Mammalia: Carnivora) in the Americas: Past to Present

Home , American cheetah, Cyonasua, Dire wolf, Machairodontinae

Journal of Mammalian Evolution https://doi.org/10.1007/s10914-020-09496-8

ORIGINAL PAPER

Environmental Drivers and Distribution Patterns of Carnivoran Assemblages (Mammalia: Carnivora) in the Americas: Past to Present

Andrés Arias-Alzate1,2 & José F. González-Maya3 & Joaquín Arroyo-Cabrales4 & Rodrigo A. Medellín5 & Enrique Martínez-Meyer2

# Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract Understanding species distributions and the variation of assemblage structure in time and space are fundamental goals of biogeography and ecology. Here, we use an ecological niche modeling and macroecological approach in order to assess whether constraints patterns in carnivoran richness and composition structures in replicated assemblages through time and space should reflect environmental filtering through ecological niche constraints from the Last Inter-glacial (LIG), Last Glacial Maximum (LGM) to the present (C) time. Our results suggest a diverse distribution of carnivoran co-occurrence patterns at the continental scale as a result of spatial climatic variation as an important driver constrained by the ecological niches of the species. This influence was an important factor restructuring assemblages (more directly on richness than composition patterns) not only at the continental level, but also from regional and local scales and this influence was geographically different throughout the space in the continent. These climatic restrictions and disruption of the niche during the environmental changes at the LIG-LGM-C transition show a considerable shift in assemblage richness and composition across the Americas, which suggests an environmental filtering mainly during the LGM, explaining between 30 and 75% of these variations through space and time, with more accentuated changes in North than South America. LGM was likely to be critical in species functional adaptation and distribution and therefore on assemblage structuring and rearranging from continental to local scales through time in the continent. Still, extinction processes are the result of many interacting factors, where climate is just one part of the picture.

Keywords Ecological niche . Carnivoran . Extinction . Climate change . Late Pleistocene . Paleodistributions

Electronic supplementary material The online version of this article Introduction (https://doi.org/10.1007/s10914-020-09496-8) contains supplementary material, which is available to authorized users. Understanding how species are distributed, their determinants * Andrés Arias-Alzate and constraints, and how they are spatially structured in as- [email protected] semblages through time and space is a fundamental goal of macroecology and ecology (Brown et al. 1995;Ferrazetal. 2012; Agosta and Bernardo 2013). Species can vary in size (in 1 Facultad de Ciencias y Biotecnología, Universidad CES, Calle 10A # 22-04, Medellín, Colombia mammals, which span 12 orders of magnitude), location, shape, and occupancy (Gaston 2003;Daviesetal.2009)and 2 Instituto de Biología, Universidad Nacional Autónoma de México, Circuito exterior s/n, Ciudad Universitaria, Coyoacán, are the result of responses to ecological rules such as climatic CP04510 México City, Mexico conditions, species dispersal abilities, historical events, phylo- 3 Proyecto de Conservación de Aguas y Tierras, ProCAT Colombia/ genetic inertia, and/or biotic interactions (Dynesius and Internacional, Carrera 11 # 96-43, Of. 303, Bogotá, Colombia Jansson 2000; Blomberg and Garland 2002; Steinitz et al. 4 Laboratorio de Arqueozoología “Ticul Álvarez Solórzano”, 2006; Davies et al. 2009; Blois et al. 2013, 2014). Subdirección de Laboratorios y Apoyo Académico, Instituto Understanding the main drivers of changes in such parameters Nacional de Antropología e Historia, Moneda # 16, Col. Centro, can also shed light in terms of local and global extinctions. An 06006 México City, Mexico important issue is whether these processes such as climatic 5 Instituto de Ecología, Universidad Nacional Autónoma de México, conditions have the same effect to structure assemblage pat- Circuito exterior s/n, Ciudad Universitaria, Coyoacán, terns and species distributions at different spatial and temporal CP04510 México City, Mexico J Mammal Evol scales (i.e., from continental to regional and/or to local scales) implications in the evolutionary patterns that took place con- (Arias-Alzate et al. 2017). However, accurate information on tinentally, thanks to its high taxonomic diversity and well- past and present distributional patterns for many species is resolved phylogeny (Goswami and Friscia 2010;Nyakatura often scarce at broad scales, and the mechanisms, determi- and Bininda-Emonds 2012). Here, we used an ecological nants, and constraints of these ranges at these scales are still niche modeling and macroecological approach in order to as- poorly understood (Graham 2001;Prevostietal.2005; sess the influence of past climate change on carnivoran assem- Martínez-Meyer et al. 2004; Davies et al. 2009; Nogués- blage richness and composition patterns from continental to Bravo 2009). Nevertheless, the understanding of the relation- local levels over the last 130 K years in the Americas and if ship between the influence of environmental drivers and as- this influence reflects environmental filtering through ecolog- semblages measures (i.e., richness and composition) have re- ical niche constraints. It is important to note that relatively ceived more attention, especially when extreme climatic little attention has been given to understanding the underlying events (e.g., glaciations) can change the outcome of the eco- causes of non-random patterns such as geographical distribu- logical patterns and processes (Thibault and Brown 2008; tion and assemblage structuring (Collins et al. 2011;Gotelli Davies et al. 2011). and Ulrich 2012; Blois et al. 2014); therefore, revealing these These patterns are largely governed by environmental fac- patterns and some of their mechanisms would improve our tors (e.g., climatic conditions) that define part of the funda- knowledge on the macroecological history of carnivoran as- mental niche of the species, where the biotic interactions are semblages in the Americas. usually less perceptible (Martínez-Meyer et al. 2004;Soberón and Nakamura 2009; Lorenzen et al. 2011; Peterson et al. 2011;Levinskyetal.2013). However, elucidating the influ- Materials and Methods ence of these processes on species coexistence and distribution is not straightforward because non-random species asso- Species Geographical Distribution (GD) Patterns Data ciations are not necessarily caused only by climatic factors nor species interactions. All these processes can operate indepen- Species GD patterns came from an ecological niche modeling dently or in synergy to determine assemblages patterns (ENM) approach done by Arias-Alzate (2016) and Arias- through time and space (Martínez-Meyer et al. 2004;Blois Alzate et al. (2017), which were estimated as described below. et al. 2014; Giarla and Jansa 2015). Therefore, the central The species studied were based on the assumption that all issue is how to differentiate or elucidate the influence pro- extant species were also present during the late Pleistocene, and duced by biotic interactions or dispersal limitations (either the extinct species disappeared at the end of the Pleistocene- by barriers or by movement capacity) from those produced early Holocene approximately 12Kyr (Webb 2006; Davies by environmental filtering (Dynesius and Jansson 2000; et al. 2009). The species criteria for that study included taxo- Svenning and Skov 2007; Blois et al. 2014; Giarla and Jansa nomic validity, proper chronological dating, and reported in- 2015). In this sense, the ecological niche played an important formation about the species’ presence during the study period. role to disentangle the combined effects of abiotic (e.g., cli- The list and the taxonomy for living species were based on matic) and biotic factors and could help us to understand in- Wilson and Reeder (2005), Wilson and Mittermeier (2009), directly about the effects of others historical factors such as and IUCN Red List of Threatened Species (version 2018.2.). barriers and accessible areas as strong forces on assemblage’s For extinct species, the taxonomy was based on the fossil re- dynamics and coexistence at different scales (Davis and Shaw cords and following Berta (1985),Barnettetal.(2005), 2001; Martínez-Meyer et al. 2004; Bofarull et al. 2008; Cisneros (2005), Cione et al. (2007), Prevosti and Rincón Peterson et al. 2011;Levinskyetal.2013; Soares 2013). (2007), Soibelzon and Prevosti (2008), Arroyo-Cabrales et al. Climate is an important determinant of species ecological (2010), and Ferrusquía-Villafranca et al. (2010). niche and its geographic expression, thus, helping to under- The study area was defined as the entire American conti- stand extinction processes through range shifts and bottleneck nent; this represents the area of accessibility M sensu Peterson events (Martínez-Meyer et al. 2004;Nogués-Bravoetal. et al. (2011), as all historical and ecological processes of colo- 2008; Arias-Alzate 2016). However, extinction processes are nization and dispersal of the species of interest occurred in the the result of many interacting factors (e.g., humans) that affect continent (during and after the Great American Biotic species differentially and could act differentially or asynchro- Interchange; GABI). Namely, during the GABI in several in- nously with additive effects at different spatial and temporal dependent migration events (in a step-like pattern) the extant scales, where climate is just one part of the picture (Thomas and extinct species or their ancestor (from Holarctic origin) et al. 2004; Cione et al. 2007;Araujoetal.2017; Arias-Alzate dispersed from North America to South America through the et al. 2017). Panamanian Isthmus, or subsequently diversified in South The order Carnivora represents an excellent group to assess America (most of these species are endemic) and colonized questions related to the biogeography and ecological niche Central and North America later (e.g., Eira barbara, Procyon JMammalEvol cancrivorus, Speothos venaticus; Webb 2006; Prevosti and the case of living species, the models were generated using Soibelzon 2012; Bacon et al. 2015; Prevosti and Forasiepi the current presence records over M1 (Current bioclimatic 2018). It is important to note that the Andes mountain range layers), and hindcast over the past climatic layers and validat- had already emerged by this time (i.e., GABI), reaching heights ed them with the fossil records (M1 to M1 with LIG climatic of over 3500 m since 10 Ma (Gregory-Wodzicki 2000), so layers, and M1 to M2 with LGM climatic layers). For extinct these areas did not represent an important barrier for these species, given that fossil records are scarce and to avoid losing species during migration events, as both small-carnivoran, information, the proper procedure was to generate the models mesocarnivoran, and large- and hypercarnivoran species were using all records for extinct species together to calibrate and represented during these migration events, where species colo- validate the models over the Inter-glacial climatic layer period nized even higher altitudes (e.g., the Andean weasel, Mustela (M1 with LIG climatic layers) and then project them over the frenata; the grison, Galictis vittata; the tayra, Eira Barbara;the Last Glacial Maximum time (M2 with LGM climatic layers). mountain lion, Puma concolor; the jaguar, Panthera onca;the As the objective was not focused on model or algorithm com- spectacled bear, Tremarctos ornatus, for some examples), and parison, it was opted for one modeling method (MaxEnt higher latitudes (e.g., the dire wolf, Canis dirus†; the jaguar, 3.3.3 K; Phillips et al. 2006; Phillips and Dudík 2008). This Panthera onca mesembrina†; the spectacled bear, Tremarctos algorithm has shown good performance and is robust with ornatus; the ancestor of the South American short-faced bears, small sample sizes (Elith and Graham 2009; Peterson et al. Arctotherium tarijense†, Arctotherium wingei†,forsomeex- 2011;Santika2011; Muscarella et al. 2014). ENM with amples). Thus, many of the barriers that these species face or MaxEnt estimates the suitability of condition across the land- faced were mainly climatic and / or biological in nature (i.e., scape by contrasting the environmental conditions where the species competition). We also highlighted the use of two dif- species has been recorded, against a sample of background ferent Ms., one for LIG and C times (M1) and one for the LGM pixels across the study area via a Bayesian procedure of model (M2) as during this period the continental area was wider given fitting under the maximum entropy principle (Phillips and the lowering of sea level that exposed areas that are now sub- Dudík 2008). The results during the niche modelling proce- merged (Varela and Fariña 2016). dure are frequently interpreted as probability of presence The subsequent analyzes performed by Arias-Alzate (Phillips et al. 2017) or as the areas where ecological condi- (2016) and Arias-Alzate et al. (2017)weredonefollowinga tions are suitable for the establishment of the species (the geoinformatics approach (Arroyo-Cabrales et al. 2010) where potential distribution of the species, sensu Peterson et al. a database containing confirmed records of extant and extinct 2011). Detailed explanations of MaxEnt and new terrestrial carnivoran species was constructed (G+, sensu implementations can be found elsewhere (Phillips et al. Peterson et al. 2011). The records were obtained through a 2006, 2017; Phillips and Dudík 2008; Elith et al. 2011; detailed search of the scientific literature (i.e., ISI Web of Merow et al. 2013). To test for model discriminatory ability

Science and Google Scholar) through key words (i.e., and performance, the AUCTest (Area under the ROC Curve Pleistocene, Carnivora, Felidae, Canidae, and specific and sci- based on the validation data) was used, which is implemented entific names), online databases, and museum specimens from in MaxEnt (Phillips et al. 2017). Despite that AUC has been online databases of biological collections and our own records criticized as a method for evaluating model quality, especially from extensive fieldwork in some countries (e.g., Colombia, to compare models from different algorithms (Lobo et al.

Costa Rica, Mexico). Then, sets of climatic data were con- 2008; Peterson et al. 2007), the AUCTest still has been shown structed from available past projections and current bioclimat- to be a useful measure for ordinal score and unique models in ic variables at the global scale, specifically climatic layers agreement with previous works (McPherson et al. 2004; from three periods: 1) Last Inter-glacial (LIG, ~120–140 Kyr Thuiller et al. 2005; Elith et al. 2011; Marino et al. 2011; BP; Otto-Bliesner et al. 2006), 2) Last Glacial Maximum Santika 2011; Muscarella et al. 2014; Phillips et al. 2017). (LGM, ~21 Kyr BP; Farrera et al. 1999; Braconnot et al. Statistical significance of models was evaluated with a bino- 2007), and 3) Current (C, Hijmans et al. 2005). All layers were mial test, which determines whether the model differs from in a 0.08333° resolution (~10 arcmin). These layers were null expectations. processed and extracted using the Ms. (i.e., M1 and M2) described above. Assemblage Hot Spots, Composition, and Richness Afterwards, with this previous information, Arias-Alzate Patterns (2016) and Arias-Alzate et al. (2017) performed the ENM via maximum entropy approach with the following parame- In order to assess whether constraints patterns in richness terization. One hundred replicate models using five hundred (species number) and composition (species identity) structures iterations per replicate for each species using a random 70:30 in the assemblages through time and spaces should reflect split (bootstrap method) of the total number of occurrence environmental filtering through ecological niche constraints, records (G+) for calibration and validation, respectively. In herein we generated a grid of 7760 1 × 1 degree cells over the J Mammal Evol continental scale for these three periods (LIG, LGM, and C). should be removed (Fotheringham et al. 1998;O’Brien 2007). With these grids for each time and with the extant and extinct Once the best model was selected, we tested for spatial differ- species GD (aforementioned in the previous section), an over- ences from random expectation: a Moran’s I spatial autocorrela- lapping count analysis was performed using the Hawth tools tion test of the residuals was used (Brunsdon et al. 2010)inorder (Beyer 2004) in order to extract the species co-occurrence as a to assess if other important variables are potentially missing in proxy of the species assemblage richness and composition the model. We also estimated the Koenke studentized Breusch- patterns present in each cell at each time in the Americas. Pagan statistic (K(BP)) and its probability, in order to assess the To validate if the carnivoran assemblages’ measures reliability of standard errors when heteroscedasticity is present. In (i.e., richness and composition) generated from individual cases where the K(BP) was significant, we used the robust prob- species GD could provide accurate richness and composi- ability instead of the raw probability estimation. Significant tion patterns, we then made a Cross-validation of these heteroscedasticity and stationarity indicate that the effects of the estimation (niche-based GD) with the assemblage’spat- drivers (environmental drivers) over the assemblage richness and terns estimated with the extant carnivoran species extent composition shifts would be different in magnitude and that of occurrences (EOO) of the IUCN. These assemblage pat- changes are not homogeneous along geographic space. terns based on EOO were estimated using the same proce- Likewise, to explore if spatial mismatching for each period dure mentioned above (see Online Resource 2 for a more occurred and if selected models did not perform adequately detailed explanation). Afterwards, we validated and esti- (i.e., showing at least one important variable was missing from mated the accuracy of these approaches comparing both the model), we performed a hot-spots analysis using the resid- assemblages’ patterns using a Pearson correlation test and uals of selected models based on the Getis-Ord Gi* test by the Jaccard index with a p < 0.001 statistically significance estimating Z-values (i.e., standard deviations) and its associ- at 99% confidence level (see Online Resource 2). All geo- ated probability for each cell on the continent (Getis and Ord graphic analyses were performed on a Geographic 2010;OrdandGetis2010; González-Maya et al. 2016a, Information System using ArcGIS 9.3 software (ESRI 2016b). This analysis identifies where clusters of high or 2001) and the statistical analysis were performed using low richness values are more marked and are significant the Infostat software (Di Rienzo et al. 2016). We then es- (p < 0.05) than one from theoretical complete random expec- timated the values for four environmental determinants in tation as the null hypothesis (Green and Ebdon 1977). each cell for each period based on the aforementioned Mapping these high and low values allowed us to highlight global scale climatic variables. These climatic conditions if these carnivoran species clusters were different for each have proven to be useful and informative as drivers of period in the continent and thus indicating if one or more mammalian biodiversity patterns and ecosystem function- explanatory variables are missing in the model for that cell ing at global scale in previous studies (Croft 2001;Safi (i.e., partial spatial fitting of the overall model) and could et al. 2011; González-Maya 2015): i) mean precipitation, mediate and play an important role on these patterns (Getis ii) precipitation seasonality, iii) mean temperature, and iv) and Ord 2010; Ord and Getis 2010; González-Maya et al. temperature range. 2016a, 2016b). To assess assemblage richness and composition shift pat- Afterwards, we performed a geographically weighted re- terns over the continent for the transition between periods gression (GWR) to identify the spatial influence of the envi- (LIG-LGM and LGM-C), we first generated an Ordinary ronmental factors within the continental grid cells over each Least Squares (OLS) regression in order to select the best assemblage measure (i.e., influence from continental to local explaining models (shifts in assemblages’ richness and com- level), allowing us to identify a spatial heterogeneity (e.g., position patterns) using the variable combinations (González- heteroscedasticity and stationarity) due to heterogeneous en- Maya et al. 2016a, 2016b). For each measure (richness and vironmental influence over the continent for the transition composition) we generated all possible variable combinations between periods (LIG-LGM and LGM-C; González-Maya with no replacement, resulting in 15 possible models, and then et al. 2016a, 2016b). As the influence of different variables selected the best competing models based on the lower Akaike is likely spatially defined, we selected GWR as an appropriate Information Criterion values (AIC) (Wagenmaker and Farrell method capable of identifying heterogeneity patterns at our 2004; González-Maya et al. 2016a, 2016b). We used the R2 as spatial scales (Fotheringham et al. 2002;Brunsdonetal. an indicator of the proportion of variation that is explained by 2010; González-Maya et al. 2016a, 2016b). All analyses and the resulting models (higher R2 values were preferred). spatial statistical tests (using the spatial statistics tools, consid- After selecting the best models, we assessed multicollinearity ered significant at p < 0.05 at 95% confidence level) were to see which variables are potentially redundant in influencing performed on a Geographic Information System using model patterns by extracting the estimated coefficients and ArcGIS 9.3 software (ESRI 2001). All data generated and Variance Inflation Factor (VIF); VIF values greater than 7.5 are analyzed here are available on reasonable request from the considered suspicious and redundant, so one of the two variables first author. JMammalEvol

Results America. Likewise, during the LGM a significant change was observed towards the west coast of North America, Central Species Geographical Distribution (GD) Patterns Data America, and South America, particularly in the Andes, and in the north and central regions (Fig. 2). For all (~88 spp.) but eight carnivoran species, no potential Correlation among assemblages’ richness patterns (niche distribution models for each period were considered due to base vs EOO) was high (Pearson = 0.90, p <0.001)(Fig. 1, lack of records (Mustela africana, Cuon alpinus, Speothos Online Resource 2). Both analyses showed relatively high pacivorus,andEnhydra macrodonta), due to ranges restricted richness from mid to low latitudes in the tropical zone and in to islands (Urocyon litoralis and Procyon pygmea, which are the Andes region and low richness toward high latitudes in the recently separated from two continental species that were iso- continent (Fig. 1). On the other hand, regarding the composi- lated after the last glacial maximum), and due to a recently tion patterns, both approaches provided relatively accurate described new species and separated from Nasuella olivacea characterizations across the majority of the grid cells with high (Nasuella meridensis) and Leopardus trigrinus (Leopardus Jaccard values (Fig. 2 Online Resource 2). Even though, little guttullus). In terms of model performance for all species, spatial differences among approaches persist. Thus, this niche

AUCTest values were all higher than 0.81, indicating a good base characterization gives us good reliability for interpreting performance and model discriminatory ability. With respect to the other estimations (see Online Resource 2 for a more the model validation, all species models presented a statistical detail). significance (p <0.05). Carnivoran assemblages’ richness shows a latitude structuring pattern and a significant shift from the LIG to C period Assemblage Hot Spots, Composition, and Richness across the Americas (Fig. 3). The best model (model with the Patterns lower AIC value, Table 1) that accounted for and explained richness shifts includes mean temperature, temperature range, We found a diverse distribution of carnivoran co-occurrence and precipitation seasonality as important drivers for richness patterns at the continental scale as a result of spatial climatic patterns at continental scales (LIG-LGM OLS R2 = 0.328; variation constrained by the ecological niches of the species LGM-C OLS R2 = 0.433 (Table 1, Fig. 3). The LIG had a (Fig. 1). The areas with more stable climatic conditions over greater carnivoran species richness concentrated from low to the Neotropics upon these periods suggest that these condi- mid latitudes in North and South America in the subtropical tions enable the persistence of some assemblages to present regions. However, during the LIG-LGM transition, due to times (Fig. 2). The spatial hot spots analysis allowed us to glacial conditions, low temperatures and drier environments, identify statistically significant and noticeable changes in as- this pattern changed with high loss of species from north and semblages’ structure during the LIG-LGM-C transitions. The central of North America (Fig. 3). During the LGM-C transi- results show high and marked hot spots during the LIG at mid tion the biogeographic assemblages' richness patterns tend to to low latitudes in North America and in the Andes of South recover toward the temperate areas of North America (i.e.,

Fig. 1 Carnivoran assemblages’ richness patterns over the last 130 K years in America. a.LIG,b.LGM,andc. C periods. Note the high richness reduction in North America during LGM J Mammal Evol

Fig. 2 Hot spots of carnivoran assemblages across Americas over the last 130 K years. a. LIG, b.LGM,andc. C periods. Z-scores accounts for standard deviations of the Getis-Ord Gi test. Red dots indicate significant (p < 0.05) and more marked associations grasslands, savannas, temperate coniferous forests, temperate Discussion broadleaf, and mixed forests) but with species loss in the Neotropics (Fig. 3). On the contrary, carnivoran assemblages’ Understanding carnivoran species distributions and how they composition structure shows a different pattern. The best mod- are spatially structured has received greater attention in recent el that accounted for and explained composition shifts includ- years (Goswami and Friscia 2010;Levinskyetal.2013). Most ed mean temperature, temperature range, mean precipitation, carnivoran lineages in the Americas were holarctic species and precipitation seasonality as main drivers (LIG-LGM OLS with a lower diversity during the Miocene-Pliocene compared R2 = 0.274; LGM-C OLS R2 = 0.346, (Table 2,Fig.3). with the Pleistocene (Marshall et al. 1982;Prevostiand During the LIG-LGM phase the environmental drivers caused Reguero 2000; Webb 2006; Prevosti and Soibelzon 2012; greater shifts in carnivoran composition at mid to high lati- Prevosti and Forasiepi 2018). This group was one of the most tudes in North America, and at mid latitudes and in the west- successful among mammalian species to participate in the ern coasts of South America, suggesting an environmental Great American Biotic Interchange (GABI as a complex filtering. During the LGM-C transition, the main environmen- process that persists today; Dundas 1999;Graham2001; tal effects were in the Neartic region and the austral zone of Webb 2006; Johnson et al. 2006; Rincón et al. 2011; South America (Fig. 3). Prevosti et al. 2013; Bacon et al. 2015). However, some of For both parameters (richness and composition patterns) these species and other terrestrial fauna faced the “Megafauna the K(BP) the statistic indicated heteroscedasticity and station- Extinction,” which apparently affected mostly the mammalian arity of the models. Showing, in this way, a heterogeneous faunas of the Nearctic, Neotropics, and Australasia regions influence of the environmental drivers on the assemblages’ (Bofarull et al. 2008; Prevosti and Soibelzon 2012). richness and composition patterns. This relationship changes However, some species persist until today. when the magnitude of drivers changes causing the patterns Our results show that carnivoran co-occurrence patterns at not to be constant across geographic space and time in the continental scales in the Americas at the end of the Pleistocene Americas (Tables 1, 2). These patterns suggest a greater asso- are, in part, the result of important environmental drivers ciation and a heterogeneous environmental influence on constrained by the ecological niche stability of the species, carnivoran richness patterns at local-regional levels within which represents the ranges of the species on the geography. the continent for the transition between periods (LIG-LGM This suggests that these significant climatic variations were GWR R2 =0.67;LGM-CGWRR2=0.75;Table1,Fig.4). critical in range contractions, mainly from high to mid lati- However, composition patterns show more moderate environ- tudes in North America and in the Andes and mid latitudes in mental influences at local-regional levels, but with an impor- South America during the LIG-LGM and LGM-C transitions, tant pattern towards north central North America (LIG-LGM causing substantial changes on carnivoran assemblages’ rich- GWR R2 = 0.57; LGM-C GWR R2 =0.62;Table2). ness and composition at the continental scale. These spatial JMammalEvol

Fig. 3 Influences of environmental drivers on carnivoran assemblages’ the number of species loss and gain, and c. and d. represents the richness shift patterns: a. LIG-LGM, b. LGM-C, and composition shift percentage of change during the transition between periods) patterns: c.LIG-LGM,d. LGM-C. (the a. and b. legend represents analyses allowed us to identify noticeable changes that appar- 2004;Dyke2005;Araújoetal.2006;BloisandHadly2009; ently match with the biomes displaced southward in North Croitor and Brugal 2010; Diniz-Filho et al. 2009; Li et al. America as the continental ice sheets (i.e., Laurentide Ice 2014; Villavicencio et al. 2016). However, it is important to Sheet) grew and partially waned at decamillennial intervals note that some other important factors that mainly act at local (Dyke 2005). Equally, these changes match with the shifting to regional scales could be missing from our approach (i.e., biomes of South America during the LGM; the Andes was species interaction, human influence, resources) and could covered in some parts by ice and deserts, and the north and help to explain the remaining proportion of variation mainly central regions were a mosaic of open forest and open savanna seen in the composition shifts during these transitions (Fig. 4). (Cione et al. 2009). These richness and composition changes For example, competitive exclusion of similar and closely from LIG-LGM and LGM-C transitions highlight the role of related species is likely to result from competition for space climatic alterations and these environmental drivers (mean and resources because species that occupy separate ranges precipitation, precipitation seasonality, mean temperature, then occurred in association during the LGM, or human im- and temperature range) as some of the primary factors with pacts and reduced herbivore populations acting at different important effects over carnivoran assemblages that made them intensities were also critical, increasing even more the extinc- more susceptible to local extinctions. tion risk (Dyke 2005; Cione et al. 2009; Diniz-Filho et al. These climatic alterations have also been proposed as im- 2009;Lietal.2014; Villavicencio et al. 2016). Nevertheless, portant factors altering the distribution and diversity patterns as the climatic variations also act differentially and it is a of other vertebrate and plants communities in other regions taxon-specific process, given that the species respond accord- (Graham and Mead 1987; Root 1988;Lundbergetal.2000; ing to their physiological and ecological characteristics, phy- Lyons 2003; Eronen and Rook 2004; Svenning and Skov logenetic inertia, dispersal and differential colonization J Mammal Evol

Table 1 Results for best competing and selected models testing for Factor, ORL: ordinary least squares; GWR: Geographic weighted environmental drivers influence on carnivoran assemblages’ richness regression; AIC: Akaike Information Criterion; K(BP): Koenker’s shift patterns using a 7760 1 × 1 degree cells in the Americas. Co: studentized Breusch-Pagan Statistic. * indicates the selected models Coefficient; SE: Standard Error; P: p value; VIF: Variance Inflation

LIG-LGM Models

Model Variable Co SE p Robust t Robust p VIF OLSR2 GWR R2 AIC K(BP) K(BP)- P

1* Intercept −0.599 0.132 <0.05 −6.256 0.000 0.328 0.67 53,117.48 495.590 <0.05 Mean Temperature 0.443 0.010 <0.05 42.345 0.000 1.282 Temperature range 0.063 0.005 <0.05 14.336 0.000 1.284 Precipitation seasonality 0.115 0.003 <0.05 32.499 0.000 1.011 2Intercept −0.617 0.133 <0.05 −6.421 0.000 0.328 53,118.28 513.678 <0.05 Mean Temperature 0.437 0.011 <0.05 36.850 0.000 Temperature range 0.062 0.005 <0.05 14.047 0.000 Mean precipitation 0.000 0.000 0.274 1.116 0.264 Precipitation seasonality 0.113 0.003 <0.05 28.301 0.000 Intercept 0.250 0.114 <0.05 2.420 0.016 0.315 53,265.06 152.355 <0.05 3 Mean Temperature 0.373 0.010 <0.05 37.468 0.000 Mean precipitation 0.001 0.000 <0.05 2.960 0.003 Precipitation seasonality 0.114 0.004 <0.05 29.189 0.000 4 Intercept 0.336 0.110 <0.05 3.194 0.001 0.314 53,272.2 104.903 <0.05 Mean Temperature 0.386 0.009 <0.05 43.973 0.000 Precipitation seasonality 0.119 0.003 <0.05 34.603 0.000 LGM-C Models 1* Intercept −1.679 0.096 <0.05 −17.387 0.000 0.433 0.75 49,996.54 207.303 <0.05 Mean Temperature 0.465 0.007 <0.05 66.180 0.000 1.024 Temperature range 0.127 0.008 <0.05 13.063 0.000 2.626 Precipitation seasonality 0.059 0.006 <0.05 7.727 0.000 2.612 2Intercept −1.667 0.096 <0.05 −17.511 0.000 0.433 49,996.6 158.532 <0.05 Mean Temperature 0.460 0.008 <0.05 57.452 0.000 Temperature range 0.126 0.008 <0.05 13.121 0.000 Mean precipitation 0.000 0.000 0.16 1.235 0.217 Precipitation seasonality 0.057 0.006 <0.05 7.154 0.000 Intercept −1.887 0.094 <0.05 −21.374 0.000 0.426 50,088.73 535.589 <0.05 3 Mean Temperature 0.460 0.008 <0.05 58.442 0.000 Temperature range 0.183 0.005 <0.05 42.762 0.000 Mean precipitation 0.001 0.000 <0.05 3.484 0.001 4Intercept −1.948 0.093 <0.05 −21.881 0.000 0.425 50,102.07 690.205 <0.05 Mean Temperature 0.475 0.007 <0.05 67.932 0.000 Temperature range 0.191 0.005 <0.05 51.072 0.000

capacity (which is different in relation to carnivoran species) approaches at the continental scale (Gittleman and Gompper (Blomberg and Garland 2002; Canto et al. 2010; Croitor and 2005;Schipperetal.2008; Polly 2013; Fergnani and Brugal 2010; Lorenzen et al. 2011; Agosta and Bernardo Ruggiero 2015). The smallest carnivoran ranges are mostly 2013; Prevosti and Forasiepi 2018), we cannot highlight ac- present in the tropical region, and the largest ranges are found cording to our results that these patterns shown herein are a from low to high latitudes. These patterns seem to follow in general rule (via ecological niches stability) for the other part Rapoport’s rule, which reflects the seasonal variability of groups (e.g., others mammals (large herbivores), birds and high latitude environments and other climatic oscillations reptiles) with similar co-occurrence patterns at continental (e.g., Milankovitch oscillations; Stevens 1989;Dynesiusand scale, but it is a plausible possibility. Jansson 2000; Davies et al. 2011). This could explain the Otherwise, the results display an important latitudinal gra- greatest assemblage richness pattern towards the tropics dur- dient (i.e., current richness pattern) consistent with other ing the LIG and C times (although, also high richness can be JMammalEvol

Table 2 Results for best competing and selected models testing for Variance Inflation Factor, ORL: ordinary least squares; GWR: environmental drivers influence on composition shift patterns Geographic weighted regression; AIC: Akaike Information Criterion; carnivoran assemblages using a 7760 1 × 1 degree cells in the K(BP): Koenker’s studentized Breusch-Pagan Statistic. * indicates the Americas. Co: Coefficient; SE: Standard Error; P: p value; VIF: selected models

LIG-LGM Models

Model Variable Co SE p Robust t Robust p VIF R2 GWR R2 AIC K(BP K(BP)-P

1* Intercept 0.353 0.006 <0.05 73.82 0.00 0.274 0.57 4446.84 513.678 <0.05 Mean Temperature −0.008 0.000 <0.05 −15.39 0.00 1.636 Temperature range 0.005 0.000 <0.05 23.57 0.00 1.317 Mean precipitation 0.000 0.000 <0.05 9.71 0.00 1.504 Precipitation seasonality 0.005 0.000 <0.05 27.03 0.00 1.195 2 Intercept 0.360 0.006 <0.05 74.52 0.00 0.265 4538.55 495.590 <0.05 Mean Temperature −0.006 0.000 <0.05 −11.85 0.00 Temperature range 0.005 0.000 <0.05 25.74 0.00 Precipitation seasonality 0.005 0.000 <0.05 32.58 0.00 Intercept 0.339 0.006 <0.05 66.23 0.00 0.188 5312.81 391.469 <0.05 3 Mean Temperature −0.009 0.000 <0.05 −16.24 0.00 Temperature range 0.005 0.000 <0.05 22.73 0.00 Mean precipitation 0.000 0.000 <0.05 20.37 0.00 4 Intercept 0.423 0.005 <0.05 82.54 0.00 0.226 4941.16 152.355 <0.05 Mean Temperature −0.013 0.000 <0.05 −28.84 0.00 Mean precipitation 0.000 0.000 <0.05 12.54 0.00 Precipitation seasonality 0.005 0.000 <0.05 27.29 0.00 LGM-C Models 1* Intercept 0.367 0.004 <0.05 77.63 0.00 0.346 0.62 2231.26 158.532 <0.05 Mean Temperature 0.015 0.000 <0.05 35.91 0.00 1.314 Temperature range 0.001 0.000 <0.05 2.07 0.04 2.634 Mean precipitation 0.000 0.000 <0.05 −14.26 0.00 1.537 Precipitation seasonality −0.007 0.000 <0.05 −22.01 0.00 2.786 2 Intercept 0.374 0.004 <0.05 75.20 0.00 0.323 2497.7 207.303 <0.05 Mean Temperature 0.012 0.000 <0.05 29.50 0.00 Temperature range 0.001 0.000 0.15 1.17 0.24 Precipitation seasonality −0.009 0.000 <0.05 −24.55 0.00 Intercept 0.395 0.004 <0.05 74.92 0.00 0.284 2935.16 535.589 <0.05 3 Mean Temperature 0.015 0.000 <0.05 30.57 0.00 Temperature range −0.006 0.000 <0.05 −26.40 0.00 Mean precipitation 0.000 0.000 <0.05 −18.05 0.00 4 Intercept 0.368 0.004 <0.05 79.24 0.00 0.345 2235.08 133.698 <0.05 Mean Temperature 0.015 0.000 <0.05 36.46 0.00 Mean precipitation 0.000 0.000 <0.05 −14.17 0.00 Precipitation seasonality −0.007 0.000 <0.05 −34.60 0.00

seen in temperate latitudes in North America), as greatest many of the holarctic species had to migrate south escaping ranges (e.g., dire wolf, (Canis dirus†), saber cats (Smilodon from unsuitable areas (Davis and Shaw 2001; Dyke 2005; fatalis† and S. populator†), jaguar (Panther onca), cougar Davies et al. 2011). This high richness pattern in the late (Puma concolor), and bush dog (Speothos venaticus)) are Pleistocene is consistent with previously suggested estima- confluent with the narrower ranges at more equatorial lati- tions, being one of the highest in richness worldwide in pro- tudes (e.g., Bassariscus spp., Bassaricyon spp., and Potos portion to the continental area, and significantly higher than flavus ranges). This was also evident at the LGM for the tro- the richness recorded during middle Holocene and the current pics where geographical ranges tended to be more stable due time as shown herein (Cione et al. 2003; Prevosti and to less drastic variation in environmental conditions and that Vizcaíno 2006). J Mammal Evol

Fig. 4 Geographically weighted regression showing the differential and LGM, b. LGM-C. Composition shift patterns: c. LIG-LGM, d. LGM- local influence by environmental drivers on carnivoran assemblages C. (legend represents Local R2 values) during the transition between periods. Richness shift patterns: a. LIG-

Even though it appears that the composition shifts at these southward to the southern United States and northern Mexico scales where slightly less influenced by these climatic drivers and westward to the western coast of the United States and due to the low variation explained by the composition models Canada. Yet, species with both temperate and tropical distri- compared with the richness pattern (see Table 2), it is clear that butions where less affected, so these shift patterns were appar- in some regions both richness and composition pattern shifted ently less severe in South America from low to mid latitudes. dramatically. For example, these shifts were more pronounced These results are consistent with the impact of Quaternary in the Nearctic region because much of this area was covered climate oscillations on other mammals in other regions by the Laurentian Glacier during the LGM, including Canada (Davies et al. 2011; Lorenzen et al. 2011; Levinsky et al. and a large portion of the United States (Dyke 2005). Thus, 2013; Blois et al. 2014). The loss of jaguar populations dis- carnivoran ranges responded accordingly with most species’ tributed at higher latitudes (P. onca augusta from North characteristics shifting larger distances to more suitable areas, America and P. onca mesembrina from South America) and JMammalEvol range constriction from high latitudes towards the tropics with carnivoran species due to their energetic constraint, biological a more stable range is an example (Arias-Alzate et al. 2017). characteristics, area and population requirements (e.g., such as It is important to note that this minor effect on composition low population size, low reproductive rates, and large home patterns perhaps could also be explained because after this last ranges) and open habitats adaptation (O’Regan et al. 2002; period of time (LGM-C transition) not considerable Cione et al. 2007; Canto et al. 2010, Lorenzen et al. 2011; carnivoran species turnover occurred as previously happened Agosta and Bernardo 2013; Levinsky 2010;Levinskyetal. (Prevosti and Soibelzon 2012; Silvestro et al. 2015). For ex- 2013; Arias-Alzate et al. 2017). Therefore, such species were ample, Prevosti and Soibelzon (2012) suggested a first turn- inherently more vulnerable to energetic constraint and range over during the Pre-Ilinonian (North America) and Rio LLico size reductions, due to the reduction of climatically suitable (South America) Ages/stages (~478–424 ka), and a second areas in the LIG-LGM transition, the reason why most larger turnover during the Ilinonian (North America) and Santa species needed to hold large distribution ranges in order to Maria (South America) Ages/stages (~200–130 ka). maintain viable populations to persist and to avoid bottlenecks Apparently, these ages/stages correspond with two glacial pe- events and extinction risk (Canto et al. 2010; Lorenzen et al. riods proposed by Porter (1981) and Cohen and Gibbard 2011; Agosta and Bernardo 2013; Arias-Alzate et al. 2017). (2010), where species replacement emerged apparently by Interestingly, more ecologically tolerant, eurytopic, and competition with the entrance of new carnivoran species with flexible species with a more stable range (i.e., more continu- similar ecological trait-space niches (e.g., Cyonasua merani ous and less fragmented ranges) could persist in the continent vs Nasua nasua; Galictis henningi vs Galicitis cuja; (Cione et al. 2007; Arias-Alzate et al. 2017). These patterns Lyncodon bosei vs Lyncodon patagonicus; Arctotherium are consistent with the evolutionary dynamics and the ecolog- angustidens vs Arctotherium tarijense; Brachyprotoma ical structure of carnivoran species in Europe, where the small optusata vs Mephitis mephitis; Martes diluviana vs Martes body-size species with high reproductive rates and ecological pennanti; Smilodon gracilis vs Smilodon fatalis; Arctodus plasticity (e.g., such as some felids and mustelids) ensured pristinus vs Arctodus simus). their success during the late Pleistocene dramatic ecological This glacial history has long been considered as an impor- changes (Croitor and Brugal 2010). It is to say that species like tant factor in shaping diversity patterns worldwide (Davies lynxes (Lynx rufus and L. canadensis) that suffered drastic et al. 2009, 2011), where the last glacial period is not the range reductions, and the jaguar (Panthera onca) and puma exception. These environmental factors shown herein and (Puma concolor) that likely lost their higher latitudes popula- their disturbances events (e.g., glaciations) are important tions persist up to the present. The latter two species main- drivers that influence these ecological and evolutionary tained a more continued ranges in the tropical region and carnivoran distribution responses, which are likely to be more recolonized North America after the LGM via founder effect detectable at broader spatial scales (Martínez-Meyer et al. (Culver et al. 2000; Arias-Alzate et al. 2017). Besides, in these 2004; Soberón and Nakamura 2009;Morrisetal.2010; suitable areas, mesocarnivores and perhaps other non- Davies et al. 2011;Lorenzenetal.2011;Levinskyetal. specialized carnivorans kept a variety of prey that allowed 2013; Blois et al. 2014). Here, we provide insights that the them to endure, unlike the more hypercarnivorous species environmental drivers dynamically and differentially affected adapted to feed on megaherbivores that also experienced the richness and composition patterns in the Americas (more reduction of their suitable areas summed to other anthropo- marked on richness than composition), not only at the conti- genic effects (Cione et al. 2007; Villavicencio et al. 2016). nental level, but also from regional and local scales as recently In this sense, it is possible that many of these carnivoran suggested (Blois and Hadly 2009;Wiszetal.2013). These species that persisted through the last climate change “extinc- results are consistent with evidence on composition of tion filter” represent the current set of species best suited to European mammalian communities of the past 20 million face natural environmental changes. However, as the models years, which suggests that this composition structure remained assessed herein where not explained entirely by environmen- constant despite the significant richness shifts of the dominant tal drivers, past climate change and filter effects played just a herbivore assemblages, as has also been proposed for the part of the puzzle about the carnivoran assemblages’ structure mammalian communities in Australia (Jernvall and Fortelius patterns and extinction event in Americas. It is important to 2004; Prideaux et al. 2007; Blois and Hadly 2009). note that other important forces, associated with biotic inter- These carnivoran patterns can be viewed as a description of actions and competitive effects among multiple clades, and the environmental and ecological niche constraints that have even with human interactions (although human migration contributed to shape the lineages’ geographical distributions and its effects was not uniform throughout the Americas, (Cardillo et al. 2006;Daviesetal.2009; Palombo et al. 2009; Barnosky and Lindsey 2010) could mediate and play an im- Agosta and Bernardo 2013), suggesting an environmental fil- portant role on these patterns from local and regional levels as tering during the LIG-LGM-C transition. This environmental has been proposed in other studies (De Vivo and Carmignotto filtering apparently affected more the hyper-large and large 2004; Araújo and Luoto 2007; Cione et al. 2009; Croitor and J Mammal Evol

Brugal 2010; Davies et al. 2011;Wiszetal.2013; Silvestro Arroyo-Cabrales J, Polaco OJ, Johnson E, Ferrusquía-Villafranca I et al. 2015; Villavicencio et al. 2016; Araujo et al. 2017). (2010) A perspective on mammal biodiversity and zoogeography in the late Pleistocene of México. Quat Int 212:187–197. https:// However, recent works suggested that extinction may have doi.org/10.1016/j.quaint.2009.05.012 started in the north before human arrival and occurred later Bacon CD, Silvestro D, Jaramillo C, Tilston, B, Chakrabarty P (2015) in higher latitudes after humans were present and where cli- Biological evidence supports an early and complex emergence of – mate changes were more severe (Hubbe et al. 2007;Barnosky the isthmus of Panama. Proc Natl Acad Sci USA 112:6110 6115. https://doi.org/10.1073/pnas.1423853112 and Lindsey 2010; Villavicencio et al. 2016). Barnett R, Barnes I, Phillips MJ, Martin LD, Harington CR, Leonard JA, Although fossil records are always limited by local condi- Cooper A (2005) Evolution of the extinct sabretooths and the tions and taphonomy, and therefore spatial and temporal dis- American cheetah-like cat. Curr Biol 15:R589–R590. https://doi. tribution could be biased (Nogués-Bravo et al. 2008), here, org/10.1016/j.cub.2005.07.052 instead of presenting fine temporal or spatial resolution analy- Barnosky AD, Lindsey EL (2010) Timing of Quaternary megafaunal extinction in South America in relation to human arrival and climate ses, due to these limitations, we aimed to show a general trend change. Quat Int 217:10–29. https://doi.org/10.1016/j.quaint.2009. of responses of the group. Here we provide plausible hypoth- 11.017 eses and new insights regarding history and spatial patterns of Berta A (1985) The status of Smilodon in North and South America. – american carnivoran species during the last 130 Kyrs. Climatic Contrib Sci Nat Hist Mus LA County 370:1 15 Beyer HL (2004) Hawth’sanalysistoolsforArcGIS.http://www. variation and the ecological niche constraints are in part an spatialecology.com/htools important role of carnivoran assemblages’ restructuring from Blois JL, Gotelli NJ, Behrensmeyer AK, Faith TS, Lyons K, Williams continental to local scales in the Americas. Environmental fil- JW, Amatangelo Kl, Bercovici A, Du A, Eronen JT, Graves GR, Jud tering is likely a key factor in the extinction context, increasing N, Labandeira C, Looy CV, McGill B, Patterson D, Potts R, Riddle the susceptibility to extinction of remaining populations, espe- B, Terry R, Tóth A, Villaseñor A, Wing S (2014) A framework for evaluating the influence of climate, dispersal limitation, and biotic cially if other stressors were operating synergistically as cur- interactions using fossil pollen associations across the late rent times perfectly evidence. Quaternary. Ecography 37:1095–1108. https://doi.org/10.1111/ ecog.00779 Acknowledgements We would like to thank H. Zarza, Joe J. Figel, and Blois JL, Hadly EA (2009) Mammalian response to Cenozoic climatic A. Townsend Peterson for support, pre-review, and insightful comments change. Annu Rev Earth Planet Sci 37:181–208. https://doi.org/10. on the manuscript. A. Arias-Alzate acknowledges the Posgrado en 1146/annurev.earth.031208.100055 Ciencias Biológicas, the Instituto de Biología-UNAM, and the scholar- Blois JL, Williams JW, Fitzpatrick MC, Ferrier S, Veloz SD, He F, Liu Z, ship and financial support for the Doctoral Degree Program in Biological Manion G (2013) Modeling the climatic drivers of spatial patterns in Sciences provided by the Consejo Nacional de Ciencia y Tecnología of vegetation composition since the last glacial maximum. Ecography México (CONACyT) (scholarship 280993). We also thank the reviewers 36:460–473. https://doi.org/10.1111/j.1600-0587.2012.07852.x for their comments, which helped to improve the manuscript. Blomberg SP, Garland T Jr (2002) Tempo and mode in evolution: phylogenetic inertia, adaptation and comparative methods. J Evol Biol 15:899–910. https://doi.org/10.1046/j.1420-9101. 2002.00472.x Bofarull A, Royo A, Fernández MH, Ortiz-Jaureguizar E, Morales J References (2008) Influence of continental history on the ecological special- ization and macroevolutionary processes in the mammalian assemblage of South America: differences between small and large Agosta SJ, Bernardo J (2013) New macroecological insights into func- mammals. BMC Evol Biol 8:97. https://doi.org/10.1186/1471- tional constraints on mammalian geographical range size. Proc Biol 2148-8-97 Sci 280:20130140. https://doi.org/10.1098/rspb.2013.0140 Braconnot P, Otto-Bliesner B, Harrison S, Joussaume S, Peterchmitt J-Y, Araujo BBA, Oliveira-Santos LGR, Lima-Ribeiro MS, Diniz-Filho F, Abe-Ouchi A, Crucifix M, Driesschaert E, Fichefet Th, Hewitt CD, Fernandez FAS (2017) Bigger kill than chill: the uneven roles of Kageyama M, Kitoh A, Laîné A, Loutre M-F, Marti O, Merkel U, humans and climate on late Quaternary megafaunal extinctions. Ramstein G, Valdes P, Weber SL, Yu Y, Zhao Y (2007) Results of Quat Int 431:216–222. https://doi.org/10.1016/j.quaint.2015.10.045 PMIP2 coupled simulations of the mid-Holocene and last glacial Araújo MB, Luoto M (2007) The importance of biotic interactions for maximum – Part 2: Feedbacks with emphasis on the location of modelling species distributions under climate change. Glob Ecol the ITCZ and mid- and high latitudes heat budget. Clim Past 3: Biogeogr 16:743–753. https://doi.org/10.1111/j.1466-8238.2007. 279–296. https://doi.org/10.5194/cp-3-279-2007 00359.x Brown JH, Mehlman DW, Stevens GC (1995) Spatial variation in Araújo MB, Thuiller W, Pearson RG (2006) Climate warming and the abundance. Ecology 76:2028–2043. https://doi.org/10.2307/ decline of amphibians and reptiles in Europe. J Biogeogr 33:1712– 1941678 1728. https://doi.org/10.1111/j.1365-2699.2006.01482.x Brunsdon C, Fotheringham AS, Charlton ME (2010) Geographically Arias-Alzate A (2016) Patrones de simpatría y efectos del cambio weighted regression: a method for exploring spatial nonstationarity. climático en carnívoros (Mammalia: Carnivora) de América. Tesis Geogr Anal 28:281–298. https://doi.org/10.1111/j.1538-4632.1996. de Doctorado, Universidad Nacional Autónoma de México, México tb00936.x City Canto J, Yáñez J, Rovira J (2010) Estado actual del conocimiento de los Arias-Alzate A, González-Maya JF, Arroyo-Cabrales J, Martínez-Meyer mamíferos fósiles de Chile. Estud Geol 66:255–284. https://doi.org/ E (2017) Wild felid range shift due to climatic constraints in the 10.3989/egeol.39778.055 Americas: a bottleneck explanation for extinct felids? J Mammal Cardillo M, Mace GM, Gittleman JL, Purvis A (2006) Latent extinction Evol 24:427–438. https://doi.org/10.1007/s10914-016-9350-0 risk and the future battlegrounds of mammal conservation. Proc Natl JMammalEvol

Acad Sci USA 103:4157–4161. https://doi.org/10.1073/pnas. Elith J, Graham CH (2009) Do they? How do they? WHY do they differ? 0510541103 On finding reasons for differing performances of species distribution Cione AL, Tonni EP, Bargo S, Bond M, Candela MA, Carlini AA, models. Ecography 32:66–77. https://doi.org/10.1111/j.1600-0587. Deschamps CM, Dozo MT, Goin FJ, Montalvo CI, Nasif N, 2008.05505.x Noriega JI, Jaureguizar EO, Pascual R, Prado JL, Reguero AM, Elith J, Phillips SJ, Hastie T, Dudík M, Chee YE, Yates CJ (2011) A Scillato-yané GJ, Soibelzon L, Verzi DH, Carolina E, Vizcaíno SF statistical explanation of MaxEnt for ecologists. Divers Distrib 17: (2007) Mamíferos continentales del Mioceno tardío a la actualidad 43–57. https://doi.org/10.1111/j.1472-4642.2010.00725.x en la Argentina: cincuenta años de estudios. Ameghiniana 11:257– Eronen JT, Rook L (2004) The Mio-Pliocene European primate fossil 278 record: dynamics and habitat tracking. J Hum Evol 47:323–341. Cione AL, Tonni EP, Soibelzon L (2003) The broken zig-zag: Late https://doi.org/10.1016/j.jhevol.2004.08.003 Cenozoic large mammal and tortoise extinction in South America. ESRI (2001) ArcGIS 9.3. Environmental Systems Research Institute, Rev del Mus Argentino Ciencias Nat 5:1–19 Redlands, California Cione AL, Tonni EP, Soibelzon L (2009) Did humans cause the late Farrera I, Harrison SP, Prentice IC, Ramstein G, Guiot J, Bartlein PJ, Pleistocene-early Holocene mammalian extinctions in South Bonnefille R, Bush M, Cramer W, von Grafenstein U, Holmgren America in a context of shrinking open areas? In: Haynes G (ed) K, Hooghiemstra H, Hope G, Jolly D, Lauritzen S-E, Ono Y, Pinot Vertebrate Paleobiology and Paleoanthropology. Springer S, pp S, Stute M, Yu G (1999) Tropical climates at the last glacial maxi- 125–144 mum: a new synthesis of terrestrial palaeoclimate data. I. Vegetation, Cisneros JC (2005) New Pleistocene vertebrate fauna from El Salvador. lake-levels and geochemistry. Clim Dyn 15:823–856. https://doi. Rev Bras Paleontol 8:239–255. https://doi.org/10.4072/rbp.2005.3. org/10.1007/s003820050317 09 Fergnani PN, Ruggiero A (2015) Ecological diversity in south American Cohen KM, Gibbard PL (2010) Global chronostratigraphical correlation mammals: their geographical distribution shows variable associa- table for the last 2.7 million years. Utrecht University repository tions with phylogenetic diversity and does not follow the latitudinal Collins MD, Simberloff D, Connor EF (2011) Binary matrices and check- richness gradient. PLoS One 10:e0128264. https://doi.org/10.1371/ erboard distributions of birds in the Bismarck archipelago. J journal.pone.0128264 Biogeogr 38:2373–2383. https://doi.org/10.1111/j.1365-2699. Ferraz K, Beisiegel B, De Paula R, Sana DA, De Campos CB, De 2011.02506.x Oliveira TG, Desbiez ALJ (2012) How species distribution models – Croft DA (2001) Cenozoic environmental change in South America as can improve cat conservation-jaguars in Brazil. CATNews 38 42 indicated by mammalian body size distributions (cenograms). Ferrusquía-Villafranca I, Arroyo-Cabrales J, Martínez-Hernández E, Divers Distrib 7:271–287. https://doi.org/10.1046/j.1366-9516. Gama-Castro J, Ruiz-González J, Polaco OJ, Johnson E (2010) 2001.00117.x Pleistocene mammals of Mexico: a critical review of regional chronofaunas, climate change response and biogeographic provin- Croitor R, Brugal J-P (2010) Ecological and evolutionary dynamics of the ciality. Quat Int 217:53–104. https://doi.org/10.1016/j.quaint.2009. carnivore community in Europe during the last 3 million years. Quat 11.036 Int 212:98–108. https://doi.org/10.1016/j.quaint.2009.06.001 Fotheringham AS, Charlton ME, Brunsdon C (1998) Geographically Culver M, Johnson WE, Pecon-Slattery J, O’Brien SJ (2000) Genomic weighted regression: a natural evolution of the expansion method ancestry of the American puma (Puma concolor). J Hered 91:186– for spatial data analysis. Environ Plan A Econ Sp 30:1905–1927. 197. https://doi.org/10.1093/jhered/91.3.186 https://doi.org/10.1068/a301905 Davies TJ, Buckley LB, Grenyer R, Gittleman JL (2011) The influence of Fotheringham, AS, Brunsdon C, Charlton M (2002) Geographically past and present climate on the biogeography of modern mammal Weighted Regression: The Analysis of Spatially Varying diversity. Philos Trans R Soc B Biol Sci 366:2526–2535. https://doi. Relationships. John Wiley and Sons, Chichester org/10.1098/rstb.2011.0018 Gaston KJ (2003) The Structure and Dynamics of Geographic Ranges. Davies TJ, Purvis A, Gittleman JL (2009) Quaternary climate change and Oxford University Press, Oxford – the geographic ranges of mammals. Am Nat 174:297 307. https:// Getis A, Ord JK (2010) The analysis of spatial association by use of doi.org/10.1086/603614 distance statistics. Geogr Anal 24:189–206. https://doi.org/10. Davis MB, Shaw RG (2001) Range shifts and adaptive responses to 1111/j.1538-4632.1992.tb00261.x – Quaternary climate change. Science 292:673 679 Giarla TC, Jansa SA (2015) The impact of Quaternary climate oscillations De Vivo M, Carmignotto AP (2004) Holocene vegetation change and the on divergence times and historical population sizes in Thylamys – mammal faunas of South America and Africa. J Biogeogr 31:943 opossums from the Andes. Mol Ecol 24:2495–2506. https://doi. 957. https://doi.org/10.1111/j.1365-2699.2004.01068.x org/10.1111/mec.13173 Di Rienzo JA, Casanoves F, Balzarini MG, Gonzalez L, Tablada M, Gittleman JL, Gompper ME (2005) Plight of predators: the importance of Robledo CW (2016) InfoStat. Grupo InfoStat, FCA, Universidad carnivores for understanding patterns of biodiversity and extinction Nacional de Córdoba risk. In: Barbosa P, Castellanos I (eds) Ecology of Predator-Prey Diniz‐Filho JAF, Rodríguez MA, Bini LM, Olalla‐Tarraga MA, Cardillo Interactions. Oxford University Press, Oxford, pp 371–388 M, Nabout JC, Hortal J, Hawkins BA (2009) Climate history, hu- González-Maya JF (2015) Conservación, diversidad funcional y riesgo manimpactsandglobalbodysize of Carnivora(Mammalia: de extinción en mamíferos neotropicales a múltiples escalas. Tesis Eutheria) at multiple evolutionary scales. J Biogeogr 36:2222– de Doctorado, Universidad Nacional Autónoma de México, México 2236. https://doi.org/10.1111/j.1365-2699.2009.02163.x City Dundas R (1999) Quaternary records of the dire wolf, Canis dirus,in González-Maya JF, Arias-Alzate A, Granados-Peña R, Mancera- North and South America. Boreas 28:375–385 Rodríguez NJ, Ceballos G (2016a) Environmental determinants Dyke AS (2005) Late Quaternary vegetation history of northern North and spatial mismatch of mammal diversity measures in Colombia. America based on pollen, macrofossil, and faunal remains. Anim biodivers Conserv 39:77–87. https://doi.org/10.32800/abc. Géographie Phys Quat 59:211. https://doi.org/10.7202/014755ar 2016.39.0077 Dynesius M, Jansson R (2000) Evolutionary consequences of changes in González-Maya JF, Víquez-R LR, Arias-Alzate A, Belant J, Ceballos G species’ geographical distributions driven by Milankovitch climate (2016b) Spatial patterns of species richness and functional diversity oscillations. Proc Natl Acad Sci USA 97:9115–9120. https://doi.org/ in Costa Rican terrestrial mammals: implications for conservation. 10.1073/pnas.97.16.9115 Divers Distrib 22:43–56. https://doi.org/10.1111/ddi.12373 J Mammal Evol

Goswami A, Friscia A (eds) (2010) Carnivoran Evolution. Cambridge Lyons SK (2003) A quantitative assessment of the range shifts of University Press, New York Pleistocene mammals. J Mammal 84:385–402. https://doi.org/10. Gotelli NJ, Ulrich W (2012) Statistical challenges in null model analysis. 1644/1545-1542(2003)084<0385:AQAOTR>2.0.CO;2 Oikos 121:171–180. https://doi.org/10.1111/j.1600-0706.2011. Marino J, Bennett M, Cossios D, Iriarte A, Lucherini M, Pliscoff P, 20301.x Sillero-Zubiri C, Villalba L, Walker S (2011) Bioclimatic constraints Graham RW (2001) Late Quaternary biogeography and extinction of to Andean cat distribution: a modelling application for rare species. proboscideans in North America. In: Cavaretta, G, Giola P, Mussi Divers Distrib 17:311–322. https://doi.org/10.1111/j.1472-4642. M, Palombo MR (eds) The World of Elephants (Lat Terra Degli 2011.00744.x Elephanti). Consiglio Nazionale della Richerche, Rome, pp 707– Marshall LG, Webb SD, Sepkoski JJ, Raup DM (1982) Mammalian 709 evolution and the great American interchange. Science 215:1351– Graham RW, Mead JI (1987) Environmental fluctuations and evolution of 1357. https://doi.org/10.1126/science.215.4538.1351 mammalian faunas during the last deglaciation in North America. In: Martínez-Meyer E, Peterson AT, Hargrove WW (2004) Ecological niches North America and Adjacent Oceans During the Last Deglaciation. as stable distributional constraints on mammal species, with impli- Geological Society of America, Boulder, pp 371–402 cations for Pleistocene extinctions and climate change projections Green JR, Ebdon D (1977) Statistics in geography: a practical approach. for biodiversity. Glob Ecol Biogeogr 13:305–314. https://doi.org/10. Appl Stat 26:341. https://doi.org/10.2307/2346984 1111/j.1466-822X.2004.00107.x Gregory-Wodzicki KM (2000) Uplift history of the central and northern McPherson JM, Jetz W, Rogers DJ (2004) The effects of species’ range Andes: a review. Geol Soc Am Bull 112:1091–1105. https://doi.org/ sizes on the accuracy of distribution models: ecological phenome- 10.1130/0016-7606(2000)112<1091:UHOTCA>2.0.CO;2 non or statistical artefact? J Appl Ecol 41:811–823. https://doi.org/ Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high 10.1111/j.0021-8901.2004.00943.x resolution interpolated climate surfaces for global land areas. Int J Merow C, Smith MJ, Silander JA (2013) A practical guide to MaxEnt for ’ Climatol 25:1965–1978. https://doi.org/10.1002/joc.1276 modeling species distributions: what it does, and why inputs and – Hubbe A, Hubbe M, Neves W (2007) Early Holocene survival of mega- settings matter. Ecography 36:1058 1069. https://doi.org/10.1111/j. fauna in South America. J Biogeogr 34:1642–1646. https://doi.org/ 1600-0587.2013.07872.x 10.1111/j.1365-2699.2007.01744.x Morris AB, Graham CH, Soltis DE, Soltis PS (2010) Reassessment of Jernvall J, Fortelius M (2004) Maintenance of trophic structure in fossil phylogeographical structure in an eastern north American tree using Monmonier’s algorithm and ecological niche modelling. J Biogeogr mammal communities: site occupancy and taxon resilience. Am Nat – 164:614–624. https://doi.org/10.1086/424967 37:1657 1667. https://doi.org/10.1111/j.1365-2699.2010.02315.x Muscarella R, Galante PJ, Soley-Guardia M, RA Kass JM, Uriarte M, Johnson WE, Eizirik E, Peco-Slattery J, Murphy WJ, Antunes A, Teeling Anderson RP (2014) ENMeval: an R package for conducting spa- E, O’Brien SJ (2006) The late Miocene radiation of modern Felidae: tially independent evaluations and estimating optimal model com- a genetic assessment. Science 311:73–77. https://doi.org/10.1126/ plexity for Maxent ecological niche models. Methods Ecol Evol 5: science.1122277 1198–1205. https://doi.org/10.1111/2041-210X.12261 Levinsky I (2010) Species Distributions and Climate Change-Linking the Nogués-Bravo D (2009) Predicting the past distribution of species climat- Past and the Future. PhD Dissertation, University of Copenhagen, ic niches. Glob Ecol Biogeogr 18:521–531. https://doi.org/10.1111/ Copenhagen j.1466-8238.2009.00476.x Levinsky I, Araújo MB, Nogués-Bravo D, Haywood AM, Valdes Nogués-Bravo D, Rodríguez J, Hortal J, Batra P, Araújo MB (2008) PJ, Rahbek C (2013) Climate envelope models suggest Climate change, humans, and the extinction of the woolly mam- spatio-temporal co-occurrence of refugia of African birds moth. PLoS Biol 6:e79. https://doi.org/10.1371/journal.pbio. and mammals. Glob Ecol Biogeogr 22:351–363. https://doi. 0060079 org/10.1111/geb.12045 Nyakatura K, Bininda-Emonds OR (2012) Updating the evolutionary Li X, Jiang G, Tian H, Xu L, Yan C, Wang Z, Wei F, Zhang Z (2014) history of Carnivora (Mammalia): a new species-level supertree Human impact and climate cooling caused range contraction of large complete with divergence time estimates. BMC Biol 10:12. https:// mammals in China over the past two millennia. Ecography 38(1): doi.org/10.1186/1741-7007-10-12 74–82. https://doi.org/10.1111/ecog.00795 O’Brien RM (2007) A caution regarding rules of thumb for variance Lobo JM, Jiménez-Valverde A, Real R (2008) AUC: a misleading mea- inflation factors. Qual Quant 41:673–690. https://doi.org/10.1007/ sure of the performance of predictive distribution models. Glob Ecol s11135-006-9018-6 Biogeogr 17:145–151. https://doi.org/10.1111/j.1466-8238.2007. O’Regan HJ, Turner A, Wilkinson DM (2002) European quaternary 00358.x refugia: a factor in large carnivore extinction? J Quaternary Sci 17: Lorenzen ED, Nogués-Bravo D, Orlando L, Weinstock J, Binladen J, 789–795. https://doi.org/10.1002/jqs.693 Marske KA, Ugan A, Borregaard MK, Gilbert MTP, Nielsen R, Ord JK, Getis A (2010) Local spatial autocorrelation statistics: distribu- Ho SYW, Goebel T, Graf KE, Byers D, Stenderup JT, Rasmussen tional issues and an application. Geogr Anal 27:286–306. https:// M, Campos PF, Leonard JA, Koepfli K-P, Froese D, Zazula G, doi.org/10.1111/j.1538-4632.1995.tb00912.x Stafford TW, Aaris-Sørensen K, Batra P, Haywood AM, Otto-Bliesner BL, Brady EC, Clauzet G, Tomas R, Levis S, Kothavala Z Singarayer JS, Valdes PJ, Boeskorov G, Burns JA, Davydov SP, (2006) Last glacial maximum and Holocene climate in CCSM3. J Haile J, Jenkins DL, Kosintsev P, Kuznetsova T, Lai X, Martin Climatol 19:2526–2544. https://doi.org/10.1175/JCLI3748.1 LD, McDonald HG, Mol D, Meldgaard M, Munch K, Stephan E, Sablin M, SommerRS, SipkoT, Scott E, Suchard MA, Tikhonov A, Palombo MR, Alberdi MT, Azanza B, Giovinazzo C, Prado JL, Sardella Willerslev R, Wayne RK, Cooper A, Hofreiter M, Sher A, Shapiro R (2009) How did environmental disturbances affect carnivoran diversity? A case study of the Plio-Pleistocene Carnivora of the B, Rahbek C, Willerslev E (2011) Species-specific responses of late – Quaternary megafauna to climate and humans. Nature 479:359– North-Western Mediterranean. Evol Ecol 23:569 589. https://doi. 364. https://doi.org/10.1038/nature10574 org/10.1007/s10682-008-9256-2 Peterson AT, Papeş M, Eaton M (2007) Transferability and model eval- Lundberg P, Ranta E, Ripa J, Kaitala V (2000) Population variability in uation in ecological niche modeling: a comparison of GARP and space and time. Trends Ecol Evol 15:460–464. https://doi.org/10. Maxent. Ecography 30:550–560. https://doi.org/10.1111/j.2007. 1016/S0169-5347(00)01981-9 0906-7590.05102.x JMammalEvol

Peterson AT, Soberón J, Pearson RG, Anderson RP, Martínez-Meyer E, Schipper J, Chanson JS, Chiozza F, Cox NA, Hoffmann M, Katariya V, Nakamura M, Bastos-Araújo M (2011) Ecological Niches and Lamoreux J, Rodrigues ASL, Stuart SN, Temple HJ, Baillie J, Geographic Distributions. Princeton University Press, Princeton Boitani L, Lacher TE, Mittermeier RA, Smith AT, Absolon D, Phillips SJ, Anderson RP, Dudík M, Schapire RE, Blair ME (2017) Aguiar JM, Amori G, Bakkour N, Baldi R, Berridge RJ, Bielby J, Opening the black box: an open-source release of Maxent. Black PA, Blanc JJ, Brooks TM, Burton JA, Butynski TM, Catullo Ecography 40:887–893. https://doi.org/10.1111/ecog.03049 G, Chapman R, Cokeliss Z, Collen B, Conroy J, Cooke JG, da Phillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy model- Fonseca GAB, Derocher AE, Dublin HT, Duckworth JW, ing of species geographic distributions. Ecol Modell 190:231–259. Emmons L, Emslie RH, Festa-Bianchet M, Foster M, Foster S, https://doi.org/10.1016/j.ecolmodel.2005.03.026 Garshelis DL, Gates C, Gimenez-Dixon M, Gonzalez S, Phillips SJ, Dudík M (2008) Modeling of species distributions with Gonzalez-Maya JF, Good TC, Hammerson G, Hammond PS, Maxent: new extensions and a comprehensive evaluation. Happold D, Happold M, Hare J, Harris RB, Hawkins CE, Ecography 31:161–175. https://doi.org/10.1111/j.0906-7590.2008. Haywood M, Heaney LR, Hedges S, Helgen KM, Hilton-Taylor 5203.x C, Hussain SA, Ishii N, Jefferson T A, Jenkins RKB, Johnston Polly PD (2013) Tiptoeing through the trophics: geographic variation in CH, Keith M, Kingdon J, Knox DH, Kovacs KM, Langhammer P, carnivoran locomotor ecomorphology in relation to environment. In: Leus K, Lewison R, Lichtenstein G, Lowry LF, Macavoy Z, Mace Goswami A, Friscia A (eds) Carnivoran Evolution. Cambridge GM, Mallon DP, Masi M, McKnight MW, Medellin RA, Medici P, University Press, Cambridge, pp 374–410 MillsG,MoehlmanPD,MolurS,MoraA,NowellK,OatesJF, Porter SC (1981) Pleistocene glaciation in the southern Lake District of Olech W, Oliver WRL, Oprea M, Patterson BD, Perrin WF, Chile. Quaternary Res 16:263–292. https://doi.org/10.1016/0033- Polidoro BA, Pollock C, Powel A, Protas Y, Racey P, Ragle J, 5894(81)90013-2 Ramani P, Rathbun G, Reeves RR, Reilly SB, Reynolds JE, Prevosti FJ, Reguero M (2000) Catalogo de tipos de vertebrados fósiles Rondinini C, Rosell-Ambal RG, Rulli M, Rylands AB, Savini S, del museo de la Plata. I. Carnivora. Serie técnica y didáctica del Schank CJ, Sechrest W, Self-Sullivan C, Shoemaker A, Sillero- MuseodelaPlata,BuenosAires Zubiri C, De Silva N, Smith DE, Srinivasulu C, Stephenson PJ, van Strien N, Talukdar BK, Taylor BL, Timmins R, Tirira DG, Prevosti FJ, Vizcaíno S (2006) Paleoecology of the large carnivore guild Tognelli MF, Tsytsulina K, Veiga LM, Vie J-C, Williamson EA, from the late Pleistocene of Argentina. Acta Palaeontol Polonica 51: Wyatt SA, Xie Y, Young BE (2008) The status of the world’sland 407–422 and marine mammals: diversity, threat, and knowledge. Science Prevosti FJ, Rincón AD (2007) A new fossil canid assemblage from the 322:225–230. https://doi.org/10.1126/science.1165115 late Pleistocene of northern South America: the canids of the Inciarte Silvestro D, Antonelli A, Salamin N, Quental TB (2015) The role of clade asphalt pit (Zulia, Venezuela), fossil record and biogeography. J competition in the diversification of North American canids. Paleontol 81:1053–1065. https://doi.org/10.1666/pleo05-143.1 Proceedings of the National Academy of Sciences 112 (28):8684– Prevosti FJ, Soibelzon L (2012) Evolution of the South American carni- 8689. https://doi.org/10.1073/pnas.1502803112 vores (Mammalia, Carnivora): a paleontological perspective. In: Soares S de A (2013) The role of competition in structuring ant commu- Patterson BD, Costa LP (eds) Bones, Clones, and Biomes: An nities: a review. Oecologia Aust 17:271–281. https://doi.org/10. Extended History of Recent Neotropical Mammals. University of 4257/oeco.2013.1702.08 Chicago Press, Chicago, pp 102–122 Soberón J, Nakamura M (2009) Niches and distributional areas: concepts, Prevosti FJ, Forasiepi AM (2018) Evolution of South American methods, and assumptions. Proc Natl Acad Sci USA 106:19644– Mammalian Predators during the Cenozoic: Paleobiogeographic 19650. https://doi.org/10.1073/pnas.0901637106 and Paleoenvironmental Contingencies. Springer International Soibelzon L, Prevosti FJ (2008) Los carnívoros (Carnivora, Mammalia) Publishing, Cham terrestres del Cuaternario de América del Sur. In: Pons GX, Vicens Prevosti FJ, Noriega J, Garcia-Esponda C, Ferrero B (2005) Primer D (eds) Geomorfología Litoral i Quaternari. Homenatge a Joan registro de Dusicyon gymnocercus (Fisher, 1814) (Carnivora: Cuerda Barceló. Monografías de la Sociedad de Historia Natural Canidae) en el pleistoceno de Entre Ríos (Argentina). Revista de las Baleares, Palma de Mallorca, pp 49–68 – Española Paleontología 20:159 167 Steinitz O, Heller J, Tsoar A, Rotem D, Kadmon R (2006) Environment, Prevosti FJ, Forasiepi A, Zimicz N (2013) The evolution of the Cenozoic dispersal and patterns of species similarity. J Biogeogr 33:1044– terrestrial mammalian predator guild in South America: competition 1054. https://doi.org/10.1111/j.1365-2699.2006.01473.x – or replacement? J Mammal Evol 20:3 21. https://doi.org/10.1007/ Stevens GC (1989) The latitudinal gradient in geographical range: how so s10914-011-9175-9 many species coexist in the tropics. The American Naturalist 133(2): Prideaux GJ, Roberts RG, Megirian D, Westaway KE, Hellstrom JC, 240–256 Olley JM (2007) Mammalian responses to Pleistocene climate Svenning J-C, Skov F (2004) Limited filling of the potential range in change in southeastern Australia. Geology 35(1):33–36. https:// European tree species. Ecol Lett 7:565–573. https://doi.org/10. doi.org/10.1130/G23070A.1 1111/j.1461-0248.2004.00614.x Rincón AD, Prevosti FJ, Parra GE (2011) New saber-toothed cat records Svenning J-C, Skov F (2007) Could the tree diversity pattern in Europe be (Felidae: Machairodontinae) for the Pleistocene of Venezuela, and generated by postglacial dispersal limitation? Ecol Lett 10:453–460. the great American biotic interchange. J Vertebr Paleontol 31:468– https://doi.org/10.1111/j.1461-0248.2007.01038.x 478. https://doi.org/10.1080/02724634.2011.550366 Thibault KM, Brown JH (2008) Impact of an extreme climatic event on Root T (1988) Environmental factors associated with avian distributional community assembly. Proc Natl Acad Sci USA 105:3410–3415. boundaries. J Biogeogr 15:489. https://doi.org/10.2307/2845278 https://doi.org/10.1073/pnas.0712282105 Safi K, Cianciaruso M V, Loyola RD, Brito D, Armour-Marshall K, Thomas CD, Cameron A, Green RE,BakkenesM,BeaumontLJ, Diniz-Filho JAF (2011) Understanding global patterns of mamma- Collingham YC, Erasmus BFN, De Siqueira MF, Grainger A, lian functional and phylogenetic diversity. Philos Trans R Soc B Hannah L, Hughes L, Huntley B, Van Jaarsveld AS, Midgley GF, Biol Sci 366:2536–2544. https://doi.org/10.1098/rstb.2011.0024 Miles L, Ortega-Huerta M a, Peterson a t, Phillips OL, Williams SE Santika T (2011) Assessing the effect of prevalence on the predictive (2004) Extinction risk from climate change. Nature 427(6970):145– performance of species distribution models using simulated data. 148. https://doi.org/10.1038/nature02121 Glob Ecol Biogeogr 20:181–192. https://doi.org/10.1111/j.1466- Thuiller W, Lavorel S, Araujo MB (2005) Niche properties and geograph- 8238.2010.00581.x ical extent as predictors of species sensitivity to climate change. J Mammal Evol

Glob Ecol Biogeogr 14:347–357. https://doi.org/10.1111/j.1466- Wilson D, Mittermeier RA (2009) Handbook of the Mammals of the 822X.2005.00162.x World. Vol. I. Carnivores. Lynx Edicions, Barcelona Varela L, Fariña RA, (2016) Co-occurrence of mylodontid sloths and Wilson DE, Reeder DM (eds) (2005) Mammal species of the world: a insights on their potential distributions during the late Pleistocene. taxonomic and geographic reference. Third Edition. Johns Hopkins Quaternary Res 85(1):66–74. https://doi.org/10.1016/j.yqres.2015. University Press, Baltimore 11.009 Wisz MS, Pottier J, Kissling WD, Pellissier L, Lenoir J, Damgaard CF, Villavicencio NA, Lindsey EL, Martin FM, Borrero LA, Moreno PI, Dormann CF, Forchhammer MC, Grytnes JA, Guisan A, Heikkinen Marshall CR, Barnosky AD (2016) Combination of humans, cli- RK, Høye TT, Kühn I, Luoto M, Maiorano L, Nilsson MC, mate, and vegetation change triggered late Quaternary megafauna Normand S, Öckinger E, Schmidt NM, Termansen M, extinction in the Última Esperanza region, southern Patagonia, Timmermann A, Wardle DA, Aastrup P, Svenning JC (2013) The Chile. Ecography 39:125–140. https://doi.org/10.1111/ecog.01606 role of biotic interactions in shaping distributions and realised as- Wagenmaker EJ, Farrell S (2004) AIC model selection using Akaike semblages of species: implications for species distribution model- weights. Psychon Bull Rev 11:192–196 ling. Biological Reviews 88(1):15–30. https://doi.org/10.1111/j. Webb SD (2006) The great American biotic interchange: patterns and 1469-185X.2012.00235.x processes. Ann Missouri Bot Gard 93:245–257