Snow-cover remote sensing of conifer tree recovery in high-severity burn patches

Abstract

The number of large, high-severity wildfires has been increasing across the western United States over the last several decades. It is not fully understood how changes in the frequency of large, severe wildfires may impact the resilience of conifer forests, due to alterations in regeneration success or failure. Our research investigates 30 years of conifer recovery patterns within 34 high-severity wildfire complexes (1988–1991) of the Northern Rocky Mountains. We evaluate the capability of snow-cover Landsat to characterize conifer tree recolonization of high-severity burn patches. Snow-cover images isolate conifer-specific vegetation signals by diminishing spectral contributions from soil and deciduous vegetation. The presence of conifer regeneration was successfully classified by snow-cover Landsat at >10% canopy cover at 98% accuracy and modeled at 3-year intervals post-fire. Spectral detectability of regenerating conifer vegetation began 11–19 years post-fire, varying across forest types. Thirty years post-fire, 65% of the total high-severity burn area had been recolonized by conifer trees, with differences observed between forest types: 72% of lodgepole pine, 77% of Douglas-fir, and 44% of fir-spruce severely burned areas containing conifer regeneration. Projected recovery timelines to pre-fire conifer vegetation also differed between lodgepole pine (29.5 years), Douglas-fir (36.9 years), and fir-spruce forests (48.7 years), as estimated from snow-cover NDVI trends. Although we generally documented patterns of conifer resilience, we also identified reduced likelihoods of recovery within high-severity burn patches exhibiting greater area-to-perimeter ratios, aridity, south-facing aspects, slopes, and elevation. Snow-cover Landsat imagery was shown to improve the characterization of post-fire forest recovery and may be applied to support forest restoration decision-making following high-severity wildfire.