NOVEL STATISTICAL APPROACH TO EVALUATE SPATIAL DISTRIBUTION OF PM FROM SPECIFIC SOURCE CATEGORIES
This task addresses aspects of NRC recommendations 10A and 10B. Positive matrix factorization (PMF) is a new statistical techniques for determining the daily contribution to PM mass of specific source categories (auto exhaust, smelters, suspended soil, secondary sulfate, etc.). Investigators in Finland have developed a novel statistical technique for applying PMF to multiple sites. A contractor will apply this technique to a St. Louis data set of PM2.5 and PM15 mass and composition (as determined by x-ray flourescent analysis), as well as some gaseous pollutant data. The St. Louis RAPS-RAMS data set is unique in having XRF elemental analysis for about two years at 10 sites with some sites having 6 hour resolution.Results: Information of the spatial distribution of the PM concentration due to various components and from various source categories. This will determine how well the value at one site can represent exposure for the entire community. A data set of daily source category contributions, provided by this task, will be used in epidemiologic analyses. Relevance: This analysis will provide important information on exposure error in acute community time series epidemiology for PM. Exposure assessment is an integral part of developing risk assessments which NCEA includes as part of its primary mission. This information will also improve the science base used to support PM regulations and improve the quality and reliability of Risk Assessment and Mitigation.