You are here

Harvard Forest >

Harvard Forest Symposium Abstract 2010

  • Title: Informed priors improve demographic matrix model forecasts for northern pitcher plants (Sarracenia purpurea L.)
  • Primary Author: Sydne Record (Harvard Forest)
  • Additional Authors: Aaron Ellison (Harvard University); Nicholas Gotelli (University of Vermont (UVM))
  • Abstract:

    Ecologists often use short-term demographic measurements to model population dynamics, but the resulting population projections may have high levels of uncertainty. One way to reduce uncertainty in these models is to include information from similar studies as prior data in a Bayesian framework. This study compares the results of stochastic Bayesian demographic matrix models for northern pitcher plants (Sarracenia purpurea) using uninformed and informed priors. We fit a stochastic matrix model to demographic parameters measured annually at Molly Bog, VT (1998-2008), and use comparable measurements at Hawley Bog, MA (1998-2005) to construct a series of increasingly informed prior distributions. Matrix sensitivity and elasticity estimates were not affected greatly by using uninformed versus informed priors. However, using increasingly informed priors resulted in incrementally higher estimates of instantaneous growth rate (uninformed λ = 0.934, informed λ = 0.948) and lower estimates of quasi-extinction risk for small populations (uninformed p = 0.58, informed p = 0.26). Informed priors decreased uncertainty in estimates of the instantaneous growth rate and quasi-extinction risk. Using informed prior distributions can reduce uncertainty in demographic forecasts, although the method may be sensitive to variation among sites in underlying population parameters.

  • Research Category: Ecological Informatics and Modelling
    Physiological Ecology, Population Dynamics, and Species Interactions