Often in regionally aggregated spatiotemporal models, a single variance parameter is used to capture variability in the spatial structure of the model, ignoring the impact that spatially-varying factors may have on the variability in the underlying process. We extend existing methodologies to allow for region-specific variance components in our analysis of monthly asthma hospitalization rates in California counties, introducing a heteroscedastic CAR model that can greatly improve the fit of our spatiotemporal process. After demonstrating the effectiveness of our new model via simulation, we reanalyze the asthma hospitalization data and note a number of important findings.
Keywords: Bayesian methods; Gaussian process; Gradients; Markov chain Monte Carlo; Spatial process models.