Size-, species-, and site-specific tree growth responses to climate variability in old-growth subalpine forests

Tree-ring data have become widely used to model tree growth responses to climate variability and gain insight about the potential effects of global warming on forests. We capitalized on a rare opportunity to develop growth-climate models using tree-ring data collected from all trees (>4 cm in diameter at breast height) within 50 x 50 m plots established in subalpine old-growth forests of western Canada. Our objective was to determine how tree growth responses to climate vary among tree size classes, species, and sites. We modeled relationships between times series of annual basal area increment (Delta BA) and yearly climate variables for individual trees; this approach obviated key statistical criticisms of "traditional" tree-ring analysis methods. Time series of annual basal area increment were detrended a priori for size, age, legacy, and competition effects. We found that the overall climate signal in our time series of Delta BA was weak; <6% of the interannual variance was explained by climate variables. Nevertheless, there were clear patterns in climate-growth relationships related to tree size and species. Relationships between Delta BA and five climate variables increased in strength with tree size class; large trees were most sensitive to annual climate fluctuations and accounted for similar to 71% of the overall climate effect on growth across all trees and sites. In all stands, Delta BA variance explained by climate variables was stable over the 20th century for large trees but decreased in the 1940s for small trees, indicating a temporal reduction in sensitivity to annual fluctuations in five climate variables. In coastal forests, relationships between Delta BA and climate for Callitropsis nootkatensis were significantly different in direction and magnitude than those of co-occurring Pinaceae species. The effect of climate on tree growth was idiosyncratic among stands and could not be discriminated by forest type (coastal vs. interior). Our individual-tree modeling approach adds to a growing body of research providing novel insights about the complexities of tree growth responses to climate variability and the challenges associated with predicting future tree growth and forest productivity using tree-ring data.