Landscape heterogeneity compensates for fuel reduction treatment effects on Northern flying squirrel populations

Abstract

In the dry forests of the western United States frequent fires historically maintained a diversity of habitats in multiple seral stages. Over the past century, fire suppression and preferential harvest of large trees has led to a densification and homogenization of forests, making them more prone to larger and more severe wildfires. In response, fuel reduction treatments have become common practice in the management of dry western forests. However, the effect of fuel reduction treatments on late seral forest species, such as the Northern flying squirrels, remains a management concern.We captured and marked Northern flying squirrels within mixed conifer forest in the Stanislaus–Tuolumne Experimental Forest (California) on a continuous trapping grid (∼1400 traps) spanning a 120-ha study landscape in which 24 4-ha units were subject to different fuel reduction treatments (variable thin, even thin, and control, all with or without prescribed burning). The study spanned two pre-thinning and three post-thinning years. We divided the study landscape into three blocks (two with treatments, one control only). For each block we analyzed data with spatial capture–recapture models to estimate Northern flying squirrel density, and tested whether canopy closure before and after thinning and percent area burned were important predictors of density.Northern flying squirrel densities varied from 0.168 (SE 0.086) to 0.808 (SE 0.094) individuals/ha across blocks and years. Densities varied by year, independent of treatments. Percent area burned was not an important predictor of density. The effect of canopy closure was variable, but more consistently positive after thinning reduced overall canopy closure. When considered by treatment type, densities were highest in control and burn only units, and lowest in thinned units.Whereas thinning had negative effects on Northern flying squirrel density on the scale of a thinning treatment unit, our results suggest that these effects were largely absorbed by the heterogeneous landscape, as animals shifted their distribution into un-thinned areas without a decline in overall density. This highlights the need to incorporate the landscape context when evaluating the effects of forest management on wildlife.