Mismatches between the fine spatial scales at which ecological processes operate and the coarser scales at which climate data are often available are increasingly recognized as a limitation to robust projections of climate change impacts on species. Recruitment has been implicated as a critical bottleneck in tree population dynamics, and may be decoupled from coarse-scale regional climate due to the buffering effect of the forest canopy. The goal of this study was to determine whether seedling recruitment patterns of two Pacific Northwest tree species are better explained by fine-scale, below-canopy climate than coarse-scale climate estimates across their ranges. We addressed this objective by utilizing a relatively novel climate dataset from a long-term network of below-canopy, plot-level climate stations located in Douglas-fir forests extending from the Oregon coast to the eastern slope of the Cascade Range in Oregon. Biologically-relevant climate variables including soil moisture and temperature and air moisture and temperature were evaluated against PRISM climate estimates in a series of hierarchical Bayesian Poisson regression models used to quantify the relationships between climate and seedling densities of two focal species – western hemlock and Douglas-fir. Model performance was compared across models calibrated on below-canopy climate data versus PRISM data.
Results/Conclusions
PRISM and plot-level climate data were moderately correlated at all sites, but climatic buffering by the forest canopy was apparent when comparing variability in both data sets. In general, below-canopy, plot-level climate variables showed reduced variability relative to analogous PRISM variables and demonstrated reduced temperature extremes and greater water availability. Correlations between temperature variables across the two datasets were stronger than for moisture variables. Preliminary model results indicate that below-canopy, plot-level climate is slightly more strongly associated with patterns of seedling density than PRISM climate estimates, but differences in model performance were minimal. The greatest differences in model performance were observed when soil temperature and moisture within the upper 15cm of soil were used to represent plot-level, below-canopy climate. Overall, these findings indicate that improvements in understanding climatic impacts by using fine-scale climate data may largely depend upon the biological relevance of available climate variables. Analogous variables at different spatial scales may produce qualitatively similar results, but inclusion of more proximal variables such as soil temperature and moisture may improve predictions of climatic impacts on seedling recruitment.