Using medium term reflectivity statistics in radar QPE

The increasing availability of radar data on-line and from archive has made it much easier to collect and analyze medium to long term statistics of data from the radars in networks. Statistics of reported reflectivity as a function of range and azimuth are showing value in several areas of radar QPE. Most obviously, the statistics show the effect of partial and complete blocking by terrain and features on the terrain. The statistics often show subtle obscuration problems that were not predicted in calculations based on digital terrain maps and allow the development of better blockage maps. The distributions of reflectivity as a function of range has also shown hardware problems with individual radars in the Canadian weather network. For example, developing problems with TR cells show a distinctive pattern. Finally, the distributions of reflectivity show the interaction between the vertical profiles of precipitation and the minimum detectable signal (MDS) level of each radar. Kitchen and Jackson (1993), indicated that a complete failure to detect precipitation was a significant source of rainfall underestimation at long range, but the problem varies by radar and meteorological situation. Using statistics of observed reflectivity for a particular situation it may be possible to identify a maximum range beyond which the probability of detection is so low that radar QPE is essentially non-recoverable.