A new module for the tracking of radar-derived precipitation with model-derived winds

Abstract. A new approach for the nowcasting of precipitation has been developed at the German Weather Service combining extrapolation techniques and Numerical Weather Prediction (NWP) for a lead time range of several hours. Radar-derived precipitation fields serve as input data for a tracking algorithm using model-derived wind data. The composite precipitation field is derived from the precipitation scans which are performed every five minutes at the 16 German radar stations. The data are corrected from clutter and shading effects. The tracking of this radar-derived precipitation field is performed using the temporally and spatially resolved horizontal wind fields at different pressure levels provided by the Local Model Europe (LME). The optimal wind field is derived from minimization of the least-squares difference between a linear combination of model wind data from different pressure levels and the linear displacement vectors calculated via pattern recognition from previous radar measurements. An area-preserving displacement of the precipitation fields is realized by eliminating the wind field divergence and by omitting the dynamical evolution of the precipitation fields. Advection is performed using the fourth-order Bott scheme. Forecasted data comprise precipitation rates for every five minutes lead time as well as hourly sums of precipitation. The verification of a case study's results against radar precipitation measurements lead to a mean Equitable Threat Score (ETS) of 70%, 46%, and 38% for the first, second, and third forecast hour, respectively.