Efficient Parameterizations for Multiscale Multivariate Spatio-Temporal Data

• Activity Number: 150
• Type: Invited
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Christopher K. Wikle* and Jonathan R. Bradley and Scott H. Holan
• Companies: University of Missouri and University of Missouri and University of Missouri
• Keywords: dynamics ; high-dimensional ; spatio-temporal ; basis functions ; multivariate ; areal

Abstract

Many data sources report related variables of interest that are also referenced over multiple geographic regions and time; however, there are relatively few general statistical methods that one can readily use that incorporate dependencies over different spatial locations, spatial scales, time points, and variables. Additionally, many multivariate spatio-temporal areal datasets are extremely high-dimensional, which leads to practical issues when formulating statistical models. We use the multivariate spatio-temporal mixed effects model (MSTM) in a fully Bayesian framework to analyze data of this type. Moreover, we introduce the use of Obled and Creutin eigenfunctions within this framework to allow for multivariate spatio-temporal data observed on different spatial scales as well as efficiently parameterized transition operators. We provide a demonstration of our approach using various environmental datasets.