A 95 GHz airborne cloud radar: Statistics of cloud reflectivity and analysis of beam-filling errors for a proposed spaceborn cloud radar
The Microwave Remote Sensing Laboratory (MIRSL) at the University of Massachusetts, in collaboration with The Jet Propulsion Laboratory, has developed a new radar designed to facilitate measurements of the radiative properties of clouds. Details of this new design are described with particular emphasis on improvements from previous systems.
This radar system was used to collect cloud data during three experiment campaigns. During these experiments, reflectivity data from all prevalent cloud types were collected over a wide range of geographical locations. The observations were then used to examine the reflectivity vs. altitude and temperature characteristics and layer structure for various types of cloud complexes. To increase the representation of tropical cirrus clouds in the composite data set, the airborne data was supplemented with data from MCTEX collected by the UMass CPRS radar.
All observations were classified into four classes of clouds and histograms of altitude and temperature vs. reflectivity were used to demonstrate the reflectivity characteristics of various clouds types. Statistics of layer base and top altitudes, thickness and number of layers were also computed. Also, the relationship between cirrus cloud thickness, reflectivity and ice water path (IWP) is examined.
The data sets from the four experiments were then used to address performance issues for a spaceborne radar. The problem of cloud detectability is discussed and an analysis of the ice water content (IWC) estimation error resulting from spatial inhomogeneity is presented. The fraction of clouds thinner than one range gate of the CloudSat radar was found to be 14% for all data sets combined. The data sets are used to simulate satellite radar pixels and the distributions of errors in IWC estimates due to inhomogeneity are calculated. On average, 40% of the pixels were partially filled and the relative IWC error was 24%. The distribution of the relative errors vs. IWC values indicated that the largest relative error occurred at vary small values of IWC and the mean error for all experiments was only 15% for IWC values larger than 10 –3 gm3.
0799: Remote sensing