Schlüsselwörter
(Englisch)
|
Meteorological simulation; hydrological model; heavy precipitation; flood forecasting
|
Alternative Projektnummern
(Englisch)
|
EU project number: ENV4-CT97-0552
|
Forschungsprogramme
(Englisch)
|
EU-programme: 4. Frame Research Programme - 3.1 Environment
|
Kurzbeschreibung
(Englisch)
|
See abstract
|
Weitere Hinweise und Angaben
(Englisch)
|
Full name of research-institution/enterprise:
Bundesamt für Meteorologie und Klimatologie
Bereichsleiter Unterstützung
|
Partner und Internationale Organisationen
(Englisch)
|
Università di Brescia, Dipartimento di Ingegneria Civile (DICBS);
CNRS, Université Paul Sabatier, Laboratoire d' Aerologie (LA), Toulouse-F; Universität München, Institut für Geographie und Geographische Fernerkundung (IGGF), München-D; Politecnico di Milano, Dipartimento di Ingegneria Idraulica, Ambientale e del Rilevamento (DIIAR), Milano-I; FISBAT-CNR, Bologna-I; VISTA - Remote Sensing Applications in Geosciences, Wessling-D; DLR-Institut für Physik der Atmosphäre, Oberpfaffenhofen-D; Geographisches Institut der ETH, Zürich-CH; Atmospheric Environment Service - Environment Canada (AES), Downsview-CND; University of Waterloo, Department of Civil Engineering (UW), Waterloo-CND
|
Abstract
(Englisch)
|
The objective of the project RAPHAEL is the coupling of meteorological and hydrological models to show the potential of such a system for improved flood forecasting. Two catchment areas have been selected as target areas for the project: The Ticino-Toce watershed in southern Switzerland/northern Italy (6599 km2) on the south side of the Alps, and the Ammer watershed (712 km2) southwest of Munich on the north side of the Alps. Episodes of a few weeks duration within the years 1993 to 1997 with strong precipitation and high runoff have been selected for atmospheric modeling.
The collection of meteorological data from existing archives as a basis for the hydrological modeling was coordinated by the Swiss Meteorological Institute (SMI). The RAPHAEL data base contains the necessary surface data for hydrological modeling and verification of meteorological models. To ease data exchange, SMI established common data formats for all RAPHAEL meteorological data files and the meteorological model output files. All available stations and observations are summarized in the meteorological data report.
The four groups running meteorological forecast models agreed on a simulation strategy for short periods of high precipitation, the so-called RAPHAEL events, which are time windows within the episodes. These strategies were designed in a way to produce continuous precipitation input for hydrological runoff prediction.
The hydrostatic operational numerical weather prediction model SM (Swiss Model) of the SMI was run for all events in both analysis and forecast mode. The conclusions from evaluating the simulations with a view for the hydrological applications are as follows:
The precipitation in both the operational forecast and analysis versions of the SM are in general quite good at the scale of the Ticino at Miorina (6599 km2) watershed. However, there is a tendency of the operational forecasts to overestimate the rainfall amount, especially with heavy precipitation peaks in the grid-scale part of the modelled precipitation. Slight underestimation of precipitation occurs in some cases in the analysis-driven simulation mode.
The location of the rainfall depends on the wind field, which in turn is influenced by the lateral boundary values of the driving model EM (Europa Model of Deutscher Wetterdienst). Therefore the general expectation is that the analysis-driven simulations be superior to the proper forecasts. Nevertheless there are cases where the forecast is better (Brig case).
The catchment average precipitation is sensitive to the location of the rainfall. This is already visible in the case of the medium-size Ticino at Miorina watershed (6599 km2). The error in average precipitation due to positioning errors grows with decreasing size of the catchment. The Ammer area is small compared to the grid size of the SM. A small spatial shift in the precipitation field can lead to large deviations of the average precipitation falling into the watershed. The Ammer watershed (712 km2) is at the lower limit of the scale of watershed sizes, for which meteorological model forecast with a resolution of 14 km can be sensibly used.
Relatively small changes in the overall precipitation were found when using a more sophisticated precipitation parameterization scheme which includes a prognostic variable for cloud ice. Little effect was detected with respect to precipitation distribution as well as to precipitation amount. This is probably due to the scales resolved by the SM with its 14km mesh, higher resolution models have exhibited more sensitivity. It turned out that the simulations are more sensitive to other parameters of the model configuration such as the number of vertical layers (20 versus 40 layers).
Evapotranspiration provides an important moisture supply for convective weather systems. Therefore it is expected that a good representation of this process as provided by the hydrological prediction models has a beneficial impact on the quality of precipitation simulation. However, in our cases it turned out to be of minor importance which indicates that the events investigated in RAPHEL were advectively dominated and less driven by local surface-atmosphere transfer fluxes.
The coupled meteorological-hydrological runoff forecasts provided some promising results although there is still a way to go to obtain reliable (flash) flood forecasts. The event-based results are presented and discussed by our hydrology partner group of ETH. The RAPHAEL project clearly indicated the potential of such coupled forecasts and that this methodology is the path to follow. Both, future developments in meteorological and hydrological forecast models and their foreseeable (partial) integration will increase our ability to forecast critical river runoff events.
|
Datenbankreferenzen
(Englisch)
|
Swiss Database: Euro-DB of the
State Secretariat for Education and Research
Hallwylstrasse 4
CH-3003 Berne, Switzerland
Tel. +41 31 322 74 82
Swiss Project-Number: 97.0069-1
|
