Harvard Forest >

Harvard Forest Symposium Abstract 2013

• Title: Cross-site analysis of tree demographics and carbon reserves across eastern US forests
• Primary Author: Joshua Mantooth (Boston University)
• Additional Authors: Michael Dietze (Harvard University)
• Abstract:

  Improving predictions of how forests of the eastern United States will respond to future global change requires a better understanding of tree demographic rates and their drivers. In order to improve our understanding of tree demographic rates, the Dietze Laboratory at Boston University established a network of forest inventory plots across ten sites in the eastern US (2011-2012), including Harvard Forest. These plots are being used to capture demographic rates of growth, mortality and recruitment across large taxonomic and spatial scales. These data are being used to answer questions about the importance of various global change and disturbance agents driving eastern US forest dynamics, the importance of trait versus environmental variations in driving demographic dynamics and the relationship between demographic rates and tree carbon reserves – in the form on non-structural carbohydrates (NSC).
  
  
  
  It has been suggested that tree NSC play a large role in reducing tree mortality rates by providing a carbon reserve buffer during stressor events. However, data on NSC reserves are limited to small spatial, temporal and taxonomic scales. We are currently collecting and analyzing NSC data from Harvard Forest and nine other sites to better understand the correlations between NSC and observed demographic rates – specifically growth and mortality. In addition, we are constructing models of plant carbon allocation that best explain the observed patterns of NSC at Harvard Forest and across the eastern US. The ability of these models to predict forest dynamics will be tested against the current allocation scheme of the Ecosystem Demography model (ED2).
  
  
  
  Additional current analyses are focused on growth responses of adult trees, detected from tree rings, and utilize hierarchical Bayesian state-space modeling to understand how trees are responding to environmental covariates at multiple spatial and temporal scales. We have begun to construct a predictive model of tree growth that incorporates our observed effects at the individual, plot and site level, with the planned incorporation of landscape-scale effects in the near future. This model will improve our understanding of how trees grow in response to their environment and will inform future modeling efforts using ED2.

• Research Category: Regional Studies