Abstract
A method is proposed for calibrating radar estimates over the entire detection area using more than one conventional radar and raingauge datum. In this algorithm, two parameters are used to correct two major causes of errors in radar estimating, namely, (1) errors arising from the instability of the sensitivity of radar hardware, and (2) errors based upon the vertical profile of precipitation. By use of these parameters and the field of the height of the radar beam, a correction factor, which is an estimate of the ratio of the actual precipitation to the radar estimate on each grid of the precipitation field, is described with a simple function. The correction factors are determined hourly with the least-squares method by comparing radar estimates not only with raingauge measurements, but also with radar estimates from different radars. The method works well for the various types of precipitation, which change according to the nature of disturbances or to the season. The correction method is found to be accurate enough for operational use. By correcting radar estimates with this algorithm, the discontinuity found around the borders of domains for respective radars in the composite map is improved. This method is useful not only over land areas, but also over a wide range of the sea, where there is no other proper method for calibrating, especially in operational use. This algorithm is used as a part of the process for producing radar-AMeDAS precipitation charts, which the Japan Meteorological Agency provides operationally as fields of hourly precipitation amounts.
