Improved algorithm for MODIS satellite retrievals of aerosol optical depths over western North America

Authors

Easan Drury, Harvard University
Daniel J. Jacob, Harvard University
Jun Wang, University of Nebraska - Lincoln
Robert J.D. Spurr, RT Solutions, Inc.
Kelly Chance, Harvard-Smithsonian Center for Astrophysics

Document Type

Article

Date of this Version

2008

Citation

Drury, E., D. J. Jacob, J. Wang, R. J. D. Spurr, and K. Chance (2008), Improved algorithm for MODIS satellite retrievals of aerosol optical depths over western North America, J. Geophys. Res., 113, D16204, doi:10.1029/2007JD009573.

Comments

Copyright 2008 by the American Geophysical Union. Used by permission.

Abstract

Quantitative evaluation of chemical transport models (CTMs) with aerosol optical depth (AOD) products retrieved from satellite backscattered reflectances can be compromised by inconsistent assumptions of aerosol optical properties and errors in surface reflectance estimates. We present an improved AOD retrieval algorithm for the MODIS satellite instrument using locally derived surface reflectances and CTM aerosol optical properties for the 0.47, 0.65, and 2.13 μm MODIS channels. Assuming negligible atmospheric reflectance at 2.13 μm in cloud-free conditions, we derive 0.65/2.13 surface reflectance ratios at 1° X 1.25° horizontal resolution for the continental United States in summer 2004 from the subset of top-of-atmosphere (TOA) reflectance data with minimal aerosol reflectance. We obtain a mean ratio of 0.57 ± 0.10 for the continental United States, with high values over arid regions and low values over the Midwest prairies. The higher surface reflectance ratios explain the high AOD bias over arid regions found in previous MODIS retrievals. We calculate TOA reflectances for each MODIS scene using local aerosol optical properties from the GEOS-Chem CTM, and fit these reflectances to the observed MODIS TOA reflectances for a best estimate of AODs for that scene. Comparison with coincident ground-based (AERONET) AOD observations at 16 sites in the western and central United States in summer 2004 shows poor correlation in the daily data but the correlation improves as averaging time increases. Averaging over the available coincident observations (n = 11–44 days) results in strong correlations (R_0.47μm = 0.90, R_0.65μm = 0.67) and a 19% low bias, representing considerable improvement over the operational MODIS AOD products in this region.