Reversible Jump MCMC for Inference in a Deterministic Individual-Tree-Based Growth Model for Studying Forest Dynamics

• Activity Number: 366
• Type: Contributed
• Date/Time: Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Jarrett Barber*+ and Darren E. Gemoets and Kiona Ogle
• Companies: University of Wyoming and University of Wyoming and University of Wyoming
• Keywords: Carbon dynamics ; dynamical midel ; forest dynamics ; individual-based model ; reversible jump MCMC

Abstract

We use a Bayesian approach to estimate input parameters of a deterministic individual-based model (IBM) of tree growth. Some input values lead to output that is discordant with observed tree mortality, i.e., the IBM deterministically kills trees it shouldn't, or vice-versa. Rather than change the IBM, which preserves important biological constraints, we introduce stochastic tree quantities (e.g., heights) otherwise determined by the IBM output if not for the discordance. As a result, the dimension of the estimation problem depends on the IBM inputs. We use reversible jump MCMC (RJMCMC) to address the trans-dimensional problem that arises. Though the possible number of discordances is finite, the complexity of the IBM leads us to consider trans-dimensional moves indexed by values in the uncountable input parameter space. Thus, we present a relatively unique application of RJMCMC.