IMAGe Seminar - Opening the black box: Statistical differential equation modeling using homogenization

02/19/2015 - 2:30pm
Mesa Laboratory - Vislab

Mevin Hooten, Colorado State University

Statistical models using partial differential equations (PDEs) to describe dynamically evolving natural systems are appearing in the scientific literature with some regularity in recent years. Such studies seek to characterize the dynamics of temporal or spatio-temporal phenomena such as weather and climate systems, invasive species, ecological consumer-resource interactions, biological community evolution, and animal resource selection. In the spatio-temporal setting data are often available at varying scales and the integration of a PDE may be computationally infeasible over the spatial support of interest. We present an approach to impose computationally advantageous changes of support in statistical implementations of PDE models and demonstrate its utility using a form of PDE known as "ecological diffusion." This method provides an emulator that is model specific and consistent with the inherent mathematics, and thus differs from phenomenological emulators such as those constructed using Gaussian processes.

Bio: Dr. Mevin Hooten is Associate Professor with a joint appointment in the Departments of Fish, Wildlife, and Conservation Biology and Statistics at Colorado State University as well as Assistant Unit Leader of the Colorado Cooperative Fish and Wildlife Research Unit. He earned a PhD in Statistics from the University of Missouri in 2006. His main research is on spatio-temporal statistical modeling of ecological processes, focusing on applications in animal movement, disease ecology, landscape genetics, optimal monitoring, paleoecology/climate, and ecosystem/environmental dynamics. Additional details can be found on this website: http://warnercnr.colostate.edu/~hooten/

Thursday, February 19, 2015
Location: Mesa Laboratory VisLab
Time: 2:30pm