Research on flood level forecasting in tidal river based on mixture regressive model

• Published in: 2021 2nd International Conference on Big Data & Artificial Intelligence & Software Engineering (ICBASE) pp. 363 - 371
• Main Authors: Shan, Lijie, Tang, Bin, Zhang, Yaolan, Zhou, Huan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2021
• Subjects: Adaptation models; Data models; Feiyun River Basin; flood level forecasting; Hydraulic systems; Mathematical models; mixture regressive model; predictive factor; Predictive models; Sea measurements; Statistical analysis; tidal river
• DOI: 10.1109/ICBASE53849.2021.00074

In order to solve the problem of complex hydrological situation and flood level forecasting of tidal river in coastal areas, a mixture regressive model was built based on the method of combining mathematical statistical analysis with physical causes. The predictive factor database was formed by flood processes which were calculated by using the runoff generation and concentration model for hydrological regionalization, discharge process of reservoir and tidal level process. On this basis, the mixture regression model was built by coupling stepwise multiple regression method and threshold regression model. The results show that, compared with the stepwise multiple regression method, the mixture regression model can effectively improve the accuracy of flood level forecasting in tidal river. The certainty factor and qualified rate of verification flood are 0.88 and 76.94% respectively. Mixture regressive model is a method of strong practicability and maneuverability in terms of simple principle and simple input conditions, which can provide reference to the work of the flood forecasting in coastal areas.