Chang and Li [2002] proposed a new cloud microphysics retrieval technique that can estimate the vertical profile of droplet effective radius (DER) for water clouds using multispectral near-infrared (NIR) measurements. The underlying principle of the retrieval technique is that radiance measurements at distinct multi-NIR wavelengths possess different penetration depths inside the cloud and this conveys certain information on the DER vertical profile (DVP). However, this information is insufficient to retrieve any shape of DVP and thus a linear DVP was assumed. In this study, three DVPs are examined: (1) as in Chang and Li [2002], a linear DVP proportional to the in-cloud optical depth, (2) a linear DVP proportional to the height within the cloud, and (3) a DVP where the liquid water content (LWC) within the cloud varies linearly with height. The latter two assumptions are in closer conformity with in-situ observations. Algorithms that can retrieve both the DVP and cloud liquid water path (LWP) are presented. The cloud LWPs derived based on the retrieved DVPs are more sound than those obtained from assuming a vertical-constant DER profile. To enhance the DVP retrievals, a split-window technique is presented to better estimate the amount of above-cloud precipitable water (PW). The retrieval algorithms are applied to the MODIS Level-1B 1-km data and presently tested for two stratiform cloud cases observed over the north-central Oklahoma where independent cloud microphysics data are available from the United States Department of Energy's Atmospheric Radiation Measurement (ARM) Program. Good agreements in the retrieved DER profile, LWP, and above-cloud PW are found in a preliminary demonstration of the new approach. Sensitivity of the retrieved DER profile to uncertainties in the above-cloud PW and surface albedos is also discussed.
