Climate-change induced forest dieback has been documented as a recent widespread phenomenon, with pervasive mortality having profound impacts on ecosystem services and natural forest functioning. The recent prolonged drought and scale of tree die-off in California (>100 million trees from 2010-2016) are unprecedented in modern history. Tree death of this magnitude can transform regional landscapes and have severe effects on forest function and associated ecosystem services. Ongoing drought and predicted alteration of precipitation patterns due to climate change are likely to lead to continued mortality and loss of species, which is most precarious for rare, endemic species, filling narrow yet essential ecosystem roles. Monterey pine (Pinus radiata), an endemic to coastal forests of California and Baja Mexico, is susceptible to the non-native pine pitch canker pathogen (Fusarium circinatum) and recently experienced widespread mortality following chronic drought stress. We used a long-term (14 year), permanent plot network within this drought-impacted endemic species to evaluate 1) the relative importance of exogenous vs. endogenous factors in shaping forest demography, 2) the role of precipitation in the direction and magnitude of population dynamics, and 3) the predicted impact of climate change on species persistence.
Antecedent precipitation was identified as the primary driver of both adult mortality and regeneration (zero-inflated negative binomial model, p<0.001) of Monterey pine. A threshold of <400mm/year led to dramatic increases in mortality and decreases in regeneration the following year. Monthly fog cover was significant in the mortality zero-inflated model, indicating that an increase in fog decreases the probability of mortality (p<0.05). Recruitment was additionally explained by the frost-free period (days, p<0.05), with regeneration decreasing with increases in length (>300 days). Forest density did not influence demographic processes (p>0.05), but pine pitch canker incidence was a significant driver of mortality (p<0.001). Sampling locations were geographically close so not included as predictors; location was, however, significant in mortality (X2=276, p<0.05) but not significance in recruitment (X2=97, p>0.10). Tree class (seedling, sapling, mature) did not have a significant relationship with mortality (X2=136, p>0.05). The prevalence of disease significantly influenced demographic patterns over time, but, in all models, regional climatic variation was the primary determinant of mortality and recruitment in this endemic conifer. Climate projections (RCP85) predict a decrease in total annual precipitation (<400mm) and increase in the frost-free period (>330 days) for this region, suggesting continued mortality and low recruitment into the future.