Environmental Problems, Spatial Modeling, and Bayesian Inference

The goal of this presentation is to consider several environmental problems which can be usefully addressed using spatial process models and to detail the implementation of and benefits of fitting such models within a Bayesian framework. First, we review the formulation of general classes of hierarchical spatial models introducing spatial random effects modeled through spatial processes. We then briefly discuss simulation based strategies for the fitting of such models.

We then consider four illustrative problems as follows:
(1) multivariate modeling of pollution surfaces using coregionalization
(2) spatio-temporal analysis of pollution surfaces using dynamic models with spatially varying coefficients
(3) detecting gradients in pollution surfaces using directional derivative processes (4)population-adjusted toxin exposure analysis using covariate-weighted spatial CDF's.