COS 175-1 - Impact of seasonal-to-annual climate on post-fire regeneration of western larch

Abstract: Fire plays a critical role in forests of the western United States (US), but as wildfire and climate deviate from historical patterns, increasing fire activity may significantly alter forest ecosystems. To understand the impacts of changing climate and wildfire activity on conifer forests, we studied the impact of seasonal-to-annual climate on post-fire regeneration of western larch (Larix occidentalis), across the northwestern US.

We used a stratified random sampling design to identify 57 sites across the climatic range of western larch, within 32 fires that burned at moderate-to-high severity from 2000-2015. Each site was within 100-m of a seed source, to maximize the likelihood of post-fire regeneration. We destructively sampled 1651 seedlings (1-70/site) within variable-width belt transects (12-520 m²). In the lab, the precise year of germination for each seedling was determined using dendrochronological methods. Estimated germination years were used to calculate annual regeneration rates (no. seedlings ha−1·yr−1). We modeled the probability of regeneration in any year using boosted regression trees, defined by annual regeneration rates exceeding the 25th percentile across all years and sites (167 seedlings ha−1·yr−1). Predictors in the model included static variables – distance-to-seed-source, satellite-derived fire severity– and dynamic variables – time-since-fire and seasonal-to-annual post-fire climate metrics reflecting temperature and water availability during germination.

Of the 681 cumulative post-fire years sampled across sites, 45% included western larch regeneration. The majority of regeneration occurred at sites with north to north-east aspects, and 87% within five years after fire. These patterns are consistent with moisture requirements and shade-intolerant traits of western larch. Our statistical model had high predictive skill, with an AUC of 0.88 and accuracy of 0.72. Both static and dynamic variables helped explain post-fire regeneration. Time-since-fire, distance-to-seed-source, and maximum surface temperature during the summer all had negative relationships with regeneration probability, with 51%, 30%, and 8% relative influence, respectively. Whereas average minimum soil moisture during the summer and satellite-derived burn severity had positive relationships with regeneration probability, with 6% and 5% relative influence, respectively.

Our results suggest that climate warming will modestly lower the probability of western larch regeneration after wildfires. These impacts may be reduced by pre-fire management that increases mature tree survival – thus, increasing seed availability. In areas with a lack of post-fire seed sources, our results highlight the short window of opportunity for replanting efforts, within the first five years after fire, and suggest higher likelihood of survival on northerly aspects.