Rain measurement with SIR-C/X-SAR

Abstract

Rain measurement is needed over many parts of the world: oceans and less-developed areas without ground-based radars and extensive rain-gauge networks. The presence of several synthetic-aperture radars (SARs) imaging the Earth suggests that these may be used to get an idea of the rainfall in the needed areas. An experiment was conducted using SIR-C/X-SAR that flew in 1994 to determine how well we can measure rain with such systems. The experiment reported here involved measurements over the southwest Pacific Ocean and the Amazon rain forest. In the former case, a method was developed to measure the rain rate using the strength of the backscatter from rain. In the latter, we used estimates of the attenuation through the rain. Backscatter methods suffer from lack of knowledge of the appropriate ZR relations. In addition, for SARs one must estimate independently the rain height and turbulent velocity. Furthermore, some method must provide information on the back-ground “clutter” from the surface. Other problems arise in dealing with attenuation through the rain cell. Here we found reasonable rain rates from 10 mm/h to 200 mm/h in convective cells over the ocean. Without correction of the C-band calibration, an unreasonable peak rain rate of 375 m/s appeared, but we believe a correction is necessary. We present results in terms of two ZR relations and both with and without correction for C-band calibration. Some discrepancies between C-band and X-band measurements remain to be explained. The attenuation method can be used where obvious rain shadows occur, as over the Amazon. Here the difference between echo in the shadow and in the unshadowed background can be used to estimate rain from the attenuation. Fewer assumptions are needed for this method than for the backscatter method, and results appear more consistent between C-band and X-band. Rates from 77 mm/h to 128 mm/h were measured in storm cells. The results show that spaceborne SARs can measure rain, but the accuracy of the measurements is unknown. Accurate absolute calibration is necessary for the backscatter method, as well as sound estimates of surface clutter, storm height, and turbulent velocities. Further study is needed, particularly for the backscatter method.