Decrease in growth preceding death in Abies amabilis (Pacific Silver Fir)

Background/Question/Methods

Old-growth forests in the Pacific Northwest are important because of their ability to sequester carbon. A third of carbon in this biome is bound up in trees, thus understanding how climate influences tree mortality is critical to predicting how old-growth forest ecosystems will respond to climate change. Moreover, the link between climate and tree mortality is also important for forecasting range shifts, as range contractions at lower and southern range limits only occurs when all trees within those populations have died.  Recent studies document increasing tree mortality in the western US over the last two decades, but it remains unclear whether recent warming is responsible. To address these questions, we studied tree mortality at Mt. Rainier National Park, Washington, USA, which harbors extensive old-growth forests. We investigated the relationship between growth and mortality in Pacific silver fir (Abies amabilis) at Mt. Rainier at three different elevations (ranging from 1000 to 1600 m), spanning the altitudinal range of this species. Increment cores were collected and analyzed from recently deceased individuals as well as living individuals of comparable sizes at the same elevation. From the increment cores, we quantified yearly growth in each individual tree. 

Results/Conclusions

We found significant differences between average yearly growth in live trees and dead trees that extended several years previously, suggesting long periods of reduced vigor prior to death. Differences between growth of living and recently dead trees were greatest at 1500 m (n=35, p=0.0054) and 1600 m (n=44, p<<.001), and least at 1000 m (n=44, p=0.029) across all size classes. However, when a deceased and a living individual of comparable size were paired, the growth immediately preceding death differed significantly from the growth observed in the healthy, living individual at 1600 m (p<<0.001; df=22) and 1500 m (p=.001; df=18), but not at 1000 m (p= 0.13; df=11). This suggests that trees at their upper range limit die more slowly, perhaps more influenced by environmental factors than those at lower elevations. In contrast, sudden mortality events (such as disease or disturbance) may be more important at lower range limits. In total, our results suggest that mortality responses to climate change may be difficult to predict.