The 2012-2016 drought and associated wildfire and bark beetle herbivory killed on the order of 100 million trees in California. This drought offers a useful window into the region’s near future, since similar droughts are projected to occur more frequently as precipitation becomes more variable and temperatures increase. Accordingly, what we can learn from this event about variation in forest vulnerability to drought-associated disturbance has value for projecting these forests’ future spatial patterns and composition. In this study, we ask whether Sierra Nevada pine and mixed conifer forest drought vulnerability can be predicted from pre-drought fluctuations in forest green-ness (EVI). We then ask whether this severe drought altered tree recruitment patterns in severely fire-disturbed areas, potentially affecting the future trajectories of forest communities.
We analyzed time series of EVI data to quantify and map pre-drought fluctuations in forest green-ness. To assess drought effects on regeneration, we surveyed plots in which fire disturbance was followed by wet years, and another set of plots in which fire occurred shortly before the 2012-2016 drought (525 plots total). We counted seedlings of trees and shrubs and evaluated distance to nearest seed source. We tested associations between post-fire weather and recruitment using mixed models.
Results/Conclusions
Higher drought resistance in forest stands was weakly predictable from the time series of EVI from “normal” periods before the onset of drought, while overall greenness, likely associated with greater tree biomass and/or density, was a stronger predictor of mortality rates (R2 = 0.15). This result is consistent with findings that pre-drought growth patterns of individual trees is only sometimes predictive of subsequent mortality risk, and with findings that higher forest density is associated with greater vulnerability to drought, particularly when higher density occurs in drier areas.
Analysis of regeneration patterns revealed that tree species recruitment patterns were not strongly associated with drought severity. The strongest predictors of post-fire seedling abundance were not drought intensity following fire (variance explained = 4%), but rather local climatic conditions and local adult tree community (variance explained = 57%). This result suggests biological inertia in forest composition that may slow the adjustment of forest to future climate. Post-fire drought was, however, associated with relatively stronger recruitment of shrubs and grasses (50% cover increase) and weaker recruitment of some tree species (decline up to 50% for white fir), suggesting a potential biome shift or retraction of forest in a drier future.