Reduced Rank Spatial Model With Temporal Confounding Effects

• Activity Number: 523
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Rajib Paul*+ and Eddy Campbell and John Parslow and Lawrence Murray and Emlyn Jones
• Companies: Western Michigan University and The Commonwealth Scientific and Industrial Research Organisation and The Commonwealth Scientific and Industrial Research Organisation and The Commonwealth Scientific and Industrial Research Organisation and The Commonwealth Scientific and Industrial Research Organisation
• Keywords: Cross Covariance ; Process Convolution ; Threshold Estimation

Abstract

Recent advances in technology facilitate collecting data for ocean processes, such as, salinity, temperature, dissolved oxygen, and chlorophyll. Glider is one of those easy and cheap devices for collecting data. In this paper, we analyze the data collected on the aforementioned processes over Storm Bay in Central Tasmania. The datasets exhibit high spatial resolution, but each location was observed at different time points. Due to advection, these processes can change over a short period of time. Hence, we cannot ignore the time effect completely and use a pure spatial model. We adjust for the time effects by using an autoregressive process of order one in our reduced rank spatial model. Our inferences on the model parameters are based on an MCMC algorithm.