Nonlinear Modeling of Ecosystem Dynamics Using Computer Models

• Activity Number: 101
• Type: Invited
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: ENAR
• Author(s): William Leeds*+ and Christopher Wikle
• Companies: University of Missouri and University of Missouri
• Keywords: Nonlinear dynamical systems ; Bayesian hierarchical models ; Computer models

Abstract

This talk focuses on multivariate spatio-temporal dynamical processes for environmental systems. These processes are often nonlinear in time and nonstationary in space. At times, we have some information as to the functional form of these dynamical relationships, in the form of deterministic computer models. These computer models range in complexity, and often models that include important dynamics (e.g., bio-physical interactions, advection, diffusion, etc.) are computationally expensive. We consider a framework for including dynamics from computationally expensive deterministic computer models into a Bayesian hierarchical model (BHM). While the methodology is motivated by ecosystem dynamics, it would be useful to anyone modeling a multivariate space-time process that exhibits nonlinear behavior and for which some a priori knowledge is available concerning the process dynamics. In particular, it may be helpful to those who want to incorporate dynamics from a deterministic computer model into a BHM.