Enhancing the usability of weather radar data for the statistical analysis of extreme precipitation events

Abstract. Spatially explicit quantification on design storms is essential for flood risk assessment and planning. Due to the limited temporal data availability from weather radar data, are usually estimated basis of rainfall records a few precipitation stations only that have substantially long time coverage. To achieve regional picture, these station-based estimates spatially interpolated, incorporating large source uncertainty due typical low station density, in particular short event durations. In this study we present method estimate with return period up 100 years statistically extended estimates, based ideas frequency analyses subsequent bias correction. Associated uncertainties quantified using an ensemble-sampling approach event-based bootstrapping. With resulting dataset, compile various periods durations federal state Baden Württemberg, Germany. We compare our findings two reference datasets interpolated estimates. find transition spatial patterns rather random (short-duration events, 15 min) more structured, orographically influenced pattern (long-duration 24 h) seems be much realistic weather-radar-based product. However, absolute magnitude storms, although bias-corrected, still generally lower product, which should addressed future studies detail.