The use of kriging to estimate monthly ozone exposure parameters for the Southeastern United States

This paper explores the feasibility of (1) using kriging to predict the monthly mean of daily 7-h mean (0900-1559) O3 concentrations, (2) using kriging to estimate the per cent of hourly mean O3 concentrations equal to or greater than 0.07 ppm (137 microg m(-3)) for a specific month, and (3) developing a quantitative relationship between the monthly mean of the daily 7-h (0900-1559) average O3 concentration and the monthly number of hourly concentrations > or = 0.08p ppm (157 microg m(-3)). We found that kriging can be used to estimate the (1) monthly mean of daily 7-h mean O3 concentrations and (2) the percentage of hourly concentrations for a given month > or = 0.07 ppm when sufficient spatial coverage was available. However, the per cent > or = 0.07 ppm parameter exhibited much greater relative variability than the monthly 7-h exposure index. A strong statistical association was found between the monthly number of occurrences > or = 0.08 ppm and monthly 7-h mean concentrations above 0.05 ppm (98 microg m(-3)). Because of the variability that cumulative indices, such as the monthly percentage of hourly concentrations > or = 0.07 ppm , exhibit from site to site, it appears that whether kriging techniques or mathematical regressions are used to estimate the number of elevated O3 hourly concentrations above selected thresholds, large uncertainties associated with the predicted values will exist. These large uncertainties will make it difficult to accurately estimate vegetation effects caused by ambient levels of O3. However, if a generalized quantitative relationship between repeated occurrences of hourly mean concentrations > or = 0.07 ppm or > or = 0.08 and vegetation effects can be developed, it may be possible, using kriged monthly values accompanied with confidence intervals, to identify those areas where vegetation may be at risk. However, before it will be possible to implement such an approach, researchers will have to better quantify the relationship between realistic O3 exposures and vegetation effects.