J.C. Bartholmes, E.Todini
Dipartimento di Scienze della Terra e Geologico-Ambientali, Università di Bologna Via Zamboni, 67 40126-BOLOGNA Tel. 051 2094537 Fax 055 2094522 e-mail jens@geomin.unibo.it
The project EFFS (European Flood Forecasting System) aims to develop a flood forecasting system for the major river basins all over Europe. To extend the forecasting and thus the warning time in a significant way (up to 10 days) meteorological forecasting data from the ECMWF (European Centre for Medium Weather Forecast) will be used as input to hydrological models. For this purpose it is fundamental to have a good rainfall-runoff model. For the river Po basin we choose the TOPKAPI model (Ciarapica, Todini 1998). TOPKAPI is a physically based rainfall-runoff model that maintains its physical significance passing from hillslope to large basin scale, enlarging the dimension of the raster cells in which the basin is subdivided for calculation purposes. There are two versions of this model : a distributed and a lumped one (the later is still in development). The aim of the distributed version - for the Po river basin we are working with 73140 rectangular cells of 1km*1km - is to reproduce the spatial variability and to lead to a better understanding of scaling effects on meteorological data used as well as of physical phenomena and parameters. This is needed to develop a relationship between saturated area and water stored in the topmost layer of the soil. This is needed as an essential input to the lumped version. By now the TOPKAPI model has been applied successfully to basins of smaller and medium size (up to 8000 km2). The present work will show its proves to be also a valuable tool for bigger basins such as the Po river. An advantage of the TOPKAPI model is its physical basis. It doesn't need a "real" calibration in the common sense of the expression. The few calibration work that has to be done is due to the unavoidable approximation in the input data representing various phenomena. This makes working on hourly timesteps possible, as for the calibration very few historical data is needed. Sufficient data in this resolution exists for the Po river only from 1994 . The data from groundbased meteorological sensors will only be used for the calibration period and the verification purposes, whereas later on only meteorological forecast data will be used as continuos input. Mesoscale forecast (from ECMWF) as well as LAM models (DWD,DMI) will be used and their behaviour will be studied.
Supplementary URL: