Publications

Tao, MH; Chen, LF; Wang, ZF; Wang, J; Che, HZ; Xu, XG; Wang, WC; Tao, JH; Zhu, H; Hou, C (2017). Evaluation of MODIS Deep Blue Aerosol Algorithm in Desert Region of East Asia: Ground Validation and Intercomparison. JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES, 122(19), 10329-10340.

Abstract
The abundant dust particles from widespread deserts in East Asia play a significant role in regional climate and air quality. In this study, we provide a comprehensive evaluation of the widely used Moderate Resolution Imaging Spectroradiometer (MODIS) Deep Blue (DB) aerosol retrievals in desert regions of East Asia using ground-based observations over eight sites of the China Aerosol Remote Sensing Network (CARSNET). Different from their well-characterized performance in urban and cropland areas around the globe, DB aerosol optical depth (AOD) retrievals exhibit underestimation across the deserts in East Asia. We found that 38%-96% of satellite values fall out of an expected-error envelope of (0.05+20%AOD(CARSNET)), with the worst performance in Taklimakan Desert. In particular, DB retrievals erroneously give a nearly constant low values of 0.05 in Taklimakan Desert when AOD is below 0.5, which does not match with variation of moderate dust plumes. Comparison with Multi-angle Imaging SpectroRadiometer AOD shows that a similar underestimation is prevalent over the extensive deserts. Inversion of sky light measurements show that single scattering albedos of the yellow dust in East Asia are mostly below 0.9 at 440nm, much lower than the whiter and "redder" dust models applied in the DB algorithm. On the other hand, overestimation of surface reflectance dominantly contributes to the significant low constant AOD values in MODIS DB retrievals in Taklimakan Desert. These large biases, however, can be substantially reduced by considering unique characteristics of aerosols and surface over the arid regions in East Asia.

DOI:
10.1002/2017JD026976

ISSN:
2169-897X