Characterizing the differential sensitivity of tree species and forest types to past weather variability using dendrochronological techniques

Background/Question/Methods

Tree biomass growth results from physiological processes that are sensitive to annual variations in temperature, precipitation and potential evapotranspiration. Dendrochronological analyses aimed at past climate reconstructions often employ standardization techniques that magnify the signal of tree growth variance that is correlated to climate. This approach tends to factor out variance in growth caused by characteristics such as tree size, age, species, competitive status, and stand density. However, the relative variance in growth attributed to climate fluctuations vs. forest structural characteristics is rarely quantified. We asked (a.) how traditional dendrochronological techniques can help predict forest biomass growth, and (b.) whether the variance in growth due to weather variability is as important as other factors.

We reconstructed tree, species and stand biomass increments using a census dataset of tree-rings from eight forest communities covering the temperate and boreal forests in northern Minnesota. We used mixed models to predict annual biomass growth from tree size and age, stand density and aggregate growth trends, species and local site characteristics, and variations in annual precipitation, temperature and summer moisture deficit using PRISM climate data (1895-2009). We used dendrochronological response function analysis to guide climate variable selection.

Results/Conclusions

We found that mean tree biomass growth depended most strongly (and positively) on tree size, age, and competitive status, while weather variability typically explained less than 5% of total variance in growth. Growth relationships varied significantly by species, with higher than average growth seen in species such as Acer saccharum and Populus tremuloides and lower than average growth in Fraxinus nigra and Thuja occidentalis. Variations in biomass growth around the mean were significantly related to the summer ratio of precipitation to PET, though growth often lagged summer conditions by one to two years depending on species and site. Variables selected by response function analysis did not always perform better than standard seasonal climate variables. While weather fluctuations significantly affect growth, our results suggest that changes in forest composition and structure are more important to predict short-term productivity responses to climate change.