Computational methods for the spatial prediction of space-time pollution fields
by
Melissa Dobbie
Department of Statistics, University of British Columbia, Canada
Coauthors: Jim Zidek (University of British Columbia), Nhu Le (University of British Columbia and BC Cancer Agency), Li Sun (Ericsson Berkeley Research Center)

Studies have shown that elevated levels of the pollutant PM₁₀ are associated with acute negative health impacts, and consequently, atmospheric concentrations of these particles are systematically monitored in most urban areas in developed countries. Our study will focus on a network of monitoring sites at a small number of locations in the Greater Vancouver Regional District in British Columbia, Canada. We describe an approach for predicting the average daily concentrations of PM₁₀ levels for unmonitored sites in this region and present details of our corresponding software. The software is relatively user-friendly due to its interactive nature, extensive commentary, and thorough on-line documentation for new syntax. It is also flexible, enabling customization, enhancement and thus broad utilization.

Date received: April 4, 2002