Climate Variability and Change—Hydrological Impacts (Proceedings of the Fifth FRIEND World Conference held at Havana, Cuba, November 2006), IAHS Publ. 308, 2006, 151–157.

---

Multi-model technique for low flow forecasting

MARY-JEANNE ADLER & VALENTINA UNGUREANU

National Administration “Apele Romane” – National Institute of Hydrology and Water Management, 97 Sos Bucuresti-Ploiesti, Sector 2, Bucharest, Romania

adler@hidro.ro

Abstract In forecasting low flows, four models, two determinist-statistic models (the VIDRA Model and a modified Stanford Model), a statistical model based on hydrometeorological analogues and an ARIMA class of model, were used to simulate the non-stationary and seasonal nature of the monthly flows. The result obtained was a large range of forecasted values which required a procedure in order to reach a decision on the most appropriate forecast. A solution was found in the development of a packet of procedures including both forecasting models and a decision procedure. This procedure is weight-based and uses the model errors on the last time step in its computing, so that with the help of model-weighting it is possible to incorporate in the procedure those models which best match the actual data (multi-model results optimization procedure).

Key words low flow forecast; multi-model procedure; forecasting decision support

---