornutus 1300 100 1200 7.69 Extirpated Extirpated Euphonia musica 88800 0 88800 0.00 Extirpated Extirpated Eurochelidon sirintarae 2300 2300 0 100.00 Current Current Euscarthmus rufomarginatus 1924500 15200 1909300 0.79 Included Included Eutrichomyias rowleyi 600 600 0 100.00 Current Current Falco araea 200 0 200 0.00 Extirpated Extirpated Falco femoralis 12687300 1025800 11661500 8.09 Included Not all models Falco punctatus 800 400 400 50.00 Extirpated Extirpated Formicivora iheringi 21200 400 20800 1.89 Extirpated Extirpated Foudia flavicans 100 100 0 100.00 Extirpated Extirpated Foudia rubra 1700 1700 0 100.00 Current Current Foudia sechellarum 0 0 0 NA Extirpated Extirpated Francolinus gularis 280400 157600 122800 56.21 Included Included Francolinus nahani 2200 300 1900 13.64 Extirpated Extirpated Fulica cornuta 379300 174700 204600 46.06 Included Included Galbula pastazae 22500 700 21800 3.11 Included Included Gallicolumba erythroptera 1300 1300 0 100.00 Current Current Gallicolumba rubescens 500 500 0 100.00 Current Current Gallicolumba sanctaecrucis 1900 100 1800 5.26 Extirpated Extirpated Gallinago stricklandii 343800 11500 332300 3.34 Included Not all models Gallirallus sylvestris 0 0 0 NA Extirpated Extirpated Gallus gallus 5093900 0 5093900 0.00 Extirpated Extirpated Garrulax courtoisi 400 0 400 0.00 Extirpated Extirpated Garrulus lidthi 1200 300 900 25.00 Extirpated Extirpated 168

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Geospiza difficilis 2400 1200 1200 50.00 Included Historical Geospiza magnirostris 8100 700 7400 8.64 NA data NA data Geothlypis speciosa 500 0 500 0.00 Extirpated Extirpated Geotrygon caniceps 114900 86800 28100 75.54 Included Included Glaucis dohrnii 109300 108900 400 99.63 Current Current Grallaria alleni 7800 100 7700 1.28 Extirpated Extirpated Grallaria gigantea 8700 6300 2400 72.41 Included Included Grallaria kaestneri 900 700 200 77.78 Current Current Grallaria milleri 500 0 500 0.00 Extirpated Extirpated Grallaria rufocinerea 7600 2200 5400 28.95 Included Included Grallaricula cucullata 2600 900 1700 34.62 Included Historical Grus leucogeranus 350000 245200 104800 70.06 Included Included Guaruba guarouba 366200 263900 102300 72.06 Included Included Gubernatrix cristata 747800 22400 725400 3.00 Included Not all models Gyalophylax hellmayri 20100 7500 12600 37.31 Included Included Gymnogyps californianus 387900 286700 101200 73.91 Included Included Gymnomyza aubryana 1200 0 1200 0.00 Extirpated Extirpated Gyps bengalensis 5976400 1282600 4693800 21.46 Included Included Gyps coprotheres 1157900 294200 864600 25.41 Included Included Gyps tenuirostris 1660600 995700 664900 59.96 Included Included Habia atrimaxillaris 2300 1200 1100 52.17 Included Historical Hapalopsittaca amazonina 29800 4500 25600 15.10 Topology Topology Hapalopsittaca fuertesi 900 900 0 100.00 Current Current Hapalopsittaca pyrrhops 10300 0 10300 0.00 Extirpated Extirpated 169

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Harpyhaliaetus coronatus 3912300 741300 3171000 18.95 Included Not all models Heliodoxa gularis 25500 5500 23000 21.57 Topology Topology Hemignathus flavus 1400 1000 400 71.43 Current Current Hemignathus kauaiensis 1400 1200 200 85.71 Current Current Hemignathus lucidus 4000 3900 100 97.50 Current Current Hemignathus munroi 7800 7700 100 98.72 Current Current Hemignathus parvus 1500 1100 400 73.33 Current Current Hemignathus virens 6400 100 6300 1.56 Extirpated Extirpated Hemitriccus furcatus 10900 3800 7100 34.86 Included Included Hemitriccus kaempferi 7700 100 7700 1.30 Extirpated Extirpated Hemitriccus mirandae 4100 2400 1700 58.54 Included Included Herpsilochmus pectoralis 16000 7200 8800 45.00 Included Included Himatione sanguinea 16300 0 16300 0.00 Extirpated Extirpated Houbaropsis bengalensis 271900 185400 86500 68.19 Included Included Hylocryptus erythrocephalus 15400 0 15400 0.00 Extirpated Extirpated Hymenolaimus malacorhynchos 200600 145900 54700 72.73 Included Included Hypopyrrhus pyrohypogaster 8700 5000 3700 57.47 Included Included Hypothymis coelestis 179800 12100 167700 6.73 Included Included Hypsipetes olivaceus 1700 1400 300 82.35 Current Current Ibycter americanus 8440600 361700 8078900 4.29 Included Included Icterus leucopteryx 11600 200 11400 1.72 Extirpated Extirpated Icterus oberi 0 0 0 NA Extirpated Extirpated Iodopleura pipra 152800 22400 130400 14.66 Included Included Ixos siquijorensis 5400 4800 600 88.89 Included Included 170

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Jacamaralcyon tridactyla 403900 336600 67300 83.34 Included Included Jacana spinosa 1047400 4100 1043300 0.39 Included Included Junco hyemalis 10167200 300 10167000 0.00 Extirpated Extirpated Laniisoma elegans 196300 24000 172300 12.23 Included Included Lathrotriccus euleri 10273300 200 10273100 0.00 Extirpated Extirpated Lathrotriccus griseipectus 65500 52600 13600 80.31 Topology Topology Lepidopyga lilliae 200 100 100 50.00 Extirpated Extirpated Leptasthenura xenothorax 2500 100 2400 4.00 Extirpated Extirpated Leptoptilos dubius 420000 324900 95100 77.36 Included Included Leptotila ochraceiventris 41700 35500 6200 85.13 Included Included Leptotila wellsi 0 0 0 NA Extirpated Extirpated Leucopeza semperi 0 0 0 NA Extirpated Extirpated Lipaugus lanioides 203600 2700 200900 1.33 Included Included Lipaugus weberi 900 900 0 100.00 Current Current Loxigilla portoricensis 9200 200 9000 2.17 Extirpated Extirpated Loxioides bailleui 1600 1500 100 93.75 Current Current Loxops caeruleirostris 0 0 0 NA Extirpated Extirpated Loxops coccineus 10500 10100 400 96.19 Current Current Macgregoria pulchra 5500 1300 4200 23.64 Included Included Macroagelaius subalaris 8000 7500 500 93.75 Included Included Macronyx sharpei 2900 200 2700 6.90 Extirpated Extirpated Margarops fuscus 6100 600 5500 9.84 NA data NA data Megapodius layardi 6800 100 6700 1.47 Extirpated Extirpated Megapodius nicobariensis 2400 1000 1400 41.67 NA data NA data 171

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Megascops nudipes 9600 200 9400 2.08 Extirpated Extirpated Megaxenops parnaguae 670000 11200 658800 1.67 Included Included Melamprosops phaeosoma 0 0 0 NA Extirpated Extirpated Meleagris ocellata 163400 22000 141400 13.46 Included Included Mergus octosetaceus 1297500 1264500 33300 97.46 Included Included Merops breweri 1629900 5000 1624900 0.31 Included Included Merulaxis stresemanni 100 100 0 100.00 Extirpated Extirpated Micrastur plumbeus 64400 1900 62500 2.95 Included Included Microhierax erythrogenys 256900 4800 252100 1.87 Included Included Mimus trifasciatus 200 200 0 100.00 Extirpated Extirpated Mohoua ochrocephala 29200 11500 17700 39.38 Included Included Myadestes lanaiensis 1600 1600 0 100.00 Current Current Myadestes obscurus 6800 4200 2600 61.76 Included Included Myadestes palmeri 200 200 0 100.00 Extirpated Extirpated Myiarchus semirufus 32100 29700 2400 92.52 Included Included Myrmeciza griseiceps 17800 300 17500 1.69 Extirpated Extirpated Myrmeciza ruficauda 46100 45700 400 99.13 Current Current Myrmoborus melanurus 19600 1200 18400 6.12 Included Not all models Myrmotherula fluminensis 100 100 0 100.00 Extirpated Extirpated Myrmotherula grisea 32500 0 32500 0.00 Extirpated Extirpated Myrmotherula minor 11700 6600 5100 56.41 Included Included Myrmotherula unicolor 54900 20200 34700 36.79 Included Included Myrmotherula urosticta 14900 11700 3200 78.52 Included Included Myzomela rubratra 2300 600 1700 26.09 Included Historical 172

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Nemosia rourei 200 200 0 100.00 Extirpated Extirpated Neochmia ruficauda 1393900 242900 1242900 17.43 Topology Topology Neopelma aurifrons 8200 3700 4500 45.12 Included Included Nesoenas mayeri 1700 1600 100 94.12 Current Current Nestor meridionalis 89700 18000 71700 20.07 Included Included Nothoprocta taczanowskii 22000 5100 16900 23.18 Included Included Nothura minor 39900 37000 2900 92.73 Included Included Notiomystis cincta 0 0 0 NA Extirpated Extirpated Numenius borealis 492600 492600 0 100.00 Current Current Odontophorus atrifrons 2500 400 2100 16.00 Extirpated Extirpated Odontophorus hyperythrus 42100 40800 1300 96.91 Included Included Odontophorus strophium 2600 1500 1100 57.69 Included Historical Ognorhynchus icterotis 144600 144600 0 100.00 Current Current Onychorhynchus coronatus 6924600 24300 6922600 0.35 Topology Topology Onychorhynchus occidentalis 31700 24300 7400 76.66 Included Included Oreomystis bairdi 1400 1300 100 92.86 Current Current Oreomystis mana 1000 100 900 10.00 Extirpated Extirpated Oreophasis derbianus 7600 100 7500 1.32 Extirpated Extirpated Oreopholus ruficollis 1880900 100 1880800 0.01 Extirpated Extirpated Oreothraupis arremonops 16100 2300 13800 14.29 Included Included Ortalis erythroptera 22200 3000 19200 13.51 Included Included Otus brookii 38800 9200 29600 23.71 Included Not all models Otus insularis 100 0 100 0.00 Extirpated Extirpated Oxyura jamaicensis 9160800 19200 9141600 0.21 Included Not all models 173

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Oxyura leucocephala 3977500 14800 3962700 0.37 Included Included Pachycephala rufogularis 105800 4500 101300 4.25 Included Included Padda oryzivora 135500 36300 99200 26.79 Included Included Palmeria dolei 900 900 0 100.00 Current Current Paradisaea rudolphi 27200 5100 22100 18.75 Included Included Pardalotus quadragintus 6100 6000 100 98.36 Current Current Paroreomyza maculata 1200 1200 0 100.00 Current Current Paroreomyza montana 1800 1500 300 83.33 Current Current Parus nuchalis 66400 4000 62400 6.02 Included Included Passer hispaniolensis 5014300 200 5014100 0.00 Extirpated Extirpated Patagioenas inornata 206100 155200 50900 75.30 Included Included Patagioenas oenops 8200 2400 5800 29.27 Included Included Pauxi pauxi 35700 1800 33900 5.04 Included Included Pauxi unicornis 59600 44200 15400 74.16 Included Included Pavo muticus 1901100 913300 987800 48.04 Included Included Penelope albipennis 2900 0 2900 0.00 Extirpated Extirpated Penelope barbata 16300 100 16200 0.61 Extirpated Extirpated Penelope ochrogaster 74200 43200 31000 58.22 Included Included Penelope ortoni 77100 25400 51700 32.94 Included Included Penelope perspicax 300 200 100 66.67 Extirpated Extirpated Penelope pileata 977700 11900 965800 1.22 Included Included Penelopides mindorensis 8400 8400 700 100.00 Topology Topology Penelopides panini 27700 1000 26700 3.61 Included Included Petroica traversi 0 0 0 NA Extirpated Extirpated 174

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Phapitreron cinereiceps 700 0 700 0.00 Extirpated Extirpated Philesturnus carunculatus 0 0 0 NA Extirpated Extirpated Phylloscartes paulista 478200 31100 473400 6.50 Topology Topology Phylloscartes roquettei 178600 200 178600 0.11 Extirpated Extirpated Phytotoma raimondii 16900 14200 2700 84.02 Included Included Picathartes oreas 385000 0 385000 0.00 Extirpated Extirpated Picoides borealis 903000 297600 605400 32.96 Included Included Picumnus fulvescens 77700 1000 76700 1.29 Included Included Picumnus steindachneri 7000 0 7000 0.00 Extirpated Extirpated Picus squamatus 782000 77500 704500 9.91 Included Included Pipile jacutinga 1090400 1053100 37300 96.58 Included Included Pipile pipile 200 0 200 0.00 Extirpated Extirpated Pipilo erythrophthalmus 6786400 300 6786100 0.00 Extirpated Extirpated Pipilo maculatus 3629500 300 3629200 0.01 Extirpated Extirpated Piprites pileata 45900 1400 44500 3.05 Included Included Platyrinchus leucoryphus 430400 4700 425700 1.09 Included Included Platyspiza crassirostris 7300 0 7300 0.00 Extirpated Extirpated Podiceps taczanowskii 200 100 100 50.00 Extirpated Extirpated Poephila cincta 827800 332000 495800 40.11 Included Included Poliocephalus rufopectus 148300 61100 87200 41.20 Included Included Pomarea dimidiata 100 100 0 100.00 Extirpated Extirpated Pomarea iphis 100 0 100 0.00 Extirpated Extirpated Pomarea mendozae 1000 1000 0 100.00 Current Current Poospiza cinerea 53200 36100 17100 67.86 Included Included 175

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Poospiza rubecula 1500 800 700 53.33 Included Historical Porphyrio hochstetteri 5700 5100 600 89.47 Included Included Primolius maracana 3846400 214200 3632200 5.57 Included Not all models Prioniturus verticalis 700 0 700 0.00 Extirpated Extirpated Procnias tricarunculatus 22500 0 22500 0.00 Extirpated Extirpated Prosopeia personata 10500 100 10400 0.95 Extirpated Extirpated Psephotus chrysopterygius 3000 200 2800 6.67 Extirpated Extirpated Pseudibis davisoni 61200 13500 47700 22.06 Included Included Pseudocolopteryx dinelliana 182700 1300 181400 0.71 Included Included Pseudonestor xanthophrys 200 0 200 0.00 Extirpated Extirpated Psittacula eques 1200 1100 100 91.67 Current Current Psittirostra psittacea 16100 15800 300 98.14 Current Current Psophodes nigrogularis 209400 139700 69700 66.71 Included Included Ptilinopus jambu 72700 27000 45700 37.14 Included Included Ptilinopus perousii 21300 500 20800 2.35 NA data NA data Ptilinopus rarotongensis 200 100 100 50.00 Extirpated Extirpated Pyriglena atra 5000 200 4800 4.00 Extirpated Extirpated Pyrilia pyrilia 201500 48600 152900 24.12 Included Included Pyrrhura calliptera 5400 4300 1100 79.63 Included Included Pyrrhura cruentata 54200 43800 10400 80.81 Included Included Pyrrhura griseipectus 2800 2100 700 75.00 Included Historical Quelea erythrops 2570200 200 2570200 0.01 Extirpated Extirpated Rallus antarcticus 110100 900 109200 0.82 Included Not all models Rallus semiplumbeus 1600 800 800 50.00 Included Historical 176

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Ramphocinclus brachyurus 1300 1300 0 100.00 Current Current Regulus calendula 9264000 300 9263700 0.00 Extirpated Extirpated Rhaphidura sabini 1776400 4100 1772300 0.23 Included Not all models Rhinomyias albigularis 24000 600 23400 2.50 Included Included Rhipidura semirubra 1700 1600 100 94.12 Current Current Rhopornis ardesiacus 2800 300 2500 10.71 Extirpated Extirpated Salpinctes obsoletus 5425900 0 5425900 0.00 Extirpated Extirpated Saltator rufiventris 27500 3300 24200 12.00 Included Included Sarcoramphus papa 14395200 133900 14261300 0.93 Included Not all models Saxicola macrorhynchus 554400 463900 90500 83.68 Included Included Sephanoides fernandensis 200 200 0 100.00 Extirpated Extirpated Serinus xantholaemus 1600 1300 300 81.25 Current Current Simoxenops striatus 20400 800 19600 3.92 Included Included Siphonorhis americana 0 0 0 NA Extirpated Extirpated Siptornopsis hypochondriaca 7300 0 7300 0.00 Extirpated Extirpated Sitta canadensis 7144600 300 7144300 0.00 Extirpated Extirpated Spizella wortheni 0 0 0 NA Extirpated Extirpated Sporophila caerulescens 6284300 2973700 3310600 47.32 Included Not all models Sporophila falcirostris 44000 2900 41100 6.59 Included Included Sporophila frontalis 82500 6700 75800 8.12 Included Included Sporophila melanops 800 800 0 100.00 Current Current Sporophila nigrorufa 117100 400 116700 0.34 Extirpated Extirpated Starnoenas cyanocephala 110900 82700 28200 74.57 Included Included Sterna acuticauda 3389500 916500 2473000 27.04 Included Included 177

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Strigops habroptila 18000 18000 0 100.00 Current Current Strix occidentalis 1221000 15600 1205400 1.28 Included Not all models Sturnella defilippii 4900 0 4900 0.00 Extirpated Extirpated Sylvia atricapilla 10109800 200 10109600 0.00 Extirpated Extirpated Sylvia melanocephala 1924200 0 1924200 0.00 Extirpated Extirpated Synallaxis infuscata 800 800 0 100.00 Current Current Synallaxis tithys 7500 2400 5100 32.00 Included Included Syndactyla ruficollis 18500 200 18300 1.08 Extirpated Extirpated Sypheotides indicus 2444900 2048100 396800 83.77 Included Included Tachycineta euchrysea 11300 3700 7600 32.74 Included Included Tangara fastuosa 15600 0 15600 0.00 Extirpated Extirpated Tangara peruviana 84300 74600 9700 88.49 Included Included Taoniscus nanus 63300 5500 57800 8.69 Included Included Taphrolesbia griseiventris 4000 0 4000 0.00 Extirpated Extirpated Terenura sharpei 2500 300 2200 12.00 Extirpated Extirpated Terenura sicki 13800 13700 100 99.28 Current Current Terpsiphone corvina 0 0 0 NA Extirpated Extirpated Thinocorus rumicivorus 806600 100 806500 0.01 Extirpated Extirpated Thripophaga berlepschi 1600 400 1200 25.00 Extirpated Extirpated Thripophaga macroura 145200 139600 5600 96.14 Included Included Todiramphus gambieri 0 0 0 NA Extirpated Extirpated Touit melanonotus 7800 700 7100 8.97 Included Not all models Touit stictopterus 34700 2500 32200 7.20 Included Included Touit surdus 63800 31200 32600 48.90 Included Included 178

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Toxostoma guttatum 0 0 0 NA Extirpated Extirpated Triclaria malachitacea 49500 19500 30000 39.39 Included Included Troglodytes aedon 25662600 2800 25659800 0.01 Included Included Turdoides bicolor 1133000 31200 1101800 2.75 Included Included Turdoides hindei 24200 10200 14000 42.15 Included Included Turdus lherminieri 1900 500 1400 26.32 Included Historical Turdus swalesi 4900 700 4200 14.29 Included Included Turdus xanthorhynchus 200 100 100 50.00 Extirpated Extirpated Turnix varius 2289000 16200 2272800 0.71 Included Included Tympanuchus cupido 3292100 2961000 378400 89.94 Topology Topology Tympanuchus pallidicinctus 453200 453200 63100 100.00 Topology Topology Tyrannus caudifasciatus 212000 100 211900 0.05 Extirpated Extirpated Tyrannus cubensis 110800 107600 3200 97.11 Included Included Tyto longimembris 4280900 10400 4270500 0.24 Included Included Vanellus gregarius 2900300 1400800 1864900 48.30 Topology Topology Vanellus macropterus 7700 7700 0 100.00 Current Current Vermivora bachmanii 808400 808400 0 100.00 Current Current Vestiaria coccinea 16200 200 16000 1.23 Extirpated Extirpated Vini australis 4100 400 3700 9.76 Extirpated Extirpated Vini kuhlii 400 100 300 25.00 Extirpated Extirpated Vini peruviana 1700 100 1600 5.88 Extirpated Extirpated Vini ultramarina 300 100 200 33.33 Extirpated Extirpated Vireo atricapilla 455600 179000 276600 39.29 Included Included Vireo masteri 4700 1600 3100 34.04 Included Included 179

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Xanthopsar flavus 523400 447800 75600 85.56 Included Included Xenornis setifrons 6200 200 6000 3.23 Extirpated Extirpated Xenospingus concolor 31100 800 30300 2.57 Included Included Xenospiza baileyi 200 200 0 100.00 Extirpated Extirpated Xiphocolaptes falcirostris 100500 14700 85800 14.63 Included Included Xipholena atropurpurea 43200 41100 2100 95.14 Included Included Zaratornis stresemanni 45300 700 44600 1.55 Included Included Zenaida aurita 232500 400 232100 0.17 Extirpated Extirpated Zimmerius villarejoi 1800 100 1700 5.56 Extirpated Extirpated Zosterops albogularis 100 0 100 0.00 Extirpated Extirpated Zosterops chloronothus 1600 1600 0 100.00 Current Current Zosterops modestus 200 200 0 100.00 Extirpated Extirpated Mammalia Acerodon jubatus 154600 11100 143500 7.18 Included Included Acinonyx jubatus 2812800 23400 2789400 0.83 Included Not all models Aegialomys galapagoensis 500 500 0 100.00 Current Current Alcelaphus buselaphus 6855700 97100 6758600 1.42 Included Included Ammospermophilus nelsoni 8600 6100 2500 70.93 Included Included Axis porcinus 361000 88200 272800 24.43 Included Included Babyrousa babyrussa 13300 600 12700 4.51 Included Included Babyrousa celebensis 135300 15200 120100 11.23 Included Included Barbastella barbastellus 3803900 46100 3757800 1.21 Included Included Bettongia tropica 400 100 300 25.00 Extirpated Extirpated Blastocerus dichotomus 4088700 3779100 319300 92.43 Topology Topology 180

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Bos javanicus 180600 4200 176400 2.33 Included Included Canis aureus 27113600 62800 27050800 0.23 Included Not all models Canis lupus 51540100 360600 51179500 0.70 Included Not all models Canis simensis 7100 800 6300 11.27 Included Included Capreolus capreolus 7257100 29900 7227200 0.41 Included Included Capricornis milneedwardsii 2923100 1200 2921900 0.04 Included Not all models Cephalophus adersi 2800 0 2800 0.00 Extirpated Extirpated Cephalophus harveyi 355200 3800 351400 1.07 Included Not all models Cervus nippon 350800 700 350100 0.20 Included Not all models Chaetodipus rudinoris 118300 0 118300 0.00 Extirpated Extirpated Choeropsis liberiensis 142800 15700 127100 10.99 Included Not all models Coleura seychellensis 200 0 200 0.00 Extirpated Extirpated Conilurus penicillatus 252800 234700 18100 92.84 Included Included Cricetulus migratorius 7676800 15800 7661000 0.21 Included Included Crocidura canariensis 2500 0 2500 0.00 Extirpated Extirpated Crocidura sicula 25900 200 25700 0.77 Extirpated Extirpated Dendrolagus pulcherrimus 3300 2900 400 87.88 Current Current Dicerorhinus sumatrensis 8000 200 7800 2.50 Extirpated Extirpated Diceros bicornis 5164700 1251800 3912900 24.24 Included Included Dipodomys stephensi 2800 0 2800 0.00 Extirpated Extirpated Emballonura semicaudata 16400 14100 2300 85.98 Included Included Eptesicus fuscus 13171200 8100 13163100 0.06 Included Not all models Equus africanus 100300 12100 88200 12.06 Included Included Gazella dorcas 9904700 1000 9903700 0.01 Included Not all models 181

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Haeromys pusillus 51600 11200 40400 21.71 Included Included Hapalomys longicaudatus 2100 1500 600 71.43 Included Historical Helarctos malayanus 3224200 1719400 1504800 53.33 Included Included Hipposideros cervinus 1965500 600 1964900 0.03 Included Included Hystrix cristata 5344400 0 5344400 0.00 Extirpated Extirpated Isoodon macrourus 1220200 128200 1092000 10.51 Included Not all models Kobus kob 3002800 17700 2985100 0.59 Included Included Komodomys rintjanus 16200 14000 2200 86.42 Included Included Lasiurus borealis 4884500 6200 4878300 0.13 Included Not all models Leptailurus serval 12813200 290700 12522500 2.27 Included Not all models Lynx pardinus 26400 25400 1000 96.21 Included Included Macaca arctoides 1422400 20300 1402100 1.43 Included Included Macaca fuscata 157500 400 157100 0.25 Extirpated Extirpated Macaca nigra 5800 200 5600 3.45 Extirpated Extirpated Mandrillus leucophaeus 43900 1000 42900 2.28 Included Not all models Martes zibellina 6752000 17400 6734600 0.26 Included Included Mastacomys fuscus 65500 2400 63100 3.66 Included Included Melursus ursinus 2778900 1273900 1505000 45.84 Included Included Mesocapromys nanus 4100 2300 1800 56.10 Included Not all models Microtus cabrerae 119100 3700 115400 3.11 Included Included Microtus ochrogaster 3244300 36400 3207900 1.12 Included Not all models Muntiacus muntjak 1495000 600 1494400 0.04 Included Not all models Mustela lutreola 2307900 1878000 438400 81.37 Topology Topology Myotis nattereri 5868300 1800 5866500 0.03 Included Not all models 182

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Myrmecophaga tridactyla 13240300 307200 12933100 2.32 Included Included Nomascus leucogenys 51700 25300 26400 48.94 Included Included Nyctimene cephalotes 239200 28200 211000 11.79 Included Included Oryx beisa 1088700 39700 1049000 3.65 Included Not all models Oryzomys couesi 1493200 11300 1481900 0.76 Included Included Panthera pardus 22187800 59700 22129700 0.27 Topology Topology Pecari tajacu 15939300 573100 15366200 3.60 Included Not all models Perognathus alticolus 1000 0 1000 0.00 Extirpated Extirpated Peromyscus madrensis 400 100 300 25.00 Extirpated Extirpated Phascogale pirata 40500 100 40400 0.25 Extirpated Extirpated Procolobus badius 437300 84400 353000 19.30 Included Included Procolobus kirkii 1700 900 800 52.94 Included Historical Procolobus rufomitratus 1333400 100 1333300 0.01 Extirpated Extirpated Pseudantechinus mimulus 5200 200 5000 3.85 Extirpated Extirpated Pseudomys desertor 5159800 1640800 3519000 31.80 Included Included Pseudomys shortridgei 31900 400 31500 1.25 Extirpated Extirpated Pteralopex atrata 4700 0 4700 0.00 Extirpated Extirpated Pteralopex flanneryi 14100 3400 10700 24.11 Included Included Pteralopex taki 3500 800 2700 22.86 Included Historical Pteropus melanotus 8000 500 7500 6.25 NA data NA data Pteropus niger 4200 2500 1700 59.52 Included Included Reithrodontomys raviventris 1000 0 1000 0.00 Extirpated Extirpated Rhinolophus canuti 5700 300 5400 5.26 Extirpated Extirpated Rhinolophus pusillus 3927400 300 3927100 0.01 Extirpated Extirpated 183

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Rucervus eldii 84700 400 84300 0.47 Extirpated Extirpated Rusa marianna 108300 4600 103700 4.25 Included Included Sminthopsis aitkeni 1500 1400 100 93.33 Current Current Spermophilus mohavensis 20500 0 20500 0.00 Extirpated Extirpated Sus barbatus 589700 22500 567200 3.82 Included Included Sus cebifrons 2800 100 2700 3.57 Extirpated Extirpated Sus celebensis 181900 600 181300 0.33 NA data NA data Sus philippensis 64200 900 63300 1.40 Included Included Sus scrofa 27910500 200 27910300 0.00 Extirpated Extirpated Tapirus bairdii 828400 77300 751100 9.33 Included Included Tapirus terrestris 13024300 1779700 11261400 13.66 Included Included Tayassu pecari 13754800 2445100 11312100 17.78 Included Included Thylogale browni 225100 2500 222600 1.11 Included Not all models Thylogale brunii 108600 9900 98700 9.12 Included Included Thylogale calabyi 900 200 700 22.22 Extirpated Extirpated Trachypithecus geei 7300 1800 5500 24.66 Included Included Tragulus napu 1403400 600 1402800 0.04 Included Included Ursus arctos 29841800 2353300 27488600 7.89 Included Included Ursus thibetanus 7045200 3737600 3307600 53.05 Included Included Wilfredomys oenax 119400 3800 115600 3.18 Included Included Zaglossus bruijnii 91700 1800 89900 1.96 Included Included Reptilia Acanthodactylus schreiberi 10800 2300 8500 21.30 Included Included Chalcides mauritanicus 4000 0 4000 0.00 Extirpated Extirpated 184

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species (class) Hist range km2 Ext range km2 Cur range km2 % contract Global Species Coronella girondica 1204800 4000 1200800 0.33 Included Included Crotalus horridus 2048700 3700 2045000 0.18 Included Not all models Cyrtopodion russowii 1334600 6800 1327800 0.51 Included Not all models Gallotia auaritae 100 100 0 100.00 Extirpated Extirpated Iguana delicatissima 4700 1900 2800 40.43 Included Included Lacerta bilineata 816300 100 816200 0.01 Extirpated Extirpated Lacerta media 862200 400 861800 0.05 Extirpated Extirpated Lacerta schreiberi 141200 1100 140100 0.78 Included Included Phrynocephalus persicus 127900 200 127700 0.16 Extirpated Extirpated Phrynosoma cornutum 1801100 16100 1785000 0.89 Included Not all models Phrynosoma douglasii 374100 2000 372100 0.53 Included Included Pituophis ruthveni 74400 73000 1400 98.12 Included Included Psammodromus microdactylus 7900 4000 3900 50.63 Included Included Regina septemvittata 1019100 1700 1017400 0.17 Included Included Timon lepidus 621100 200 620900 0.03 Extirpated Extirpated Timon tangitanus 145200 1200 144000 0.83 Included Not all models Trapelus savignii 36500 100 36400 0.27 Extirpated Extirpated Vipera ursinii 67900 200 67700 0.29 Extirpated Extirpated

185

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Appendix S3.2. Supplementary methods Distributions with contraction Distribution maps of terrestrial vertebrates with observed range contraction were obtained from the International Union for Conservation of Nature (International Union for Conservation of Nature 2010a). From all terrestrial vertebrate species with available spatial data (N=23,743), we selected areas defined by the IUCN as native in origin, resident or breeding season in seasonality and in the presence categories of extant, probably extant, possibly extinct, or extinct (Based in IUCN, maps of the historical range of a species are a combination of polygons coded as Extant, Probably Extant, Possibly Extinct, and Extint, IUCN 2010b). Original presence categories of the IUCN were reclassified into current (extant and probably extant) and extirpated (extinct and probably extinct), defining the historical range as the sum of both, and defining a species with range contraction as a species with categories of current and extirpated (N=628, Table A4.1). We used an equal area projection (Cylindrical Equal Area projection) in ArcMap 9.3 (ESRI 2008), and rasterized, for each distribution, the current areas and the extirpated areas, using Idrisi Taiga (Clark Labs 2009). The use of raster format is common in biodiversity studies at range scale, including others that analyzed range contraction (Ceballos and Ehrlich 2002, Yackulic et al. 2011), as it facilitates the management of spatial calculations in large datasets. We applied a resolution grid cell of 100 km2 (e.g., proximally 10 km cell side).

Calculation of land use We reclassified the original categories of the MODIS (MCD12Q1) Land Cover Product (Oak Ridge National Laboratory Distributed Active Archive Center 2010) to define the variable Land use used in our analyses (Table A4.3). Original data at 1km grid cell resolution were projected into Cylindrical Equal Area with ArcMap 9.3 (ESRI 2008) using the same registration point as distributions rasterized (x and y coordinates, in the output space, used for pixel alignment) to obtain the alignment of pixels between the land use map and the distributions. These values were used to calculate the percentage of land use for the 100 km2 grid cells used in the analyses.

186

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Calculation of human density Gridded Population of the World, Version 3 (GPWv3) are a times series of rasterized data on population densityes (humans per square kilometer) for the years 1990, 1995 and 2000 (Center for International Earth Science Information Network - CIESIN - Columbia University 2005). We used data from the year 2000. Original data at a resolution of 2.5 arc- minute grid cellswere projected it into a resolution of 100 km2 using Cylindrical Equal Area projection with ArcMap 9.3 (ESRI 2008) using the same registration point as distributions rasterized (x and y coordinates, in the output space, used for pixel alignment) to obtain the alignment of pixels between the human density map and the distributions.

Calculation of distance to Equator The best distance-preserving projections available at a global scale have a root-mean- square logarithmic distance error of σ ≈ 0.340 (Gott et al. 2007). These errors are biased with latitude and/or longitude (Mena Berrios 2008) and could affect our estimates of distances. Therefore, we calculated the spatial position of each grid cell center over the spheroid of the Earth (using the spheroid World Geodetic System 1984, WGS84), and then calculated the distance between each grid cell center and the point of the same longitude from the equatorial line as the geodetic distance, that is the distance between two unprojected points on the Earth's surface. Geodetic distances were calculated in Visual Fox Pro applying the Vicenty’s algorithm (Vincenty 1975) over the spheroid WGS84. Estimates generated using this algorithm agreed to within 0.115 mm for distances between 10 to 18,000 km (Thomas and Featherstone 2005) and thus, provide accurate estimates of varying distances at global scales.

Analysis of spatial autocrrelation To asses if our random variable (block of four contiguous cells) was correctly controlling for any potential spatial autocorrelation, we calculated the Moran Index of residuals for the species models as described in the main text, and for models fitted without the random variable. Results indicate that, allthought in some models did not reduced the spatial autocorrelation, in general using blocks of four contiguous cells reduced the spatial autocorrelation in models (table A4.5 and figure A4.3).

187

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A4.2 Descriptive statistics for the distributions used in global and species analysis. We report the mean, standard deviation (SD), median and range (minimum and maximum) of the total area (in km2), as well as the percentage of contraction (% Contraction). Range Mean SD Median Range Global models Historical 1,813,997 5,914,033 89,700 1,400-62,356,700 Extirpated 135,133 413,801.60 11,100 500-3,737,600 Current 1,678,926 5,823,124 44,500 500-62,354,900 % Contraction 34.17 33.43 23.64 0.002-99.28 Species models Historical 13,062,600 4,056,032 82,700 4,100-32,639,400 Extirpated 152,100 438,853.90 14,100 500-3,737,600 Current 1,154,120 3,916,562 364,500 500-3,238,600 % Contraction 37.539 34.02 27.079 0.002-99.28

Table A4.3. Reclassification of the original categories of the MODIS (MCD12Q1) Land Cover Product (Oak Ridge National Laboratory Distributed Active Archive Center 2010) used to define the human land cover (1 for land use, 0 for non land use) variable used in our analyses. Original categories Original code Land use Water 0 No data Evergreen needleleaf forest 1 0 Evergreen broadleaf forest 2 0 Deciduous needleleaf forest 3 0 Deciduous broadleaf forest 4 0 Mixed forest 5 0 Closed shrublands 6 0 Open shrublands 7 0 Woody savannas 8 0 Savannas 9 0 Grasslands 10 0 Permanent wetlands 11 0 Croplands 12 1 Urban and built-up lands 13 1 Cropland/natural vegetation mosaics 14 1 Snow and ice 15 0 Barren 16 0

188

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.4. Results from the species regression models (GLMM) used to evaluate the importance within species of climate change (Mod_clim) and land use (Mod_use) and the multi-impact additive process (Mod_clim_use). We also include a null model used as a reference to evaluate model fit. We report model coefficients (best estimates and their SE), AICc, and model weight (based on comparing models for each species). NA indicates variables not included in the model or that model did not converge (indicated also by NA in model comparation columns). Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Bubulcus ibis Mod_use NA NA NA NA NA NA Bubulcus ibis Mod_clim NA NA -2.98 15.540 23.88 0.25 Bubulcus ibis Mod_clim_use -8.36 64.794 -3.08 14.346 25.82 0.09 Bubulcus ibis Mod_null NA NA NA NA 21.93 0.66 Canis lupus Mod_use NA NA NA NA NA NA Canis lupus Mod_clim NA NA 2.02 0.478 3366.42 0.70 Canis lupus Mod_clim_use 0.32 0.593 1.99 0.483 3368.14 0.30 Canis lupus Mod_null NA NA NA NA 3383.14 0.00 Coracias garrulus Mod_use 1.75 0.972 NA NA 954.32 0.21 Coracias garrulus Mod_clim NA NA -9.85 8.308 951.99 0.68 Coracias garrulus Mod_clim_use NA NA NA NA NA NA Coracias garrulus Mod_null NA NA NA NA 955.74 0.10 Dendrocygna viduata Mod_use 3.15 10.339 NA NA 16.69 0.20 Dendrocygna viduata Mod_clim NA NA 7.03 19.465 16.61 0.21 Dendrocygna viduata Mod_clim_use 3.54 10.479 7.69 24.392 18.52 0.08 Dendrocygna viduata Mod_null NA NA NA NA 14.76 0.52 Ursus arctos Mod_use 2.30 0.147 NA NA 28632.89 0.00 Ursus arctos Mod_clim NA NA 4.56 0.205 28188.69 0.00 Ursus arctos Mod_clim_use 0.95 0.154 4.31 0.224 28154.22 1.00 Ursus arctos Mod_null NA NA NA NA 28837.97 0.00

189

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Troglodytes aedon Mod_use 2.37 5.789 NA NA 37.59 0.21 Troglodytes aedon Mod_clim NA NA 1.38 5.510 37.68 0.20 Troglodytes aedon Mod_clim_use 3.83 7.160 2.87 6.382 39.40 0.08 Troglodytes aedon Mod_null NA NA NA NA 35.74 0.52 Canis aureus Mod_use 0.28 0.000 NA NA 822.09 0.00 Canis aureus Mod_clim NA NA -29.23 0.000 807.53 1.00 Canis aureus Mod_clim_use NA NA NA NA NA NA Canis aureus Mod_null NA NA NA NA 820.46 0.00 Coccyzus americanus Mod_use -4.42 0.001 NA NA 37.33 0.00 Coccyzus americanus Mod_clim NA NA 170.94 51.887 26.42 0.72 Coccyzus americanus Mod_clim_use -3.46 25.375 170.13 65.679 28.39 0.27 Coccyzus americanus Mod_null NA NA NA NA 35.55 0.01 Pecari tajacu Mod_use 2.75 0.386 NA NA 5961.36 1.00 Pecari tajacu Mod_clim NA NA NA NA NA NA Pecari tajacu Mod_clim_use NA NA NA NA NA NA Pecari tajacu Mod_null NA NA NA NA 6002.33 0.00 Oxyura jamaicensis Mod_use -0.66 3.598 NA NA 163.68 0.27 Oxyura jamaicensis Mod_clim NA NA NA NA NA NA Oxyura jamaicensis Mod_clim_use NA NA NA NA NA NA Oxyura jamaicensis Mod_null NA NA NA NA 161.72 0.73 Sarcoramphus papa Mod_use 1.48 0.000 NA NA 1327.51 0.37 Sarcoramphus papa Mod_clim NA NA -0.62 0.000 1328.72 0.20 Sarcoramphus papa Mod_clim_use NA NA NA NA NA NA Sarcoramphus papa Mod_null NA NA NA NA 1327.19 0.43 190

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Tayassu pecari Mod_use 1.99 0.210 NA NA 27656.13 0.51 Tayassu pecari Mod_clim NA NA 0.01 0.158 27734.15 0.00 Tayassu pecari Mod_clim_use 2.04 0.215 0.22 0.160 27656.22 0.49 Tayassu pecari Mod_null NA NA NA NA 27732.13 0.00 Myrmecophaga tridactyla Mod_use 3.98 0.546 NA NA 3200.69 0.00 Myrmecophaga tridactyla Mod_clim NA NA -5.35 1.565 3223.33 0.00 Myrmecophaga tridactyla Mod_clim_use 3.51 0.547 -4.42 1.547 3186.02 1.00 Myrmecophaga tridactyla Mod_null NA NA NA NA 3247.51 0.00 Eptesicus fuscus Mod_use NA NA NA NA NA NA Eptesicus fuscus Mod_clim NA NA NA NA NA NA Eptesicus fuscus Mod_clim_use NA NA NA NA NA NA Eptesicus fuscus Mod_null NA NA NA NA 70.22 1.00 Tapirus terrestris Mod_use 1.96 0.243 NA NA 20715.19 0.73 Tapirus terrestris Mod_clim NA NA -0.21 0.179 20771.20 0.00 Tapirus terrestris Mod_clim_use 1.95 0.245 -0.04 0.181 20717.14 0.27 Tapirus terrestris Mod_null NA NA NA NA 20770.53 0.00 Leptailurus serval Mod_use NA NA NA NA NA NA Leptailurus serval Mod_clim NA NA NA NA NA NA Leptailurus serval Mod_clim_use NA NA NA NA NA NA Leptailurus serval Mod_null NA NA NA NA 2703.42 1.00 Falco femoralis Mod_use NA NA NA NA NA NA Falco femoralis Mod_clim NA NA -0.99 0.377 9125.46 0.00 Falco femoralis Mod_clim_use -2.13 0.639 -1.10 0.376 9112.07 1.00 Falco femoralis Mod_null NA NA NA NA 9131.35 0.00 191

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Aramus guarauna Mod_use -0.55 2.845 NA NA 248.68 0.20 Aramus guarauna Mod_clim NA NA 0.07 2.249 248.72 0.20 Aramus guarauna Mod_clim_use -0.55 2.880 0.01 2.259 250.69 0.07 Aramus guarauna Mod_null NA NA NA NA 246.72 0.53 Oxyura leucocephala Mod_use 4.05 3.663 NA NA 154.08 0.26 Oxyura leucocephala Mod_clim NA NA 2.95 3.040 154.91 0.17 Oxyura leucocephala Mod_clim_use 5.19 3.925 3.56 2.636 153.72 0.32 Oxyura leucocephala Mod_null NA NA NA NA 154.19 0.25 Gazella dorcas Mod_use 8.79 0.002 NA NA 19.32 0.28 Gazella dorcas Mod_clim NA NA NA NA NA NA Gazella dorcas Mod_clim_use NA NA NA NA NA NA Gazella dorcas Mod_null NA NA NA NA 18.14 0.51 Ibycter americanus Mod_use 4.56 0.554 NA NA 3628.32 0.04 Ibycter americanus Mod_clim NA NA -1.39 0.576 3671.31 0.00 Ibycter americanus Mod_clim_use 4.61 0.551 -1.49 0.554 3622.10 0.96 Ibycter americanus Mod_null NA NA NA NA 3675.93 0.00 Dendrocopos canicapillus Mod_use 2.86 15.054 NA NA 13.08 0.25 Dendrocopos canicapillus Mod_clim NA NA NA NA NA NA Dendrocopos canicapillus Mod_clim_use 3.24 16.254 201.43 156.702 14.91 0.10 Dendrocopos canicapillus Mod_null NA NA NA NA 11.12 0.66 Ara ararauna Mod_use 3.58 5.397 NA NA 54.37 0.22 Ara ararauna Mod_clim NA NA 0.69 4.499 54.68 0.19 Ara ararauna Mod_clim_use 3.76 5.595 0.88 4.701 56.33 0.08 Ara ararauna Mod_null NA NA NA NA 52.71 0.51 192

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Cricetul us migratorius Mod_use -0.27 2.503 NA NA 201.27 0.17 Cricetulus migratorius Mod_clim NA NA 2.14 2.200 200.24 0.28 Cricetulus migratorius Mod_clim_use -0.12 2.682 2.14 2.211 202.24 0.10 Cricetulus migratorius Mod_null NA NA NA NA 199.28 0.45 Capreolus capreolus Mod_use -0.32 1.483 NA NA 395.42 0.07 Capreolus capreolus Mod_clim NA NA 3.46 2.010 391.22 0.54 Capreolus capreolus Mod_clim_use -0.74 1.532 3.45 1.969 392.99 0.22 Capreolus capreolus Mod_null NA NA NA NA 393.46 0.17 Ursus thibetanus Mod_use 19.59 0.240 NA NA 40152.47 0.00 Ursus thibetanus Mod_clim NA NA 18.24 0.215 39894.59 0.00 Ursus thibetanus Mod_clim_use 13.78 0.310 11.44 0.291 38134.59 1.00 Ursus thibetanus Mod_null NA NA NA NA 40041.46 0.00 Ara macao Mod_use 5.12 0.608 NA NA 3084.72 0.56 Ara macao Mod_clim NA NA 0.09 0.517 3140.89 0.00 Ara macao Mod_clim_use 5.29 0.621 -0.65 0.530 3085.18 0.44 Ara macao Mod_null NA NA NA NA 3138.92 0.00 Alcelaphus buselaphus Mod_use -3.45 4.446 NA NA 927.42 0.30 Alcelaphus buselaphus Mod_clim NA NA 0.81 0.924 927.86 0.24 Alcelaphus buselaphus Mod_clim_use -181.11 0.002 1.27 0.460 1273.43 0.00 Alcelaphus buselaphus Mod_null NA NA NA NA 926.60 0.46 Martes zibellina Mod_use 1.91 2.571 NA NA 231.01 0.12 Martes zibellina Mod_clim NA NA 4.52 3.721 228.46 0.44 Martes zibellina Mod_clim_use 0.11 2.687 4.49 3.794 230.46 0.16 Martes zibellina Mod_null NA NA NA NA 229.42 0.27 193

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Sporophila caerulescens Mod_use -67.21 0.000 NA NA 24295.62 0.00 Sporophila caerulescens Mod_clim NA NA 29.31 0.292 23187.48 1.00 Sporophila caerulescens Mod_clim_use NA NA NA NA NA NA Sporophila caerulescens Mod_null NA NA NA NA 24683.90 0.00 Gyps bengalensis Mod_use -0.44 0.235 NA NA 11962.71 0.15 Gyps bengalensis Mod_clim NA NA 0.55 0.357 11963.91 0.08 Gyps bengalensis Mod_clim_use -0.61 0.247 0.85 0.372 11959.67 0.69 Gyps bengalensis Mod_null NA NA NA NA 11964.22 0.07 Myotis nattereri Mod_use -147.96 99.810 NA NA 39.81 0.65 Myotis nattereri Mod_clim NA NA NA NA NA NA Myotis nattereri Mod_clim_use NA NA NA NA NA NA Myotis nattereri Mod_null NA NA NA NA 43.89 0.08 Diceros bicornis Mod_use -0.76 0.692 NA NA 11143.93 0.00 Diceros bicornis Mod_clim NA NA 2.04 0.240 11077.34 0.48 Diceros bicornis Mod_clim_use -0.96 0.705 2.07 0.242 11077.22 0.52 Diceros bicornis Mod_null NA NA NA NA 11143.31 0.00 Pseudomys desertor Mod_use 151.31 10.256 NA NA 14689.34 0.00 Pseudomys desertor Mod_clim NA NA -4.52 0.562 14962.02 0.00 Pseudomys desertor Mod_clim_use 152.97 9.197 -4.59 0.580 14545.57 1.00 Pseudomys desertor Mod_null NA NA NA NA 15111.29 0.00 Lasiurus borealis Mod_use -4.91 10.420 NA NA 60.73 0.32 Lasiurus borealis Mod_clim NA NA NA NA NA NA Lasiurus borealis Mod_clim_use NA NA NA NA NA NA Lasiurus borealis Mod_null NA NA NA NA 59.23 0.68 194

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Tyto longimembris Mod_use -0.25 3.345 NA NA 122.43 0.20 Tyto longimembris Mod_clim NA NA -0.44 3.994 122.42 0.20 Tyto longimembris Mod_clim_use -0.17 3.439 -0.39 4.111 124.42 0.07 Tyto longimembris Mod_null NA NA NA NA 120.43 0.53 Coturnicops noveboracensis Mod_use 0.74 8.288 NA NA 20.05 0.19 Coturnicops noveboracensis Mod_clim NA NA 3.66 14.390 19.94 0.21 Coturnicops noveboracensis Mod_clim_use -0.28 8.887 3.70 14.089 21.94 0.08 Coturnicops noveboracensis Mod_null NA NA NA NA 18.06 0.53 Anthus spragueii Mod_use -0.93 0.575 NA NA 2177.70 0.13 Anthus spragueii Mod_clim NA NA -0.98 0.710 2178.43 0.09 Anthus spragueii Mod_clim_use -1.44 0.603 -1.58 0.708 2174.35 0.69 Anthus spragueii Mod_null NA NA NA NA 2178.47 0.09 Harpyhaliaetus coronatus Mod_use 1.96 0.359 NA NA 7342.05 0.00 Harpyhaliaetus coronatus Mod_clim NA NA NA NA NA NA Harpyhaliaetus coronatus Mod_clim_use 1.36 0.346 -5.75 0.984 7261.58 1.00 Harpyhaliaetus coronatus Mod_null NA NA NA NA 7366.88 0.00 Primolius maracana Mod_use -0.61 0.000 NA NA 2343.61 0.00 Primolius maracana Mod_clim NA NA NA NA NA NA Primolius maracana Mod_clim_use -1.20 0.759 -5.32 1.882 2324.33 1.00 Primolius maracana Mod_null NA NA NA NA 2342.62 0.00 Barbastella barbastellus Mod_use 0.46 0.000 NA NA 598.07 0.02 Barbastella barbastellus Mod_clim NA NA 3.11 1.368 591.50 0.66 Barbastella barbastellus Mod_clim_use 0.42 1.463 3.13 1.398 593.42 0.25 Barbastella barbastellus Mod_null NA NA NA NA 596.15 0.06 195

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Sterna acuticauda Mod_use -2.71 0.349 NA NA 7518.96 0.73 Sterna acuticauda Mod_clim NA NA -0.96 0.465 7581.83 0.00 Sterna acuticauda Mod_clim_use -2.71 0.360 0.01 0.440 7520.96 0.27 Sterna acuticauda Mod_null NA NA NA NA 7584.28 0.00 Microtus ochrogaster Mod_use -0.76 1.461 NA NA 363.18 0.02 Microtus ochrogaster Mod_clim NA NA 251.01 25.618 355.43 0.93 Microtus ochrogaster Mod_clim_use NA NA NA NA NA NA Microtus ochrogaster Mod_null NA NA NA NA 361.45 0.05 Helarctos malayanus Mod_use 13.67 0.302 NA NA 21579.72 0.00 Helarctos malayanus Mod_clim NA NA 3.42 0.125 26934.14 0.00 Helarctos malayanus Mod_clim_use 13.36 0.305 1.07 0.134 21515.67 1.00 Helarctos malayanus Mod_null NA NA NA NA 27787.89 0.00 Kobus kob Mod_use -1.82 5.097 NA NA 199.18 0.19 Kobus kob Mod_clim NA NA 1.87 2.594 198.78 0.23 Kobus kob Mod_clim_use -2.22 5.179 1.96 2.517 200.50 0.10 Kobus kob Mod_null NA NA NA NA 197.36 0.48 Capricornis milneedwardsii Mod_use 2.82 8.201 NA NA 26.68 0.21 Capricornis milneedwardsii Mod_clim NA NA 17.07 53.372 26.23 0.26 Capricornis milneedwardsii Mod_clim_use NA NA NA NA NA NA Capricornis milneedwardsii Mod_null NA NA NA NA 24.78 0.53 Acinonyx jubatus Mod_use NA NA NA NA NA NA Acinonyx jubatus Mod_clim NA NA -13.46 12.563 229.52 0.68 Acinonyx jubatus Mod_clim_use NA NA NA NA NA NA Acinonyx jubatus Mod_null NA NA NA NA 231.05 0.32 196

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Melursus ursinus Mod_use 20.41 0.435 NA NA 14511.02 0.00 Melursus ursinus Mod_clim NA NA 0.93 0.271 15339.83 0.00 Melursus ursinus Mod_clim_use 20.11 0.441 1.89 0.381 14488.42 1.00 Melursus ursinus Mod_null NA NA NA NA 15349.61 0.00 Sypheotides indicus Mod_use 0.94 0.339 NA NA 5665.00 0.65 Sypheotides indicus Mod_clim NA NA 0.51 0.456 5671.26 0.03 Sypheotides indicus Mod_clim_use 0.90 0.346 0.27 0.459 5666.65 0.28 Sypheotides indicus Mod_null NA NA NA NA 5670.52 0.04 Turnix varius Mod_use -1.62 4.042 NA NA 184.83 0.20 Turnix varius Mod_clim NA NA 1.57 2.150 184.48 0.23 Turnix varius Mod_clim_use -0.55 4.417 1.41 2.454 186.47 0.09 Turnix varius Mod_null NA NA NA NA 183.05 0.48 Leptoptilos dubius Mod_use -1.26 4.812 NA NA 890.87 0.35 Leptoptilos dubius Mod_clim NA NA -0.02 1.070 892.83 0.13 Leptoptilos dubius Mod_clim_use -1.55 0.987 0.82 1.271 892.44 0.16 Leptoptilos dubius Mod_null NA NA NA NA 890.83 0.36 Crotalus horridus Mod_use 1.66 5.973 NA NA 57.22 0.21 Crotalus horridus Mod_clim NA NA NA NA NA NA Crotalus horridus Mod_clim_use NA NA NA NA NA NA Crotalus horridus Mod_null NA NA NA NA 55.31 0.55 Hipposideros cervinus Mod_use -1.32 7.502 NA NA 37.52 0.16 Hipposideros cervinus Mod_clim NA NA 13.20 14.007 36.13 0.31 Hipposideros cervinus Mod_clim_use -0.90 6.966 13.12 14.885 38.11 0.12 Hipposideros cervinus Mod_null NA NA NA NA 35.56 0.42 197

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Euscarthmus rufomarginatus Mod_use -6.36 9.166 NA NA 183.76 0.25 Euscarthmus rufomarginatus Mod_clim NA NA 1.71 2.220 184.08 0.22 Euscarthmus rufomarginatus Mod_clim_use -5.16 9.240 0.96 2.289 185.57 0.10 Euscarthmus rufomarginatus Mod_null NA NA NA NA 182.71 0.43 Pavo muticus Mod_use 23.36 0.001 NA NA 8632.63 0.00 Pavo muticus Mod_clim NA NA 21.22 0.000 8396.70 0.00 Pavo muticus Mod_clim_use 7.09 0.704 18.20 0.681 8260.37 0.00 Pavo muticus Mod_null NA NA NA NA 8119.73 1.00 Gyps tenuirostris Mod_use -0.65 0.339 NA NA 6307.96 0.00 Gyps tenuirostris Mod_clim NA NA 0.81 0.379 6306.77 0.00 Gyps tenuirostris Mod_clim_use -1.46 0.407 1.66 0.465 6295.92 0.99 Gyps tenuirostris Mod_null NA NA NA NA 6309.56 0.00 Phrynosoma cornutum Mod_use 2.70 0.000 NA NA 392.14 0.71 Phrynosoma cornutum Mod_clim NA NA NA NA NA NA Phrynosoma cornutum Mod_clim_use NA NA NA NA NA NA Phrynosoma cornutum Mod_null NA NA NA NA 393.90 0.29 Rhaphidura sabini Mod_use 31.93 0.001 NA NA 67.58 0.94 Rhaphidura sabini Mod_clim NA NA NA NA NA NA Rhaphidura sabini Mod_clim_use NA NA NA NA NA NA Rhaphidura sabini Mod_null NA NA NA NA 72.93 0.06 Merops breweri Mod_use 6.59 0.001 NA NA 65.56 0.32 Merops breweri Mod_clim NA NA 5.44 9.222 66.56 0.19 Merops breweri Mod_clim_use 5.53 4.684 4.51 8.814 67.10 0.15 Merops breweri Mod_null NA NA NA NA 65.40 0.34 198

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Amaurolimnas concolor Mod_use 2.33 3.282 NA NA 97.28 0.23 Amaurolimnas concolor Mod_clim NA NA -0.46 3.484 97.69 0.19 Amaurolimnas concolor Mod_clim_use 2.39 3.312 -0.72 3.906 99.25 0.09 Amaurolimnas concolor Mod_null NA NA NA NA 95.71 0.50 Muntiacus muntjak Mod_use -1.57 7.123 NA NA 37.41 0.05 Muntiacus muntjak Mod_clim NA NA 150.08 73.020 31.83 0.82 Muntiacus muntjak Mod_clim_use NA NA NA NA NA NA Muntiacus muntjak Mod_null NA NA NA NA 35.47 0.13 Oryzomys couesi Mod_use 1.06 3.473 NA NA 100.85 0.20 Oryzomys couesi Mod_clim NA NA 0.13 2.989 100.94 0.19 Oryzomys couesi Mod_clim_use 1.08 3.569 -0.08 3.107 102.85 0.07 Oryzomys couesi Mod_null NA NA NA NA 98.94 0.53 Macaca arctoides Mod_use 18.80 3.597 NA NA 315.61 0.00 Macaca arctoides Mod_clim NA NA -12.48 2.086 343.08 0.00 Macaca arctoides Mod_clim_use 13.80 3.221 -7.61 2.245 302.69 1.00 Macaca arctoides Mod_null NA NA NA NA 391.88 0.00 Tragulus napu Mod_use 139.40 72.822 NA NA 15.36 0.61 Tragulus napu Mod_clim NA NA 3.48 14.194 20.48 0.05 Tragulus napu Mod_clim_use 139.10 46.300 2.69 35.774 17.35 0.22 Tragulus napu Mod_null NA NA NA NA 18.59 0.12 Gyps coprotheres Mod_use -17.28 7.953 NA NA 2511.52 0.55 Gyps coprotheres Mod_clim NA NA 1.16 0.475 2521.52 0.00 Gyps coprotheres Mod_clim_use -15.13 7.820 0.57 0.481 2511.97 0.44 Gyps coprotheres Mod_null NA NA NA NA 2525.47 0.00 199

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Cyrtopodion russowii Mod_use -1.88 13.563 NA NA 77.41 0.27 Cyrtopodion russowii Mod_clim NA NA NA NA NA NA Cyrtopodion russowii Mod_clim_use NA NA NA NA NA NA Cyrtopodion russowii Mod_null NA NA NA NA 75.44 0.73 Mergus octosetaceus Mod_use 4.55 2.400 NA NA 508.67 0.16 Mergus octosetaceus Mod_clim NA NA -2.98 1.448 508.26 0.19 Mergus octosetaceus Mod_clim_use 4.45 2.449 -2.79 1.522 505.94 0.61 Mergus octosetaceus Mod_null NA NA NA NA 511.55 0.04 Strix occidentalis Mod_use -0.32 5.422 NA NA 244.73 0.27 Strix occidentalis Mod_clim NA NA NA NA NA NA Strix occidentalis Mod_clim_use NA NA NA NA NA NA Strix occidentalis Mod_null NA NA NA NA 242.72 0.73 Isoodon macrourus Mod_use 4.17 1.254 NA NA 1527.84 0.96 Isoodon macrourus Mod_clim NA NA NA NA NA NA Isoodon macrourus Mod_clim_use NA NA NA NA NA NA Isoodon macrourus Mod_null NA NA NA NA 1534.00 0.04 Coronella girondica Mod_use 1.05 4.759 NA NA 53.68 0.19 Coronella girondica Mod_clim NA NA -5.40 17.107 53.39 0.22 Coronella girondica Mod_clim_use 0.75 4.793 -5.16 16.897 55.36 0.08 Coronella girondica Mod_null NA NA NA NA 51.73 0.51 Turdoides bicolor Mod_use 9.59 3.803 NA NA 384.36 0.58 Turdoides bicolor Mod_clim NA NA -1.41 1.823 388.81 0.06 Turdoides bicolor Mod_clim_use 9.02 3.990 -0.65 1.841 386.23 0.23 Turdoides bicolor Mod_null NA NA NA NA 387.48 0.12 200

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Chaetornis striata Mod_use -30.30 1.264 NA NA 3712.48 0.00 Chaetornis striata Mod_clim NA NA 2.53 0.702 4505.24 0.00 Chaetornis striata Mod_clim_use -38.90 1.686 33.80 1.824 3493.52 1.00 Chaetornis striata Mod_null NA NA NA NA 4516.95 0.00 Ardeotis nigriceps Mod_use 20.07 0.540 NA NA 5033.14 0.00 Ardeotis nigriceps Mod_clim NA NA 1.02 0.684 4963.66 0.00 Ardeotis nigriceps Mod_clim_use 2.76 0.446 0.50 0.675 4927.85 1.00 Ardeotis nigriceps Mod_null NA NA NA NA 4963.92 0.00 Pipile jacutinga Mod_use 9.23 1.651 NA NA 767.34 0.61 Pipile jacutinga Mod_clim NA NA -0.03 0.723 824.52 0.00 Pipile jacutinga Mod_clim_use 9.55 1.707 -0.83 0.771 768.21 0.39 Pipile jacutinga Mod_null NA NA NA NA 822.50 0.00 Oryx beisa Mod_use -9.82 21.863 NA NA 371.31 0.24 Oryx beisa Mod_clim NA NA 0.79 1.475 371.45 0.23 Oryx beisa Mod_clim_use NA NA NA NA NA NA Oryx beisa Mod_null NA NA NA NA 369.73 0.53 Lissotriton helveticus Mod_use -4.94 12.416 NA NA 26.28 0.22 Lissotriton helveticus Mod_clim NA NA -0.43 5.541 26.54 0.19 Lissotriton helveticus Mod_clim_use -4.95 12.447 0.04 5.349 28.28 0.08 Lissotriton helveticus Mod_null NA NA NA NA 24.55 0.51 Jacana spinosa Mod_use -20.89 9.011 NA NA 52.85 0.61 Jacana spinosa Mod_clim NA NA 1.47 4.743 58.14 0.04 Jacana spinosa Mod_clim_use -20.65 9.160 1.03 5.362 54.81 0.23 Jacana spinosa Mod_null NA NA NA NA 56.24 0.11 201

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Regina septemvittata Mod_use -1.07 8.491 NA NA 23.26 0.20 Regina septemvittata Mod_clim NA NA -1.03 10.319 23.27 0.20 Regina septemvittata Mod_clim_use -1.37 8.280 -1.42 10.754 25.24 0.07 Regina septemvittata Mod_null NA NA NA NA 21.28 0.53 Alectrurus tricolor Mod_use 0.14 0.650 NA NA 3515.68 0.00 Alectrurus tricolor Mod_clim NA NA -2.37 0.523 3491.26 0.70 Alectrurus tricolor Mod_clim_use -0.38 0.671 -2.40 0.524 3492.97 0.30 Alectrurus tricolor Mod_null NA NA NA NA 3513.73 0.00 Alytes obstetricans Mod_use -1.61 6.799 NA NA 33.36 0.27 Alytes obstetricans Mod_clim NA NA NA NA NA NA Alytes obstetricans Mod_clim_use NA NA NA NA NA NA Alytes obstetricans Mod_null NA NA NA NA 31.42 0.73 Penelope pileata Mod_use 0.42 1.580 NA NA 450.08 0.10 Penelope pileata Mod_clim NA NA 2.42 1.590 447.09 0.46 Penelope pileata Mod_clim_use 0.20 1.588 2.40 1.596 449.07 0.17 Penelope pileata Mod_null NA NA NA NA 448.15 0.27 Picoides borealis Mod_use 1.36 0.488 NA NA 3792.63 0.28 Picoides borealis Mod_clim NA NA -1.19 0.566 3795.22 0.08 Picoides borealis Mod_clim_use 1.23 0.496 -1.00 0.563 3791.04 0.62 Picoides borealis Mod_null NA NA NA NA 3798.41 0.02 Tapirus bairdii Mod_use 1.17 0.798 NA NA 1149.63 0.20 Tapirus bairdii Mod_clim NA NA 1.24 0.690 1148.43 0.36 Tapirus bairdii Mod_clim_use 0.91 0.817 1.09 0.705 1149.21 0.25 Tapirus bairdii Mod_null NA NA NA NA 1149.72 0.19 202

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Poephila cincta Mod_use 6.09 0.000 NA NA 2938.46 0.00 Poephila cincta Mod_clim NA NA -123.94 0.033 1918.26 0.00 Poephila cincta Mod_clim_use 79.75 0.030 -126.87 0.028 1884.33 1.00 Poephila cincta Mod_null NA NA NA NA 2943.50 0.00 Alectrurus risora Mod_use 2.72 1.187 NA NA 880.37 0.00 Alectrurus risora Mod_clim NA NA -10.94 2.008 856.53 0.65 Alectrurus risora Mod_clim_use -1.48 1.553 -11.31 3.999 857.74 0.35 Alectrurus risora Mod_null NA NA NA NA 884.28 0.00 Picus squamatus Mod_use -17.95 9.400 NA NA 966.18 0.65 Picus squamatus Mod_clim NA NA 0.56 0.732 975.35 0.01 Picus squamatus Mod_clim_use -17.67 9.403 0.59 0.716 967.54 0.33 Picus squamatus Mod_null NA NA NA NA 973.92 0.01 Gubernatrix cristata Mod_use 1.04 1.576 NA NA 296.60 0.31 Gubernatrix cristata Mod_clim NA NA NA NA NA NA Gubernatrix cristata Mod_clim_use NA NA NA NA NA NA Gubernatrix cristata Mod_null NA NA NA NA 295.01 0.69 Pseudocolopteryx dinelliana Mod_use -4.98 18.119 NA NA 25.87 0.20 Pseudocolopteryx dinelliana Mod_clim NA NA 4.08 0.002 25.75 0.22 Pseudocolopteryx dinelliana Mod_clim_use -3.47 20.312 3.34 9.714 27.70 0.08 Pseudocolopteryx dinelliana Mod_null NA NA NA NA 24.08 0.50 Saxicola macrorhynchus Mod_use 4.29 1.664 NA NA 969.31 0.50 Saxicola macrorhynchus Mod_clim NA NA -2.48 1.131 979.37 0.00 Saxicola macrorhynchus Mod_clim_use 3.96 1.647 -1.84 1.309 969.31 0.50 Saxicola macrorhynchus Mod_null NA NA NA NA 982.08 0.00 203

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Empidonax fulvifrons Mod_use -3.35 2.638 NA NA 1303.66 0.38 Empidonax fulvifrons Mod_clim NA NA -0.53 0.797 1305.60 0.15 Empidonax fulvifrons Mod_clim_use -3.22 2.643 -0.37 0.800 1305.44 0.16 Empidonax fulvifrons Mod_null NA NA NA NA 1304.07 0.31 Cisticola eximius Mod_use 0.81 2.024 NA NA 224.74 0.13 Cisticola eximius Mod_clim NA NA -11.66 15.595 221.74 0.56 Cisticola eximius Mod_clim_use NA NA NA NA NA NA Cisticola eximius Mod_null NA NA NA NA 222.89 0.31 Megaxenops parnaguae Mod_use 6.51 2.272 NA NA 364.46 0.66 Megaxenops parnaguae Mod_clim NA NA -0.25 1.130 371.00 0.03 Megaxenops parnaguae Mod_clim_use 6.56 2.295 0.18 1.174 366.44 0.25 Megaxenops parnaguae Mod_null NA NA NA NA 369.05 0.07 Anodorhynchus hyacinthinus Mod_use 0.03 1.943 NA NA 1370.61 0.20 Anodorhynchus hyacinthinus Mod_clim NA NA 0.03 1.064 1370.61 0.20 Anodorhynchus hyacinthinus Mod_clim_use 0.04 1.941 0.03 1.065 1372.61 0.07 Anodorhynchus hyacinthinus Mod_null NA NA NA NA 1368.61 0.53 Agriornis albicauda Mod_use -8.36 18.087 NA NA 738.90 0.22 Agriornis albicauda Mod_clim NA NA -0.36 0.958 739.13 0.20 Agriornis albicauda Mod_clim_use -7.65 18.546 -0.26 0.975 740.83 0.08 Agriornis albicauda Mod_null NA NA NA NA 737.28 0.50 Xanthopsar flavus Mod_use 0.69 0.001 NA NA 987.11 0.13 Xanthopsar flavus Mod_clim NA NA -1.32 0.770 984.79 0.42 Xanthopsar flavus Mod_clim_use -0.46 1.231 -1.57 1.039 986.65 0.17 Xanthopsar flavus Mod_null NA NA NA NA 985.64 0.28 204

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Vireo atricapilla Mod_use 3.72 1.136 NA NA 1991.10 0.00 Vireo atricapilla Mod_clim NA NA -36.23 2.365 1994.20 0.00 Vireo atricapilla Mod_clim_use 16.00 1.737 -21.17 1.934 1963.92 1.00 Vireo atricapilla Mod_null NA NA NA NA 1999.78 0.00 Sus barbatus Mod_use -0.37 4.216 NA NA 224.85 0.19 Sus barbatus Mod_clim NA NA -1.42 2.330 224.60 0.22 Sus barbatus Mod_clim_use 0.77 4.540 -1.25 2.350 226.52 0.08 Sus barbatus Mod_null NA NA NA NA 222.86 0.51 Pelobates syriacus Mod_use -0.30 6.549 NA NA 33.05 0.19 Pelobates syriacus Mod_clim NA NA 1.60 5.079 32.95 0.20 Pelobates syriacus Mod_clim_use -0.31 6.740 1.61 5.096 34.95 0.08 Pelobates syriacus Mod_null NA NA NA NA 31.05 0.53 Cyanolanius madagascarinus Mod_use -0.17 31.187 NA NA 12.88 0.20 Cyanolanius madagascarinus Mod_clim NA NA -0.21 11.235 12.88 0.20 Cyanolanius madagascarinus Mod_clim_use -0.16 31.189 -0.20 11.253 14.88 0.07 Cyanolanius madagascarinus Mod_null NA NA NA NA 10.88 0.53 Procolobus badius Mod_use 1.56 1.022 NA NA 835.23 0.00 Procolobus badius Mod_clim NA NA 4.03 1.301 823.34 0.41 Procolobus badius Mod_clim_use 2.00 1.165 4.12 1.287 822.61 0.59 Procolobus badius Mod_null NA NA NA NA 835.35 0.00 Platyrinchus leucoryphus Mod_use 1.05 2.083 NA NA 170.00 0.22 Platyrinchus leucoryphus Mod_clim NA NA -0.24 1.938 170.25 0.19 Platyrinchus leucoryphus Mod_clim_use 1.08 2.083 -0.32 1.958 171.98 0.08 Platyrinchus leucoryphus Mod_null NA NA NA NA 168.26 0.51 205

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Jacamaralcyon tridactyla Mod_use 1.81 1.331 NA NA 831.03 0.11 Jacamaralcyon tridactyla Mod_clim NA NA -1.52 0.807 828.81 0.33 Jacamaralcyon tridactyla Mod_clim_use 2.21 1.384 -1.62 0.789 828.22 0.44 Jacamaralcyon tridactyla Mod_null NA NA NA NA 830.90 0.12 Grus leucogeranus Mod_use NA NA NA NA NA NA Grus leucogeranus Mod_clim NA NA 1.34 3.238 132.21 0.21 Grus leucogeranus Mod_clim_use 606.63 295.604 1.34 3.228 134.20 0.08 Grus leucogeranus Mod_null NA NA NA NA 130.38 0.52 Gymnogyps californianus Mod_use 1.57 1.120 NA NA 1293.92 0.34 Gymnogyps californianus Mod_clim NA NA -0.32 0.000 1295.92 0.13 Gymnogyps californianus Mod_clim_use 1.92 1.172 -0.67 0.609 1294.68 0.24 Gymnogyps californianus Mod_null NA NA NA NA 1294.26 0.29 Pleurodeles waltl Mod_use 0.06 3.365 NA NA 59.77 0.19 Pleurodeles waltl Mod_clim NA NA 2.19 4.761 59.52 0.21 Pleurodeles waltl Mod_clim_use -0.52 3.459 2.42 5.031 61.50 0.08 Pleurodeles waltl Mod_null NA NA NA NA 57.77 0.52 Fulica cornuta Mod_use -0.85 14.829 NA NA 1763.39 0.03 Fulica cornuta Mod_clim NA NA 1.40 0.539 1757.18 0.64 Fulica cornuta Mod_clim_use -4.13 15.093 1.42 0.542 1759.08 0.25 Fulica cornuta Mod_null NA NA NA NA 1761.40 0.08 Ara militaris Mod_use 1.20 0.001 NA NA 963.27 0.25 Ara militaris Mod_clim NA NA 0.22 0.757 963.91 0.18 Ara militaris Mod_clim_use 1.18 0.001 0.05 0.001 965.27 0.09 Ara militaris Mod_null NA NA NA NA 962.00 0.47 206

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Phrynosoma douglasii Mod_use -2.83 21.086 NA NA 32.98 0.20 Phrynosoma douglasii Mod_clim NA NA -3.80 16.917 32.89 0.21 Phrynosoma douglasii Mod_clim_use -0.93 19.911 -3.64 17.250 34.88 0.08 Phrynosoma douglasii Mod_null NA NA NA NA 31.01 0.52 Guaruba guarouba Mod_use 2.09 1.554 NA NA 1114.38 0.09 Guaruba guarouba Mod_clim NA NA 1.67 0.738 1111.77 0.33 Guaruba guarouba Mod_clim_use 2.40 1.562 1.79 0.748 1110.90 0.50 Guaruba guarouba Mod_null NA NA NA NA 1114.53 0.08 Axis porcinus Mod_use -1.74 0.917 NA NA 905.16 0.46 Axis porcinus Mod_clim NA NA 0.90 0.990 908.18 0.10 Axis porcinus Mod_clim_use -1.73 0.920 0.79 0.917 906.45 0.24 Axis porcinus Mod_null NA NA NA NA 906.95 0.19 Cephalophus harveyi Mod_use 2.16 4.702 NA NA 62.51 0.29 Cephalophus harveyi Mod_clim NA NA NA NA NA NA Cephalophus harveyi Mod_clim_use NA NA NA NA NA NA Cephalophus harveyi Mod_null NA NA NA NA 60.68 0.71 Cervus nippon Mod_use 0.88 13.124 NA NA 16.23 0.27 Cervus nippon Mod_clim NA NA NA NA NA NA Cervus nippon Mod_clim_use NA NA NA NA NA NA Cervus nippon Mod_null NA NA NA NA 14.24 0.73 Asthenes anthoides Mod_use 178.35 191.884 NA NA 10.22 0.73 Asthenes anthoides Mod_clim NA NA NA NA NA NA Asthenes anthoides Mod_clim_use 178.35 193.778 73.37 708.313 12.22 0.27 Asthenes anthoides Mod_null NA NA NA NA 34.13 0.00 207

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Gallinago stricklandii Mod_use NA NA NA NA NA NA Gallinago stricklandii Mod_clim NA NA NA NA NA NA Gallinago stricklandii Mod_clim_use NA NA NA NA NA NA Gallinago stricklandii Mod_null NA NA NA NA 164.62 1.00 Houbaropsis bengalensis Mod_use 0.71 0.000 NA NA 1268.57 0.24 Houbaropsis bengalensis Mod_clim NA NA 0.66 0.696 1268.68 0.23 Houbaropsis bengalensis Mod_clim_use 0.59 0.000 0.52 0.000 1269.97 0.12 Houbaropsis bengalensis Mod_null NA NA NA NA 1267.58 0.40 Francolinus gularis Mod_use -2.16 0.825 NA NA 1353.63 0.00 Francolinus gularis Mod_clim NA NA 21.91 1.148 1430.51 0.00 Francolinus gularis Mod_clim_use -3.44 0.885 4.16 0.936 1327.43 1.00 Francolinus gularis Mod_null NA NA NA NA 1359.96 0.00 Cyanoramphus novaezelandiae Mod_use NA NA NA NA NA NA Cyanoramphus novaezelandiae Mod_clim NA NA 1.70 5.638 39.26 0.21 Cyanoramphus novaezelandiae Mod_clim_use NA NA NA NA NA NA Cyanoramphus novaezelandiae Mod_null NA NA NA NA 37.33 0.55 Microhierax erythrogenys Mod_use 6.03 10.370 NA NA 58.33 0.26 Microhierax erythrogenys Mod_clim NA NA 0.41 3.156 59.07 0.18 Microhierax erythrogenys Mod_clim_use 6.04 10.306 -0.10 3.129 60.32 0.09 Microhierax erythrogenys Mod_null NA NA NA NA 57.08 0.47 Conilurus penicillatus Mod_use -5.27 7.314 NA NA 204.92 0.17 Conilurus penicillatus Mod_clim NA NA -3.32 3.030 203.97 0.28 Conilurus penicillatus Mod_clim_use -3.95 14.886 -3.27 3.356 205.90 0.11 Conilurus penicillatus Mod_null NA NA NA NA 203.05 0.44 208

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Nyctimene cephalotes Mod_use 1.64 1.932 NA NA 252.44 0.25 Nyctimene cephalotes Mod_clim NA NA 0.10 1.406 253.09 0.18 Nyctimene cephalotes Mod_clim_use 1.79 2.043 -0.30 1.481 254.40 0.09 Nyctimene cephalotes Mod_null NA NA NA NA 251.10 0.48 Coryphaspiza melanotis Mod_use 4.33 1.560 NA NA 700.57 0.35 Coryphaspiza melanotis Mod_clim NA NA -3.51 2.133 703.16 0.10 Coryphaspiza melanotis Mod_clim_use 3.70 1.562 -3.09 2.235 699.78 0.52 Coryphaspiza melanotis Mod_null NA NA NA NA 705.51 0.03 Thylogale browni Mod_use NA NA NA NA NA NA Thylogale browni Mod_clim NA NA -2.53 13.638 26.05 0.28 Thylogale browni Mod_clim_use NA NA NA NA NA NA Thylogale browni Mod_null NA NA NA NA 24.12 0.72 Cacatua haematuropygia Mod_use 3.01 2.141 NA NA 307.66 0.33 Cacatua haematuropygia Mod_clim NA NA -1.01 1.231 309.66 0.12 Cacatua haematuropygia Mod_clim_use 3.83 2.269 -1.78 1.332 307.75 0.32 Cacatua haematuropygia Mod_null NA NA NA NA 308.35 0.23 Psophodes nigrogularis Mod_use 0.98 0.924 NA NA 786.97 0.01 Psophodes nigrogularis Mod_clim NA NA 3.04 1.179 779.40 0.66 Psophodes nigrogularis Mod_clim_use 0.64 0.946 2.99 1.200 780.94 0.30 Psophodes nigrogularis Mod_null NA NA NA NA 786.10 0.02 Patagioenas inornata Mod_use 4.33 0.000 NA NA 837.40 0.73 Patagioenas inornata Mod_clim NA NA 0.36 0.608 863.48 0.00 Patagioenas inornata Mod_clim_use 4.36 0.000 -0.09 0.000 839.38 0.27 Patagioenas inornata Mod_null NA NA NA NA 861.82 0.00 209

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Lipaugus lanioides Mod_use 4.21 0.001 NA NA 48.66 0.23 Lipaugus lanioides Mod_clim NA NA 0.27 3.722 49.04 0.19 Lipaugus lanioides Mod_clim_use 4.38 0.001 0.45 0.001 50.64 0.08 Lipaugus lanioides Mod_null NA NA NA NA 47.05 0.50 Pyrilia pyrilia Mod_use 2.69 0.897 NA NA 918.15 0.00 Pyrilia pyrilia Mod_clim NA NA 2.02 0.475 907.19 0.03 Pyrilia pyrilia Mod_clim_use 2.71 0.965 2.02 0.490 900.51 0.97 Pyrilia pyrilia Mod_null NA NA NA NA 925.70 0.00 Hymenolaimus malacorhynchos Mod_use 9.62 2.995 NA NA 1390.67 0.00 Hymenolaimus malacorhynchos Mod_clim NA NA 2.10 0.441 1380.82 0.01 Hymenolaimus malacorhynchos Mod_clim_use 7.95 2.893 1.87 0.443 1372.31 0.99 Hymenolaimus malacorhynchos Mod_null NA NA NA NA 1405.26 0.00 Laniisoma elegans Mod_use 0.13 2.166 NA NA 305.14 0.05 Laniisoma elegans Mod_clim NA NA 3.74 2.455 300.37 0.58 Laniisoma elegans Mod_clim_use 0.00 2.526 3.74 2.457 302.37 0.22 Laniisoma elegans Mod_null NA NA NA NA 303.14 0.15 Hypothymis coelestis Mod_use 2.53 3.409 NA NA 145.50 0.25 Hypothymis coelestis Mod_clim NA NA 0.53 1.910 146.13 0.18 Hypothymis coelestis Mod_clim_use 2.47 3.443 0.26 1.939 147.48 0.09 Hypothymis coelestis Mod_null NA NA NA NA 144.21 0.48 Bos javanicus Mod_use -3.21 0.001 NA NA 55.73 0.18 Bos javanicus Mod_clim NA NA -9.77 25.244 55.14 0.24 Bos javanicus Mod_clim_use -3.46 19.590 -9.76 24.979 57.10 0.09 Bos javanicus Mod_null NA NA NA NA 53.78 0.48 210

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Amazona oratrix Mod_use 1.49 1.237 NA NA 592.47 0.06 Amazona oratrix Mod_clim NA NA -2.51 1.071 588.15 0.51 Amazona oratrix Mod_clim_use 1.46 1.303 -2.47 1.075 588.83 0.36 Amazona oratrix Mod_null NA NA NA NA 592.00 0.07 Meleagris ocellata Mod_use 2.06 1.237 NA NA 412.17 0.41 Meleagris ocellata Mod_clim NA NA -0.25 1.171 414.72 0.12 Meleagris ocellata Mod_clim_use 2.11 1.241 -0.44 1.190 414.04 0.16 Meleagris ocellata Mod_null NA NA NA NA 412.76 0.31 Lithobates warszewitschii Mod_use 3.82 4.037 NA NA 43.52 0.28 Lithobates warszewitschii Mod_clim NA NA -0.61 2.010 44.55 0.17 Lithobates warszewitschii Mod_clim_use 3.73 4.104 -0.19 2.083 45.51 0.10 Lithobates warszewitschii Mod_null NA NA NA NA 42.64 0.44 Amazona vinacea Mod_use 1.57 1.548 NA NA 565.24 0.06 Amazona vinacea Mod_clim NA NA 2.56 1.181 561.00 0.53 Amazona vinacea Mod_clim_use 1.55 1.669 2.57 1.209 562.16 0.30 Amazona vinacea Mod_null NA NA NA NA 564.23 0.11 Acerodon jubatus Mod_use 1.13 3.085 NA NA 111.61 0.20 Acerodon jubatus Mod_clim NA NA -1.04 1.943 111.43 0.21 Acerodon jubatus Mod_clim_use 1.61 3.199 -1.24 1.962 113.16 0.09 Acerodon jubatus Mod_null NA NA NA NA 109.74 0.50 Iodopleura pipra Mod_use 1.05 2.863 NA NA 222.76 0.20 Iodopleura pipra Mod_clim NA NA -0.40 1.580 222.83 0.20 Iodopleura pipra Mod_clim_use 1.03 2.823 -0.40 1.575 224.70 0.08 Iodopleura pipra Mod_null NA NA NA NA 220.89 0.52 211

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Poliocephalus rufopectus Mod_use -8.22 7.942 NA NA 568.23 0.00 Poliocephalus rufopectus Mod_clim NA NA -96.23 0.001 524.48 0.56 Poliocephalus rufopectus Mod_clim_use -9.19 0.001 -93.31 0.001 524.97 0.44 Poliocephalus rufopectus Mod_null NA NA NA NA 568.39 0.00 Timon tangitanus Mod_use NA NA NA NA NA NA Timon tangitanus Mod_clim NA NA 6.82 0.002 19.24 0.30 Timon tangitanus Mod_clim_use NA NA NA NA NA NA Timon tangitanus Mod_null NA NA NA NA 17.52 0.70 Thripophaga macroura Mod_use 3.24 1.426 NA NA 358.28 0.64 Thripophaga macroura Mod_clim NA NA -0.07 0.736 364.18 0.03 Thripophaga macroura Mod_clim_use 3.23 1.424 -0.10 0.768 360.26 0.24 Thripophaga macroura Mod_null NA NA NA NA 362.19 0.09 Choeropsis liberiensis Mod_use -2.47 4.317 NA NA 167.76 0.05 Choeropsis liberiensis Mod_clim NA NA 16.85 20.063 162.12 0.84 Choeropsis liberiensis Mod_clim_use NA NA NA NA NA NA Choeropsis liberiensis Mod_null NA NA NA NA 166.25 0.11 Lacerta schreiberi Mod_use -7.96 23.766 NA NA 28.97 0.21 Lacerta schreiberi Mod_clim NA NA 0.91 5.812 29.18 0.19 Lacerta schreiberi Mod_clim_use -7.92 0.001 0.60 0.001 30.96 0.08 Lacerta schreiberi Mod_null NA NA NA NA 27.21 0.52 Crax blumenbachii Mod_use 3.08 3.159 NA NA 79.98 0.00 Crax blumenbachii Mod_clim NA NA -31.14 17.593 67.11 0.73 Crax blumenbachii Mod_clim_use 0.42 4.644 -30.91 18.625 69.10 0.27 Crax blumenbachii Mod_null NA NA NA NA 79.00 0.00 212

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Babyrousa celebensis Mod_use 1.53 2.827 NA NA 176.81 0.21 Babyrousa celebensis Mod_clim NA NA -0.47 1.601 176.96 0.19 Babyrousa celebensis Mod_clim_use 2.04 2.937 -0.80 1.678 178.57 0.09 Babyrousa celebensis Mod_null NA NA NA NA 175.06 0.51 Padda oryzivora Mod_use 0.18 1.822 NA NA 372.26 0.18 Padda oryzivora Mod_clim NA NA -0.89 1.182 371.64 0.24 Padda oryzivora Mod_clim_use 0.39 1.824 -0.93 1.189 373.59 0.09 Padda oryzivora Mod_null NA NA NA NA 370.27 0.48 Amazona leucocephala Mod_use 1.16 1.140 NA NA 372.17 0.00 Amazona leucocephala Mod_clim NA NA -4.83 2.199 361.89 0.57 Amazona leucocephala Mod_clim_use 1.45 1.230 -4.68 2.024 362.52 0.42 Amazona leucocephala Mod_null NA NA NA NA 371.20 0.01 Wilfredomys oenax Mod_use -10.61 22.333 NA NA 58.15 0.23 Wilfredomys oenax Mod_clim NA NA 1.57 6.161 58.50 0.19 Wilfredomys oenax Mod_clim_use -10.79 23.167 1.75 6.642 60.07 0.09 Wilfredomys oenax Mod_null NA NA NA NA 56.57 0.50 Microtus cabrerae Mod_use -2.28 6.717 NA NA 65.73 0.21 Microtus cabrerae Mod_clim NA NA 0.05 2.367 65.88 0.19 Microtus cabrerae Mod_clim_use -2.28 6.712 0.04 2.367 67.73 0.08 Microtus cabrerae Mod_null NA NA NA NA 63.88 0.52 Geotrygon caniceps Mod_use 4.28 1.015 NA NA 593.23 0.13 Geotrygon caniceps Mod_clim NA NA 2.46 0.825 603.90 0.00 Geotrygon caniceps Mod_clim_use 3.99 1.032 2.02 0.859 589.34 0.87 Geotrygon caniceps Mod_null NA NA NA NA 612.11 0.00 213

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Accipiter gundlachi Mod_use 5.99 1.563 NA NA 323.95 0.58 Accipiter gundlachi Mod_clim NA NA 2.40 1.021 341.24 0.00 Accipiter gundlachi Mod_clim_use 5.47 1.563 1.18 1.036 324.62 0.42 Accipiter gundlachi Mod_null NA NA NA NA 345.71 0.00 Tyrannus cubensis Mod_use 0.10 1.663 NA NA 136.30 0.15 Tyrannus cubensis Mod_clim NA NA -1.53 1.328 134.84 0.30 Tyrannus cubensis Mod_clim_use 1.43 1.895 -2.07 1.536 136.28 0.15 Tyrannus cubensis Mod_null NA NA NA NA 134.30 0.40 Aratinga euops Mod_use 2.74 1.101 NA NA 370.77 0.65 Aratinga euops Mod_clim NA NA 0.17 0.739 377.10 0.03 Aratinga euops Mod_clim_use 2.80 1.130 -0.19 0.764 372.71 0.25 Aratinga euops Mod_null NA NA NA NA 375.16 0.07 Starnoenas cyanocephala Mod_use 4.65 0.915 NA NA 716.63 0.32 Starnoenas cyanocephala Mod_clim NA NA 1.55 0.618 742.82 0.00 Starnoenas cyanocephala Mod_clim_use 4.48 0.907 1.25 0.676 715.13 0.68 Starnoenas cyanocephala Mod_null NA NA NA NA 747.32 0.00 Rallus antarcticus Mod_use 6.17 6.190 NA NA 22.02 0.33 Rallus antarcticus Mod_clim NA NA NA NA NA NA Rallus antarcticus Mod_clim_use NA NA NA NA NA NA Rallus antarcticus Mod_null NA NA NA NA 20.62 0.67 Rusa marianna Mod_use 2.21 3.927 NA NA 64.54 0.21 Rusa marianna Mod_clim NA NA 1.88 3.193 64.46 0.22 Rusa marianna Mod_clim_use 1.81 3.969 1.66 3.276 66.25 0.09 Rusa marianna Mod_null NA NA NA NA 62.86 0.48 214

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Thylogale brunii Mod_use 6.68 5.829 NA NA 112.20 0.24 Thylogale brunii Mod_clim NA NA -2.64 3.158 112.44 0.22 Thylogale brunii Mod_clim_use 6.75 6.152 -2.63 2.997 113.42 0.13 Thylogale brunii Mod_null NA NA NA NA 111.15 0.41 Chondrohierax wilsonii Mod_use 6.54 3.704 NA NA 83.57 0.58 Chondrohierax wilsonii Mod_clim NA NA 1.02 2.818 88.19 0.06 Chondrohierax wilsonii Mod_clim_use 6.55 3.719 -0.04 2.862 85.57 0.21 Chondrohierax wilsonii Mod_null NA NA NA NA 86.33 0.15 Colaptes fernandinae Mod_use 6.47 1.814 NA NA 247.24 0.58 Colaptes fernandinae Mod_clim NA NA 0.29 1.126 267.55 0.00 Colaptes fernandinae Mod_clim_use 7.28 2.069 -1.51 1.368 247.87 0.42 Colaptes fernandinae Mod_null NA NA NA NA 265.62 0.00 Pachycephala rufogularis Mod_use -0.97 2.846 NA NA 105.70 0.22 Pachycephala rufogularis Mod_clim NA NA -0.36 2.978 105.81 0.21 Pachycephala rufogularis Mod_clim_use -3.31 0.001 -0.38 0.001 110.76 0.02 Pachycephala rufogularis Mod_null NA NA NA NA 103.82 0.56 Ara ambiguus Mod_use 3.02 3.578 NA NA 82.86 0.23 Ara ambiguus Mod_clim NA NA -1.38 2.883 83.23 0.19 Ara ambiguus Mod_clim_use 3.93 3.728 -2.12 2.998 84.26 0.12 Ara ambiguus Mod_null NA NA NA NA 81.50 0.46 Carduelis cucullata Mod_use -1.77 2.357 NA NA 209.82 0.10 Carduelis cucullata Mod_clim NA NA -2.90 1.820 206.54 0.50 Carduelis cucullata Mod_clim_use -1.01 2.403 -2.81 1.848 208.36 0.20 Carduelis cucullata Mod_null NA NA NA NA 208.37 0.20 215

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Xiphocolaptes falcirostris Mod_use 1.09 8.942 NA NA 167.10 0.18 Xiphocolaptes falcirostris Mod_clim NA NA 1.72 2.235 166.42 0.25 Xiphocolaptes falcirostris Mod_clim_use -0.16 10.370 1.73 2.259 168.42 0.09 Xiphocolaptes falcirostris Mod_null NA NA NA NA 165.12 0.48 Tangara peruviana Mod_use 0.33 3.034 NA NA 161.20 0.00 Tangara peruviana Mod_clim NA NA -12.58 4.410 146.54 0.72 Tangara peruviana Mod_clim_use 1.13 3.424 -12.46 4.755 148.43 0.28 Tangara peruviana Mod_null NA NA NA NA 159.21 0.00 Equus africanus Mod_use -1.52 7.992 NA NA 195.87 0.16 Equus africanus Mod_clim NA NA -1.83 1.799 194.69 0.29 Equus africanus Mod_clim_use -1.23 8.294 -1.82 1.801 196.67 0.11 Equus africanus Mod_null NA NA NA NA 193.90 0.43 Callaeas cinereus Mod_use 3.76 8.131 NA NA 215.90 0.12 Callaeas cinereus Mod_clim NA NA 2.31 1.628 213.48 0.42 Callaeas cinereus Mod_clim_use -1.43 7.644 2.43 1.768 215.44 0.16 Callaeas cinereus Mod_null NA NA NA NA 214.14 0.30 Lithobates tarahumarae Mod_use -9.41 0.001 NA NA 29.20 0.20 Lithobates tarahumarae Mod_clim NA NA -2.99 0.001 29.12 0.20 Lithobates tarahumarae Mod_clim_use -4.62 67.806 -2.81 13.164 31.12 0.08 Lithobates tarahumarae Mod_null NA NA NA NA 27.23 0.52 Anthus nattereri Mod_use 2.54 1.980 NA NA 304.82 0.00 Anthus nattereri Mod_clim NA NA -47.05 4.461 227.62 0.67 Anthus nattereri Mod_clim_use 1.94 2.660 -46.74 5.090 229.06 0.33 Anthus nattereri Mod_null NA NA NA NA 304.46 0.00 216

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Zaglossus bruijnii Mod_use -5.43 133.831 NA NA 29.25 0.20 Zaglossus bruijnii Mod_clim NA NA 0.50 0.001 29.24 0.20 Zaglossus bruijnii Mod_clim_use -6.78 0.001 0.53 0.001 31.24 0.07 Zaglossus bruijnii Mod_null NA NA NA NA 27.25 0.53 Nestor meridionalis Mod_use 0.06 8.216 NA NA 280.58 0.18 Nestor meridionalis Mod_clim NA NA 0.84 1.089 279.99 0.24 Nestor meridionalis Mod_clim_use -1.26 9.316 0.88 1.111 281.97 0.09 Nestor meridionalis Mod_null NA NA NA NA 278.58 0.49 Corvus leucognaphalus Mod_use 6.51 1.316 NA NA 314.82 0.00 Corvus leucognaphalus Mod_clim NA NA 8.21 2.279 299.09 0.00 Corvus leucognaphalus Mod_clim_use 5.08 1.487 7.16 2.414 285.84 1.00 Corvus leucognaphalus Mod_null NA NA NA NA 348.93 0.00 Cotinga maculata Mod_use -2.63 8.581 NA NA 32.03 0.19 Cotinga maculata Mod_clim NA NA -12.04 0.002 31.69 0.23 Cotinga maculata Mod_clim_use -3.07 9.177 -13.18 21.689 33.58 0.09 Cotinga maculata Mod_null NA NA NA NA 30.11 0.50 Sporophila frontalis Mod_use 4.20 2.972 NA NA 106.63 0.30 Sporophila frontalis Mod_clim NA NA 2.62 2.870 107.72 0.17 Sporophila frontalis Mod_clim_use 5.44 3.721 3.40 3.221 106.88 0.27 Sporophila frontalis Mod_null NA NA NA NA 106.93 0.26 Picumnus fulvescens Mod_use 2.66 0.002 NA NA 16.02 0.20 Picumnus fulvescens Mod_clim NA NA 0.36 9.075 16.11 0.19 Picumnus fulvescens Mod_clim_use 2.87 9.695 -0.55 9.916 18.02 0.07 Picumnus fulvescens Mod_null NA NA NA NA 14.11 0.53 217

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Penelope ortoni Mod_use 58.31 0.001 NA NA 317.26 0.57 Penelope ortoni Mod_clim NA NA 1.14 0.990 341.53 0.00 Penelope ortoni Mod_clim_use 58.41 12.070 1.39 1.180 317.83 0.43 Penelope ortoni Mod_null NA NA NA NA 340.90 0.00 Chloropeta gracilirostris Mod_use 1.30 9.936 NA NA 19.20 0.27 Chloropeta gracilirostris Mod_clim NA NA NA NA NA NA Chloropeta gracilirostris Mod_clim_use NA NA NA NA NA NA Chloropeta gracilirostris Mod_null NA NA NA NA 17.22 0.73 Pituophis ruthveni Mod_use 21.03 9.240 NA NA 91.88 0.02 Pituophis ruthveni Mod_clim NA NA 22.22 9.938 89.34 0.06 Pituophis ruthveni Mod_clim_use 23.54 12.004 20.28 10.937 83.87 0.92 Pituophis ruthveni Mod_null NA NA NA NA 100.31 0.00 Penelope ochrogaster Mod_use 0.57 7.489 NA NA 382.98 0.20 Penelope ochrogaster Mod_clim NA NA 0.28 1.885 382.96 0.20 Penelope ochrogaster Mod_clim_use 0.86 7.854 0.33 1.925 384.95 0.07 Penelope ochrogaster Mod_null NA NA NA NA 380.99 0.53 Zaratornis stresemanni Mod_use 22.48 26.311 NA NA 28.49 0.23 Zaratornis stresemanni Mod_clim NA NA 1.17 5.038 28.90 0.19 Zaratornis stresemanni Mod_clim_use 23.92 36.999 -0.42 7.313 30.48 0.08 Zaratornis stresemanni Mod_null NA NA NA NA 26.95 0.50 Ptilinopus jambu Mod_use 58.30 0.001 NA NA 200.03 0.68 Ptilinopus jambu Mod_clim NA NA 2.20 1.346 259.56 0.00 Ptilinopus jambu Mod_clim_use 45.51 17.979 2.66 3.749 201.58 0.32 Ptilinopus jambu Mod_null NA NA NA NA 260.17 0.00 218

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Carpornis melanocephala Mod_use 2.77 1.836 NA NA 356.20 0.05 Carpornis melanocephala Mod_clim NA NA 3.12 1.398 351.94 0.40 Carpornis melanocephala Mod_clim_use 2.84 1.860 3.06 1.408 351.45 0.51 Carpornis melanocephala Mod_null NA NA NA NA 356.54 0.04 Amazona viridigenalis Mod_use 0.56 1.633 NA NA 205.09 0.20 Amazona viridigenalis Mod_clim NA NA -0.47 1.250 205.06 0.20 Amazona viridigenalis Mod_clim_use 0.68 1.662 -0.55 1.244 206.90 0.08 Amazona viridigenalis Mod_null NA NA NA NA 203.21 0.51 Crypturellus erythropus Mod_use 9.45 2.232 NA NA 307.00 0.28 Crypturellus erythropus Mod_clim NA NA 1.47 0.722 336.29 0.00 Crypturellus erythropus Mod_clim_use 9.39 2.230 1.56 0.819 305.12 0.72 Crypturellus erythropus Mod_null NA NA NA NA 338.85 0.00 Aratinga chloroptera Mod_use 3.51 1.241 NA NA 387.26 0.25 Aratinga chloroptera Mod_clim NA NA 1.60 0.654 389.88 0.07 Aratinga chloroptera Mod_clim_use 3.07 1.251 1.34 0.680 385.24 0.68 Aratinga chloroptera Mod_null NA NA NA NA 394.16 0.01 Parus nuchalis Mod_use 1.66 6.236 NA NA 51.72 0.00 Parus nuchalis Mod_clim NA NA 339.94 9.618 28.45 0.73 Parus nuchalis Mod_clim_use -0.49 13.863 340.20 19.749 30.45 0.27 Parus nuchalis Mod_null NA NA NA NA 49.82 0.00 Mastacomys fuscus Mod_use 1.08 10.381 NA NA 35.50 0.19 Mastacomys fuscus Mod_clim NA NA 3.59 0.002 35.09 0.23 Mastacomys fuscus Mod_clim_use -1.42 0.002 3.84 0.002 37.07 0.08 Mastacomys fuscus Mod_null NA NA NA NA 33.51 0.50 219

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Aphelocoma coerulescens Mod_use 0.91 1.028 NA NA 355.76 0.25 Aphelocoma coerulescens Mod_clim NA NA 0.43 1.193 356.39 0.18 Aphelocoma coerulescens Mod_clim_use 0.90 1.117 0.02 1.269 357.76 0.09 Aphelocoma coerulescens Mod_null NA NA NA NA 354.52 0.47 Sus philippensis Mod_use -0.32 10.319 NA NA 19.27 0.20 Sus philippensis Mod_clim NA NA -0.21 6.060 19.27 0.20 Sus philippensis Mod_clim_use -0.21 11.034 -0.16 6.491 21.27 0.07 Sus philippensis Mod_null NA NA NA NA 17.27 0.53 Micrastur plumbeus Mod_use 0.83 12.115 NA NA 49.82 0.10 Micrastur plumbeus Mod_clim NA NA 4.08 3.399 46.71 0.46 Micrastur plumbeus Mod_clim_use 3.80 12.668 4.15 3.429 48.64 0.18 Micrastur plumbeus Mod_null NA NA NA NA 47.83 0.26 Touit surdus Mod_use 24.08 3.734 NA NA 312.49 0.00 Touit surdus Mod_clim NA NA 24.45 2.870 250.25 0.05 Touit surdus Mod_clim_use 8.83 3.679 24.97 3.643 244.36 0.95 Touit surdus Mod_null NA NA NA NA 327.55 0.00 Taoniscus nanus Mod_use -2.34 7.354 NA NA 58.27 0.20 Taoniscus nanus Mod_clim NA NA -0.47 3.872 58.38 0.19 Taoniscus nanus Mod_clim_use -3.35 8.068 -1.25 3.875 60.17 0.08 Taoniscus nanus Mod_null NA NA NA NA 56.39 0.52 Pseudibis davisoni Mod_use 65.65 19.506 NA NA 139.86 0.00 Pseudibis davisoni Mod_clim NA NA 4.86 4.182 141.73 0.00 Pseudibis davisoni Mod_clim_use 49.45 0.001 59.02 0.001 119.01 1.00 Pseudibis davisoni Mod_null NA NA NA NA 141.67 0.00 220

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Atrichornis rufescens Mod_use 3.50 2.267 NA NA 373.79 0.25 Atrichornis rufescens Mod_clim NA NA 1.45 0.875 373.37 0.31 Atrichornis rufescens Mod_clim_use 2.81 2.356 1.20 0.894 373.86 0.25 Atrichornis rufescens Mod_null NA NA NA NA 374.43 0.19 Claravis godefrida Mod_use 5.09 8.610 NA NA 57.81 0.18 Claravis godefrida Mod_clim NA NA 4.20 4.996 57.02 0.27 Claravis godefrida Mod_clim_use 5.35 9.293 4.36 5.351 58.62 0.12 Claravis godefrida Mod_null NA NA NA NA 56.16 0.42 Pauxi unicornis Mod_use 0.72 18.375 NA NA 198.95 0.19 Pauxi unicornis Mod_clim NA NA -0.55 1.411 198.80 0.21 Pauxi unicornis Mod_clim_use 0.75 20.664 -0.55 1.410 200.80 0.08 Pauxi unicornis Mod_null NA NA NA NA 196.95 0.52 Strabomantis bufoniformis Mod_use 6.10 7.481 NA NA 36.79 0.00 Strabomantis bufoniformis Mod_clim NA NA -41.59 17.985 25.60 0.71 Strabomantis bufoniformis Mod_clim_use 5.88 11.357 -38.38 18.995 27.44 0.28 Strabomantis bufoniformis Mod_null NA NA NA NA 35.91 0.00 Charitospiza eucosma Mod_use -1.21 7.334 NA NA 61.28 0.19 Charitospiza eucosma Mod_clim NA NA -8.38 43.953 60.90 0.23 Charitospiza eucosma Mod_clim_use -2.66 8.085 -9.38 44.724 62.74 0.09 Charitospiza eucosma Mod_null NA NA NA NA 59.31 0.50 Crax globulosa Mod_use 3.45 6.418 NA NA 322.38 0.23 Crax globulosa Mod_clim NA NA 0.18 1.695 322.80 0.19 Crax globulosa Mod_clim_use 3.60 6.468 0.35 1.662 324.34 0.09 Crax globulosa Mod_null NA NA NA NA 320.81 0.50 221

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Pyrrhura cruentata Mod_use -2.52 3.864 NA NA 141.46 0.18 Pyrrhura cruentata Mod_clim NA NA -2.37 2.291 140.48 0.29 Pyrrhura cruentata Mod_clim_use -2.28 4.035 -2.32 2.333 142.16 0.13 Pyrrhura cruentata Mod_null NA NA NA NA 139.87 0.40 Myrmotherula unicolor Mod_use -1.71 2.772 NA NA 235.61 0.18 Myrmotherula unicolor Mod_clim NA NA -1.53 1.391 234.66 0.29 Myrmotherula unicolor Mod_clim_use -1.28 2.848 -1.45 1.418 236.43 0.12 Myrmotherula unicolor Mod_null NA NA NA NA 234.03 0.40 Cinclodes aricomae Mod_use 151.94 27.215 NA NA 67.74 0.31 Cinclodes aricomae Mod_clim NA NA 0.78 3.671 69.04 0.16 Cinclodes aricomae Mod_clim_use 151.79 135.698 0.05 21.611 69.74 0.11 Cinclodes aricomae Mod_null NA NA NA NA 67.09 0.42 Ommatotriton vittatus Mod_use 0.21 17.545 NA NA 9.38 0.27 Ommatotriton vittatus Mod_clim NA NA NA NA NA NA Ommatotriton vittatus Mod_clim_use NA NA NA NA NA NA Ommatotriton vittatus Mod_null NA NA NA NA 7.38 0.73 Poospiza cinerea Mod_use 0.45 3.512 NA NA 177.31 0.11 Poospiza cinerea Mod_clim NA NA 5.15 6.523 175.04 0.34 Poospiza cinerea Mod_clim_use 5.69 5.246 6.99 7.071 175.50 0.27 Poospiza cinerea Mod_null NA NA NA NA 175.32 0.29 Pristimantis caryophyllaceus Mod_use 4.66 2.316 NA NA 127.12 0.25 Pristimantis caryophyllaceus Mod_clim NA NA 3.75 2.147 126.36 0.36 Pristimantis caryophyllaceus Mod_clim_use 3.20 2.402 2.78 2.102 126.54 0.33 Pristimantis caryophyllaceus Mod_null NA NA NA NA 130.21 0.05 222

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Haeromys pusillus Mod_use 5.54 3.651 NA NA 137.67 0.30 Haeromys pusillus Mod_clim NA NA 1.88 1.801 138.39 0.21 Haeromys pusillus Mod_clim_use 5.18 4.017 1.65 1.893 138.89 0.16 Haeromys pusillus Mod_null NA NA NA NA 137.55 0.32 Nomascus leucogenys Mod_use 6.19 4.243 NA NA 273.76 0.15 Nomascus leucogenys Mod_clim NA NA -1.35 1.493 275.28 0.07 Nomascus leucogenys Mod_clim_use 77.68 16.135 -10.61 3.864 270.72 0.67 Nomascus leucogenys Mod_null NA NA NA NA 274.21 0.12 Triclaria malachitacea Mod_use 2.67 2.404 NA NA 209.16 0.00 Triclaria malachitacea Mod_clim NA NA 42.32 7.373 90.10 0.50 Triclaria malachitacea Mod_clim_use 10.93 7.509 37.33 7.139 90.13 0.50 Triclaria malachitacea Mod_null NA NA NA NA 208.40 0.00 Piprites pileata Mod_use -3.50 11.182 NA NA 31.93 0.21 Piprites pileata Mod_clim NA NA -0.05 5.002 32.07 0.19 Piprites pileata Mod_clim_use -4.42 12.347 1.02 5.255 33.89 0.08 Piprites pileata Mod_null NA NA NA NA 30.07 0.52 Mandrillus leucophaeus Mod_use NA NA NA NA NA NA Mandrillus leucophaeus Mod_clim NA NA 2.61 8.305 19.08 0.28 Mandrillus leucophaeus Mod_clim_use NA NA NA NA NA NA Mandrillus leucophaeus Mod_null NA NA NA NA 17.21 0.72 Sporophila falcirostris Mod_use 3.55 6.601 NA NA 35.28 0.20 Sporophila falcirostris Mod_clim NA NA -4.91 12.260 35.13 0.22 Sporophila falcirostris Mod_clim_use 4.76 8.582 -5.14 11.641 36.79 0.10 Sporophila falcirostris Mod_null NA NA NA NA 33.54 0.48 223

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Crax alberti Mod_use 4.50 3.057 NA NA 74.19 0.43 Crax alberti Mod_clim NA NA -0.24 1.518 76.91 0.11 Crax alberti Mod_clim_use 4.82 3.286 0.56 1.691 76.08 0.17 Crax alberti Mod_null NA NA NA NA 74.94 0.29 Asthenes heterura Mod_use 2.74 3.268 NA NA 144.46 0.18 Asthenes heterura Mod_clim NA NA 1.33 1.029 143.36 0.32 Asthenes heterura Mod_clim_use 1.55 3.712 1.20 1.078 145.19 0.13 Asthenes heterura Mod_null NA NA NA NA 143.08 0.37 Electron carinatum Mod_use 3.32 3.654 NA NA 85.45 0.25 Electron carinatum Mod_clim NA NA 0.95 2.574 86.10 0.18 Electron carinatum Mod_clim_use 3.21 3.705 0.63 2.534 87.39 0.10 Electron carinatum Mod_null NA NA NA NA 84.23 0.47 Xipholena atropurpurea Mod_use 1.01 5.825 NA NA 47.86 0.20 Xipholena atropurpurea Mod_clim NA NA -0.09 4.127 47.89 0.20 Xipholena atropurpurea Mod_clim_use 1.30 6.155 -0.53 4.775 49.85 0.07 Xipholena atropurpurea Mod_null NA NA NA NA 45.89 0.53 Odontophorus hyperythrus Mod_use 131.47 150.472 NA NA 65.01 0.19 Odontophorus hyperythrus Mod_clim NA NA 11.55 10.707 64.15 0.30 Odontophorus hyperythrus Mod_clim_use 128.69 170.007 12.55 11.299 64.40 0.26 Odontophorus hyperythrus Mod_null NA NA NA NA 64.54 0.24 Leptotila ochraceiventris Mod_use 12.47 7.857 NA NA 120.19 0.30 Leptotila ochraceiventris Mod_clim NA NA 7.48 5.841 121.85 0.13 Leptotila ochraceiventris Mod_clim_use 13.63 10.101 7.73 6.406 119.13 0.51 Leptotila ochraceiventris Mod_null NA NA NA NA 123.73 0.05 224

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Carduelis yarrellii Mod_use NA NA NA NA NA NA Carduelis yarrellii Mod_clim NA NA -1.74 6.469 22.34 0.24 Carduelis yarrellii Mod_clim_use 2271.92 6627238.395 -0.56 61.086 23.97 0.11 Carduelis yarrellii Mod_null NA NA NA NA 20.40 0.65 Craugastor ranoides Mod_use 4.41 2.706 NA NA 238.46 0.46 Craugastor ranoides Mod_clim NA NA 0.27 0.965 241.49 0.10 Craugastor ranoides Mod_clim_use 4.41 2.726 0.22 1.010 240.42 0.17 Craugastor ranoides Mod_null NA NA NA NA 239.57 0.27 Nothura minor Mod_use 1.87 6.756 NA NA 54.19 0.20 Nothura minor Mod_clim NA NA -0.27 3.726 54.26 0.19 Nothura minor Mod_clim_use 1.93 0.002 -0.22 0.002 56.19 0.07 Nothura minor Mod_null NA NA NA NA 52.26 0.53 Apteryx haastii Mod_use 184.42 431.774 NA NA 179.80 0.46 Apteryx haastii Mod_clim NA NA 0.28 0.971 182.81 0.10 Apteryx haastii Mod_clim_use 183.80 90.757 0.05 1.004 181.80 0.17 Apteryx haastii Mod_null NA NA NA NA 180.89 0.27 Otus brookii Mod_use NA NA NA NA NA NA Otus brookii Mod_clim NA NA 0.83 1.732 135.17 0.24 Otus brookii Mod_clim_use -79.04 180.636 1.24 1.741 135.79 0.18 Otus brookii Mod_null NA NA NA NA 133.39 0.58 Pauxi pauxi Mod_use -8.12 3.169 NA NA 83.46 0.00 Pauxi pauxi Mod_clim NA NA -17.66 6.651 56.30 0.48 Pauxi pauxi Mod_clim_use 9.47 7.632 -22.36 8.515 56.13 0.52 Pauxi pauxi Mod_null NA NA NA NA 90.39 0.00 225

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Touit stictopterus Mod_use 6.68 0.003 NA NA 50.39 0.23 Touit stictopterus Mod_clim NA NA 1.08 2.995 50.80 0.19 Touit stictopterus Mod_clim_use 6.43 7.835 0.85 3.109 52.32 0.09 Touit stictopterus Mod_null NA NA NA NA 48.92 0.49 Myiarchus semirufus Mod_use 0.58 6.692 NA NA 87.08 0.02 Myiarchus semirufus Mod_clim NA NA 19.58 8.188 79.55 0.69 Myiarchus semirufus Mod_clim_use -1.14 6.921 19.69 8.579 81.52 0.26 Myiarchus semirufus Mod_null NA NA NA NA 85.10 0.04 Onychorhynchus occidentalis Mod_use 3.19 2.961 NA NA 172.64 0.22 Onychorhynchus occidentalis Mod_clim NA NA 1.64 1.477 172.40 0.25 Onychorhynchus occidentalis Mod_clim_use 3.99 3.133 1.90 1.460 172.67 0.21 Onychorhynchus occidentalis Mod_null NA NA NA NA 171.84 0.32 Bucco noanamae Mod_use 6.08 4.726 NA NA 116.10 0.33 Bucco noanamae Mod_clim NA NA -0.14 1.377 117.71 0.15 Bucco noanamae Mod_clim_use 6.28 4.834 -0.39 1.455 118.03 0.13 Bucco noanamae Mod_null NA NA NA NA 115.72 0.40 Xenospingus concolor Mod_use 4.48 6.367 NA NA 37.14 0.21 Xenospingus concolor Mod_clim NA NA -3.29 9.851 37.15 0.21 Xenospingus concolor Mod_clim_use 6.34 8.181 -4.35 11.970 38.71 0.09 Xenospingus concolor Mod_null NA NA NA NA 35.44 0.49 Dasyornis brachypterus Mod_use -2.21 3.771 NA NA 82.38 0.20 Dasyornis brachypterus Mod_clim NA NA -1.94 2.680 82.07 0.24 Dasyornis brachypterus Mod_clim_use -1.62 4.202 -1.73 2.787 83.94 0.09 Dasyornis brachypterus Mod_null NA NA NA NA 80.70 0.47 226

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Dacnis nigripes Mod_use 6.83 6.004 NA NA 47.73 0.32 Dacnis nigripes Mod_clim NA NA 2.25 4.017 48.94 0.17 Dacnis nigripes Mod_clim_use 6.66 6.367 1.20 3.762 49.63 0.12 Dacnis nigripes Mod_null NA NA NA NA 47.37 0.38 Rana sierrae Mod_use -11.00 16.098 NA NA 74.70 0.00 Rana sierrae Mod_clim NA NA 222.45 6443770.039 59.23 0.73 Rana sierrae Mod_clim_use NA NA NA NA NA NA Rana sierrae Mod_null NA NA NA NA 73.05 0.00 Mohoua ochrocephala Mod_use 261.55 50.272 NA NA 212.86 0.47 Mohoua ochrocephala Mod_clim NA NA 14.00 2.394 229.97 0.00 Mohoua ochrocephala Mod_clim_use 233.14 59.418 2.27 1.552 212.65 0.53 Mohoua ochrocephala Mod_null NA NA NA NA 231.19 0.00 Biatas nigropectus Mod_use 66.32 15.456 NA NA 125.50 0.15 Biatas nigropectus Mod_clim NA NA 0.26 2.236 125.19 0.18 Biatas nigropectus Mod_clim_use 6.57 5.035 -0.64 2.388 124.93 0.20 Biatas nigropectus Mod_null NA NA NA NA 123.21 0.47 Saltator rufiventris Mod_use -0.12 10.929 NA NA 94.00 0.19 Saltator rufiventris Mod_clim NA NA -0.63 2.041 93.90 0.20 Saltator rufiventris Mod_clim_use 0.83 10.514 -0.67 2.115 95.89 0.08 Saltator rufiventris Mod_null NA NA NA NA 92.00 0.53 Lynx pardinus Mod_use 52.63 12.194 NA NA 33.38 0.31 Lynx pardinus Mod_clim NA NA -553.52 75.528 34.28 0.20 Lynx pardinus Mod_clim_use 50.96 13.635 -2023.00 13.467 35.09 0.13 Lynx pardinus Mod_null NA NA NA NA 33.13 0.35 227

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Penelopides panini Mo d_use -5.93 9.323 NA NA 27.99 0.10 Penelopides panini Mod_clim NA NA -15.55 7.371 24.66 0.52 Penelopides panini Mod_clim_use -1.76 9.506 -15.08 7.578 26.62 0.19 Penelopides panini Mod_null NA NA NA NA 26.72 0.19 Amazona rhodocorytha Mod_use 1.71 0.003 NA NA 53.69 0.20 Amazona rhodocorytha Mod_clim NA NA 0.59 3.307 53.72 0.20 Amazona rhodocorytha Mod_clim_use 1.77 7.143 0.62 3.315 55.66 0.07 Amazona rhodocorytha Mod_null NA NA NA NA 51.75 0.53 Paradisaea rudolphi Mod_use -4.68 15.341 NA NA 84.86 0.21 Paradisaea rudolphi Mod_clim NA NA -0.08 2.533 85.02 0.19 Paradisaea rudolphi Mod_clim_use -4.73 15.359 -0.15 2.626 86.86 0.08 Paradisaea rudolphi Mod_null NA NA NA NA 83.02 0.52 Amazona pretrei Mod_use -1.53 2.945 NA NA 166.28 0.00 Amazona pretrei Mod_clim NA NA 307.91 27.170 62.32 0.73 Amazona pretrei Mod_clim_use 0.27 4.601 307.02 30.131 64.32 0.27 Amazona pretrei Mod_null NA NA NA NA 164.55 0.00 Dicaeum haematostictum Mod_use -81.21 37.408 NA NA 19.86 0.13 Dicaeum haematostictum Mod_clim NA NA -15.32 8.904 16.94 0.55 Dicaeum haematostictum Mod_clim_use -0.63 12.209 -15.18 8.846 18.93 0.20 Dicaeum haematostictum Mod_null NA NA NA NA 19.94 0.12 Carduelis siemiradzkii Mod_use -0.47 3.092 NA NA 107.85 0.15 Carduelis siemiradzkii Mod_clim NA NA 2.51 2.060 106.18 0.33 Carduelis siemiradzkii Mod_clim_use -1.10 3.113 2.65 2.127 108.05 0.13 Carduelis siemiradzkii Mod_null NA NA NA NA 105.88 0.39 228

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Cinclus schulzi Mod_use 8.59 6.810 NA NA 61.34 0.32 Cinclus schulzi Mod_clim NA NA 1.25 3.104 62.70 0.16 Cinclus schulzi Mod_clim_use 8.33 6.913 1.05 3.484 63.25 0.12 Cinclus schulzi Mod_null NA NA NA NA 60.84 0.40 Turdoides hindei Mod_use -5.80 2.105 NA NA 209.46 0.19 Turdoides hindei Mod_clim NA NA -2.12 0.845 213.48 0.03 Turdoides hindei Mod_clim_use -5.15 2.034 -1.97 0.918 206.66 0.78 Turdoides hindei Mod_null NA NA NA NA 218.45 0.00 Coracina ostenta Mod_use -81.21 30.038 NA NA 19.86 0.13 Coracina ostenta Mod_clim NA NA -15.31 8.530 16.93 0.55 Coracina ostenta Mod_clim_use -0.80 12.196 -15.13 9.539 18.93 0.20 Coracina ostenta Mod_null NA NA NA NA 19.93 0.12 Rhinomyias albigularis Mod_use -81.21 30.038 NA NA 19.86 0.13 Rhinomyias albigularis Mod_clim NA NA -15.31 8.530 16.93 0.55 Rhinomyias albigularis Mod_clim_use -0.80 12.196 -15.13 9.539 18.93 0.20 Rhinomyias albigularis Mod_null NA NA NA NA 19.93 0.12 Aceros waldeni Mod_use 4.04 4.218 NA NA 99.27 0.29 Aceros waldeni Mod_clim NA NA -0.18 1.992 100.14 0.19 Aceros waldeni Mod_clim_use 41.55 14.528 -6.88 5.738 105.52 0.01 Aceros waldeni Mod_null NA NA NA NA 98.15 0.51 Ortalis erythroptera Mod_use 8.44 7.131 NA NA 58.17 0.32 Ortalis erythroptera Mod_clim NA NA 2.07 3.033 59.34 0.18 Ortalis erythroptera Mod_clim_use 8.01 7.573 1.23 3.131 60.02 0.13 Ortalis erythroptera Mod_null NA NA NA NA 57.80 0.38 229

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Galbula pastazae Mod_use -8.97 47.970 NA NA 23.28 0.18 Galbula pastazae Mod_clim NA NA 2.64 3.495 22.54 0.26 Galbula pastazae Mod_clim_use 4.35 21.053 2.83 3.705 24.51 0.10 Galbula pastazae Mod_null NA NA NA NA 21.32 0.47 Atelopus varius Mod_use -2.39 3.205 NA NA 100.83 0.24 Atelopus varius Mod_clim NA NA -0.27 1.491 101.37 0.18 Atelopus varius Mod_clim_use -2.37 3.252 -0.07 1.520 102.82 0.09 Atelopus varius Mod_null NA NA NA NA 99.40 0.49 Nothoprocta taczanowskii Mod_use -52.72 56.514 NA NA 95.94 0.32 Nothoprocta taczanowskii Mod_clim NA NA -0.94 2.953 97.36 0.16 Nothoprocta taczanowskii Mod_clim_use -51.35 54.748 -0.67 2.924 97.88 0.12 Nothoprocta taczanowskii Mod_null NA NA NA NA 95.48 0.40 Gyalophylax hellmayri Mod_use 0.08 7.889 NA NA 93.80 0.19 Gyalophylax hellmayri Mod_clim NA NA 1.71 3.319 93.54 0.22 Gyalophylax hellmayri Mod_clim_use 1.51 8.561 1.89 3.498 95.51 0.08 Gyalophylax hellmayri Mod_null NA NA NA NA 91.80 0.51 Simoxenops striatus Mod_use -758.55 0.011 NA NA 43.19 0.24 Simoxenops striatus Mod_clim NA NA 2.39 2.465 46.07 0.06 Simoxenops striatus Mod_clim_use -1102.81 410.204 11.95 5.879 41.25 0.62 Simoxenops striatus Mod_null NA NA NA NA 45.23 0.09 Myrmoborus melanurus Mod_use -224.31 86.895 NA NA 44.47 0.95 Myrmoborus melanurus Mod_clim NA NA 6.38 6.372 51.60 0.03 Myrmoborus melanurus Mod_clim_use NA NA NA NA NA NA Myrmoborus melanurus Mod_null NA NA NA NA 51.57 0.03 230

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Clytorhynchus vitiensis Mod_use -1.82 0.003 NA NA 12.49 0.27 Clytorhynchus vitiensis Mod_clim NA NA NA NA NA NA Clytorhynchus vitiensis Mod_clim_use NA NA NA NA NA NA Clytorhynchus vitiensis Mod_null NA NA NA NA 10.50 0.73 Agapornis nigrigenis Mod_use 345.66 126.952 NA NA 58.78 0.55 Agapornis nigrigenis Mod_clim NA NA -0.18 3.981 63.03 0.07 Agapornis nigrigenis Mod_clim_use 345.74 141.572 0.00 5.010 60.78 0.20 Agapornis nigrigenis Mod_null NA NA NA NA 61.03 0.18 Artamus mentalis Mod_use -2.95 28.704 NA NA 12.52 0.27 Artamus mentalis Mod_clim NA NA NA NA NA NA Artamus mentalis Mod_clim_use NA NA NA NA NA NA Artamus mentalis Mod_null NA NA NA NA 10.54 0.73 Phytotoma raimondii Mod_use 1.38 12.832 NA NA 47.17 0.20 Phytotoma raimondii Mod_clim NA NA 0.73 4.387 47.15 0.20 Phytotoma raimondii Mod_clim_use 1.05 12.937 0.68 4.419 49.14 0.07 Phytotoma raimondii Mod_null NA NA NA NA 45.18 0.53 Chlorochrysa nitidissima Mod_use -6.27 4.830 NA NA 48.45 0.00 Chlorochrysa nitidissima Mod_clim NA NA -9.25 0.019 34.23 1.00 Chlorochrysa nitidissima Mod_clim_use -6.27 4.845 -0.11 1.150 50.44 0.00 Chlorochrysa nitidissima Mod_null NA NA NA NA 49.83 0.00 Coccyzus rufigularis Mod_use 5.09 12.196 NA NA 21.86 0.21 Coccyzus rufigularis Mod_clim NA NA 1.43 6.854 22.05 0.19 Coccyzus rufigularis Mod_clim_use 5.51 13.516 1.64 7.459 23.80 0.08 Coccyzus rufigularis Mod_null NA NA NA NA 20.10 0.51 231

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Emballonura semicaudata Mod_use 2.07 14.086 NA NA 18.89 0.20 Emballonura semicaudata Mod_clim NA NA -1.14 6.389 18.89 0.20 Emballonura semicaudata Mod_clim_use 2.35 14.444 -1.34 6.416 20.85 0.07 Emballonura semicaudata Mod_null NA NA NA NA 16.92 0.53 Komodomys rintjanus Mod_use -2.16 7.985 NA NA 40.78 0.20 Komodomys rintjanus Mod_clim NA NA 0.77 4.375 40.82 0.20 Komodomys rintjanus Mod_clim_use -2.02 7.837 0.71 4.399 42.76 0.07 Komodomys rintjanus Mod_null NA NA NA NA 38.85 0.53 Hyloscirtus colymba Mod_use -58.26 488.710 NA NA 31.64 0.22 Hyloscirtus colymba Mod_clim NA NA -0.63 4.539 31.69 0.21 Hyloscirtus colymba Mod_clim_use NA NA NA NA NA NA Hyloscirtus colymba Mod_null NA NA NA NA 29.71 0.57 Dysithamnus plumbeus Mod_use -2.93 3.196 NA NA 131.89 0.25 Dysithamnus plumbeus Mod_clim NA NA -1.18 1.554 132.20 0.21 Dysithamnus plumbeus Mod_clim_use -2.63 3.374 -0.93 1.594 133.53 0.11 Dysithamnus plumbeus Mod_null NA NA NA NA 130.83 0.43 Rana muscosa Mod_use 1.49 10.876 NA NA 44.62 0.19 Rana muscosa Mod_clim NA NA NA NA NA NA Rana muscosa Mod_clim_use NA NA NA NA NA NA Rana muscosa Mod_null NA NA NA NA 42.64 0.50 Oreothraupis arremonops Mod_use 8.01 3.621 NA NA 82.84 0.53 Oreothraupis arremonops Mod_clim NA NA -1.57 2.225 88.43 0.03 Oreothraupis arremonops Mod_clim_use 9.39 4.489 -2.49 2.425 83.53 0.37 Oreothraupis arremonops Mod_null NA NA NA NA 87.03 0.07 232

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Herpsilochmus pectoralis Mod_use 4.80 5.881 NA NA 95.24 0.00 Herpsilochmus pectoralis Mod_clim NA NA 68.43 17.541 48.35 0.00 Herpsilochmus pectoralis Mod_clim_use 349.35 0.003 374.93 0.003 35.31 1.00 Herpsilochmus pectoralis Mod_null NA NA NA NA 93.88 0.00 Anairetes alpinus Mod_use -9.45 112.698 NA NA 78.01 0.17 Anairetes alpinus Mod_clim NA NA 2.21 0.002 77.15 0.26 Anairetes alpinus Mod_clim_use -42.10 121.879 2.47 2.641 79.01 0.10 Anairetes alpinus Mod_null NA NA NA NA 76.02 0.46 Calyptophilus frugivorus Mod_use -20.52 5.896 NA NA 104.60 0.01 Calyptophilus frugivorus Mod_clim NA NA 0.43 2.111 98.43 0.18 Calyptophilus frugivorus Mod_clim_use -52.52 0.002 22.66 0.002 97.07 0.35 Calyptophilus frugivorus Mod_null NA NA NA NA 96.47 0.47 Duellmanohyla uranochroa Mod_use 25.90 25.123 NA NA 35.73 0.36 Duellmanohyla uranochroa Mod_clim NA NA NA NA NA NA Duellmanohyla uranochroa Mod_clim_use NA NA NA NA NA NA Duellmanohyla uranochroa Mod_null NA NA NA NA 36.05 0.31 Myrmotherula urosticta Mod_use 5.06 6.027 NA NA 89.50 0.27 Myrmotherula urosticta Mod_clim NA NA -0.59 2.999 90.35 0.17 Myrmotherula urosticta Mod_clim_use 4.97 6.038 -0.20 3.132 91.50 0.10 Myrmotherula urosticta Mod_null NA NA NA NA 88.39 0.46 Pteralopex flanneryi Mod_use NA NA NA NA NA NA Pteralopex flanneryi Mod_clim NA NA -3.77 11.922 43.45 0.22 Pteralopex flanneryi Mod_clim_use NA NA NA NA NA NA Pteralopex flanneryi Mod_null NA NA NA NA 41.76 0.51 233

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Babyrousa babyrussa Mod_use -2.21 22.596 NA NA 15.63 0.20 Babyrousa babyrussa Mod_clim NA NA 1.64 6.597 15.58 0.20 Babyrousa babyrussa Mod_clim_use -3.48 22.255 1.98 6.617 17.54 0.08 Babyrousa babyrussa Mod_null NA NA NA NA 13.64 0.53 Brotogeris pyrrhoptera Mod_use 0.20 3.348 NA NA 85.00 0.18 Brotogeris pyrrhoptera Mod_clim NA NA 1.10 1.587 84.54 0.23 Brotogeris pyrrhoptera Mod_clim_use 0.07 3.350 1.10 1.593 86.54 0.09 Brotogeris pyrrhoptera Mod_null NA NA NA NA 83.00 0.50 Tachycineta euchrysea Mod_use 1.85 5.116 NA NA 58.67 0.17 Tachycineta euchrysea Mod_clim NA NA 2.96 2.788 57.64 0.28 Tachycineta euchrysea Mod_clim_use 2.14 0.002 3.03 0.002 59.50 0.11 Tachycineta euchrysea Mod_null NA NA NA NA 56.80 0.43 Cotinga ridgwayi Mod_use 3.94 3.545 NA NA 67.13 0.31 Cotinga ridgwayi Mod_clim NA NA 0.53 2.217 68.44 0.16 Cotinga ridgwayi Mod_clim_use 3.98 3.624 -0.10 2.350 69.13 0.11 Cotinga ridgwayi Mod_null NA NA NA NA 66.49 0.42 Myrmotherula minor Mod_use -2.07 4.265 NA NA 84.27 0.16 Myrmotherula minor Mod_clim NA NA -2.98 3.012 82.85 0.33 Myrmotherula minor Mod_clim_use -0.13 4.773 -2.96 3.165 84.85 0.12 Myrmotherula minor Mod_null NA NA NA NA 82.52 0.39 Hemitriccus furcatus Mod_use 3.83 6.111 NA NA 52.54 0.22 Hemitriccus furcatus Mod_clim NA NA -0.29 3.220 52.91 0.19 Hemitriccus furcatus Mod_clim_use 5.29 7.181 -1.56 3.848 54.36 0.09 Hemitriccus furcatus Mod_null NA NA NA NA 50.92 0.50 234

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Dysithamnus occidentalis Mod _use -11.01 0.006 NA NA 64.68 0.00 Dysithamnus occidentalis Mod_clim NA NA -14.56 4.756 53.92 0.72 Dysithamnus occidentalis Mod_clim_use 3.18 19.586 -14.72 4.908 55.89 0.27 Dysithamnus occidentalis Mod_null NA NA NA NA 62.91 0.01 Acanthodactylus schreiberi Mod_use 111.29 0.002 NA NA 29.59 0.27 Acanthodactylus schreiberi Mod_clim NA NA 5.17 10.884 30.89 0.14 Acanthodactylus schreiberi Mod_clim_use 113.89 58.237 20.68 16.377 29.44 0.29 Acanthodactylus schreiberi Mod_null NA NA NA NA 29.38 0.30 Anthocephala floriceps Mod_use -6.87 9.625 NA NA 19.89 0.24 Anthocephala floriceps Mod_clim NA NA 2.62 6.043 20.29 0.19 Anthocephala floriceps Mod_clim_use -7.18 9.403 3.57 6.695 21.58 0.10 Anthocephala floriceps Mod_null NA NA NA NA 18.50 0.47 Chasiempis sandwichensis Mod_use -4.09 3.368 NA NA 93.96 0.30 Chasiempis sandwichensis Mod_clim NA NA 0.82 0.002 94.90 0.19 Chasiempis sandwichensis Mod_clim_use -4.30 3.488 0.85 1.023 95.21 0.16 Chasiempis sandwichensis Mod_null NA NA NA NA 93.60 0.36 Eleothreptus candicans Mod_use 55.01 20.477 NA NA 43.11 0.00 Eleothreptus candicans Mod_clim NA NA 3.09 7.318 46.27 0.00 Eleothreptus candicans Mod_clim_use 109.89 35.935 103.04 1440.753 15.11 1.00 Eleothreptus candicans Mod_null NA NA NA NA 44.63 0.00 Neopelma aurifrons Mod_use 30.69 12.792 NA NA 52.40 0.22 Neopelma aurifrons Mod_clim NA NA -0.30 3.132 55.49 0.05 Neopelma aurifrons Mod_clim_use 64.80 27.714 -21.57 14.002 50.31 0.61 Neopelma aurifrons Mod_null NA NA NA NA 53.50 0.12 235

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Ammospermophilus nelsoni Mod_ use 2.08 0.002 NA NA 94.28 0.14 Ammospermophilus nelsoni Mod_clim NA NA 4.08 4.791 91.65 0.51 Ammospermophilus nelsoni Mod_clim_use 0.99 0.001 3.16 0.001 93.48 0.20 Ammospermophilus nelsoni Mod_null NA NA NA NA 94.08 0.15 Amazona ventralis Mod_use 7.27 14.905 NA NA 24.53 0.24 Amazona ventralis Mod_clim NA NA 141.80 281.321 24.59 0.24 Amazona ventralis Mod_clim_use NA NA NA NA NA NA Amazona ventralis Mod_null NA NA NA NA 23.02 0.52 Hypopyrrhus pyrohypogaster Mod_use -0.23 2.000 NA NA 110.66 0.20 Hypopyrrhus pyrohypogaster Mod_clim NA NA -0.01 0.946 110.67 0.20 Hypopyrrhus pyrohypogaster Mod_clim_use -0.23 2.012 0.00 0.944 112.66 0.07 Hypopyrrhus pyrohypogaster Mod_null NA NA NA NA 108.67 0.53 Amazona agilis Mod_use 4.90 2.459 NA NA 80.26 0.43 Amazona agilis Mod_clim NA NA 1.82 1.366 83.41 0.09 Amazona agilis Mod_clim_use 5.01 2.638 1.81 1.549 80.30 0.42 Amazona agilis Mod_null NA NA NA NA 83.91 0.07 Bangsia melanochlamys Mod_use 32.10 12.155 NA NA 93.58 0.69 Bangsia melanochlamys Mod_clim NA NA -1.06 0.844 119.12 0.00 Bangsia melanochlamys Mod_clim_use 31.10 12.024 -0.40 0.631 95.17 0.31 Bangsia melanochlamys Mod_null NA NA NA NA 118.86 0.00 Grallaria gigantea Mod_use 112.09 35.371 NA NA 42.13 0.66 Grallaria gigantea Mod_clim NA NA -0.25 3.007 48.45 0.03 Grallaria gigantea Mod_clim_use 112.16 42.932 -0.21 4.344 44.13 0.24 Grallaria gigantea Mod_null NA NA NA NA 46.46 0.08 236

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Macroagelaius subalaris Mod_use 2.44 18.956 NA NA 18.76 0.20 Macroagelaius subalaris Mod_clim NA NA 0.72 0.005 18.76 0.20 Macroagelaius subalaris Mod_clim_use 1.99 19.542 0.40 7.279 20.75 0.07 Macroagelaius subalaris Mod_null NA NA NA NA 16.78 0.53 Patagioenas oenops Mod_use -6.66 13.529 NA NA 50.93 0.22 Patagioenas oenops Mod_clim NA NA -2.35 5.702 51.06 0.21 Patagioenas oenops Mod_clim_use -6.59 14.461 -2.13 5.723 52.70 0.09 Patagioenas oenops Mod_null NA NA NA NA 49.35 0.48 Psammodromus microdactylus Mod_use -25.47 9.571 NA NA 44.49 0.26 Psammodromus microdactylus Mod_clim NA NA 0.23 2.873 52.45 0.00 Psammodromus microdactylus Mod_clim_use -40.96 15.041 15.47 8.669 42.47 0.72 Psammodromus microdactylus Mod_null NA NA NA NA 50.46 0.01 Coeligena prunellei Mod_use 112.64 33.128 NA NA 32.18 0.15 Coeligena prunellei Mod_clim NA NA 0.08 3.285 47.70 0.00 Coeligena prunellei Mod_clim_use 157.17 0.002 -27.66 0.002 28.73 0.85 Coeligena prunellei Mod_null NA NA NA NA 45.70 0.00 Grallaria rufocinerea Mod_use 1.39 10.579 NA NA 43.39 0.20 Grallaria rufocinerea Mod_clim NA NA 0.18 3.515 43.40 0.20 Grallaria rufocinerea Mod_clim_use 1.32 10.831 0.10 3.600 45.39 0.07 Grallaria rufocinerea Mod_null NA NA NA NA 41.40 0.53 Touit melanonotus Mod_use 4.62 15.312 NA NA 18.73 0.25 Touit melanonotus Mod_clim NA NA NA NA NA NA Touit melanonotus Mod_clim_use 5.69 89.783 221.50 10366819.760 20.38 0.11 Touit melanonotus Mod_null NA NA NA NA 16.82 0.64 237

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Trachypithecus geei Mod_use 9.41 4.276 NA NA 50.27 0.00 Trachypithecus geei Mod_clim NA NA -16.52 6.070 37.81 0.60 Trachypithecus geei Mod_clim_use 20.81 7.344 -17.20 6.930 38.66 0.40 Trachypithecus geei Mod_null NA NA NA NA 56.93 0.00 Synallaxis tithys Mod_use 50.58 24.245 NA NA 43.25 0.44 Synallaxis tithys Mod_clim NA NA 2.59 3.180 50.04 0.01 Synallaxis tithys Mod_clim_use 28.88 27.527 7.32 5.210 42.97 0.51 Synallaxis tithys Mod_null NA NA NA NA 48.66 0.03 Collocalia orientalis Mod_use 147.36 95.872 NA NA 22.11 0.02 Collocalia orientalis Mod_clim NA NA 461.01 11184810.667 24.32 0.01 Collocalia orientalis Mod_clim_use NA NA NA NA NA NA Collocalia orientalis Mod_null NA NA NA NA 30.59 0.00 Canis simensis Mod_use 8.91 14.412 NA NA 24.29 0.14 Canis simensis Mod_clim NA NA -1.58 7.789 24.90 0.10 Canis simensis Mod_clim_use 169.95 55.224 -61.66 24.766 21.78 0.49 Canis simensis Mod_null NA NA NA NA 22.96 0.27 Myadestes obscurus Mod_use -1.37 5.868 NA NA 65.55 0.02 Myadestes obscurus Mod_clim NA NA 5.37 3.134 58.59 0.66 Myadestes obscurus Mod_clim_use 2.95 7.098 5.34 3.086 60.42 0.26 Myadestes obscurus Mod_null NA NA NA NA 63.60 0.05 Bolitoglossa robusta Mod_use 35.78 46.545 NA NA 20.68 0.56 Bolitoglossa robusta Mod_clim NA NA 11.01 6.220 23.08 0.17 Bolitoglossa robusta Mod_clim_use 77.96 54.780 -6.04 9.917 22.26 0.25 Bolitoglossa robusta Mod_null NA NA NA NA 27.60 0.02 238

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Branta sandvicensis Mod_use -8.71 3.291 NA NA 50.68 0.65 Branta sandvicensis Mod_clim NA NA -0.20 0.913 61.53 0.00 Branta sandvicensis Mod_clim_use -9.69 3.670 0.76 0.927 51.96 0.34 Branta sandvicensis Mod_null NA NA NA NA 59.57 0.01 Apalharpactes reinwardtii Mod_use 15.32 32.701 NA NA 21.56 0.22 Apalharpactes reinwardtii Mod_clim NA NA 2.12 9.777 21.86 0.19 Apalharpactes reinwardtii Mod_clim_use 14.93 38.379 0.30 9.868 23.56 0.08 Apalharpactes reinwardtii Mod_null NA NA NA NA 19.94 0.50 Porphyrio hochstetteri Mod_use 1608.37 1448.155 NA NA 43.01 0.27 Porphyrio hochstetteri Mod_clim NA NA -0.75 0.908 43.75 0.19 Porphyrio hochstetteri Mod_clim_use NA NA NA NA NA NA Porphyrio hochstetteri Mod_null NA NA NA NA 42.43 0.36 Macgregoria pulchra Mod_use -978.04 2626.869 NA NA 40.23 0.04 Macgregoria pulchra Mod_clim NA NA 14.38 6.131 34.13 0.78 Macgregoria pulchra Mod_clim_use -8632.06 0.002 12.63 0.002 41.42 0.02 Macgregoria pulchra Mod_null NA NA NA NA 37.32 0.16 Pyrrhura calliptera Mod_use -0.87 7.300 NA NA 33.25 0.20 Pyrrhura calliptera Mod_clim NA NA 0.56 4.707 33.25 0.20 Pyrrhura calliptera Mod_clim_use -1.85 9.316 1.10 5.772 35.21 0.07 Pyrrhura calliptera Mod_null NA NA NA NA 31.27 0.53 Basileuterus conspicillatus Mod_use -10.43 6.645 NA NA 69.09 0.61 Basileuterus conspicillatus Mod_clim NA NA -1.04 1.282 75.08 0.03 Basileuterus conspicillatus Mod_clim_use -11.49 8.755 -1.06 1.549 70.47 0.31 Basileuterus conspicillatus Mod_null NA NA NA NA 73.86 0.06 239

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Ixos siquijorensis Mod_use 47.46 27.983 NA NA 9.46 0.45 Ixos siquijorensis Mod_clim NA NA -2.35 12.191 12.39 0.10 Ixos siquijorensis Mod_clim_use 46.84 1091.571 -4.47 1299.438 11.38 0.17 Ixos siquijorensis Mod_null NA NA NA NA 10.45 0.27 Turdus swalesi Mod_use 106.21 51.796 NA NA 23.09 0.12 Turdus swalesi Mod_clim NA NA -15.45 8.223 20.05 0.56 Turdus swalesi Mod_clim_use 6.65 14.348 -13.83 8.418 21.77 0.24 Turdus swalesi Mod_null NA NA NA NA 24.05 0.08 Iguana delicatissima Mod_use 14.06 10.812 NA NA 17.57 0.72 Iguana delicatissima Mod_clim NA NA 6.28 4.531 25.30 0.02 Iguana delicatissima Mod_clim_use 13.76 10.970 0.75 7.649 19.56 0.27 Iguana delicatissima Mod_null NA NA NA NA 30.00 0.00 Atelopus chiriquiensis Mod_use -2.76 4.502 NA NA 43.08 0.22 Atelopus chiriquiensis Mod_clim NA NA -1.15 0.002 43.19 0.21 Atelopus chiriquiensis Mod_clim_use -2.68 0.002 -0.08 0.002 45.07 0.08 Atelopus chiriquiensis Mod_null NA NA NA NA 41.45 0.49 Vireo masteri Mod_use 30.47 17.222 NA NA 51.83 0.71 Vireo masteri Mod_clim NA NA -0.08 0.897 65.45 0.00 Vireo masteri Mod_clim_use 30.24 17.270 0.38 0.788 53.60 0.29 Vireo masteri Mod_null NA NA NA NA 63.45 0.00 Pteropus niger Mod_use -92.69 31.773 NA NA 13.58 0.36 Pteropus niger Mod_clim NA NA -1.12 6.319 15.62 0.13 Pteropus niger Mod_clim_use -93.95 38.725 7.94 10.732 15.33 0.15 Pteropus niger Mod_null NA NA NA NA 13.65 0.35 240

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Hemitriccus mirandae Mod_use -75.81 50.740 NA NA 21.08 0.64 Hemitriccus mirandae Mod_clim NA NA -0.74 4.232 36.67 0.00 Hemitriccus mirandae Mod_clim_use -116.95 60.416 -8.50 10.009 22.26 0.36 Hemitriccus mirandae Mod_null NA NA NA NA 34.70 0.00 Capito hypoleucus Mod_use 3.31 8.929 NA NA 24.64 0.20 Capito hypoleucus Mod_clim NA NA 0.35 4.999 24.74 0.19 Capito hypoleucus Mod_clim_use 3.36 9.091 0.35 4.994 26.63 0.08 Capito hypoleucus Mod_null NA NA NA NA 22.75 0.53 Mesocapromys nanus Mod_use 82.40 39.576 NA NA 22.39 0.00 Mesocapromys nanus Mod_clim NA NA 47.09 32.930 9.38 1.00 Mesocapromys nanus Mod_clim_use NA NA NA NA NA NA Mesocapromys nanus Mod_null NA NA NA NA 22.72 0.00 Pteralopex taki Mod_use NA NA NA NA NA NA Pteralopex taki Mod_clim NA NA -0.38 6.735 18.62 0.20 Pteralopex taki Mod_clim_use NA NA NA NA NA NA Pteralopex taki Mod_null NA NA NA NA 16.62 0.53 Didunculus strigirostris Mod_use 1785.53 957155.673 NA NA 16.06 0.56 Didunculus strigirostris Mod_clim NA NA NA NA NA NA Didunculus strigirostris Mod_clim_use NA NA NA NA NA NA Didunculus strigirostris Mod_null NA NA NA NA 16.50 0.44 Atlapetes fuscoolivaceus Mod_use 16.08 8.973 NA NA 32.39 0.66 Atlapetes fuscoolivaceus Mod_clim NA NA -0.84 1.621 38.74 0.03 Atlapetes fuscoolivaceus Mod_clim_use 16.40 9.191 -0.32 1.048 34.29 0.25 Atlapetes fuscoolivaceus Mod_null NA NA NA NA 37.05 0.06 241

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Cistothorus apolinari Mod_use 436.28 57.475 NA NA 10.17 0.72 Cistothorus apolinari Mod_clim NA NA 0.16 4.315 24.63 0.00 Cistothorus apolinari Mod_clim_use 386.69 74.192 3.68 11.361 12.11 0.27 Cistothorus apolinari Mod_null NA NA NA NA 22.63 0.00 Odontophorus strophium Mod_use 1.41 16.646 NA NA 30.38 0.20 Odontophorus strophium Mod_clim NA NA -0.14 4.028 30.38 0.20 Odontophorus strophium Mod_clim_use 1.72 17.831 -0.26 4.211 32.37 0.07 Odontophorus strophium Mod_null NA NA NA NA 28.39 0.53 Lithobates vibicarius Mod_use -1.39 10.907 NA NA 18.66 0.20 Lithobates vibicarius Mod_clim NA NA 0.49 5.219 18.67 0.20 Lithobates vibicarius Mod_clim_use -2.51 12.913 1.02 6.265 20.64 0.07 Lithobates vibicarius Mod_null NA NA NA NA 16.68 0.53 Columbina cyanopis Mod_use NA NA NA NA NA NA Columbina cyanopis Mod_clim NA NA -2.98 10.630 24.45 0.28 Columbina cyanopis Mod_clim_use NA NA NA NA NA NA Columbina cyanopis Mod_null NA NA NA NA 22.56 0.72 Pyrrhura griseipectus Mod_use -146.66 42.458 NA NA 17.36 0.21 Pyrrhura griseipectus Mod_clim NA NA -1.08 5.256 18.61 0.11 Pyrrhura griseipectus Mod_clim_use -307.81 53.138 -100.63 20.167 16.11 0.39 Pyrrhura griseipectus Mod_null NA NA NA NA 16.65 0.30 Grallaricula cucullata Mod_use 0.07 7.405 NA NA 21.63 0.20 Grallaricula cucullata Mod_clim NA NA -1.05 6.274 21.60 0.20 Grallaricula cucullata Mod_clim_use 0.29 7.397 -1.09 6.356 23.60 0.07 Grallaricula cucullata Mod_null NA NA NA NA 19.63 0.53 242

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Geospiza difficilis Mod_use 130.90 75.218 NA NA 18.37 0.28 Geospiza difficilis Mod_clim NA NA -19.15 9.747 18.27 0.29 Geospiza difficilis Mod_clim_use 90.38 52.482 -11.82 8.894 19.95 0.13 Geospiza difficilis Mod_null NA NA NA NA 18.20 0.30 Myzomela rubratra Mod_use NA NA NA NA NA NA Myzomela rubratra Mod_clim NA NA -2.12 15.554 12.66 0.27 Myzomela rubratra Mod_clim_use NA NA NA NA NA NA Myzomela rubratra Mod_null NA NA NA NA 10.68 0.73 Habia atrimaxillaris Mod_use 3.71 8.318 NA NA 27.12 0.20 Habia atrimaxillaris Mod_clim NA NA 1.49 2.409 26.92 0.22 Habia atrimaxillaris Mod_clim_use 3.60 8.116 1.48 2.402 28.72 0.09 Habia atrimaxillaris Mod_null NA NA NA NA 25.33 0.49 Hapalomys longicaudatus Mod_use 143.51 76.882 NA NA 17.95 0.22 Hapalomys longicaudatus Mod_clim NA NA 2.66 9.860 18.49 0.17 Hapalomys longicaudatus Mod_clim_use 213.61 0.002 -35.59 0.002 18.36 0.18 Hapalomys longicaudatus Mod_null NA NA NA NA 16.57 0.44 Turdus lherminieri Mod_use NA NA NA NA NA NA Turdus lherminieri Mod_clim NA NA NA NA NA NA Turdus lherminieri Mod_clim_use NA NA NA NA NA NA Turdus lherminieri Mod_null NA NA NA NA 20.31 1.00 Poospiza rubecula Mod_use 29.88 41.462 NA NA 22.80 0.00 Poospiza rubecula Mod_clim NA NA NA NA NA NA Poospiza rubecula Mod_clim_use NA NA NA NA NA NA Poospiza rubecula Mod_null NA NA NA NA 19.38 0.02 243

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Species Model Land use Climate Change Model comparation Coef. SE Coef. SE AICc Weight Aythya innotata Mod_use 21.45 55.960 NA NA 18.37 0.07 Aythya innotata Mod_clim NA NA 1803.18 44397977.251 15.38 0.31 Aythya innotata Mod_clim_use NA NA NA NA NA NA Aythya innotata Mod_null NA NA NA NA 16.54 0.17 Procolobus kirkii Mod_use 179.56 118.284 NA NA 12.26 0.39 Procolobus kirkii Mod_clim NA NA -84.43 0.004 13.11 0.26 Procolobus kirkii Mod_clim_use 179.19 103.776 -536.93 142.722 14.26 0.14 Procolobus kirkii Mod_null NA NA NA NA 13.57 0.20 Anthracoceros montani Mod_use NA NA NA NA NA NA Anthracoceros montani Mod_clim NA NA -1.07 7.858 15.55 0.27 Anthracoceros montani Mod_clim_use NA NA NA NA NA NA Anthracoceros montani Mod_null NA NA NA NA 13.57 0.73 Isthmohyla angustilineata Mod_use 64.91 33.312 NA NA 17.94 0.59 Isthmohyla angustilineata Mod_clim NA NA 3.35 1.679 19.86 0.23 Isthmohyla angustilineata Mod_clim_use 4.21 4.766 2.83 1.734 20.97 0.13 Isthmohyla angustilineata Mod_null NA NA NA NA 22.83 0.05 Rallus semiplumbeus Mod_use -1.30 1.839 NA NA 27.95 0.17 Rallus semiplumbeus Mod_clim NA NA 1.58 1.221 26.62 0.33 Rallus semiplumbeus Mod_clim_use -0.84 2.005 1.46 1.237 28.44 0.13 Rallus semiplumbeus Mod_null NA NA NA NA 26.47 0.36 Chasiempis sclateri Mod_use 9.28 5.785 NA NA 14.71 0.84 Chasiempis sclateri Mod_clim NA NA NA NA NA NA Chasiempis sclateri Mod_clim_use NA NA NA NA NA NA Chasiempis sclateri Mod_null NA NA NA NA 17.96 0.16 244

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.5. Results comparing the autocorrelation aapying Morna Index in original data (Data) and residuals from the species models based on regression models controlling the spatialautocorrelation (GLMM) and in GLM without control for the spatial autocorrelation. Due to the calculation of spatial autocorrelation requires a high use of RAM memory, the autocorrelation for 27 species with big ranges could not be calculated (Dendrocygna viduata, Ursus arctos, Troglodytes aedon, Coccyzus americanus, Tayassu pecari, Myrmecophaga tridactyla, Tapirus terrestris, Aramus guarauna, Oxyura leucocephala, Ibycter americanus, Ara ararauna, Cricetulus migratorius, Capreolus capreolus, Ursus thibetanus, Ara macao, Alcelaphus buselaphus, Martes zibellina, Gyps bengalensis, Diceros bicornis, Pseudomys desertor, Tyto longimembris, Coturnicops noveboracensis, Barbastella barbastellus, Sterna acuticauda, Helarctos malayanus, Kobus kob, Melursus ursinus ). Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Anthus spragueii Bird 0.24 0.01 0.01 0.01 0.01 0.24 0.23 0.24 0.24 Sypheotides indicus Bird 0.10 0.00 0.00 0.00 0.00 0.11 0.11 0.11 0.10 Turnix varius Bird 0.20 0.19 0.19 0.19 0.19 0.20 0.20 0.20 0.20 Leptoptilos dubius Bird 0.52 0.49 0.49 0.48 0.49 0.50 0.52 0.50 0.52 Hipposideros cervinus Mammal 0.02 0.00 0.00 0.00 0.00 0.02 0.02 0.02 0.02 Euscarthmus rufomarginatus Bird 0.11 0.10 0.10 0.10 0.10 0.11 0.11 0.12 0.11 Pavo muticus Bird 0.36 0.03 0.03 0.03 0.02 0.35 0.32 0.33 0.36 Gyps tenuirostris Bird 0.39 0.02 0.02 0.02 0.02 0.38 0.38 0.35 0.39 Merops breweri Bird 0.17 0.14 0.15 0.14 0.16 0.15 0.16 0.15 0.17 Amaurolimnas concolor Bird 0.22 0.21 0.21 0.21 0.21 0.22 0.22 0.22 0.22 Oryzomys couesi Mammal 0.18 0.17 0.17 0.17 0.17 0.18 0.18 0.18 0.18 Macaca arctoides Mammal 0.16 0.00 0.00 0.00 0.00 0.16 0.16 0.16 0.16 Tragulus napu Mammal 0.02 0.00 0.00 0.00 0.00 0.02 0.02 0.03 0.02 Gyps coprotheres Bird 0.57 0.55 0.55 0.55 0.56 0.50 0.56 0.50 0.57 Mergus octosetaceus Bird 0.19 0.00 0.00 0.00 0.00 0.15 0.17 0.14 0.19 Coronella girondica Reptile 0.11 0.10 0.10 0.10 0.10 0.11 0.11 0.11 0.11 245

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Turdoides bicolor Bird 0.08 0.06 0.07 0.06 0.07 0.06 0.07 0.06 0.08 Chaetornis striata Bird 0.49 0.05 0.10 0.02 0.10 0.41 0.47 0.36 0.49 Ardeotis nigriceps Bird 0.22 0.01 0.01 0.01 0.01 0.20 0.22 0.20 0.22 Pipile jacutinga Bird 0.09 0.00 0.00 0.00 0.00 0.09 0.09 0.09 0.09 Lissotriton helveticus Amphibian 0.04 0.03 0.03 0.03 0.03 0.05 0.04 0.05 0.04 Jacana spinosa Bird 0.04 -0.01 0.00 -0.01 0.00 0.05 0.04 0.05 0.04 Regina septemvittata Reptile 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 Alectrurus tricolor Bird 0.34 0.02 0.02 0.02 0.02 0.34 0.29 0.29 0.34 Penelope pileata Bird 0.04 0.00 0.00 0.00 0.00 0.04 0.04 0.04 0.04 Picoides borealis Bird 0.19 0.01 0.01 0.01 0.01 0.18 0.18 0.18 0.19 Tapirus bairdii Mammal 0.14 0.00 0.00 0.00 0.00 0.14 0.14 0.14 0.14 Poephila cincta Bird 0.55 0.06 0.01 0.01 0.06 0.54 0.38 0.37 0.55 Alectrurus risora Bird 0.40 0.38 0.38 0.37 0.39 0.36 0.30 0.29 0.40 Picus squamatus Bird 0.29 0.01 0.01 0.01 0.01 0.24 0.28 0.23 0.29 Pseudocolopteryx dinelliana Bird 0.27 0.23 0.22 0.23 0.24 0.27 0.27 0.27 0.27 Saxicola macrorhynchus Bird 0.27 0.01 0.01 0.01 0.01 0.21 0.23 0.19 0.27 Empidonax fulvifrons Bird 0.54 0.03 0.03 0.03 0.03 0.48 0.53 0.48 0.54 Megaxenops parnaguae Bird 0.06 0.00 0.00 0.00 0.00 0.06 0.06 0.05 0.06 Anodorhynchus hyacinthinus Bird 0.29 0.05 0.05 0.05 0.05 0.29 0.29 0.29 0.29 Agriornis albicauda Bird 0.38 0.03 0.03 0.03 0.03 0.38 0.38 0.38 0.38 Xanthopsar flavus Bird 0.27 0.01 0.01 0.01 0.01 0.24 0.25 0.24 0.27 Vireo atricapilla Bird 0.38 0.01 0.01 0.05 0.01 0.35 0.26 0.21 0.38 246

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Sus barbatus Mammal 0.39 0.36 0.36 0.36 0.36 0.39 0.38 0.38 0.39 Pelobates syriacus Amphibian 0.16 0.14 0.14 0.14 0.14 0.16 0.15 0.15 0.16 Cyanolanius madagascarinus Bird 0.06 0.05 0.05 0.05 0.05 0.06 0.06 0.06 0.06 Procolobus badius Mammal 0.50 0.03 0.03 0.03 0.03 0.48 0.42 0.39 0.50 Platyrinchus leucoryphus Bird 0.06 0.00 0.00 0.00 0.00 0.06 0.06 0.06 0.06 Jacamaralcyon tridactyla Bird 0.36 0.01 0.01 0.01 0.01 0.29 0.37 0.29 0.36 Grus leucogeranus Bird 0.16 0.01 0.01 0.01 0.01 0.16 0.15 0.15 0.16 Gymnogyps californianus Bird 0.38 0.01 0.01 0.01 0.01 0.38 0.35 0.35 0.38 Pleurodeles waltl Amphibian 0.15 0.00 0.00 0.00 0.00 0.15 0.15 0.15 0.15 Fulica cornuta Bird 0.45 0.02 0.02 0.02 0.02 0.45 0.43 0.43 0.45 Ara militaris Bird 0.81 0.77 0.77 0.77 0.77 0.80 0.81 0.80 0.81 Phrynosoma douglasii Reptile 0.10 0.09 0.09 0.09 0.09 0.10 0.10 0.10 0.10 Guaruba guarouba Bird 0.34 0.01 0.01 0.01 0.01 0.32 0.33 0.30 0.34 Axis porcinus Mammal 0.53 0.15 0.15 0.15 0.15 0.50 0.53 0.50 0.53 Houbaropsis bengalensis Bird 0.28 0.01 0.01 0.01 0.01 0.25 0.28 0.26 0.28 Francolinus gularis Bird 0.37 0.02 0.03 0.02 0.02 0.37 0.36 0.35 0.37 Microhierax erythrogenys Bird 0.11 0.10 0.10 0.10 0.10 0.10 0.11 0.10 0.11 Conilurus penicillatus Mammal 0.47 0.43 0.39 0.39 0.43 0.46 0.44 0.44 0.47 Nyctimene cephalotes Mammal 0.57 0.54 0.54 0.54 0.54 0.56 0.57 0.55 0.57 Coryphaspiza melanotis Bird 0.63 0.08 0.08 0.08 0.07 0.55 0.57 0.53 0.63 Cacatua haematuropygia Bird 0.24 0.02 0.01 0.03 0.01 0.24 0.25 0.24 0.24 Psophodes nigrogularis Bird 0.42 0.02 0.02 0.02 0.02 0.41 0.41 0.40 0.42 247

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Patagioenas inornata Bird 0.23 0.00 0.00 0.00 0.00 0.22 0.22 0.22 0.23 Lipaugus lanioides Bird 0.13 0.10 0.11 0.10 0.11 0.12 0.13 0.12 0.13 Pyrilia pyrilia Bird 0.18 0.00 0.00 0.00 0.00 0.18 0.18 0.18 0.18 Hymenolaimus malacorhynchos Bird 0.11 0.00 0.00 0.00 0.00 0.10 0.11 0.11 0.11 Laniisoma elegans Bird 0.40 0.02 0.02 0.02 0.02 0.40 0.29 0.29 0.40 Bos javanicus Mammal 0.32 0.29 0.28 0.28 0.30 0.32 0.31 0.30 0.32 Hypothymis coelestis Bird 0.32 0.29 0.30 0.29 0.30 0.30 0.32 0.30 0.32 Amazona oratrix Bird 0.48 0.02 0.02 0.02 0.02 0.46 0.46 0.44 0.48 Meleagris ocellata Bird 0.13 0.00 0.00 0.00 0.00 0.12 0.12 0.11 0.13 Lithobates warszewitschii Amphibian 0.06 -0.01 -0.01 -0.01 -0.01 0.06 0.06 0.07 0.06 Amazona vinacea Bird 0.49 0.10 0.09 0.09 0.10 0.48 0.44 0.44 0.49 Acerodon jubatus Mammal 0.32 0.01 0.01 0.01 0.01 0.30 0.30 0.30 0.32 Iodopleura pipra Bird 0.38 0.36 0.36 0.36 0.36 0.38 0.38 0.37 0.38 Poliocephalus rufopectus Bird 0.68 0.63 0.20 0.20 0.64 0.64 0.60 0.58 0.68 Thripophaga macroura Bird 0.04 -0.01 -0.01 -0.01 -0.01 0.04 0.03 0.03 0.04 Lacerta schreiberi Reptile 0.14 0.13 0.13 0.13 0.13 0.14 0.14 0.14 0.14 Crax blumenbachii Bird 0.01 -0.01 -0.01 -0.01 -0.01 0.01 0.02 0.02 0.01 Padda oryzivora Bird 0.53 0.03 0.03 0.03 0.03 0.53 0.53 0.53 0.53 Babyrousa celebensis Mammal 0.32 0.01 0.01 0.01 0.01 0.31 0.32 0.31 0.32 Amazona leucocephala Bird 0.13 0.00 0.00 0.00 0.00 0.13 0.13 0.13 0.13 Wilfredomys oenax Mammal 0.50 0.43 0.45 0.42 0.46 0.47 0.49 0.45 0.50 Microtus cabrerae Mammal 0.26 0.00 0.00 0.00 0.00 0.26 0.26 0.26 0.26 248

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Geotrygon caniceps Bird 0.26 0.01 0.01 0.01 0.01 0.22 0.25 0.22 0.26 Accipiter gundlachi Bird 0.18 0.00 0.00 0.00 0.00 0.15 0.18 0.14 0.18 Aratinga euops Bird 0.14 0.00 0.00 0.00 0.00 0.12 0.12 0.10 0.14 Starnoenas cyanocephala Bird 0.17 0.00 0.00 0.00 0.00 0.15 0.18 0.16 0.17 Tyrannus cubensis Bird 0.08 -0.01 -0.01 -0.01 -0.01 0.08 0.08 0.07 0.08 Thylogale brunii Mammal 0.58 0.49 0.50 0.46 0.54 0.51 0.52 0.45 0.58 Rusa marianna Mammal 0.26 0.22 0.21 0.21 0.23 0.25 0.24 0.23 0.26 Chondrohierax wilsonii Bird 0.21 0.00 0.00 0.00 0.00 0.18 0.21 0.18 0.21 Colaptes fernandinae Bird 0.12 0.00 0.00 0.00 0.00 0.11 0.11 0.11 0.12 Pachycephala rufogularis Bird 0.18 0.01 0.01 0.01 0.01 0.18 0.18 0.18 0.18 Ara ambiguus Bird 0.25 0.00 0.00 0.00 0.00 0.25 0.25 0.25 0.25 Carduelis cucullata Bird 0.50 0.01 0.01 0.01 0.01 0.47 0.50 0.47 0.50 Xiphocolaptes falcirostris Bird 0.47 0.44 0.43 0.43 0.44 0.47 0.44 0.44 0.47 Tangara peruviana Bird 0.19 0.00 0.00 0.00 0.00 0.18 0.15 0.15 0.19 Equus africanus Mammal 0.35 0.01 0.01 0.01 0.01 0.33 0.32 0.30 0.35 Callaeas cinereus Bird 0.12 0.00 0.00 0.00 0.00 0.12 0.12 0.12 0.12 Lithobates tarahumarae Amphibian 0.22 0.19 0.19 0.19 0.19 0.22 0.21 0.21 0.22 Anthus nattereri Bird 0.50 0.44 0.19 0.19 0.46 0.47 0.40 0.39 0.50 Zaglossus bruijnii Mammal 0.19 0.18 0.18 0.18 0.18 0.19 0.19 0.19 0.19 Nestor meridionalis Bird 0.39 0.01 0.01 0.01 0.01 0.39 0.37 0.37 0.39 Corvus leucognaphalus Bird 0.07 0.00 -0.01 -0.01 0.00 0.06 0.05 0.05 0.07 Cotinga maculata Bird 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 249

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Sporophila frontalis Bird 0.36 0.00 0.00 0.02 0.00 0.36 0.36 0.36 0.36 Picumnus fulvescens Bird 0.05 0.04 0.05 0.04 0.05 0.05 0.05 0.04 0.05 Penelope ortoni Bird 0.44 0.00 0.01 0.00 0.01 0.27 0.44 0.27 0.44 Pituophis ruthveni Reptile 0.03 -0.01 -0.01 -0.01 -0.01 0.03 0.03 0.02 0.03 Penelope ochrogaster Bird 0.41 0.36 0.36 0.36 0.36 0.41 0.41 0.40 0.41 Zaratornis stresemanni Bird 0.12 -0.02 -0.02 -0.02 -0.02 0.12 0.12 0.12 0.12 Ptilinopus jambu Bird 0.89 0.16 0.74 0.19 0.82 0.57 0.80 0.57 0.89 Carpornis melanocephala Bird 0.53 0.01 0.01 0.01 0.01 0.35 0.41 0.30 0.53 Amazona viridigenalis Bird 0.17 0.00 0.00 0.00 0.00 0.18 0.17 0.17 0.17 Crypturellus erythropus Bird 0.18 -0.01 0.00 -0.01 0.00 0.19 0.18 0.19 0.18 Aratinga chloroptera Bird 0.19 0.00 0.00 0.00 0.00 0.19 0.17 0.17 0.19 Parus nuchalis Bird 0.70 0.65 0.03 0.03 0.66 0.69 0.26 0.26 0.70 Mastacomys fuscus Mammal 0.32 0.29 0.28 0.28 0.29 0.32 0.29 0.28 0.32 Aphelocoma coerulescens Bird 0.03 -0.01 -0.01 -0.01 -0.01 0.03 0.03 0.03 0.03 Micrastur plumbeus Bird 0.14 0.00 -0.01 -0.01 0.00 0.14 0.14 0.14 0.14 Sus philippensis Mammal 0.25 0.22 0.22 0.22 0.22 0.25 0.25 0.25 0.25 Touit surdus Bird 0.74 0.31 0.27 0.25 0.64 0.66 0.59 0.57 0.74 Taoniscus nanus Bird 0.80 0.77 0.77 0.76 0.77 0.79 0.80 0.76 0.80 Atrichornis rufescens Bird 0.14 0.00 0.00 0.00 0.00 0.14 0.14 0.14 0.14 Pseudibis davisoni Bird 0.68 0.19 0.56 0.15 0.63 0.55 0.60 0.53 0.68 Claravis godefrida Bird 0.63 0.49 0.51 0.46 0.54 0.57 0.52 0.48 0.63 Pauxi unicornis Bird 0.54 0.06 0.06 0.06 0.06 0.54 0.54 0.54 0.54 250

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Strabomantis bufoniformis Amphibian 0.22 -0.03 0.07 0.07 -0.03 0.22 0.22 0.22 0.22 Crax globulosa Bird 0.58 0.49 0.49 0.49 0.49 0.57 0.58 0.57 0.58 Charitospiza eucosma Bird 0.69 0.64 0.63 0.63 0.64 0.69 0.66 0.64 0.69 Myrmotherula unicolor Bird 0.67 0.03 0.03 0.03 0.03 0.66 0.63 0.61 0.67 Cinclodes aricomae Bird 0.14 -0.01 -0.01 -0.01 -0.01 0.14 0.14 0.14 0.14 Pyrrhura cruentata Bird 0.32 0.03 0.03 0.03 0.02 0.31 0.29 0.28 0.32 Poospiza cinerea Bird 0.54 0.50 0.48 0.44 0.50 0.54 0.41 0.36 0.54 Nomascus leucogenys Mammal 0.51 0.07 0.07 0.02 0.07 0.44 0.44 0.32 0.51 Pristimantis caryophyllaceus Amphibian 0.34 0.00 0.00 0.00 0.00 0.33 0.34 0.32 0.34 Haeromys pusillus Mammal 0.68 0.57 0.60 0.54 0.64 0.57 0.63 0.55 0.68 Triclaria malachitacea Bird 0.78 0.69 0.22 0.20 0.71 0.73 0.48 0.43 0.78 Piprites pileata Bird 0.35 0.33 0.33 0.33 0.33 0.36 0.35 0.36 0.35 Sporophila falcirostris Bird 0.58 0.51 0.55 0.47 0.56 0.52 0.51 0.47 0.58 Asthenes heterura Bird 0.24 -0.01 -0.01 -0.01 -0.01 0.23 0.23 0.23 0.24 Crax alberti Bird 0.10 -0.01 -0.01 -0.01 -0.01 0.09 0.10 0.09 0.10 Xipholena atropurpurea Bird 0.19 0.15 0.16 0.15 0.16 0.18 0.19 0.18 0.19 Electron carinatum Bird 0.69 0.58 0.61 0.57 0.61 0.55 0.66 0.53 0.69 Odontophorus hyperythrus Bird 0.08 0.00 0.00 0.00 0.00 0.07 0.08 0.08 0.08 Leptotila ochraceiventris Bird 0.30 0.00 0.00 0.00 0.00 0.23 0.25 0.20 0.30 Craugastor ranoides Amphibian 0.27 0.00 0.00 0.00 0.00 0.22 0.26 0.22 0.27 Nothura minor Bird 0.35 0.30 0.29 0.30 0.29 0.36 0.34 0.36 0.35 Apteryx haastii Bird 0.26 0.00 0.00 0.00 0.00 0.26 0.26 0.25 0.26 251

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Pauxi pauxi Bird 0.09 -0.02 -0.02 -0.02 -0.02 0.09 0.09 0.09 0.09 Touit stictopterus Bird 0.56 0.43 0.48 0.42 0.49 0.46 0.55 0.46 0.56 Myiarchus semirufus Bird 0.10 -0.01 -0.02 -0.02 -0.01 0.08 0.10 0.08 0.10 Onychorhynchus occidentalis Bird 0.26 0.00 0.00 0.00 0.00 0.22 0.27 0.22 0.26 Bucco noanamae Bird 0.03 -0.02 -0.02 -0.02 -0.02 0.03 0.02 0.02 0.03 Xenospingus concolor Bird 0.05 -0.05 -0.05 -0.05 -0.05 0.04 0.05 0.04 0.05 Dacnis nigripes Bird 0.21 -0.01 -0.02 -0.01 -0.02 0.21 0.20 0.19 0.21 Mohoua ochrocephala Bird 0.30 -0.01 -0.01 -0.01 -0.01 0.25 0.24 0.21 0.30 Rana sierrae Amphibian 0.07 -0.01 -0.01 -0.01 -0.01 0.07 0.07 0.07 0.07 Dasyornis brachypterus Bird 0.24 0.00 0.00 0.00 0.00 0.24 0.23 0.23 0.24 Biatas nigropectus Bird 0.64 0.19 0.57 0.50 0.57 0.56 0.64 0.56 0.64 Penelopides panini Bird 0.10 -0.01 0.03 0.03 -0.01 0.10 0.10 0.10 0.10 Saltator rufiventris Bird 0.27 0.00 0.00 0.00 0.00 0.27 0.27 0.27 0.27 Amazona rhodocorytha Bird 0.46 0.40 0.40 0.40 0.41 0.43 0.44 0.40 0.46 Paradisaea rudolphi Bird 0.51 0.46 0.46 0.46 0.46 0.49 0.51 0.49 0.51 Lynx pardinus Mammal 0.25 0.15 -0.04 0.14 -0.05 0.23 0.18 0.17 0.25 Amazona pretrei Bird 0.50 0.44 0.11 0.11 0.44 0.50 0.29 0.29 0.50 Cinclus schulzi Bird 0.18 -0.04 -0.04 -0.04 -0.04 0.16 0.17 0.16 0.18 Carduelis siemiradzkii Bird 0.35 0.01 0.01 0.01 0.01 0.31 0.32 0.28 0.35 Dicaeum haematostictum Bird 0.09 0.03 0.04 0.04 -0.01 0.09 0.09 0.09 0.09 Turdoides hindei Bird 0.19 0.05 0.05 0.05 0.05 0.19 0.18 0.18 0.19 Coracina ostenta Bird 0.09 0.03 0.04 0.04 -0.01 0.09 0.09 0.09 0.09 252

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Rhinomyias albigularis Bird 0.09 0.03 0.04 0.04 -0.01 0.09 0.09 0.09 0.09 Aceros waldeni Bird 0.52 0.43 0.47 0.18 0.47 0.49 0.51 0.47 0.52 Galbula pastazae Bird 0.14 -0.02 -0.02 -0.02 -0.02 0.14 0.14 0.14 0.14 Ortalis erythroptera Bird 0.31 0.28 0.30 0.28 0.31 0.25 0.28 0.24 0.31 Nothoprocta taczanowskii Bird 0.49 0.00 -0.01 0.00 -0.01 0.46 0.46 0.43 0.49 Atelopus varius Amphibian 0.45 0.01 0.01 0.01 0.01 0.44 0.45 0.44 0.45 Simoxenops striatus Bird 0.07 -0.04 -0.03 -0.04 -0.03 0.07 0.07 0.07 0.07 Gyalophylax hellmayri Bird 0.57 0.50 0.50 0.50 0.50 0.57 0.56 0.56 0.57 Agapornis nigrigenis Bird 0.26 -0.01 -0.01 -0.01 -0.01 0.25 0.25 0.24 0.26 Chlorochrysa nitidissima Bird 0.15 0.12 -0.06 0.13 0.15 0.12 0.15 0.13 0.15 Coccyzus rufigularis Bird 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 0.09 Phytotoma raimondii Bird 0.57 0.50 0.50 0.50 0.50 0.57 0.57 0.57 0.57 Dysithamnus plumbeus Bird 0.30 0.00 0.00 0.00 0.00 0.31 0.30 0.30 0.30 Emballonura semicaudata Mammal 0.58 0.48 0.47 0.47 0.48 0.59 0.50 0.49 0.58 Komodomys rintjanus Mammal 0.51 0.44 0.44 0.44 0.44 0.50 0.50 0.49 0.51 Oreothraupis arremonops Bird 0.24 -0.04 -0.04 -0.03 -0.04 0.19 0.24 0.19 0.24 Rana muscosa Amphibian 0.09 -0.02 -0.02 -0.02 -0.02 0.09 0.07 0.07 0.09 Herpsilochmus pectoralis Bird 0.79 0.67 0.37 0.03 0.69 0.78 0.46 0.26 0.79 Anairetes alpinus Bird 0.55 -0.03 0.01 0.02 -0.03 0.55 0.55 0.54 0.55 Myrmotherula urosticta Bird 0.26 0.00 0.00 0.00 0.00 0.26 0.24 0.24 0.26 Duellmanohyla uranochroa Amphibian 0.10 -0.02 -0.02 -0.02 -0.03 0.11 0.09 0.09 0.10 Calyptophilus frugivorus Bird 0.49 0.27 0.45 0.12 0.45 0.48 0.50 0.46 0.49 253

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Pteralopex flanneryi Mammal 0.59 0.55 0.55 0.55 0.55 0.58 0.53 0.53 0.59 Babyrousa babyrussa Mammal 0.28 0.27 0.26 0.25 0.27 0.26 0.27 0.24 0.28 Brotogeris pyrrhoptera Bird 0.37 0.00 0.00 0.00 0.00 0.30 0.22 0.20 0.37 Cotinga ridgwayi Bird 0.32 -0.02 -0.01 -0.02 -0.01 0.32 0.31 0.31 0.32 Myrmotherula minor Bird 0.55 0.01 0.06 0.06 0.00 0.53 0.54 0.53 0.55 Tachycineta euchrysea Bird 0.83 -0.01 -0.01 -0.01 -0.01 0.82 0.61 0.60 0.83 Dysithamnus occidentalis Bird 0.44 -0.01 0.37 0.37 -0.01 0.43 0.43 0.42 0.44 Hemitriccus furcatus Bird 0.65 0.56 0.58 0.54 0.58 0.64 0.64 0.63 0.65 Acanthodactylus schreiberi Reptile 0.71 0.16 0.63 0.16 0.62 0.44 0.58 0.43 0.71 Anthocephala floriceps Bird 0.30 -0.08 -0.07 -0.08 -0.07 0.29 0.22 0.22 0.30 Chasiempis sandwichensis Bird 0.10 0.02 0.02 0.02 0.02 0.10 0.10 0.10 0.10 Bangsia melanochlamys Bird 0.11 0.03 0.04 0.03 0.04 0.03 0.12 0.04 0.11 Hypopyrrhus pyrohypogaster Bird 0.34 0.18 0.18 0.18 0.18 0.34 0.35 0.34 0.34 Grallaria gigantea Bird 0.37 0.18 -0.02 0.18 -0.02 0.36 0.37 0.35 0.37 Eleothreptus candicans Bird 0.85 0.48 0.77 -0.02 0.77 0.80 0.86 0.32 0.85 Ammospermophilus nelsoni Mammal 0.10 0.02 -0.02 -0.01 0.02 0.10 0.10 0.10 0.10 Amazona agilis Bird 0.23 0.06 0.04 0.04 0.06 0.26 0.23 0.26 0.23 Coeligena prunellei Bird 0.18 0.06 -0.03 0.04 -0.03 0.10 0.18 0.10 0.18 Neopelma aurifrons Bird 0.66 0.29 0.56 0.13 0.56 0.54 0.65 0.48 0.66 Patagioenas oenops Bird 0.33 -0.02 -0.02 -0.02 -0.02 0.28 0.29 0.26 0.33 Macroagelaius subalaris Bird 0.20 0.19 0.18 0.19 0.19 0.21 0.20 0.21 0.20 Psammodromus microdactylus Reptile 0.70 0.21 0.61 0.25 0.61 0.52 0.70 0.51 0.70 254

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Sciname Class Moran Index

Data GLMM GLM Use Climate Aditive Null Use Climate Aditive Null

Grallaria rufocinerea Bird 0.74 0.68 0.68 0.68 0.68 0.73 0.74 0.72 0.74 Synallaxis tithys Bird 0.23 -0.03 -0.02 -0.01 -0.02 0.15 0.15 0.09 0.23 Branta sandvicensis Bird 0.30 0.13 0.26 0.13 0.26 0.13 0.29 0.13 0.30 Trachypithecus geei Mammal 0.13 0.00 -0.04 -0.04 -0.01 0.08 0.14 0.08 0.13 Canis simensis Mammal 0.42 0.38 0.37 0.14 0.39 0.30 0.41 0.29 0.42 Collocalia orientalis Bird 0.29 0.12 0.08 -0.04 -0.03 0.12 0.08 -0.04 0.29 Myadestes obscurus Bird 0.19 0.01 0.00 0.00 0.02 0.18 0.18 0.17 0.19 Bolitoglossa robusta Amphibian 0.12 -0.05 -0.05 -0.05 -0.03 0.01 0.12 -0.01 0.12 Apalharpactes reinwardtii Bird 0.54 0.45 0.49 0.45 0.49 0.39 0.52 0.39 0.54 Porphyrio hochstetteri Bird 0.06 0.06 0.05 0.06 0.06 0.07 0.05 0.06 0.06 Basileuterus conspicillatus Bird 0.06 -0.02 0.00 -0.02 0.00 0.05 0.06 0.05 0.06 Macgregoria pulchra Bird 0.31 -0.10 0.25 0.09 -0.10 0.30 0.29 0.29 0.31 Pyrrhura calliptera Bird 0.21 -0.04 -0.04 -0.04 -0.04 0.21 0.18 0.19 0.21 Ixos siquijorensis Bird 0.00 -0.01 0.00 -0.01 0.00 -0.02 0.00 -0.02 0.00 Turdus swalesi Bird 0.30 0.06 0.15 0.12 -0.10 0.18 0.29 0.19 0.30 Iguana delicatissima Reptile 0.30 -0.05 0.04 -0.05 0.12 -0.04 0.08 -0.04 0.30 Vireo masteri Bird 0.11 0.08 0.07 0.07 0.07 0.08 0.11 0.07 0.11 Capito hypoleucus Bird 0.20 0.23 0.23 0.23 0.23 0.20 0.20 0.20 0.20 Atelopus chiriquiensis Amphibian 0.37 -0.02 -0.02 -0.02 -0.02 0.37 0.37 0.37 0.37 Pteropus niger Mammal 0.83 -0.03 0.76 -0.03 0.79 0.46 0.82 0.40 0.83 Hemitriccus mirandae Bird 0.63 0.21 0.57 0.19 0.57 0.37 0.62 0.37 0.63

255

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.6. Distribution of historical, current, and extirpated ranges by continent for species included in the global analysis. Asia North America Europe Africa South America Oceania Australia Historical 78 83 21 38 142 25 14 Current 78 83 21 38 142 25 14 Extirpated 66 76 14 29 124 23 9

Table A3.7. Distribution of historical, current, and extirpated ranges by biome for species included in the global analysis. Tropical & Tropical & Tropical & Temperate Temperat Boreal Tropical & Subtropical Temperate Subtropic Subtropical Dry Subtropical Broadleaf & e Conifer Forests Grasslands, Savannas Grasslands, al Moist Broadleaf Coniferous Mixed Forest Forests /Taiga & Shrublands Savannas & Broadleaf Forests Forests Shrublands Forests Historical 271 150 77 74 58 19 141 60 Current 264 126 70 69 54 19 121 49 Extirpated 247 104 41 33 21 4 88 31

Flooded Montane Tundra Mediterranean Deserts & Mangroves Lakes Rock and Grasslands Grasslands Forests, Xeric Ice & Savannas & Woodlands & Shrublands Shrublands Scrub Historical 88 105 9 44 125 141 45 15 Current 72 89 9 42 106 115 44 15 Extirpated 46 44 1 25 73 91 5 0

256

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.8. Distribution of historical, current, and extirpated ranges by continent for species model. Asia North America Europe Africa South America Oceania Australia Historical 60 63 14 20 116 17 12 Current 60 63 13 19 116 16 12 Extirpated 55 58 10 16 104 16 8

Table A3.9. Distribution of historical, current, and extirpated ranges by biome for species model Tropical & Tropical & Tropical & Temperate Temperat Boreal Tropical & Subtropical Temperate Subtropical Subtropical Dry Subtropical Broadleaf & e Conifer Forests Grasslands, Savannas Grasslands, Moist Broadleaf Coniferous Mixed Forest Forests /Taiga & Shrublands Savannas & Broadleaf Forests Forests Shrublands Forests Historical 213 123 61 49 37 10 109 34 Current 208 104 55 44 34 10 92 26 Extirpate d 199 92 34 26 16 4 73 20

Flooded Montane Tundra Mediterranean Deserts & Mangroves Lakes Rock and Grasslands Grasslands & Forests, Xeric Ice & Savannas Shrublands Woodlands & Shrublands Scrub Historical 64 74 4 28 94 113 24 6 Current 48 60 4 26 78 89 24 6 Extirpated 43 39 1 19 59 78 4 0

257

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.10. Classification of species based in estimate association and in standard error. Best model Pattern N species based in Species within each category with IC not including estimate association “0”

1*SE 1,96*SE

N species (%) N species (%)

Null - 126 - - Climate Supporting 24 21 (87.50%) 14 (58.33%) Not Supporting 19 19 (100.00%) 12 (63.16%) Use Supporting 34 32 (94.12%) 20 (58.82%) Not Supporting 10 10 (100.00%) 5 (50.00%) Additive Supporting 24 22 (91.67%) 12 (50.00%) Supporting only clim prediction 7 7 (100.00%) 6 (85.71%) Supporting only use prediction 15 15 (100.00%) 12 (80.00%)

Not supporting any 3

262 TOTAL

258

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.11. Linear models describing change in environmental variables with the latitude (in km) by impact. Impact Coefficients Model ∆Latitude (km) ∆Elevation (m) ∆Slope (%) Intercept Dist_Equator Intercept Dist_Equator Intercept Dist_Equator Climate change -69.8 (239.80) 0.158 (0.076)* 246 (108.4)* -0.031 (0.0345) 4.13 (2.488) -0.001237(0.000792) Land use -213.5 (113.9)† -0.015 (0.056) -53.9 (54.95) 0.04222 (0.02677) 4.08 -0.001119 (0.0007848) (1.611)* Additive -91.0 (153.60) 0.118 (0.055)* 320 (100.1)** -0. 21 (0.036) 6.659 0.0001868 (0.0008807) (2.446)* **P < 0.01; *P < 0.05; †P < 0.10

259

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Table A3.12. Results from the regression analyses based on regression models (GLMM) to evaluate the importance of each descriptor of climate change (Thermal_limit_qX, being “X” one of the five quantiles used as thermal limit) for range contraction in terrestrial vertebrates. We report model coefficients (best estimates and their SE), AIC, ΔAIC (difference in AIC with the best model). Model Coefficients Model comparison N=335 Thermal_limit_qX AIC ΔAIC

Mod_Thermal_limit_q0.75 1.89 (0.356)* 910.25 0.00 Mod_Thermal_limit_q0.80 1.93 (0.366)* 910.28 0.02 Mod_Thermal_limit_q0.85 1.96 (0.379)* 911.25 1.00 Mod_Thermal_limit_q0.90 1.97 (0.404)* 913.97 3.72 Mod_Thermal_limit_q0.95 2.14 (0.469)* 916.39 6.14 *P < 0.05

Table A3.14. Results from the regression analyses based on regression models (GLMM) to evaluate the importance of each descriptor of land use (Land cover and Human density) for range contraction in terrestrial vertebrates. We report model coefficients (best estimates and their SE), AIC, ΔAIC (difference in AIC with the best model). Dashes indicate variables not included in the model. Model Coefficients Model comparison N=335 Land cover Human density AIC ΔAIC

Land_Cover 2.05 (0.363)* - 906.63 0.00 Hum_density - 0.0023 (0.00054)* 909.11 2.48 *P < 0.05

260

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Figure A3.1. Distribution frequencies of the area of historical distributions, the area of extirpated distributions, the area of current distributions and the percentage of contraction of species included in the global analysis.

261

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Figure A3.2. Distribution frequencies of the area of historical distributions, the area of extirpated distributions, the area of current distributions and the percentage of contraction of species included in the species analysis.

262

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Figure A3.3. Distribution of frequencies of the Moran Index of residuals for species models ussing GLM wich not include a random variable for control the spatial autocorrelation (Cli_glm for climate model, Use_glm for land use model, Additive_glm for model including climate and land use impacts as additive), for residuals of species models using GLMM which include a random variable to control for the spatial autocorrelation (Cli for climate model, Use for land use model, Additive for model including climate and land use impacts as additive) and original data (Data). Moran Index ranges between -1, indicating a negative spatial autocorrelation, and 1 indicating positive spatial autocorrelation and with 0 indicating absent of spatial autocorrelation.

263

3. The roles of climate change and land use in recent terrestrial vertébrate range contractions

Figure A3.4. The distribution for the extinct area, the current area and the historical area of the distributions included in the species analysis. Ends of the whiskers represent the lowest datum still within the 1.5 interquartile range (IQR) of the lower quartile, and the highest datum still within the 1.5 IQR of the upper quartile (Tukey boxplot).

264

265

Chapter 4: Biases in comparative analyses of extinction risk: mind the gap

266

4. Biasis in comparative analysis of extinction ris: mind the gap

González-Suárez, M; Lucas, P.M; Revilla, E. (2012) Biases in comparative analyses of extinction risk: mind the gap. Journal of Animal Ecology, 81: 1211–1222. DOI:10.1111/j.1365-2656.2012.01999.x

267

4. Biasis in comparative analysis of extinction ris: mind the gap Abstract

1. Comparative analyses are used to address the key question of what makes a species more prone to extinction by exploring the links between vulnerability and intrinsic species’ traits and/or extrinsic factors. This approach requires comprehensive species data but information is rarely available for all species of interest. As a result comparative analyses often rely on subsets of relatively few species which are assumed to be representative samples of the overall studied group. 2. Our study challenges this assumption and quantifies the taxonomic, spatial, and data type biases associated to the quantity of data available for 5415 mammalian species using the freely available life-history database PanTHERIA. 3. Moreover, we explore how existing biases influence results of comparative analyses of extinction risk by using subsets of data that attempt to correct for detected biases. In particular, we focus on links between four species’ traits commonly linked to vulnerability (distribution range area, adult body mass, population density and gestation length) and conduct univariate and multivariate analyses to understand how biases affect model predictions. 4. Our results show important biases in data availability with ~22% of mammals completely lacking data. Missing data, which appear to be not missing at random, occur frequently in all traits (14-99% of cases missing). Data availability is explained by intrinsic traits, with larger mammals occupying bigger range areas being the best studied. Importantly, we find that existing biases affect the results of comparative analyses by overestimating the risk of extinction and changing which traits are identified as important predictors. 5. Our results raise concerns over our ability to draw general conclusions regarding what makes a species more prone to extinction. Missing data represent a prevalent problem in comparative analyses and unfortunately, because data are not missing at random conventional approaches to fill data gaps are not valid or present important challenges. These results show the importance of making appropriate inferences from comparative analyses by focusing on the subset of species for which data are available. Ultimately, addressing the data bias problem requires greater investment in data collection and dissemination, as well as the development of methodological approaches to effectively correct existing biases.

268

4. Biasis in comparative analysis of extinction ris: mind the gap

Keywords: Data Imputation, Extinction Risk, Life-history Traits, Phylogenetic Generalised Linear Models, PhyloPars.

Introduction

An important priority for conservation biology is to understand what makes a species or population more likely to become extinct. A popular and appealing answer is based on comparative analyses that explore the links between species vulnerability to extinction and intrinsic ecological and life-history species’ traits (Cardillo et al. 2008, Fisher and Owens 2004, Fritz et al. 2009, Pinsky et al. 2011, Purvis et al. 2000) or extrinsic factors (Cardillo et al. 2004, Forester and Machlis 1996, Kerr and Currie 1995). This approach requires large databases describing species traits (or extrinsic factors) in a format suitable for comparative analyses. Compiling such databases takes considerable effort from multiple dedicated researchers, who ideally make their complete databases publicly available allowing future research (Jones et al. 2009). However, any efforts to gather information are limited by the fact that data are not available for all species in all locations, something that has been previously recognized by other authors (Fisher et al. 2003, Luck 2007, Matthews et al. 2011, Nakagawa and Freckleton 2008). As a result, gathered data represent only a subset of species and locations which traditional comparative analyses implicitly assume are a representative sample of the taxon or group of interest (but see Matthews et al. 2011). Our study challenges this assumption testing the hypothesis that studied species, those for which data are available, are not a random sample of the global biodiversity and that this bias affects results from comparative studies. In particular, we address three objectives: 1) to describe existing biases associated with the number and type of data available in a mammalian comparative dataset; 2) to test the hypothesis that life-history, ecological and behavioural traits are associated with greater data availability because some traits can facilitate, or complicate, research and make species more or less appealing as study subjects (Matthews et al. 2011); and 3) to investigate if existing biases affect the results and conclusions of comparative analyses linking intrinsic species’ traits and vulnerability to extinction. Specifically, we compare results from standard

269

4. Biasis in comparative analysis of extinction ris: mind the gap phylogenetically-informed comparative analyses based on different subsets of species, some of which attempt to control biases. Currently available tools for comparative analyses can be broadly classified into phylogenetic and non-phylogenetic regressions (Bielby et al. 2010). Phylogenetic methods are more commonly used, and include regressions using phylogenetic independent contrasts (Felsenstein 1985), a popular approach despite its unrealistic assumptions about Brownian trait evolution (Blomberg et al. 2003), and generalised regressions, such as phylogenetic generalised least square models (PGLSs, Martins and Hansen 1997), which provide a flexible alternative with fewer assumptions. Non-phylogenetic methods include regression trees (Breiman 1984) which have fewer data requirements but can be unstable and fail to account for phylogenetic relationships (Bielby et al. 2010). All of these tools are limited by data availability because they generally require complete data for all predictors. Therefore, exploring patterns with multiple predictors requires either interpolating missing data, which may introduce biases if data are not missed at random (Little and Rubin 2002), or eliminating all species with any missing data, which can bias estimates and reduces the sample size considerably (Nakagawa and Freckleton 2008). For example, a well-cited study by Cardillo et al. (2006) drew inferences from <20% of the extant species in some analyses, while analyses for this study were in some cases limited to <12% of the species of interest. If those species with available data are not a random sample, conclusions may not apply to the broad group of interest and inferences need to be made carefully. In recent years many authors have contributed to develop large databases suitable for comparative analyses which describe life history traits in diverse taxa including birds, mammals, amphibians, fish, and angiosperms (Bielby et al. 2008, Froese and Pauly 2000, Jones et al. 2009, Sekercioglu et al. 2004, Sodhi et al. 2008). For this study we decided to focus on mammals for several reasons. First, mammals are arguably the best studied group with many species of conservation, economic, and social interest. Second, the links between species traits and vulnerability have been extensively investigated in mammals with multiple comparative studies published showing how traits such as adult body mass, distribution range area, gestation length, and population density are linked to vulnerability to extinction (Brashares 2003, Cardillo 2003, Cardillo et al. 2008, Cardillo et al. 2006, Cardillo et al. 2005, Cardillo et al. 2004, Davidson et al. 2009, Fagan et al. 2001, Fritz et al. 2009, Purvis et al. 2000). Finally, we had free online access (http://www.utheria.org/) to a large mammalian life- 270

4. Biasis in comparative analysis of extinction ris: mind the gap history dataset, PanTHERIA (Jones et al. 2009), which was also used in several recent comparative studies (e.g., Bininda-Emonds et al. 2007, Cardillo et al. 2008, Davies et al. 2008, Fritz et al. 2009). In this study we show that species’ ecology, life history and morphology explain variation in the quantity of data collected. Data appear to be not missing at random, and thus applying imputations techniques to fill data gaps may be difficult. Moreover, existing biases affect estimates obtained from comparative analyses suggesting the predictive ability of currently used models may be limited. Although our results are limited to mammalian species, the existence of data biases that can affect comparative analyses is likely common to other taxa. Overall, these findings highlight the importance of explicitly considering data biases in comparative analyses and ultimately, the need for gathering and publishing basic natural history data even if currently deemed “old-fashion”.

Materials and Methods

Database PanTHERIA is comprised of two files: the median dataset and the raw data file. The median dataset includes an entry for each of the 5415 mammalian species recognized by the Wilson and Reeder’s (2005) taxonomy with calculated median values for 30 variables describing morphology, development, reproduction, ecology and spatial data (Jones et al. 2009). These median values were calculated from a varying number of estimates gathered from the literature (Jones et al. 2009). The raw data file includes these individual literature estimates which we used to estimate the total number and the type of data available for each mammalian species. Because we expect entries in the raw data file to represent a reasonably random sample of the literature we equate more data entries with more available published data. Certainly this relationship is likely not exact because some published sources are easier to access than others and, as Jones et al. (2009) discuss, the database may include some duplicate entries. However, we assume any bias associated with finding data in the literature and duplicate entries is minor compared with the bias in data collection and publication. In the raw data files species names were tracked onto the Wilson & Reeder mammalian taxonomy (2005) based on the synonyms file provided by Jones et al. (2009). The final file

271

4. Biasis in comparative analysis of extinction ris: mind the gap includes all 5415 extant mammalian species but 1211 species have zero data entries (i.e., no literature records were available for those species).

Are there biases in data availability? Taxonomic and phylogenetic bias Phylogenies capture the evolutionary history of a group of species better than taxonomy, but generally there is also more uncertainty associated with phylogenies due to unresolved or inconsistent relationships among taxa. For that reason, we explored the potential for biases, i.e., related species having greater similarity in the total number of data entries than expected by chance, using both taxonomy, as defined by Wilson & Reeder (2005), and phylogeny, based on the best date estimates of the mammalian supertree (Fritz et al. 2009). Because tip branches in the phylogeny were not fully resolved, we generated 10 trees with randomly resolved polytomies using the procedure multi2di (package ape in R). Parameter estimates were identical for all trees indicating that how polytomies were resolved did not influence results. We used non-parametric Kruskal-Wallis tests to compare data counts among orders, families and genera. We addressed the question of phylogenetic bias using the parameter λ (Pagel 1999a), which characterizes the phylogenetic correlation in the total number of data entries available per species (log-transformed). We used the procedures corPagel (package ape in R) and gls (nlme package in R) to define an intercept-only model with data availability as the dependent variable. Following Freckleton et al. (2002) we compared log likelihood estimates to determine if the estimate of λ was significantly different from 0 (0 indicates no phylogenetic correlation in the data).

Other biases To assess biases in the type of data we grouped the original 25 variables listed by the raw data file into five data groups: Ecology, Morphology, Development, Reproduction, and Spatial (Table A4.1 in Supporting Information) and compared the average number of entries per variable per species among groups. We explored biases in data availability related to threat category using the 2008 Red List classification (International Union for Conservation of Nature 2010b). Threat classification was available for 5288 species in our database, including 731 listed as Data Deficient. Finally, to explore spatial biases we obtained global 272

4. Biasis in comparative analysis of extinction ris: mind the gap distribution maps of terrestrial mammals from the IUCN spatial database (International Union for Conservation of Nature 2010b). We used data from the 4847 species recognized by Wilson & Reeder (2005) and with range areas defined as presence “extant” or “probably extant”. Maps were projected in the cylindrical equal area projection and onto a grid equivalent to 2º x 2º near the equator (Hurlbert and Jetz 2007). For each grid cell we calculated: species richness, as the total number of distinct species’ ranges overlapping any area of the cell; data richness, as the mean number of data entries per species occupying the cell; and the coefficient of variation in the number of data entries among all species occupying the cell. Data richness reflects the average data availability expected for any species occupying a cell, whereas the coefficient of variation indicates the difference in data availability among species within the same cell.

Do species traits explain the bias in data availability? To explore if intrinsic species traits could explain data availability, we used 28 variables describing life-history, behavioural and ecological traits and an estimate of distribution range area provided in the median dataset (Jones et al. 2009). Analyzing these data presented us with a series of challenges. First, due to abundant missing data in the 29 variables considered, using AIC model selection approaches was not possible as models based on different datasets are not comparable (Burnham and Anderson 2002). Therefore, we initially defined univariate models to explore relationships between data availability and species’ traits using all available data for each trait. We then defined a multivariate model based on a reduced dataset which included only data-rich (<950 missing cases. Table 4.1) and non-highly correlated variables to reduce the effects of collinearity (r<0.80, Variance Inflation Factor, VIF<5. Table A4.2). The reduced dataset includes 10 quantitative and three categorical variables that describe intrinsic species traits but represents only a limited number of species (as many data are missing). Thus, multivariate results may not reflect patterns common to all mammals. A second challenge was whether, and how, to incorporate non-independence of species data due to evolutionary relationships (McNab 2003, Purvis 2008). We followed three different approaches using PGLSs, taxonomically-corrected generalised linear mixed models (GLMMs), and non-corrected regression trees. Phylogenetic correction is generally preferable to taxonomic correction (if a good phylogeny exists); however our estimate of data availability best fits a negative binomial distribution which, to our knowledge, cannot be 273

4. Biasis in comparative analysis of extinction ris: mind the gap

Table 4.1. Coefficient estimates for univariate GLMMs and PGLSs describing the number of data entries available in the mammalian database PanTHERIA as a function of intrinsic species traits. N indicates number (and percentage) of species with available data for each trait from the total 5415 mammals studied. Variable N (%) Coefficient (SE) GLMMs PGLSs Activity cycle 1657 (30.6) Nocturnal 732 · · Crepuscular, cathemeral 486 0.57 (0.067)** 0.45 (0.071)** Diurnal 439 0.32 (0.081)** 0.13 (0.094) Terrestriality 2634 (48.6) Fossorial 1144 · · Above ground 1490 -0.29 (0.062)** 0.06 (0.075) Trophic level 2159 (39.9) Herbivore 781 · · Omnivore 739 0.20 (0.058)** 0.23 (0.061)** Carnivore 639 -0.29 (0.069)** -0.26 (0.088)* Neonate body mass‡ 1083 (20.0) 0.09 (0.027)** 0.13 (0.031)** Weanling body mass ‡ 487 (9.0) 0.13 (0.012)** 0.09 (0.045)† Adult body mass ‡ 3539 (65.4) 0.30 (0.020)** 0.27 (0.036)** Neonate head-body length‡ 226 (4.2) 0.24 (0.090)* 0.31 (0.123)* Weanling head-body length 47 (0.9) 0.40 (0.222) 0.28 (0.264) ‡ Adult head-body length ‡ 1939 (35.8) 0.64 (0.065)** 0.70 (0.123)** Adult forearm length‡ 903 (16.7) 1.80 (0.283)** 1.29 (0.300)** Teat number‡ 639 (11.8) 0.22 (0.224) 0.57 (0.243)* Age at eye opening‡ 474 (8.8) -0.01 (0.028) -0.22 (0.139) Weaning age‡ 1161 (21.4) 0.21 (0.057)** -0.06 (0.084) Sexual maturity age‡ 1049 (19.4) 0.22(0.050)** -0.06 (0.082) Age at first birth‡ 445 (8.2) 0.22 (0.073)* 0.15 (0.114) Dispersal age‡ 143 (2.6) 0.09 (0.116) 0.06 (0.138) Maximum longevity‡ 1011 (18.7) 0.42 (0.068)** 0.81 (0.102)** Gestation length‡ 1359 (25.1) 0.19 (0.058)* 0.07 (0.115) Interbirth interval‡ 695 (12.8) 0.18 (0.066)* 0.07 (0.111) Litter size‡ 2498 (46.1) 0.08 (0.091) 0.93 (0.122)** Litters per year‡ 893 (16.5) 0.10 (0.122) 0.40 (0.154)* Diet breadth 2159 (39.9) 0.16 (0.015)** 0.14 (0.014)** Habitat breadth 2722 (50.3) 0.30 (0.048)** 0.30 (0.045)** Group size‡ 388 (7.2) 0.12 (0.044)* 0.19 (0.050)** Social group size‡ 705 (13.0) 0.42 (0.072)** 0.46 (0.089)** Population density‡ 954 (17.6) -0.09 (0.021)** 0.02 (0.034) Home range‡ 705 (13.0) 0.07 (0.018)** 0.11 (0.029)** Home range individual‡ 624 (11.5) 0.07 (0.019)** 0.10 (0.030)** Distribution range area‡ 4664 (86.1) 0.64 (0.016)** 0.22 (0.006)**

†P<0.10, *P<0.05, **P<0.001, ‡ Log10-transformed

274

4. Biasis in comparative analysis of extinction ris: mind the gap modelled using frequentist phylogenetic models. Thus, we defined PGLSs log-transforming our dependent variable (using procedures corPagel and gls in R, Development Core Team 2011). However, analyses based on transformed dependent variables are problematic (O’Hara and Kotze 2010), thus we also fitted GLMMs including taxonomic random effects (nested effects of order, family and genus) with a negative binomial distribution (in SAS 9.2). We fitted univariate and multivariate PGLSs and GLMMs. Finally, using the complete dataset we built a regression tree with the procedure rpart (package rpart in R), log-transforming the number of data available to meet the normality requirement. Missing data were handled with surrogate splits as described by Breiman (1984). The tree was pruned using 40 sets of 10-fold cross-validations to produce an optimal tree based on the modal number of splits corresponding to the 1 SE rule (Breiman 1984, De'ath and Fabricius 2000). By comparing the results from all three alternative, imperfect methods we aimed to assess the overall agreement and ideally identify some general (non-method dependent) patterns to explain data availability.

How do data biases affect comparative analyses? To understand how data limitations affect our understanding of the links between intrinsic species traits and vulnerability to extinction, we explored how results from a standard comparative approach, PGLSs, differed among distinct datasets. For simplicity, we selected a priori four species’ traits that have been consistently linked to vulnerability to extinction in mammals: adult body mass, distribution range area, gestation length and population density (see Introduction for a reference list). We defined a group including all species with estimates available for all four traits in the median dataset of PanTHERIA (N=636). This group represents the total sample available for multivariate regression analyses based on the four traits. We compared this group with two other datasets: the PanTHERIA dataset which includes all data available for each trait (see Table 4.1 for number of species per trait), and an imputed dataset with data on all four traits for 5016 mammalian species. The imputed dataset was populated using the phylogenetic data imputation technique implemented in the program PhyloPars (Bruggeman et al. 2009). We used available data from PanTHERIA assuming allometric relationship among traits and the phylogeny supertree mentioned above (Fritz et al. 2009). The supertree describes phylogenetic relationships for 5016 species listed by Wilson and Reeder (2005) and thus data for the remaining 399 species could not be imputed using 275

4. Biasis in comparative analysis of extinction ris: mind the gap this approach. When running PhyloPars we assumed no phenotypic variation, i.e., no measurement error, in order to avoid re-estimation of already available data. Leave-one-out cross-validation analyses were used to estimate bias (mean differences between observed and estimated values) and absolute error (mean of the absolute differences between observed and estimated) for each trait. Both bias (-0.005 to 0.002) and absolute errors (0.06-0.88) were generally low. To compare datasets we first plotted the distribution of values in each of the four traits for the subset group, the PanTHERIA dataset, and the imputed dataset. Second, using PGLSs we estimated the relationship between the four traits and vulnerability to extinction as defined by the IUCN Red List category (International Union for Conservation of Nature 2010b). We used the same phylogenetic supertree with randomly resolved polytomies. For each trait we defined univariate PGLSs for the sample of 622 species with data available for all four traits, with a Red List category (not including Data Deficient) and represented in the phylogenetic tree (hence forth the “multivariate subset”). To explore the range of expected values, univariate PGLSs were also defined for 500 samples of 622 species each which were selected at random from the imputed dataset. This analysis was repeated drawing random samples from the PanTHERIA dataset. Following Purvis et al. (2000) the Red List categories were converted to a continuous index: Least Concern=0, Near Threatened=1, Vulnerable=2, Endangered=3, Critically Endangered=4, and Extinct in the Wild/Extinct=5. For analyses that included data on range area we removed species listed under IUCN Red List criteria B (small geographic range or area of occupancy) to avoid circularity. This resulted in a subset of 584 species which were compared to 500 random samples of 584 species each. We also defined multivariate PGLSs including all four traits for different subgroups of the multivariate subset (drawing 300 random replicates per subgroup). Each subgroup was defined to conform to the distribution of values observed for given trait in the imputed dataset, while trying to maximize the number of species per subgroup (the number was limited when the distributions were very dissimilar). Defining a subgroup that conformed to all four trait distributions at once was not possible; thus, we defined four subgroups each conforming to a single trait and ran four separate multivariate PGLSs. We compared parameter estimates of these subgroups with those obtained from a multivariate PGLS based on the entire multivariate subset. Sampling to conform to the distribution of a given trait often altered the distribution of the other traits and in some cases increased the deviation from the 276

4. Biasis in comparative analysis of extinction ris: mind the gap general distribution observed with all data. This analysis was repeated conforming to the distribution of values in the PanTHERIA dataset (non-imputed data).

Results

Are there biases in data availability? We found no data for 1211 (22.4%) of the 5415 extant mammalian species recognized by Wilson (2005), whereas 438 species have a single entry in the raw data file, and 401 have two entries. On the other hand, the highest number of data entries is 443 for the deer mouse (Peromyscus maniculatus, Rodentia). Missing data are prevalent, occurring in all studied traits, often at high frequencies (14-99% of missing cases, Table 1), and with no species having complete data for all 29 studied traits (all species are missing information for at least one trait).

Taxonomic and phylogenetic bias The number of data entries per species differs among orders (Kruskal-Wallis test, χ2 = 880.1, df = 28, P < 0.0001. Fig. A4.1), families (χ2 = 1351.1, df = 152, P < 0.0001), and genera (χ2 = 2463.0, df = 1229, P < 0.0001). Species without data belong primarily to the orders Rodentia and Soricomorpha (no data for 29% and 46% of their species respectively). On the other hand, among carnivores and ungulates <7% of species have no data. Analyses based on phylogenetic relationships also indicate similar data availability among closely related species. The estimate of the parameter λ (0.510), which characterizes the degree of phylogenetic correlation, is significantly different from 0 (χ2= 1379.8, df=5015, P<0.001).

Other biases More information is available for some types of data than others (Kruskal-Wallis test, χ2 = 3851.6, df = 4, P < 0.0001. Fig. A4.2). Generally, morphological data are the most abundant (1.86 ± 2.34, mean ± SD number of entries per species), followed by reproduction (1.27 ± 2.44), ecology (0.94 ± 1.61), development (0.66 ± 1.40) and spatial data (0.33 ± 1.19). However, there is some variation among taxa in the relative abundance of each data type (e.g., spatial data are not the least abundant for the Erinaceomorpha; Fig. A4.2). We also find

277

4. Biasis in comparative analysis of extinction ris: mind the gap differences in data availability among the different threat categories defined by the IUCN Red List (Kruskal-Wallis test, χ2 > 502.5, df = 3, P < 0.0001. Fig. A4.3). In particular, the number of entries per species is higher for non-threatened species (classified as Least Concern or Non Threatened, N=3420 species) than for threatened species (classified as Critically Endangered, Endangered, or Vulnerable, N=1070). In fact, there is no data available for 29% of the mammalian species classified as Endangered or Critically Endangered; whereas only 16% of the Least Concern and 14% of the Near Threatened species lack data (23% of Vulnerable species lack data). Not surprisingly, there are fewer data for species classified as Data Deficient (N= 731) or Extinct (N=67), with many species in these categories having no data at all (39% and 48% respectively). Finally, there is evidence of spatial biases in data availability with species living in the northern hemisphere, particularly in North America and Europe, being considerably better studied than those in tropical or Southern regions (Fig. 4.1B). This contrasts with the pattern of species richness (Fig. 4.1A), so that, on average, there are fewer studies per species in those areas with the highest diversity of mammals. Moreover, less-studied areas often have high variation in data availability among species (Fig. 4.1C) and include relatively high percentages (20-60%) of species with <3 data entries (very poorly studied species).

Do species traits explain the bias in data availability? Univariate analyses identify many traits as significant in explaining data availability (Table 4.1). Results from GLMMs and PGLSs are generally similar, although some traits are identified as significant under one approach but not the other (Table 4.1). Nevertheless, there are not contradictory results, i.e., no traits identified as significant have opposite estimated effects. Results from the multivariate models, based on data from 266 species, are very similar using taxonomic and phylogenetic correction, although significance is marginal for some traits using phylogenetic correction (Table 4.2). Both approaches suggest that the number of data entries is generally greater for diurnal or crepuscular mammals with larger body mass, bigger litter sizes, earlier sexual maturation age, and longer life spans. Species with more data also have a wider distribution range area and live above ground at higher population densities. Regression coefficients are generally similar in univariate and multivariate analyses, although as expected some variables are significant in the univariate analyses but not in the 278

4. Biasis in comparative analysis of extinction ris: mind the gap multivariate models (Table 4.1). In addition, there is a qualitative change in the estimated effect based on GLMMs for three traits (terrestriality, sexual maturity age and population density). Noticeably, these traits are not significant in the univariate PGLSs. In order to test if the changes are simply due to changes in sample size and composition (i.e., which species are

Fig. 4.1. Global distribution of terrestrial mammalian species and data availability. Panels: A) species richness calculated as the total number of terrestrial mammals per cell; B) data richness calculated as the mean number of data entries per terrestrial mammal per cell; C) coefficient of variation in the number of data entries per terrestrial mammal per cell. Species range distribution maps were obtained from the IUCN spatial database. Grid size is equivalent to 2º x 2º near the equator. represented), we re-ran univariate GLMMs analyses using only the 266 species included in the multivariate models. Using these data, all three coefficients are not significantly different from zero suggesting there is an effect of sample size and composition, but also a possible interaction among variables in the multivariate model that lead to significant coefficients. The non-corrected approach based on regression trees gives similar results, although fewer traits 279

4. Biasis in comparative analysis of extinction ris: mind the gap are associated with data availability. The final tree reveals that the highest mean number of data entries is associated with species having larger distribution areas and greater body mass (Fig. 4.2). On the other hand, species with smaller distribution area, small body mass, and reduced habitat breadth have the fewest data entries. The tree explains 34.9% of the variance

Table 4.2. Coefficient estimates for multivariate GLMMs and PGLSs describing the number of data entries available in the mammalian database PanTHERIA as a function of intrinsic species traits. Models are based on 266 species for which data were available (from the total 5415 mammals studied). Variable Coefficient (SE) GLMMs PGLSs Activity cycle Nocturnal · · Crepuscular, cathemeral 0.21 (0.072)* 0.13 (0.075)† Diurnal 0.24 (0.084)* 0.18 (0.094)† Terrestriality Fossorial · · Above ground 0.22 (0.088)* 0.17 (0.094)† Trophic level Herbivore · · Omnivore 0.05 (0.094) 0.06 (0.095) Carnivore -0.08 (0.132) 0.05 (0.131) Adult body mass ‡ 0.18 (0.061)* 0.19 (0.069)* Weaning age‡ 0.01 (0.126) -0.07 (0.145) Sexual maturity age‡ -0.33 (0.146)* -0.29 (0.154)† Maximum longevity‡ 0.50 (0.19)* 0.66 (0.190)** Gestation length‡ -0.05 (0.130) -0.14 (0.189) Litter size‡ 0.63 (0.176)** 0.90 (0.195)** Diet breadth -0.03 (0.025) -0.02 (0.023) Habitat breadth -0.01 (0.053) 0.03 (0.056) Population density‡ 0.16 (0.040)** 0.11 (0.042)* Distribution range area‡ 0.16 (0.039)** 0.19 (0.042)**

†P<0.10, *P<0.05, **P<0.001, ‡ Log10-transformed in the number of entries (calculated as 1-relative error). Overall, based on the three different approaches, two variables appear as clearly linked with data availability: adult body mass and distribution range area. Other traits such as activity pattern, terrestriality, sexual maturity age, maximum longevity, litter size, population density, and habitat breadth are also likely relevant predictors.

280

4. Biasis in comparative analysis of extinction ris: mind the gap

Although these analyses had to be limited to species with trait data, the main observed relationships appear to extend to those species missing data. For example, there is a strong and positive relationship between the median adult body mass in each taxonomic family (calculated from available species data) and the proportion of species with body mass data in that family (Spearman correlation r=0.39, p<0.001, N=153 families; for families with 10 or more species r=0.45, p<0.001, N=69 families). In other words, families with the

Fig. 4.2. Regression tree showing the number of data entries available in the mammalian database PanTHERIA based on diverse intrinsic species traits. On each node the threshold value and name of the splitting trait are indicated. Data on the leaves (represented by circles) provide the average number of data entries and the number of species in the group. smaller, on average, species such as rodents or shrews have fewer species with body mass estimates available. Therefore, missing data appear to not be missing at random, but rather the likelihood of having information on a given trait is affected by the trait value itself (smaller species are less likely to have body mass estimates). The relationship is not as clear for distribution range area (Spearman correlation r=0.15, p=0.06, N=153 families), likely because range area is available for many more species (considering only families with species missing data, r=0.32, p=0.02, N=79 families) and because range is a more flexible “trait” (less constrained by evolution). In fact, estimates of range area are more variable among species within a family (median CV within families=0.17) than estimates of adult body mass (0.10).

281

4. Biasis in comparative analysis of extinction ris: mind the gap

How do data biases affect comparative analyses? From the 5415 extant mammals, adult body mass data are only available for 65.4% of the species, range area for 86.1%, population density for 17.6%, and gestation length for 25.1%. However, the subset of species with data on all four traits is much smaller, including only 636 species (11.7%). From this subgroup, 622 species (the multivariate subset) also have Red List status and defined phylogenetic relationships; thus, the subset of species available for multivariate PGLSs is quite small compared to the overall mammalian biodiversity. In addition, species with data on all four traits are not a representative sample of all mammals (Fig. 3). These species represent 22 of the 29 mammalian orders (notably excluding all 84 species from the order Cetacea) and 91 of the 153 families. The multivariate subset includes <7% of the species from the most populous families (Muridae, n=730, and Cricetidae n=681) but >51% of the 35 canids (Canidae). In general, the subset includes mammals with higher body mass (following a bimodal distribution), larger range areas, lower population densities and longer gestation periods (Fig. 4.3).

Fig. 4.3. Distribution of values for four traits consistently linked to vulnerability to extinction in mammals for all species (imputed dataset, black solid line), for the PanTHERIA dataset (grey solid line), and for the 636 species with data on all four traits (dotted line). Sample sizes for the

282

4. Biasis in comparative analysis of extinction ris: mind the gap

PanTHERIA dataset are: adult body mass N=3539, distribution range area N= 4664, population density N=954, gestation length N=1359. Imputed data is available for 5016 species.

Fig. 4.4. Values of the intercept and slope coefficients estimated in univariate PGLSs explaining Red List status as a function of A) adult body mass, B) distribution range area, C) population density, and D) gestation length. The grey surfaces show the distribution of parameter estimates obtained for all mammals (imputed dataset) calculated using 500 subsets of 622 species each drawn at random (for range area subsets had 584 species after excluding those listed under the IUCN small range criteria). The black arrows indicate the parameter estimates obtained for the multivariate subset (species with Red List status and data available for all four traits. See Fig. 4.3). For illustration purposes the arrow points are placed along the z-axis at the point of intersection with the grey surface. Note the values in the intercept-axis in panel A are reversed.

Univariate PGLSs associating Red List status with each of the four traits show that the parameter estimates obtained for the multivariate subset are generally not representative of the relationships expected for all mammals (Fig. 4.4) and they would be rarely, if ever, observed when using representative (random) samples. Intercept values estimated using the multivariate subset are higher than those based on all data. Therefore, analyses based on the 283

4. Biasis in comparative analysis of extinction ris: mind the gap subset appear to overestimate the baseline extinction risk (measured by Red List status). In addition, although there are no changes in slope sign (the relationships between each trait and Red List status are qualitatively the same), the slope values, which estimate the strength of the relationship, vary. In particular, the increase in Red List status predicted in response to a reduction in range area or population density is more pronounced for the multivariate subset than when considering all data, suggesting that the subset overestimates the influence of these traits on the extinction risk. On the other hand, the multivariate subset appears to underestimate the rate of increase in Red List status associated with longer gestation periods. The estimates of the relationship between body mass and Red List status are similar for all data and the multivariate subset. Results are qualitatively the same when the multivariate subset results are compared with random samples from the PanTHERIA dataset (non- imputed data. Fig. A4.4).

Fig. 5. Multivariate PGLSs parameter estimates for subgroups of mammals with Red List status and data available for the four traits listed but selected to conform to the distribution of all mammalian data for each trait (imputed dataset. See Fig. 4.3). Symbols are the mean estimate with error bars representing ±2 SD from 300 independent random replicates (N in the legend indicates the number of species in each replicate. The sample size was determined to maximize the number of species while conforming to the distribution of all data). Error bars overlapping with zero indicate the relationship between the trait and the Red List status is ambiguous. Horizontal black lines represent the mean parameter estimates of a single PGLS based on the entire multivariate subset (622 species with data on all four traits). Grey bands represent ±1SE. Grey bands overlapping with zero indicate a trait is not linked to vulnerability to extinction.

284

4. Biasis in comparative analysis of extinction ris: mind the gap

Parameter estimates from multivariate PGLSs defined for subgroups conforming to all mammal distributions differ from those calculated for the multivariate subset (Fig. 4.5). In particular, conforming to the distributions of all available data for any trait affects the expected relationship between body mass and Red List status. This relationship is significantly positive for the multivariate subset but not for the conforming subgroups, indicating body mass may not be a good predictor of extinction risk in mammals after all. Similarly, conforming to the distribution of body mass or population density generally reduces the estimated effect of range area on vulnerability to extinction, so that a reduction in range area is not predicted to increase Red List status as rapidly. On the other hand, when conforming to the distribution of range area, the rate at which Red List status increases with a reduction in population density is greater. Finally, although gestation length has been identified as important by several previous studies and our own univariate analyses, the coefficient estimate is not significantly different from zero in any of our multivariate models, suggesting gestation length may not be strongly associated to Red List status when other factors are taken into account. For other traits, the univariate and multivariate coefficients predict the same general relationships. We found no evidence of collinearity in the multivariate models (VIF <1.55).

Discussion

Are there biases in data availability? Our results show an important bias in data availability for life-history, ecological and behavioural traits in mammalian species, arguably the best-studied group of organisms. We found that some species are better studied than others and that biases have taxonomic and phylogenetic signals, so that related species have similar data availability (Amori and Gippoliti 2000). As a result, some groups tend to be data-poor, e.g., Rodentia, while others are generally well-studied, e.g., Artiodactyla (Fig. A4.1). Surprisingly, we found that non- threatened species are better studied than mammals of conservation concern (Fig. A4.3), with data completely lacking for nearly one third of the most threatened species (Endangered or Critically Endangered). Although conservation biologists have been at work for some time,

285

4. Biasis in comparative analysis of extinction ris: mind the gap we still know more about common species, possibly those of direct economic importance such as pests or game, than species at risk of extinction. In addition to biases in the amount of data available we also identify biases in the type of data gathered. For example, morphological data, e.g., adult body mass, are more often available than spatial data, e.g., home range size (Fig. A4.2). These biases are likely due to technological constraints, which have limited our ability to track small species such as bats (Holland and Wikelski 2009), and/or financial or logistic limitations associated with obtaining different types of data. Importantly, scarcity of data for some traits likely affects the results of comparative studies because a lack of power in the analyses may limit our ability to recognize traits as important. In fact, we find that traits linked most often to vulnerability to extinction are also those with data for more species and which appear to explain data availability (i.e., body mass and range area). We also identify important biases in the spatial distribution of data availability (Fig. 1, Amori and Gippoliti 2000). Regions of higher species richness, where mean data availability per species is lowest, largely correspond to tropical areas where the highest abundance of threatened species also occurs (Schipper et al. 2008). In these regions there is also a greater disparity in data availability among species, so that the limited number of studies conducted concentrate on a few of the species present, likely those easier to study or more attractive, while many species remain poorly known. In contrast, areas with higher data availability and where species are more uniformly studied (lower variation in data availability among them) are predominantly in developed countries where fewer endangered species are found (Schipper et al. 2008) but more resources are invested in research leading to more data collection and publication (World Bank 2011).

Do species traits explain the bias in data availability? Our analyses show that existing biases in data availability are in part explained by intrinsic species traits, which presumably influence the ease and attractiveness of a species as a study organism (Matthews et al. 2011). Interestingly, traits associated with higher data availability (Tables 4.1, 4.2 and Fig. 4.2) do not appear to define a single group of species but may represent two general types of well-studied mammals: the big, long-lived mammals occupying large range areas and the smaller mammals with an early maturation age and large litter sizes. The former, for which there is overall the largest amount of data, probably 286

4. Biasis in comparative analysis of extinction ris: mind the gap correspond to charismatic megafauna, such as the giraffe (Giraffa camelopardalis with 275 data entries), whereas the second group includes common species with a fast life-history more suitable for ecological experiments and manipulations such as the deer mouse (Peromyscus maniculatus with 443 entries). The bimodal distribution of body mass values in Fig. 3 also supports the existence of two types of best-studied mammals. Interestingly, a bimodal pattern in body mass has been also reported regarding vulnerability to extinction (Cardillo et al. 2005, Davidson et al. 2009). In fact, these studies describe a threshold around 3-5 kg of body mass, which coincides well with the ~3.5 kg split in our regression tree explaining data availability.

How do data biases affect comparative analyses? We find that the existence of biases in data availability has worrying consequences for comparative analyses. Multivariate analyses that consider several traits associated to vulnerability to extinction are likely limited to a skewed subset of species which are not a representative sample of all mammals. For example, we show that well-studied species appear to be larger, have bigger range areas, longer gestation periods and live at lower population densities than those less studied. Comparative analyses that partly correct these biases give different results than analyses based on the skewed subset, indicating that ignoring existing biases in the data available has consequences for our understanding of how species traits influence vulnerability to extinction. Our study does not imply that previous conclusions are necessarily mistaken or erroneous, but rather raises concern over our ability to accurately make broad inferences with the available data. For example, large body size is perhaps identified as a trait associated with higher risk of extinction because we have data primarily for the big and rare vs. the small and common.

Potential solutions and recommendations We have identified important data biases in the mammalian life-history literature which appear to reflect a pattern of data “not missing at random”, that is the probability of not having information for a trait depends on the unobserved values of that trait (Little and Rubin 2002). This presents a great challenge for analyzing these data because as we have seen here deleting species with missing data greatly reduces the available sample size and introduces biases in model estimates. However, conventional techniques to fill gaps (such as multiple 287

4. Biasis in comparative analysis of extinction ris: mind the gap imputation) generally assume data are missing at random or completely at random (Little and Rubin 2002, Nakagawa and Freckleton 2008). For data “not missing at random” it is possible to use imputation but a clear understanding of the mechanism causing the missing data is generally necessary. However missing data in PanTHERIA are missing likely as a result of multiple mechanisms. For example, some species may be harder to study due to their life- history, while others may simply live in areas where research is complicated by the topology or political situation. In addition, basic research in some areas may be published in non- English journals or in publication formats not as readily available to researchers compiling databases. Therefore, filling data gaps in PanTHERIA using conventional approaches may be challenging. Alternatively, missing data may be inferred based on expected relationships among traits and phylogenetic relationships (Bruggeman et al. 2009, Pagel 1999b). We applied this approach in this study, but the method is not without challenges. First, one must have a complete phylogeny, yet phylogenies are rarely complete. Even for well-studied species such as mammals we found ~9% of extant species are not reflected in the most updated phylogeny. Second, these methods assume relatively simple relationships among traits (e.g., allometric) and evolutionary models (i.e., Brownian evolution) which may not be realistic for many ecological and behavioural traits (Blomberg et al. 2003). Finally, interpolation based on a skewed sampled may generate biased datasets, so inferences should be made with caution. For example, exploratory analyses with PhyloPars (González-Suárez, M. unpublished data) suggest that estimates of body mass for species with missing mass in the order Carnivora (N=34) can differ up to 2 order of magnitude when imputation is done using only data for small carnivores (≤ 3 kg) vs. only data for large carnivores (> 3 kg). Imputed values are always larger when estimated from the large carnivore dataset. Interestingly, biasing the dataset by body mass (imputing data based on large vs. small carnivores) also changes the estimates of missing data for range size, population density and gestation length. The implications of these challenges for our own analyses are that we cannot accurately quantify biases because we cannot truly know mammalian diversity. However, our results are consistent using imputed data or only the available data in PanTHERIA, thus we feel there is strong evidence that biases exist and that we can show their general direction. In conclusion, our results highlight the need for gathering additional data because many species, even within well-known taxa, are poorly studied and imputing missing data is 288

4. Biasis in comparative analysis of extinction ris: mind the gap very challenging. In addition, obtained data must be made available to others (Costello 2009). Efforts such as PanTHERIA (Jones et al. 2009), which make published data readily available in a convenient format, are key to understanding general patterns because individual researchers are limited in their ability to gather the large amounts of data needed for broad comparative analyses. As we see it, reducing data biases requires both augmenting data collection and encouraging data dissemination. In addition, comparative analyses need to acknowledge and explicitly address the bias in data availability, making inferences that are appropriate to the available data (i.e., restricted to the subset of species used in the analyses). We suggest that comparative analyses incorporate approaches to explore the consequences of existing biases. For example, as done here, authors may use resampling techniques to incorporate uncertainty, or compare results from univariate and multivariate models as the former may include considerably more species. In addition, sensitivity analyses of imputed datasets based on diverse missingness-causing mechanisms may be conducted (Little and Rubin 2002). As we see it, the long-term solution is an increase in data availability but meanwhile comparative studies should acknowledge the limitations in the existing information by implementing approaches that account for data biases, and authors should be cautious with their conclusions. Analyses may also focus on the best-studied groups, such as ungulates and carnivores, for which more data are available and for which conclusions, even if not as general, may not be as biased. Finally, incorporating extrinsic factors associated with extinction risk is essential to fully understand why some species are more vulnerable than others. However, data on extrinsic factors influencing vulnerability are also likely biased and should be also explored with caution.

Acknowledgments

Our work would not have been possible without all the researchers that contributed to PanTHERIA by conducting and publishing mammalian research and by defining and populating the database. Kate Jones kindly provided us with access to the raw data file. Assistance from Jorn Bruggeman was essential to generate the imputed dataset with PhyloPars. We would also like to thank Tim Coulson, Miguel Clavero, Miguel Delibes, Peter Thrall and two anonymous reviewers for valuable suggestions to improve earlier versions of

289

4. Biasis in comparative analysis of extinction ris: mind the gap this manuscript. This work was funded by the European Community’s Seventh Framework Programme (FP7⁄2007-2013) under grant agreement nº 235897, the Spanish Ministry of Science and Innovation (CGL2009-07301⁄BOS and BES-2010-034151).

290

4. Biasis in comparative analysis of extinction ris: mind the gap Supporting material

Fig. A4.1. Central tendency (mean and median) and maximum number of data entries per species for each mammalian taxonomic order. Bars represent mean, horizontal lines median, and stars the maximum number of data entries. Numbers in parenthesis indicate the total number of extant species in each order.

291

4. Biasis in comparative analysis of extinction ris: mind the gap

Fig. A4.2. Availability of different types of data (mean and SD number of data entries per species) for mammalian orders with at least 10 extant species (excluding the 1211 species for which no data was available in any category).

Fig. A4.3. Number of data entries (mean and SD) for the different categories of threat defined by the IUCN Red List for 5288 mammalian species. Different letters indicate significant differences (P < 0.001) among groups based on pairwise Wilcoxon test with Bonferroni correction.

292

4. Biasis in comparative analysis of extinction ris: mind the gap

Fig. A4.4. Values of the intercept and slope estimated in univariate GLMMs explaining Red List status as a function of A) adult body mass, B) distribution range area, C) population density, and D) gestation length. The grey surfaces show the distribution of parameter estimates obtained from the PanTHERIA dataset calculated using 500 subsets of 622 species each drawn at random from all species with available data for each trait (for range area subsets had 584 species after excluding those listed under the IUCN small range criteria). The black arrows indicate the parameter estimates obtained for the multivariate subset (species with Red List status and data available for all four traits. See Fig. 3). For illustration purposes the arrow points are placed along the z-axis at the point of intersection with the grey surface.

293

4. Biasis in comparative analysis of extinction ris: mind the gap

Fig. A4.5. Multivariate PGLSs parameter estimates for subgroups of mammals with Red List status and data available for the four traits listed but selected to conform to the distribution of all available data in PanTHERIA for each trait (see Fig. 3). Symbols are the mean estimate with error bars representing ± 2 SD from 300 independent random replicates (N in the legend indicates the number of species in each replicate. The sample size was determined to maximize the number of species while conforming to the distribution of all available data). Error bars overlapping with zero indicate the relationship between the trait and the Red List status is ambiguous. Horizontal black lines represent the mean parameter estimates of a single PGLS based on the entire multivariate subset (species with data on all four traits). Grey bands represent ± 1SE. Grey bands overlapping with zero indicate a trait is not linked to vulnerability to extinction.

294

4. Biasis in comparative analysis of extinction ris: mind the gap

Table A4.1. Grouping of the original PanTHERIA variables as listed in the raw data file into five groups for our analyses. Three original variables were not included in any group: growth data and mortality data (logical variables listed as yes or no); and metabolic rate.

Data group Variables as listed in PanTHERIA raw data file Development Age at Eye Opening Average Lifespan Weaning Age Maximum Longevity Sexual Maturity Age Reproduction Age at First Birth Gestation Length Interbirth Interval Litter Size Litters Per Year Ecology Activity Cycle Diet Habitat Layer Group Composition & Size Population Density Morphology Adult Limb Length Body Mass Head-Body Length Teat Number Spatial Dispersal Age Migratory Behaviour Ranging Behaviour

295

4. Biasis in comparative analysis of extinction ris: mind the gap

Table A4.2. Correlations among species median estimates of life-history, ecological and behavioural traits obtained from PanTHERIA. All variables were log10-transformed except diet breadth and habitat breadth. Very high correlations (r>0.8) are indicated in bold. Variance Inflation Factors (VIF) describe collinearity among traits for the sample used in the multivariate models (N=266. Table 2). VIF were calculated with the corvif procedure (library AED in R, Development Core Team 2011, R Development Core Team 2013). As a general rule of thumb, collinearity is considered to be a problem with VIF>10 (Belsley 2004).

Adult Neonatal Sexual Range Adult Gestation Max. Population Weaning Diet body Litter size body maturity VIF area length length longevity density age breadth mass mass age Range area 1.00 1.44

Adult body mass 0.02 1.00 4.08

Adult length 0.17 0.99 1.00 —

Gestation length 0.01 0.68 0.72 1.00 3.19

Litter size 0.05 -0.31 -0.43 -0.74 1.00 2.91

Max. longevity 0.06 0.72 0.71 0.73 -0.64 1.00 4.80

Neonatal body — 0.07 0.89 0.91 0.83 -0.50 0.71 1.00 mass Population 2.64 -0.19 -0.76 -0.77 -0.60 0.46 -0.64 -0.67 1.00 density Sexual maturity 3.96 -0.07 0.71 0.71 0.72 -0.58 0.78 0.66 -0.59 1.00 age Weaning age -0.10 0.62 0.60 0.56 -0.54 0.59 0.50 -0.46 0.75 1.00 3.15

Diet breadth -0.01 0.22 0.10 -0.04 0.20 -0.01 0.10 0.05 -0.01 0.00 1.00 1.15 Habitat breadth -0.06 0.01 -0.16 -0.27 0.33 -0.19 -0.07 0.29 -0.16 -0.22 0.19 1.19

296

297

Conclusions

1. In the majority of terrestrial vertebrates, extirpated areas were more likely to be heavily humanized, as expected under the refuge null model (species remain in areas where human impact is low). This occurred at both range and fragment scales. 2. Species also showed that extirpated areas were explained by the relative position within the range, but differing among scales. At range scale, extirpated areas were more likely to be near the border of the range for initial stages of range contraction. When more than half of the range has been lost, extirpated areas were more likely to be in the core of the historical range, as predicted by the contagion model. At fragment scale, extirpated areas were more likely to be located at the border of the historical border but this probability of extirpation was weaker as contraction advanced to the final stages, an intermediate prediction of both the contagion and the demographic null models. 3. Earlier null models of range contraction have focused on single processes –basic population rules and simple threat dynamics. Single-process null-models do not provide adequate baselines, at least for terrestrial vertebrates most probably because species persistence may be influenced by multiple external threats and intrinsic processes. 4. There is a need to develop more realistic null models to use as baselines. Without departing from the objective of simplicity, we propose to combine simple key elements already identified as relevant to define new multi-process null models of range contraction. 5. Range area is negatively associated with vulnerability to extinction. Additionally, the inclusion of diverse spatial descriptors of distribution range, namely shape, number of fragments, and fragment size heterogeneity improved our capacity to predict extinction risk. We detected complex relationships, revealed by multiple interaction terms, indicating that simple rules of thumb associating vulnerability to the spatial configuration of the range are likely to be inadequate. 6. Most species distribution ranges are spatially complex, often formed by multiple fragments with different shapes. These spatial descriptors of distribution ranges have a clear ecological basis, are simple to calculate and only require range distribution

298

Conclusions

maps. In spite of the uncertainty associated with the definition of species ranges and their heterogeneous quality, their availability for most terrestrial vertebrates makes them a suitable source of information to be used in vulnerability assessments. 7. Currently there is a special interest to evaluate and predict the consequences of climate change on biodiversity. All in all we found that biodiversity loss in terrestrial vertebrates is better explained by land use impact. Efforts to evaluate and predict future biodiversity losses should consider the impact of land use changes, even if the main focus is on other drivers. 8. There is a gap in knowledge on how simultaneous global change drivers affect range contraction. We observed that including both climate change and land use impacts, with or without their interaction, described range contraction in a better way than models considering only one driver. We should consider diverse impacts to evaluate and predict biodiversity loss and to properly associate patterns of change to a specific driver. 9. Species with contraction patterns associated to climate change showed: (1) range contraction towards higher latitudes particularly marked at intermediate to higher latitudes; (2) contraction towards higher altitudes mainly at low latitudes; and (3) a weak tendency for contraction towards steeper areas at low latitudes. 10. Species with contraction patterns associated to land use showed: (1) consistent range contraction towards lower latitudes; (2) no clear tendency in elevation changes; and (3) a tendency to remain in steeper areas when at low latitudes. 11. Relatively few species had contraction patterns best explained by both drivers (additive model) but these species showed clearer environmental patterns with range contraction towards higher latitudes particularly at intermediate to higher latitudes, and consistent changes towards higher elevations and steeper areas. Species associated with both drivers showed a high percentage of contraction and were more likely to be in higher risk Red List categories. 12. Missing data represent a prevalent problem in comparative analyses and unfortunately, because data are not missing at random conventional approaches often used to fill data gaps are not valid or present important challenges.

299

13. Addressing the data bias problem requires greater investment in data collection and dissemination, as well as the development of methodological approaches to effectively correct existing biases.

300

301

Bibliography Bibliography

Ackerly, D. D. 2003. Community assembly, niche conservatism, and adaptive evolution in changing environments. — Int. J. Plant Sci. 164: S165–S184. Alroy, J. 2001. A multispecies overkill simulation of the end-Pleistocene megafaunal mass extinction. — Science 292: 1893-1896. Ameztegui, A. et al. 2016. Land‐use legacies rather than climate change are driving the recent upward shift of the mountain tree line in the Pyrenees. — Global Ecol. Biogeogr. 25: 263-273. Amori, G. and Gippoliti, S. 2000. What do mammalogists want to save? Ten years of mammalian conservation biology. — Biodivers. Conserv. 9: 785-793. Araújo, M. B. et al. 2005. Validation of species-climate impact models under climate change. — Global Change Biol. 11: 1504-1513. Araújo, M. B. and Rozenfeld, A. 2014. The geographic scaling of biotic interactions. — Ecography 37: 406–415. Armour, K. C. et al. 2016. Southern Ocean warming delayed by circumpolar upwelling and equatorward transport. — Nature Geoscience 9: 549–554. Barnosky, A. et al. 2011. Has the Earth's sixth mass extinction already arrived? — Nature 471: 51-57. Bartlett, L. J. et al. 2016. Robustness despite uncertainty: regional climate data reveal the dominant role of humans in explaining global extinctions of Late Quaternary megafauna. — Ecography 39: 152-161. Bascompte, J. and Solé, R. V. 1995. Rethinking complexity: modelling spatiotemporal dynamics in ecology. — Trends Ecol. Evol. 10: 361-366. Bascompte, J. and Solé, R. V. 1996. Habitat fragmentation and extinction thresholds in spatially explicit models. — J. Anim. Ecol. 65: 465-473. Bascompte, J. and Solé, R. V. (eds.) 1998. Modeling spatiotemporal dynamics in ecology. — Springer. Belsley, D. A. 2004. Regression diagnostics: identifying influential data and sources of collinearity. — Wiley-Interscience. Bennett, K. D. and Provan, J. 2008. What do we mean by 'refugia'? — Quaternary Science Reviews 8: 2449–2455. Bielby, J. et al. 2010. Modelling extinction risk in multispecies data sets: phylogenetically independent contrasts versus decision trees. — Biodivers. Conserv. 19: 113-127. Bielby, J. et al. 2008. Predicting susceptibility to future declines in the world's frogs. — Conservation Letters 1: 82-90. Bininda-Emonds, O. R. P. et al. 2007. The delayed rise of present-day mammals. — Nature 446: 507-512. Blomberg, S. P. et al. 2003. Testing for phylogenetic signal in comparative data: Behavioral traits are more labile. — Evolution 57: 717-745. Bonebrake, T. C. et al. 2010. Population decline assessment, historical baselines, and conservation. — Conservation Letters 3: 371–378. Botts, E. A. et al. 2013. Small range size and narrow niche breadth predict range contractions in South African frogs. — Global Ecol. Biogeogr. 22: 567–576. Brashares, J. S. 2003. Ecological, behavioral, and life-history correlates of mammal extinctions in West Africa. — Conserv. Biol. 17: 733-743. 302

Bibliography

Bregman, T. P. et al. 2014. Global patterns and predictors of bird species responses to forest fragmentation: implications for ecosystem function and conservation. — Biol. Conserv. 169: 372–383. Breiman, L. 1984. Classification and regression trees. — Wadsworth International Group. Brook, B. W. et al. 2008. Synergies among extinction drivers under global change. — Trends Ecol. Evol. 23: 453-460. Brown, J. H. 1971. Mammals on mountaintops: nonequilibrium insular biogeography. — Am. Nat. 105: 467-478. Brown, J. H. 1984. On the Relationship between Abundance and Distribution of Species. — The American Naturalist 124: 255-279. Brown, J. H. 1995. Macroecology. — University of Chicago Press, Chicago. Brown, J. H. and Kodricbrown, A. 1977. Turnover rates in insular biogeography: effect of immigration on extinction1. — Ecology 58: 445-449. Brown, J. H. et al. 1996. The geographic range: size, shape, boundaries, and internal structure. — Annu. Rev. Ecol. Syst. 27: 597-623. Bruggeman, J. et al. 2009. PhyloPars: estimation of missing parameter values using phylogeny. — Nucleic Acids Res. 37: W179-W184. Buckley, L. B. and Jetz, W. 2007. Environmental and historical constraints on global patterns of amphibian richness. — Proceedings of the Royal Society of London B Biological Sciences 274: 1167–1173. Burnham, K. P. and Anderson, D. R. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. — Springer, New York. Burrows, M. T. et al. 2014. Geographical limits to species-range shifts are suggested by climate velocity. — Nature 507: 492-495. Butchart, S. H. M. et al. 2010. Global Biodiversity: Indicators of Recent Declines. — Science 328: 1164-1168. Butt, N. et al. 2015. Cascading effects of climate extremes on vertebrate fauna through changes to low‐latitude tree flowering and fruiting phenology. — Global Change Biol. 21: 3267–3277. Buuveibaatar, B. et al. 2016. Human activities negatively impact distribution of ungulates in the Mongolian Gobi. — Biol. Conserv. 203: 168-175. Cahill, A. E. et al. 2012. How does climate change cause extinction? — Proc. R. Soc. Lond., Ser. B: Biol. Sci. 280: 1-9. Cardillo, M. 2003. Biological determinants of extinction risk: why are smaller species less vulnerable? — Anim. Conserv. 6: 63-69. Cardillo, M. et al. 2008. The predictability of extinction: biological and external correlates of decline in mammals. — Proc. R. Soc. Lond., Ser. B: Biol. Sci. 275: 1441-1448. Cardillo, M. et al. 2006. Latent extinction risk and the future battlegrounds of mammal conservation. — Proceedings of the National Academy of Sciences of the United States of America 103: 4157-4161. Cardillo, M. et al. 2005. Multiple causes of high extinction risk in large mammal species. — Science 309: 1239-1241. Cardillo, M. et al. 2004. Human population density and extinction risk in the world's carnivores. — PLoS Biol. 2: 0909-0914. Carvalheiro, L. G. et al. 2013. Species richness declines and biotic homogenisation have slowed down for NW-European pollinators and plants. — Ecol. Lett. 16: 870-878. Caughley, G. et al. 1988. The edge of the range. — J. Anim. Ecol. 57: 771-785.

303

Bibliography

Cazelles, K. et al. 2015. On the integration of biotic interaction and environmental constraints at the biogeographical scale. — Ecography 38: 001–011. CBD (ed.) 2010. (Secretariat of the Convention on Biological Diversity). Global Biodiversity Outlook 3. Ceballos, G. and Ehrlich, P. R. 2002. Mammal population losses and the extinction crisis. — Science 296: 904-907. Center for International Earth Science Information Network - CIESIN - Columbia University, a. C. I. d. A. T.-C. 2005. Gridded Population of the World, Version 3 (GPWv3). Clark Labs, C. U. 2009. Idrisi 16: The Taiga Edition. Clavero, M. et al. 2009. Prominent role of invasive species in avian biodiversity loss. — Biol. Conserv. 142: 2043–2049. Clavero, M. and García-Berthou, E. 2005. Invasive species are a leading cause of animal extinctions. — TRENDS in Ecology and Evolution Clavero, M. and Hermoso, V. 2015. Historical data to plan the recovery of the European eel. — J. Appl. Ecol. 52: 960-968. Clavero, M. and Revilla, E. 2014. Biodiversity data: Mine centuries-old citizen science. — Nature 510: 35. Costello, M. J. 2009. Motivating online publication of data. — Bioscience 59: 418-427. Cowley, M. J. R. et al. 2001. Density–distribution relationships in British butterflies. I. The effect of mobility and spatial scale. — J. Anim. Ecol. 70: 410–425. Crees, J. J. et al. 2016. Millennial-scale faunal record reveals differential resilience of European large mammals to human impacts across the Holocene. — Proc. R. Soc. B 283: 2015 2152. Cunningham, H. R. et al. 2016. Abiotic and biotic constraints across reptile and amphibian ranges. — Ecography 39: 1-8. Channell, R. and Lomolino, M. V. 2000a. Dynamic biogeography and conservation of endangered species. — Nature 403: 84-86. Channell, R. and Lomolino, M. V. 2000b. Trajectories to extinction: spatial dynamics of the contraction of geographical ranges. — J. Biogeogr. 27: 169-179. Chen, I. C. C. et al. 2011. Rapid range shifts of species associated with high levels of climate warming. — Science 333: 1024-1026. David, H. R. et al. 2003. Estimates of minimum viable population sizes for vertebrates and factors influencing those estimates. — Biol. Conserv. 113: 23-34. David Tilman and Kareiva, P. (eds.) 1997. Spatial Ecology: The Role of Space in Population Dynamics and Interspecific Interactions. — Princeton University Press. Davidson, A. D. et al. 2009. Multiple ecological pathways to extinction in mammals. — Proc. Natl. Acad. Sci. USA 106: 10702-10705. Davidson, A. M. et al. 2011. Do invasive species show higher phenotypic plasticity than native species and, if so, is it adaptive? A meta‐analysis. — Ecol. Lett. 14: 419–431. Davies, T. J. et al. 2008. Phylogenetic trees and the future of mammalian biodiversity. — Proceedings of the National Academy of Sciences of the United States of America 105: 11556-11563. Davis, A. J. et al. 1998. Making mistakes when predicting shifts in species range in response to global warming. — Nature 391: 783-786. Davis, M. B. and Shaw, R. G. 2001. Range shifts and adaptive responses to Quaternary climate change. — Science 292: 673-679. De'ath, G. and Fabricius, K. E. 2000. Classification and regression trees: a powerful yet simple technique for ecological data analysis. — Ecology 81: 3178-3192. 304

Bibliography

Development Core Team 2011. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Devictor, V. et al. 2010. Defining and measuring ecological specialization. — J. Appl. Ecol. 47: 15-25. Devictor, V. et al. 2008. Birds are tracking climate warming, but not fast enough. — Proceedings of the Royal Society B: Biological Sciences 275: 2743-2748. Di Fonzo, M. M. I. et al. 2016. Patterns of mammalian population decline inform conservation action. — J. Appl. Ecol. 53: 1046-1054. Diamond, J. M. 1975. The island dilemma: lessons of modern biogeographic studies for the design of natural reserves. — Biological Conserv 7: 129-145. Dietz, H. and Edwards, P. J. 2006. Recognition that causal processes change during plant invasion helps explain conflicts in evidence. — Ecology 87: 1359–1367. Ellis, E. C. and Ramankutty, N. 2008. Putting people in the map: anthropogenic biomes of the world. — Putting people in the map: anthropogenic biomes of the world doi: 10.1890/070062. ESRI 2008. ArcMap 9.3 Redlands, CA: Environmental Systems Research Institute. Evans, K. L. et al. 2005. The roles of extinction and colonization in generating species- energy relationships. — J. Anim. Ecol. 74: 498-507. Fagan, W. F. et al. 2001. Characterizing population vulnerability for 758 species. — Ecol. Lett. 4: 132-138. Felsenstein, J. 1985. Phylogenies and the comparative method. — Am. Nat. 125: 1-15. Fisher, D. O. 2011. Trajectories from extinction: where are missing mammals rediscovered? — Global Ecol. Biogeogr. 20: 415-425. Fisher, D. O. et al. 2003. Extrinsic versus intrinsic factors in the decline and extinction of Australian marsupials. — Proc. R. Soc. Lond. B Biol. Sci. 270: 1801-1808. Fisher, D. O. and Owens, I. P. F. 2004. The comparative method in conservation biology. — Trends Ecol. Evol. 19: 391-398. Forester, D. J. and Machlis, G. E. 1996. Modeling human factors that affect the loss of biodiversity. — Conserv. Biol. 10: 1253-1263. Forman, R. T. T. 1995. Some general principles of landscape and regional ecology. — Landscape Ecol. 10: 133-142. Franco, A. M. A. et al. 2006. Impacts of climate warming and habitat loss on extinctions at species' low‐latitude range boundaries. — Global Change Biol. 12: 1545-1553. Freckleton, R. P. et al. 2002. Phylogenetic analysis and comparative data: a test and review of evidence. — Am. Nat. 160: 712-726. Frishkoff, L. O. et al. 2016. Climate change and habitat conversion favour the same species. — Ecol. Lett. 19: 1081-1090. Fritz, S. A. et al. 2009. Geographical variation in predictors of mammalian extinction risk: big is bad, but only in the tropics. — Ecol. Lett. 12: 538-549. Froese, R. and Pauly, D. 2000. FishBase 2000: concepts, design and data sources. — ICLARM. Gallien, L. et al. 2010. Predicting potential distributions of invasive species: where to go from here? — Divers. Distrib. 16: 331-342. Gaston, K. J. 1990. Patterns in the geographical ranges of species. — Biol. Rev. Camb. Philos. Soc. 65: 105-129. Gaston, K. J. 1991. How Large Is a Species' Geographic Range? — Oikos 61: 434-438. Gaston, K. J. 1994a. Geographic range sizes and trajectories to extinction. — Biodivers. Lett. 2: 163-170. 305

Bibliography

Gaston, K. J. (ed.) 1994b. Rarity. — Chapman & Hall. Gaston, K. J. 1996. Species-range-size distributions: patterns, mechanisms and implications. — Trends Ecol. Evol. 11: 197-201. Gaston, K. J. 2003. The Structure and Dynamics of Geographic Ranges. — Oxford University Press, Oxford. Gaston, K. J. 2008. Biodiversity and extinction: the dynamics of geographic range size. — Progress in Physical Geography 32: 678-683. Gaston, K. J. 2009. Geographic range limits: achieving synthesis. — Proc. R. Soc. Lond., Ser. B: Biol. Sci. 276: 1395-1406. Gaston, K. J. and Blackburn, T. M. 1995. Birds, body size and the threat of extinction. — Philos. Trans. R. Soc. Lond., Ser. B: Biol. Sci. 347: 205-212. Gaston, K. J. and Curnutt, J. L. 1998. The dynamics of abundance-range size relationships. — Oikos 81: 38-44. Gaston, K. J. and Fuller, R. A. 2009. The sizes of species' geographic ranges. — J. Appl. Ecol. 46: 1-9. Gellrich, M. et al. 2007. Agricultural land abandonment and natural forest re-growth in the Swiss mountains: A spatially explicit economic analysis. — Agric., Ecosyst. Environ. 118: 93-108. Giam, X. et al. 2011. Local geographic range predicts freshwater fish extinctions in Singapore. — J. Appl. Ecol. 48: 356-363. Gilman, S. E. et al. 2010. A framework for community interactions under climate change. — Trends Ecol. Evol. 25: 325-331. Gilpin, M. E. and Soulé, M. E. 1986. Minimum viable populations: processes of species extinction. — In: Soulé, M. E. (ed), Conservation Biology: The Science of Scarcity and Diversity. Sinauer Associates, pp. 19-34. Gilroy, J. J. et al. 2014. Effect of scale on trait predictors of species responses to agriculture. — Conserv. Biol. 29: 463-472. Gómez, A. and Lunt, D. H. 2007. Refugia within refugia: patterns of phylogeographic concordance in the Iberian Peninsula. — In: Weiss, S. and Ferrand, N. (eds), Phylogeography of southern European refugia. Springer, pp. 155-188. González-Suárez, M. et al. 2013. Which intrinsic traits predict vulnerability to extinction depends on the actual threatening processes. — Ecosphere 4 (6): 76. González-Suárez, M. et al. 2012. Biases in comparative analyses of extinction risk: mind the gap. — The Journal of animal ecology 81: 1211-1222. González-Suárez, M. and Revilla, E. 2013. Variability in life-history and ecological traits is a buffer against extinction in mammals. — Ecol. Lett. 16: 242-251. González-Suárez, M. and Revilla, E. 2014. Generalized drivers in the Mammalian endangerment process. — PLoS One 9: e90292. Gotelli, N. J. 2001. Research frontiers in null model analysis. — Global Ecol. Biogeogr. 10: 337-343. Gott, R. J. et al. 2007. Map Projections Minimizing Distance Errors. — Cartographica: The International Journal for Geographic Information and Geovisualization 42: 219-234. Gray, S. M. et al. 2015. Temperature and trophic structure are driving microbial productivity along a biogeographical gradient. — Ecography 38: 001-009. Grinnell, J. 1917. The niche-relationships of the California Thrasher. — Auk Cambridge Mass 34: 427-433. Gross, S. J. and Price, T. D. 2000. Determinants of the northern and southern range limits of a warbler. — J. Biogeogr. 27: 869-878. 306

Bibliography

Guisan, A. and Thuiller, W. 2005. Predicting species distribution: offering more than simple habitat models. — Ecol. Lett. 8: 993-1009. Gyllenberg, M. and Hanski, I. 1992. Single-species metapopulation dynamics: a structured model. — Theor. Popul. Biol. 42: 35-61. Gynther, I. et al. 2016. Confirmation of the extinction of the Bramble Cay melomys Melomys rubicola on Bramble Cay, Torres Strait: results and conclusions from a comprehensive survey in August–September 2014. Unpublished report to the Department of Environment and Heritage Protection, Queensland Government. Brisbane (Australia). Hannah, L. et al. 2014. Fine-grain modeling of species' response to climate change: holdouts, stepping-stones, and microrefugia. — Trends in Ecology and Evolution 29: 390-397. Hanski and Gyllenberg 1997. Uniting Two General Patterns in the Distribution of Species. — Science 275: 397-400. Hanski, I. 1998. Metapopulation dynamics. — Nature 396: 41-49. Hanski, I. (ed.) 1999. Metapopulation Ecology. — Oxford Univiversity Press. Hanski, I. and Gyllenberg, M. 1993. Two general metapopulation models and the core- satellite species hypothesis. — Am. Nat. 142: 17-41. Hanski, I. et al. 1996. Minimum viable metapopulation size. — Am. Nat. 147: 527-541. Hanski, I. and Ovaskainen, O. 2000. The metapopulation capacity of a fragmented landscape. — Nature 404: 755-758. Harris, G. and Pimm, S. L. 2008. Range size and extinction risk in forest birds. — Conservation biology : the journal of the Society for Conservation Biology 22: 163- 171. Harris, I. et al. 2014. Updated high‐resolution grids of monthly climatic observations–the CRU TS3. 10 Dataset. — International Journal of Climatology 34: 623-642. Hayward, M. W. 2011. Using the IUCN Red List to determine effective conservation strategies. — Biodivers. Conserv. 20: 2563–2573. Heim, N. A. and Peters, S. E. 2011. Regional Environmental Breadth Predicts Geographic Range and Longevity in Fossil Marine Genera. — PLoS One 6: e18946. Hemerik, L. et al. 2006. The eclipse of species ranges. — Acta Biotheor. 54: 255-266. Hengeveld, R. and Haeck, J. 1982. The distribution of abundance. I. Measurements. — J. Biogeogr. 9: 303-316. Hill, J. K. et al. 2001. Impacts of landscape structure on butterfly range expansion. — Ecology … 4: 313-321. Hill, J. K. et al. 1999. Climate and habitat availability determine 20th century changes in a butterfly's range margin. — Proceedings of the Royal Society of London B: Biological Sciences 266: 1197-1206. Hinder, S. L. et al. 2012. Changes in marine dinoflagellate and diatom abundance under climate change. — Nature Climate Change 2: 271-275. Hirzel, A. H. and Le Lay, G. 2008. Habitat suitability modelling and niche theory. — J. Appl. Ecol. 45: 1372-1381. Hoffmann, M. et al. 2010. The Impact of Conservation on the Status of the World's Vertebrates. — Science 330: 1503-1509. Holt, R. D. and Barfield, M. 2009. Trophic interactions and range limits: the diverse roles of predation. — Proceedings of the Royal Society B: Biological Sciences 276: 1435- 1442. Holland, R. A. and Wikelski, M. 2009. Studying the migratory behavior of individual bats: current techniques and future directions. — J. Mammal. 90: 1324-1329. 307

Bibliography

Hollings, T. et al. 2016. Disease‐induced decline of an apex predator drives invasive dominated states and threatens biodiversity. — Ecology 97: 394-405. Hulme, P. E. 2009. Trade, transport and trouble: managing invasive species pathways in an era of globalization. — J. Appl. Ecol. 46: 10-18. Huntley, B. et al. 1989. Climatic control of the distribution and abundance of beech (Fagus L.) in Europe and North America. — J. Biogeogr. 16: 551-560. Hurlbert, A. H. and Jetz, W. 2007. Species richness, hotspots, and the scale dependence of range maps in ecology and conservation. — Proc. Natl. Acad. Sci. USA 104: 13384- 13389. Hurlbert, A. H. and Liang, Z. 2012. Spatiotemporal variation in avian migration phenology: citizen science reveals effects of climate change. — PLoS One doi: 10.1371/journal.pone.0031662. Hurtt, G. C. et al. 2011. Harmonization of land-use scenarios for the period 1500-2100: 600 years of global gridded annual land-use transitions, wood harvest, and resulting secondary lands. — Clim. Change 109: 117-161. International Union for Conservation of Nature 2010a. IUCN red list of threaten species Version 2010.4. http://www.iucnredlist.org/ (accessed February 08, 2010). International Union for Conservation of Nature 2010b. IUCN Red List of Threatened Species. Version 2010.4. International Union for Conservation of Nature 2010c. METADATA: Digital Distribution Maps on The IUCN Red List of Threatened Species™ http://www.iucnredlist.org. IPCC 2007. Climate change 2007: impacts, adaptation and vulnerability. — Cambridge Univ. Press. IPCC (ed.) 2014. Climate Change 2013: The physical cience basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovermental Panel on Climate Change. — Cambridge University Press. IPCC 2015. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on climate change. — In: Barros, V. R. and Field, C. B. (eds), Climate change 2014: impacts, adaptation, and vulnerability. . Cambridge University Press. IUCN 2010. IUCN Red List of Threatened Species. Version 2010.4. http://www.iucnredlist.org>. Downloaded on 27 October 2010 IUCN 2012. IUCN Red List Categories and Criteria. Version 3.1. IUCN 2015. The IUCN Red List of Threatened Species. Jackson, S. T. and Betancourt, J. L. 2009. Ecology and the ratchet of events: climate variability, niche dimensions, and species distributions. — Proceedings of the National Academy of Sciences 106: 19685–19692. Jakobs, G. et al. 2004. Introduced plants of the invasive Solidago gigantea (Asteraceae) are larger and grow denser than conspecifics in the native range. — Divers. Distrib. 10: 11-19. Janecka, J. E. et al. 2014. Loss of genetic diversity among ocelots in the United States during the 20th century linked to human induced population reductions. — PloS one 9: e89384. Jetz, W. et al. 2007. Projected impacts of climate and land-use change on the global diversity of birds. — PLoS Biol. 5: Jiménez-Valverde, A. et al. 2011. Use of niche models in invasive species risk assessments. — Biol. Invasions 13: 2785-2797.

308

Bibliography

Johnston, A. et al. 2013. Observed and predicted effects of climate change on species abundance in protected areas. — Nature Climate Change 3: 1055-1061. Jones, H. L. and Diamond, J. M. 1976. Short-time-base studies of turnover in breeding bird populations on the California Channel Islands. — The Condor 78: 526-549. Jones, H. P. et al. 2008. Severity of the effects of invasive rats on seabirds: a global review. — Conserv. Biol. 22: 16-26. Jones, K. E. et al. 2009. PanTHERIA: a species-level database of life history, ecology, and geography of extant and recently extinct mammals. — Ecology 90: 2648-2648. Kehoe, L. et al. 2015. Global patterns of agricultural land-use intensity and vertebrate diversity. — Divers. Distrib. 21: 1308-1318. Keppel, G. et al. 2012. Refugia: identifying and understanding safe havens for biodiversity under climate change. — Global Ecol. Biogeogr. 21: 393-404. Kerr, J. T. and Currie, D. J. 1995. Effects of human activity on global extinction risk. — Conserv. Biol. 9: 1528-1538. Kerr, J. T. et al. 2015. Climate change impacts on bumblebees converge across continents. — Science 349: 177-180. Kier, G. et al. 2009. A global assessment of endemism and species richness across island and mainland regions. — Proceedings of the National Academy of Sciences USA 106: 9322-9327. Koch, P. L. and Barnosky, A. D. 2006. Late Quaternary Extinctions: State of the Debate. — Annual Review of Ecology, Evolution, and Systematics 37: 215-250. Kozak, K. H. and Wiens, J. J. 2016. Testing the Relationships between Diversification, Species Richness, and Trait Evolution. — Syst. Biol. 0: 1-14. Kunin, W. E. and Gaston, K. J. 1993. The biology of rarity: patterns, causes and consequences. — Trends Ecol. Evol. 8: 298-301. Laliberte, A. S. and Ripple, W. J. 2004. Range contractions of North American carnivores and ungulates. — Bioscience 54: 123-138. Lane, J. E. et al. 2012. Delayed phenology and reduced fitness associated with climate change in a wild hibernator. — Nature 489: 554-557. Laurance, W. F. 1991. Ecological Correlates of Extinction Proneness in Australian Tropical Rain Forest Mammals. — Conserv. Biol. 5: 79-89. Laurance, W. F. et al. 2009. Impacts of roads and linear clearings on tropical forests. — Trends in Ecology and Evolution 24: 659-669. Lavergne, S. et al. 2010. Biodiversity and climate change: integrating evolutionary and ecological responses of species and communities. — Annu. Rev. Ecol. Syst. 41: 321- 350. Lawton, J. H. 1993. Range, population abundance and conservation. — Trends in Ecology and Evolution 8: 409-413. Lea, J. et al. 2016. Recognition and management of ecological refugees: A case study of the Cape mountain zebra. — Biol. Conserv. 203: 207-215. Lenoir, J. and Svenning, J. C. 2015. Climate‐related range shifts–a global multidimensional synthesis and new research directions. — Ecography 38: 15-28. Lester, S. E. et al. 2007. The relationship between dispersal ability and geographic range size. — Ecol. Lett. 10: 745-758. Levine, J. M. et al. 2003. Mechanisms underlying the impacts of exotic plant invasions. — Proceedings of the Royal Society B: Biological Sciences 270: 775-781. Levins, R. 1969. Some demographic and genetic consequences of environmental heterogeneity for biological control. — Bull. ent. Soc. Am. 15: 237-240. 309

Bibliography

Li, J. et al. 2016. Climate refugia of snow leopards in High Asia. — Biol. Conserv. 203: 188- 196. Li, X. et al. 2015. Human impact and climate cooling caused range contraction of large mammals in China over the past two millennia. — Ecography 38: 74–82. Lieury, N. et al. 2016. Geographically isolated but demographically connected: Immigration supports efficient conservation actions in the recovery of a range-margin population of the Bonelli's eagle in France. — Biol. Conserv. 195: 272-278. Lindenmayer, D. B. et al. 2014. Complex responses of birds to landscape‐level fire extent, fire severity and environmental drivers. — Divers. Distrib. 20: 467-477. Little, R. J. A. and Rubin, D. B. 2002. Statistical analysis with missing data. — Wiley. Lomolino, M. V. and Channell, R. 1995. Splendid isolation - Patterns of geographic range collapse in endangered mammals. — J. Mammal. 76: 335-347. Lomolino, M. V. and Channell, R. 1998. Range collapse, re-introductions, and biogeographic guidelines for conservation. — Conserv. Biol. 12: 481-484. Longobardi, P. et al. 2016. Deforestation Induced Climate Change: Effects of Spatial Scale. — PLoS One 11: e0153357. Lorenzen, E. D. et al. 2011. Species-specific responses of Late Quaternary megafauna to climate and humans. — Nature 479: 359-364. Lucas, P. M. et al. 2016. Toward multifactorial null models of range contraction in terrestrial vertebrates. — Ecography doi: 10.1111/ecog.01819. Luck, G. W. 2007. A review of the relationships between human population density and biodiversity. — Biological Reviews 82: 607-645. MacArthur, B. H. and Wilson, E. O. 1963. An equilibrium theory of insular zoogeography. — Evol. Dev. 17: 373-387. MacArthur, R. H. and Wilson, E. O. 1967. The Theory of Island Biogeography. — Princeton University Press, Princeton, N.J. MacDonald, D. et al. 2000. Agricultural abandonment in mountain areas of Europe: Environmental consequences and policy response. — J. Environ. Manage. 59: 47-69. Maclean, I. et al. 2016. Fine-scale climate change: modelling spatial variation in biologically meaningful rates of warming. — Global Change Biol. doi: 10.1111/gcb.13343. Maclean, I. et al. 2011. Predicting changes in the abundance of African wetland birds by incorporating abundance–occupancy relationships into habitat association models. — Divers. Distrib. 17: 480-490. Maklakov, A. A. et al. 2011. Brains and the city: big-brained passerine birds succeed in urban environments. — Biol. Lett. 7: 730–732. Marco, M. and Santini, L. 2015. Human pressures predict species’ geographic range size better than biological traits. — Global Change Biol. 21: 2169-2178. Martins, E. P. and Hansen, T. F. 1997. Phylogenies and the comparative method: a general approach to incorporating phylogenetic information into the analysis of interspecific data. — Am. Nat. 149: 646-667. Matthews, L. J. et al. 2011. Primate extinction risk and historical patterns of speciation and extinction in relation to body mass. — Proceedings of the Royal Society of London B Biological Sciences 278: 1256-1263. Maxwell, S. L. et al. 2016. Biodiversity: The ravages of guns, nets and bulldozers. — Nature 536: 143-145. Mayr, E. (ed.) 1963. Animal Species and Evolution. — Belknap Press. McGeoch, M. A. et al. 2010. Global indicators of biological invasion: species numbers, biodiversity impact and policy responses. — Divers. Distrib. 16: 95-108. 310

Bibliography

McKinney, M. L. 1997. Extinction vulnerability and selectivity: combining ecological and paleontological views. — Annu. Rev. Ecol. Syst. 28: 495-516. McNab, B. K. 1994. Energy Conservation and the Evolution of Flightlessness in Birds. — Am. Nat. 144: 628-642. McNab, B. K. 2003. Standard energetics of phyllostomid bats: The inadequacies of phylogenetic-contrast analyses. — Comparative Biochemistry and Physiology Part A Molecular & Integrative Physiology 135A: 357-368. Medina, F. M. et al. 2011. A global review of the impacts of invasive cats on island endangered vertebrates. — Global Change Biol. 17: 3503–3510. Mehlman, D. W. 1997. Change in avian abundance across the geographic range in response to environmental change. — Ecol. Appl. 7: 614-624. Mena Berrios, J. B. 2008. Geodesia Superior. — Centro Nacional de Información Geográfica. Menéndez, R. and González‐Megías, A. 2014. Climate change and elevational range shifts: evidence from dung beetles in two European mountain ranges. — Global Ecol. Biogeogr. 23: 646-657. Merrill, R. M. et al. 2008. Combined effects of climate and biotic interactions on the elevational range of a phytophagous insect. — J. Anim. Ecol. 77: 145-155. Morin, X. and Lechowicz, M. J. 2013. Niche breadth and range area in North American trees. — Ecography 36: 300-312. Morrison, L. W. et al. 2004. Potential global range expansion of the invasive fire ant, Solenopsis invicta. — Biol. Invasions 6: 183-191. Morueta-Holme, N. et al. 2015. Strong upslope shifts in Chimborazo's vegetation over two centuries since Humboldt. — Proceedings of the national Academy of Sciences USA 112: 12741–12745. Morueta-Holme, N. et al. 2013. Habitat area and climate stability determine geographical variation in plant species range sizes. — Ecol. Lett. 16: 1446-1454. Moss, R. H. et al. 2010. The next generation of scenarios for climate change research and assessment. — Nature 463: 747-756. Mottet, A. et al. 2006. Agricultural land-use change and its drivers in mountain landscapes: A case study in the Pyrenees. — Agric., Ecosyst. Environ. 114: 296-310. Moussalli, A. et al. 2009. Variable responses of skinks to a common history of rainforest fluctuation: concordance between phylogeography and palaeo‐distribution models. — Mol. Ecol. 18: 483-499. Muhlfeld, C. C. et al. 2014. Invasive hybridization in a threatened species is accelerated by climate change. — Nature Climate Change 4: 620-624. Murray, B. R. and Dickman, C. R. 2000. Relationships between body size and geographical range size among Australian mammals: has human impact distorted macroecological patterns? — Ecography 23: 92-100. Murray, K. A. et al. 2011. Integrating species traits with extrinsic threats: closing the gap between predicting and preventing species declines. — Proc. R. Soc. Lond., Ser. B: Biol. Sci. 278: 1515-1523. Nakagawa, S. and Freckleton, R. P. 2008. Missing inaction: the dangers of ignoring missing data. — Trends Ecol. Evol. 23: 592-596. Narendra S. Goel and Richter-Dyn, N. 1974. Stochastic models in biology. — Academic Press. Nelson, G. C. et al. 2014. Climate change effects on agriculture: Economic responses to biophysical shocks. — Proceedings of the National Academy of Sciences USA 111: 3274-3279. 311

Bibliography

Newbold, T. et al. 2015. Global effects of land use on local terrestrial biodiversity. — Nature 520: 45-50. Newbold, T. et al. 2014. Functional traits, land‐use change and the structure of present and future bird communities in tropical forests. — Global Ecol. Biogeogr. 23: 1073-1084. O’Hara, R. B. and Kotze, D. J. 2010. Do not log-transform count data. — Methods Ecol. Evol. 1: 118-122. Oak Ridge National Laboratory Distributed Active Archive Center 2010. MODIS (MCD12Q1) Land Cover Product http://webmap.ornl.gov ORNL DAAC, Oak Ridge, Tennessee, USA. Accessed September, 2013. Ockendon, N. et al. 2014. Mechanisms underpinning climatic impacts on natural populations: altered species interactions are more important than direct effects. — Global Change Biol. 20: 2221-2229. Ohlemüller, R. et al. 2008. The coincidence of climatic and species rarity: high risk to small- range species from climate change. — Biol. Lett. 4: 568-572. Olalla-Tárraga, M. Á. et al. 2011. Climatic niche conservatism and the evolutionary dynamics in species range boundaries: global congruence across mammals and amphibians. — J. Biogeogr. 38: 2237–2247. Olalla‐Tárraga, M. et al. 2016. Contrasting evidence of phylogenetic trophic niche conservatism in mammals worldwide. — J. Biogeogr. doi: 10.1111/jbi.12823. Oliver, T. H. et al. 2015. Interacting effects of climate change and habitat fragmentation on drought-sensitive butterflies. — Nature Climate Change 5: 941-945. Oliver, T. H. and Morecroft, M. D. 2014. Interactions between climate change and land use change on biodiversity: attribution problems, risks, and opportunities. — WIREs Clim Change 5: 317–335. Organization, W. M. Orr, M. R. and Smith, T. B. 1998. Ecology and speciation. — Trends in Ecology and Evolution 13: 502-506. Orzechowski, E. A. et al. 2015. Marine extinction risk shaped by trait-environment interactions over 500 million years. — Global Change Biol. 21: 3595-3607. Ovaskainen, O. 2002. Long-term persistence of species and the SLOSS problem. — J. Theor. Biol. 218: 419–433. Pagel, M. 1999a. Inferring the historical patterns of biological evolution. — Nature 401: 877- 884. Pagel, M. 1999b. The maximum likelihood approach to reconstructing ancestral character states of discrete characters on phylogenies. — Syst. Biol. 48: 612-622. Parmesan, C. 1996. Climate and species' range. — Nature 382: 765-766. Parmesan, C. et al. 2005. Empirical perspectives on species borders: from traditional biogeography to global change. — Oikos 108: 58-75. Parmesan, C. et al. 1999. Poleward shifts in geographical ranges of butterfly species associated with regional warming. — Nature 399: 579-583. Parmesan, C. and Yohe, G. 2003. A globally coherent fingerprint of climate change impacts across natural systems. — Nature 421: 37-42. Pearman, P. B. et al. 2008. Niche dynamics in space and time. — Trends Ecol. Evol. 23: 149- 158. Pearson, R. G. and Dawson, T. P. 2003. Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful? — Global Ecol. Biogeogr. 12: 361-371.

312

Bibliography

Pease, C. M. et al. 1989. A model of population growth, dispersal and evolution in a changing environment. — Ecology 70: 1657-1664. Pechmann, J. H. K. and Wilbur, H. M. 1994. Putting declining amphibian populations in perspective: natural fluctuations and human impacts. — Herpetologica 50: 65-84. Pekin, B. K. and Pijanowski, B. C. 2012. Global land use intensity and the endangerment status of mammal species. — Divers. Distrib. 18: 909-918. Pelletier, J. 2000. Model assessments of the optimal design of nature reserves for maximizing species longevity. — J. Theor. Biol. 202: 25-32. Pereira, H. M. et al. 2010. Scenarios for global biodiversity in the 21st century. — Science 330: 1496-1501. Peters, H. et al. 2014. Identifying species at extinction risk using global models of anthropogenic impact. — Global Change Biol. 21: 618–628. Peterson, A. T. et al. 1999. Conservatism of Ecological Niches in Evolutionary Time. — Science 285: 1265-1267. Pilfold, N. W. et al. 2016. Migratory response of polar bears to sea ice loss: to swim or not to swim. — Ecography 39: 001–011. Pimentel, D. et al. 2000. Environmental and economic costs of nonindigenous species in the United States. — Bioscience 50: 53-65. Pimm, S. et al. 2006. Human impacts on the rates of recent, present, and future bird extinctions. — Proceedings of the National Academy of Sciences USA 103: 10941- 10946. Pimm, S. L. et al. 2014. The biodiversity of species and their rates of extinction, distribution, and protection. — Science 344: 1246752. Pimm, S. L. et al. 1988. On the risk of extinction. — Am. Nat. 132: 757-785. Pino, J. et al. 2009. Floristic homogenization by native ruderal and alien plants in north‐east Spain: the effect of environmental differences on a regional scale. — Global Ecol. Biogeogr. 18: 573-5574. Pinsky, M. L. et al. 2011. Unexpected patterns of fisheries collapse in the world's oceans. — Proceedings of the National Academy of Sciences USA 108: 8317-8322. Plieninger, T. et al. 2006. Traditional land-use and nature conservation in European rural landscapes. — Environ. Sci. Policy 9: 317-321. Pomara, L. Y. et al. 2014. Demographic consequences of climate change and land cover help explain a history of extirpations and range contraction in a declining snake species. — Global Change Biol. 20: 2087–2099. Pounds, J. A. et al. 2006. Widespread amphibian extinctions from epidemic disease driven by global warming. — Nature 439: 161-167. Pöyry, J. et al. 2009. Species traits explain recent range shifts of Finnish butterflies. — Global Change Biol. 15: 732-743. Purvis, A. 2008. Phylogenetic approaches to the study of extinction. — Annu. Rev. Ecol. Syst. 39: 301-319. Purvis, A. et al. 2000. Predicting extinction risk in declining species. — Proceedings of the Royal Society of London B Biological Sciences 267: 1947-1952. Pysek, P. et al. 2004. Alien plants in checklists and floras: towards better communication between taxonomists and ecologists. — Taxon 53: 131-143. Quinn, J. F. and Hastings, A. 1987. Extinction in subdivided habitats. — Conserv. Biol. 3: 198-208. R Development Core Team 2013. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 313

Bibliography

Ramankutty, N. et al. 2008. Farming the planet: 1. Geographic distribution of global agricultural lands in the year 2000. — Global Biogeochemical Cycles 22: GB1003. Randin, C. F. et al. 2009. Climate change and plant distribution: local models predict high- elevation persistence. — Global Change Biol. 15: 1557-1569. Reinhart, K. O. et al. 2003. Plant–soil biota interactions and spatial distribution of black cherry in its native and invasive ranges. — Ecol. Lett. 6: 1046–1050. Richard T. T. Forman and Godron, M. (eds.) 1986. Landscape Ecology. — John Wiley and Sons. Richardson, D. M. and Rejmánek, M. 2011. Trees and shrubs as invasive alien species–a global review. — Divers. Distrib. 17: 788–809. Robinson, N. et al. 2014. EarthEnv-DEM90: A nearly-global, void-free, multi-scale smoothed, 90m digital elevation model from fused ASTER and SRTM data. — ISPRS Journal of Photogrammetry and Remote Sensing 87: 57–67. Rödder, D. and Lötters, S. 2009. Niche shift versus niche conservatism? Climatic characteristics of the native and invasive ranges of the Mediterranean house gecko (Hemidactylus turcicus). — Global Ecol. Biogeogr. 18: 674–687. Rodriguez, A. and Delibes, M. 1992. Current range and status of the Iberian lynx Felis pardina Temminck, 1824 in Spain. — Biol. Conserv. 61: 189-196. Rodrıguez, A. and Delibes, M. 2003. Population fragmentation and extinction in the Iberian lynx. — Biol. Conserv. 109: 321-331. Rodríguez, A. and Delibes, M. 2002. Internal structure and patterns of contraction in the geographic range of the Iberian lynx. — Ecography 25: 314-328. Rodriguez, J. P. 2002. Range contraction in declining North American bird populations. — Ecol. Appl. 12: 238-248. Root, T. 1988. Environmental Factors Associated with Avian Distributional Boundaries. — J. Biogeogr. 15: 489-505. Root, T. L. et al. 2003. Fingerprints of global warming on wild animals and plants. — Nature 421: 57-60. Runge, C. A. et al. 2015. Geographic range size and extinction risk assessment in nomadic species. — Conserv. Biol. 29: 865-876. Safi, K. and Pettorelli, N. 2010. Phylogenetic, spatial and environmental components of extinction risk in carnivores. — Global Ecol. Biogeogr. 19: 352-362. Sagarin, R. D. and Gaines, S. D. 2002. The 'abundant centre' distribution: to what extent is it a biogeographical rule? — Ecol. Lett. 5: 137-147. Sagarin, R. D. et al. 2006. Moving beyond assumptions to understand abundance distributions across the ranges of species. — Trends Ecol. Evol. 21: 524-530. Sala, O. E. et al. 2000. Global biodiversity scenarios for the year 2100. — Science 287: 1770- 1774. Scharlemann, J. P. W. et al. 2004. Land‐use trends in Endemic Bird Areas: global expansion of agriculture in areas of high conservation value. — Global Change Biol. 10: 2046– 2051. Schipper, J. et al. 2008. The status of the world's land and marine mammals: diversity, threat, and knowledge. — Science 322: 225-230. Scholl, J. P. and Wiens, J. J. 2016. Diversification rates and species richness across the Tree of Life. — Proceedings of the Royal Society B: Biological Sciences 283: 20161334. Seabrook, L. et al. 2014. Determining range edges: habitat quality, climate or climate extremes? — Divers. Distrib. 20: 95-106.

314

Bibliography

Sekercioglu, C. H. et al. 2004. Ecosystem consequences of bird declines. — Proceedings of the National Academy of Sciences USA 101: 18042-18047. Sekercioglu, C. H. et al. 2008. Climate change, elevational range shifts, and bird extinctions. — Conserv. Biol. 22: 140-150. Selwood, K. E. et al. 2014. The effects of climate change and land-use change on demographic rates and population viability. — Biological Reviews 90: 837-853. Sexton, J. P. et al. 2009. Evolution and ecology of species range limits. — Evolution and ecology of species range limits 40: 415-436. Shoo, L. P. et al. 2005. Potential decoupling of trends in distribution area and population size of species with climate change. — Global Change Biol. 11: 1469–1476. Simberloff, D. 1998. Small and declining populations. — In: Sutherland, W. J. (ed), Conservation science and action. Blackwell Science., pp. 116-134. Simberloff, D. S. and Abele, L. G. 1976. Island biogeography theory and conservation practice. — Science 191: 285-286. Skelly, D. K. et al. 1999. Long-term distributional dynamics of a Michigan amphibian assemblage. — Ecology 80: 2326-2337. Slatyer, R. A. et al. 2013. Niche breadth predicts geographical range size: a general ecological pattern. — Ecol. Lett. 16: 1104-1114. Smale, D. A. and Wernberg, T. 2013. Extreme climatic event drives range contraction of a habitat-forming species. — Proceedings of the Royal Society B: Biological Sciences 280: 20122829. Smissen, P. J. et al. 2013. Mountain barriers and river conduits: phylogeographical structure in a large, mobile lizard (Varanidae: Varanus varius) from eastern Australia. — J. Biogeogr. 40: 1729-1740. Soberon, J. 2007. Grinnellian and Eltonian niches and geographic distributions of species. — Ecol. Lett. 10: 1115-1123. Sodhi, N. S. et al. 2008. Correlates of extinction proneness in tropical angiosperms. — Divers. Distrib. 14: 1-10. Stanton, J. C. et al. 2014. Warning times for species extinctions due to climate change. — Global Change Biol. 21: 1066–1077. Steffen, W. et al. 2015a. The trajectory of the Anthropocene: The Great Acceleration. — The Anthropocene Review 2: 81-98. Steffen, W. et al. 2015b. Planetary boundaries: Guiding human development on a changing planet. — Science 347: 1259855. Strijker, D. 2005. Marginal lands in Europe—causes of decline. — Basic Appl. Ecol. 6: 99- 106. Sunday, J. M. et al. 2011. Global analysis of thermal tolerance and latitude in ectotherms. — Proceedings of the Royal Society Biological Sciences 278: 1823-1830. Sunday, J. M. et al. 2012. Thermal tolerance and the global redistribution of animals. — Nature Climate Change 2: 686-690. Thomas, C. and Featherstone, W. 2005. Validation of Vincenty’s Formulas for the Geodesic Using a New Fourth-Order Extension of Kivioja’s Formula. — J. Surv. Eng. 131: 20- 26. Thomas, C. et al. 2006. Range retractions and extinction in the face of climate warming. — Trends Ecol. Evol. 21: 415-416. Thomas, C. D. et al. 2008. Where within a geographical range do species survive best? A matter of scale. — Insect Conserv. Divers. 1: 2-8. Thomas, C. D. et al. 2004. Extinction risk from climate change. — Nature 427: 145-148. 315

Bibliography

Thomas, C. D. et al. 2011. A framework for assessing threats and benefits to species responding to climate change. — Methods Ecol. Evol. 2: 125-142. Thompson, K. et al. 1999. Range size, dispersal and niche breadth in the herbaceous flora of central England. — J. Ecol. 87: 150-155. Thornton, P. K. et al. 2014. Climate variability and vulnerability to climate change: a review. — Global Change Biol. 20: 3313-3328. Thuiller, W. et al. 2005a. Niche properties and geographical extent as predictors of species sensitivity to climate change. — Global Ecol. Biogeogr. 14: 347-357. Thuiller, W. et al. 2005b. Climate change threats to plant diversity in Europe. — Proceedings of the National Academy of Sciences USA 102: 8245-8250. Tian, Y. Q. et al. 2001. Estimating solar radiation on slopes of arbitrary aspect. — Agricultural and Forest Meteorology 109: 67-74. Tingley, M. W. et al. 2009. Birds track their Grinnellian niche through a century of climate change. — Proceedings of the National Academy of Sciences USA 106: 19637- 19643. Travis, J. M. J. 2003. Climate change and habitat destruction: a deadly anthropogenic cocktail. — Proc. R. Soc. Lond., Ser. B: Biol. Sci. 270: 467-473. Turvey, S. T. et al. 2015. Historical data as a baseline for conservation: reconstructing long- term faunal extinction dynamics in Late Imperial–modern China. — Proc. R. Soc. B 282: 20151299. Turvey, S. T. et al. 2016. Holocene range collapse of giant muntjacs and pseudo‐endemism in the Annamite large mammal fauna. — J. Biogeogr. doi: 10.1111/jbi.12763. Tylianakis, J. M. et al. 2008. Global change and species interactions in terrestrial ecosystems. — Ecol. Lett. 11: 1351-1363. Tzedakis, P. C. et al. 2002. Buffered tree population changes in a Quaternary refugium: evolutionary implications. — Science 297: 2044-2047. Urban, M. C. 2015. Accelerating extinction risk from climate change. — Science 384: 571- 573. Urban, M. C. et al. 2016. Improving the forecast for biodiversity under climate change. — Science 353: Václavík, T. and Meentemeyer, R. K. 2012. Equilibrium or not? Modelling potential distribution of invasive species in different stages of invasion. — Divers. Distrib. 18: 73-83. van Kleunen, M. et al. 2010. A meta-analysis of trait differences between invasive and non- invasive plant species. — Ecol. Lett. 13: 235-245. Venette, R. C. et al. 2010. Pest risk maps for invasive alien species: a roadmap for improvement. — Bioscience 60: 349-362. Vilà, M. et al. 2011. Ecological impacts of invasive alien plants: a meta‐analysis of their effects on species, communities and ecosystems. — Ecol. Lett. 14: 702-708. Vilà, M. and Ibáñez, I. 2011. Plant invasions in the landscape. — Landscape Ecol. 26: 461- 472. Villavicencio, N. A. et al. 2016. Combination of humans, climate, and vegetation change triggered Late Quaternary megafauna extinction in the Última Esperanza region, southern Patagonia, Chile. — Ecography 39: 125-140. Vincenty, T. 1975. Direct and inverse solutions of geodesics on the ellipsoid with application of nested equations. — Survey Review 22: 88-93. Warren, M. S. et al. 2001. Rapid responses of British butterflies to opposing forces of climate and habitat change. — Nature 414: 65-69. 316

Bibliography

Watson, J. et al. 2016. Catastrophic Declines in Wilderness Areas Undermine Global Environment Targets. — Curr. Biol. doi: 10.1016/j.cub.2016.08.049. Webb, T. J. and Gaston, K. J. 2000. Geographic range size and evolutionary age in birds. — Proceedings of the Royal Society Biological Sciences 267: 1843-1850. Weigelt, P. et al. 2015. Global patterns and drivers of phylogenetic structure in island floras. — Scientific reports doi: 10.1038/srep12213. Whittaker, R., J. et al. 2001. Scale and species richness: towards a general, hierarchical theory of species diversity. — J. Biogeogr. 28: 453-470. Whittaker, R. H. 1956. Vegetation of the Great Smoky Mountains. — Ecol. Monogr. 26: 1- 69. Wiens, J. J. et al. 2010. Niche conservatism as an emerging principle in ecology and conservation biology. — Ecol. Lett. 13: 1310-1324. Wilson, D. E. and Reeder, D. M. 2005. Mammal species of the world: a taxonomic and geographic reference. — Johns Hopkins University Press. Wilson, R. J. et al. 2005. Changes to the elevational limits and extent of species ranges associated with climate change. — Ecol. Lett. 8: 1138-1146. Wilson, R. J. et al. 2004. Spatial patterns in species distributions reveal biodiversity change. — Nature 432: 393-396. Williams, Y. M. et al. 2006. Niche breadth and geographical range: ecological compensation for geographical rarity in rainforest frogs. — Biol. Lett. 2: 532-535. Willis, K. J. and Bhagwat, S. A. 2009. Biodiversity and Climate Change. — Science 326: 806-807. World Bank 2011. Research and development expenditure (% of GDP) Wright, S. J. and Hubbell, S. P. 1983. Stochastic extinction and reserve size - A focal species apporach. — Oikos 41: 466-476. Yackulic, C. B. et al. 2011. Anthropogenic and environmental drivers of modern range loss in large mammals. — Proceedings of the National Academy of Sciences USA 108: 4024-4029. Yu, F. et al. 2016. Climatic niche breadth can explain variation in geographical range size of alpine and subalpine plants. — International Journal of Geographical Information Science doi: 10.1080/13658816.2016.1195502. 1-23. Zar, J. H. 1999. Biostatistical analysis, 4/e. — Pearson Education India. Zeidberg, L. D. and Robison, B. H. 2007. Invasive range expansion by the Humboldt squid, Dosidicus gigas, in the eastern North Pacific. — Proceedings of the National Academy of Sciences USA 104: 12948–12950.

317

318

Agradecimientos Agradecimientos

Todavía recuerdo como si fuese ayer el día en el que estando trabajando en casa de mis padres con los modelos de hábitat de la cabra montesa y sonó el móvil, al descolgarlo dijeron desde el otro lado “Hola, soy Eloy Revilla de la Estación Biológica de Doñana”. ¡Menuda sorpresa, la persona de la que había leído tantos trabajos sobre el tejón, el lince y otros carnívoros estaba al otro lado!! Había intentado contactar con él anteriormente para hablar sobre su proyecto que tenía asociada una FPI pero había sido imposible. El título de ese proyecto me sonaba un poco a chino “Factores mecanísticos que afectan a la vulnerabilidad de especies”. En esa charla supo explicarme la temática del proyecto, las oportunidades y la flexibilidad de la tesis, las ventajas del grupo y a la par que me ocultaba los futuros quebraderos de cabeza de trabajar con extensas bases de datos y con programación. Eloy muchas gracias por esa llamada, por confiar en mí, por los ánimos y el apoyo durante estos años. Desde ese momento, y tras algunos contratiempos iniciales empezó mi andadura por la Estación Biológica de Doñana. La primera visita no pudo empezar de mejor forma, recuerdo que vinieron a buscarme mi directora Manuela y Clara a la estación de tren en aquel mítico todoterreno portugués y fuimos a casa de Eloy donde habían organizado una barbacoa. Si hubiera tenido alguna duda (que no era el caso) con esto se hubiera terminado de disipar. Allí conocí a gran parte del grupo de carnívoros. En primer lugar tengo que agradecer el apoyo que siempre he tenido de mis padres, me han apoyado mental y ecónomamente con los estudios y cuando he tenido momentos difíciles ellos seguían confiando y animándome para continuar. Aunque este apoyo ha sido unánime dentro de la familia, también les debo mucho a mis tíos Juan y Lourdes, que son como unos padres para mí y también han estado animándome y ayudándome en todo lo posible. Y por supuesto a mi familia nazarena Baena- que siempre estaban preocupándose por cómo estaba y con los que he disfrutado muchos días de cenas y de la feria de Dos Hermanas de la que guardo un bonito recuerdo. A todos los integrantes del grupo de carnívoros, a Clara, Miguel Delibes, Miguel Ángel Rodríguez, Miguel Clavero, Marcello, Miguel Jácome, Laura, Javitxu, y a los que me habré dejado muchísimas gracias. Pero hay una persona muy especial dentro de ese grupo sin el cual sería todo un kaos, estoy hablando de Sofi, que siempre soluciona todo y tiene una sonrisa en la cara, muchísimas gracias por todo Sofi. A todos aquellos con los que he compartido estos años despacho, a Carol, a Andrea, a Alberto, a Estrella, a Rubén, a Isa, a Bego, Sete muchísimas gracias y siento haberos hecho soportar un clima polar durante estos años. Y a una persona que le tengo un cariño especial, a mi hermanita, a Ester Polaina, perdón Doctora Polaina, que tantos momentos hemos pasado odiando a los programadores de Esri, diciéndole miles de improperios al ordenador/arcmap y por todos los ratitos que hemos pasado fuera en festivales, de tapeo, etc, muchísimas gracias, te deseo lo mejor. Muchas gracias al Ministerio de Competitividad por la beca de formación de Personal Investigador y por las dos ayudas para estancias breves en centros de investigación extranjeros. Many thanks to the British Ecological Society for the grant to attend to the European Ecological Federation Meeting. Muchas gracias al personal de la Estación 319

Agradecimientos

Biológica de Doñana que aportan su granito de arena y hacen que todo funcione, en especial a los servicios informáticos. Al equipo del Laboratorio de Sistemas de Información Geográfica y teledetección, LAST-EBD, David, Isabel y Ricardo, por los quebraderos de cabeza y por ayudarme a solucionar mis problemas, muchísimas gracias. A Eduardo Narbona, mi tutor que me ha ayudado siempre con todo los trámites administrativos y me ha mandado su apoyo. Muchísimas gracias a mi comité de tesis, Miguel Delibes, Luís Brotons y Miguel Ángel Olalla-Tárraga, que me recomendó y aconsejó para que llegara a buen puerto la presente tesis. A todos los del grupillo de las tres, a los que siguen ahí y a los que no están, a Luismi, Echegaray, J, Edu. Sara (Gracias por el Endnote), Paloma, Irene, Jesús Gómez, Alazne, Antonio, Alex, Fran, Mari, Marina, Mar, Martina Vane, Miguel, Rafa, Hyeun-Ji Lee. Dicen que el roce hace el cariño y claro, tantos años comiendo en “El comedero” pues se os coge algo de cariño ;). Hemos pasado momentos estupendos tanto dentro como fuera, en conciertos, viajes exprés, cenitas, tapeo, etc. También hemos pasado momentos no tan buenos como aquel día en el que tras recalentar un pollo reseco estuve a punto de palmarla (no sé si hubiera muerto antes porque no podía respirar o por las collejas que me dabais); menos mal que apareció Pablo Villalba y ejecutó a la perfección la maniobra de Heimlich, muchas gracias por salvarme la vida. Muchas gracias a Rosita por ese cariño y ese apoyo, los cafés, las charlas, por todos los momentos que hemos pasado tanto dentro como fuera de la EBD, en viajes, conciertos, festivales, etc. eres una persona muy especial. Muchas gracias a Noa por su apoyo y cariño, porque es la parte más heavy y sensible del grupo, por aguantar todas las tonterías que hago/digo con su animal de compañía (no me refiero a Lua, bueno ni a las tortugas, bueno ni a Pepa). Y por supuesto muchísimas gracias a mi alma gemela, a Burrac, gracias por estar ahí animándome, por las bromas, por mi vuelta sano y salvo de festivales veraniegos, por compartir tantos y tantos momentos y por los que seguiremos compartiendo en un futuro, eres un tío muy grande, y gracias por depositar mi tesis. A toda la gente que he conocido de la EBD estos años, a Violeta, Ana Montero, Cande, Néstor, Pablo González, María, Alex, Álvaro, etc etc, seguro que me dejo a alguien!! A Juan Herrero, profesor durante el tercer curso de Ciencias ambientales, y que me ha dirigido el proyecto de introducción a la investigación, el proyecto fin de carrera y el proyecto de fin de máster, hemos compartido innumerables jornadas de campo con grandes madrugones y buenísimos almuerzos y que sobre todo es un gran amigo. Aunque no me ha dirigido la ha sido la persona con la que me inicié en el campo de la investigación, todo lo que he aprendido/hecho con él, sus consejos y ánimos han sido clave en primer lugar para la consecución de la beca y en segundo para el desarrollo de la tesis. Muchísimas gracias, espero que nuestra amistad perdure y que sigamos trabajando juntos. Many thanks to Kevin Gaston and Hugh Possingham for give me the opportunity to visit their research groups and for help with my research. And many thanks to the people from the Spatial Ecology Lab (University of Queensland), to Martina Di Fonzo, Salit, Andrea, Hsien-Yung Lin, Tara. A very special acknowledges to Viv & Corey, for their help for their friendliness and in summary because they are great persons, I wish all the best for you and for your little and gorgeous baby. 320

Agradecimientos

Una de las cosas que más me emociona cuando recuerdo estos años son todos los ratos que he pasado con el Coco´s Team. Muchos de vosotros sabéis que muchas veces no podía acudir o acudía muy tarde a los sitios porque estaba entrenando o porque estaba en estado comatoso después del entreno. El Coco´s Team es mucho más que un equipo de atletismo, es una gran familia encabezada por José Ceballos Coco, comúnmente conocido como “Coco”, al que le debo mucho y le admiro enormemente. Coco, antiguo atleta y actual sargento del cuerpo de bomberos de Sevilla dedica gran parte de su tiempo libre a entrenar a un grupo de atletas de Sevilla y alrededores por la mera recompensa de vernos hacer atletismo y disfrutar haciendo deporte, es de esas personas que hacen que tengamos un mundo mejor. Durante estos años entrenando hemos pasado estupendos ratos y tanto Coco como el resto de compañeros han sabido transmitir los valores que representa el deporte: esfuerzo, constancia, compañerismo y afán de superación. Con todos tengo recuerdos muy especiales, los chistes de Migue y la réplica de Fran “Es bueno eh, es bueno”, Bea regateándole a Coco “2,5 kg menos que es que no puedo” cuando está levantando 80 kg, cuando a Migue se le partió la goma haciendo supervelocidad “shhhhh ¡¡plash!!”, me duele a mi sólo de pensarlo, la vez que Javi me ganó el 100 metros entrando a lo Ussain Bolt (¡¡me debes la revancha!!). Por supuesto no sólo los recuerdos se ciñen a los entrenos, las cenas “light” que hacíamos en el Pura-gula o en el hamburguesería tienen también muy gratos recuerdos, muchos incluyen posibles records de comida/bebida por parte de Marín, como aquella cena en la que te bebiste unos dos litros de fanta naranja. También recuerdo con mucho cariño la vez que fuimos a comprar mi moto y las clases que me disteis Migue y Javi. Por supuesto guardo muy buenos recuerdos de Alba, anteriormente conocida como la hermana de Javi, por todas esos desayunos en el Papelón, cuando hicimos el herbario, y demás charlas y cafés, porque eres mi moñas preferida muchas gracias (espero que te esté cayendo una lagrimita). A todo el equipo del Coco´s Team, a Fran, Marcos, Plata, Bachero, Isco, Ricardo, Mude, Edu, Álvaro, a la pequeña Irene que me animaba diciéndome que ya no corría como un robot, a Beli, por permitirnos monopolizar a Coco y preocuparse por nosotros y en especial a Coco por ser como es, a todos os deseo lo mejor y os agradezco enormemente todos esos momentos. Y por supuesto, muchas gracias a una persona muy especial, a Silvia, por todos esos buenos momentos, por todos los ánimos en los momentos difíciles, eres una chica genial, muchísimas gracias. Finalmente, a la persona que sin duda más se merece esta sección que es mi directora Manuela, siempre ha estado animándome, ha tenido tiempo para mí y ha sabido guiarme y dirigirme a la perfección, y ha invertido muchísimo tiempo en corregir y supervisarme. Pero Manuela ha sido y es mucho más que una directora de tesis, siempre ha estado ahí cuando he tenido algún problema o necesitaba ayuda para algo Manuela muchísimas gracias, y siento que te sacara de quicio algunas veces por algunos despistes, como las faltas en inglés o los comentarios en los textos ¡¡ya te libras de mi!!;). Muchas gracias y te deseo lo mejor.

Gracias a todos por estos maravillosos años

321