A prototype for automated wind gust warnings

In 2022, the German Meteorological Service (DWD) started the program RainBoW (“Risk-based, Application-oriented and INdividualizaBle delivery of Optimized Weather warnings”) aiming at a rework of the warning system. Among others, one important goal of the program is to extend the forecast horizon of warnings in order to inform authorities and the general public early-on about possible hazardous weather. This so-called “warning trend” is envisioned for all warning parameters, but has been explored first for wind gusts. In this contribution, the outline of the prototype for wind gusts, that is already available and running in a test mode, will be presented.

One of the main features of the prototype is its modularity, i.e. it consists of several different components which are tied together with modern messaging techniques. The wind gust prototype is based on data of DWD's inhouse model ICON, but is easily extensible to other data sources. It combines ensemble data from all ICON model versions to allow the construction of a warning trend up to 7 days. This comprises the local area version for Germany (ca. 2 km spatial resolution and forecast up to 48 hours), the EU version (ca. 13 km spatial resolution and forecast up to 120 hours) and the global version (ca. 26 km resolution and forecast up to 180 hours). After spatial and temporal aggregation of the individual model data, exceedance probabilities for given warning thresholds are computed and combined into a single, coherent warning trend dataset. Since this warning trend dataset changes with each new model run, an additional post-processing step is necessary to reduce the inherent jumpiness. Overall, the prototype thereby provides a smooth, automated wind gust warning trend.