Identification of clutter echoes using a fuzzy logic technique

Non-meteorological targets introduce significant biases in the estimation of rainfall that limit the use of radar information for quantitative purposes. Moreover, the location, size and intensity of the echoes caused by these targets are not constant, but depend on the propagation conditions. It is thus necessary to discriminate meteorological echoes from clutter in order to improve not only precipitation estimates but to avoid negative effects on automatic algorithms based on radar information (such as nowcasting techniques). In this study an algorithm for clutter identification using fuzzy logic concepts has been developed. It is based on the analysis of some statistics that highlight clutter characteristics (low vertical extension, high horizontal variability, small Doppler velocities,...). Important differences between ground clutter and sea clutter echoes have been found (as already shown by Steiner and Smith 1999, their pdf curves are significantly different). Therefore, the developed algorithm will analyze echoes measured over land differently than echoes measured over the sea (different statistics are taken into account). The proposed technique has been implemented (a) over some selected radar scans corresponding to illustrative situations and (b) over a long series of around 15000 scans measured between 01 June 2003 and 30 September 2003. Results achieved using the proposed algorithm will be compared against the results obtained (a) using a mean clutter map and (b) from Doppler-processed data. This analysis is carried out in terms of accumulated rainfall fields and in comparison against the measurements of a network of rain gauges.