Expansion of the agricultural frontier in the largest South American Dry Forest: Identifying priority conservation areas for snakes before everything is lost


Authors: María Soledad Andrade-Díaz aff001;  Juan Andrés Sarquis aff002;  Bette A. Loiselle aff003;  Alejandro R. Giraudo aff002;  Juan Manuel Díaz-Gómez aff001
Authors place of work: Instituto de Bio y Geociencias del Noroeste Argentino (Consejo Nacional de Investigaciones Científicas y Técnicas—Universidad Nacional de Salta), Rosario de Lerma, Salta, Argentina aff001;  Instituto Nacional de Limnología (Consejo Nacional de Investigaciones Científicas y Técnicas–Universidad Nacional del Litoral), Ciudad Universitaria, Santa Fe, Argentina aff002;  Department of Wildlife Ecology and Conservation, Center for Latin American Studies, University of Florida, Gainesville, FL, United States of America aff003;  Facultad de Humanidades y Ciencias (Universidad Nacional del Litoral), Ciudad Universitaria, Santa Fe, Argentina aff004
Published in the journal: PLoS ONE 14(9)
Category: Research Article
doi: 10.1371/journal.pone.0221901

Summary

Conservation planning relies on integrating existing knowledge, social-environmental contexts, and potential threats to identify gaps and opportunities for action. Here we present a case study on how priority areas for conservation can be determined using existing information on biodiversity occurrence and threats. Specifically, our goals are: (1) to model the ecological niche of twelve endemic snake species in the Dry Chaco Forest, (2) to quantify the impact of the deforestation rates on their distributions, (3) to propose high priority areas for conservation in order to improve the actual protected area system, and (4) to evaluate the influence of the human footprint on the optimization of selected priority areas. Our results demonstrate that Argentinian Dry Chaco represent, on average, ~74% of the distribution of endemic snake species and deforestation has reduced suitable areas of all snake species in the region. Further, the current protected areas are likely insufficient to conserve these species as only very low percentages (3.27%) of snakes’ ranges occur within existing protected areas. Our models identified high priority areas in the north of the Chaco forest where continuous, well-conserved forest still exists. These high priority areas include transition zones within the foothill forest and areas that could connect patches of forest between the western and eastern Chaco forest. Our findings identify spatial priorities that minimize conflicts with human activities, a key issue for this biodiversity hotspot area. We argue that consultation with stakeholders and decision-makers are urgently needed in order to take concrete actions to protect the habitat, or we risk losing the best conservation opportunities to protect endemic snakes that inhabit the Argentinian Dry Chaco.

Keywords:

Biology and life sciences – Organisms – Eukaryota – Animals – Vertebrates – Amniotes – Reptiles – Squamates – Snakes – Ecology – Ecosystems – Forests – Biodiversity – Forest ecology – Ecology and environmental sciences – Conservation science – Ecology – Ecosystems – Forests – Biodiversity – Forest ecology – Terrestrial environments – Forests – Deforestation – Habitats

Introduction

The loss and fragmentation of native forests due to the expansion of the agricultural frontier have affected the abundance, diversity and distribution of species in tropical and subtropical ecosystems [13]. At a global scale, the expansion and intensification of agriculture [4,5] were achieved by prioritizing economic and political interests. In contrast, ecological attributes, such as distinctive habitats, species richness and population abundances, connections between the different patches of remaining forest, among others, often were not considered when decisions were made regarding transformation of land. The failure to consider the maintenance of structure and functionality of forests' ecosystems in environmental policies and decision-making regarding development may result in the loss of socio-economic benefits that these natural ecosystems provide to human populations [6,7]. Moreover, the existing protected areas in Latin America are insufficient and generally, they were not designed focusing on biodiversity conservation goals [8]. Currently, in many places the expansion of the agricultural frontier has enclosed protected areas, thereby transforming them into ‘islands’ where plant and animal populations become isolated [9]. It is in this dynamic scenario of changing land use and inadequate protected area systems, where studies are needed to predict species distributions in order to identify priority areas for conservation and propose effective strategies for their long-term protection.

The tropical dry biomes, currently considered among the most worldwide endangered ecosystems, exemplifies a system where land use changes are rampant and existing protected areas are insufficient to achieve conservation objectives. Further, these biomes have received relatively little attention from both ecologists and conservationists [1012]. In fact, the few studies focused on the identification of conservation areas for these ecosystems showed that current representativeness levels (i.e., represent the range of expected biological variation) are inadequate, including the tropical seasonally dry forests [13], the Cerrado [14,15] and the Gran Chaco [16]. The ‘Gran Chaco’ is the second largest forest in South America, after the Amazon rainforest [17]. This ecoregion corresponds to Chaco Biogeographic Province encompassing both the Dry and Humid Chaco sub eco-regions [18,19] and it extends through Argentina, Bolivia, Paraguay and southwestern Brazil [20]. From an ecosystem service perspective, this forest is extremely important as it harbors one of the largest extra-tropical carbon stocks in the southern hemisphere [2123]. Despite being characterized by heterogeneous environments and high species diversity [20], this forest is currently considered as one of the world’s most threatened wooded subtropical ecosystems as a consequence of an intense anthropogenic disturbance as a consequence of logging and agricultural activities [12,22,24,25]. In fact, the main driver of high deforestation rates (200,000 ha/year -1) in the Chaco forest is agribusiness expansion [12,26]. This transformation produces high carbon emissions [27], landscape configuration changes [28], habitat degradation [29], and species extinctions [1]. Likewise, despite the fact that Argentina has the largest area of Chaco (approx. 62% of the total area) [26], we observed that only 1.7% of the Dry Chaco is within the current Protected Areas system (PAs) [30,31]. Moreover, increasing evidence indicates that distribution and survival of inhabiting species on dry ecosystems could also be also affected by climate change which could lead to widespread reduction of current species richness and ecological integrity (e.g. [3234]). Therefore, in order to identify priority areas and to develop management decisions and conservation strategies, information on species distributions and their vulnerability to land use change is needed [7,3537].

Establishing a priority areas network that adequately represents the biodiversity within the Dry Chaco with clear conservation goals, while at the same time compatible with the sustained human development is urgently needed [16]. In this sense, different conservation planning schemes have been developed over the last decade [3840] promoting a representative and connected network of PAs that contributes to the viability of biodiversity and ecosystems [41]. These approaches are based on the distribution of key biodiversity features and anthropic variables, and identifying the most important sites for conservation and sites, which are compatible with sustainable use [42,43]. From this perspective, the integration of species-level surrogates is necessary to ensure that critical habitats and ecosystems within the region are not missed [13,16,44,45]. Snakes have received little attention in land use changes studies [46,47], despite the fact that habitat loss and degradation are one of the main factors driving declines in reptile populations [4851]. Some studies have shown that habitat loss and the spatial patterns of remaining vegetation affect the physical condition, as well as the distribution of snakes in the landscape [52,53]. Moreover, snakes are killed by humans, despite their important role as prey and predators in ecosystems [54]. Indeed, unpopular animals like snakes have attracted less attention than charismatic vertebrates [5557], and as a consequence, conservation action plans rarely consider snakes. While PAs are important to preserve snake communities unaffected by human modifications, few reserves have been established that represent snakes adequately [53,58,59]. The Dry Chaco is not an exception to this scenario. In fact, compared to other taxonomic groups, we know little about the natural history, ecology and distribution of most snake species associated with this region [53]. Thus, analyzing the suitable habitats of endemic snakes in the Dry Chaco forest is a crucial step to evaluate and identify conservation priority areas that require immediate protection actions [47].

Nevertheless, delimiting the distribution of a species is a complex task that involves many determining both factors, both biotic and a biotic, which are difficult to assess through fieldwork [60,61]. Generally, species distribution ranges are frequently represented by polygons [62] and are often used for conservation decisions [16]. However, these approaches suffer from the effects of multi-level conflicts among scales and resolutions and are likely to frequently include many areas not holding populations or exclude others where populations actually occur [60,63]. Thus, data-driven techniques (e.g., Ecological Niche Models [ENMs]) have been developed to predict the potential distribution of species, via identifying suitable areas for species occurrence, as well as pointing to the most relevant bioclimatic variables that predict occurrence [61,64,65].These approaches offer widely accepted methods for summarizing species' distributional patterns for conservation applications [63,66]. Therefore, the use of ENMs and systematic conservation planning methods are useful to achieve conservation goals as they provide critical information on spatial areas required for species conservation [67] and guide conservation decision-making processes regarding actions and policies needed for the long-term protection of biodiversity.

The main objective of this study is to analyze the relationship between the expansion of the agricultural frontier and the distribution of endemic snake species in the Dry Chaco forest. To accomplish this, we use geographical information of 12 snake species endemic to Dry Chaco forest as focal group, and employ ENMs and conservation planning tools. We address the following questions: (1) what are the predicted geographic distributions of the endemic snakes in Argentinean dry Chaco?; (2) which bioclimatic variables that best determine these snake species distributions?; (3) how has agricultural expansion impacted distribution of snake species?; (4) what are the priority areas for conservation of these endemic snakes and what proportion of these areas are represented in the current protected areas system? [6871]; and, (5) how has human activity (human footprint) influenced the spatial prioritization of an optimal/ representative priority area network?

Materials and methods

Occurrence data

We selected 12 snake species, which are endemic to the Dry Chaco forest as focal taxa (Table 1). We obtained occurrence records from reviewing museum collections, reliable literature records and from high-intensity sampling in Argentina for over 28 years (S1 Table). These data were complemented with information obtained from the Global Biodiversity Facility network (GBIF) [72]. Using multiple data sources provided complementarity and served to minimize biases in our data [59,73].Because the information of occurrence records for species has many shortcomings, we filtered the information to avoid inadequate taxonomy and identify problematic or imprecise locality records [74]. We compared the spatial distribution of records obtained with the ranges for species, checking this information with help of specialists (Giraudo pers. obs.), and removed all those mismatched records. For specimens without geographic coordinates, we used Google Earth to define a georeferenced location using information provided with the record. We omitted, however, those records where geographical information could not be verified. We compared these museum and on-line data with expert maps and our knowledge [53] and determined that these data resulted in well-represented geographic coverage of the focal taxa. The combination between expert maps that are in fact an excellent resource for delimiting the broad areas outside which a species is not expected to occur [75] and the well-understood distributions by specialists could be considered the best approach to define empirical geographical distributions [76,77]. Then, using 'ecospat' package in R software [7880], we removed records repeated in multiple sources and retained only unique localities within a vicinity of 10 km2.

Tab. 1.

Summary of the bioclimatic variables with their percentage contributions (%) in models predicting snake distribution.

<h2>Summary of the bioclimatic variables with their percentage contributions (%) in models predicting snake distribution.</h2>

Climatic data

The dry Chaco forest [18] is seasonal with a dry winter and a rainy summer [20]. We obtained data layers from the WorldClim v1.4 at 30” spatial resolution, which includes a set of 19 climatic variables summarizing aspects of precipitation and temperature for the earth's surface for the period 1950–2000 [81]. After a review of the natural history of the species, a jackknife test of variable importance provided by Maxent (i.e. variables with highest gain when used in isolation and variables that decreased the gain the most when they were omitted) and a Pearson correlation analysis between all the variables, we discarded highly correlated variables (r > 0.8) and selected the following seven variables: BIO2 = Mean Diurnal Range; BIO4 = Temperature Seasonality; BIO5 = Maximum Temperature of Warmest Month; BIO6 = Minimum Temperature of Coldest Month; BIO13 = Precipitation of Wettest Month; BIO14 = Precipitation of Driest Month; and BIO15 = Precipitation Seasonality. Layers were prepared and processed using QGIS software v. 2.16.2 [82].

Ecological niche models

Because dispersal plays a crucial role in the distribution of organisms and should be considered in species modeling [83], we defined an area for model calibration (or M in BAM diagram; [61]) that reflects the accessible historical area for each species. This calibration area was created by a background polygon that corresponded to the extension area for dry Chaco ecoregion. Such consideration assumed that this region represent the species’ tolerance limits, historical barriers to dispersal, and intrinsic need for certain abiotic conditions [84].

We modeled habitat suitability––based on ENM principle––for each species using a maximum entropy method, MaxEnt v3.3.3K [85]. MaxEnt software calculates the potential geographic distribution for each species linking spatial records and bioclimatic variables [85,86]. Although recent studies have shown that there are uncertainties when forecasting species distributions depending on the algorithm used [87,88], we decided to use MaxEnt over other available methods given its proven high performance and suitability for presence-only data [64,89,90]. We used 'ENMeval' package [91] to perform a calibration protocol assessing the model complexity. For this step, models for species with small data sets (<10 records), we applied the n-1 jackknife method proposed by Pearson et al. [89], where each occurrence was used for testing once, while the other records were used to train the model [89,91,92]. For species with more than 10 records, we used an equivalent of the 'cross-validate' method in MaxEnt software, where the 'random k-fold' method partitions occurrence localities randomly into a specific number of bins (in this case we used k = 5 bins). We ran models (using randomly sub-sampled 50% of the data as testing) under varying model response types (feature classes: L, Q, LQ), different values of regularization multiplier values (RM: 1, 2.2, 4.6) [93] and performing 500 iterations with 500 replicates for bootstrap analysis [94]. We selected only linear and quadratic features classes because in general we had species with few records [95]. With the ENM evaluate function we obtained the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) plot [96] based on the test data (AUCTEST). To quantify the degree to which models overfit the data, we calculated three metrics: (1) the difference between training and testing AUC (AUCDIFF), (2) the Minimum Training Presence 10 omission rate (ORMTP) and, (3) 10% training omission rate (OR10). The final best model was selected using the Akaike Information Criterion (AICc) [91]. Likewise, we used the platform NICHE TOOLBOX to obtain the values of partial-ROC [94,97] for each final model obtained. Unlike AUC, Partial-ROC allows a differential weighting of omission and commission errors and focuses on meaningful predictions for model evaluation. Finally, the obtained continuous models for the suitability conditions of species were converted into binary maps (presence/absence) considering a threshold value that maximized the True Skill Statistics (TSS). For that, we elected predictions corresponding to 5% of omission error.

Spatial analyses

To evaluate how land use change affected the potential distribution of snake species in the Argentinian Dry Chaco, we performed an analysis considering two approaches: (a) total reduction area for the species and (b) estimating the reduction area per year from 1976 to 2017. We used a map of the expansion of the agricultural frontier from 1976 to 2017 for Argentinian Dry Chaco Forest (available at http://geoportal.idesa.gob.ar/). For this step, we considered only the natural intact forest areas, whereas disturbed areas (i.e. urban areas, deforested areas, farming areas, and pastureland for cattle ranching) were not included [98,99]. Then, we assessed the importance of PAs for the obtained models calculating (in km2) the proportion of potential distribution areas within the current PAs system. Shape file of PAs were downloaded from the National Geographic Institute (IGN-ARGENTINA, available in http://www.ign.gob.ar/). Likewise, to assess if the current PAs system contained the most suitable bioclimatic conditions for the species, we performed a Kolmogorov-Smirnov (KS) test in R [100] verifying if values differed significantly between within and without protected areas [101].

Spatial conservation prioritization analyses

We identified conservation priority areas using ZONATION 4.0 Software [40]. ZONATION algorithms generate a hierarchical priority ranking of the landscape based on the biological values of the spatial units (cells). Overall, ZONATION uses a raster for each biodiversity feature (herein snake species and anthropic variables), where each pixel contains information of either the occurrence or intensity of each feature [102]. The way the value of ‘loss of conservation’ is aggregated across features (occurring in a pixel) depends on the so-called ‘cell-removal’ rules. In this sense, the software produces a complementarity-based and balanced ranking of conservation priority over the entire landscape maximizing the species' occurrence and considering the different ‘penalization’ variables used [40,102]. For a more detailed explanation about the use of ZONATION see Di Minin et al. [102]. In this work, we compared scenarios with two different options of marginal aggregate loss: (1) the Additive Benefit Function (ABF), which emphasizes species richness minimizing extinction risk; and (2) the Core Area Zonation (CAZ), which emphasizes areas with both the highest suitability scores and the lowest uncertainty values for each species [102104].

During the priority analysis, Zonation considers species priorities (weights), land cost, habitat quality, measures of connectivity, etc. [69,105,106]. We used as biodiversity features the probability of occurrence maps for each snake species and we weighted species assigning a high value (5) for threatened species and a low value (1) for those categorized as low conservation concern according to the IUCN (2018) and Giraudo et al. [53]. To promote the selection of optimal areas for current PAs expansion, we used as a hierarchical mask [102] a layer of national and provincial PAs (IUCN and UNEP-WCMC 2012). In this sense, the program identifies the best part of the landscape for an optimal and balanced expansion of existing PAs (which are preferably selected as the first option in the analysis), and also to compensate for specific ecological losses and satisfy the targets with minimum cost [102]. Given that most snake species cannot adequately be protected inside highly modified areas, we assigned negative weights or ‘penalization’ value to pixels covered by crops or urbanized areas. This last step prevented the software from selecting highly modified areas and assigning high conservation values to such areas. In this sense, considering that human influence tends to diminish habitat quality, and therefore, the potentiality for conservation, we used the Argentina Human Footprint layer [107] as a negative variable, ‘penalizing’ those pixels with high human influence. This Human Footprint layer (HII) was created considering data layers of human population pressure, human land use and infrastructure, and human access [108]. We assigned negative weights to these features (i.e. pixels in highly modified areas) so that the sum of the positive and negative weights was zero, allowing a balanced solution for prioritization [104,109]. In order to evaluate the influence of expansion of the agricultural frontier in the largest area of the Dry Forest, determining the relative importance of current PAs and identify complementary priority conservation areas for long-term protection of endemic snakes, we performed four alternative analyses in ZONATION. For the first priority areas analysis, we used only all snake species distributions, equally weighted, to obtain information about the most important areas to conserve in an ideal scenario without deforestation and fragmentation habitat. Then, in a second approach, we did an analysis considering all species distributions, also equally weighted, but with the PAs as hierarchical mask, which allowed us to identify where complementary areas to PAs occur, assuming fragmentation and habitat loss were not yet a current problem for the species. Thirdly, to identify the best opportunities for an optimal and balanced expansion of existing PAs which compensates specific ecological losses and satisfies the targets with minimum cost, we developed a priority areas analysis considering all species distributions, individual weighted according to the IUCN status, with the PAs as hierarchical mask and using the Human Footprint layer as a negative variable. Finally, to determine the relative importance of current PAs within the species distribution considering the expansion of the agricultural frontiers, we performed an analysis based on species distribution, individual weighted, using the Human Footprint as negative variable but without the PAs as hierarchical mask. After running these prioritization analyses, we plotted performance curves for all approaches considering the general patterns. These performance curves quantify the proportion of the original occurrences retained for each biodiversity feature at each top fraction of the landscape selected for conservation [40,102]. Finally, we determined the representativeness of (1) the current PAs network and (2) the top priority 17% of the landscape under protection (this percentage represents the AICHI targets; [110]).

Results

We obtained 706 historical occurrence records for the 12 focal endemic snake species from which 360 independently records were used in ENMs (Table 1; S1 Table). The number of occurrence records used for each species and the percentage of contribution for bioclimatic variables in the models are summarized in Table 1. Overall, we observed that the maximum temperature of the warmest month (BIO 05) was the most important variable for almost all species (except E. albertguentheri and P. baroni). In these latter two species, precipitation seasonality (BIO 15) and diurnal temperate range (BIO 02) emerged as the most relevant variables respectively.

In general, models for each snake species performed well against the validation data (Fig 1 and Table 2). The Partial ROC bootstrap tests indicated significant ratio values (mean AUC ratios ≥1.4), low standard deviations and significant p-values (p <0.001) in all the models. Likewise, the Jackknife test (considering the ROC plot [96] based on the test data [AUCTEST]) showed that models also tended to be statistically significant (p < 0.01) for those species with <10 occurrence records. These results demonstrate that all models performed better than random (Table 2), and, thus, performed well in estimating potential distributions of species.

<h2>ENMs for snake species in the Argentinian Dry Chaco considering deforestation since 1976 to 2017.</h2>
Fig. 1.

ENMs for snake species in the Argentinian Dry Chaco considering deforestation since 1976 to 2017.


Warmer colors (red, orange and yellow) indicate higher habitat suitability while colors from green to white indicate the lowest predicted values. Black color indicates deforested areas. (a) B. constrictor occidentalis; (b) E. alvarezi; (c) E. albertguentheri; (d) E. sagittifer modestus; (e) L. annulata pulchriceps; (f) P. baroni; (g) P. erlandi; (h) P. vittatus; (i) P. genimaculatus; (j) S. lavillai; (k) T. chaquensis; (l) X. pulcher.

Tab. 2.

Mean AUCTRAIN, AUCTEST, AUCDIFF and Partial ROC (AUCRATIO) for snake species models with the minimum AICc value.

<h2>Mean AUC<sub>TRAIN</sub>, AUC<sub>TEST</sub>, AUC<sub>DIFF</sub> and Partial ROC (AUC<sub>RATIO</sub>) for snake species models with the minimum AICc value.</h2>

Argentinian Dry Chaco has experienced a 15% reduction in area between 1976 (786,790 km2) and 2017(669,410 km2) (Fig 2). We observed that models predicted, on average, a distributional range extent of ~447,000 km2 for focal species (min 204,828 km2 [E. alvarezi]–max 748,630 km2 [P. vittatus]), with at least five species (B. constrictor occidentalis, E. sagittifer modestus, P. vittatus and X. pulcher) with large areas of distribution (>500,000 km2). The two most spatially restricted species were E. alvarezi and L. annulata pulchriceps (Table 3). Likewise, we observed that Argentinian Dry Chaco represent, on average, ~74% of species distribution species with only P. genimaculatus and T. chaquensis with less than 60% of distribution within Argentinian Chaco. However, when we considered the effects of fragmentation and habitat loss, we observed an average ~73% reduction (min 64.56%[E. alvarezi]–max 81.72% [E. albertguentheri]) in suitable areas for snakes species in the Dry Chaco forest. The three species with most habitat loss estimated were E. albertguentheri, T. chaquensis, and E. sagittifer modestus; while that snakes less affected were P. genimaculatus, P. baroni, E. alvarezi. The general trends, from 1975 to 2019, for the reduction in areal extent for the Argentinian Dry Chaco forest and species' potential distribution are summarized in the Fig 2.

<h2>Percentage of the remaining distributional area for snake species and for the Argentinian Dry Chaco forest from 1976 to 2017.</h2>
Fig. 2.

Percentage of the remaining distributional area for snake species and for the Argentinian Dry Chaco forest from 1976 to 2017.


Tab. 3.

Potential distribution area (km2) of snake species in the Dry Chaco forest.

<h2>Potential distribution area (km<sup>2</sup>) of snake species in the Dry Chaco forest.</h2>

Using IUCN criteria, all of our focal snake species are not considered to be threatened or have not been evaluated (Table 4) [53]. Yet, we found that only a small portion of snakes’ ranges occur within existing PAs from Argentina (Table 4, Fig 3). On average, we observed that 3.27% (min: 0.21 [E. alvarezi]–max: 5.9 [P. genimaculatus] %) of predicted distributions by species are within some category of protection. Also, for all focal species, the areas inside PAs had significantly lower suitability values compared with the areas outside PAs (Table 4).

<h2>Potential distribution areas (in km<sup>2</sup>) of snake species in Dry Chaco forest.</h2>
Fig. 3.

Potential distribution areas (in km2) of snake species in Dry Chaco forest.


Blue bars: distributional range predicted for each snake species in the Dry Chaco forest; Red bars: distributional range predicted for species in Argentinian Dry Chaco forest considering habitat loss/deforestation; Black bars: distributional range predicted within protected areas.

Tab. 4.

Comparison between the suitability values (Mean± SD) outside and inside PAs.

<h2>Comparison between the suitability values (Mean± SD) outside and inside PAs.</h2>

Prioritization scenarios for snake species based on ZONATION analyses are shown in Fig 4. As a general pattern, the priority areas change when the cost of the HII is considered. The current PAs cover 1.69% of the Argentinian Chaco forest. In all scenarios, this percentage represented on average ~3% (max. Control scenario: 2.99% in both ABF and CAZ; min. Scenario 2: 1.58%ABF and 1.56%CAZ) of the predicted distributional ranges of snakes. When we modeled the AICHI landscape goals of 17% of area protected, all snake species benefited markedly from increase in protected areas. The value, not surprisingly, was highest (~28%) in the scenario without deforestation and lower under more realistic scenarios that incorporated land use change and species priorities [Scenario 2: 20.38% ABF and 20.19% CAZ] (Fig 5A and 5B). Further, the representativeness of distribution of snakes’ species was greatest in HII scenarios [Scenario 3: 23.05% ABF and 20.32% CAZ] (Fig 5A and 5B). Moreover, considering a 17% protection goal, the priority areas identified in Scenario 2 showed spatial patterns consistent for both algorithms and slightly different to that from Scenario 3, which included more priority areas in the center (ABF) and south (CAZ) of the Dry Chaco forest (Fig 6). Finally, consensus areas among scenarios can be found adjacent to current PAs while other PAs are not found to be adjacent with priority selected areas regardless of model scenario (Fig 6).

<h2>Priority areas for focal snakes distributed in the Argentinian Dry Chaco under different prioritization scenarios.</h2>
Fig. 4.

Priority areas for focal snakes distributed in the Argentinian Dry Chaco under different prioritization scenarios.


In all scenarios we compared the outputs that emphasized species richness minimizing extinction risk [Added Benefit Function, ABF (left)] and emphasized areas with both the highest suitability scores and the lowest uncertainty values for each species [Core Area Zonation CAZ (right)]. (a-b) Control scenario (species equally weighted); (c-d) Scenario 1: species equally weighted + PAs; (e–f) Scenario 2: species individual weighted + PAs + cost (Human Footprint—HII); (g-h) Scenario 3: species individual weighted + cost (HII). The color scheme shows the nested ranking on a map. Ranking the biological value of the site: 2% high priority (light red); 2–5% (dark red); 5–10% (orange); 10–25% (yellow); 25–50% (light blue); and 50–80% (blue); 80–100% (or the least valuable 20%) (dark blue).

Fig. 5.

Performance curves of the spatial prioritization solutions for both removal rules ABF [Added Benefit Function] (a) and CAZ [Core Area Zonation] (b). Each line represents a scenario [Control scenario: species equally weighted; Scenario 1: species equally weighted + PAs; Scenario 2: species individual weighted + PAs + cost (Human Footprint-HII) and; Scenario 3: species individual weighted + cost (HII)] showing the proportion of landscape protected and its corresponding average species range protected. The vertical lines represent current proportion of land in PAs (1.69%) and the recommended AICHI target (17%) [110].

<h2>Maps showing the 17% of priority conservation for ABF (Added Benefit Function) and CAZ (Core Area Zonation) in Scenario 2 [Species individual weighted + PAs + cost (Human Footprint-HII)] and; Scenario 3 [Species individual weighted + cost (HII)].</h2>
Fig. 6.

Maps showing the 17% of priority conservation for ABF (Added Benefit Function) and CAZ (Core Area Zonation) in Scenario 2 [Species individual weighted + PAs + cost (Human Footprint-HII)] and; Scenario 3 [Species individual weighted + cost (HII)].


Red pixels correspond to the consensus areas selected by both CAZ and ABF prioritization schemes for conservation snakes and which also complement current PAs system of the Argentinian Dry Chaco forest. Those selected by only one prioritization scheme are shown in green (ABF) or orange (CAZ).

Discussion

Here we present for the first time how expansion of the agricultural frontier has potentially impacted the distribution of endemic snake species in the Dry Chaco forest. Further, we evaluate the degree with which those species are represented in the current protected area system in the Argentinian Chaco and propose locations of high priority areas for snake conservation in the region. In the paragraphs below we discuss some of the key results including the reduction of potentially suitable habitat for snakes, but also the potentially limited value of the current protected area system to conserve quality habitats for endemic snakes of Dry Chaco forest.

Impacts of deforestation and expansion of the agricultural frontiers

The Dry Chaco is severely threatened due to the significant transformation of forest for agriculture and timber production. In the last decades, the growing demand for agricultural products and the new “extreme weather resistant” variety of crops have led to an exponential increase in areas under crop production [22,111]. In Argentina, which harbors a large portion of Dry Chaco forest, deforestation alone has resulted in a 15% reduction (117,380 km2) in the past four decades (Fig 2). Less easily measured is the percent of Dry Chaco areas, which have not been transformed, but may have experienced degradation due to escaped fires, logging, introduction of invasive species, and other factors. These habitat loss and degradation processes, together with the limited percent of Chaco in protected status (1.69%), highlight the present need to direct conservation planning efforts that consider and evaluate the trade-offs between conservation and economic development.

Evaluating the impact of deforestation and expansion of agricultural frontier on snakes is hampered by the limited information on the natural history and distributions of many snake species in the Dry Chaco forest. Indeed, for the majority of our focal species, IUCN does not have sufficient information to categorize the conservation status of the species (Table 4). Particularly under such conditions, the use of ENMs to predict potential distributions has become a valuable tool in conservation planning, and has been used for other taxa in Argentinean Dry Chaco context [16,112]. Using results from ENMs, we demonstrated that the increase in deforestation rates in the Argentinian Dry Chaco over the last four decades have resulted in a disproportionate decrease in the distribution of nearly all snake species analyzed (Fig 1 and Fig 2). This decreasing distribution has also been demonstrated for Leptodactylus species (Anura) with at least 25% of their geographic range occurring in the Dry Chaco. In this work, the potential distribution area of these species decreased between 16% and 10% approaching with the deforestation rate from 1976 to 2013 [37]. Also, this loss was even more relevant for the distribution of species, which had already been transformed to agriculture in the last 40 years experienced. However, not all species respond in a similar way to forest loss and degradation as differences in natural history among the species may result in greater or lower risk. For example, generalist or fossorial snakes, such as E. sagittifer modestus, P. vittatus and X. pulcher, which tolerate a wide range of bioclimatic conditions, may be more able to adapt to transformed and degraded environments [46,113]. Nonetheless, these areas which likely include periurban or urban zones [47], can represent other risks as a result of negative interactions with humans which tend to fear and indiscriminately kill snakes [55,114]. In contrast, specialist snake species with restricted bioclimatic conditions, such as B. constrictor occidentalis and E. alvarezi, are expected to have increased vulnerability with changes in habitat conditions [46,113]. For example, studies have shown that loss of habitat affects body condition, clutch size and testicular volume of B. constrictor occidentalis [52,53]. Consequently, the natural history of the species is important to consider in management and conservation planning, but was beyond the scope of the current study.

Role of argentinian protected areas for snake conservation

The current protected area system in Argentinian Dry Chaco is likely inadequate for long-term protection of most of the endemic snake species studied here for several reasons. First, only 1.69% of the Argentinian Dry Chaco forest is currently under protection. Despite this, the current PAs system represents, on average, 3.27% of snakes predicted distributions. Yet, this is still a very small amount of the range of the potential distribution in the Dry Chaco. In some cases, such as the threatened (AM) E. alvarezi, only 0.21% of their predicted range is includes in the PAs (Fig 3). Secondly, our results indicated that the areas inside the PAs generally contained lower suitability values compared with areas outside them (Table 4). Therefore, the current PAs not only provide very limited spatial protection for these endemic snakes, but also protect habitats of apparently less optimal bioclimatic conditions. These results are consistent with studies that have analyzed the role of PAs in the protection of other groups of vertebrates in the Chaco forest. For example, a study for 63 vertebrate species (21 amphibians, 20 mammals and 22 birds) with at least 70% of their distributions within the Gran Chaco, found that current PAs represent on average 9% of the total distribution of endemic species [47]; also, the PAs protects around 5% of the potential distribution of Leptodactylus species (Anura) [37]. Moreover, comparing the control map with the three alternatives scenarios (Fig 4), our spatial conservation prioritization revealed that the current PAs system of the Argentinian Dry Chaco did not capture the most suitable areas for endemic species of snakes, and in fact, were notable in their failure to align with more suitable locations.

The PAs in Argentina are apparently inadequate to protect these endemic snakes and other vertebrates, this is not surprising given that information on biodiversity and habitat were not used in the creation of the PAs. Further reducing the conservation value of these PAs is the increasing fragmentation of land adjacent to these lands, as well as poor legal-institutional security and weak or inexistent management programs [8]. Lack of planning and design in selection of PAs has resulted in inefficiency not only in the investment of human and financial resources but also in terms of conservation value of the PAs [8,30].

The prioritization analyses conducted in this study were consistent in identifying the northwestern and central parts of the Dry Chaco as high priority areas and the southern Chaco, where the forest is much more degraded due to anthropogenic actions [25,115], as the areas with the lowest priority ranking (Fig 4). These scenarios differed in a number of ways including whether or not they included information on current land use (i.e., incorporated a Human Footprint Index), included PAs in prioritization selection, considered species equally or ranked them according to threat status. Further, how areas were selected for each scenario was done in two ways–minimizing species extinction risk (ABF) and incorporating information about suitability of areas and uncertainty (CAZ). When we compared the performance curves of the different scenarios, the representativeness of the current PAs system covered, on average, only a small portion of the distribution of endemic snake species [Control: 2.99 (ABF) and 2.99 (CAZ); Scenario 1: 1.78% (ABF) and 1.72% (CAZ); Scenario 2: 1.58% (ABF) and 1.56% (CAZ); Scenario 3: 2.80% (ABF) and 2.50% (CAZ)] (Fig 5A and 5B).

We suggest that the most realistic scenario from which to draw conclusions and guide future conservation planning is Scenario 2, which incorporated species individual weighted + PAs + Human Footprint (HII) as a cost layer. This scenario results in the most realistic solution by selecting well-conserved Dry Chaco forest areas for snake conservation in Argentina (Fig 4E and 4F). When 17% AICHI targets for land protection were incorporated, this scenario conserved an average 20.38% (ABF) to 20.19% (CAZ) of snakes’ potential distribution; Scenario 3 presented similar result for CAZ outputs; while the other scenarios (Control and Scenario 1) accounted for above 20% of protection but ignoring current patterns of fragmentation and habitat loss (Fig 5A and 5B). For both ABF and CAZ outputs in Scenario 2, the highest areas of prioritization were located on continuous native forest remains, many of which are located in transitional areas between Dry Chaco and southern tropical Andean forest (i.e., Austral Yungas ecoregion). These transition zones with foothill forests may serve as corridors to connect patches of remaining native forest between the western Dry Chaco and Yungas forests. Currently, due to agricultural expansion, however, the connection between Yungas-Chaco has been reduced from 1,035 km to only 162 km [116118] and, as a consequence, this transition zone is highly threatened [26,119]. Overall, our priority areas coincided with a spatial prioritization analysis made for the Gran Chaco using ZONATION [16]. However, in this study, reptiles were not considered, thus our results complement and highlight the relevance of these areas.

The adequacy of the current PAs for snakes conservation is highlighted by comparing priority area selection between Scenarios 2 and 3 (Fig 6). While selection is constrained to complementing PAs in Scenario 2 (i.e., adds onto existing PAs), no such constraint exists in Scenario 3. As a consequence, there is less overlap of priority areas with current PAs in Scenario 3. This likely occurs because of generally lower suitability values for snakes found inside protected areas (Table 4). Further, releasing the constraint of parks in selecting priority areas appears to increase the sensitivity to selection rules (ABF and CAZ) although considerable consensus occurs between the two (Fig 6). This difference is likely also a result of the importance of considering suitability in site selection and the existing human impact in the current PAs system. According to that, Argentinian government should invest economic resources to working in the strengthening of action plans and management of the PAs. Therefore, these results highlight the need to not only conserve existing protected areas but also expand the current PAs to complement and update the current environmental policies.

Our results, while highlighting several areas of concern and making some specific recommendations, have a number of limitations. Like other studies that have applied species distribution models, our study also is limited by (1) lack of data for some species, (2) potential geographic bias of data and, (3) lack of true absences [86,90,96]. Although we have worked on solving these issues, we recognize the limitations presented by the low number of records for some species. In our approach, we evaluated each model with parameters (feature classes and regularization multiplied) designed to reduce model complexity and improve model performance.

Final considerations

The Chaco forest is one of the most threatened ecosystems in the world, with a high rate of deforestation and fragmentation [16]. The spatial pattern of human activity is a relevant factor in conservation planning. With increasing human activity, the possibilities to adequately conserve biodiversity are decreasing [120]. However, our spatial prioritization approach demonstrates that it is still possible to complement the current protected areas network and promote habitat connectivity. Our finding identifies spatial priorities that minimize conflicts with human activities, a key issue in a threatened ecoregion characterized by the rapid transformation of natural areas due to the advancing agricultural frontier. Although this study was carried out taking into account the agricultural expansion in Argentina, it should be noted that our prioritization results identified high priority areas located near Bolivia and Paraguay. The current situation of the Chaco forest in Bolivia and especially in Paraguay is quite similar to Argentina [16,121]. Further study should extend this analysis into Bolivia and Paraguay and include more taxa, climate change scenarios, regional habitat loss and remaining habitat connections in the entire Gran Chaco region [16,20,29,36,37,52]. In addition, collaborative work among countries is urgently needed to complement existing protected areas and to generate comprehensive plans for expanding and connecting priority areas in the whole ecosystem. We need immediate concrete actions taken by stakeholders and decision-makers, or, we will risk losing the best opportunities to conserve the biodiversity of the Dry Chaco forest.

Supporting information

S1 Table [xls]
Records of snakes species for Dry Chaco forest after filtering.


Zdroje

1. Foley JA, DeFries R, Asner GP, Barford C, Bonan G, Carpenter SR, et al. Global Consequences of Land Use. Science (80-). 2005;309: 570–574. doi: 10.1126/science.1111772 16040698

2. Sala OE, Van Vuuren D, Pereira HM, Lodge D, Alder J, Cumming GS, et al. Biodiversity across Scenarios. Millennium Ecosystem Assessment Volume 2: Scenarios Assessment. 2005. doi: 10.1890/080023

3. Bennett AF, Saunders DA. Habitat fragmentation and landscape change. In: Sodhi NS, Ehrlich PR, editors. Conservation Biology for All. New York: Oxford University press; 2010. doi: 10.1093/acprof:oso/9780199554232.003.0006

4. Baudron F, Giller KE. Agriculture and nature: Trouble and strife? Biol Conserv. Elsevier Ltd; 2014;170: 232–245. doi: 10.1016/j.biocon.2013.12.009

5. Laurance WF, Sayer J, Cassman KG. Agricultural expansion and its impacts on tropical nature. Trends Ecol Evol. Elsevier Ltd; 2014;29: 107–116. doi: 10.1016/j.tree.2013.12.001 24388286

6. Kortsch S, Primicerio R, Fossheim M, Dolgov A V, Aschan M. Climate change alters the structure of arctic marine food webs due to poleward shifts of boreal generalists. R Soc Publ. 2015;B 282. doi: 10.5061/dryad.73r6j

7. Nadeau CP, Urban MC, Bridle JR. Climates Past, Present, and Yet-to-Come Shape Climate Change Vulnerabilities. Trends Ecol Evol. Elsevier Ltd; 2017;32: 786–800. doi: 10.1016/j.tree.2017.07.012 28844791

8. Castaño-Uribe C. Diagnóstico y situación actual de las áreas protegidas en América Latina y el Caribe. Informe Regional. Bariloche, Argentina; 2007. doi: 10.1515/pralin-2016-0004.PBML

9. Garcia D. Efectos biológicos de la fragmentación de hábitats: nuevas aproximaciones para resolver un viejo problema. Ecosistemas. 2011;20: 1–10. doi: 10.7818/18

10. Miles L, Newton AC, DeFries RS, Ravilious C, May I, Blyth S, et al. A global overview of the conservation status of tropical dry forests. J Biogeogr. 2006;33: 491–505. doi: 10.1111/j.1365-2699.2005.01424.x

11. Portillo-Quintero CA, Sánchez-Azofeifa GA. Extent and conservation of tropical dry forests in the Americas. Biol Conserv. Elsevier Ltd; 2010;143: 144–155. doi: 10.1016/j.biocon.2009.09.020

12. Aide TM, Clark ML, Grau HR, López-Carr D, Levy MA, Redo D, et al. Deforestation and Reforestation of Latin America and the Caribbean (2001–2010). Biotropica. 2013;45: 262–271. doi: 10.1111/j.1744-7429.2012.00908.x

13. Prieto-Torres DA, Nori J, Rojas-Soto OR. Identifying priority conservation areas for birds associated to endangered Neotropical dry forests. Biol Conserv. Elsevier; 2018;228: 205–214. doi: 10.1016/j.biocon.2018.10.025

14. Strassburg BBN, Brooks T, Feltran-Barbieri R, Iribarrem A, Crouzeilles R, Loyola R, et al. Moment of truth for the Cerrado hotspot. Nat Ecol Evol. 2017;1: 13–15. doi: 10.1038/s41559-017-0099 28812670

15. Ballesteros-Mejia L, Lima JS, Collevatti RG. Spatially-explicit analyses reveal the distribution of genetic diversity and plant conservation status in Cerrado biome. Biodivers Conserv. Springer Netherlands; 2018; 1–18. doi: 10.1007/s10531-018-1588-9

16. Nori J, Torres R, Lescano JN, Cordier JM, Periago ME, Baldo D. Protected areas and spatial conservation priorities for endemic vertebrates of the Gran Chaco, one of the most threatened ecoregions of the world. Divers Distrib. 2016;22: 1212–1219. doi: 10.1111/ddi.12497

17. Eva HD, Belward AS, De Miranda EE, Di Bella CM, Gond V, Huber O, et al. A land cover map of South America. Glob Chang Biol. 2004;10: 731–744. doi: 10.1111/j.1529-8817.2003.00774.x

18. Cabrera AL. Regiones Fitogeográficas Argentinas. In: Kugler W, editor. Enciclopedia Argentina de agricultura y jardinería Tomo II; Fascículo 1. Acme S.A.C.I. Editorial; 1994. pp. 1–85.

19. Dinerstein E, Olsen DM, Graham DJ, Webster AL, Primm SA, Book-binder MP, et al. A conservation assessment of the terrestrial ecoregions of Latin America and the Caribbean. Washington D. C., USA: World Bank, WWF; 1995.

20. Periago ME, Chillo V, Ojeda RA. Loss of mammalian species from the South American Gran Chaco: Empty savanna syndrome? Mamm Rev. 2015;45: 41–53. doi: 10.1111/mam.12031

21. Abril A, Barttfeld P, Bucher EH. The effect of fire and overgrazing disturbes on soil carbon balance in the Dry Chaco forest. For Ecol Manage. 2005;206: 399–405. doi: 10.1016/j.foreco.2004.11.014

22. Grau HR, Gasparri NI, Aide TM. Agriculture expansion and deforestation in seasonally dry forests of north-west Argentina. Environ Conserv. 2005;32: 140–148. doi: 10.1017/S0376892905002092

23. Powell PA, Nanni AS, Názaro MG, Loto D, Torres R, Gasparri NI. Characterization of forest carbon stocks at the landscape scale in the Argentine Dry Chaco. For Ecol Manage. Elsevier; 2018;424: 21–27. doi: 10.1016/j.foreco.2018.04.033

24. Hoekstra JM, Boucher TM, Ricketts TH, Roberts C. Confronting a biome crisis: Global disparities of habitat loss and protection. Ecol Lett. 2005;8: 23–29. doi: 10.1111/j.1461-0248.2004.00686.x

25. Frate L, Acosta ATR, Cabido M, Hoyos L, Carranza ML. Temporal changes in forest contexts at multiple extents: Three decades of fragmentation in the Gran Chaco (1979–2010), Central Argentina. PLoS One. 2015;10: 1–17. doi: 10.1371/journal.pone.0142855 26630387

26. Gasparri NI, Grau HR, Manghi E. Carbon pools and emissions from deforestation in extra-tropical forests of northern Argentina between 1900 and 2005. Ecosystems. 2008;11: 1247–1261. doi: 10.1007/s10021-008-9190-8

27. Houghton RA. Revised estimates of the annual net flux of carbon to the atmosphere from changes in land use and land management 1850–2000. Tellus, Ser B Chem Phys Meteorol. 2003;55: 378–390. doi: 10.1034/j.1600-0889.2003.01450.x

28. Skole D, Tucker C. Tropical Deforestation and Habitat Fragmentation in the Amazon: Satellite Data From 1978 to 1988.”. Science (80-). 1993;260: 1905–1910.

29. Gasparri NI, Grau HR. Deforestation and fragmentation of Chaco dry forest in NW Argentina (1972–2007). For Ecol Manage. 2009;258: 913–921. doi: 10.1016/j.foreco.2009.02.024

30. Brown AD, Pacheco S, Lomáscolo T, Malizia L. Ecorregión Yungas. Situación ambiental en los bosques andinos yungueños. In: Brown A, Martínez Ortíz U, Acerbi M, Corcuera J, editors. La situación ambiental argentina 2005. Buenos Aires, Argentina: Fundación Vida Silvestre Argentina; 2006. pp. 53–56.

31. Izquierdo AE, Grau HR. Agriculture adjustment, land-use transition and protected areas in Northwestern Argentina. J Environ Manage. Elsevier Ltd; 2009;90: 858–865. doi: 10.1016/j.jenvman.2008.02.013 18439743

32. Collevatti RG, Lima-Ribeiro MS, Diniz-Filho JAF, Oliveira G, Dobrovolski R, Terribile LC. Stability of Brazilian Seasonally Dry Forests under Climate Change: Inferences for Long-Term Conservation. Am J Plant Sci. 2013;04: 792–805. doi: 10.4236/ajps.2013.44098

33. Prieto-Torres DA, Navarro-Sigüenza AG, Santiago-Alarcon D, Rojas-Soto OR. Response of the endangered tropical dry forests to climate change and the role of Mexican Protected Areas for their conservation. Glob Chang Biol. 2016;22: 364–379. doi: 10.1111/gcb.13090 26367278

34. Hidasi-Neto J, Joner DC, Resende F, Monteiro L de M, Faleiro FV, Loyola RD, et al. Climate change will drive mammal species loss and biotic homogenization in the Cerrado Biodiversity Hotspot. Perspect Ecol Conserv. 2019;17: 57–63. doi: 10.1016/j.pecon.2019.02.001

35. Jiménez-Valverde A, Lobo JM. Threshold criteria for conversion of probability of species presence to either–or presence–absence. Acta oecologica. 2007;31: 361–369. doi: 10.1016/j.actao.2007.02.001

36. Torres R, Gasparri NI, Blendinger PG, Grau HR. Land-use and land-cover effects on regional biodiversity distribution in a subtropical dry forest: a hierarchical integrative multi-taxa study. Reg Environ Chang. 2014;14: 1549–1561. doi: 10.1007/s10113-014-0604-1

37. Medina RG, Ponssa ML, Aráoz E. Environmental, land cover and land use constraints on the distributional patterns of anurans: Leptodacylusspecies (Anura, Leptodactylidae) from Dry Chaco. PeerJ. 2016;4: e2605. doi: 10.7717/peerj.2605 27833796

38. Ciarleglio M, Wesley Barnes J, Sarkar S. ConsNet: New software for the selection of conservation area networks with spatial and multi-criteria analyses. Ecography (Cop). 2009;32: 205–209. doi: 10.1111/j.1600-0587.2008.05721.x

39. Sarkar S, Illoldi-Range P. Systematic conservation planning: An updated protocol. Nat a Conserv. 2010;8: 19–26. doi: 10.4322/natcon.00801003

40. Moilanen A, Pouzols FM, Meller L, Veach V, Arponen A, Leppänen J, et al. Spatial conservation planning methods and software ZONATION. User Manual. 2014.

41. Watson JE, Grantham HS, Wilson KA, Possingham HP. Systematic conservation planning: past, present and future. In: Ladle RJ, Whittaker R., editors. Conservation biogeography. Oxford: Wi. 2011. pp. 136–160.

42. Kukkala AS, Arponen A, Maiorano L, Moilanen A, Thuiller W, Toivonen T, et al. Matches and mismatches between national and EU-wide priorities: Examining the Natura 2000 network in vertebrate species conservation. Biol Conserv. Elsevier Ltd; 2016;198: 193–201. doi: 10.1016/j.biocon.2016.04.016

43. Brum FT, Graham CH, Costa GC, Hedges SB, Penone C, Radeloff VC, et al. Global priorities for conservation across multiple dimensions of mammalian diversity. Proc Natl Acad Sci. 2017;114: 7641–7646. doi: 10.1073/pnas.1706461114 28674013

44. Lessmann J, Muñoz J, Bonaccorso E. Maximizing species conservation in continental Ecuador: A case of systematic conservation planning for biodiverse regions. Ecol Evol. 2014;4: 2410–2422. doi: 10.1002/ece3.1102 25360277

45. Prieto-Torres DA, Rojas-Soto OR. Reconstructing the Mexican Tropical Dry Forests via an autoecological niche approach: Reconsidering the ecosystem boundaries. PLoS One. 2016;11: 1–22. doi: 10.1371/journal.pone.0150932 26968031

46. Reed Richard Shine RN. Lying in wait for extinction: Ecological correlates of conservation status among Australian elapid snakes. Conserv Biol. 2002;16: 451–461. doi: 10.1046/j.1523-1739.2002.02283.x

47. Nori J, Carrasco PA, Leynaud GC. Venomous snakes and climate change: Ophidism as a dynamic problem. Clim Change. 2014;122: 67–80. doi: 10.1007/s10584-013-1019-6

48. Gibbons JW, Scott DE, Ryan TJ, Buhlmann KA, Tuberville TD, Metts BS, et al. The Global Decline of Reptiles, Déjà Vu Amphibians. Bioscience. 2000;50: 653–666. doi: 10.1641/0006-3568(2000)050[0653:TGDORD]2.0.CO;2

49. Ribeiro MC, Metzger JP, Martensen AC, Ponzoni FJ, Hirota MM. The Brazilian Atlantic Forest: How much is left, and how is the remaining forest distributed? Implications for conservation. Biol Conserv. Elsevier Ltd; 2009;142: 1141–1153. doi: 10.1016/j.biocon.2009.02.021

50. Schneider-Maunoury L, Lefebvre V, Ewers RM, Medina-Rangel GF, Peres CA, Somarriba E, et al. Abundance signals of amphibians and reptiles indicate strong edge effects in Neotropical fragmented forest landscapes. Biol Conserv. Elsevier Ltd; 2016;200: 207–215. doi: 10.1016/j.biocon.2016.06.011

51. Sunny A, González-Fernández A, D’Addario M. Potential distribution of the endemic imbricate alligator lizard (Barisia imbricata imbricata) in highlands of central Mexico. Amphib Reptil. 2017;38: 225–231. doi: 10.1163/15685381-00003092

52. Cardozo G, Chiaraviglio M. Landscape changes influence the reproductive behaviour of a key “capital breeder” snake (Boa constrictor occidentalis) in the Gran Chaco region, Argentina. Biol Conserv. Elsevier Ltd; 2008;141: 3050–3058. doi: 10.1016/j.biocon.2008.08.026

53. Giraudo AR, Arzamendia V, Bellini GP, Bessa CA, Cinthia C, Cardozo G, et al. Categorización del estado de conservación de las Serpientes de la República Argentina. Cuad Herpetol. 2012;26: 303–326.

54. Akani GC, Eniang EA, Ekpo IJ, Angelici FM, Luiselli L. Food Habits of the Snake Psammophis phillipsi from the Continuous Rain-Forest Region of Southern Nigeria (West Africa). J Herpetol. 2003;37: 208–211. Available: http://www.jstor.org/stable/1565855

55. Shine R, Fitzgerald M. Conservation and reproduction of an endangered species: the broad- headed snake, Hoplocephalus bungaroides (Elapidae). Aust Zool. 1989;25: 65–67. doi: 10.1016/j.taap.2007.10.016

56. Dodd CKJ. Strategies for snake conservation. In: Seigel RA, Collins JT, editors. Snakes: Ecology and Behavior. McGraw-Hil. Nueva York; 1993. pp. 363–393.

57. Naggs F. Saving living diversity in the face of the unstoppable 6th mass extinction: A call for urgent international action. Popul Sustain. 2007;1: 67–81.

58. Arzamendia V, Giraudo AR. Using biodiversity patterns for assessment and design protected areas: snakes of Santa Fe province (Argentina) as example. Rev Chil Hist Nat. 2004;77: 335–348. doi: 10.4067/S0716-078X2004000200011

59. Arzamendia V, Giraudo AR. A panbiogeographical model to prioritize areas for conservation along large rivers. Divers Distrib. 2012;18: 168–179. doi: 10.1111/j.1472-4642.2011.00829.x

60. Mota-Vargas C, Rojas-Soto OR. The importance of defining the geographic distribution of species for conservation: The case of the Bearded Wood-Partridge. J Nat Conserv. Elsevier GmbH.; 2012;20: 10–17. doi: 10.1016/j.jnc.2011.07.002

61. Soberón J, Peterson AT. Interpretation of models of fundamental ecological niches and species’ distributional areas. Biodivers Informatics. 2005;2: 1–10. doi: 10.17161/bi.v2i0.4

62. Li B V., Pimm SL. China’s endemic vertebrates sheltering under the protective umbrella of the giant panda. Conserv Biol. 2016;30: 329–339. doi: 10.1111/cobi.12618 26332026

63. Peterson AT, Navarro-Sigüenza AG, Gordillo A. Assumption-versus data-based approaches to summarizing species’ ranges. Conserv Biol. 2018;32: 568–575. doi: 10.1111/cobi.12801 27488935

64. Elith J, Graham CH, Anderson RP, Dudík M, Ferrier S, Guisan A, et al. Novel methods improve prediction of species’ distributions from occurrence data. Ecography (Cop). 2006;29: 129–151. doi: 10.1111/j.1432-1033.1987.tb13499.x

65. Peterson T. Predicting species’ geographic distributions based on ecological niche modeling. Condor. 2001;103: 599–605. doi: 10.1650/0010-5422(2001)103[0599:PSGDBO]2.0.CO;2

66. Araújo MB, Garcia RA, Naimi B, Rahbek C, Barbosa AM, Anderson RP, et al. Standards for distribution models in biodiversity assessments. Sci Adv. 2019;5. doi: 10.1126/sciadv.aat4858 30746437

67. Kremen C, Cameron A, Moilanen A, Phillips SJ, Thomas CD, Beentje H, et al. Aligning conservation priorities across taxa in Madagascar with high-resolution planning tools. Science (80-). 2008;320: 222–226. doi: 10.1126/science.1155193 18403708

68. Pressey RL, Cabeza M, Watts ME, Cowling RM, Wilson KA. Conservation planning in a changing world. Trends Ecol Evol. Elsevier; 2007;22: 583–92. doi: 10.1016/j.tree.2007.10.001 17981360

69. Moilanen A, Franco AMA, Early RI, Fox R, Wintle B, Thomas CD. Prioritizing multiple-use landscapes for conservation: Methods for large multi-species planning problems. Proc R Soc B Biol Sci. 2005;272: 1885–1891. doi: 10.1098/rspb.2005.3164 16191593

70. Ferrier S, Wintle BA. Quantitative approaches to spatial conservation prioritization: matching the solution to the need. Moilanen A, Wilson KA, Possingham H, editors. Oxford University Press; Oxford.; 2009; 1–15.

71. Kukkala AS, Moilanen A. Core concepts of spatial prioritisation in systematic conservation planning. Biol Rev. 2013;88: 443–464. doi: 10.1111/brv.12008 23279291

72. Global Biodiversity Information Facility—GBIF.org. GBIF Occurrence Download: Boa constrictor (2nd February 2018) https://doi.org/10.15468/dl.yjywdt; Epicrates alvarezi (28th August 2017) https://doi.org/10.15468/dl.yjgu4r; Erythrolamprus guentheri (26th September 2017) https://doi.org/10.15468/dl.45loag; Erythrolamprus sagittifer (20th March 2018) https://doi.org/10.15468/dl.pysagt; Leptodeira annulata subsp. pulchriceps (28th August 2017) https://doi.org/10.15468/dl.gv1yot; Philodryas baroni (27th September 2017) https://doi.org/10.15468/dl.fqhypr; Philodryas mattogrossensis (15th September 2017) https://doi.org/10.15468/dl.vtvhym; Phimophis vittatus (22nd March 2018) https://doi.org/10.15468/dl.vl1s5p; Psomophis genimaculatus (27th September 2017) https://doi.org/10.15468/dl.hwhitd; Sibynomorphus lavillai (27th September 2017) https://doi.org/10.15468/dl.bgt1wt; Thamnodynastes chaquensis (27th September 2017) https://doi.org/10.15468/dl.n7pijg; Xenodon pulcher (27th September 2017) https://doi.org/10.15468/dl.qreavk

73. Arzamendia V, Giraudo AR. Influence of large South American rivers of the Plata Basin on distributional patterns of tropical snakes: A panbiogeographical analysis. J Biogeogr. 2009;36: 1739–1749. doi: 10.1111/j.1365-2699.2009.02116.x

74. Hortal IN, Lobo JM, Jimenez-Valverde A. Limitations of Biodiversity Databases: Case Study on Seed-Plant Diversity in Tenerife, Canary Islands. 2007;21: 853–863. doi: 10.1111/j.1523-1739.2007.00686.x 17531062

75. Merow C, Wilson AM, Jetz W. Integrating occurrence data and expert maps for improved species range predictions. Glob Ecol Biogeogr. 2016;26: 243–258. doi: 10.1111/geb.12539

76. Loiselle BA, Jørgensen PM, Consiglio T, Jiménez I, Blake JG, Lohmann LG, et al. Predicting species distributions from herbarium collections: Does climate bias in collection sampling influence model outcomes? J Biogeogr. 2008;35: 105–116. doi: 10.1111/j.1365-2699.2007.01779.x

77. Merow C, Smith MJ, Edwards TC, Guisan A, Mcmahon SM, Normand S, et al. What do we gain from simplicity versus complexity in species distribution models? Ecography (Cop). 2014;37: 1267–1281. doi: 10.1111/ecog.00845

78. Boria RA, Olson LE, Goodman SM, Anderson RP. Spatial filtering to reduce sampling bias can improve the performance of ecological niche models. Ecol Modell. Elsevier B.V.; 2014;275: 73–77. doi: 10.1016/j.ecolmodel.2013.12.012

79. Broennimann O, Cola V Di, Petitpierre B, Breiner F, Scherrer D, D`Amen M, et al. Ecospat: spatial ecology miscellaneous methods. Package “ecospat”. (R Packag Version 30). 2018;2. doi: 10.1111/ecog.02671

80. Di Cola V, Broennimann O, Petitpierre B, Breiner FT, D’Amen M, Randin C, et al. ecospat: an R package to support spatial analyses and modeling of species niches and distributions. Ecography (Cop). 2017;40: 774–787. doi: 10.1111/ecog.02671

81. Fick SE, Hijmans RJ. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int J Climatol. 2017;37: 4302–4315. doi: 10.1002/joc.5086

82. QGIS DevelopmentTeam. QGIS Geographic information system. Open source geospatial foundation project. 2016.

83. Barve N, Barve V, Jiménez-Valverde A, Lira-Noriega A, Maher SP, Peterson AT, et al. The crucial role of the accessible area in ecological niche modeling and species distribution modeling. Ecol Modell. Elsevier B.V.; 2011;222: 1810–1819. doi: 10.1016/j.ecolmodel.2011.02.011

84. Olson DM, Dinerstein E, Wikramanayake ED, Burgess ND, Powell GVN, Underwood EC, et al. Terrestrial Ecoregions of the World: A New Map of Life on Earth. Bioscience. 2001;51: 933. doi: 10.1641/0006-3568(2001)051[0933:TEOTWA]2.0.CO;2

85. Phillips SJ, Andersonb RP, Schapired RE. Maximum entropy modeling of species geographic distributions. Ecol Model 190. 2006; 231–259. doi: 10.1016/j.ecolmodel.2005.03.026

86. Phillips SJ, Dudík M, Elith J, Graham CH, Lehmann A, Leathwick JR, et al. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. 2009;19: 181–197. doi: 10.1890/07-2153.1 19323182

87. Heikkinen RK, Luoto M, Araujo MB, Virkkala R, Thuiller W, Sykes MT. Methods and uncertainties in bioclimatic envelope modelling under climate change. Prog Phys Geogr. 2006;30: 751–777. doi: 10.1177/0309133306071957

88. Qiao H, Soberón J, Peterson AT. No silver bullets in correlative ecological niche modelling: Insights from testing among many potential algorithms for niche estimation. Methods Ecol Evol. 2015;6: 1126–1136. doi: 10.1111/2041-210X.12397

89. Pearson RG, Raxworthy CJ, Nakamura M, Peterson AT. Predicting species distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. J Chem Technol Metall. 2007;34: 102–117. doi: 10.1111/j.1365-2699.2006.01594.x

90. Wisz MS, Hijmans RJ, Li J, Peterson AT, Graham CH, Guisan A, et al. Effects of sample size on the performance of species distribution models. Divers Distrib. 2008;14: 763–773. doi: 10.1111/j.1472-4642.2008.00482.x

91. Muscarella R, Galante PJ, Soley-Guardia M, Boria RA, Kass JM, Uriarte M, et al. ENMeval: An R package for conducting spatially independent evaluations and estimating optimal model complexity for Maxentecological niche models. Methods Ecol Evol. 2014;5: 1198–1205. doi: 10.1111/2041-210X.12261

92. Shcheglovitova M, Anderson RP. Estimating optimal complexity for ecological niche models: A jackknife approach for species with small sample sizes. Ecol Modell. Elsevier B.V.; 2013;269: 9–17. doi: 10.1016/j.ecolmodel.2013.08.011

93. Phillips SJ, Dudík M. Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. 2008;31: 161–175. doi: 10.1111/j.2007.0906–7590.05203.x

94. Peterson AT, Papeş M, Soberón J. Rethinking receiver operating characteristic analysis applications in ecological niche modeling. Ecol Modell. 2008;213: 63–72. doi: 10.1016/j.ecolmodel.2007.11.008

95. Merow C, Smith MJ, Silander JA. A practical guide to MaxEnt for modeling species’ distributions: What it does, and why inputs and settings matter. Ecography (Cop). 2013;36: 1058–1069. doi: 10.1111/j.1600-0587.2013.07872.x

96. Mc Pherson J, Jetz W, Rogers D. The effects of species’ range sizes on the accuracy of distribution models. J Appl Ecol. 2004;41: 811–823.

97. Osorio-Olvera L. NicheToolBox R package. [Internet]. 2016.

98. Ortega-Andrade HM, Prieto-Torres DA, Gómez-Lora I, Lizcano DJ. Ecological and geographical analysis of the distribution of the Mountain Tapir (Tapirus pinchaque) in Ecuador: Importance of protected areas in future scenarios of global warming. PLoS One. 2015;10: 1–20. doi: 10.1371/journal.pone.0121137 25798851

99. Lizcaíno D, Prieto-Torres D, Ortega-Andrade M. Distribución de la danta de montaña (Tapirus pinchaque) en Colombia: importancia de las áreas no protegidas para la conservación en escenarios de cambio climático. In: Payán E, Lasso C, Castaño-Uribe C, editors. Conservación de grandes vertebrados en áreas no protegidas de Colombia, Venezuela y Brasil. Serie Edit. 2015. pp. 115–129.

100. R Core Team. R: A Language and Environment for Statistical Computing. In: R Foundation for Statistical Computing, Vienna, Austria. 2012.

101. Prieto-Torres DA, Pinilla-Buitrago G. Estimating the potential distribution and conservation priorities of Chironectes minimus (Zimmermann, 1780) (Didelphimorphia: Didelphidae). Therya. 2017;8: 131–144. doi: 10.12933/therya-17-478

102. Di Minin E, Veach V, Lehtomaki J, Pouzols F., Moilanen A. A quick introduction to Zonation. Unigrafia OY, Helsinki.; 2014.

103. Moilanen A. Landscape Zonation, benefit functions and target-based planning: Unifying reserve selection strategies. Biol Conserv. 2007;134: 571–579. doi: 10.1016/j.biocon.2006.09.008

104. Moilanen A, Anderson BJ, Eigenbrod F, Heinemeyer A, Roy DB, Gillings S, et al. Balancing alternative land uses in conservation prioritization. Ecol Appl. 2011;21: 1419–1426. doi: https://doi.org/10.1890/10-1865.1 21830691

105. Moilanen A. Two paths to a suboptimal solution—once more about optimality in reserve selection. Biol Conserv. 2008;141: 1919–1923. doi: 10.1016/j.biocon.2008.04.018

106. Moilanen A, Wintle BA, Elith J, Burgman M. Uncertainty analysis for regional-scale reserve selection. Conserv Biol. 2006;20: 1688–1697. doi: 10.1111/j.1523-1739.2006.00560.x 17181804

107. Lizarraga L, Monguillot J. Mapa de Huella Humana para Argentina. Argentina; 2018.

108. Sanderson EW, Jaiteh M, Levy MA, Redford KH, Wannebo A V, Woolmer G. The Human Footprint and the Last of the Wild. Bioscience. 2002;52: 891. doi: 10.1641/0006-3568(2002)052[0891:THFATL]2.0.CO;2

109. Faleiro F V., Machado RB, Loyola RD. Defining spatial conservation priorities in the face of land-use and climate change. Biol Conserv. Elsevier Ltd; 2013;158: 248–257. doi: 10.1016/j.biocon.2012.09.020

110. UNEP 2010. Report of the Tenth Meeting of the Conference of the Parties to the Convention on Biological Diversity. Held in Nagoya, Japan on 18–29 October 2010.

111. Geist HJ, Lambin EF. Proximate Causes and Underlying Driving Forces of Tropical Deforestation. Bioscience. 2002;52: 143. doi: 10.1641/0006-3568(2002)052[0143:pcaudf]2.0.co;2

112. Dobrovolski R, Loyola RD, Guilhaumon F, Gouveia SF, Diniz-filho JAF. Global agricultural expansion and carnivore conservation biogeography. Elsevier Ltd; 2013;165: 162–170. doi: 10.1016/j.biocon.2013.06.004

113. Segura C, Feriche M, Pleguezuelos JM, Santos X. Specialist and generalist species in habitat use: Implications for conservation assessment in snakes. J Nat Hist. 2007;41: 2765–2774. doi: 10.1080/00222930701664203

114. Giraudo A. Deforestación, defaunación como consecuencia de las actividades humanas en la llanura del Chaco Argentino. El Chaco sin bosques: La Pampa o el desierto del futuro. 2009. doi: 10.1017/CBO9781107415324.004

115. Hoyos LE, Cingolani AM, Zak MR, Vaieretti M V., Gorla DE, Cabido MR. Deforestation and precipitation patterns in the arid Chaco forests of central Argentina. Appl Veg Sci. 2013;16: 260–271. doi: 10.1111/j.1654-109X.2012.01218.x

116. Pacheco S, Malizia LR, Cayuela L. Effects of climate change on subtropical forests of South America. Trop Conserv Sci. 2010;3: 423–437. doi: 10.1177/194008291000300407

117. Schmidt M. Situación de la tierra en la provincia de Salta. Una aproximación al contexto previo al Ordenamiento Territorial de Bosques Nativos. Estud Rural. 2012;2.

118. Pacheco S, Brown AD. La biodiversidad de la ecorregión de las Yungas ¿es sustentable a largo plazo? In: Brown AD, Martínez Ortíz U, Acerbi M, Corcuera J, editors. La Situación Ambiental Argentina 2005. Fundación. 2006. pp. 59–61.

119. Brown AD, Malizia LR. Las Selvas Pedemontanas de las Yungas. Fundación Pro-Yungas. 2004

120. Zwiener VP, Padial AA, Marques MCM, Peterson AT, Faleiro F V, Loyola R. Planning for conservation and restoration under climate and land use change in the Brazilian Atlantic Forest. Divers Distrib. 2017; 1–12. doi: 10.1111/ddi.12588

121. Hansen MC, Potapov P V., Moore R, Hancher M, Turubanova SA, Tyukavina A, et al. High-resolution global maps of 21st-century forest cover change. Science (80-). 2013;342: 850–853. doi: 10.1126/science.1244693 24233722


Článek vyšel v časopise

PLOS One


2019 Číslo 9