PubAg

Main content area

Sub-daily temporal reconstruction of extreme precipitation events using NWP model simulations

Author:
Bližňák, Vojtěch, Kašpar, Marek, Müller, Miloslav, Zacharov, Petr
Source:
Atmospheric research 2019 v.224 pp. 65-80
ISSN:
0169-8095
Subject:
atmospheric precipitation, radar, rain gauges, simulation models, weather forecasting
Abstract:
An alternative correction procedure for a posteriori improvement of quantitative precipitation re-forecast generated by a numerical weather prediction (NWP) model COSMO in a high temporal resolution is presented. The main motivation is to provide reliable precipitation re-analyses from the NWP model, which will enable a more detailed analysis of historical extreme precipitation events (EPEs). The procedure adjusts the model precipitation sums by daily rain gauge measurements and corrects the localization of 10-min model precipitation totals based on the highest correlation coefficient between the 24-h model and observed precipitation sums. The results show that the NWP model COSMO usually well predicts the occurrence of EPEs, but its spatial localization is not always accurate. This is usually a case of more localized convective precipitation, where the impact of the applied correction procedure is the most evident. A detailed temporal analysis including a calculation of correlation coefficient and Fractions Skill Score also confirmed that the re-forecast improvement is observed in a time when the highest precipitation within a given EPE occurs. Beyond this time, the effect of the correction is insignificant. In contrast, the impact of the correction is rather negligible in case the NWP model correctly locates precipitation re-forecast because uncorrected and corrected re-forecasts are very similar. Although the results stem from the seven selected EPEs only, the study provides a general performance of the proposed method and indicates that the method can be applied to historical EPEs, for which no weather radar data are available, to obtain their more accurate sub-daily temporal reconstruction.
Agid:
6336368