Nowcasting Rainfall Fields Derived from Specific Differential Phase

Abstract

AbstractShort-term automated forecasting (nowcasting) of precipitation has traditionally been done using radar reflectivity data; recent research, however, indicates that using specific differential phase Kdp has several advantages over using reflectivity for estimating rainfall. This paper presents an evaluation of the characteristics of nowcasting Kdp-based rainfall fields using the Collaborative Adaptive Sensing of the Atmosphere Kdp estimation and nowcasting methods applied to approximately 42 h of X-band radar network data. The results show that Kdp-based rainfall fields exhibit lifetimes of ~17 min as compared with ~15 min for rainfall fields derived from reflectivity Zh in a continuous (cross correlation based) sense. Categorical (skill score based) lifetimes of ~26 min were observed for Kdp-based rainfall fields as compared with ~30 min for Zh-based rainfall fields. Radar–rain gauge verification showed that Kdp-based rainfall estimates consistently outperformed Zh-based estimates out to a lead time of 30 min, but the difference between the two estimators decreased in terms of normalized standard error with increasing lead time.