水道港口2025，Vol.46Issue(4)：489-497,9.
基于数据驱动的长江口潮位预报方法研究
Study on prediction approach of tidal levels in the Yangtze estuary using data-driven algorithms
摘要
Abstract
The estuarine area,as a land-sea interaction zone,is vulnerable to floods.Its upstream river discharge and downstream tide interact strongly in tidal reach,resulting in estuarine tides with strong spatial and temporal heterogeneity.Therefore,it is difficult to predict the occurrence and development of estuarine tides.Because of the long tidal reach,the Yangtze estuarine tides formed by the interaction between river discharge and tides are highly representative and serve as an important source of regional floods.Accurate prediction of the Yangtze estuarine tides is an important basis for the supporting regional flood control and disaster reduction.In this paper,with the combination of the NS_TIDE and the data-driven Auto-regressive(AR)models,a short-term prediction study was carried out for the Yangtze estuarine tides.The method to obtain the upper and lower boundary information required by the NS_TIDE model in prediction was proposed.The NS_TIDE model was used to carry out basic prediction,and then the AR model was used to correct the short-term error of the prediction results of the NS_TIDE model,forming the prediction results of the hybrid model(NS_TIDE&AR).The test results in 2020 show that the accuracy of the 24-hour short-term prediction is between 0.08 m and 0.12 m,and the model is still robust even during the severe flood of that year.关键词
长江口/感潮河段/河口潮汐/数据驱动/非稳态调和分析/自回归模型Key words
Yangtze estuary/tidal reach/estuarine tide/data-driven algorithms/NS_TIDE/AR分类
交通工程引用本文复制引用
甘敏,陈永平,赖锡军,闻云呈,夏明嫣,储鏖..基于数据驱动的长江口潮位预报方法研究[J].水道港口,2025,46(4):489-497,9.基金项目
水利部水旱灾害防御重点实验室开放研究基金项目(KYFB202308140246) （KYFB202308140246）
国家自然科学基金项目(42171012) （42171012）
中国科学院南京地理与湖泊研究所自主部署科研项目(NIGLAS2022GS07) （NIGLAS2022GS07）
江苏省水利科技项目(2023046,2023021) （2023046,2023021）
