Assessing and communicating uncertainties in PM2.5 estimates derived from satellite aerosol optical depth
Abstract
Health and air quality stakeholders are increasingly using fine particulate matter (PM2.5) derived from satellite aerosol optical depth (AOD) to inform decisions. As one example, we are deriving PM2.5 exposure maps from satellite (MAIAC retrieval from MODIS) AOD at 1 km2 resolution for 2011 to link with health records over New York State. Our approach applies relationships between surface PM2.5 and column AOD simulated by a regional air quality model (CMAQ; 12 km2 horizontal resolution). We demonstrate that the seasonal average satellite-derived PM2.5 reveals more spatial detail and better captures ground-based observations than CMAQ. With ground-based remote measurements of AOD located near surface PM2.5 monitors, we show, on average, that the modeled PM2.5/AOD relationships introduce more uncertainty to satellite derived PM2.5 than the MAIAC AOD product retrieved from the satellite instrument. Uncertainties in modeled PM2.5/AOD relationships reflect a combination of factors, including the simulated aerosol speciation, vertical distributions, relative humidity, and aerosol optical properties. Evaluation with ground based speciated PM2.5 measurements indicates that poor representation of aerosol speciation leads to seasonality in model biases; evaluation with the 2011 DISCOVER-AQ aircraft campaign over the mid-Atlantic U.S.A. suggests that uncertainties in simulated vertical profiles contribute a large fraction of the total uncertainty in satellite derived PM2.5 on individual days. We show that satellite-derived PM2.5 is sensitive to the parameterization of optical properties such as size distribution (the largest source of uncertainty at 30%), and hygroscopicity of inorganic salt (20%). Future efforts to reduce uncertainty in geophysical approaches to derive surface PM2.5 from satellite AOD would thus benefit from improved model representation of vertical PM2.5 distributions and of aerosol optical properties. In addition to our health-motivated application, we discuss the relevance of these uncertainties to converting measured, modeled, or satellite-derived aerosol mass to visibility for assessing compliance under the regional haze rule. We will highlight ongoing efforts, in partnership with air quality managers, to communicate our findings to a broad set of stakeholder groups.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.A43E..04F
- Keywords:
-
- 0345 Pollution: urban and regional;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 0240 Public health;
- GEOHEALTHDE: 1640 Remote sensing;
- GLOBAL CHANGEDE: 6309 Decision making under uncertainty;
- POLICY SCIENCES