Accepted Manuscript Title: Using species distributions models for designing conservation strategies of Tropical Andean biodiversity under climate change Author: Juli´an Ram´ırez-Villegas Francisco Cuesta C. Christian Devenish Manuel Peralvo Andy Jarvis Carlos Arnillas PII: DOI: Reference: S1617-1381(14)00038-7 http://dx.doi.org/doi:10.1016/j.jnc.2014.03.007 JNC 25349 To appear in: Received date: Revised date: Accepted date: 14-6-2013 24-3-2014 24-3-2014 Please cite this article as: Ram´ırez-Villegas, J., Cuesta C., F., Devenish, C., Peralvo, M., Jarvis, A., & Arnillas, C.,Using species distributions models for designing conservation strategies of Tropical Andean biodiversity under climate change, Journal for Nature Conservation (2014), http://dx.doi.org/10.1016/j.jnc.2014.03.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 TITLE 2 Using species distributions models for designing conservation strategies of Tropical Andean 3 biodiversity under climate change 6 Julián Ramírez-Villegas1, 2, 3, * 7 Francisco Cuesta C.4 8 Christian Devenish5, 6 9 Manuel Peralvo4 Carlos Arnillas7 12 AFFILIATIONS 13 1 Colombia, AA6713 2 16 17 Colombia, AA6713 3 18 19 Decision and Policy Analysis (DAPA), International Center for Tropical Agriculture (CIAT) Cali, Ac ce p 15 CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS), Cali, te 14 an 11 M Andy Jarvis1, 2 d 10 cr AUTHORS us 5 ip t 4 Institute for Climate and Atmospheric Science (ICAS), School of Earth and Environment, University of Leeds, Leeds, LS2 9JT, UK 4 20 Biodiversity Department - Consorcio para el Desarrollo Sostenible de la Ecorregión Andina (CONDESAN) 21 5 BirdLife International – Americas Secretariat 22 6 School of Science and the Environment, Manchester Metropolitan University, UK 23 7 Centro de Datos para la Conservación, Universidad Agraria La Molina 24 * Corresponding author: j.r.villegas@cgiar.org 1 Page 1 of 47 ABSTRACT 26 Biodiversity in the Tropical Andes is under continuous threat from anthropogenic activities. 27 Projected changes in climate will likely exacerbate this situation. Using species distribution 28 models, we assess possible future changes in the diversity and climatic niche size of an 29 unprecedented number of species for the region. We modeled a broad range of taxa (11,012 30 species of birds and vascular plants), including both endemic and widespread species and 31 provide a comprehensive estimation of climate change impacts on the Andes. We find that if no 32 dispersal is assumed, by 2050s, more than 50% of the species studied are projected to undergo 33 reductions of at least 45% in their climatic niche, whilst 10% of species could be extinct. Even 34 assuming unlimited dispersal, most of the Andean endemics (comprising ~5% of our dataset) 35 would become severely threatened (>50% climatic niche loss). While some areas appear to be 36 climatically stable (e.g. Pichincha and Imbabura in Ecuador; and Nariño, Cauca, Valle del Cauca 37 and Putumayo in Colombia) and hence depict little diversity loss and/or potential species gains, 38 major negative impacts were also observed. Tropical high Andean grasslands (páramos and 39 punas) and evergreen montane forests, two key ecosystems for the provision of environmental 40 services in the region, are projected to experience negative changes in species richness and high 41 rates of species turnover. Adapting to these impacts would require a landscape-network based 42 approach to conservation, including protected areas, their buffer zones and corridors. A central 43 aspect of such network is the implementation of an integrated landscape management approach 44 based on sustainable management and restoration practices covering wider areas than currently 45 contemplated. 46 Keywords: Andes, biodiversity, conservation, climate change, threats, climatic niche, maxent Ac ce p te d M an us cr ip t 25 47 2 Page 2 of 47 48 49 1. Introduction 51 Despite ambitious goals to significantly reduce the rate of biodiversity loss by 2010 (CBD, 52 2007), biodiversity continues to be severely threatened (Ramirez-Villegas et al., 2012; Sachs et 53 al., 2009). These threats include over exploitation of natural resources (e.g. water, agricultural 54 soils), habitat loss and degradation, and invasive species (Butchart et al., 2010; Kim and Byrne, 55 2006). Biodiversity loss has been increasing since the second half of the 20th century, and is 56 likely to continue into the future (Kim and Byrne, 2006; MEA, 2005). With climate change 57 entailing likely increases in temperature and regional and seasonal changes in precipitation 58 (Knutti and Sedlacek, 2013), ecosystems and their services are likely to suffer additional stresses 59 (Chen et al., 2009; Feeley and Silman, 2010; Fuhrer, 2003; IPCC, 2007). cr us an M d te 60 ip t 50 The Tropical Andes tops the list of worldwide hotspots for species diversity and endemism 62 (Fjeldså et al., 1999; Gentry, 1995; Sklenár and Ramsay, 2001). For this reason, the region is 63 considered a key priority for biodiversity conservation (Brooks et al., 2006; Myers et al., 2000). 64 At the same time, the Tropical Andes have been identified as one of the most severely threatened 65 natural areas globally (Jetz et al., 2007; Mittermeier et al., 1997). During the last century, 66 concentration of human population and associated demands for goods and services in the inter- 67 Andean valleys and the inner slopes of the Andean ridges, has transformed a significant portion 68 of the natural landscape causing habitat loss and degradation followed by species extinction and 69 disruption of ecosystem functions (e.g. water-flow regulation), especially in the Northern Andes 70 (Bruinsma, 2003; Wassenaar et al., 2007; Armenteras et al., 2011; Rodriguez et al., 2013). Ac ce p 61 3 Page 3 of 47 Resource-base over-exploitation of natural resources has led to a severe land degradation process 72 (Podwojewski et al., 2002; Poulenard et al., 2001, 2004), increasing the pressure on the goods 73 and services provided by these ecosystems (Rundel and Palma, 2000). In addition, the Andes are 74 expected to undergo severe stresses over the next 100 years as a result of climate change 75 (Beaumont et al., 2011; Malcolm et al., 2006). ip t 71 cr 76 Addressing potential impacts from climate change is important because the environmental 78 impacts of human activities (Biesmeijer et al., 2006; MEA, 2005) could be exacerbated by the 79 likely rapid changes in the climate system during the 21st century (IPCC, 2007; Knutti and 80 Sedlacek, 2013). Warren et al. (2013) estimated that, in the absence of any climate change 81 mitigation strategy, large range contractions for ca. 60 % of plants and 35 % of animals could be 82 expected globally. Understanding and quantifying the extent at which climate change could 83 threaten Andean species is therefore critical since many of the species in the region occur in low 84 dense populations with narrow distribution patterns (i.e. endemics) with a high level of 85 replacement within the environmental gradients. These characteristics make the Andean biota 86 particularly sensitive to climate change disruptions. an M d te Ac ce p 87 us 77 88 Our primary objective was to assess the likely impacts of climate change on the distributions of 89 vascular plant and bird species of the Tropical Andes. Using species distributions modelling 90 techniques, we assessed the potential climatic niche of 11,012 species, and then projected them 91 under the SRES-A2 emission scenario for two periods: 2020 and 2050. Future projected changes 92 in species assemblages, including richness, turnover and range size were assessed. Lastly, the 4 Page 4 of 47 93 projected impacts in selected groups of species of Andean origin were analysed. Finally, we 94 discuss future strategies to reduce expected biodiversity loss. 95 2. Study area 97 The study area (Tropical Andes hereafter) comprises all interconnected areas above altitudes of 98 500 m within the countries of Venezuela, Colombia, Ecuador, Peru and Bolivia, plus the Sierra 99 Nevada de Santa Marta in Colombia, delimited using data from the SRTM digital elevation 100 model (Farr et al. 2007). Extending over 1.5 million km2 from 11º N to 23º S, the Tropical Andes 101 are the longest and widest mountain region in the tropics (Figure 1) (Clapperton, 1993; Fjeldså 102 and Krabbe, 1990). The morphological and bioclimatic heterogeneity of the Andes have led to 103 the formation of an enormous diversity of microhabitats favouring speciation (Mittermeier et al., 104 1997; Young et al., 2002). Moreover, their location between the lowlands of the Amazon, La 105 Chiquitanía and El Chaco to the east and the Chocó, Tumbes-Guayaquil and the arid systems of 106 the Sechura Desert to the west, has created complex dynamics of species exchange and isolation 107 (Bass et al., 2010; Young et al., 2002). The Tropical Andes harbours more than 45,000 vascular 108 plant (20,000 endemics) and 3,400 vertebrate species (1,567 endemics) in just 1 percent of the 109 Earth’s land mass (Lamoreux et al., 2006; Olson et al., 2001). cr us an M d te Ac ce p 110 ip t 96 111 3. Methods 112 We modelled the climatic niches of 11,012 species (1,555 birds and 9,457 plants) using species 113 distributions models. We modelled the climate-constrained present-day distributions of all 114 species, and projected them onto two different future periods (2020s, 2050s) and two contrasting 115 dispersal scenarios. The approach implemented here aims to evaluate the likely impacts of 5 Page 5 of 47 climate change on the widest array possible of Andean plant and bird species by mid 2020s and 117 mid 2050s and comprises the following six steps: 118 1. Assembling of species occurrence data 119 2. Generation of climate surfaces 120 3. Maximum entropy species distribution modeling 121 4. Analysis of projected climate change impacts on species assemblages 122 5. Delineation of conservation recommendations for the 2020s and 2050s us cr ip t 116 123 3.1 Species datasets 125 Presence data for 11,012 species (1,555 birds and 9,457 plants) were sourced from three 126 databases. CONDESAN, the Centro de Datos para la Conservación de la Universidad Nacional 127 Agraria La Molina (CDC-UNALM), and a previous global study (Warren et al., 2013) (W2013). 128 From the three sources, we extracted all occurrences in the five tropical Andean countries (i.e. 129 Venezuela, Colombia, Ecuador, Peru and Bolivia) of all vascular plant clades (Magnoliophyta, 130 Pteridophyta, Pinophyta, Psilophyta, Cycadophyta, Gnetophyta, Lycopodiophyta) and bird (class 131 Aves, phylum Chordata) species with at least one record within the study area (Figure 1B). By 132 including these three sources of data we ensured the inclusion of common and widespread 133 species (see Warren et al. 2013) as well as narrow-range Andean endemics and imperil species 134 (also see Sect. 4.1 for details). Ac ce p te d M an 124 135 136 CONDESAN’s database consisted of data from multiple sources. Vascular plant specimen data 137 were obtained from the Missouri Botanical Garden's Vascular Tropicos (VAST) nomenclatural 138 database (Garden, 2004), the Herbarium of the National Science Institute in Colombia (ISN) and 6 Page 6 of 47 the Catholic University Herbarium (QCA) in Ecuador. Bird species data were obtained from 140 databases belonging to the Chicago Field Museum of Natural History, Academy of Natural 141 Sciences of Philadelphia, California Academy of Sciences and the Berkeley Museum of Natural 142 History and cross-checked with BirdLife International database (version 2012). Additional data 143 were obtained from private databases (Juan Fernando Freile for Antpittas, Paul Hamec for 144 Dendroica cerulea; Cal Dodson-Lorena Endara for orchid’s records and James Luteyn's database 145 stored at the New York Botanical Garden site for Ericaceae) and published literature (Casares et 146 al., 2003; Renjifo et al., 2002; Schuchmann et al., 2001). The CDC-UNALM database was 147 produced from the review of papers and reports during the last 25 years. It also comprises field 148 reports obtained by its own research as well as data provided by other national (i.e. Peruvian) 149 researchers. The W2013 database was originally sourced from the Global Biodiversity 150 Information Facility (GBIF, available at http://data.gbif.org). Warren et al. (2013) thoroughly 151 checked the GBIF plant and animal database for location errors following the methodology of 152 Ramirez-Villegas et al. (2012), whereby the consistency of the location data is verified at both 153 geographic (using coastal and country borders) and environmental (using outlier-removal tests) 154 levels. We carefully checked bird species names using BirdLife’s taxonomy database as a 155 reference. Plant taxonomy was verified using The Plant List (http://www.theplantlist.org, see 156 Warren et al., 2013). cr us an M d te Ac ce p 157 ip t 139 158 3.2 Climate data 159 Current climate data were derived from WorldClim (Hijmans et al., 2005). WorldClim is a 160 global gridded dataset of monthly climatological means of maximum, minimum and mean 161 temperature and total precipitation developed through Thin Plate Spline interpolation of long- 7 Page 7 of 47 term (i.e. 1950-2000) weather station records (Figure 1A). There is a generally dense distribution 163 of weather stations across the core of our geographic analysis domain (Hijmans et al., 2005). 164 Using the monthly WorldClim data we derived 10 ‘bioclimatic’ indices (Busby, 1991; Rivas- 165 Martinez, 2004) (Table 1). These indices describe annual and seasonal trends and allow for an 166 adequate characterization of the species bioclimatic niches. These indices are important limiting 167 factors for growth and development of species, and have been used extensively for predicting 168 species distributions using presence-only data (Elith et al., 2006; Graham et al., 2008; Warren et 169 al., 2013). For the Andes, the 10 bioclimatic indices chosen cover aspects of both average and 170 extreme conditions of a year. In addition, the use of the ombrothermic index allows for 171 differentiating climate conditions between and across ecosystems (Rivas-Martinez, 2004). an us cr ip t 162 M 172 173 d te 174 [Table 1 here] We obtained future climate projections from the CMIP3 (Coupled Model Inter-comparison 176 Project phase 3) web data portal (https://esg.llnl.gov:8443/index.jsp) (Meehl et al., 2007). We 177 downloaded monthly time series of temperature and precipitation data for the baseline period 178 (20th century) and projections of future climate for the 21st century for the SRES-A2 emission 179 scenario for 24 different Intergovernmental Panel on Climate Change (IPCC) coupled GCMs 180 (Table 2). We chose SRES-A2 because we considered the full-mitigation SRES-B1 unlikely, and 181 because differences between SRES-A2 and SRES-A1B and SRES-A1FI by 2050s are negligible 182 (Hawkins and Sutton, 2009). Based on the availability of maximum and minimum temperature 183 data, we further selected a subset of nine GCMs (Table 2). Ac ce p 175 184 8 Page 8 of 47 185 [Table 2 here] 186 Using the complete GCM time series, for each of the GCMs, months and variables, we 188 calculated the 30 year running average over the baseline period (1961-1990) and two future 189 periods: 2020s (2010-2039) and 2050s (2040-2069), representing the early and mid- 21st century. 190 We then calculated the anomalies (deltas) of each GCM future scenario with respect to the 191 baseline period (average 1961-1990 climate) for each month, variable and period. us cr ip t 187 192 Given the significant heterogeneity in Andean climates, coarse scale GCM grids fail to represent 194 the diversity of niches where species are distributed, hence we increased the resolution of the 195 GCM data by means of empirical downscaling with the delta method (Ramirez-Villegas and 196 Jarvis, 2010). For each month, variable, and period, the respective set of GCM deltas was 197 averaged (i.e. ensemble mean). Temperature anomalies were directly added, whilst precipitation 198 anomalies were added as a relative factor to the value in WorldClim in order to avoid 199 precipitation values below zero due to the differences between the GCM simulated and 200 WorldClim observed baseline. For each of the future periods, we calculated the same bioclimatic 201 indices as for current climate data (Table 1). This yielded climate scenarios for each of the future 202 periods as an average trend of the set of available GCMs on the SRES-A2 emission scenario. M d te Ac ce p 203 an 193 204 We used the ensemble mean (rather than individual GCMs) owing to processing and storage 205 needs, and given the considerable number of species being modelled and the resolution at which 206 the models were projected (2.5 arc-min). 207 9 Page 9 of 47 3.3 Species distribution models (SDMs) 209 Species distributions were modelled using Maxent (Phillips et al., 2006; Phillips and Dudík, 210 2008), a robust bioclimatic envelope modelling techniques (Smith et al., 2013). We modelled 211 only species with at least 10 distinct locations (Ramírez-Villegas et al., 2010; Wisz et al., 2008), 212 as a compromise between model quality and sufficient coverage of limited-range species. 213 Maxent models the climate-constrained distribution of a species using presence-only data and a 214 set of environmental descriptors (Elith et al., 2010; Phillips et al., 2006). Maxent has been tested 215 extensively and has been found to suitably perform as a state-of-the-art modelling technique both 216 under current and future conditions (Costa et al., 2010; Phillips, 2008; Smith et al., 2013). an us cr ip t 208 217 Here, we followed a similar methodology to that employed by Warren et al. (2013), whereby 219 default features optimised to broad species groups were used to construct Maxent models for 220 each species (Phillips, 2008; Phillips et al., 2006; Phillips and Dudík, 2008). For each species we 221 drew 10,000 pseudo-absences from the countries where the species was reported (according to 222 our database). This was done to avoid over-fitting of the models whilst maintaining a good 223 discrimination between presence and absence of the species (Isaac et al., 2009; VanDerWal et 224 al., 2009). d te Ac ce p 225 M 218 226 Most niche modeling techniques are sensitive to the number of predictors used and Maxent is no 227 exception (Braunisch et al., 2013; Dormann, 2007; Phillips, 2008). Excess predictors in a Maxent 228 model can cause over-fitting and hence bias the responses under future scenarios by over- 229 weighting certain drivers over others (Warren and Seifert, 2010). Hence, following Warren et al. 230 (2013), we reduced the number of predictors in the Maxent model for species with low numbers 10 Page 10 of 47 of occurrences. For those species with < 40 unique data points, a set of six climate predictors was 232 used (i.e. P1, P4, P12, P15, Io and Iod2), whilst for taxa with > 40 unique data points, the 233 complete set of 10 predictors (i.e. P1, P4, P5, P6, P12, P15, P16, P17, Io and Iod2) was used. 234 This choice was a compromise between having overly-complex Maxent models for species with 235 low numbers of occurrences and having overly-simplistic models for species with very large 236 numbers of occurrences. cr ip t 231 us 237 Maxent models were fitted using cross-validation (10 iterations), each one randomly dropping 239 10-20% input points. We then assessed the model skill using the Area under the ROC (Receiver 240 Operating Characteristic) Curve of the test data (AUCTest), calculated as the average AUCTest of 241 the 10 runs. Despite known limitations (Lobo et al., 2008; Warren et al., 2013), AUCTest is a 242 useful metric for selecting Maxent models of appropriate complexity (Warren and Seifert, 2010) 243 and is a widely used model accuracy and selection criterion (Braunisch et al., 2013; Graham et 244 al., 2008; VanDerWal et al., 2009). The procedure applied here allowed us to discard species 245 with models showing low predictive skill: only models with 10-fold average test AUCTest ≥ 0.7 246 were projected onto the future climatic periods. M d te Ac ce p 247 an 238 248 We then projected the fitted models onto both the continuous WorldClim current climate 249 surfaces and the downscaled surfaces of future climate conditions (2020s and 2050s). We then 250 binned the probability distributions using the ‘prevalence threshold’ (Liu et al., 2005; 2013). This 251 threshold is defined as the average probability over all input data points used to fit the model (i.e. 252 training presence points). To reduce commission (i.e. straying too far from the actual niche of a 253 taxon) or omission (i.e. missing major species populations due to lack of observations), the 11 Page 11 of 47 254 current climate distributions of each species were further clipped within a 300 km buffer around 255 the respective input occurrence points (also see Warren et al. 2013). 256 For future climatic scenarios, species distribution maps were first binned using the prevalence 258 threshold, and then further limited using two assumptions about species’ dispersion mechanisms 259 (Jarvis et al., 2008; Thomas et al., 2004; Thuiller et al., 2005): (1) no dispersal and (2) unlimited 260 dispersal. For the no dispersal scenario, the projected future distributions were not allowed to 261 stray away from the current-climate distribution. For the unlimited dispersal scenario, all future 262 suitable areas outside the current-climate distribution were considered of the future distribution. 263 This implies that a species can migrate and occupy any new site that becomes suitable under 264 future climatic conditions. We acknowledge that unlimited dispersal is unrealistic (particularly 265 for plants), but we use this scenario to illustrate the likely impacts of climate change on diversity 266 even when the best possible conditions are assumed (e.g. through use of assisted migration, also 267 see Sect. 5.3). cr us an M d te Ac ce p 268 ip t 257 269 3.4 Assessment of climate change impacts in species assemblages 270 Species richness was calculated using the binned species distributions as the total number of 271 species in a given site (i.e. pixel) and then used to calculate changes in species richness as the 272 difference between future species richness and current species richness divided by current 273 species richness. Additionally, we calculated the species turnover for the unlimited dispersal 274 scenario (Broennimann et al., 2006). This index arises from a modification of the ‘classical’ 275 species turnover (beta-diversity) indicators (Lennon et al., 2001; Whittaker, 1960) which are 12 Page 12 of 47 276 computed in geographic space using a defined spatial neighbourhood (Broennimann et al., 2006) 277 (Eq. 1). 278 species turnover = 100 * species gain + species loss initial species richness + species gain cr 280 [Equation 1] ip t 279 This turnover index has a lower limit of zero when the ‘species gain’ and the ‘species loss’ are 282 zero (both of which are very unlikely to happen with a large set of species), and an upper limit of 283 100, when the whole set of species changes from one time period to the other (i.e. either the 284 species gain or loss equals the initial species richness and there is no loss or gain respectively). an us 281 M 285 3.5 Assessment of individual species responses to climate change 287 To estimate the sensitivity to climate change at the species level for both migration scenarios and 288 periods, we intersected the current and future climatic niches and calculated the climatic niche 289 persistence. This is defined as the percentage of area that remains suitable in relation to the total 290 area in the current climatic niche (Loehle and LeBlanc, 1996; Peterson et al., 2001). Climatic 291 niche loss and gain were first calculated as the percentage area predicted to become unsuitable or 292 suitable respectively in the future climatic niche in relation to the total area in the current 293 climatic niche (Broennimann et al., 2006). The species range change was then calculated as the 294 difference between climatic niche gain and loss. This represents the percentage of range 295 expansion or contraction in relation to the current climatic niche for each species under the future 296 scenarios. Ac ce p te d 286 297 298 4. Results 13 Page 13 of 47 4.1 Species datasets 300 Our final modelling dataset comprised 478,301 vascular plant occurrences for 9,457 species and 301 88,636 bird occurrences for 1,555 species (Figure 1B). The W2013 dataset provided the greatest 302 proportion of occurrences, with 93% of all locality points used, and holding data for 9,371 303 vascular plants species and 1,429 birds. The database from CDC-UNALM provided 4.14% of the 304 occurrence points used for 186 vascular plant and 1,316 bird species. CONDESAN’s dataset 305 contributed 2.9% of the occurrences representing 501 birds and 237 vascular plants. Despite the 306 majority of records were from the W2013, the CDC-UNALM and CONDESAN datasets 307 provided critical occurrence data for rare, endemic and narrow-range species that were poorly (if 308 at all) represented in the W2013 database (see e.g. Supplementary Figure S1 in Warren et al. 309 2013). M an us cr ip t 299 312 [Figure 1 here] te 311 d 310 4.2 Performance of species distribution models 314 Almost half of the plant (48%) and bird species (44%) had an average test AUC > 0.9, 315 suggesting a good aptitude of the models to discriminate the species’ fundamental climatic niche. 316 The average test AUC of all plant species was 0.874 (median = 0.894, SD = 0.088), while that of 317 bird species was 0.872 (median = 0.889, SD = 0.076) (Figure 2). Cross-validated runs indicated 318 that variability of AUC ranged from 0 to 13.7% for training-sets and from 0 to 38.8% for 319 evaluation sets. Relatively unstable test statistics were found for species with very low number of 320 data points (high variability in AUC across repetitions), both in training and test sets. Ac ce p 313 321 14 Page 14 of 47 322 [Figure 2 here] 323 Maxent models performance as measured by the average AUC was relatively similar for birds 325 (BD) and vascular plants (VP), on average (Figure 2). Average training VP AUC ranged from 326 0.433 to 0.999, whilst test AUC varied from 0.28 to 0.999. In a few cases (< 500 for plants and < 327 50 for birds) the AUC statistic fell below the 0.7 threshold for model quality, probably owing to 328 a combination of a limited number of species records and an asymmetric spatial distribution (i.e. 329 high spatial autocorrelation). Less than 1 % of the whole set of plant and bird species had an 330 AUC value equal to or worse than random discrimination of presences and absences (AUC ≤ 331 0.5). All species with average test AUC below 0.7 were removed from any further analyses (see 332 Sect. 3.3.1). Based on a sufficiently high AUC (i.e. > 0.7), a total of 9,062 vascular plant and 333 1,456 bird species (95.7 and 96.6% respectively) were used in all following analyses. d M an us cr ip t 324 te 334 4.3 Shifts in species richness and community turnover 336 Current species richness ranged from 0 to 452 species for birds and from 0 to 1,535 species for 337 vascular plants per pixel of 25 km2 (Figure 3). The highest concentration of plants is located on 338 the outer slopes of the Western and Eastern Andean chain, between 1,500 to 3,000 m in altitude, 339 primarily in the Andes of Colombia, Ecuador and Venezuela as well as on the inner slopes of the 340 Central Chain of Colombia (upper Magdalena river basin) (Figure 3A). Diversity of birds is 341 particularly high throughout the Peruvian Andes, in the montane forests along the Eastern ridge 342 (Range = 141-452), and in the montane forests of the north-western chain of Ecuador (Figure 343 3B). Ac ce p 335 344 15 Page 15 of 47 345 [Figure 3 here] 346 Patterns of changes in species richness show important differences depending on the dispersal 348 thresholds and the period analysed (2020 or 2050). The unlimited dispersal scenario projects an 349 upslope migration of both plant and bird species suggesting important changes in the 350 configuration of the diversity patterns of Andean biota. On the other hand, the no-dispersal 351 scenarios show a significant reduction in species richness for both plant and bird species with 352 major changes by 2050. The maximum richness values in the no dispersal scenario by 2050 353 period are 1,244 for plant species (mean = 163 ± 178) and 295 for birds (mean = 29 ± 36) per 25 354 km2 pixel (Figure 4). Areas showing the largest decreases in species richness are located along 355 the montane forests of the Eastern Andes of Bolivia and Peru between 500 and 1,200 m, on the 356 outer slopes of the Eastern Andean foothills in Colombia and Ecuador, and on the Pacific slope 357 of Northern Ecuador and southern Colombia (Figure 4). Conversely, the areas with minor 358 changes are the highlands of Peru and Bolivia (Altiplano) and the pacific slope of the Peruvian 359 Andes. cr us an M d te Ac ce p 360 ip t 347 361 Negative changes in species richness are also observed even when unlimited dispersal is 362 considered. Loss of diversity is observed from north to south of the Andes, although some 363 particular areas are worthy of more attention; areas below altitudes of 1,500 m in the east 364 Peruvian Andean mountains (i.e. central and eastern Huanuco, Pasco and Junin) seem to be 365 severely impacted (>60% loss in species richness), and the same pattern is observed in the border 366 between Ecuador and Peru, and in Nariño, Valle del Cauca, and Putumayo in Colombia. These 16 Page 16 of 47 367 changes may be attributed to the eastern margins of the mountain chain being less climatically 368 suitable in warmer climates. 369 370 ip t [Figure 4 here] 371 The projected changes in community turnover are concentrated to a large extent in the High 373 Andes of Bolivia and Peru, as well as in the foothills of the Sierra de la Macarena, Sierra Nevada 374 de Santa Marta and around the Magdalena river basin in Colombia. Significant shifts are also 375 evident in the Venezuelan Andes along the Merida chain (Figure 5). an us cr 372 376 377 M [Figure 5 here] d 378 4.4 Individual species responses 380 Increases are projected in average climatic niche size for all species under the unlimited dispersal 381 assumptions for the 2020s period (Figure 6A). As expected, more severe impacts are projected 382 for the 2050s, and this is reflected in a less pronounced increase of range size in the unlimited 383 dispersal scenario and a stronger decrease in the non-dispersal scenario (Figure 6A, B). 384 Considering an unlimited dispersal scenario, the rates of climatic niche expansion seem to be 385 high, with most of the species being highly favoured or barely affected by climate change if 386 migration in fact occurs and other non-abiotic factors remain stable (e.g. land-use patterns, pests 387 and diseases), particularly for birds. Some 45% (n=655) of bird and 41% (n=3,715) of vascular 388 plant species modelled are likely to experience an increase in their climatic niches of 100% or 389 more by 2050s (Figure 6A). By contrast, only a limited proportion of species (< 10 %) is Ac ce p te 379 17 Page 17 of 47 expected to experience no increase or a net loss in their climatic niche size. Our estimates 391 indicate that even assuming unlimited dispersal some species are expected to undergo range 392 contraction (even to the extent of extinction), thus highlighting specific sensitivities to climate 393 change. ip t 390 394 395 cr [Figure 6 here] us 396 In a no dispersal scenario, the differences between periods become more evident (Figure 6B). 398 Whilst by 2020s the maximum changes in range size are reductions of 50% and 80% for birds 399 and vascular plants, respectively, by the 2050s, species within both groups are projected to 400 experience 100% range reduction, indicating likely extinctions for a vast number of species. M an 397 d 401 To illustrate species-specific responses under future climate, we further selected and analysed 403 two contrasting genera for each species group (plants and birds). These genera were selected 404 because they are of relatively recent origin (during the Pleistocene, ca. 1 to 3 million years ago), 405 include species that are endemic to the Andes, and are classified vulnerable or critically 406 endangered by IUCN (Table 3 and 4). Many of the species of the genera Grallaria and 407 Eriocnemis (class: Aves) are projected to expand their niche by more than 100 % if dispersal was 408 assumed. In particular, the species E. cupreoventris and E. nigrivestis were found to increase 409 their niche considerably by 2020 and 2050. In the case of no-dispersal, however, these species 410 depict range contractions of 69 and 65 % (respectively) by 2050. Similar responses were found 411 for most species of the genus Grallaria, notably G. alleni, G. aplotona, G. gigantea, and G. Ac ce p te 402 18 Page 18 of 47 412 hypoleuca, for which range contractions of 59, 83, 54, and 63 % are projected by 2050s (no 413 dispersal), respectively (Table 3). 414 Similar responses are reported for the plant genera Polylepis and Gynoxis. Species such as P. 416 lanuginosa and P. tomentela showed significant increases in range size in both future scenarios 417 (unlimited migration), but rather large decreases in range size under no-migration assumptions. 418 By contrast, some species of these genera (e.g. P. incana, P. reticulate, G. buxifolia, and G. 419 caracensis) report range contractions for both dispersal scenarios and periods (Table 4). These 420 species that respond negatively even under when unlimited dispersal is allowed can be 421 considered of very high sensitivity, and perhaps also be prioritised for further research to 422 understand such sensitivities. M an us cr ip t 415 d 423 5. Discussion 425 5.1 Changes in species distribution patterns 426 Our results suggest that impacts of climate change over the Andean biota could be extremely 427 severe. This finding is in agreement with previous studies for the Andean region (Feeley and 428 Silman, 2010; Feeley et al., 2011ab; Tovar et al., 2013), other tropical areas (Hole et al., 2009; 429 Miles et al., 2004; Still et al., 1999), or globally (Warren et al., 2013). The effects of climate 430 change on the Tropical Andes can be synthesized at two different levels: the extent of the whole 431 Tropical Andes (regional level), and at the species level. At the regional level, the inner and 432 outer Andean foothills (800 – 1,500 meters) are likely to be the most affected due to a high 433 amount of species loss. In addition, the spatial patterns of species turnover demonstrate a 434 bimodal response. First, an upslope shift of several species from mid elevations to the high Ac ce p te 424 19 Page 19 of 47 Andes is expected. Second, a large west and southward displacement of species from the upper 436 areas of the northern portion of the study area (i.e. Merida, Perijá and Santa Marta) towards 437 lower latitudes and a significant climatic niche reduction of mountain-top endemics is also 438 projected. ip t 435 439 The areas that would be most affected by high absolute species turnover rates and the subsequent 441 change in the composition of communities are the montane dry forest, the Santa Marta massif, 442 the Mérida ridge, the inner slopes of the Central and Eastern ridges of the Colombian Andes and 443 the Altiplano of Peru and Bolivia (> 3,800 meters). an us cr 440 444 At the species level, the biophysical impacts of exposure to climate change are projected to be 446 highly variable. In this study, the two contrasting dispersal scenarios show extremes of a 447 spectrum of projected responses by species to climate change. For plants, it is likely that the true 448 response lies nearer the no-dispersion scenario (see also Feeley et al., 2011a), whereas for birds 449 the response may in some cases resemble that of the full-dispersion scenario. Overall, we report 450 that plant species may be more negatively affected in both magnitude and direction of range 451 change impacts than birds in both periods. The same pattern holds for both migration scenarios, 452 probably due to a greater proportion of endemic and narrow-range plant species and/or the 453 presence of isolated (meta) populations (Figure 6A) (also see Ramirez-Villegas et al. 2012), and 454 perhaps to some extent also due to incompleteness of samples for some species. Yet species 455 interactions might have a prominent role in this point. For example, species interactions can slow 456 climate tracking and produce more extinctions than predicted by climatic niche models only 457 (Urban et al. 2013); or on the contrary, broad-ranging animals might transport seeds enabling Ac ce p te d M 445 20 Page 20 of 47 458 long-distance dispersal, as documented before during the last de-glaciation period, in which trees 459 dispersed at rates of 100-1000 m year-1 (Clark, 1998). 460 The projected alteration of the spatial distribution patterns of Andean assemblages (Feeley and 462 Silman, 2010; Feeley et al., 2011a; Jetz et al., 2007) suggest the appearance of novel 463 communities adapted to non-analogous climatic conditions, which could affect the functioning of 464 Andean ecosystems (Williams and Jackson, 2007). Many shrubby and epiphyte species (e.g. 465 Solanaceae, Bromeliaceae) depend on their specialized symbiotic interactions with animals for 466 seed dispersion and pollination. Climate change effects on these organisms could cause spatial, 467 temporal, or physiological asynchronies between mutualistic species, producing changes in 468 community composition and structure (Zavaleta et al., 2003). 470 Our estimates are thus useful in gauging general trends and possible impacts, although it is very 471 likely that individual responses at the species or community level will be determined by species’ 472 ecological traits (i.e. dispersal capacity), species interactions (i.e. competition) and/or by their 473 physiological response to stresses, leading (in some cases) to different outcomes. If species are 474 sufficiently mobile they may be able to track the geographic displacement of their climatic 475 niches, or if species are capable of rapid evolutionary change or have a wide range of abiotic 476 tolerances, they may adjust to changing ecological conditions and landscapes (Broennimann et 477 al., 2006). According to Travis (2003) and Opdam and Wascher (2004), the exact nature of a 478 species’ response to different rates of climate change depends upon colonization ability and how 479 much of a generalist the species is. For species with lower colonization ability and for specialist 480 species, the threshold occurs at a lower climate change signal. In a human dominated world, Ac ce p te d 469 M an us cr ip t 461 21 Page 21 of 47 however, natural or semi-natural ecosystems are embedded in tracts of unsuitable landscape, and 482 populations of species restricted to those habitat types are spatially dissected. By consequence, 483 what is ascribed as a shifting species range is in fact the complex result of extinction of (meta) 484 populations at the warm range limit (that surpasses thresholds of species adaptability), and 485 colonization and growth of (meta) populations into regions that newly came within the cold 486 range limit (that enters the range of species adaptability). Hence, for understanding the potential 487 risks of climate change to a species, we must consider the dynamics of the populations 488 constituting the geographical range in connection to the spatial features of the landscapes across 489 the range (also see Sect. 5.3). Human land-use may be especially important in the Andes where 490 anthropogenic activities above tree line and in the piedmont may create a hard barrier to upward 491 migrations, imperilling Andean biodiversity (Feeley et al. 2010; 2011a); therefore, the 492 incorporation of a coupled model that integrates climate change scenarios together with land 493 cover change dynamics is a priority task to analyse specific responses of the Andean biota to 494 these drivers of change. cr us an M d te Ac ce p 495 ip t 481 496 5.2 Species extinction risks 497 Climatic fluctuations during the Pliocene-Pleistocene period strongly influenced the origin and 498 spatial arrangement of the majority of Andean species used in this study (Luteyn, 2002; Young et 499 al., 2002; Garcia-Moreno et al. 1999). During periods of intense climatic change in the 500 Pleistocene, epiphyte-laden evergreen vegetation remained only where conditions remained 501 stable, suggesting that ecologically stable areas may have existed during the glaciations as small 502 pockets within surrounding drier pieces of montane forest (Fjeldså, 1995; Roy et al., 1997; 503 Arctander and Fjeldså, 1997). As a consequence, many of these surviving species present in 22 Page 22 of 47 these ecosystems are endemic, with narrow habitat tolerances in conjunction with a restricted 505 distribution range (Kattan et al., 2004). These patterns and conditions constitute a perfect 506 scenario to promote higher rates of species loss and turnover under projected climate anomalies 507 such as those projected in the present study. ip t 504 508 In this context, reductions in the size of the climatic niche such as those herein projected imply 510 that a number of species may become restricted to a few sites. Species with small range sizes are 511 vulnerable to smaller stochastic events as these could affect a larger proportion of the species’ 512 total population, especially in fragmented landscapes (With and King, 1999). As a result of this, 513 extinction risks will likely intensify for a large portion of the taxa analysed here, particularly at 514 long lead times (2050s in this study). Our study, as many others, assumes that species will die 515 out within regions that are predicted to become climatically unsuitable for them (Ohlemüller et 516 al., 2006), and takes no account of species- or population-level adaptive responses that may 517 reduce negative effects (see e.g. Harte et al., 2004). Despite that, our results may be conservative 518 given that we (1) did not include habitat loss data for the Tropical Andes in the analysis (Leisher 519 et al., 2013; Ramirez-Villegas et al., 2012), (2) did not consider potential impacts of changing 520 interannual variability (e.g. frequency or intensity of drought or heat waves) in our models, and 521 (3) did not model any secondary effects such as pests, diseases or important species-level 522 interactions required for survival. Furthermore, the rather low generation times of many vascular 523 plants and some bird species will probably preclude adaptation rates from keeping pace with 524 human induced climate change. Ac ce p te d M an us cr 509 525 526 5.3 Management and conservation implications 23 Page 23 of 47 In conservation planning, irreplaceability (commonly measured as singularity) and vulnerability 528 (measured through threat processes) are among the most important dimensions to analyse 529 (Brooks et al., 2006). Several authors have depicted the Tropical Andes as being within the most 530 vulnerable regions with high irreplaceability (Brooks et al., 2006; Kattan et al., 2004; 531 Mittermeier et al., 1997), placing the region extremely important for conservation action. ip t 527 cr 532 The question of whether the current protected area system is sufficient given the challenges of 534 climate change is a critical one. A regional analysis by Ramirez-Villegas et al. (2012) showed 535 that 8 out of 16 conservation areas in South America are in the Andean highlands. According to 536 the present study, negatively impacted areas (orange to red areas in Figure 4) could lose up to 537 60% of species richness and suffer up to 100% changes in community makeup, thus, affecting 538 ecosystem functioning as well as ecosystem services to human society (Gamfeldt et al., 2008). 539 There is no question that these projected impacts will affect conservation planning during the 21st 540 century, and hence further research should focus on developing a better understanding of 541 conservation effectiveness under future climates for the Andes (Araujo et al., 2004). Tropical 542 mountain systems such as the Andes are highly variable in climate, and therefore, offer a wide 543 range of adaptation pathways for species, further increasing their value for conservation. The 544 herein projected changes in range sizes, species richness and community composition are useful 545 metrics in evaluating tools for conservation, such as for adjusting extinction risk assessments, 546 delimitation of priority conservation areas and conservation targets within protected areas. Ac ce p te d M an us 533 547 548 Using these results to identify priority areas at a medium to large scale could be particularly 549 useful, given that diversity cannot always be easily captured in a single site-specific targeting of 24 Page 24 of 47 conservation in the Andes, requiring instead, conservation actions spread throughout entire 551 biomes (Fjeldså et al., 2005; Ramirez-Villegas et al., 2012). In this context, based on Opdam and 552 Wascher (2004) we propose three major components for a conservation strategy in a warmer 553 Tropical Andes. Firstly, a focus on landscape conditions for biodiversity, where populations 554 potentially can respond to large-scale changes and disturbances. These conditions should allow 555 populations to respond to large-scale disturbances. If species distributions patterns change more 556 dynamically in space and time, local conservation management for single species will be less 557 effective. Secondly, we propose to shift in strategy from protected areas towards landscape 558 networks including protected areas, connecting zones and intermediate landscapes. Thirdly, we 559 propose a shift from a defensive conservation strategy towards a landscape development 560 strategy. A static approach of establishing isolated reserves surrounded by a highly unnatural 561 landscape is not an effective strategy under a climate change scenario. Given the intense land use 562 changes in the Andes, the sensitivity of Andean species to climatic changes, and the fact we are 563 globally already committed to at least +2 °C warming, we must accept that conservation of 564 biodiversity is only effective if we dynamically integrate it in the development of the entire 565 landscape, based on coalitions with other functions such as the identification of key areas for 566 provision of ecosystem services, heterogeneity, and landscape permeability (Brooks et al., 2006). cr us an M d te Ac ce p 567 ip t 550 568 Regional policy and planning should aim at improving landscape connectivity. Amongst the 569 most evident conservation planning strategies is the establishment of reserves. Particularly under 570 climate change, the inclusion of new areas seems to be a relevant, albeit challenging, task 571 (Hannah et al., 2007). Land tenure issues, poverty, development gaps between rural and urban 572 areas, the demand for natural resources, and an economic model oriented toward extraction (e.g. 25 Page 25 of 47 mining) make the establishment of new conservation areas difficult in the Andes. In the absence 574 of such possibilities, the appropriate articulation of national reserves with other conservation 575 sub-systems such as protective forests, indigenous territories, civil society reserves, and sub- 576 national protected areas could be an appropriate mechanism of action. In addition, significant 577 attention should be paid to the design (or adjustment) of the Andean protected area system. We 578 recommend the following criteria be taken into account: cr ip t 573 • Maintain the connectivity across the elevation, moisture and edaphic gradient (Killeen 580 and Solórzano, 2008). These gradients are critical for maintaining beta diversity and 581 response capacity (Thuiller et al., 2008). an us 579 • Incorporate ecotone diversity in the design of conservation areas. The landscapes within 583 these areas are characterized by habitat mosaics that reflect differences in soil humidity, 584 productivity, among others. These mosaics are occupied by species assembled in 585 communities that reflect the presence of micro-environmental constraints in an area 586 where climate stress is the overriding macro-environmental characteristic. These 587 populations may have genetic traits distinct from core populations pre-adapting them to 588 the physiological stress of climate change (Killen and Solórzano 20008). In the Tropical 589 Andes the preservation of the ecotone between the montane forest and grasslands 590 ecoystems is a fundamental adaptation measure to buffer the massive upward 591 displacement of species ranges in response to increased warming (Feeley et al. 2011b). 592 • The identification of climatically stable areas as potential biological refugia through 593 bioclimatic envelope model (see e.g. dark green areas in Figure 4 combined with dark 594 areas in Figure 3) which could act as connectors and/or corridors between current and 595 future areas of high biodiversity (Vos et al., 2008). Ac ce p te d M 582 26 Page 26 of 47 596 Improvement of landscape connectivity through the creation of biological corridors is probably 598 the most frequent recommendation in the scientific literature (Heller and Zavaleta, 2009). We 599 suggest an optimisation of spatial configuration of such corridors and an assessment of the risks 600 of these turning into channels for disease transmission and/or movement of invasive species. In 601 addition to these, a better land use planning through better and targeted government-level 602 policies is warranted in order to reduce the risks of deforestation, loss of pollination services and 603 genetic erosion in the agricultural frontier, while at the same time bolstering the dispersion and 604 population breeding between (and within) remaining habitat patches (Opdam and Wascher, 605 2004). an us cr ip t 597 M 606 5.4 Final remarks 608 Several sources of uncertainty may influence the results we provide here. These include the 609 primary biodiversity data, the climate data and the climate envelope modeling (Braunisch et al., 610 2013; Pearson et al., 2006; Ramirez-Villegas and Challinor, 2012). Although these uncertainties 611 are carried into the analysis, we argue that our results provide important insight on a globally 612 important biodiversity hotspot. Importantly, our results agree and partly complement with 613 previous regional and global studies (see Warren et al. 2013; Still et al., 1999; Thomas et al., 614 2004; Feeley and Silman, 2010). Improvement to our modeling approach for future studies may 615 be warranted through achieving better spatial representativeness of both species and climate 616 observations, the use of abundance data (in addition to presence-only data), better constraining 617 species migration patterns, the inclusion of changes interannual variability and their effects on 618 species distributions, the use of higher resolution climate models that resolve local climatic Ac ce p te d 607 27 Page 27 of 47 619 change patterns in a more detailed manner, as well as a detailed assessment of relevant local 620 processes driving extinctions. 621 Acknowledgments 623 The authors thank Héctor Tobón and Daniel Amariles, from the International Center for Tropical 624 Agriculture (CIAT) for their help in programming the automated data cleansing algorithms. 625 Authors also thank Johannes Signer, from the International Center for Tropical Agriculture 626 (CIAT) for his help in some of the processing, and María Teresa Becerra and Wouter Buytaert 627 for their useful comments and improvement on earlier versions of this manuscript. This project 628 was funded by the Andean Regional Program of the Spanish Agency for International 629 Cooperation and Development (AECID), The Mountain Partnership and the Swiss Agency for 630 Development and Cooperation (SDC) through their regional program, ECOBONA and CIMA. 631 JRV and AJ were partly supported by the CGIAR Research Program on Climate Change, 632 Agriculture and Food Security (CCAFS). Authors thank two anonymous reviewers for their 633 comments. cr us an M d te Ac ce p 634 ip t 622 635 References 636 637 638 639 640 641 642 643 644 645 646 Araújo, M.B., Cabeza, M., Thuiller, W., Hannah, L., Williams, P.H. 2004. Would climate change drive species out of reserves? An assessment of existing reserve-selection methods. Global Change Biology, 10, 1618-1626. Arctander, P., Fjeldså, J., 1997. Andean Tapaculos of the Genus Scytalopus (Rhicocryptidae): A study of speciation using DNA sequence data, Birkhauser Verlag, Basel. Armenteras, D., Gast, F., Villareal, H., 2003. Andean forest fragmentation and the representativeness of protected natural areas in the eastern Andes, Colombia. Biological Conservation 113, 245-256. Armenteras, D., Rodríguez, N., Retana, J., Morales, M. 2011 Understanding deforestation in montane and lowland forests of the Colombian Andes. Regional Environmental Change, 11, 693-705. 28 Page 28 of 47 te d M an us cr ip t Barthlott, W., Hostert, A., Kier, G., Küper, W., Kreft, H., Mutke, J., Rafiqpoor, D., Sommer, J.H., 2007. Geographic Patterns of Vascular Plant Diversity at Continental and Global Scales. Erkunde 61, 305-315. Bass, M.S., Finer, M., Jenkins, C.N., Kreft, H., Cisneros-Heredia, D.F., McCracken, S.F., Pitman, N.C.A., English, P.H., Swing, K., Villa, G., Di Fiore, A., Voigt, C.C., Kunz, T.H., 2010. Global Conservation Significance of Ecuador's Yasuní National Park. PLoS ONE 5, e8767. Beaumont, L. J., Pitman, A., Perkins, S., Zimmermann, N.E., Yoccoz, N.G., Thuiller, W. 2011. Impacts of Climate Change on the World’s Most Exceptional Ecoregions. Proceedings of the National Academy of Sciences, 108 (6): 2306–2311. Biesmeijer, J.C., Roberts, S.P.M., Reemer, M., Ohlemüller, R., Edwards, M., Peeters, T., Schaffers, A.P., Potts, S.G., Kleukers, R., Thomas, C.D., Settele, J., Kunin, W.E., 2006. Parallel Declines in Pollinators and Insect-Pollinated Plants in Britain and the Netherlands. Science 313, 351-354. Bradshaw, H.D., Wilbert, S.M., Otto, K.G., Schemske, D.W., 1995. Genetic mapping of floral traits associated with reproductive isolation in monkeyflowers (Mimulus). Nature 376, 762-765. Braunisch, V., Coppes, J., Arlettaz, R., Suchant, R., Schmid, H., Bollmann, K., 2013. Selecting from correlated climate variables: a major source of uncertainty for predicting species distributions under climate change. Ecography, no-no. Broennimann, O., Thuiller, W., Hughes, G., Midgley, G.F., Alkemade, J.M.R., Guisan, A., 2006. Do geographic distribution, niche property and life form explain plants' vulnerability to global change? Global Change Biology 12, 1079-1093. Brooks, T.M., Mittermeier, R.A., da Fonseca, G.A.B., Gerlach, J., Hoffmann, M., Lamoreux, J.F., Mittermeier, C.G., Pilgrim, J.D., Rodrigues, A.S.L., 2006. Global Biodiversity Conservation Priorities. Science 313, 58-61. Bruinsma, J., 2003. World agriculture: Towards 2015/2030; An FAO perspective. Earthscan Publications Ltd., London. Busby, J.R., 1991. BIOCLIM—a bioclimate analysis and prediction system. Plant Protection Quarterly 6, 8-9. Butchart, S.H.M., Walpole, M., Collen, B., van Strien, A., Scharlemann, J.P.W., Almond, R.E.A., Baillie, J.E.M., Bomhard, B., Brown, C., Bruno, J., Carpenter, K.E., Carr, G.M., Chanson, J., Chenery, A.M., Csirke, J., Davidson, N.C., Dentener, F., Foster, M., Galli, A., Galloway, J.N., Genovesi, P., Gregory, R.D., Hockings, M., Kapos, V., Lamarque, J.F., Leverington, F., Loh, J., McGeoch, M.A., McRae, L., Minasyan, A., Morcillo, M.H., Oldfield, T.E.E., Pauly, D., Quader, S., Revenga, C., Sauer, J.R., Skolnik, B., Spear, D., Stanwell-Smith, D., Stuart, S.N., Symes, A., Tierney, M., Tyrrell, T.D., Vié, J.-C., Watson, R., 2010. Global Biodiversity: Indicators of Recent Declines. Science 328, 11641168. Casares, C., Boyla, K., Davidson, I., 2003. Threatened birds of the America – The ICBP/UICN Red Data Book. . Smithsonian Institution Press, Washington and London. In cooperation with the International Council for Bird Preservation, Cambridge, UK 1150 pp. CBD, 2007. Cross -Roads of Life on Earth -- Exploring means to meet the 2010 Biodiversity Target. Solution- oriented scenarios for Global Biodiversity Outlook 2, Montreal, Canada 90 pp. Ac ce p 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 29 Page 29 of 47 te d M an us cr ip t Chen, I.-C., Shiu, H.-J., Benedick, S., Holloway, J.D., Chey, V.K., Barlow, H.S., Hill, J.K., Thomas, C.D., 2009. Elevation increases in moth assemblages over 42 years on a tropical mountain. Proceedings of the National Academy of Sciences 106, 1479-1483. Clapperton, C.M., 1993. Quaternary geology and geomorphology of South America. Elsevier, Amsterdam and New York. Clark, J.S. 1998. Why trees migrate so fast: confronting theory with dispersal biology and the paleorecord. The American Naturalist, 152, 204-224. Costa, G., Nogueira, C., Machado, R., Colli, G., 2010. Sampling bias and the use of ecological niche modeling in conservation planning: a field evaluation in a biodiversity hotspot. Biodiversity and Conservation 19, 883-899. Devenish, C., Diaz-Fernandez, D.F., Clay, R.P., Davidson, I., Zabala, I.Y., 2009. Important Bird Areas of the Americas: Priority Sites for Biodiversity Conservation. BirdLife International, Cambridge, UK. Dormann, C.F., 2007. Promising the future? Global change projections of species distributions. Basic and Applied Ecology 8, 387-397. Elith, J., Graham, C., Anderson, R., Dudík, M., Ferrier, S., Guisan, A., Hijmans, R., Huettmann, F., Leathwick, J., Lehmann, A., Li, J., Lohmann, L., Loiselle, B., Manion, G., Moritz, C., Nakamura, M., Nakazawa, Y., Overton, J., Peterson, A.T., Phillips, S., Richardson, K., Scachetti-Pereira, R., Schapire, R., Soberón, J., Williams, S., Wisz, M., Zimmermann, N., 2006. Novel methods improve prediction of species distributions from occurrence data. Ecography 29, 129-151. Elith, J., Phillips, S.J., Hastie, T., Dudík, M., Chee, Y.E., Yates, C.J., 2010. A statistical explanation of MaxEnt for ecologists. Diversity and Distributions 17, 43-57. Feeley, K.J., 2010. The conservation value of secondary forests for tropical nocturnal bird species. Animal Conservation 13, 16-18. Feeley, K.J., Silman, M.R., 2010. Land-use and climate change effects on population size and extinction risk of Andean plants. Global Change Biology 16, 3215-3222. Feeley, K. J., Silman, M.R., Bush, M.B., Farfan, W., Garcia Cabrera, K., Malhi, Y., Meir, P., Salinas Revilla, N., Raurau Quisiyupanqui, M.N., Saatchi, S. 2011. Upslope Migration of Andean Trees. Journal of Biogeography 38, 783–791. Finer, M., Jenkins, C.N. 2012. Proliferation of Hydroelectric Dams in the Andean Amazon and Implications for Andes-Amazon Connectivity. PloS One, 7, e35126. Fjeldså, J., 1995. Geographical patterns of neoendemic and older relict species of Andean forest birds: The significance of ecologically stable areas, New York, USA. Fjeldså, J., Lambin, E., Mertens, B. 1999. Correlation between Endemism and Local Ecoclimatic Stability Documented by Comparing Andean Bird Distributions and Remotely Sensed Land Surface Data. Ecography, 22, 63-78. Fjeldså, J., Álvarez, M.D., Lazcano, J.M., León, B., 2005. Illicit Crops and Armed Conflict as Constraints on Biodiversity Conservation in the Andes Region. AMBIO: A Journal of the Human Environment 34, 205-211. Fjeldså, J., Krabbe, N., 1990. Birds of the High Andes: A Manual to the Birds of the Temperate Zone of the Andes and Patagonia, South America. Apollo Books, Copenhagen, Denmark 880 pp. Fjeldså, J., Lovett, J.C., 1997. Biodiversity and environmental stability. Biodiversity and Conservation 6, 315-323. Ac ce p 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 30 Page 30 of 47 te d M an us cr ip t Fuhrer, J., 2003. Agroecosystem responses to combinations of elevated CO2, ozone, and global climate change. Agriculture, Ecosystems & Environment 97, 1-20. Gamfeldt, L., Hillebrand, H., Jonsson, P.R., 2008. Multiple Functions Increase the Importance of Biodiversity for Overall Ecosystem Functioning. Ecology 89, 1223-1231. García-Moreno, J., Arctander, P., Fjeldså, J. (1999) Strong Diversification at the Treeline among Metallura Hummingbirds. The Auk, 116, 702-711. Garden, M.B., 2004. VAScular Tropicos (VAST) nomenclatural database in: Garden, M.B. (Ed.). Gentry, A.H., 1995. Patterns of diversity and floristic composition in Neotropical montane forests. New York Botanical Garden, Bronx. Graham, C.H., Elith, J., Hijmans, R.J., Guisan, A., Townsend Peterson, A., Loiselle, B.A., The Nceas Predicting Species Distributions Working, G., 2008. The influence of spatial errors in species occurrence data used in distribution models. Journal of Applied Ecology 45, 239-247. Hannah, L., Midgley, G., Andelman, S., Araújo, M., Hughes, G., Martinez-Meyer, E., Pearson, R., Williams, P., 2007. Protected area needs in a changing climate. Frontiers in Ecology and the Environment 5, 131-138. Hansen, D.M., Donlan, C.J., Griffiths, C.J., Campbell, K.J., 2010. Ecological history and latent conservation potential: large and giant tortoises as a model for taxon substitutions. Ecography 33, 272-284. Harrison, P.A., Berry, P.M., Butt, N., New, M., 2006. Modelling climate change impacts on species’ distributions at the European scale: implications for conservation policy. Environmental Science & Policy 9, 116-128. Harte, J., Ostling, A., Green, J.L., Kinzig, A., 2004. Biodiversity conservation: Climate change and extinction risk. Nature 430. Hawkins, E., Sutton, R., 2009. The Potential to Narrow Uncertainty in Regional Climate Predictions. Bulletin of the American Meteorological Society 90, 1095-1107. Heller, N.E., Zavaleta, E.S., 2009. Biodiversity management in the face of climate change: A review of 22 years of recommendations. Biological Conservation 142, 14-32. Henderson-Sellers, A., Gornitz, V., 1984. Possible climatic impacts of land cover transformations, with particular emphasis on tropical deforestation. Climatic Change 6, 231-257. Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., Jarvis, A., 2005. Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25, 1965-1978. Hole, D.G., Willis, S.G., Pain, D.J., Fishpool, L.D., Butchart, S.H.M., Collingham, Y.C., Rahbek, C., Huntley, B., 2009. Projected impacts of climate change on a continent-wide protected area network. Ecology Letters 12, 420-431. IPCC, 2007. IPCC Fourth Assessment Report: Climate Change 2007 (AR4). IPCC, Geneva, Switzerland. Isaac, J.L., Vanderwal, J., Johnson, C.N., Williams, S.E., 2009. Resistance and resilience: quantifying relative extinction risk in a diverse assemblage of Australian tropical rainforest vertebrates. Diversity and Distributions 15, 280-288. Jarvis, A., Lane, A., Hijmans, R.J., 2008. The effect of climate change on crop wild relatives. Agriculture, Ecosystems & Environment 126, 13-23. Ac ce p 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 31 Page 31 of 47 te d M an us cr ip t Jetz, W., Wilcove, D.S., Dobson, A.P., 2007. Projected Impacts of Climate and Land-Use Change on the Global Diversity of Birds. PLoS Biol 5, e157. Kattan, G.H., Franco, P., Rojas, V., Morales, G., 2004. Biological diversification in a complex region: a spatial analysis of faunistic diversity and biogeography of the Andes of Colombia. Journal of Biogeography 31, 1829-1839. Killeen, T.J., Solórzano, L.A., 2008. Conservation strategies to mitigate impacts from climate change in Amazonia. Philosophical Transactions of the Royal Society B: Biological Sciences 363, 1881-1888. Kim, K., Byrne, L., 2006. Biodiversity loss and the taxonomic bottleneck: emerging biodiversity science. Ecological Research 21, 794-810. Knutti, R., Sedlacek, J., 2013. Robustness and uncertainties in the new CMIP5 climate model projections. Nature Clim. Change 3, 369-373. Lamoreux, J.F., Morrison, J.C., Ricketts, T.H., Olson, D.M., Dinerstein, E., McKnight, M.W., Shugart, H.H., 2006. Global tests of biodiversity concordance and the importance of endemism. Nature 440, 212-214. Langhammer, P.F., Bakarr, M.I., Bennun, L.A., Brooks, T.M., Clay, R.P., Darwall, W., De Silva, N., Edgar, G.J., Eken, G., Fishpool, L.D.C., da Fonseca, G.A.B., Foster, M.N., Knox, D.H., Matiku, P., Radford, E.A., Rodrigues, A.S.L., Salaman, P., Sechrest, W., Tordoff, A.W., 2007. Identification and gap analysis of key biodiversity areas: targets for comprehensive protected area systems, Gland, Switzerland. Leisher, C., Touval, J., Hess, S., Boucher, T., Reymondin, L. 2013. Land and Forest Degradation inside Protected Areas in Latin America. Diversity 5, 779–795. Lennon, J.J., Koleff, P., GreenwooD, J.J.D., Gaston, K.J., 2001. The geographical structure of British bird distributions: diversity, spatial turnover and scale. Journal of Animal Ecology 70, 966-979. Liu, C., Berry, P.M., Dawson, T.P., Pearson, R.G., 2005. Selecting thresholds of occurrence in the prediction of species distributions. Ecography 28, 385-393. Liu, C., White, M., Newell, G., 2013. Selecting thresholds for the prediction of species occurrence with presence-only data. Journal of Biogeography 40, 778-789. Lobo, J.M., Jiménez-Valverde, A., Real, R., 2008. AUC: a misleading measure of the performance of predictive distribution models. Global Ecology and Biogeography 17, 145-151. Loehle, C., LeBlanc, D., 1996. Model-based assessments of climate change effects on forests: a critical review. Ecological Modelling 90, 1-31. Luteyn, J., 2002. Diversity, adaptation, and endemism in neotropical Ericaceae: biogeographical patterns in the Vaccinieae. The Botanical Review 68, 55-87. Malcolm, J.R., Canran, L., Neilson, R. P. Hansen, L., Hannah, L. 2006. Global Warming and Extinctions of Endemic Species from Biodiversity Hotspots. Conservation Biology 20 (2): 538–548. Mawdsley, J.R., O’Malley, R., Ojima, D.S., 2009. A Review of Climate-Change Adaptation Strategies for Wildlife Management and Biodiversity Conservation. Una Revisión de las Estrategias de Adaptación al Cambio Climático para el Manejo de Vida Silvestre y Conservación de la Biodiversidad. Conservation Biology 23, 1080-1089. MEA, 2005. Millennium Ecosystem Assessment: Ecosystems and human well-being: Biodiversity synthesis. World Resources Institute, Washington, DC, USA. Ac ce p 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 32 Page 32 of 47 te d M an us cr ip t Meehl, G.A., Covey, C., Taylor, K.E., Delworth, T., Stouffer, R.J., Latif, M., McAvaney, B., Mitchell, J.F.B., 2007. THE WCRP CMIP3 Multimodel Dataset: A New Era in Climate Change Research. Bulletin of the American Meteorological Society 88, 1383-1394. Miles, L., Grainger, A., Phillips, O. 2004. The impact of global climate change on tropical forest biodiversity in Amazonia. Global Ecology and Biogeography 13 (6): 553-565. Mittermeier, R.A., Gil, P.R., Mittermeier, C.G., 1997. Megadiversity: Earth’s Biologically Wealthiest Nations. Conservation International, Cemex. Mittermeier, R.A., Mittermeier, C.G., Brooks, T.M., Pilgrim, J.D., Konstant, W.R., da Fonseca, G.A.B., Kormos, C., 2003. Wilderness and biodiversity conservation. Proceedings of the National Academy of Sciences of the United States of America 100, 10309-10313. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853-858. Ohlemüller, R., Gritti, E.S., Sykes, M.T., Thomas, C.D., 2006. Towards European climate risk surfaces: the extent and distribution of analogous and non-analogous climates 1931– 2100. Global Ecology and Biogeography 15, 395-405. Olson, D.M., Dinerstein, E., 1998. The Global 200: A representation approach to conserving the Earth's most biologically valuable ecoregions. Conservation Biology 12, 502-505. Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D'amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, C.J., Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P., Kassem, K.R., 2001. Terrestrial Ecoregions of the World: A New Map of Life on Earth. BioScience 51, 933-938. Opdam, P., Wascher, D., 2004. Climate change meets habitat fragmentation: linking landscape and biogeographical scale levels in research and conservation. Biological Conservation 117, 285-297. Pearson, R.G., Thuiller, W., Araújo, M.B., Martinez-Meyer, E., Brotons, L., McClean, C., Miles, L., Segurado, P., Dawson, T.P., Lees, D.C., 2006. Model-based uncertainty in species range prediction. Journal of Biogeography 33, 1704-1711. Peterson, A.T., Sánchez-Cordero, V., Soberón, J., Bartley, J., Buddemeier, R.W., NavarroSigüenza, A.G., 2001. Effects of global climate change on geographic distributions of Mexican Cracidae. Ecological Modelling 144, 21-30. Phillips, S.J., 2008. Transferability, sample selection bias and background data in presence-only modelling: a response to Peterson et al. (2007). Ecography 31, 272-278. Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of species geographic distributions. Ecological Modelling 190, 231-259. Phillips, S.J., Dudík, M., 2008. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31, 161-175. Pielke, R.A., 2005. Land Use and Climate Change. Science 310, 1625-1626. Podwojewski, P., Poulenard, J., Zambrana, T., Hofstede, R., 2002. Overgrazing effects on vegetation cover and properties of volcanic ash soil in the páramo of Llangahua and La Esperanza (Tungurahua, Ecuador). Soil Use and Management 18, 45-55. Poulenard, J., Podwojewski, P., Janeau, J., Collinet, J. 2001. Runoff and soil erosion under rainfall simulation of Andisols from the Ecuadorian Páramo: effect of tillage and burning. CATENA, 45, 185-207. Ac ce p 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 33 Page 33 of 47 te d M an us cr ip t Poulenard, J., Michel, J.C., Bartoli, F., Portal, J.M., Podwojewski, P. 2004. Water repellency of volcanic ash soils from Ecuadorian páramo: effect of water content and characteristics of hydrophobic organic matter. European Journal of Soil Science, 55, 487-496. Ramirez-Villegas, J., Challinor, A., 2012. Assessing relevant climate data for agricultural applications. Agricultural and Forest Meteorology 161, 26-45. Ramirez-Villegas, J., Jarvis, A., 2010. Disaggregation of Global Circulation Model Outputs. Decision and Policy Analysis Working Paper No. 2, Decision and Policy Analysis Working Paper. International Center for Tropical Agriculture (CIAT), Cali, Colombia. Ramirez-Villegas, J., Jarvis, A., Touval, J., 2012. Analysis of threats to South American flora and its implications for conservation. Journal for Nature Conservation 20, 337-348. Ramírez-Villegas, J., Khoury, C., Jarvis, A., Debouck, D.G., Guarino, L., 2010. A Gap Analysis Methodology for Collecting Crop Genepools: A Case Study with Phaseolus Beans. PLoS ONE 5, e13497. Rivas-Martinez, S. 2004. Global Bioclimatics. Clasificación Bioclimática de la Tierra. Version 2004. Phytosociological Research Center. Available at http://www.globalbioclimatics.org Renjifo, L.M., Franco-Maya, A.M., Amaya-Espinel, J.D., Kattan, G., López-Lanús, B., 2002. Libro rojo de aves de Colombia. . Instituto de Investigaciones de Recursos Biológicos Alexander von Humboldt y Ministerio del Medio Ambiente, Bogota Colombia. Rivas-Martinez, S., 2004. Global Bioclimatics (Clasificación Bioclimática de la Tierra). Version 27-08-2004, p. 29. Rodríguez, E., Armenteras, P., Alumbreros. 2013. Land use and land cover change in the Colombian Andes: dynamics and future scenarios. Journal of Land Use Science, 8, 154174. Roy, M.S., da Silva, J.M.C., Arctander, P., Garcia-Moreno, J., Fjeldså, J., 1997. The speciation of South American and African birds in montane regions, in: Mindell, D.P. (ed.), Avian molecular evolution and systematics. Academic Press, San Diego, US. Rundel, P.W., Palma, B., 2000. Preserving the Unique Puna Ecosystems of the Andean Altiplano. Mountain Research and Development 20, 262-271. Sachs, J.D., Baillie, J.E.M., Sutherland, W.J., Armsworth, P.R., Ash, N., Beddington, J., Blackburn, T.M., Collen, B., Gardiner, B., Gaston, K.J., Godfray, H.C.J., Green, R.E., Harvey, P.H., House, B., Knapp, S., Kümpel, N.F., Macdonald, D.W., Mace, G.M., Mallet, J., Matthews, A., May, R.M., Petchey, O., Purvis, A., Roe, D., Safi, K., Turner, K., Walpole, M., Watson, R., Jones, K.E., 2009. Biodiversity Conservation and the Millennium Development Goals. Science 325, 1502-1503. Schuchmann, K.-L., Weller, A.-A., Heynen, I., 2001. Systematics and biogeography of the Andean genus Eriocnemis (Aves: Trochilidae). Journal of Ornithology 142, 433-481. Sklenár, P., Ramsay, P.M. 2001. Diversity of zonal páramo plant communities in Ecuador. Diversity and Distributions, 7, 113-124. Smith, A.B., Santos, M.J., Koo, M.S., Rowe, K.M.C., Rowe, K.C., Patton, J.L., Perrine, J.D., Beissinger, S.R., Moritz, C., 2013. Evaluation of species distribution models by resampling of sites surveyed a century ago by Joseph Grinnell. Ecography, no-no. Stern, N., 2006. The Economics of Climate Change: The Stern Review. Cambridge University Press, London 712 pp. Still, C.J., Foster, P.N., Schneider, S.H. 1999. Simulating the effect of climate change on tropical montane cloud forests. Nature 398, 608-610. Ac ce p 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 34 Page 34 of 47 te d M an us cr ip t Thomas, C.D., Cameron, A., Green, R.E., Bakkenes, M., Beaumont, L.J., Collingham, Y.C., Erasmus, B.F.N., de Siqueira, M.F., Grainger, A., Hannah, L., Hughes, L., Huntley, B., van Jaarsveld, A.S., Midgley, G.F., Miles, L., Ortega-Huerta, M.A., Townsend Peterson, A., Phillips, O.L., Williams, S.E., 2004. Extinction risk from climate change. Nature 427, 145-148. Thuiller, W., Albert, C., Araújo, M.B., Berry, P.M., Cabeza, M., Guisan, A., Hickler, T., Midgley, G.F., Paterson, J., Schurr, F.M., Sykes, M.T., Zimmermann, N.E., 2008. Predicting global change impacts on plant species' distributions: Future challenges. Perspectives in Plant Ecology, Evolution and Systematics 9, 137-152. Thuiller, W., Lavorel, S., Araújo, M.B., Sykes, M.T., Prentice, I.C., 2005. Climate change threats to plant diversity in Europe. Proceedings of the National Academy of Sciences of the United States of America 102, 8245-8250. Tovar, C., Arnillas, C.A., Cuesta, F., Buytaert, W., 2013. Diverging Responses of Tropical Andean Biomes under Future Climate Conditions. PLoS ONE 8, e63634. Travis, J.M.J., 2003. Climate change and habitat destruction: a deadly anthropogenic cocktail. Proceedings of the Royal Society of London. Series B: Biological Sciences 270, 467-473. UNFCCC, 2012. Conference of Parties: Report of the Conference of the Parties on its seventeenth session, held in Durban from 28 November to 11 December 2011, available at http://unfccc.int/resource/docs/2011/cop17/eng/09a02.pdf. Durban, South Africa. VanDerWal, J., Shoo, L.P., Graham, C., Williams, S.E., 2009. Selecting pseudo-absence data for presence-only distribution modeling: How far should you stray from what you know? Ecological Modelling 220, 589-594. Vos, C.C., Berry, P., Opdam, P., Baveco, H., Nijhof, B., O’Hanley, J., Bell, C., Kuipers, H., 2008. Adapting landscapes to climate change: examples of climate-proof ecosystem networks and priority adaptation zones. Journal of Applied Ecology 45, 1722-1731. Warren, D.L., Seifert, S.N., 2010. Ecological niche modeling in Maxent: the importance of model complexity and the performance of model selection criteria. Ecological Applications 21, 335-342. Warren, R., VanDerWal, J., Price, J., Welbergen, J.A., Atkinson, I., Ramirez-Villegas, J., Osborn, T.J., Jarvis, A., Shoo, L.P., Williams, S.E., Lowe, J., 2013. Quantifying the benefit of early climate change mitigation in avoiding biodiversity loss. Nature Clim. Change advance online publication. Wassenaar, T.D., Ferreira, S.M., Van Aarde, R.J., 2007. Flagging Aberrant Sites and Assemblages in Restoration Projects. Restoration Ecology 15, 68-76. Whittaker, R.H., 1960. Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs 30, 279-338. Williams, J.W., Jackson, S.T., 2007. Novel climates, no-analog communities, and ecological surprises. Frontiers in Ecology and the Environment 5, 475-482. Williams, J.W., Jackson, S.T., Kutzbach, J.E., 2007. Projected distributions of novel and disappearing climates by 2100 AD. Proceedings of the National Academy of Sciences 104, 5738-5742. Wisz, M.S., Hijmans, R.J., Li, J., Peterson, A.T., Graham, C.H., Guisan, A., Group, N.P.S.D.W., 2008. Effects of sample size on the performance of species distribution models. Diversity and Distributions 14, 763-773. With, K.A., King, A.W., 1999. Extinction Thresholds for Species in Fractal Landscapes Ac ce p 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 35 Page 35 of 47 Umbrales de Extinción para Especies en Paisajes Fraccionados. Conservation Biology 13, 314326. Young, K., Ulloa, C., Luteyn, J., Knapp, S., 2002. Plant evolution and endemism in Andean South America: An introduction. The Botanical Review 68, 4-21. Zavaleta, E.S., Shaw, M.R., Chiariello, N.R., Mooney, H.A., Field, C.B., 2003. Additive effects of simulated climate changes, elevated CO2, and nitrogen deposition on grassland diversity. Proceedings of the National Academy of Sciences 100, 7650-7654. ip t 961 962 963 964 965 966 967 968 969 Ac ce p te d M an us cr 970 36 Page 36 of 47 970 971 972 ip t 973 974 cr 975 Table 1 List of bioclimatic variables used in the modeling ID Variable name an P1 Annual mean temperature us 976 Units °C °C P5 Maximum temperature of warmest month °C P6 Minimum temperature of coldest month °C te P12 Annual precipitation d M P4 Temperature seasonality (standard deviation) P15 Precipitation seasonality (coefficient of variation) mm % mm P17 Precipitation of Driest quarter mm Ac ce p P16 Precipitation of Wettest quarter Io Ombrothermic index mm °C-1 Iod2 Ombrothermic index of the driest 2-months of the driest quarter mm °C-1 977 978 37 Page 37 of 47 Table 2 List of all and available GCMs and principal characteristics (resolutions) Model Country Atmosphere** Ocean** A2* Norway T63, L31 1.5x0.5, L35 CCCMA-CGCM3.1 (T47) Canada T47 (3.75x3.75), L31 1.85x1.85, L29 CCCMA-CGCM3.1 (T63) Canada T63 (2.8x2.8), L31 1.4x0.94, L29 CNRM-CM3 France T63 (2.8x2.8), L45 1.875x(0.5-2), L31 CSIRO-Mk3.0 Australia T63, L18 1.875x0.84, L31 A CSIRO-Mk3.5 Australia T63, L18 1.875x0.84, L31 A GFDL-CM2.0 USA 2.5x2.0, L24 1.0x(1/3-1), L50 A GFDL-CM2.1 USA 2.5x2.0, L24 1.0x(1/3-1), L50 A GISS-AOM USA 4x3, L12 GISS-MODEL-EH USA 5x4, L20 GISS-MODEL-ER USA 5x4, L20 IAP-FGOALS1.0-G China 2.8x2.8, L26 1x1, L16 INGV-ECHAM4 Italy T42, L19 2x(0.5-2), L31 INM-CM3.0 Russia 5x4, L21 2.5x2, L33 IPSL-CM4 France 2.5x3.75, L19 2x(1-2), L30 MIROC3.2-HIRES Japan T106, L56 0.28x0.19, L47 MIROC3.2-MEDRES Japan T42, L20 1.4x(0.5-1.4), L43 MIUB-ECHO-G Germany/Korea T30, L19 T42, L20 Germany T63, L32 1x1, L41 Japan T42, L30 2.5x(0.5-2.0) USA T85L26, 1.4x1.4 1x(0.27-1), L40 A NCAR-PCM1 USA T42 (2.8x2.8), L18 1x(0.27-1), L40 A UKMO-HADCM3 UK 3.75x2.5, L19 1.25x1.25, L20 UKMO-HADGEM1 UK 1.875x1.25, L38 1.25x1.25, L20 MRI-CGCM2.3.2A Ac ce p NCAR-CCSM3.0 cr us an M te MPI-ECHAM5 A ip t BCCR-BCM2.0 d 978 4x3, L16 5x4, L13 5x4, L13 A A 979 *A: Monthly maximum and minimum temperature available **Horizontal (T) resolution indicates number of cells 980 in which the globe was divided. Vertical (L) resolution indicates the number of layers in which the atmosphere was 981 divided. When a model is developed with different latitudinal and longitudinal resolutions, the respective cellsizes 982 (LonxLat) in degrees are provided instead of a unique value. 983 38 Page 38 of 47 983 984 985 987 ip t 986 Table 3 Change in distributional range for the Andean bird genera Eriocnemis and Grallaria. Species Endemic to Andes2 Elevation range (m)3 Eriocnemis alinae Eriocnemis cupreoventris LC NT - 2300-2800 1950-3000 Full -16.8 149.4 Null -23.3 -44.8 Full -32.6 101.2 Null -37.0 -68.6 Eriocnemis derbyi Eriocnemis luciani Eriocnemis mosquera Eriocnemis nigrivestis NT LC LC CR EC 2500-3600 2800-3800 1200-3600 1700-3500 -31.3 41.8 -17.8 261.4 -45.3 -13.6 -20.8 -30.0 18.0 -9.4 -34.0 92.1 -48.3 -30.3 -37.9 -65.0 Eriocnemis vestita LC - 2800-3500 8.4 -29.7 -1.7 -52.0 - 1800-2500 46.7 -31.5 3.1 -59.1 PE 2150-3000 1300-2350 38.8 50.9 -21.6 -16.7 -12.5 -8.4 -46.9 -47.6 VU B1a+b(i,ii,iii) - 1200-2600 > 500 -26.1 > 500 -54.0 LC - 200-3000 10.0 -31.3 2.4 -50.8 Grallaria gigantea Ac ce p Grallaria guatimalensis LC LC 2020 us an M d Grallaria erythroleuca Grallaria flavotincta VU B1a+b(i,ii,iii) te Grallaria alleni 988 989 990 991 992 993 994 cr Range change (%)3 IUCN 2010 category1 2050 Grallaria haplonota Grallaria hypoleuca Grallaria nuchalis LC LC LC - 700-2000 1400-2300 1900-3150 11.0 170.3 73.1 -55.1 -12.7 -10.3 -18.9 71.1 25.9 -82.7 -63.0 -36.4 Grallaria quitensis Grallaria ruficapilla LC LC - 2200-4500 1200-3600 -8.4 28.6 -38.2 -15.5 -48.5 18.1 -66.6 -35.0 Grallaria rufocinerea VU B1a+b(i,ii,iii) - 2200-3150 11.1 -31.1 60.7 -42.2 Grallaria rufula LC - 2300-3650 30.8 -25.9 10.4 -52.9 Grallaria squamigera Grallaria watkinsi LC LC - 2000-3800 600-1700 5.6 43.6 -26.0 -20.8 -21.8 33.7 -50.7 -49.9 1 Status of the species according to the IUCN red list of threatened species: LC: least concern, NT: nearthreatened, VU: vulnerable, EN: endangered, CR: critically endangered. Additional criteria as in http://www.iucnredlist.org/static/categories_criteria_3_1 2 Country where endemic, if endemic to the Andes. EC: Ecuador, PE: Peru, BO: Bolivia 3 Range change under different periods and for two dispersal scenarios. Full: unlimited dispersal, Null: no dispersal Species in bold depict range contractions (either by 2020 or 2050) regardless of migration assumptions. 39 Page 39 of 47 995 996 997 Table 4 Change in distributional range for the Andean plant genera Gynoxis and Polylepis. ip t 998 Range Change (%)3 Species IUCN 2010 category Endemic to Andes Elevation range (m) Gynoxis acostae Gynoxis asterotricha Gynoxis baccharoides LC n/a VU D(ii) EC - 2700-4300 3100-4100 3300-4200 Full > 500 > 500 233.3 Null -36.8 -21.0 -41.4 Full > 500 > 500 109.6 Null -84.0 -65.5 -69.2 Gynoxis buxifolia Gynoxis caracensis Gynoxis cuicochensis n/a LC NT PE EC 2500-4100 2800-4335 2500-4050 -12.8 -13.3 90.9 -21.9 -69.0 -21.7 -52.1 -39.6 53.8 -56.9 -81.3 -39.3 Gynoxis fuliginosa Gynoxis hallii Gynoxis miniphylla n/a LC NT EC EC 2700-4150 2500-4100 3100-4000 -7.3 266.4 223.8 -26.7 -17.7 -36.6 -35.1 198.6 44.6 -52.6 -39.6 -64.4 Gynoxis oleifolia Gynoxis parvifolia Gynoxis psilophylla Gynoxis reinaldii LC n/a n/a n/a PE BO - 3380-4900 2900-4100 2800-3900 2400-3300 -58.8 > 500 > 500 165.2 -81.6 -22.1 -7.6 -44.9 -90.1 > 500 > 500 226.1 -94.5 -42.5 -14.6 -64.5 EC 2900-4286 55.5 -15.8 21.4 -37.6 cr us an M d Polylepis incana no - 2450-3800 -39.1 -64.8 -55.8 -83.3 Polylepis lanuginosa VU B1abIII EC 2600-3630 > 500 -26.1 > 500 -49.1 Polylepis pauta Polylepis reticulata Polylepis sericea Polylepis besseri Polylepis racemosa Polylepis tomentella no VU A4c no no no no EC - 2700-4200 3200-4450 2500-3900 2500-4100 2900-4500 2800-4700 8.3 -28.9 -39.1 12.8 23.8 71.9 -59.7 -52.3 -63.6 -24.5 -16.4 -7.2 -61.1 -31.3 -52.6 8.4 30.2 59.0 -87.5 -81.3 -83.8 -32.4 -31.5 -16.2 no - 2700-4800 -38.0 -60.3 -46.7 -73.0 te VU B1ab(iii) Ac ce p 999 1000 1001 1002 1003 1004 1005 1006 1007 2050 Gynoxis sodiroi Polylepis weberbaueri 1 2020 Status of the species according to the IUCN red list of threatened species: LC: least concern, NT: nearthreatened, VU: vulnerable, EN: endangered, CR: critically endangered. Additional criteria as in http://www.iucnredlist.org/static/categories_criteria_3_1 2 Country where endemic, if endemic to the Andes. EC: Ecuador, PE: Peru, BO: Bolivia 3 Range change under different periods and for two dispersal scenarios. Full: unlimited dispersal, Null: no dispersal Species in bold depict range contractions (either by 2020 or 2050) regardless of migration assumptions. 40 Page 40 of 47 FIGURE CAPTIONS 1009 Figure 1 Study area. A. Elevation (in meters) across the tropical Andes countries overlaid with locations 1010 of weather stations in WorldClim; B. Number of modelling occurrences in 0.5 degree cells and key sites 1011 with high projected impacts (mentioned throughout the text). 1012 Figure 2 Evaluation of Maxent models. Distribution of the Area under the ROC Curve (AUC) for A. All 1013 vascular plants; B. All birds. Training AUC values are plotted for training (grey bars) and test (black bars) 1014 sets. AUC values of individual species are averages of 10 cross-validated runs with 10-20% of the input 1015 points drawn randomly. 1016 Figure 3 Modeled current species richness for A. Vascular plants and B. birds in the Tropical Andes as 1017 derived by the sum of binned species distributions models. Values are counts of species occurring in a 25 1018 km2 pixel. 1019 Figure 4 Spatial patterns of changes in species richness for birds and vascular plants under both migration 1020 scenarios and time periods. Values are percentage change in species richness from the present-day value 1021 shown in Figure 3. 1022 Figure 5 Species turnover for birds and vascular plants, for both periods. Community turnover can only 1023 be calculated for scenarios that somehow assume migration as this calculation requires that species can 1024 move to more suitable environments whenever possible. Values are percentages of change in community 1025 turnover as calculated by Eq. 1 (see Sect. 3.4 for details). 1026 Figure 6 Climate change impacts on individual species. Change in range size for birds (white bars) and 1027 vascular plants (grey bars) for A. Unlimited dispersal and B. No dispersal, for the SRES-A2 emission 1028 scenario and both periods (2020s and 2050s) (outliers have been removed from the plot for easier 1029 visualization). Box plots were constructed with n=1,456 and n=9,062 for birds and vascular plants, 1030 respectively. Ac ce p te d M an us cr ip t 1007 1008 1031 1032 41 Page 41 of 47 Ac ce pt ed M an us cr i Figure 1 Page 42 of 47 Ac ce pt ed M an us cr i Figure 2 Page 43 of 47 Ac ce pt ed M an us cr i Figure 3 Page 44 of 47 cr ip t Figure 4 2020s No dispersal 2050s ce pt Ac Vascular plants ed Birds M an 2050s us Unlimited dispersal 2020s Page 45 of 47 Ac ce p te d M an us cr ip t Figure 5 Page 46 of 47 Ac ce pt ed M an us cr i Figure 6 Page 47 of 47