Evaluation of radar-based precipitation estimates during a flood event using rain gauge validation

This study uses dual-polarization radar data from two weather radars to assess the accuracy of estimating rainfall rate (R) from radar reflectivity (Z) using classical empirical Z–R relationships (Marshall–Palmer, Wexler, and Doumoulin–Cogombles). Three additional polarimetric relationships incorporating differential reflectivity (ZDR) were also analyzed. Calculations were carried out for seven altitude levels, representing a novel approach to verifying the vertical structure of the precipitation field. Radar data were validated against measurements from a rain gauge network, enabling an assessment of spatiotemporal consistency and rainfall estimation errors. The analysis covers measurements collected during the flood event that occurred in Poland in September 2024 (Genoa low-pressure system). The results indicate that for the analyzed extreme flood event, the most reliable precipitation totals were obtained using the polarimetric relationship based on ZDR (RMSE=2.20 mm±0.90 mm, MAE=1.84 mm±0.73 mm, and Bias=-0.67 mm±0.81 mm). It was demonstrated that under the studied extreme conditions, the proposed ZDR3 relationship exhibits approximately 69% lower Bias compared with the standard operational Marshall-Palmer method. The findings confirm the potential of polarimetric methods for rainfall estimation and their applicability in operational settings during similar extreme weather events, particularly in early warning and crisis management systems.