Jaguar habitat connectivity, 540 m

A raster of Circuitscape output for habitat connectivity in western Mexico and the southwest United States. Data are provided at 540 m resolution.

We estimated resistance based on degree of human modification to identify pinch points where natural areas, regardless of other characteristics (i.e., canopy cover, topography) are most constrained by human influence. We rescaled degree of human modification (H) to range from 0 to 1, then applied a Power(10) transformation and converted to integer values using a multiplication factor of 10,000, as Circuitscape has difficulty solving networks with narrow ranges of values. We identified centroids of each 75th percentile core area as source nodes for Circuitscape models. Use of centroids as source nodes rather than complete core polygons not only speeds implementation, but also allows estimation of current flow within core areas rather than treating them as uniformly zero-resistance patches (e.g., Dickson et al. 2016). Designation of S75 cores resulted in delineation of one very large core in the center of the study area. In the case of this core, we ‘shrunk’ core edges inward by 900 m to separate it into two discrete patches at its narrowest point, then identified the centroid of each of the two resulting patches. Use of multiple points to represent the north and south portions of this large, central core resulted in better distribution of current flow throughout its area.

We implemented Circuitscape in one-to-all mode, in which one node at a time is charged with current and all other nodes are grounded, iterating and summing over all nodes (McRae et al. 2013). This mode allows estimation of ecological flow from each potential source to all potential destinations in the landscape, which we suggest is most realistic for representing natal dispersal processes. Circuitscape is computationally intensive, and speed and memory requirements increase exponentially with number of raster cells in the landscape and number of nodes connected. In order to achieve reasonable computation time (days rather than months), we reduced the analysis resolution to 540-m and limited connections among nodes to those < 100 km apart.

For additional information, please see the project report, which can be downloaded at http://www.csp-inc.org/wp-content/uploads/2017/04/CSP_mapping-jaguar-habitat.pdf