Covariance Estimation for Natural Spatio-Temporal Processes

• Activity Number: 254
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Michael Horrell*+
• Companies: University of Chicago
• Keywords: Remote sensing ; Gaussian random fields ; Composite Likelihood

Abstract

Analysis of remotely sensed natural processes requires use of flexible models and computational techniques that can be applied to large datasets. We develop new models for natural processes that capture physical movement of the process as well as other structures of typical interest in spatio-temporal statistics. The nature of the models we develop requires use of large datasets for certain covariance parameters to be estimated. We fit these models to atmospheric data collected by satellites using a fitting procedure related to composite likelihood methods.