Recent and large-scale tree die-off in the southwestern U.S. highlights the importance of understanding the underlying processes of tree mortality in order to predict the impacts of climate variability and change on forest ecosystems. The severe droughts and associated pulses of tree mortality in the Southwest during the 1950’s and the early 2000’s provide a rich source of material for evaluating and improving our knowledge of tree mortality close to arid treeline. In this study, we examine the relationship between diameter growth and mortality for pinyon pine (Pinus edulis) to develop statistical models of mortality risk through different generations of trees and across the species’ extensive range. Our sampling design combines tree-ring records of growth from archival material with new tree-ring samples across a latitudinal gradient in New Mexico. This novel approach allows us to compare the growth histories of dead vs. surviving trees throughout the region and during two distinct episodes of drought-induced tree mortality. Using this dataset, we address three key questions: (i) can growth-based statistical models reliably capture mortality risk across gradients of climate and severity in stand-level mortality? (ii) can we consistently predict mortality for trees living now and trees that lived and died in the past? (iii) are growth-based mortality models for semi-arid species improved by incorporating the influence of competition and stand density on growth history?
Results/Conclusions
Preliminary results show that growth before death is highly variable between and within sites, with some trees dying after a quick reduction in growth, some succumbing after long declines, while others are able to withstand decades of very low growth followed by recovery as environmental conditions improve. The effect of neighboring trees on mortality is significant at some sites, pointing to the importance of density-dependent processes such as competition and insect outbreaks in some but not all cases. Our findings highlight the complexity as well as the potential of modeling mortality in pinyon pine using annual growth as a key predictor. Ongoing work is quantifying differences in growth between living and dead trees across sites, and will identify the best growth-related predictors for regression modeling. Results from these models for the region and for individual sites will be presented.
