ABSTRACT
This chapter presents the data-model integration framework. Data-model integration methods, e.g. data assimilation, can be selected and applied to construct additional members. In the data-model integration framework outlined, each data source is used to derive catchment parameter values or hydrometeorological input time series for a hydrological model. The proposed ensemble-based data-model integration framework may lead to extensive computing power requirements, especially with the use of a high resolution distributed model. In the Ensemble stage, all the simulation results, using multiple data sources and data-model integration methods, are wrapped in a multi-model ensemble simulation and the performance is assessed. Performance is assessed by comparing simulation results with observations of the output variables. To account for performance differences between ensemble members, weighting methods can be applied. Relative operating characteristic diagram is used to analyse the probabilistic performance of the ensemble simulation in identifying an event threshold exceedance.
