You are here

Harvard Forest >

Harvard Forest Symposium Abstract 2015

  • Title: Using codispersion analysis to quantify forest spatial pattern
  • Primary Author: Hannah Buckley (Lincoln University)
  • Additional Authors: Bradley Case (Lincoln University, New Zealand); Aaron Ellison (Harvard University)
  • Abstract:

    Spatial patterns of forest trees reflect the cumulative history of the stand: species’ environmental tolerances and intra- and interspecific interactions overlain on environmental variation. To identify the scales at which these processes operate, we can quantify patterns using spatial pattern analysis. In particular, analyzing bivariate spatial patterns (species-species or species-environment) allows us to hypothesize which processes are important in forest stand structure and the spatial scale on which these processes occur. In many cases, environmental gradients dominate spatial patterns in forest stands, but the most widely-used spatial pattern analyses assume that underlying spatial patterns are isotropic (the same in all directions). In this research, we use codispersion analysis to detect and quantify anisotropic patterns and their relevant spatial scales in bivariate data. Simulated data illustrate the range of patterns that codispersion analysis can detect, including anisotropy. Analysis of species co-occurrence patterns within the Harvard Forest 35-ha forest dynamics plot show positive and negative covariation among the four most common tree species. Comparison of observed patterns to those simulated using null models shows that these patterns were all significantly different from random expectation. Our analysis shows that codispersion analysis is an effective method for quantifying anisotropic, bivariate patterns in forest stand data.

  • Research Category: Ecological Informatics and Modelling
    Large Experiments and Permanent Plot Studies