Premium
Detection of cloud‐affected AIRS channels using an adjacent‐pixel approach
Author(s) -
Joiner J.,
Poli P.,
Frank D.,
Liu H. C.
Publication year - 2004
Publication title -
quarterly journal of the royal meteorological society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.744
H-Index - 143
eISSN - 1477-870X
pISSN - 0035-9009
DOI - 10.1256/qj.03.93
Subject(s) - cloud computing , atmospheric infrared sounder , remote sensing , depth sounding , pixel , environmental science , data assimilation , computer science , moderate resolution imaging spectroradiometer , residual , satellite , meteorology , troposphere , algorithm , geology , artificial intelligence , geography , physics , oceanography , operating system , astronomy
High‐spectral‐resolution infrared sounders such as the Atmospheric InfraRed Sounder (AIRS), flying on the National Aeronautics and Space Administration Earth Observing System (EOS) Aqua satellite, provide information about vertical temperature and humidity structure that is potentially useful for data assimilation and numerical weather prediction. Tropospheric channels from infrared sounders are frequently affected by cloud. The methods currently used operationally to account for cloud effects are screening to eliminate cloud‐contaminated data and cloud‐clearing. For either approach, it is important to determine which channels peak sufficiently above the cloud so that they are not contaminated. Depending on the sounding, different combinations of channels are cloud‐contaminated, thus making cloud detection difficult. This paper proposes a new method of identifying clear or unaffected channels using adjacent pixels. Unlike other proposed or implemented methods, this approach does not rely heavily on having accurate background information about the atmospheric state or estimates of its error. The method also does not make any assumptions about cloud spectral properties. Instead, it assumes that clouds will produce adjacent‐pixel variability. The approach requires an estimate of clear‐scene adjacent‐pixel homogeneity that can be obtained with real data. We apply the methodology to simulated AIRS data using Monte Carlo experiments. We then apply the algorithm to real AIRS data and use brightness temperature departure statistics as an independent check for residual cloud contamination. We also make qualitative comparisons with cloud properties derived from the EOS Aqua Moderate‐Resolution Imaging Spectroradiometer instrument. Copyright © 2004 Royal Meteorological Society