Assessing the Information Content of AIRS and MODIS Radiances for Retrieving Ice Cloud Properties

Accurate satellite cloud property retrievals are crucial for evaluating how well climate models represent clouds, which remain as the largest sources of uncertainty in today's climate projections. Improvements in the use of Atmospheric InfraRed Sounder (AIRS) and Moderate Resolution Imaging Spectroradiometer (MODIS) measurements are particularly valuable, as these instruments have accumulated more than a decade of coincident observations from which climatic variations in cloud properties may be extractable. This study utilizes two metrics of information, the Shannon information content and the Kullback-Leibler divergence, to assess the optimal AIRS/MODIS channel combinations for retrieving ice cloud properties over a wide range of cloud morphologies and environmental conditions such as surface type, atmospheric temperature and humidity, and the presence of aerosols. To evaluate how measurement information varies with cloud properties, multiple years of ice cloud data from ground-based radar/lidar cloud retrievals were objectively categorized with the K-means clustering algorithm. This yielded hundreds of ice cloud types that were used to generate both information metrics to identify the optimal channels for each. The optimal channel configuration for global ice cloud retrievals will be presented based on the product of the selected channels weighted by the occurrence frequency of the associated cloud type. In light of previous studies that have demonstrated the value of cloud boundary information from active sensors, we also examine how the information content and optimal channel configuration are modified when CloudSat or CALIPSO observations are included in the retrieval.