Limitations of microphysical property retrievals in marine water clouds based on satellite-based passive remote sensing

In satellite-based passive remote sensing, the cloud droplet effective radius (re), which has a strong influence on cloud radiative effects, is derived together with cloud optical thickness (τ) from reflectance measurements at two wavelengths, so called bi-spectral method. Since the bias to the re of marine water clouds can lead to a global radiative forcing error, it is critical to identify the sources and magnitudes of errors in the retrieval. Due to the use of loop-up tables, a bi-spectral method has limitations in optical and microphysical properties retrieval. It might be problematic for optically very thin clouds, and also erroneous for the clouds having very large radius droplets or large thickness due to the upper limits of retrieval method. This study investigates the occurrence rate and sources of the failed effective radius retrievals on Moderate Resolution Imaging Spectroradiometer (MODIS) onboard Aqua satellite together with Cloud Profiling Radar (CPR) observation on CloudSat satellite. First, a typical failed retrieval case will be shown to identify the characteristics. And, we investigate potential reasons of failed retrieval by demonstrating the dependency on the relevant parameters, such as optical thickness, horizontal homogeneity and reflectances. The difference between collections of MODIS data set will also be investigated. The CPR measurement collocated with MODIS can help to identify the source of failed retrieval, especially for the large particle. Preliminary result shows that the significant fractions (about 15%) of water clouds are failed in the MODIS re retrievals.