• Flood hazard maps indicating up to 10 days ahead if certain discharge thresholds are exceeded. The rainfall-runoff simulations use several medium-range weather forecasts from the Deutsche Wetterdienst (DWD) and the European Centre of Medium-Range Weather Forecasts (ECWMF) as input.
• Maps with simulated flood event probabilities up to 10 days ahead based on full sets of the Ensemble Prediction System (EPS) from ECMWF.
• Consistent and comparable discharge simulations for entire trans-national catchments and for the whole of Europe.
The hydrological model employed in EFAS is the physically based rainfall-runoff model LISFLOOD (de Roo, 2000) that has been developed at the JRC specifically for the simulation of floods in large river catchments. The EFAS prototype foresees, however, open architecture so that any other model capable of simulating rainfall-runoff processes for entire catchments can also be plugged into the system.
There are essentially three different types of input:
• Static data that describe the river basin including topography, land-use, soil type and depth, geology, and river geometry
• Observed data to calculate the initial conditions of the flood forecasts, e.g. meteorological data, discharge data, etc.
• Medium-range weather forecasts to drive the flood model – weather forecasts over time periods longer than 3 days.
The simulated discharge forecasts are compared against threshold values that have been derived beforehand from long-term waterbalance simulations – using the same model and the same parameter sets. If the simulated discharges are exceeding these thresholds EFAS indicates the concerned river stretches.
EFAS is set-up for the whole of Europe on a 5 km grid for all river basins larger than 2000 km2 and on a 1 km grid for the Elbe and Danube river basins. Most data to define the river basin such as topography, land-use, soil type and depth, and river drainage system are available at the JRC. High-resolution river data to define cross sections or structural measures are being collected for the 1 km set-ups but not included in the 5 km set-up. The temporal resolution of the model is 1h. At present the 5 km set-up has been coarsely calibrated using observed data, and detailed calibration and validation studies for the 1km river basins are ongoing.