Determination of 3-D Cloud Ice Water Contents by Combining Multiple Data Sources from Satellite, Ground Radar and Numerical Model

This study aims at determining the 3-dimensional distribution of ice water content over a broad area near the Atmospheric Radiation Measurement Southern Great Plain site, where cloud radar and meteorological observations have been routinely conducted. Together with wind field from other measurements, the ice water content retrievals can be used to derive cloud ice water advective tendency terms needed for single-column model simulations. In this study, a Bayesian retrieval algorithm has been developed, which combines multiple data sources from satellite high-frequency microwave radiometry and ground cloud radar observations, and mesoscale numerical model analysis. The cloud radar observations allow us to infer the characteristics of vertical ice water content structures. The numerical model data are used to locate the cloud height, and the satellite data provide the information on the integrated ice water path, its horizontal distribution over a broad area, and to a lesser extent, the vertical structure of ice water content. Our approach is to retrieve the 3-dimensional cloud ice water content in a 10° by 10° area surrounding the cloud radar site by combining all the information contained in the above datasets through a Bayesian framework. Validation of the algorithm has been done by comparing the retrievals with measurements of two ground radars. The comparison results show that the mean ice water content profiles and the 2-dimensional (height-ice water content) probability density functions retrieved for 19 coincident cases agree fairly well with validation data. However, the retrieved ice water contents are generally lack of detailed vertical structures due to the low sensitivity of satellite data to the vertical variation of cloud ice.