Phantom Precipitation and Other Problems in Trmm Products

The Tropical Rainfall Measuring Mission (TRMM) satellite carries two instruments to measure precipitation: the TRMM Microwave Imager (TMI) and the Precipitation Radar (PR). Examination of instantaneous orbit data from TMI and PR reveals several problems in the TRMM products that are not readily discernible from monthly or areal averaged data sets. TMI rainfall processing algorithms yield significantly more oceanic rainfall area than the PR. Some of this areal difference is related to differences in minimum sensitivity and spatial resolution of TMI versus PR, but there are also cases where the TMI algorithms show “phantom precipitation” that is not observed by PR or ground-based radar. The TMI algorithms also produce cloud ice above the observed cloud top height for the phantom precipitation cases. Non-physical discontinuities are frequently observed in the TMI precipitation field along the boundary between coast and ocean surface types where the weighting of the TMI channels within the algorithm changes. The discontinuities in surface rain rate are often accompanied by non-physical TMI hydrometeor profiles showing precipitation ice in rain layers below the freezing level. These non-physical profiles indicate a problem with the selection of appropriate hydrometeor profiles from the algorithm database. Data from TMI and PR are compared on global and regional scales to identify the nature and impact of these TMI precipitation errors and the conditions under which the errors occur. It is important to identify errors and uncertainties in satellite estimates of global precipitation so that these can be taken into account in merged products and other applications of the data.