Hybrid Modelling for Flood Forecasting in Hydroinformatics Context

摘要

Flood forecasting, is normally based on behaviour - driven, or physically - based (process) models based on equations describing the behaviour of water bodies. Since recently models built on the basis of large amounts of collected data are gaining popularity. This became possible only during the last decade due to the extensive measuring campaigns and remote sensing tools based on the developments in hydroinformatics. This modelling approach - data - driven modelling — borrows methods from various areas related to computational intelligence. Hybrid modelling refers to combinations of various approaches and types of models. Modular models each of which is responsible for a particular hydrological condition allow for increase in accuracy. An issue here is the inclusion of human experts into the process of model building. The paper presents some of the approaches to building data -driven and hybrid committees of models, presents an algorithm involving an expert in building such models, and reports the case studies in the area of flood forecasting in China, Nepal and Italy.