Understanding demographic controls of aboveground forest carbon at a continental scale
Future changes in the global terrestrial carbon sink remain the largest uncertainty in predictive models of the Earth’s coupled carbon-climate system. Terrestrial vegetation (particularly forests) is a critical carbon reservoir, but we have limited understanding of how geographic variation in carbon storage capacity arises from continental- to global-scale differences in growth (productivity) and mortality (turnover). In forest ecosystems, demographic performance varies both among tree species and across broad environmental gradients. Unfortunately, existing terrestrial biosphere models represent plant functional diversity in a simplistic manner that may not capture how carbon storage relates to interspecific variation in tree demography. Also, terrestrial biosphere models predominantly focus on environmental controls over productivity, with less attention given to variation in mortality. It is uncertain how much these common model attributes limit our ability to predict terrestrial carbon uptake, storage, and turnover. Using demographic information from ~120 000 forest inventory plots distributed across the U.S. and Canada, we therefore asked: (1) is continental variation in growth and mortality more strongly related to environmental conditions or to species turnover? (2) how much do growth and mortality each contribute to continental variation in aboveground forest biomass?
We used hierarchical Bayesian models to statistically partition continental-scale variation in the biomass growth and mortality rates of 49 tree species groups into (1) species-independent spatial effects and (2) systematic differences in demographic performance among species. 82% of the observed variation in forest growth was explained by spatial effects that were independent of species composition, but the largest share of variation in mortality (44%) was related to species turnover across the continent. Simple differential equation models for biomass accumulation showed that observed variation in forest biomass can be explained primarily by spatial gradients in growth that are unrelated to species composition. Species-dependent patterns of mortality explained significant additional variation in biomass, with forests supporting less biomass when dominated by species with strong adverse responses to competition or susceptible to biotic disturbance agents. These empirical findings indicate that while aboveground forest carbon seems to be most strongly regulated by the effects of climate and soil on productivity, incorporating elements of interspecific variation in tree mortality into terrestrial biosphere models could further improve predictions of carbon turnover and storage.