Bark thickness is a key factor determining tree injury or mortality from fire, especially surface fire. It is used as an independent variable to predict post-fire tree mortality in the two most widely used US fire effect prediction systems: Forest Vegetation Simulator – Fire and Fuel Extension (FVS-FFE) and the First Order Fire Effects Model (FOFEM). Both of these tools are actively used by forest managers to promote and restore ecological function through prescribed fire, but species within the same habitat may respond differently to the effects of fire. Since bark thickness changes with stem size, both FVS-FFE and FOFEM estimate bark as a species-specific proportion to stem size and both tools assume bark thickness follows a simple linear allometric relationship across stem sizes. However, the accuracy of these bark thickness estimates has never been independently tested. Our objective was to evaluate the accuracy of model estimates of bark thickness across stem sizes for various tree species. To do this, we utilized a large dataset consisting of bark thickness and diameter measurements of 55 tree species native to the United States. We developed species-specific non-linear models and compared our best-fit models to the linear models in FVS-FEE and FOFEM.
Results/Conclusions
Our results indicate the accuracy of the estimates vary greatly across species, suggesting that the bark thickness and stem size relationship assume different forms depending on species and diameter at breast height. Instead of applying the same linear relationship across all species, we incorporated varied best fitted model forms to better reflect the relationship between diameter and bark thickness for each species. We found that our models had a reduced RMSE and increased R2 in comparison to the existent models in FVS-FFE and FOFEM. We recommend that these newly developed bark thickness equations be implemented into both FVS-FFE and FOFEM so that post-fire tree mortality can be better predicted.