Title: The Additional Value of Ensemble Weather Forecasts to Flood Forecasting: First Results on EFAS Forecasts for the Danube River Basin
Citation: Proceedings of the 23rd Conference of the Danube Countries on the Hydrological Forecasting and Hydrological Bases of Water Management p. 1-12
Publisher: International Hydrological Programme IHP of UNESCO
Publication Year: 2007
JRC N°: JRC34054
URI: http://www.hidmet.sr.gov.yu/slike_eng/hidro_konf/CONFERENCE%20ABSTRACTS.pdf
Type: Contributions to Conferences
Abstract: Over the past decades, severe floods in major transnational river basins have taken place in Europe, including the Rhine/Meuse floods in 1993 and 1995, the Oder flood in 1997, the Po floods in 1994 and 2000, as well as the Elbe and Danube floods in 2002 and, more recently, in spring 2006. Following the disastrous floods in the Elbe and Danube river basins in August 2002, the European Commission launched an activity to support the development of a European Flood Alert System (EFAS) to increase preparedness for large flood events. EFAS has been in development and testing phase since 2003 at the Joint Research Centre in close collaboration with the hydrological services in the Member States. The system is designed to provide additional information on potential flood situations by means of early flood warnings between 3 to 10 days in advance (http://efas.jrc.it/). This paper focuses on the additional value of weather ensemble prediction systems (EPS) to EFAS flood forecasts in the Danube river basin. Meteorological ensemble predictions from the European Centre for Medium-range Weather Forecasts (ECMWF) have been recently introduced pre-operationally in EFAS. The ECMWF-EPS consists of 51 10-day weather forecasts issued twice a day at 00:00 and 12:00 with a resolution TL255L40 (~80 km, 40 levels) - www.ecmwf.int. Contrary to single deterministic flood forecasts, flood forecasts based on EPS allow to take into account the uncertainty on reaching flood alert levels due to the different 51 possible weather developments. The incorporation of ECMWF-EPS in the EFAS platform and the statistical investigation of the probabilistic results raise a new challenge: how to quantify uncertainty and how to visualize the results in terms of probabilities of exceeding alert levels in a simple and concise way for a useful and correct statistical interpretation of the forecasted situation? EFAS EPS-based forecasts are expected to take into account the uncertainty of weather developments along the forecast range and therefore to provide additional and useful information on possible flood situations when increasing the lead-time. The first results obtained from the incorporation of ECMWF-EPS into EFAS during 2005 are presented and the tools implemented to quantify the uncertainty in exceeding EFAS alert levels in EPS-based forecasts are introduced. Maps and diagrams were developed allowing a concise and useful visualization of probabilistic results. The additional value of EPS to flood forecasting is illustrated from the investigation of case-studies related to observed flood events in the Danube river basin. The results show that EPS-based forecasts can contribute to an earlier detection of the possibility of flooding and therefore to increase preparedness for future flood events. In individual case-studies, a gain in lead-time, comparatively to the use of deterministic forecasts only, is usually observed, more or less important according to the forecasted flood event. The investigation of individual case-studies provide useful insights into the behavior of EPS-based forecasts, as well as essential information to the analysis of long-term reforecasts and the assessment of the average quality and skill of EFAS EPS-based ensemble forecasts.
JRC Institute:Institute for Environment and Sustainability

Files in This Item:
There are no files associated with this item.

Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.