Flash flood forecasts are generated using the methodology of the Enhanced Runoff Index based on Climatology (ERIC) [Raynaud et al., 2014]. It is calculated on a 1 km resolution river network whose extent matches the domain of the COSMO-LEPS forecast data.
The index is calculated based on the comparison of forecasted accumulated upstream surface runoff, UR, with the mean annual maxima from a 19 year climatology series, M, taken from COSMO-LEPS reforecast data [Fundel et al., 2010]. It is calculated for each of the ensemble members in the COSMO-LEPS forecast.
Surface runoff is calculated by multiplying forecasted precipitation data, taken from the COSMO-LEPS ensemble, with the corresponding soil moisture data produced by the LISFLOOD hydrological model [van der Knijff et al., 2012] driven with the COSMO-LEPS data. The accumulated upstream surface runoff is then calculated in every 1 km resolution river network pixel, j, where the upstream area is < 2,000 km2. This calculation is performed over three different accumulation periods, dk, of 6, 12 and 24 hours, the resulting ERIC value corresponds to the maximum over each of these three accumulation periods.
In line with other EFAS components, ERIC generates the probability for an event exceeding a 5-year or 20-year return period magnitude, which are derived from the model climatology.