Title: Monthly-, Medium-, and Short-Range Flood Warning: Testing the Limits of Predictability
Authors: THIELEN DEL POZO JuttaBOGNER KonradPAPPENBERGER FlorianKALAS MilanDEL MEDICO MauroDE ROO Arie
Citation: METEOROLOGICAL APPLICATIONS vol. 16 no. Special Issue p. 77-90
Publisher: JOHN WILEY & SONS INC
Publication Year: 2009
JRC N°: JRC48226
ISSN: 1350-4827
URI: http://publications.jrc.ec.europa.eu/repository/handle/JRC48226
DOI: 10.1002/met.140
Type: Articles in Journals
Abstract: This paper explores the boundaries of predictability of floods by combining forecasts with multiple spatial and temporal resolutions. Monthly, medium- and short range numerical weather prediction data (NWP) are input to the European Flood Alert System for a flood event that affected rivers in Romania in October 2007. The NWP¿s consist of ensembles and deterministic forecasts of different spatial resolutions and lead times from different weather prediction models. Results are explored for the individual NWP components as well as for the ensemble. For this case study, ensembles of monthly weather forecasts contribute only marginally to the early warning, although some indication is given as early as three weeks before the event. The 15 day medium-range weather forecasts produce early flood warning information 9-11 days in advance. As the event draws nearer and is in range to be captured by the higher resolution ensemble forecasts, the spatial extent of the event is being forecasted with much more precision than with the medium-range. A novel post-processing method for the discharge is applied for those stations where observations are available, corrects time shifts and improves the quantitative forecasts. The paper illustrates how combination of forecasts and post-processing improves the lead time for early flood by 2-3 days while remaining reliable also in the short-range.
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.