A bivariate space–time downscaler under space and time misalignment

• DOI: 10.1214/10-aoas351
• Publication Date: 2011-01-04T20:07:34Z
• Authors: Veronica J. Berrocal; Alan E. Gelfand; David M. Holland
• Keywords: FOS: Computer and information sciences; multivariate spatial process; coregionalization; dynamic model; Statistics - Applications; 01 natural sciences; 3. Good health; 13. Climate action; Co-kriging; kriging; Applications (stat.AP); spatially varying coefficients; 0105 earth and related environmental sciences

Ozone and particulate matter PM(2.5) are co-pollutants that have long been associated with increased public health risks. Information on concentration levels for both pollutants come from two sources: monitoring sites output complex numerical models produce surfaces over large spatial regions. In this paper, we offer a fully-model based approach fusing these sources of information the pair which is computationally feasible regions periods time. Due to association between environmental contaminants, it expected regarding one will help improve prediction other. Misalignment an obvious issue since networks contaminants only partly intersect because collection rate typically less frequent than ozone.Extending previous work in Berrocal et al. (2009), introduce bivariate downscaler provides flexible class space-time assimilation models. We discuss computational issues model fitting analyze dataset ozone season during year 2002. show modest improvement predictive performance, not surprising setting where can anticipate small gain.