Investigating the spatiotemporal characteristics of motion fields using three‐dimensional radar echoes to construct an ensemble nowcasting system

Abstract In this study, a spatial and temporal analysis of three‐dimensional motion fields was conducted to formulate an ensemble nowcasting scheme. These motion fields were derived from radar data covering the entire volume and were analyzed across space and time for five different weather phenomena in Taiwan, namely, autumn precipitation, Meiyu front, squall line, afternoon thunderstorm and typhoon. Subsequently, perturbations will be introduced into the two‐dimensional motion field to develop an ensemble nowcasting framework. The magnitude of the perturbations will be adjusted based on the spatiotemporal analysis of the three‐dimensional motion fields. The effectiveness of the ensemble nowcasting in forecasting radar echoes and accumulated rainfall will be examined using a variety of verification scores. An evaluation of this ensemble nowcasting method demonstrated its efficacy in accurately predicting precipitation occurrences, with a reliability extending nearly three hours for five afore‐mentioned weather events. Furthermore, the forecasted accumulation of rainfall closely mirrored observed patterns, indicating the scheme's ability to capture precipitation distribution accurately. In summary, the ensemble nowcasting scheme mitigates uncertainties associated with motion field and exhibits superior forecasting capabilities compared with the original extrapolation method.