COS 262-4 - Integrated population models for focal wildlife species inform landscape-scale forest restoration

Abstract: In the USA, landscape-scale forest restoration on public lands aims to create more resilient ecosystems by reintroducing processes (e.g., fire) altered by human activity. These programs track the outcomes of restoration using a variety of methods, including population trends of focal wildlife species that serve as one proxy for forest ecosystem recovery. However, these programs are just beginning to compile enough data to assess long-term wildlife population trends. The goals of this project are to 1) determine which life stages of a focal bird species are most impacted by forest restoration and 2) understand how these stage-specific responses shape the focal species’ population trends. We evaluate the effect of three landscape-scale restoration projects located in Oregon and Idaho, USA on the focal species, white-headed woodpecker (Dryobates albolarvatus). We used a two-step approach to generate an integrated population model (IPM) that relies on the combined information from 10 years of data that follow multiple life stages, accompanied by data on mediating landscape conditions. Specifically, we used nest tracking data to determine the number of eggs produced by adults, egg survival, and nestling survival, combined with N-occupancy modeling of adult abundance. The first step involved fitting Bayesian statistical models to data on each life stage individually to estimate the effects of restoration and habitat covariates on each stage. Then, we used the posterior estimates of these variables to propagate uncertainty in those parameters in the IPM, which combined all the nest stages and adult occupancy into one model. We found that egg and nestling survival are influenced by forest restoration, with stage-specific differences. Egg and nestling survival increased in sites that experienced local (0.4 ha) forest restoration of either prescribed burning or tree harvesting (eggs: β = 1.46, p-value = 0.004; nestlings: β = 2.5, p-value < 0.001). However, as the percent of restoration treatment on the landscape (100 ha) surrounding nests increased (“restoration extent” increased), survival decreased (eggs: = β -0.54, p-value = 0.03; nestlings: β = -1.01, p-value = 0.001). This study provides evidence of the efficacy of landscape-scale restoration, and it identifies life stages that are sensitive to past and current management decisions as well as how these sensitivities shape population trends. More broadly, this study provides a model for data analysis and integration for landscape-scale restoration efforts that rely on focal species to evaluate forest restoration outcomes.