2020 ESA Annual Meeting (August 3 - 6)

PS 13 Abstract - Vegetation dynamics in frequently burned mixed conifer forests, northern Sierra Nevada Range, California

Asha Paudel1, Scott H. Markwith1, Michelle Coppoletta2 and Kyle Merriam3, (1)Geosciences, Florida Atlantic University, Boca Raton, FL, (2)US Forest Service, (3)US Forest Service, Plumas, CA
Background/Question/Methods

Dry mixed conifer forests in the Sierra Nevada range, CA, are increasingly facing threats from frequent mixed-severity reburns exacerbated by massive fuel accumulation and increasing drought and high temperature frequencies. Shrub dominance after high severity fire and standing dead snags have been presumed to be great problems for achieving the targeted historical heterogeneous forest type. Monitoring understory and overstory vegetation interactions with each other, their environment, and ecological processes is important before application of any management actions. To address this importance our question was, how do the understory and overstory vegetation pattern at the plot-level conifer forests change from pre- to post- reburn samples after second mixed-severity fire compared to single mixed-severity fire and inburned plots? We compared density changes in seedlings, saplings, and trees, and percent cover change in shrub among once- and twice-burned and unburned plots in response to fire severity and various biophysical explanatory variables. A total of 134 circular plots were resampled using the USDA Forest Service common stand exam (CSE) protocol. Random Forest machine learning technique was used to identify the most important variables and the taking the most important variables the relationship with the dependent variables were tested for statistical significance using non-parametric regression analysis followed by margin analysis in stata.

Results/Conclusions

The Random Forest machine learning technique showed that pre-reburn shrub cover, as well as pre-reburn density of live mixed conifer trees and saplings were important explanatory variables for determining post-reburn shrub cover and density of trees/saplings. Post-reburn seedling density was influenced by elevation and post-reburn shrub cover, with a negligible effect of reburn fire severity. Using the most important variables from Random Forest, non-parametric regression analysis was conducted, followed by a margins analysis in Stata, which showed that there is a significant effect of low to moderate reburn severity on post reburn mixed conifer tree density when the pre-reburn forest has <100 trees per ha. Based on the margins analysis, post-reburn regeneration was negatively affected by moderate severity reburns at higher elevation (2000 m) when shrub cover <50 %. Low to moderate severity has significant negative effect on post-reburn regeneration at a combined elevation of 1500 m and shrub cover >50 %. Based on this we can conclude that natural regeneration of mixed conifers can vary depending on the combination of shrub dominance and elevation, and even low tree density forests may suffer significant density reductions in low to moderate severity fire.