Modeling Potential Movement Corridors for Jaguars in the U.S.-Mexico Borderlands Using Least-Cost Path Analysis

Abstract

The development of security infrastructure at the U.S.-Mexico borderlands threatens to degrade functional connectivity for many Neotropical taxa whose range limits converge here. One such species is the jaguar (Panthera onca), in its northernmost range in the borderlands of Arizona and Sonora. An important conservation tool for this carnivore has been the implementation of corridors to reduce isolation of habitat fragments. However, less studies have focused on the potential effects of border wall development on jaguar dispersal between habitat cores. This study presents a connectivity model for jaguar conservation in support of gap placement decisions using a least-cost path analysis within a geographic information system (GIS). Landscape characteristics considered to most affect jaguar movement were combined into a cost surface using values derived from expert opinion, and optimal routes of travel across the surface were calculated. The effects of a border wall were simulated by defining the wall as an input barrier to the model, repeating least-cost path analysis, and comparing movement corridors with and without the wall. Results indicate that two predicted least-cost paths connecting habitat cores in Sonora to those in Arizona are disrupted, leaving only one lowest cost neighboring path between an existing gap 11.9 km wide. In addition, a continuous border wall in the east would isolate habitat in Arizona from the northernmost breeding population in Sonora. Conservation planners could use these results to implement effective mitigation strategies should security infrastructure development in the region continue. 

Introduction

Habitat fragmentation at a variety of spatial scales has been acknowledged as a primary cause of the decline of many species worldwide (Epps et al., 2005; Puyravaud et al., 2016; Walker & Craighead, 1997). The loss of functional connectivity through habitat fragmentation generally leads to smaller and more isolated animal populations (Walker & Craighead, 1997), thereby impeding exchange of individuals among populations and accelerating the loss of genetic diversity (Hedrick, 2009; Epps et al., 2005). To reduce isolation of habitat fragments, conservation biologists have recommended maintaining functional connectivity through the use of corridors (Beier & Noss, 1998). Ideally, corridors increase movement and genetic exchange between populations in habitat fragments and rescue populations from extinction (Alexander et al., 2016). Although corridor conservation strategies have been widely applied (e.g., LaRue & Nielsen, 2008; Puyravaud et al., 2016), less studies have focused on the potential effects of border wall development on transboundary connectivity at the U.S.-Mexico borderlands. Transboundary connectivity is especially important to conservation in this region as several geographic provinces converge to produce the range limits of many Neotropical taxa (Flesch et al., 2009). 

One such species is the jaguar (Panthera onca), an endangered species (USFWS, 2014; Gonzalez-Gallina et al., 2018) and the largest cat in the New World, in its northernmost range in the borderlands of Arizona and Sonora. Near the species' northern limits, Sonora supports the northernmost breeding population of jaguars from which individual males disperse into its historical range in Southern Arizona (Chambers et al., 2022). To assess the effects of border wall development on jaguar movement, a least-cost path analysis using a geographic information system (GIS) was selected. Rabinowitz & Zeller (2010) used least-cost path analysis and ArcGIS v9 to identify corridors connecting known populations across the jaguar's range. Similarly, Chambers et al. (2022) measured the effects of border barriers on the required energy expenditure of individual jaguars near Nogales, Arizona and Sonora using the ArcGIS Path Distance tools. The objective of this study was to model movement corridors for jaguars between habitat cores at the U.S.-Mexico borderlands in support of gap placement decisions. In particular, the following questions are addressed:

1.  Do border barriers significantly affect interpopulation connectivity for the jaguar?

2.  What gap placement decision best optimizes transboundary movement between habitat cores?

Methods

A. Study Area

Using jaguar habitat core polygons derived from previously confirmed jaguar observations (Sanderson & Fisher, 2011 2013; Theobald et al., 2017), eight habitat cores were selected at the borderlands of Southern Arizona and Sonora (Figure 1). The cores represent the top 25% of suitable jaguar habitat (Theobald et al., 2017) and cover a total area of 1,439.4 km2. Lowland vegetation in the Arizona-Sonora borderlands is dominated by three vegetation communities: semidesert grassland at higher elevations to the north and east, Arizona Upland desert scrub at moderate elevations, and lower Colorado River Valley desert scrub at the lowest elevations in the west (Flesch et al., 2009). The area was selected because of its jaguar occurrences (Babb et al., 2022) and relevant importance to guide conservation in the region. 

B. Data

Five landscape characteristics considered to most affect jaguar movement and survival were selected: elevation, land cover type, percent tree cover, distance from roads, and the presence of a border wall (Table 1). Elevation, land cover type, and percent tree cover are closely related to the movement behavior of jaguars (Rabinowitz & Zeller, 2010), whereas distance from roads is considered to be correlated with human persecution of jaguars (Rabinowitz, 2005). In addition, the presence of a border wall is considered to impede movement because a jaguar must circumvent the wall rather than walking directly across it (Chambers et al., 2022).

To describe elevation in the study area, Shuttle Radar Topography Mission (SRTM) 1-Arc Second Global tiles were mosaiced into a single digital elevation model (DEM) using the ArcGIS Mosaic To New Raster tool. This same process was applied to the Global Landsat Tree Canopy dataset to create a single raster for percent tree cover. The Land Cover of North America dataset, as well as the North American Roads and U.S.-Mexico Border Fence vector datasets, were clipped to the study area using the ArcGIS Analysis and Data Management tools. All layers were standardized to the same projection and the raster layers were resampled to a 30 m2 resolution.

C. Least-Cost Path Analysis

A least-cost path analysis was performed using ModelBuilder in ArcGIS Pro 3.3. Least-cost path modelling is based upon a cost surface raster where values within the raster are used to represent the costs associated with traversing different parts of the landscape (Etherington, 2016; Esri, 2024). To create the cost surface, a distance-to-roads raster was first derived from the North American Roads vector dataset using the ArcGIS Distance Accumulation tool. The DEM, land cover, percent tree cover, and distance-to-roads layers were then transformed to a common scale with the ArcGIS Reclassify tool utilizing the landscape characteristics and cost values from Rabinowitz & Zeller (2010) derived from expert opinion (Table 2). Cost values ranged from 0 (no cost to jaguar movement) to 10 (a high cost for jaguar movement). The four layers were combined by adding them in Raster Calculator, and the output was reclassified so that all cells whose sums were above 25 (the average cumulative cost indicating a barrier to jaguar movement (Rabinowitz & Zeller, 2010)) received NoData values to represent a barrier in the surface (Esri, 2024).

To determine the optimal routes of travel across the cost surface, the ArcGIS Optimal Region Connections tool in Spatial Analyst was used to calculate the optimal connectivity network of least-cost paths between the eight habitat cores. This method uses a simple algorithm to calculate the cumulative cost of moving through the landscape: for any given movement from cell Ni to cell Ni+1, the cumulative cost is calculated as the cost to reach cell Ni plus the average cost to move through cell Ni and Ni+1 (Adriaensen et al., 2003). The least-cost paths layer was used as an input for the ArcGIS Pairwise Buffer tool to delineate potential movement corridors. While no empirical data exist on the width at which corridors fully lose their functionality, Rabinowitz & Zeller (2010) suggested corridors for jaguar to be at least 10 km in width at any point along their length to avoid becoming genetic bottlenecks. Least-cost paths were therefore buffered by 10 km to ensure sufficient width for jaguar movement, survival, and genetic exchange. Finally, the effects of the border wall were simulated by setting a buffer distance value of 5 km around the wall (ensuring a minimum gap width of 10 km) and defining it as an input barrier to the model, repeating least-cost path analysis, and comparing corridors with and without the wall. 

Results

The results from the least-cost path analysis represent potential areas that could be used by a dispersing male jaguar from Sonora into its historical range in Southern Arizona. The optimal connectivity network of least-cost paths predicted that with and without the border wall, all eight jaguar habitat cores are linked by dispersal. With the border wall, however, two predicted least-cost paths connecting habitat cores in Sonora to those in Southern Arizona are disrupted, leaving only one lowest cost neighboring path between an existing gap ~ 11.9 km wide in the east (Figure 2). Furthermore, a continuous border wall in the east would isolate habitat cores in Arizona because movement from Sonora was predicted only through areas to the north and northeast. Although an additional gap ~ 21.2 km wide exists in the west, most of its adjacent landscape has a cumulative cost above 25 and so this gap was avoided by the model. The total length of predicted movement corridors with the border wall is 775.5 km, whereas without a border wall the total length of predicted movement corridors is 734.9 km. The total area of all eight jaguar habitat cores is 1,439.4 km2, while the total area of the corridors connecting these cores with and without the wall is 8,221.5 km2 and 7,798.3.5 km2, respectively.

Discussion

The predicted movement behavior for the jaguar suggests that the border wall has immediate effects on dispersal and could significantly degrade interpopulation connectivity should its development continue. Although all eight habitat cores are linked by dispersal with and without the wall, comparison of the least-cost paths in both scenarios implies a weakening of the linkages among these cores with the wall. A continuous wall in the east would also isolate habitat in Arizona from the northernmost breeding population in Sonora. This would, in turn, impede natural movement of individual male dispersers and potentially accelerate the loss of genetic diversity. Furthermore, Peters et al. (2018) suggested that a continuous border-wide wall could fragment an estimated 34% of U.S. nonflying native terrestrial animal species. It is possible that the 11.9 km wide gap located in this study enables transboundary movement at a sufficient size for corridor implementation not only for the jaguar, but for other Neotropical and Nearctic taxa whose range limits converge here (Flesch et al., 2009). Detailed information on the distributions and movement behavior of these taxa are therefore required to fully assess the likely effects of border wall development and to implement effective mitigation strategies. 

While relevant data on feasible dispersal distances for jaguars are limited, Rabinowitz & Zeller (2010) cites the occasional jaguar traversing corridors ranging from 3 to 1607 km in length throughout their range. Lengths of movement corridors predicted in this analysis with and without the wall are within this range and are therefore assumed feasible. However, the total length of movement corridors with the wall increased by 40.6 km. Chambers et al. (2022) found that a surplus distance of only 3 to 9 km can increase the energy expenditure of an individual jaguar by nearly 43.5 kcal. As conservation planners are working on effective corridors linking populations that minimize energy expenditure (Chambers et al., 2022), the two un-walled least-cost paths from this study could provide a basis for corridor implementation. Finally, results in this analysis suggest that the existing 11.9 km wide gap best optimizes transboundary movement with current border infrastructure.

Conclusion

There is great importance in modelling least-cost paths for jaguars because this analysis allows for the identification of potential movement corridors, which is important to long-term management and planning for jaguar populations in the U.S.-Mexico borderlands (Rabinowitz & Zeller, 2010; Chambers et al., 2022). Conservation planners could use these results to implement effective mitigation strategies to include the provision of gaps should security infrastructure development continue. Although this study focused on physical barriers, additional factors such as associated lighting, vehicle traffic, and human activity may warrant consideration.