Evaluation of cost analyses

While Rabinowitz et al. [20] used the lowest 0.10% of grid values (or cell values) in the LCC as the cutoff for area to be included in determining corridors in a wide-range model for jaguars, this level was determined to be insufficient at the small scale in this study because it prevented connectivity with the largest protected area in the central region, Reserva de Biósfera Yabotí/PP Esmeralda. For example, among the five species, the oncilla and bush dog achieved complete connectivity with this area when the cutoff was set to 0.3% and 0.4%, respectively. When the cutoff was expanded to include the lowest 0.15% and 0.20% of grid values, connectivity with this area was achieved; specifically, via a northern (0.15% and 0.20%) and northwestern (0.20%) connection. However, this also resulted in an expansion in the breadth of area included in the corridor; therefore, this study used the LCP and home range information to determine the optimal balance between corridor length and width [10,11,19]. We aimed to make the species-specific corridors wide enough to support sufficient territories for these wide-ranging species [11], while keeping the total area at a level that could be feasibly implemented. We allowed flexibility in the species-specific width of the corridor, however, aimed to maintain a minimum width of 14 km around the LCP, values that mimicked the maximum potential home range values were used to establish predictor value grids (Table 1).

With an overlay of the five species-specific LCP-LCC models, it was possible to narrow the overall width of the corridor and define two multispecies corridors that could be used to set conservation priorities: a primary (1°; 7 km) and a secondary (2°; 14 km). Together, both multispecies corridors range between 1–3× the minimal home range (50 km²) of a jaguar, puma, and bush dog, with some areas in the combined 1°- 2° corridor extending beyond 3× the minimal home range.

The quality of area in these multispecies corridors was evaluated by quantifying the overlap with the species-specific ENMs, PSR, and degree of habitat fragmentation/modification. In addition, we quantified the overlap of habitat type with PSR level. Areas that crossed through protected areas (i.e., currently managed) were eliminated, so final values reflect land needing priority management. Together these data identified areas within the 1° and 2° corridors that need protection of current habitat and those that need habitat restoration.