Sensitivity of a Coupled Atmospheric-Hydrologic Probabilistic Forecast Scheme of South Asia Valleys

We present two case studies of forecast river discharge over the Indus and Ganges valleys for 2010 and 2011 summer seasons. Discharge is estimated utilizing meteorological forcings from the European Centre for Medium Range Weather Forecast (ECMWF) Ensemble Prediction System (EPS) coupled to macroscale hydrological parameterizations from the Variable Infiltration Capacity (VIC) model. The ECMW EPS was chosen based on its skill in predicting rainfall over the region up to 6-8 days in advance [Webster, Toma and Kim, 2011]. The use of EPS in hydrological and flood forecasting provides the end user with the opportunity of making decisions based on probability of occurrence of extreme events. This approach is becoming more common operationally and semi-operationally [Cloke and Pappenberger 2009]. Moreover, the assumed equi-probability of meteorological predictions may not be linearly translated into the hydrological probabilistic forecasts [Bartholmes et al. 2009], meaning that sensitivity analyses are critical for improvement of system uncertainties, so the use of the hydrological model in ensemble mode is important.

Flood warning systems are often initialized by binary decision regimes, i.e. forecast discharge persistently exceeding a maximum alert level at medium-range lead times (2-10 days). The probability of an extreme event is expressed based on the number of ensemble members exceeding a fixed level, or series of levels associated with increasing risk and developed based on historical forecasts (hindcast). We comment on the applicability of this approach in relation to extreme discharge events in South Asia and examine the skill of the forecast. When there is increased uncertainty in forecasting the meteorological parameters, an ensemble clustering based on large-scale atmospheric circulation patterns is also explored.

Cloke, H.L. and F. Pappenberger, 2009: Ensemble flood forecasting: A review, Journal of Hydrology, 375, 613-626.

Bartholmes, J.C., J. Thielen, M.H. Ramos, and S. Gentilini, 2009: The European flood alert system EFAS—Part 2: Statistical skill assessment of probabilistic and deterministic operational forecasts, Hydrology and Earth System Sciences, 13, 141-153.

Webster, P. J., V. E., Toma and H. M. Kim, 2011: Were the 2010 Pakistan floods predictable? Geophys. Res. Lett. 38, L04806, doi:10.1029/2010GL046346