Natural tree regeneration dynamics a decade after the Storrie Fire in the Lassen National Forest

Large scale, high-severity fires are increasing in the western United States. Despite this trend, there have been few studies investigating post-fire tree regeneration, even fewer have focused on tree regeneration after large wildfires in the southern Cascades-northern Sierra Nevada. We established a study in the 2000 Storrie Fire, a 23,000 ha wildfire that occurred in the Lassen and Plumas National Forests in northern California. We used a stratified sampling design to quantify post-fire vegetation dynamics across four levels of burn-severity (unchanged, low-severity, medium-severity, high-severity) and three vegetation types (mixed conifer, low-elevation fir, high-elevation fir) on the Lassen during the summers of 2009 and 2010. Conifer seedlings were measured within five replicates of clustered 3.59 m radius plots. Within each plot we recorded biotic and abiotic factors that we hypothesized to influence seedling growth and establishment. We found abundant mature conifer seedlings across much of the landscape. Oaks were prevalent in the lower elevation strata and responded to fire with heavy sprouting. Median conifer seedling densities varied substantially by burn severity, ranging from 1,918 seedlings ha-1 in the unchanged units; 4,838 seedlings ha-1 in the low-severity units; 6,484 seedlings ha-1 in the medium-severity units; and 710 seedlings ha-1 in the high-severity units. The majority (80%) of high-severity units sampled were "understocked" according to regional density thresholds. We fitted a Zero-Inflated Negative Binomial (ZINB) linear model to determine significant predictors of mature conifer seedling density. The developed model best accounted for the high level of residual variance inherent to seedling densities across the landscape. In addition to the strata blocking (burn severity × vegetation type), significant positive predictors included the topographic index (a transform of slope and aspect), slope position (lower-, mid- and upper-slope), available growing space (a transform of free above-ground growing space and shrub cover), average shrub height and the stand perimeter-to-area ratio. A linear model was also developed to better understand seedling stocking across the landscape. Interpretation of model coefficients is useful in aiding land management decisions and better understanding regeneration dynamics in response to large wildfires that are increasingly common in fire-prone western landscapes.