Open AccessValidation of two‐channel VIRS retrievals of aerosol optical thickness over ocean and quantitative evaluation of the impact from potential subpixel cloud contamination and surface wind effect
Author(s) -
Zhao Tom X.P.,
Laszlo Istvan,
Holben Brent N.,
Pietras Christophe,
Voss Kenneth J.
Publication year - 2003
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/2002jd002346
Subject(s) - aeronet , subpixel rendering , environmental science , remote sensing , aerosol , satellite , wind speed , meteorology , advanced very high resolution radiometer , pixel , geology , computer science , physics , astronomy , computer vision
TRMM/CERES‐VIRS Single Satellite Footprint (SSF) data and AERONET Sun/sky radiometer observations from 1998 have been combined to validate SSF aerosol optical thickness (τ) retrievals over ocean along with a quantitative evaluation of the effects of potential subpixel cloud contamination and surface wind on the satellite τ retrievals. Potential subpixel cloud contamination is verified in Visible/Infrared Scanner (VIRS) SSF aerosol retrievals and constitutes a major source of systematic and random errors of the retrieval algorithm as determined from comparisons with AERONET observations. A positive correlation between the surface wind speed (which determines the roughness of the ocean surface) and the SSF τ has been observed for large surface wind speed. The validation results imply this correlation represents the real relationship between the surface wind and the wind‐driven aerosols rather than the disturbing effect of the surface reflectance associated with the rough ocean surface. After the potential subpixel cloud contamination is minimized and the effects of large surface wind are removed in the τ match‐ups, the positive biases in the SSF τ (compared to AERONET τ) for mean conditions have been reduced from 0.05 to 0.02 in VIRS channel 1 (0.63 μm) and 0.05 to 0.03 in channel 2 (1.61 μm). Random errors have also been reduced from 0.09 to 0.06 at 0.63 μm, and from 0.06 to 0.05 at 1.61 μm. The validation results support the application of the SSF aerosol data in radiation and climate studies as well as supply useful guidance for the adjustment and improvement of the aerosol retrieval algorithm.