Search

There is significant public health interest in gaining a better understanding of the health effects associated with particulate matter (PM) of differing sizes and compositions. Unfortunately the PM monitoring data needed for such studies are limited. Satellite observations of aerosol optical depth (AOD) have been used to overcome this limitation, but existing techniques using AOD to estimate surface concentrations of PM have limited ability to differentiate particle size and type, and the spatial resolution is often too coarse to evaluate PM variability on local scales. Fortunately there has been a significant breakthrough in aerosol product from the Multi-angle Imaging SpectroRadiometer (MISR) instrument onboard the NASA Terra satellite. Recent computational advances have made it possible to derive a “local mode” MISR aerosol product at a spatial resolution of 4.4 km that maintains the quality of the 17.6 km resolution Version 22 operational aerosol product. MISR is uniquely capable of providing total AOD as well as AOD fractionated by size (small, medium, and large), enabling estimation of size-resolved surface PM concentrations. Total and size-fractionated AOD from the MISR 4.4 km aerosol product were used to develop spatio-temporal models to estimate surface concentrations of PM 2.5 , PM 10 , and PM 2.5 chemical species over Southern California. We found, and confirmed by leave-one-site-out cross validation, that PM 2.5 was best estimated with a spatio-temporal model of AOD small + medium that also included adjustment for relative humidity and wind speed (R 2 = 0.67, CV R 2 = 0.51), while PM 10 was best estimated from AOD large with adjustment for dew point and wind speed (R 2 = 0.76, CV R 2 = 0.44). Total AOD was most strongly associated with PM 2.5 components SO 4 2 − (R 2 = 0.74, CV R 2 = 0.69) and NO 3 − (R 2 = 0.72, CV R 2 = 0.39). The best fitting models were applied to all available MISR aerosol retrievals over the region, generating surfaces of estimated size-resolved PM concentrations that will be a great asset to the environmental science and public health communities. [ABSTRACT FROM AUTHOR]

Research efforts to better characterize the differential toxicity of PM 2.5 (particles with aerodynamic diameters less than or equal to 2.5 μm) speciation are often hindered by the sparse or non-existent coverage of ground monitors. The Multi-angle Imaging SpectroRadiometer (MISR) aboard NASA's Terra satellite is one of few satellite aerosol sensors providing information of aerosol shape, size and extinction globally for a long and continuous period that can be used to estimate PM 2.5 speciation concentrations since year 2000. Currently, MISR only provides a 17.6 km product for its entire mission with global coverage every 9 days, a bit too coarse for air pollution health effects research and to capture local spatial variability of PM 2.5 speciation. In this study, generalized additive models (GAMs) were developed using MISR prototype 4.4 km-resolution aerosol data with meteorological variables and geographical indicators, to predict ground-level concentrations of PM 2.5 sulfate, nitrate, organic carbon (OC) and elemental carbon (EC) in Southern California between 2001 and 2015 at the daily level. The GAMs are able to explain 66%, 62%, 55% and 58% of the daily variability in PM 2.5 sulfate, nitrate, OC and EC concentrations during the whole study period, respectively. Predicted concentrations capture large regional patterns as well as fine gradients of the four PM 2.5 species in urban areas of Los Angeles and other counties, as well as in the Central Valley. This study is the first attempt to use MISR prototype 4.4 km-resolution AOD (aerosol optical depth) components data to predict PM 2.5 sulfate, nitrate, OC and EC concentrations at the sub-regional scale. In spite of its low temporal sampling frequency, our analysis suggests that the MISR 4.4 km fractional AODs provide a promising way to capture the spatial hotspots and long-term temporal trends of PM 2.5 speciation, understand the effectiveness of air quality controls, and allow our estimated PM 2.5 speciation data to be linked with common spatial units such as census tract or zip code in epidemiological studies. This modeling strategy needs to be validated in other regions when more MISR 4.4 km data becoming available in the future. [ABSTRACT FROM AUTHOR]