Aerosol and Clouds
Utilizing the MODIS 1.38 μm channel for cirrus cloud optical thickness retrievals: Algorithm and retrieval uncertainties
Article first published online: 18 DEC 2010
Copyright 2010 by the American Geophysical Union.
Journal of Geophysical Research: Atmospheres (1984–2012)
Volume 115, Issue D24, 27 December 2010
How to Cite
2010), Utilizing the MODIS 1.38 μm channel for cirrus cloud optical thickness retrievals: Algorithm and retrieval uncertainties, J. Geophys. Res., 115, D24209, doi:10.1029/2010JD014872., and (
- Issue published online: 18 DEC 2010
- Article first published online: 18 DEC 2010
- Manuscript Accepted: 20 OCT 2010
- Manuscript Revised: 12 OCT 2010
- Manuscript Received: 6 AUG 2010
- cirrus clouds;
- remote sensing;
 The cloud products from the Moderate Resolution Imaging Spectroradiometers (MODIS) on Terra and Aqua have been widely used within the atmospheric research community. The retrieval algorithms, however, oftentimes have difficulty detecting and retrieving thin cirrus, due to sensitivities to surface reflectance. Conversely, the 1.38 μm channel, located within a strong water vapor absorption band, is quite useful for detecting thin cirrus clouds since the signal from the surface can be blocked or substantially attenuated by the absorption of atmospheric water vapor below cirrus. This channel, however, suffers from nonnegligible attenuation due to the water vapor located above and within the cloud layer. Here we provide details of a new technique pairing the 1.38 μm and 1.24 μm channels to estimate the above/in-cloud water vapor attenuation and to subsequently retrieve thin cirrus optical thickness (τ) from attenuation-corrected 1.38 μm reflectance measurements. In selected oceanic cases, this approach is found to increase cirrus retrievals by up to 38% over MOD06. For these cases, baseline 1.38 μm retrieval uncertainties are estimated to be between 15 and 20% for moderately thick cirrus (τ > 1), with the largest error source being the unknown cloud effective particle radius, which is not retrieved with the described technique. Uncertainties increase to around 90% for the thinnest clouds (τ < 0.5) where instrument and surface uncertainties dominate.