High Resolution Bayesian Space-Time Modelling for Ozone Concentration Levels
Sujit K. Sahu
School of Mathematics, University of Southampton, United Kingdom

Ground-level ozone is a pollutant that is a significant health risk, especially for children with asthma. It also damages crops, trees and other vegetation. It is a main ingredient of urban smog. To evaluate exposure to ozone levels, the United States Environmental Protection Agency (USEPA) has developed a primary and a secondary air quality standard. To assess compliance to these standards, the USEPA collects ozone concentration data continuously from several networks of sparsely and irregularly spaced monitoring sites throughout the US. Data obtained from these sparse networks must be processed using spatial and spatio-temporal methods to check compliance to the ozone standards at an unmonitored site in the vast continental land mass of the US.

This talk will first discuss the two air quality standards for ozone levels and then will develop high resolution Bayesian space-time models which can be used to assess compliance. Predictive inference properties of several rival modelling strategies for both spatial interpolation and temporal forecasting will be compared and illustrated with simulation and real data examples. A number of large real life ozone concentration data sets observed over the eastern United States will also be used to illustrate the Bayesian space-time models. Several prediction maps from these models for the eastern US, published and used by the USEPA, will be discussed.

Keywords: Space-time modelling; Bayesian Spatial Prediction; Auto-regressive model; Dynamic Linear Model

Biography: Sujit K. Sahu is a senior lecturer in statistics in the School of Mathematics, University of Southampton.He obtained his Ph.D. from the University of Connecticut in 1994. Before joining Southampton in 1999 he worked for the Cambridge and Cardiff Universities. He is interested in understanding uncertainty through the analysis of large and complex space-time data sets and the development of predictive models using Bayesian methods.