Improving knowledge on precipitation distribution during historical extremes using highresolution NWP model

0503746 - ÚFA 2020 CH eng A - Abstract
Bližňák, Vojtěch - Kvak, Róbert - Kašpar, Marek - Müller, Miloslav
Improving knowledge on precipitation distribution during historical extremes using highresolution NWP model.
UrbanRain18: 11th International Workshop on Precipitation in Urban Areas. Zurich: ETH Zurich, 2018.
[International Workshop on Precipitation in Urban Areas /11./. 05.12.2018-07.12.2018, Pontresina]
Institutional support: RVO:68378289
Keywords : NWP model simulation * extreme precipitation
OECD category: Meteorology and atmospheric sciences
https://www.ethz.ch/content/dam/ethz/special-interest/conference-websites-dam/urbanrain-dam/papers/UrbanRain18_ABSTRACT_Bliznak.pdf

Extreme precipitation events (EPEs) belong to the most studied natural hazards because of their
extra high impacts on the human society. However, observational data with a high temporal and
spatial resolution along with a sufficient level of accuracy is available for the last several years only
and is completely missing for historical EPEs.
The presented contribution will be aimed at the employment of high-resolution numerical weather
prediction (NWP) model simulations and daily rain gauge measurements to create a temporal
reconstruction of historical EPEs, which will enable to evaluate the selected EPEs in terms of their
sub-daily extremity. The EPEs are selected based on a return period of their maximal daily
precipitation. The selected EPEs will be simulated by the non-hydrostatic NWP model COSMO
with a spatial resolution of 2.8 km and temporal step of 10-minutes (raw NWP model re-forecasts).
The corrected 10-min NWP model re-forecasts will be then prepared in three steps. First,
interpolated precipitation sums from daily rain gauge measurements (RG) and the 24-h
precipitation sums forecasted by the NWP model COSMO (RF) will be used for the adjustment
procedure (Sokol, 2003; Bližňák et al, 2018) to obtain 24-h adjusted model precipitation sum
(RFA). The second step includes a calculation of the correlation coefficients between RG and RF
in a specific radius around a given grid box. The pixel with the highest value of correlation
coefficient determines the direction of the shift, which is applied for all 10-min raw NWP model reforecasts. Finally, shifted 10-min NWP model re-forecasts are then accumulated in 24 hour
precipitation sums, compared with RFA and obtained ratio is then used for multiplication of all 10-
min shifted NWP model re-forecasts.
Permanent Link: http://hdl.handle.net/11104/0295533