SHORT COMMUNICATION The use of native vegetation as a proxy for habitat may overestimate habitat availability in fragmented landscapes Mauricio Almeida-Gomes . Jayme Augusto Prevedello . Renato Crouzeilles Received: 3 August 2015 /Accepted: 24 November 2015 / Published online: 15 December 2015 � Springer Science+Business Media Dordrecht 2015 Abstract Context Native vegetation is often used as a proxy for habitat to estimate habitat availability in land- scapes. This approach may lead to incorrect estimates of the impacts of habitat loss and fragmentation on species, which have not been thoroughly quantified so far. Objectives We quantified to what extent the loss of native vegetation reflect actual habitat loss by native species in landscapes. We tested the hypothesis that habitat availability declines at greater rates than native vegetation and thus is overestimated when it is quantified on the basis of native vegetation. Methods Using simulations, we quantified how the loss of native vegetation in artificial and real land- scapes affects habitat availability for species with different habitat requirements. We contrasted a gen- eralist species, which uses all native vegetation, with 10 habitat-specialist species classified into three categories (interior, patchy and riparian species). Results Habitat availability generally declined at greater rates than native vegetation for all specialist species. This pattern was apparent for different specialist species in a broad range of landscape types. Interior species always lost habitat availability more rapidly than the generalist species. Most riparian species lost habitat availability more rapidly than the generalist species. Responses of patchy species were more complex, depending on their dispersal abilities and landscape structure. Conclusions Habitat availability is likely to be overestimated when native vegetation is used as proxy for habitat, because habitat availability will generally decline at greater rates than native vegetation. There- fore, a species-centered approach should be adopted when estimating habitat availability in landscapes. Keywords Extinction risks � Fragmented landscapes � Habitat reachability � Habitat loss � Probability of connectivity � Species-centered approach � Species persistence Mauricio Almeida-Gomes, Jayme Augusto Prevedello and Renato Crouzeilles have contributed equally to this manuscript. M. Almeida-Gomes (&) � R. Crouzeilles Laboratory of Vertebrates, Department of Ecology, Federal University of Rio de Janeiro, Avenida Carlos Chagas Filho 373, Cidade Universitária, Rio de Janeiro, RJ 21941-902, Brazil e-mail: almeida.gomes@yahoo.com.br J. A. Prevedello Laboratory of Landscape Ecology and Conservation, Department of Ecology, State University of São Paulo, Rua do Matão, São Paulo, SP 05508-900, Brazil Present Address: J. A. Prevedello Laboratory of Landscape Ecology, State University of Rio de Janeiro, Rio de Janeiro, RJ 20550-900, Brazil R. Crouzeilles International Institute for Sustainability, Rio de Janeiro, RJ 22460-320, Brazil 123 Landscape Ecol (2016) 31:711–719 DOI 10.1007/s10980-015-0320-3 http://orcid.org/0000-0001-7938-354X http://crossmark.crossref.org/dialog/?doi=10.1007/s10980-015-0320-3&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1007/s10980-015-0320-3&domain=pdf Introduction The process of habitat loss and fragmentation is the main driver of the current worldwide biodiversity decline (Fahrig 2003). Although all species depend on habitat and thus may be affected by habitat loss and fragmentation, there is large variability in their actual responses to this process (Ewers and Didham 2006). A major challenge for ecologists and conservation biologists is to understand the causes of such variabil- ity (Henle et al. 2004). The variability in species responses to habitat loss and fragmentation could result, at least in part, from the mismatch between how species and ecologists perceive the landscape (Betts et al. 2014). Many ecological studies have measured landscapes using human-defined cover types, assuming for example that ‘‘habitat’’ is the same as ‘‘native vegetation’’ (Fischer and Lindenmayer 2007). This assumption has been common, for example, in analysis of species- and density-area relationships, in which ‘‘area’’ is com- monly measured on the basis of native vegetation (see e.g. Connor et al. 2000; Rybicki and Hanski 2013). This approach may have led to incorrect estimates of habitat loss and fragmentation for many species (Ewers and Didham 2006; Betts et al. 2014). As an alternative, the adoption of a species-centered approach to estimate habitat amount and availability could increase our ability to predict species’ abun- dance and distribution in landscapes (Betts et al. 2014). For example, species distribution models can be used to estimate habitat suitability for particular species across landscapes (e.g. Cabeza et al. 2004; Keith et al. 2008). Adoption of the species-centered approach can be essential when studying habitat specialist species, which are frequently highly sensitive to habitat loss and fragmentation (Krämer et al. 2012; Monks and Burrows 2014). Examples are core-dependent species (e.g. birds and insects; Villard 1998; Peyras et al. 2013), riparian species occupying narrow areas within native vegetation (e.g. stream-dwelling amphibians and insects; Almeida-Gomes et al. 2014; Suhonen et al. 2014), and species with naturally patchy distributions (e.g. amphibians and butterflies; Saccheri et al. 1998; Heard et al. 2012). Some studies have shown, for example, that core-dependent species lose habitat disproportionately as native vegetation is reduced, due to an increase in the amount of edges (Villard 1998; Henle et al. 2004). In addition, even if habitat amount is reduced in the same proportion as native vegetation, this may not be the case for habitat availability. The availability of a particular habitat depends not only on its amount in the landscape but also on landscape connectivity, as disconnected habitat cannot be reached and used by individuals (Saura and Pascual-Hortal 2007). As habitat availability is likely to be critical for popula- tion persistence in fragmented landscapes (Fahrig 2003; Awade et al. 2012), it is essential to understand when it can be accurately estimated from native vegetation. Here, we quantify to what extent the loss of native vegetation reflect actual habitat loss by native species in landscapes. We tested the hypothesis that habitat availability declines at greater rates than native vegetation and thus is overestimated when it is quantified on the basis of native vegetation. To do so, we quantify how the loss and fragmentation of native vegetation affects habitat availability for species with different habitat requirements and dis- persal abilities, through an extensive modeling study in artificial and real landscapes from a biodiversity hotspot. Methods Overview To quantify the effects of the loss and fragmentation of native vegetation on habitat availability, we used artificial and real landscapes varying in the amount, degree of fragmentation and spatial extent of native vegetation. Habitat availability was quantified for one ‘‘generalist’’ species, which is able to use all native vegetation as habitat, and for species which use only particular habitats within native vegetation (i.e. ‘‘inte- rior’’, ‘‘riparian’’ and ‘‘patchy’’ species; Fig. 1).We also performed a sensitivity analysis to evaluate how the mobility of each species affected habitat availabil- ity. We restricted our analyses to hypothetical rather than real species to be able to contrast a wide range of species with different habitat requirements and disper- sal abilities. Finally, we also quantified habitat amount for each species in addition to habitat availability. 712 Landscape Ecol (2016) 31:711–719 123 Native cover 100% 80% 60% 40% 20% Generalist Interior 50m Interior 200m Patchy 50% clumped Patchy 50% scattered Patchy 20% clumped Patchy 20% scattered Riparian 40m 30% Riparian 15m 30% Riparian 40m 10% Riparian 15m 10% (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) (k) Fig. 1 Example of how the reduction of native vegetation (from 100 to 20 %) affects habitat distributions of different species, in a simulated landscape with a low degree of fragmentation. The generalist species (a) occupied all native vegetation, whereas the different specialist species (b–k) occupied only more restricted portions within the native vegetation. Interior species occupied only areas of native vegetation located at least 50 or 200 m away from edges. Patchy species occupied randomly- scattered portions (either 20 or 50 %) of the native vegetation, in a pattern either highly fragmented or highly clumped. Riparian species were restricted to the margins of rivers located within native vegetation. Two amounts of rivers in the landscape (10 or 30 %), as well as two buffer distances from river margins (15 and 40 m), were considered to delimit the habitat of these species Landscape Ecol (2016) 31:711–719 713 123 Studied landscapes We first performed simulations using artificial land- scapes generated with the modified random cluster method (Saura and Martı́nez-Millán 2000). These landscapes had 200 9 200 pixels, composed of either native vegetation or matrix pixels. We created frag- mented landscapes by sequentially removing native vegetation from an originally continuous landscape (100 % cover), in steps of 10 % (±2 %; see Fig. 1a), thus simulating the progressive loss and fragmentation of native vegetation.We removed native vegetation by converting native vegetation pixels to matrix pixels, disregarding the location of the particular habitats of each species, thus avoiding any bias in favor or against a particular species. However, removal was not random but controlled by a specific parameter, ‘‘p’’, with p = 0 resulting in a completely scattered distri- bution of native vegetation and p & 0.593 resulting in a highly clumped distribution (Saura and Martı́nez- Millán 2000). We set p to either 0.10 or 0.56, to create landscapes with high and low degree of fragmentation in the native vegetation, respectively (see Fig. 1a). We built a total of 200 neutral landscapes, encompassing 10 replicates for each combination of native vegeta- tion (from 100 to 10 %) and degree of fragmentation (p = 0.10 or 0.56). We also conducted simulations on real landscapes of the Atlantic Forest biodiversity hotspot. We ran- domly selected 100 landscapes from the entire Atlantic Forest, 10 for each percentage of native vegetation, varying from 100 to 10 % in steps of 10 % (±5 %). For riparian species, we also selected 20 landscapes with either low (10 %; n = 10) or high (30 %; n = 10) amounts of rivers (obtained from INEA-RJ). This allowed obtaining realistic distribu- tions of habitats for riparian species (see Fig. 1h–k). For generalist, interior and patchy species, we set pixel resolution in both artificial and real landscapes to 50 m, resulting in landscapes with 10,000 ha, simi- larly to many empirical and simulation studies conducted in the Atlantic Forest (e.g. Banks-Leite et al. 2010; Crouzeilles et al. 2014). We set minimum patch size to 3 ha (12 pixels) as in the most recent map of Atlantic Forest remnants (SOS Mata Atlântica and INPE 2012). For riparian species, we used the same landscapes, but assumed a pixel resolution of 10 m (as in the original database) to better depict the linear habitat of these species, resulting in 400-ha landscapes with minimum patch size of 0.12 ha. Comparisons among species were still valid because the analyses focused only on the proportional (rather than the raw) loss of habitat amount and availability for each species. We created a buffer zone of four pixels (corresponding to 40 or 200 m for riparian or interior/patchy species, respectively) around each landscape to avoid underestimating habitat amount and availability close to landscape boundaries. We treated real landscapes in ArcGis 9.3, and performed simulations in R 2.12 environment (R Development Core Team 2011). Specialist species We considered two ‘‘Interior’’ species, each occupy- ing only areas of native vegetation located at least 50 or 200 m away from edges (e.g. Ewers and Didham 2008; Banks-Leite et al. 2010; Fig. 1b, c). We considered four ‘‘Patchy’’ species, each occupying either 20 or 50 % of the landscape, in a pattern either highly fragmented or highly clumped (Fig. 1d–g). Such distribution patterns for patchy species were generated in continuous landscapes by using the modified random cluster method (Saura and Martı́- nez-Millán 2000), setting p to either 0.10 or 0.56 (high and low levels of habitat fragmentation, respectively). We also considered four ‘‘Riparian’’ species, each occupying only areas within 15 or 40 m of river margins (Almeida-Gomes et al. 2014) for landscapes with either low (10 %) or high (30 %) amounts of rivers (Fig. 1h–k). We also varied the mobility of all species due to its potential effect on habitat availability (Saura and Rubio 2010). First, we varied inter-patch dispersal abilities of species as short or large relative to landscape size, setting median dispersal distances as either 100 or 3000 m for generalist, interior and patchy species (based on Crouzeilles et al. 2014), and as either 4 or 120 m for riparian species (based on Semlitsch and Bodie 2003). These median dispersal abilities correspond to a probability of 50 % of direct dispersal between two patches (Saura and Pascual-Hortal 2007). Riparian species had lower raw median dispersal abilities because their landscapes were correspond- ingly smaller; however, relative to landscape size, they had the same dispersal abilities as the other species. We also varied the ability of species tomove through native vegetation, by considering two extreme 714 Landscape Ecol (2016) 31:711–719 123 scenarios: (i) native vegetation is completely permeable tomovement or (ii) native vegetation and the matrix are equally permeable. In the first scenario, the distance between two habitat patches located at a same patch of native vegetation was zero, whereas in the second that distance corresponded to the Euclidean distance between the two habitat patches. Matrix permeability may affect habitat availability, but we kept it constant in all simulations, focusing on the differential use of native vegetation by each species only. Quantifying habitat availability and amount Wequantified habitat availability using the Probability of Connectivity (PC; Saura and Pascual-Hortal 2007). This index takes into account both the amount of habitat in the landscape and the functional connectivity, providing a robustmeasure of the amountofhabitat actually available to a species in a landscape (Saura and Pascual-Hortal 2007). PC has been increasingly used as an integrative metric for landscape analysis, because it requires few parameters to be computed (Saura and Pascual-Hortal 2007), is positively correlated with species occurrence in fragmented landscapes (Awade et al. 2012), and is also a good approximation method of metapopulation capacity (Crouzeilles et al. 2015). The PC index varies from 0, when no habitat is available, to 1, when the entire landscape is occupied by habitat (see Saura and Pascual- Hortal 2007). We calculated the PC index as: PC ¼ Pn i¼1 Pn j¼1 aiajp � ij A2 L where n represented the number of habitat patches in the landscape, ai and aj represented the sizes of a given pair of patches, p�ij represented the maximum proba- bility of connection between the two patches in the pair and AL 2 represented the square of the total area of the landscape (Saura and Rubio 2010). The maximum probability of connection (p�ij) was calculated by considering all the possible paths between the patches i and j (see Saura and Pascual-Hortal 2007 for details). We also quantified habitat amount for each species, i.e. the percentage of the landscape occupied by the habitat of each species. This analysis ignored the spatial configuration of habitat patches, thus allowing determining whether differences in habitat availability among species were caused by differences in the amount and/or in the configuration of habitat patches. Results Habitat availability Habitat availability decreased exponentially with the loss of native vegetation for all species in all scenarios simulated (Fig. 2). The interior species always lost habitat availability more rapidly than the generalist species, and this pattern was little affected by species’ dispersal abilities or native vegetation permeability. The differences between the generalist and the interior species were less evident in real than in neutral landscapes. Most riparian species lost habitat avail- ability more rapidly than the generalist species, especially when the native vegetation permeability was low. The only exception was the ‘‘riparian 15 m—30 %’’, which behaved similarly to the gener- alist when native vegetation permeability was high (Fig. 2, rows 1 and 2). Responses of patchy species to the loss of native vegetation were more complex, depending on their dispersal abilities, the permeability of native vegeta- tion and the degree of fragmentation (Fig. 2). When the permeability or dispersal abilities were high, all patchy species responded similarly to the generalist (Fig. 2, rows 1, 2 and 4). Otherwise, differences between patchy and the generalist species were clear, most of them losing habitat availability more rapidly than the generalist (Fig. 2, row 3). Habitat amount Habitat amount decreased linearly with the loss of native vegetation for the generalist, riparian and patchy species in both neutral and real landscapes (Fig. 3). However, for interior species, habitat amount decreased exponentially rather than linearly, occurring more rapidly than the loss of native vegetation. The differences between the generalist and the interior species were smaller and more variable in real (Fig. 3c) than in neutral landscapes (Fig. 3a, b). Discussion Our simulations show that habitat availability is overestimated in most landscapes when native vege- tation is used as proxy for habitat, as habitat availability decreases more rapidly than vegetation Landscape Ecol (2016) 31:711–719 715 123 Highly-fragmented neutral landscapes Little-fragmented neutral landscapes Real landscapes H ig h pe rm ea bi lit y Lo w d is pe rs al H ab ita t a va ila bi lit y (% ) H ig h di sp er sa l Lo w p er m ea bi lit y Lo w d is pe rs al H ig h di sp er sa l Native vegetation (%) 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 Generalist Patchy50%-Scattered Patchy50%-Clumped Patchy20%-Scattered Patchy20%-Clumped Interior50m Interior200m Riparian15m-10% Riparian15m-30% Riparian40m-10% Riparian40m-30% 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 Fig. 2 Effects of the loss of native vegetation on habitat availability (Probability of Connectivity index) for different species. Simulations were performed: (i) when native vegetation is either completely permeable to species movement or as permeable to species movement as the matrix; (ii) in highly- and little-fragmented neutral landscapes and in real landscapes of the Atlantic Forest hotspot; and (iii) for species with different dispersal abilities. Points depict the mean values of the amount of habitat recorded for each species across 10 replicate landscapes at each percentage of native vegetation. Bars represent 95 % confidence intervals calculated across the 10 replicate landscapes, and are sometimes smaller than the points. Depicted values are proportional to the mean values recorded for each species at the original landscapes (100 % of native vegetation) 716 Landscape Ecol (2016) 31:711–719 123 cover. This pattern was apparent for specialist species with different habitat requirements and mobility, in a broad range of landscape types. Therefore, our study supports the claim that a species-centered approach is essential to accurately predict organism’s responses to the loss of native vegetation cover (Betts et al. 2014). The importance of correctly identifying habitat to estimate habitat availability was evident not only when comparing generalists and specialists, but also when contrasting the different specialist species. Interior species were the only species losing habitat amount (in addition to habitat availability) more rapidly than the generalist. This occurs because interior species lose not only those habitat areas that are removed together with native vegetation, but also areas located near habitat edges, as acknowledged in previous studies (e.g. Banks-Leite et al. 2010; Vil- laseñor et al. 2014). Our study also shows the decline in habitat availability for interior species, which may contribute to explain their disappearance from some highly fragmented landscapes (e.g. Banks-Leite et al. 2010). Contrary to the interior species, the riparian and patchy species lost habitat amount in the same proportion as the generalist. However, for the riparian species, habitat availability was drastically reduced by the loss of native vegetation, especially when native vegetation was less permeable to species movement. The linear configuration of habitat patches for these species makes them vulnerable to even small losses of native vegetation, which may break habitat apart causing abrupt decreases in patch sizes and increases in patch isolation (see Fig. 1). Such structural changes may reduce habitat availability (Crouzeilles et al. 2014) and possibly population viability (e.g. Öckinger et al. 2010; Mari et al. 2014), and thus may be a key mechanism driving population declines of many stream-dwelling amphibians and reptiles in frag- mented landscapes (Gardner et al. 2007; Almeida- Gomes et al. 2014). However, protection of riparian zones by law may reduce the loss of native vegetation and thus the loss of habitat availability for riparian species (Metzger et al. 2010). Patchy species differed from the generalist only when their mobility was low. When this occurred, H ab ita t a m ou nt (% ) Native vegetation (%) 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100(a) 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100(b) Generalist Patchy50%-Scattered Patchy50%-Clumped Patchy20%-Scattered Patchy20%-Clumped Interior50m Interior200m Riparian15m-10% Riparian15m-30% Riparian40m-10% Riparian40m-30% 100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100(c) bFig. 3 Effects of the loss of native vegetation on the amount of habitat for different species. Simulations encompassed highly- fragmented neutral landscapes (a), little-fragmented neutral landscapes (b), or real landscapes of the Atlantic Forest hotspot (c). Symbols as in Fig. 2 Landscape Ecol (2016) 31:711–719 717 123 most patchy species lost habitat availability more rapidly than the generalist, due to the abrupt decrease in connectivity, as reported by Andrén (1994) for random landscapes. Our simulations suggest that responses of patchy species to the loss of native vegetation cover will be complex, depending on their mobility and the amount and degree of aggregation of their original habitat. This variability may explain why some butterfly and plant species with patchy distribu- tions disappear with the loss of native vegetation, whereas others are able to survive even in highly fragmented landscapes (e.g. Devictor et al. 2008; Brückmann et al. 2010). Moreover, some real patchy species could have higher dispersal abilities than generalist species as a result of adaptive responses to habitat patchiness (Fahrig 2007), which could make them equally or even less sensitive to the loss of native cover than more generalist species. Our findings have important implications for the management and conservation of specialist species in fragmented landscapes. Habitat availability for many specialist species may have been overestimated in landscapes, considering that it is frequently estimated under the assumption that ‘‘habitat’’ is the same as ‘‘native vegetation’’ (Fischer and Lindenmayer 2007; Betts et al. 2014), or at least directly proportional to it. Similarly, many estimates of species loss in commu- nities have been made using species-area relation- ships, measuring ‘‘area’’ as ‘‘native vegetation cover’’ (Rybicki and Hanski 2013). This is likely to underes- timate species loss if many specialist species are present, considering that the true habitat availability for those species is probably overestimated. To accurately predict the extent of habitat and species loss in landscapes, more effort should be expended on estimating the original distributions of species’ habi- tats in landscapes, for example using habitat suitability models (Cabeza et al. 2004; Keith et al. 2008). Despite logistically challenging, the adoption of a species- centered approach when estimating habitat availabil- ity will certainly improve our ability to predict and mitigate the impacts of landscape changes on biodiversity. Acknowledgments We thank Greet De Coster, Hawthorne L. Beyer and Mariana M. Vale for comments on the manuscript. Financial support was provided by PNPD-CAPES (scholarship to M. A. Gomes.), FAPESP (scholarship to J. A. Prevedello, Project 2013/03457-1), and CAPES/FAPERJ/PAPD (scholarships to R. Crouzeilles). References Almeida-Gomes M, Lorini ML, Rocha CFD, Vieira MV (2014) Underestimation of extinction threat to stream-dwelling amphibians due to lack of consideration of narrow area of occupancy. Conserv Biol 28:616–619 Andrén H (1994) Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. Oikos 71:355–366 Awade M, Boscolo D, Metzger JP (2012) Using binary and probabilistic habitat availability indices derived from graph theory to model bird occurrence in fragmented for- ests. Landscape Ecol 27:185–198 Banks-Leite C, Ewers RM, Metzger JP (2010) Edge effects as the principal cause of area effects on birds in fragmented secondary forest. Oikos 119:918–926 Betts MG, Fahrig L, Hadley AS, Halstead KE, Bowman J, Robinson WD, Wiens J, Lindenmayer DB (2014) A spe- cies-centered approach for uncovering generalities in organism responses to habitat loss and fragmentation. Ecography 37:517–527 Brückmann SV, Krauss J, Steffan-Dewenter I (2010) Butterfly and plant specialists suffer from reduced connectivity in fragmented landscapes. J Appl Ecol 47:799–809 Cabeza M, Araújo MB, Wilson RJ, Thomas CD, Cowley MJR, Moilanen A (2004) Combining probabilities of occurrence with spatial reserve design. J Appl Ecol 41:252–262 Connor EF, Courtney AC, Yoder JM (2000) Individuals–area relationships: the relationship between animal population density and area. Ecology 81:734–748 Crouzeilles R, Prevedello JA, Figueiredo MSL, Lorini ML, Grelle CEV (2014) The effects of the number, size and isolation of patches along a gradient of native vegetation cover: how can we increment habitat availability? Land- scape Ecol 29:479–489 Crouzeilles R, Beyer HL, Mills M, Grelle CEV, Possingham H (2015) Incorporating habitat availability into systematic planning for restoration: a species-specific approach for Atlantic Forest mammals. Divers Distrib 21:1027–1037 Devictor V, Julliard R, Jiguet F (2008) Distribution of specialist and generalist species along spatial gradients of habitat disturbance and fragmentation. Oikos 117:507–514 Ewers RM, Didham RK (2006) Confounding factors in the detection of species responses to habitat fragmentation. Biol Rev 81:117–142 Ewers RM, Didham RK (2008) Pervasive impact of large-scale edge effects on a beetle community. PNAS 105:5426–5429 Fahrig L (2003) Effects of habitat fragmentation on biodiversity. Annu Rev Ecol Evol Syst 34:487–515 Fahrig L (2007) Non-optimal animal movement in human-al- tered landscapes. Funct Ecol 21:1003–1015 Fischer J, Lindenmayer DB (2007) Landscape modification and habitat fragmentation: a synthesis. Glob Ecol Biogeogr 16:265–280 Gardner TA, Barlow J, Peres CA (2007) Paradox, presumption and pitfalls in conservation biology: the importance of habitat change for amphibians and reptiles. Biol Conserv 138:166–179 Heard GW, Scroggie MP, Malone BS (2012) Classical metapopulation theory as a useful paradigm for the 718 Landscape Ecol (2016) 31:711–719 123 conservation of an endangered amphibian. Biol Conserv 148:156–166 Henle K, Davies KF, Kleyer M, Margules C, Settele J (2004) Predictors of species sensitivity to fragmentation. Biodi- vers Conserv 13:207–251 Keith DA, Akçakaya HR, ThuillerW,Midgley GF, Pearson RG, Phillips SJ, Regan HM, Araújo MB, Rebelo TJ (2008) Predicting extinction risks under climate change: coupling stochastic population models with dynamic bioclimatic habitat models. Biol Lett 4:560–563 Krämer B, Poniatowski D, Fartmann T (2012) Effects of land- scape and habitat quality on butterfly communities in pre- alpine calcareous grasslands. Biol Conserv 152:253–261 Mari L, Casagrandi R, Bertuzzo E, Rinaldo A, Gatto M (2014) Metapopulation persistence and species spread in river networks. Ecol Lett 17:426–434 Metzger JP, Lewinsohn TM, Joly CA, Verdade LM, Martinelli LA, Rodrigues RR (2010) Brazilian Law: full speed in reverse? Science 329:276–277 Monks A, Burrows L (2014) Are threatened plant species spe- cialists, or just more vulnerable to disturbance? J Appl Ecol 51:1228–1235 Öckinger E, Schweiger O, Crist TO, Debinski DM, Krauss J, Kuussaari M, Petersen JD, Pöyry J, Settele J, Summerville KS, Bommarco R (2010) Life-history traits predict species responses to habitat area and isolation: a cross-continental synthesis. Ecol Lett 13:969–979 Peyras M, Vespa NI, Bellocq MI, Zurita GA (2013) Quantifying edge effects: the role of habitat contrast and species spe- cialization. J Insect Conserv 17:807–820 R Development Core Team (2011) R: a language and environ- ment for statistical computing. R Foundation for Statistical Computing. Available from http://www.R-project.org. Accessed Jan 2012) Rybicki J, Hanski I (2013) Species–area relationships and extinctions caused by habitat loss and fragmentation. Ecol Lett 16:27–38 Saccheri I, Kuussaari M, Kankare M, Vikman P, Fortelius W, Hanski I (1998) Inbreeding and extinction in a butterfly metapopulation. Nature 392:491–494 Saura S, Martı́nez-Millán J (2000) Landscape patterns simula- tion with a modified random clusters method. Landsc Ecol 15:661–678 Saura S, Pascual-Hortal L (2007) A new habitat availability index to integrate connectivity in landscape conservation planning: comparison with existing indices and application to a case study. Landsc Urban Plan 83:91–103 Saura S, Rubio L (2010) A common currency for the different ways in which patches and links can contribute to habitat availability and connectivity in the landscape. Ecography 33:523–537 Semlitsch RD, Bodie R (2003) Biological criteria for buffer zones around wetlands and riparian habitats for amphibians and reptiles. Conserv Biol 17:1219–1228 SOS Mata Atlântica, INPE (2012) Atlas dos remanescentes florestais da Mata Atlântica. Fundacão SOS Mata Atlân- tica, São Paulo Suhonen J, Korkeamäki E, Salmela J, Kuitunen M (2014) Risk of local extinction of Odonata freshwater habitat general- ists and specialists. Conserv Biol 28:783–789 Villard MA (1998) On forest-interior species, edge avoidance, area sensitivity, and dogmas in avian conservation. Auk 115:801–805 Villaseñor NR, Driscoll DA, Escobar MAH, Gibbons P, Lin- denmayer DB (2014) Urbanization impacts on mammals across urban-forest edges and a predictive model of edge effects. PLoS ONE 9:e97036 Landscape Ecol (2016) 31:711–719 719 123 http://www.R-project.org The use of native vegetation as a proxy for habitat may overestimate habitat availability in fragmented landscapes Abstract Context Objectives Methods Results Conclusions Introduction Methods Overview Studied landscapes Specialist species Quantifying habitat availability and amount Results Habitat availability Habitat amount Discussion Acknowledgments References