Bayesian semiparametric space-time models

 

We propose a new class of semi-parametric Bayesian models for spatial, temporal, and spatio-temporal data, a generalized moving-average of Lévy random fields. The method is useful for building flexible spatio-temporal models that can accommodate non-Gaussian non-stationary spatio-temporal data, while keeping the computation feasible even for large data sets. For some applications the methods have advantages over Gaussian random field models and wavelet models. The methods are illustrated in an application to sulfur dioxide monitoring in the mid-Atlantic United States.