Towards a leadtime, scale and dynamical feature dependent postprocessing for wind gusts (C05)

与冬季低压系统相关的阵风是中欧最严重的自然灾害之一。在这里，我们将把最初为降水开发的空间验证方法应用于整体天气预报，以便将低压系统的空间（或时间）变化、相关的小尺度风特征以及大小和强度差异对预报误差和不确定性的贡献分开。剩下的误差主要是由局部条件、湍流和对流造成的，这些误差可以通过现代机器学习算法进行统计校正。

Wind gusts associated with wintertime low-pressure systems are amongst the most significant natural haz-ards in central Europe. Here we will adapt spatial verification methods originally developed for precipitation to ensemble weather predictions in order to separate contributions to forecast error and uncertainty from spatial (or temporal) shifts of the low-pressure system and the associated smaller-scale wind features, as well as size and intensity differences. Remaining errors will then mainly be due to local conditions, turbu-lence and convection, which can be statistically corrected training modern machine learning algorithms.