Dr. Bart Geerts, UWyo
The Precipitation Radar (PR) aboard the TRMM satellite measures in-cloud echo strength ('reflectivity'), from which rainfall rates are estimated. This instrument is accurately calibrated, but its resolution is too course to sample the true fine-scale variation of rainrates, especially in the case of deep convection. Also, the PR, by design, cannot detect very weak echoes, with reflectivities less than about 18 dBZ.
In this talk we show how higher-resolution and more sensitive radars, both airborne and ground-based, can be used to assess how well the PR can 'see' precipitating systems of various sizes, types and intensities. The cornerstone of this study is the ER-2 Doppler radar (EDOP), which looks down onto storms, as does the PR. The resolution volumes of the PR and EDOP look like slices of bread that are more or less level. In the case of a ground-based radar, the slices are standing upright. And EDOP's resolution and sensitivity allow the detection of such details as fallstreaks and thunderstorm anvils.
To allow a comparison, PR and ground radar reflectivities are remapped to the cross sections of the EDOP observations, in the case of a stratiform rainband, convection of various sizes, and a hurricane. The EDOP and ground radar data were collected as part of several TRMM Ground Validation campaigns, in particular TEFLUN (in Texas and central Florida), CAMEX-3 (targeting hurricanes), and TRMM-LBA (in the Amazon Basin). Comparisons of this kind are not trivial, (1) because TRMM overpasses, ER-2 flights and meteorological targets of interest rarely occur simultaneously, and (2) because the various radars use different coordinate systems and geolocation techniques.
In general it appears that the PR observations compare very well to high-resolution data. Both the horizontal and the vertical structures of storms are captured quite well, as are the derived products such as rain type and bright band presence/height. The PR storm height is 1-2 km below the actual storm top, as can be expected from the PR's limited sensitivity.
However echoes that are small relative to the PR footprint of 4.3 km are underestimated in strength and overestimated in size. This is a natural consequence of the beam filtering. Such small echoes are generally convective, yet because of their apparent weakness the PR may classify them as stratiform. Also, limited resolution and sensitivity may compound to significantly underestimate the storm top height of relatively small storms. Such storms, and sharp reflectivity gradients, are common in the tropical atmosphere, therefore users of TRMM products (especially the level 3 'climatological' products), should be aware of this scale dependency and its inherent limitations on rainfall estimation, rain type classification and other PR-derived variables.