Rain Measurement in Hilly Terrain with X-Band Weather Radar Systems: Accuracy of Path-Integrated Attenuation Estimates Derived from Mountain Returns

Abstract
The authors recently showed that when attenuating wavelengths are used mountain returns may allow estimation of path-integrated attenuations (PIAs) between a ground-based weather radar and a given mountain, an application of the well-known Surface Reference Technique originally proposed for spaceborne radar configurations. This information proved to be valuable for the quantitative interpretation of X-band weather radar data in terms of rainfall rate for an urban hydrological application in Marseilles, France. In this paper, a further verification of this concept is presented with the comparison of mountain-derived PIAs and direct measurements obtained by means of a receiving antenna installed in the Balcons de Belledonne mountain ridge near Grenoble, France. Maximum PIAs in the range of 8–16 dB are obtained over the considered 9-km propagation path for various rain events observed between May and July 1997. A physical model of the mountain return power is developed leading to the formulation of two mountain PIA estimators under various hypotheses concerning 1) the stability of both the radar equipment and the electromagnetic properties of the mountain surfaces and 2) the effects of the rain falling over the mountain. A geometric calculation based on the use of a digital terrain model then allows the authors to estimate both the radar-positioning errors and the rain beamfilling factors of the mountain-cluttered radar bins. The dry-weather mountain return time series are also studied, showing a good stability of the average value from one rain event to the next with, however, a marked effect of the wetting of the mountain surfaces. The time variability of these values is also characterized in order to assess the minimum detectable mountain PIA, estimated in the present case to be about 2.25 dB. Finally, a good agreement is observed between the estimated mountain PIAs and the measured ones (with correlation coefficients and regression slopes reaching 0.92 and 0.95, respectively). Assuming that the receiving antenna measurements are error free, the standard error in the mountain PIA estimation is found to be about ±2.5 dB in the considered configuration.