PS 28-138 - Influence of Delayed Tree Mortality on Post-Fire Landscapes in Oregon’s Western Cascades

Abstract: As wildfires across the western United States become increasingly frequent, severe, and large, changes in fire regimes may threaten long-term forest resilience as remnant patches of live trees post-fire, known as fire refugia, become scarce with limited successful forest regeneration following a wildfire. Post-fire delayed tree mortality (PFDTM) can substantially alter the forest extent, composition, and structure of fire refugia over time, limiting natural tree regeneration that in turn can promote forest conversions to alternative or non-forest states. Although PFDTM is well understood at the individual tree level, very little is known about the ecological patterns and processes of PFDTM across larger spatial scales. The objectives of this research are to develop and validate PFDTM estimations using moderate-resolution satellite imagery and to identify the spatio-temporal patterns and driving mechanisms of PFDTM at multiple landscape scales in the western Cascades. Pre- and post-fire Forest Inventory & Analysis (FIA) plot data along with high-resolution imagery are used to calibrate and validate remote sensing-derived metrics of PFDTM to estimate tree decline and/or mortality continuously across five megafire events from the 2020 Labor Day fires in Oregon’s western Cascades. With predictor variables that have been mechanistically linked to tree stress (i.e., topography, climate/weather, forest structure, and immediate fire effects), PFDTM is modeled using a Boosted Regression Tree (BRT) to predict basal area mortality post-fire for conifer tree species over time and analyze the magnitude, interactions, and flexible (i.e., non-linear) relationships between predictive variables and delayed mortality.

Preliminary findings support previous studies that the immediate fire effect is the strongest predictor of delayed tree mortality and that clumped, large trees experience higher delayed mortality rates than uniformly spaced trees due to increased competition. Additionally, analysis by Potential Vegetation Zones (PVZs) show that conifer species in Mountain Hemlock and Silver Fir PVZs experienced higher rates of delayed decline post-fire than species in the Douglas Fir PVZ, suggesting that higher elevation conifer species may be more susceptible to delayed post-fire effects. Such relationships can support managers for effectively controlling post-fire forest conditions, including to sustain refugia for habitat and to reduce post-fire hazards (e.g., hazardous tree removals). Given these observed patterns, PFDTM has the potential to critically limit seed source availability and consequently post-fire regeneration, however, further work is needed to identify thresholds and patterns of PFDTM that can initiate alternative or non-forest states.