You are here:
CONTENDING WITH SPACE-TIME INTERACTION IN THE SPATIAL PREDICTION OF POLLUTION: VANCOUVER'S HOURLY AMBIENT PM 10 FIELD
Citation:
Zidek, J., L. Sun, N. D. Le, AND A H. Ozkaynak. CONTENDING WITH SPACE-TIME INTERACTION IN THE SPATIAL PREDICTION OF POLLUTION: VANCOUVER'S HOURLY AMBIENT PM 10 FIELD. ENVIRONMETRICS 13(5-6):595-613, (2002).
Impact/Purpose:
The primary objective of this research is to improve current PM population exposure models to more accurately predict exposures for the general population and susceptible sub-populations. Through model improvements, a better understanding of the major factors controlling exposure to PM will be achieved. Specific objectives of this research are to:
- predict total personal exposure to PM10 and PM2.5 for the general and for susceptible sub-populations residing in different urban environments
- estimate the contribution of ambient PM to predicted total PM exposures
- determine what factors are of primary importance in determining PM exposures, including an analysis of the effects of time spent in various microenvironments and the importance of spatial variability in ambient PM concentrations
- determine what factors contribute the greatest uncertainty to model predictions and make recommendations for measurement and modeling studies to reduce these uncertainties
- predict daily and annual average exposures using single or multi-day time-activity diaries
- incorporate state-of-the-art dosimetric models of the lung into PM population exposure and dose models
- evaluate models against measured data from PM panel and other exposure measurement studies
- develop exposure and dose metrics applicable to acute and chronic environmental epidemiology studies
Description:
In this article we describe an approach for predicting average hourly concentrations of ambient PM10 in Vancouver. We know our solution also applies to hourly ozone fields and believe it may be quite generally applicable. We use a hierarchal Bayesian approach. At the primary level we model the logarithmic field as a trend model plus a Gaussian stochastic residual. That trend model depends on hourly meteorological predictors and is common to all sites. The stochastic component consists of a 24 hour vector response that we model as a multivariate AR(3) temporal process with common spatial parameters. Removing the trend and AR structure leaves "whitened" time series of vector series. With this approach ( as opposed to using 24 separate univariate time series models), little loss of spatial; correlation in these residuals compared with that in just the detrended residuals (prior to removing the AR component). Moreover, our multivariate approach enables predictions for any given hour to " borrow strength" through its correlation with adjoining hours. On this basis we develop an spatial predictive distribution for these residuals at unmonitored sites. By transforming the predicted residuals back to original data scales we can impute Vancouver's hourly PM10 field.
The U.S. Environmental Protection Agency through its Office of Research and Development partially funded the research described here under a Cooperative Agreement #CR825267-01 to Harvard University School of Public Health. It has been subjected to Agency review and approved for publication. Mention of trade names or commercial products does not constitute an endorsement or recommendation for use.