Integrated microwave and infrared precipitation analysis.
Frailey, Lisa E.
Wash, Carlyle H.
Durkee, Philip A.
MetadataShow full item record
GOES infrared (IR) data is intercompared with rain analyses from the SSM/I exponential rain algorithm for the purpose of determining thresholds and statistics from IR imagery which delineate oceanic rain area. Data from ERICA cyclogenesis cases were evaluated. Discriminant analysis was performed using IR mean cloud top temperature, standard deviation and kurtosis as discriminating variables. Resulting functions separated rain from no-rain areas with average Probability of Detection (POD) and Percentage Error (ERR) scores of 0.68 and 0.30 for development data (0.62 and 0.37 for validation data). The scheme demonstrated little skill in discriminating rain categories beyond rain/no-rain. An IR threshold scheme was used to delineate rain/no-rain areas by optimizing a set of evaluation statistics. Optimal thresholds attained a predetermined POD level of 0.60 while minimizing percent misclassification error and SSM/I-IR rain area difference. The scheme yielded average POD and ERR scores of 0.64 and 0.38 with IR thresholds from 229 to 232 K.
Approved for public release; distribution is unlimited