Assisting Bay Area Air Pollution Management of Surface PM2.5 by Using Satellite AOD Data

Surface fine particulate matter less than 2.5 microns in diameter (PM2.5) imposes numerous health risks to population around the world. Making decisions based on the surface PM2.5 concentrations by health department and other stakeholders requires more accurate quantification and characterization of the surface PM2.5 distribution, the spatial variabilities on the community scale, as well as analyses that utilize ground-based monitors, satellite information and model simulations. In this work, we will first present a visualization tool based on the ArcGIS platform that displays the satellite aerosol optical depth (AOD) with the background of geographical information. Using this visualization tool, a composite dataset is selected by choosing days with good correlations between AOD and surface PM2.5, and spatial variabilities of the columnar AOD in the Bay Area will be analyzed. Second, we will derive the surface PM2.5 based on the satellite AOD information by using regression and spatial models. A sensitivity test was conducted using two different types of regression algorithms, and the results showed that the inverse distance weighted (IDW) method has better statistical correlations between AOD and surface PM2.5 compared to the B-spline method. Lastly, we will present a survey on various PM2.5 datasets, method, and their limitations. This survey will contrast three publicly available PM2.5 datasets that are frequently used for public health purposes, including the CDC WONDER data, the Dalhousie University data, and the CDC Tracking Network data. A summary on the commonly used method to generate PM2.5 exposure datasets will be given, and this survey calls for more long-term, consistent validation on the PM2.5 exposure data, and provides a guidance to users who access publicly available PM2.5 for health exposure analyses. This work is based on our research as a member of the NASA Health and Air Quality Applied Science Team.