Evaluating the MODIS C6 multilayer cloud detection and phase algorithms through comparisons with CALIOP and CloudSat

Most heritage satellite imager-based global retrievals of cloud thermodynamic phase and optical/microphysical properties make simplified assumptions, e.g., that cloudy pixels are composed of single layer, plane-parallel clouds, while in nature clouds are much more complex. Because a field of view having multiple cloud layers represents a breakdown of the cloud retrieval forward model, discriminating pixels containing multilayer from monolayer cloud scenes is an important part of assessing retrieval quality. Here we report on an evaluation of the MODIS Collection 6/6.1 multilayer cloud classification and optical property thermodynamic phase algorithms using co-located CloudSat CLDCLASS-lidar and CALIOP 5km Cloud Layer products. While similar investigations have been performed in the past, this assessment appropriately considers the intent of the MODIS multilayer cloud identification algorithm design, namely to identify those cases in which multilayer clouds have radiative consequences that adversely affect cloud microphysical retrievals. Thus, our MODIS multilayer cloud identification comparisons with CloudSat/CALIOP are assessed in terms of several parameters such as the optical thickness of the upper cloud layer, the distance between the upper and lower cloud layers, the cloud thermodynamic phase in each cloud layer, etc.