Using ground-based observations and satellite retrievals to constrain urban-to-regional-scale air quality chemical transport modeling
Friberg, Mariel D.
MetadataShow full item record
The overarching goal of this research is to improve urban- and regional-scale air quality modeling for health risk assessment using a combination of ground-station and satellite-based measurements. The integration of near-surface air pollution concentrations, emissions-based air quality model simulations, and satellite observations of column-integrated quantities will improve the accuracy of exposure metrics and promote the appropriate use of satellite data over extended areas for long time periods, especially where ground-based air quality measurement networks are limited or lacking. In the broader sense, this information will help public health scientists, policy makers, and monitoring agencies to research and implement better control strategies and regulations. The first phase of this study (Friberg et al., 2016) demonstrated and assessed a systematic and practical approach to fusing surface-network measurements with chemical transport model (CTM) simulations to produce daily concentrations for five air pollutants of primary origin (NO2, NOx, CO, SO2, and EC), and seven secondary pollutants (O3, PM10 mass, PM2.5 mass, SO4, NH4, NO3, and OC) for use in cross-sectional epidemiological studies. A second study (Friberg et al., 2017) assessed the ability of the data fusion method to produce daily concentrations across five metropolitan areas from 2002 to 2008. In addition to the variety of pollutant types, the five cities represent a range of meteorological conditions, background aerosol conditions, population densities, and sampling-station distributions. Among the pollutant types, the primary pollutants tend to be more heterogeneously distributed over the urban regions than the secondary ones. The resulting daily spatial field estimates of air pollutant concentrations and associated correlations were not only consistent with observations, emissions, and meteorology, but substantially improved CTM-derived results in areas without observations, for most pollutants and all cities. The data fusion improved daily metrics across all pollutants with the greatest improvements occurring for O3 and PM2.5. The final study (Friberg et al., 2017, to be submitted) demonstrated and assessed an optimization technique, expanding upon the surface-station-model fusion technique, to estimate ambient PM2.5 mass and associated chemically speciated concentrations for potential use in longitudinal epidemiological studies. The newest method constrains surface PM2.5 and chemical-component CTM results, using both ground-station data to anchor speciated, near-surface aerosol concentrations, and total column aerosol optical depth (AOD). When the mid-visible AOD is high, the retrieved AOD from the Multi-angle Imaging SpectroRadiometer (MISR) Research Aerosol retrieval algorithm along with qualitative, column-effective aerosol type observations helped constrain the CTM over extended regions. The retrieved AOD had a horizontal resolution of 275m. The method was applied over a case study area in the San Joaquin Valley of California during NASA’s DISCOVER-AQ field campaign in this region, on days when there was good satellite coverage and considerable suborbital data for validation of the approach. The accuracy of estimated concentrations and evaluation of the latest MISR aerosol retrieval algorithm ability to typify urban AOD, aerosol mixtures, and aerosol airmasses were examined by comparing the results with speciated ground observations and standard model fitting statistics. The results indicate that on days with high AOD and adequate observing conditions, satellite retrievals improve simulated spatial distributions of PM2.5 and chemical component concentrations.