Forests play a major role in the terrestrial biosphere, covering only 30% of the global land area but accounting for over 50% of the carbon sequestration capacity. Modeled projections of forest function under future climate scenarios suggest that forests will need to be more efficient in their water use to maintain current levels of carbon sequestration. However, the results are often parameterized with data collected primarily from the uppermost canopy layers or leaf-level measurements that might not adequately capture the myriad of interactions that affect tree growth. This can skew our understanding of local and regional forest response and affect the accuracy of forest growth projections in regions with complex canopy structure, such as the Northeastern United States. We hypothesize that canopy position has a significant effect on the climate-growth relationships recorded radial growth, resulting in unique and distinct relationships among canopy layers. We used generalized additive models to analyze the responses of annual growth rings to climate across multiple canopy strata from four abundant Northeastern US species (Fagus grandifolia, Tsuga canadensis, Quercus rubra, and Acer rubrum). We compared the effects of species and canopy position in moderating climate-growth relationships on the growth of 1085 trees over the past century.
Results/Conclusions
We constructed four models (null, species-only, canopy-only, and species x canopy) to assess the influence of species and canopy position on the tree climate responses. Although all models explained similar amounts of variance within the dataset (exp. var. ≈ 0.52), the model accounting for canopy class differences within a species produced the most parsimonious model compared with canopy-only and species-only models, according to the Akaie Information Criteria. Climate sensitivity differed among canopy layers, and greatest differences occurred at both extremes of temperature and precipitation conditions. After having accounted for the size effect, understory trees had a more dynamic climate response than their overstory counterparts and showed increased growth at mild and moderate temperature and precipitation levels. Dominant trees showed an increased advantage during periods of increased temperature and precipitation. These results suggest that tree response to climate conditions is not synchronous across all canopy levels and reveals a need to increase sampling efforts to ensure representation of the structure of the entire forest. Doing so will provide a more accurate representation of forest response to climate change and variability than is currently being incorporated into global ecosystem model frameworks.