Modelling jaguar gene flow in fragmented landscapes offers insights into functional population connectivity

Abstract

Abstract Context Preserving functional connectivity is a conservation priority to secure the long-term viability of geographically dispersed subpopulations, such as the jaguar (Panthera onca) populations in Central America. However, managing connectivity in this region is limited due to the scarcity of local assessments of the connectivity between existing populations, some of which exhibit low levels of gene flow and genetic admixture. Objectives We selected the jaguar as a model species to understand how gene flow of large carnivores is shaped in a heavily human-impacted landscape in the Neotropics. We assessed the impact of landscape features and spatial scale on jaguar gene flow across northern Central America, predicted synoptic, landscape-wide functional connectivity across the region; and compared connectivity predictions based on models of gene flow, habitat suitability, and expert knowledge. Methods We employed genetic data based on 335 faecal samples collected from 72 individual jaguars. We parameterized gene flow resistance surfaces using linear mixed effects models and the maximum likelihood population-effects method. We evaluated nine landscape variables at six spatial scales, selecting the optimal scale and transformation for each variable according to univariate models and AIC. To predict jaguar gene flow, we developed multivariate models and implemented resistant kernels to forecast functional connectivity between jaguar populations across the study region under three dispersal distance scenarios. Furthermore, we compared the connectivity estimates based on gene flow against those based on habitat suitability and the corridors delineated by expert knowledge in the region. Results Low resistance to jaguar gene flow was associated with greater tree cover and vegetation, lower areas of built-up, and intermediate distances from water bodies. Notably, tree cover affected jaguar gene flow on a smaller scale compared to the rest of variables. Higher connectivity, indicated by lower resistance, was found within and around the largest jaguar conservation units (JCUs) such as the Reserva-de-Biosfera-Transfronteriza, Selva-Maya and Maya-Mountains. This contrasted with the smallest JCUs, like Sierra-Santa-Cruz and Sierra-de-las-Minas in eastern Guatemala, and Cordillera-Nombre-de-Dios in Honduras. Across the region, lower connectivity was observed in the Caribbean connection between eastern Guatemala and midwestern Honduras, as well as in the Honduran and Nicaraguan Miskito area. Models based on gene flow and habitat suitability were similar in their predictions of areas of high connectivity; however, the habitat suitability models predicted larger areas of low connectivity than did the gene flow models. Moreover, the expert knowledge corridors were consistent with areas of high and medium connectivity as predicted by the gene flow model. Conclusions Gene flow of jaguars is positively driven by the presence of forest and water sources, while human impact (built-up and non-vegetation areas) has a negative effect. Areas of lowest resistance largely correspond to the location of JCUs, which serve as crucial reservoirs of high-quality jaguar habitat within the region. While the largest JCU in the region displayed high connectivity, the low connectivity between the smallest JCUs (i.e. Sierra-Santa-Cruz, Sierra-de-las-Minas and Cordillera-Nombre-de-Dios) underscores the need for conservation attention in these areas. Conservation and management actions such as habitat loss-prevention/restoration and anthropogenic impact mitigation should be prioritized in the binational region of Guatemala-Honduras, a key connectivity bottleneck between the species’ northern and southern ranges. Similarly, attention is warranted in the Honduran-Nicaraguan Miskito area.