Forecasting river levels during flash floods using Data Based Mechanistic models, online data assimilation and meteorological forecasts.

Smith, Paul and Beven, Keith (2010) Forecasting river levels during flash floods using Data Based Mechanistic models, online data assimilation and meteorological forecasts. In: British Hydrological Society International symposium, 2010-07-01.

Abstract

The parsimonious time series models used within the Data-Based Mechanistic (DBM) modelling framework are readily transferred into a State-Space form allowing the implementation of data assimilation using the Kalman filter. Multiple case studies have demonstrated the effectiveness of this framework in providing probabilistic forecasts for many hydrological situations, such as flood events on large UK rivers. The recent work presented here has applied the DBM methodology to forecast floods in catchments with a rapid response to rainfall. The resulting DBM models which, amongst other adaptations, utilise meteorological predictions to increase the forecast horizon are demonstrated using a European case study.