Individual tree growth in structurally complex forests may vary considerably with tree size, species interactions, and resource heterogeneity. Moreover, whole ecosystem productivity comprises individual growth, which suggests that local tree interactions could influence ecosystem level productivity and carbon balance. Our objective was to predict individual tree growth as a function of focal tree diameter, the size and proximity of potentially interacting trees (local neighborhoods), and individual-scale soil resource availability. We expected that growth rate should: increase with focal tree diameter, be positively related to local soil resource availability, and be negatively related to the size and proximity of neighboring trees. Growth was modeled for the 10 most common tree species from 13 sites established across strong soil resource gradients in northwest lower Michigan, USA. We tested several neighborhood index and resource model formulations to test how these factors influenced individual tree growth. Tree size and neighborhood effects are often confounded in mature forests, so we also tested how size, neighborhood, and soil resources reduced growth from an extrapolated potential rate based on our data and USFS Forest Inventory Analysis data for the region.
Results/Conclusions
In partial support of our hypotheses, growth was positively associated with focal tree diameter in all species, and explained 62 - 92% growth variance. The correlations of growth to neighborhood effects and soil resources were weak, however, and explained only a maximum of 5% growth variance. For four species, neighborhood indices that recognized species identity provided the best fit to the data, suggesting a competitive hierarchy; intraspecific interactions consistently reduced growth to a greater extent than interspecific interactions for these four species. When growth was assessed as a function of extrapolated potential growth, results were similar, although neighborhood and soil resources explained not more than 4% growth variance. Overall, these results supported relatively weak relationships of neighborhood interactions and soil resources to individual growth. In contrast, site-level productivity in these forests was strongly related to soil resources. Differences between individual- and ecosystem-scale patterns likely arose from shifts in species composition. Most of the variation in growth for a given species was constrained to sites with similar resource and productivity levels, whereas the two-fold change in productivity across all sites spanned the entire resource availability gradient. These results emphasize the importance of species replacement in understanding productivity changes across resource gradients.
