Prioritizing Diurnal Variation in Aerosol Concentrations for Temporal Satellite AOD and Health Studies

Remote sensing from newer high-temporal resolution geostationary instruments is rapidly expanding. However, the utility of these data in exposure assessment and epidemiologic health studies may be greatest in regions with more substantial within-day variation of aerosols and ground-level particulate matter. To prioritize these regions of the continental USA, we analyzed aerosol optical depth (AOD) at long-running sunphotometer stations of the AERONET system during 2015-2018. We calculated the distribution of the within-day standard deviation of AOD at 470nm for station-days with >20 observations. We show that the distribution of within-day variation in AOD is highest in the mid-Atlantic, Southeastern US, and Central California AERONET stations (Figure 1) and these areas have much more within-day variation than the Mountain West where AOD from a single MODIS overpass may be more representative of within-day AOD. Within-day trends include notable heteroskedasticity and higher variation in the summer months. To compare AERONET values with remote sensing AOD retrievals, we collocated the same stations with 10 km resolution MAIAC AOD from the DSCOVR EPIC instrument (with ~6-10 potential AOD retrievals per day). The availability of data was substantially lower and the within-day standard deviation of AOD from the DSCOVR MAIAC retrievals was ~3x to 5x higher and lacked the same consistent spatial pattern, suggesting retrieval error may drown out within-day variation. Additional analyses will examine within-day variation in 2 km resolution MAIAC AOD retrievals from GOES-16. Because disentangling the significance of within-day variation in aerosol concentrations on human health requires measurements of ground conditions, we will also provide an analysis of within-day variation in EPA stations reporting hourly concentrations of fine particulate matter (PM 2.5 ). In addition to characterizing episodic and extreme events, a promising application of high-temporal resolution remote sensing of aerosols is to disentangle the contribution of varying conditions to human exposure and health outcomes. We propose a prioritization of research using geostationary aerosol estimates in regions and conditions with greater within-day variation.