Evaluation of Ice Water Content Retrievals from Cloud Radar Reflectivity and Temperature Using a Large Airborne In Situ Microphysical Database

The objective of this paper is to assess the performances of the proposed ice water content (IWC)–radar reflectivity Z and IWC–Z–temperature T relationships for accurate retrievals of IWC from radar in space or at ground-based sites, in the framework of the forthcoming CloudSat spaceborne radar, and of the European CloudNET and U.S. Atmospheric Radiation Measurement Program projects. For this purpose, a large airborne in situ microphysical database is used to perform a detailed error analysis of the IWC–Z and IWC–Z–T methods. This error analysis does not include the error resulting from the mass–dimension relationship assumed in these methods, although the expected magnitude of this error is bounded in the paper. First, this study reveals that the use of a single IWC–Z relationship to estimate IWC at global scale would be feasible up to Ϫ15 dBZ, but for larger reflectivities (and therefore larger IWCs) different sets of relationships would have to be used for midlatitude and tropical ice clouds. New IWC–Z and IWC–Z–T relationships are then developed from the large aircraft database and by splitting this database into midlatitude and tropical subsets, and an error analysis is performed. For the IWC–Z relationships, errors decrease roughly linearly from ϩ210%/Ϫ70% for IWC ϭ 10Ϫ4 g mϪ3 to ϩ75%/Ϫ45% for IWC ϭ 10Ϫ2 g mϪ3, are nearly constant (ϩ50%/Ϫ33%) for the intermediate IWCs (0.03–1 g mϪ3), and then linearly increase up to ϩ210%/Ϫ70% for the largest IWCs. The error curves have the same shape for the IWC–Z–T relationships, with a general reduction of errors with respect to the IWC–Z relationships. Comparisons with radar–lidar retrievals confirm these findings. The main improvement brought by the use of temperature as an additional constraint to the IWC retrieval is to reduce both the systematic overestimation and rms differences of the small IWCs (IWC Ͻ 0.01 g mϪ3). For the large IWCs, the use of temperature also results in a slight reduction of the rms differences but in a substantial reduction (by a factor of 2) of the systematic underestimation of the large IWCs, probably owing to a better account of the Mie effect when IWC–Z relationships are stratified by temperature.