Short‐term rainfall forecasting using radar data

Abstract This paper describes the application of two methods for short‐term rainfall forecasting using radar data. The first method is based on the linear extrapolation of the centroids of features of rainfall cells, whereas the second method utilizes cross‐correlation techniques applied to the radar rainfall fields. Storm events from a region suffering from frequent and hazardous flash floods within the 120 km umbrella of the WSR‐74 S‐band radar located at the Larissa military airport in central Greece were analysed and processed. After the suppression of the ground clutter map, the subtraction of the errors due to anomalous propagation, beam refraction and losses, and the merging of radar and raingauge rainfall data, a multiple‐step‐ahead forecast was carried out by setting lead times 40, 80, 120 and 160 minutes ahead. It was found that both methods gave results of acceptable accuracy for the 40 and 80‐minute rainfall forecasting, while for the 120‐and 160‐minute forecasting the results were of relative...