Description:
<jats:title>Abstract</jats:title>
<jats:p>A new approach to passive microwave retrievals of precipitation is described that relies on an objective dimensional reduction procedure to filter, normalize, and decorrelate geophysical background noise while retaining the majority of radiometric information concerning precipitation. The dimensional reduction also sharply increases the effective density of any a priori database used in a Bayesian retrieval scheme.</jats:p>
<jats:p>The method is applied to passive microwave data from the Tropical Rainfall Measuring Mission (TRMM), reducing the original nine channels to three “pseudochannels” that are relatively insensitive to most background variations occurring within each of seven surface classes (one ocean plus six land and coast) for which they are defined. These pseudochannels may be used in any retrieval algorithm, including the current standard Goddard profiling algorithm (GPROF), in place of the original channels. The same methods are also under development for the Global Precipitation Measurement (GPM) Core Observatory Microwave Imager (GMI).</jats:p>
<jats:p>Starting with the pseudochannel definitions, a new Bayesian algorithm for retrieving the surface rain rate is described. The algorithm uses an a priori database populated with matchups between the TRMM precipitation radar (PR) and the TRMM Microwave Imager (TMI). The explicit goal of the algorithm is to retrieve the PR-derived best estimate of the surface rain rate in portions of the TMI swath not covered by the PR. A unique feature of the new algorithm is that it provides robust posterior Bayesian probabilities of pixel-averaged rain rate exceeding various thresholds.</jats:p>
<jats:p>Validation and intercomparison of the new algorithm is the subject of a companion paper.</jats:p>