Fire-driven tree mortality is often highest within the first three years post-fire. The combined threat of drought and fire might elevate post-fire mortality beyond what is currently predicted from frequently used models such as the Ryan-Amman. California has experienced severe multi-year droughts in the recent past, particularly 2012-2015. Consequently, there is a need for readily obtained drought metrics to improve delayed post-fire mortality predictions. For this study, we used 97 Forest Inventory and Analysis (FIA) plots visited three times: pre-fire, 1-year post-fire, and 4 to 9 years post-fire. The majority of plots experienced fire in 2007, with a range from 2002 to 2008. Field crews assessed crown scorch and mineral soil char 1- year post-fire. Diameter and mortality status were measured at all 3 visits. The Palmer Drought Severity Index (PDSI) was obtained for each plot for a period covering 10 years before the 2nd post-fire plot assessment. We selected six tree species with at least 200 trees for analysis: canyon live oak, Douglas-fir, Jeffery pine, incense cedar, ponderosa pine, and white fir. Logistic regression was used to assess if PDSI, crown scorch, bark thickness, and Ryan soil char rating (low, medium, and high) would be useful for predicting mortality.
Results/Conclusions
Delayed mortality of trees that were alive at 1-year post-fire ranged from 30% in Jeffery pine to 10% Douglas-fir. For incense cedar, ponderosa pine, and white fir, PDSI was a significant predictor of mortality. PDSI had the greatest effect on ponderosa mortality. Increasing PDSI (less drought) resulted in a maximum 10% decrease in mortality odds. Crown scorch had a significant effect for six species, but the effect was more modest than PSDI. Mineral soil char was a significant predictor only for incense cedar. Bark thickness was not a significant factor. The lack of across the board drought effects highlights the difficultly of generalizing mortality trends even under a severe drought. Our results support of the inclusion of PDSI into post-fire mortality prediction models in California.