Fig 1.
Our study area encompassed central Idaho and west-central Montana surrounding the reintroduction (Alternative 1) and natural recovery (Alternative 2) recovery zones identified in the 2000 Final EIS (USFWS 2000), together identified as the Bitterroot Ecosystem (BE) for this study (top panel). We subset the BE boundary into 3 regions: BE North, Central, and South. The BE overlaps various national forests, and nearly all of BE Central and South are designated wilderness (bottom panel; see also Figs A1 –A3 in S1 Appendix).
Fig 2.
Our study represented the fourth phase of research aimed at providing science to help inform grizzly bear conservation. In the first phase, Sells et al. [20] developed and tested models for a focal population (the NCDE). In the second phase, Sells et al. [21] assessed model transferability by applying the models to nearby populations in the GYE, CYE, and SE. In the third phase, Sells et al. [22] applied the models to simulate connectivity pathways between populations in western Montana. In this present fourth phase, we applied the models to a recovery area without a current population, the BE.
Fig 3.
Maps of steps taken by simulated grizzly bears.
Simulations were initiated on the north and eastern edges of the study area for the natural recolonization scenarios and within the BE for the reintroduction scenarios.
Fig 4.
Maps of predicted habitat use for grizzly bears.
Results are shown for the natural recolonization and reintroduction scenarios, and both scenarios combined. Classes represent the quantile-binned relative habitat use values (1 = low, 10 = high), as summarized within the study area based on the number of steps taken per 300- x 300-m grid cell (Fig 3).
Fig 5.
Predicted sequence of habitat use by grizzly bears.
Results are shown for the natural recolonization (left panels), reintroduction (middle panels), and combined scenarios. Each map depicts a series of 5,000 steps from the full set of 20,000 steps taken by our simulated bears in each simulation iteration (i.e., Sequence 1 represents steps 1–5,000, Sequence 2 represents steps 5,001–10,000, etc.).
Fig 6.
Panel A: comparison of habitat values within the BE and other recovery ecosystems. White dots are median values, boxes are 50% interquartile ranges, and thin lines extend to the 95% values. Panel B: Mean values of habitat variables used by grizzly bears during simulations, measured within each class of predicted habitat use (1 = lowest, 10 = highest) within the BE study area. Where only one line is visible in Panel B’s 6th graph for building density, the recolonization results were identical to the reintroduction results (and thus behind that line); building densities were at or near 0 in the BE.
Fig 7.
Land ownerships used by grizzly bears.
The proportion of ownership or jurisdiction type in the study area is shown on the left side. On the right, the contribution of each type to classes of habitat use (1 = lowest, 10 = highest) by simulated grizzly bears, under recolonization (left), reintroduction (middle), and combined (right) scenarios.