Spatio-Temporal Calibration and the Change of Support Problem

Spatial epidemiology is fast becoming a very important branch of public health research as it offers a very unique set of tools to understand and explain various health phenomena. One focus of spatial epidemiology is understanding disease etiology while controlling and accounting for spatial correlation and dependence between study subjects and/or exposure covariates. In addition to spatial variation, time is another important covariate that very often needs to be accounted for when studying disease etiology. Spatial epidemiologic studies analyzing the effects of environmental pollutants on a range of public health outcomes thus require exposure data to be collected over time across a geographical area. These data are then used to develop complex space-time models before the eects of pollutants on the public health outcome of interest can be accurately analyzed. The complexity of these models arises from the dependence of these data over space and time, i.e. these exposure data are spatially and temporally correlated.