Death comes for every tree, but timing and location have important implications for the persistence of forest ecosystems following major disturbance. Predictions of post-fire tree mortality are typically based on observed relationships to bark thickness and degree of crown scorch. However, models (FFE-FVS e.g.) often fail to predict delayed mortality because models were developed using a limited range of tree characteristics and fire weather, and typically omit important factors (e.g., fire effects on soil and boles.) that influence delayed post-fire mortality.
The Fire Effects and Recovery Study (FERS) seeks to improve mortality predictions by leveraging US Forest Inventory and Analysis (FIA) inventories with additional assessments of fire effects on tree crowns, boles, and ground surface substrates. FIA plots provide a probabilistic sample of forest lands that experienced large wildfires in the west coast states. FERS has pre- and post-fire (1-2 years) observations from 840 FIA plots representing 130 fires that occurred between 2003 and 2016. We developed a structural equation model (SEM) to test a-priori relationships between fire effects on both forest soils and tree crowns on delayed tree mortality, using FERS data. Our preliminary model relates a soil variable (% cover, by char class, on mineral soil substrate) to crown fire effects (% scorched, burned or unburned) from 102 California fires from 2003-2015.
Results/Conclusions
The strongest positive driver was % deeply charred mineral soil cover on tree crown scorch: 0.08 and (se 0.01, p-value < 0.05). In contrast, % deep char was a negative driver of tree crown consumed (burned to black): -0.067 (0.014, p-value < 0.05). Increased crown scorch was associated with greater cover of mineral soil in more severe char classes. In contrast, the greater the proportion of crowns in a stand that were burned, the lesser was the impact of the fire on mineral soil. These findings comport with previous research showing that ground surface fire effects do not necessarily track effects on tree crowns. Results are being applied to improve the predictive power of fire effects models to include both immediate and delayed post-fire mortality by linking fire effects at the ground surface, in tree crowns and tree mortality. Our SEM framework advances post-fire management by improving our knowledge and modelling of how the interactions between soil and tree drivers of mortality influence prospects for post-fire recovery of forested ecosystems.