Predicting mortality of Pinus edulis during drought using growth and resin duct anatomy

Background/Question/Methods

Drought and insects such as bark beetles frequently interact to produce widespread tree mortality in forest ecosystems.  However, our ability to project future forest dynamics is strongly limited because the processes and thresholds leading to tree death are poorly constrained.  Tree allocation to defense may be a key element in tree avoidance of mortality, especially in semi-arid conifer forests of western North America, but there are few empirical tests of its importance in predisposing trees to die. Here we document properties of vertical resin ducts in tree rings of piñon pine (Pinus edulis), and investigate relationships between tree growth, allocation to constitutive resin ducts, and mortality. We generated radial growth and resin duct time series for 102 pairs of trees that died and survived the 2000s and 1950s droughts at four sites spanning a latitudinal gradient in New Mexico, USA. We used logistic regression to generate statistical models of mortality risk from growth indices, resin duct parameters, and combinations of the two types of metrics.

Results/Conclusions

• Recent (3-20 year) average radial growth was generally higher in trees that survived drought, and mean sensitivity – a metric of year-to-year growth variability – was lower.  However, this result was inconsistent across space and time, with differences between trees that lived and died either marginally significant or not significant overall, depending on the growth metric.
• In contrast, recent (3-year) resin duct number, average size, and area ratio (the ratio of resin duct area to xylem area) were significantly higher in trees that survived, with ~60% larger resin ducts in surviving trees.
• Models calibrated using both growth and resin duct variables had substantially more support than models based on growth or resin duct parameters alone, with correct classification rates for top models developed for all sites exceeding 80% (compared to 70% for growth-based models). 
• Allocation to resin duct defenses appears to have been more important for tree survival during the relatively warm 2000s drought that may have had amplified bark beetle pressure.
• Our results point to the importance of further investigations of tree allocation to defense in order to improve understanding of tree mortality processes, and highlight the promise of using simple correlative metrics of growth and defense to improve predictions of tree mortality.