Biodiversity, Connectivity, and Impacts of Anthropogenic Barriers on Vertebrate Communities in the Mexico-USA Borderlands

Abstract

The Sky Islands region between Mexico and the United States has been recognized as a biodiversity hotspot due to the convergence of Neartic and Neotropical species (Coronel-Arellano et al., 2018; DeBano et al., 1995; Warshall, 1995). For mammals along the US-Mexico border, the northeastern portion of Sonora and the northwestern portion of Chihuahua are considered to be corridors for jaguars (Panthera onca), ocelots (Leopardus pardalis), and black bears (Ursus americanus), all of which use riparian and montane forests to move among the Sky Islands (Atwood et al., 2011; Grigione et al., 2009; Sierra-Corona et al., 2005). Large-scale developments in the region threaten landscape connectivity and the persistence of native wildlife populations (McCormack et al., 2009; Peters et al., 2018). The US-Mexico border wall bisects the distribution range of 346 species of nonflying terrestrial and freshwater animals, including 120 species of mammals, and will impact the cross-border movement of over 100 species under current climate scenarios (Peters et al., 2018; Titley et al., 2021). In this study, we address the ongoing fragmentation of one of the most diverse ecosystems in North America, focusing on the impacts of human barriers, discussing the possibilities of emerging techniques for monitoring terrestrial mammals faster and cost-effective, and investigating the spatiotemporal patterns of apex predators in the region. In chapter one, we focused on the international border between Arizona and Mexico, which traverses a vast landscape of extreme biodiversity. We aimed to determine how mammals' diversity and habitat use change across significant landscape features: the Cajon Bonito stream, the border wall, and the most transited highway in the area. We installed 85 camera traps and recorded 31 species of terrestrial mammals from 15 families and identified two different jaguar individuals and other cryptic species: the ocelot, porcupine and beaver. Diversity analyses showed that vegetation type influenced the composition of mammals in combination with the presence of cattle and distance to the road. The presence of cars, distance to the border wall, and effort, also affected the composition as non-interacting covariates. We found that human influence has a negative effect on the habitat use of mammals. Areas closer to the border are less used by larger mammals. However, mesopredators and small herbivores were more tolerant to disturbance and used sites closer to the border and dirt roads, but they exhibited a negative association with the presence of humans and sites closer to the border wall. Also, the presence of cattle negatively affected the habitat use and detectability of large and small carnivores. Our results highlight the need to keep bi-national connectivity for mammal populations using mountain corridors and prioritize locations that provide water sources to wildlife to mitigate highway and border wall impacts. This chapter is formatted for submission to the Science of the Total Environment journal. Given that traditional monitoring methods, particularly camera trapping, can be expensive, labor-intensive, and susceptible to equipment failure or theft, especially when elusive and low-abundant species of interest are present. In the second chapter, we explored the effectiveness of environmental DNA metabarcoding for monitoring terrestrial mammals as a potentially more cost-effective and faster alternative. We surveyed a 20 km section of the perennial Cajon Bonito stream in Mexico, deployed camera traps, and collected eDNA water samples at ten locations along the stream during wet and dry seasons. The camera traps operated continuously for 31 days before each eDNA sampling event. The results indicated that camera traps detected significantly more species (18 in the wet season, 17 in the dry) than eDNA (5 and 4 species, respectively). While some species detected via eDNA were also detected using camera traps, the beaver (Castor canadensis) was exclusively detected with eDNA. We analyzed the relationship between camera trap detections and eDNA detection, considering temporal proximity between detection events and spatial proximity. However, no significant correlation was found, contrasting with findings from more humid environments. The lack of correlation suggests that the factors influencing eDNA detection in arid environments, such as DNA persistence, degradation, and transport, make this technique less reliable than in humid environments. Also, the lower structural complexity of arid vegetation compared to humid environments might reduce DNA persistence and transport. While eDNA metabarcoding shows promise as a supplementary tool, camera traps remain more effective in assessing overall mammalian diversity in this arid ecosystem. Finally, in chapter three, we investigated the spatiotemporal ecology of three apex predators, black bears (Ursus americanus), pumas (Puma concolor), and jaguars (Panthera onca) in the US-Mexico borderlands. We used a 22-month camera trapping dataset, employing 85 motion-activated cameras placed along a riparian corridor and in a grid parallel to the international border. We focused on (i) spatial overlap, (ii) seasonal activity patterns, (iii) daily activity patterns, (iv) interspecific activity overlap, and (v) the effect of the presence of predators over two prey species, the white-tailed deer (Odocoileus virginianus) and the collared peccary (Dicotyles tajacu). We found spatial overlap between the three species of predators in riparian forests, and black bears and pumas in pine-oak forests, but pumas were typically the only apex predator in the scrublands and grasslands. We found distinct seasonal activity patterns among the three predators, with black bears showing peak activity in June, pumas exhibiting consistent activity throughout the year, and jaguars demonstrating peak activity from August to January. Daily activity was predominantly diurnal for bears and did not overlap with the two felids. Puma and jaguar activity overlapped, but puma activity peaked at dusk and jaguars during the night. The habitat use of both prey species responded positively to the presence of black bears but not to the felids. We suggest that given the dominance of black bears over both felids, prey species might be using those sites as a refuge from their other predators, exploiting avoidance of bears by the felids that describe the complexity of spatiotemporal partitioning.