Habitat loss and fragmentation from expanding agricultural land use is a principle threat to biodiversity conservation (Foley, 2005; Haddad et al., 2015) as they negatively affect populations by reducing habitat availability, landscape connectivity, and the movement of individuals and genes (Fahrig, 2007, 2003). For mammals, extinction risk increases with increasing fragmentation and loss of connectivity, with increased risks for large terrestrial carnivores (Crooks, 2002; Crooks et al., 2017, 2011), which is evident in the high rates of extinction and decreases in distribution and abundance of these species worldwide (Ceballos, 2002; Ripple et al., 2014). Additionally, the negative effects on habitat availability and connectivity from agricultural expansion and the associated increased human footprint is exacerbated for large carnivores as they facilitate the hunting of carnivores and their prey, while also leading to increased persecution from conflicts with ranching interests or out of fear (Machovina et al., 2015; Ripple et al., 2014).

The implications of the loss of habitat and landscape connectivity from agricultural land conversion is evident for the jaguar (Panthera onca), the largest felid in the Americas and the apex predator throughout its distribution, whose range has contracted by ∼50% from its historic extent (de la Torre et al., 2018; Petracca et al., 2018; Thompson and Velilla, 2017). In response to the declining population and distribution of jaguars a range-wide conservation framework has developed that consists of the conservation of core populations (Jaguar Conservation Units; hereafter JCUs), largely dependent upon protected areas, linked by connectivity corridors within anthropogenic landscapes (Rabinowitz and Zeller, 2010; Sanderson et al., 2002; United Nations Development Program et al., 2019). Since most protected areas are insufficiently small to maintain viable populations in the absence of sufficient connectivity given the jaguars’ large spatial needs (Morato et al., 2016; Thompson et al., 2021), the survival and movement of jaguars through anthropogenic landscapes outside of protected areas is essential for the long-term survival of the jaguar (López-Bao et al., 2017).

Despite the contraction of the jaguar’s occurrence outside of protected areas (de la Torre et al., 2018), and anthropogenically driven loss of genetic diversity in jaguar populations throughout its range (Haag et al., 2010; Lorenzana et al., 2020, 2022; Roques et al., 2016; Srbek-Araujo et al., 2018; Wultsch et al., 2016), the large majority of research on jaguar population ecology has occurred within protected areas (Boron et al., 2016; Foster et al., 2020; Tobler and Powell, 2013). This lack of research focus upon anthropogenic landscapes in population estimation is of concern as the inferences generated from research in protected areas may not be representative of populations outside those areas and could lead to misinformed decision making (Allen et al., 2017; Tobler and Powell, 2013). Consequently, there is an inherent need for a shift in research on jaguar population ecology to focus more upon anthropogenic landscapes. This is especially true for areas outside of the Amazon where population declines have been greatest and where the conservation of remaining populations is highly dependent upon the jaguar’s movement through and survival in anthropogenic landscapes.

The negative effects of agricultural land use on the jaguar is most notable at the southern limit of its distribution where it has been extirpated from Uruguay, most of its range in Argentina, and the majority of the Atlantic forest of Brazil and Paraguay (de la Torre et al., 2018; Paviolo et al., 2016; Quiroga et al., 2014). Historically, the Dry Chaco of Argentina, Bolivia, and Paraguay has been important habitat for jaguar, and until recently the Dry Chaco of western Paraguay had been a population stronghold and connectivity hub for the species (Sanderson et al., 2002; Rabinowitz and Zeller, 2010; Thompson and Velilla, 2017). However, during the last twenty years, the Dry Chaco of Paraguay has become a deforestation hotspot, undergoing some of the world’s highest rates of deforestation, mostly driven by expanding livestock production.(Baumann et al., 2017; Caldas et al., 2015; Hansen et al., 2013; Vallejos et al., 2015), This deforestation has reduced the availability of preferred jaguar habitat, while increasing persecution and reducing prey availability, resulting in increased space use by jaguars and decreases in their abundance and population connectivity (Romero-Muñoz et al., 2020, 2019; Thompson et al., 2021, 2020; Thompson and Velilla, 2017).

As is typical of the majority of jaguar density estimates, population surveys of jaguar in the Dry Chaco have been conducted in relation to protected areas and indigenous reserves. In Bolivia, densities of 0.31–1.82 individuals/km2 were estimated (Noss et al., 2012), while in Argentina surveys showed that the jaguar is nearly extirpated in the region (Quiroga et al., 2014). Despite the extensive habitat loss and its apparent negative impacts on jaguars in the Dry Chaco (Romero-Muñoz et al., 2019), there have been no population surveys of jaguars in the region’s working landscapes, nor has there been any rigorously conducted surveys in the Paraguayan Dry Chaco or anywhere else in Paraguay. Moreover, there has not been an analysis of the effects of land use in the Dry Chaco on jaguar density and population connectivity.

Consequently, given the need to better understand the implications of widespread deforestation for the jaguar in general, and specifically in the Paraguayan Dry Chaco, we were interested in quantifying the effect of forest cover on jaguar density and population connectivity in deforestation hotspots, using the Paraguayan Dry Chaco as a case study. To do this we surveyed jaguars using camera trapping along a deforestation gradient in the Paraguayan Dry Chaco in a spatial capture-recapture (SCR) framework, taking advantage of modeling developments that allow for jointly estimating density and landscape connectivity (Morin et al., 2017; Royle et al., 2013; Sutherland et al., 2015). Hence, our objectives were to generate 1) the first spatially explicit density estimates for jaguars in Paraguay, and specifically for the Paraguayan Dry Chaco, and 2) the first estimate of the effects of habitat conversion on jaguar densities and population connectivity. Aside from important applications for the conservation of the jaguar in the Dry Chaco, our findings and the use of SCR modeling extensions have broader applications for understanding the effects of large-scale habitat loss and land management on jaguar density and population connectivity throughout the species’ distribution.

We detected between 6–13 jaguars per site and one GPS collared individual from one site was censored in the analysis after it moved out of our sampling area and was killed. The ratio of males to females photographed ranged from 0.5 to 2.25 and the number of spatial recaptures per site was 38–116 (Table 1). Of the model set, three models contained 100% of the model weights (Table S1), with the model containing a site-specific effect on σ the highest ranked supported by 71% of model weights. The second and third highest ranked models included uninformative effects of sex on detection and σ (Table S1). The highest ranking model had a significant positive effect of effort on detection (βeffort = 0.013, SE = 0.001; Table S2) and a significant negative effect of proportional forest area on landscape resistance (βδ = −1.306, SE = 0.05; Table S2).

Based upon the highest ranking model, density estimates ranged from 0.44 to 1.6 individuals/100 km2, baseline detectability was estimated to be between 32% and 66%, and estimates of σ ranged from 1.0 to 1.9 km (Table 2). Detectability increased by about 0.3 percent per day of sampling effort and landscape resistance (δ) decreased by about 12% for each 10% increase in forest area, translating to fully deforested areas having a landscape resistance 3.7 times greater compared to fully forested areas. Sex ratio estimates had high uncertainty, but the estimated probability of being a male varied from 36% to 70% across sites (Table 2).

Site-specific estimated densities, potential connectivity, and density weighted connectivity showed a general negative relationship with increasing deforested area (Fig. 2). Estimated density between the two least deforested sites (sites 1 and 2) were not significantly different at the 95% level, nor was the difference between the most deforested sites (sites 3 and 4; Fig. 3a) significant. However, estimated density for the least deforested sites (sites 1 and 2) were each significantly different from each of the most deforested sites (sites 3 and 4; Fig. 3a). The mean values of estimated potential connectivity were significantly different among all sites with the exception being between the two most deforested sites (sites 3 and 4; Fig. 3b), while the differences in the mean density-weighted connectivity were significantly different among all sites (Fig. 3c).

Fig. 2.

The relationship of proportional area deforested within the modeled state spaces with (a) density (error bars represent the 95% confidence interval), (b) standardized potential connectivity, and (c) standardized density-weighted connectivity. Boxes show the 25th to 75th percentile, vertical lines within the boxes the median, and the whiskers 1.5 times the interquartile range. Points in the boxes represent the mean value.

(0.29MB).

Fig. 3.

Difference between site-specific (a) density estimates, (b) mean estimated potential connectivity, and (c) mean estimated density-weighted connectivity. Points represent differences and bars represent the 95% confidence intervals. Where confidence intervals do not include 0 differences are significant. Percentages associated with site names represent the percent area deforested for state space of each trapping grid.

(0.27MB).

By taking advantage of newly developed modeling approaches and sampling across a gradient of deforestation intensity, our study is the first to examine the effect of landscape transformation on jaguar density and connectivity, apart from producing the first rigorous estimates of jaguar density for Paraguay. More generally, despite the recognized importance of connectivity for the long-term persistence of species (Lindenmayer et al., 2008), to date few studies have simultaneously estimated density and connectivity (Fuller et al., 2016; Pal et al., 2021), and this study is the first to do this across differing landscape structures. Despite the uncertainty in our density estimates we were able to show that estimated densities in the more deforested sites were significantly lower compared to densities in the less deforested sites. Moreover, by incorporating ecological distance in the SCR formulation, we found a strong dependence of connectivity and jaguar space use on forest area which was consistent with observed patterns in jaguar spatial ecology (Alvarenga et al., 2021; Morato et al., 2018a, 2018b; Thompson et al., 2021).

We did not find sex-specific differences in detectability or for σ which was contrary to other jaguar studies (Boron et al., 2016; Sollmann et al., 2011; Tobler et al., 2018, 2013) and not expected since space use and movements by male jaguars in the Dry Chaco, and range-wide, is considerably larger than that of females (McBride and Thompson, 2018; Morato et al., 2016; Thompson et al., 2021). The lack of sex-specific differences in detectability and σ could be explained by variations in individual space use and differences in sex ratios among sites offsetting the frequency of males and females detected across sites (Table 1). Estimates of σ were smaller than other studies (Boron et al., 2016; Sollmann et al., 2011; Tobler et al., 2013) which is attributable to our accounting for habitat-driven asymmetry in space use (Tobler et al., 2018). This was expected since the ecological distance formulation of the SCR model estimates smaller values of σ compared to the Euclidean distance formulation as the effect size of δ increases (Morin et al., 2017; Royle et al., 2013; Sutherland et al., 2015).

Jaguar occurrence in the Dry Chaco is subject to complex source-sink dynamics, driven by habitat loss, high anthropogenic mortality, and prey depletion (McBride and Thompson, 2018; Romero-Muñoz et al., 2020, 2019; Thompson et al., 2020). Consequently, our estimates of density and connectivity in relation to forest cover result from an interplay of the negative effects of high anthropogenic mortality and prey suppression acting synergistically with habitat loss. We found that connectivity in the more transformed landscapes is maintained, although greatly reduced, suggesting that populations can be augmented through immigration and that our study sites are potentially serving as populations sinks. Moreover, the rapid rate of forest loss in the Paraguayan Dry Chaco, and the longevity of jaguars, raises additional concerns that we underestimated the negative population effects of habitat loss as the true effect could be masked via a time lag (Tilman and Lehman, 1994) as has been observed for other mammals in the Dry Chaco (Semper-Pascual et al., 2018). The uncertainties over the degree of demographic influence of these mechanisms point to the need for long-term monitoring of jaguar populations to estimate vital rates and immigration, however, there are conspicuously few such estimates for jaguars (Glennie et al., 2019; Gutiérrez-González et al., 2015; Harmsen et al., 2017). Given high anthropogenic mortality and evidence of source-sink dynamics for the jaguar in the Dry Chaco (Romero-Muñoz et al., 2019), long-term monitoring of jaguars should be a priority for their conservation in the region, as well as for the species throughout its distribution.

In Paraguay the jaguar is listed as critically endangered due to its country-wide range constriction and habitat loss, with the majority of its remaining distribution being in the Dry Chaco (Giordano et al., 2017). Given the expected continuation of habitat loss, the relatively small amount of protected area in the region in relation to the jaguar’s spatial needs, and high levels of human-caused mortality (McBride and Thompson, 2018; Romero-Muñoz et al., 2019), the negative effects of deforestation on jaguar density and connectivity that we observed are of concern for the long-term conservation of the species in the Paraguayan Dry Chaco. This concern is further warranted as land use trajectories and jaguar population declines in the Dry Chaco are mirroring those that have led to the near extirpation of jaguars in the Atlantic Forest of eastern Paraguay, whereby extensive deforestation for agricultural production has confined the occurrence and movements of the remaining jaguars to protected areas and adjacent habitat patches with no connectivity among other populations (McBride and Thompson, 2019; Paviolo et al., 2016).

Within the Paraguayan Dry Chaco landowners are legally permitted to clear between 50% to 75% of forest area within their properties (Congreso de la Nación Paraguaya, 1973; Secretaria del Ambiente de Paraguay, 2001) and subsequently, our findings suggest that if private land is fully developed, the landscape outside of protected areas will be at best marginally suitable to maintain jaguars and likely serve as population sinks, while protected areas will be increasingly isolated (Romero-Muñoz et al., 2019). However, we showed that connectivity in our more transformed landscapes, albeit reduced by increasing forest loss, can be consciously maintained and we urge the enforcement of national forest conservation laws, the application of landscape planning, and the development of initiatives to incentivize forest conservation to achieve that. In addition, the high anthropogenic mortality of jaguars in our study area points to the need for habitat protection to be undertaken concurrently with efforts to reduce anthropogenic mortality. As jaguar killings are rooted in real or perceived livestock losses from depredation, enforcing and implementing relevant sections of the national jaguar conservation law (Congreso de la Nación Paraguaya, 2014) and management plan (Secretaría del Ambiente et al., 2016) to foster coexistence between the ranching community and jaguars would aid in reducing the negative effects from habitat loss through increasing jaguar survival on private ranchlands.

The anthropogenic context of jaguar conservation in the Paraguayan Dry Chaco is typical of jaguar populations throughout the species’ range and demonstrates the broad relevance of our findings as they show how jaguar density and connectivity relate to habitat loss in anthropogenic landscapes. Moreover, as jaguars exhibit strong spatial associations with forest cover and rivers (Alvarenga et al., 2021; Eriksson et al., 2022; Morato et al., 2018a, 2018b; Thompson et al., 2021), the strong asymmetrical habitat-driven space use that we found highlights the need to consider the potential negative bias induced in SCR modeling from assuming symmetrical space use when estimating jaguar density in heterogeneous landscapes, as well as the value of formally estimating connectivity within the SCR modeling framework (Morin et al., 2017; Sutherland et al., 2015; Tobler et al., 2018). Importantly, as the large majority of jaguar population studies have been in protected areas (Boron et al., 2016; Foster et al., 2020; Tobler and Powell, 2013), our results show that these studies are potentially generating skewed inferences for conservation as they fail to account for the complex spatial dynamics of populations in anthropogenic landscapes.

The case of the jaguar in the Paraguayan Dry Chaco illustrates the importance of understanding the drivers of jaguar population ecology in the landscape matrix outside of protected areas for the species’ long-term survival. Moreover, our findings demonstrate that adequately implementing range-wide initiatives for jaguar conservation (United Nations Development Program et al., 2019) will require a focus on quantifying how land use affects populations and movements of jaguars in anthropogenic landscapes, as well as on interventions promoting coexistence and habitat conservation, to achieve positive conservation outcomes in anthropogenic landscapes in Paraguay and across the jaguar’s range.

J.J.T., M.V., J.M.C.K. and J.L.C. conceptualized the project and designed the methodology; J.J.T., M.V., V.R.B. administered the project; J.J.T., M.V., H.C., N.C., V.R.B., E.B., J.M.C.K., R.T.M., and R.A. collected the data; J.J.T. and N.C. curated data; J.J.T. analyzed the data; J.J.T., M.V., and J.L.C. led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.