Verification of reconstructed historical extreme precipitation events in an hourly resolution

0510863 - ÚFA 2020 DE eng A - Abstrakt
Bližňák, Vojtěch - Kašpar, Marek - Müller, Miloslav - Zacharov, Petr, jr.
Verification of reconstructed historical extreme precipitation events in an hourly resolution.
10th European Conference on Severe Storms (ECSS 2019). Münchene: European Severe Storms Laboratory e.V. (ESSL), 2019.
[European Conference on Severe Storms (ECSS) /10./. 04.11.2019-08.11.2019, Kraków]
Institucionální podpora: RVO:68378289
Klíčová slova: heavy precipitation * historical precipitation events * retrospective forecast * numerical weather prediction * COSMO model
Obor OECD: Meteorology and atmospheric sciences
https://meetingorganizer.copernicus.org/ECSS2019/ECSS2019-66.pdf

Extreme precipitation events (EPEs) represent one of the main natural hazard in Central Europe, often accompanied
by extensive flooding that can result in loss of human life and costly damage to property, infrastructure, and
agriculture. Although more or less recent EPEs are well documented by rain gauges and/or weather radars, an
observation of historical EPEs in a high spatial and temporal resolution does not exist at all. The contribution
will present an application of the sub-daily temporal reconstruction method to the historical convective EPEs that
occurred over the Czech Republic (CR) in the second half of the 20th century. The method requires precipitation
re-forecasts (retrospective forecast) generated by a numerical weather prediction (NWP) model in a high temporal
resolution along with daily rain gauge measurements. Recently, the method employing re-forecasts of the NWP
model COSMO with a spatial resolution of 2.8 km and output temporal step of 10-minutes has been successfully
verified with adjusted radar-derived precipitation estimates for several recent EPEs, where radar observations were
available. In the presented contribution, the application of the method will be enlarged on the historical EPEs with
an insufficient spatial and temporal observation. For these events, the reconstructed 10-min precipitation will be
accumulated in 1 hour and verified with 1-h precipitation totals observed by pluviographs. Because their number is
very limited and their distribution over the CR is sparse (especially for more historical events), an employment of
their interpolated values would bring additional uncertainties. From this reason, we will verify point measurements
with gridded precipitation re-forecasts taking into account information from neighbouring pixels (e. g., enlarging
evaluation window over a given grid box). Reconstructed EPEs will then serve as a tool to evaluate their sub-daily
extremity from historical perspective.
Trvalý link: http://hdl.handle.net/11104/0301233