Real-time data assimilation


The most important issue in real time flood forecasting is to allow for the fact that in any prediction event there will be departures between observations of water levels or discharges and predicted values. Any real time forecasting system should therefore allow for an updating or data assimilation strategy, with a view to minimising the uncertainties in the updated predictions. This will be possible in real applications as long as real time observations can be made available to the forecasting system, e.g. by telemetry. A number of data assimilation strategies are available, depending on the assumptions that the modeller is prepared to make. ''The distinction between this chapter and [Model Calibration and Conditioning Uncertainty on Available Data] is fuzzy. Some methods from [Model Calibration and Conditioning Uncertainty on Available Data] be also under this heading'' ''Methods'' * [Kalman Filter] * [Extended Kalman Filter] * [Ensemble Kalman Filter] * [Sequential Monte Carlo Methods]
Optional change note:
Upload a file or image:
For editing help, see HelpPage .

FloodRiskNet