| 基于在线学习与误差修正的珠江河口区盐度预报 |
| |
| 引用本文： | 周凡涵,刘丙军.基于在线学习与误差修正的珠江河口区盐度预报[J].水文,2022,42(1):59-66. |
| |
| 作者姓名： | 周凡涵 刘丙军 |
| |
| 作者单位： | 中山大学土木工程学院;珠江水文水资源勘测中心;华南地区水循环与水安全广东省普通高校重点实验室 |
| |
| 基金项目： | 国家自然科学基金资助项目(51879289);广东省基础与应用基础研究基金联合基金重点资助项目(2019B1515120052);广州市水务科技项目(GZSWKJ-2020-2)。 |
| |
| 摘 要： | 受潮汐、径流、风速风向、地形变化等多种海陆要素交互作用,河口区盐水入侵呈高度不确定性与非线性特征,盐度预报难度较大.利用在线学习算法与误差自回归修正方法在水文预报中时效性更强的优点,构建一种耦合在线序列极限学习机-误差修正(OSELM-EC)盐水入侵预报模型,选取珠江河口区磨刀门水道为典型研究区进行逐日盐度预报.结果表...
|
| 关 键 词： | 盐水入侵 珠江河口区 误差修正 在线学习 实时预报 |
Estuarine Salinity Forecast in Pear River Based on Online Learning and Error Correction Methods |
| |
| ZHOU Fanhan,LIU Bingjun.Estuarine Salinity Forecast in Pear River Based on Online Learning and Error Correction Methods[J].Hydrology,2022,42(1):59-66. |
| |
| Authors: | ZHOU Fanhan LIU Bingjun |
| |
| Institution: | (School of Civil Engineering,Sun Yat-Sen University,Zhuhai 519082,China;Pearl River Hydrology and Water Resources Survey Center,Guangzhou 510370,China;Key Laboratory of Water Cycle and Water Security in Southern China of Guangdong High Education Institute,Zhuhai 519082,China) |
| |
| Abstract: | Due to the interaction of marine,terrestrial and meteorological factors,saltwater intrusion that often occurs in estuary presents significant uncertainty and nonlinearity,leading to the difficulty of estuarine salinity forecast.This study aims at developing a hybrid model coupling the online sequential-extreme learning machine(OSELM)and error correction(EC)for enhancing the accuracy and timeliness of salinity forecast in the Pearl River Estuary.The results show that OSELM-EC model could make full use of real-time data to effectively improve prediction accuracy.The longer forecast lead-time is,the more obvious superiority of OSELM-EC model is.The values of Nash-Sutcliffe efficiency(NSE)of OSELM model increase by 0.50%,10.73%and 18.67%respectively compared with traditional extreme learning machine(ELM)model with respect to 1,3 and 5 days forecast lead-time,and the NSE values respectively increase by 0.35%,4.55%and 16.54%compared with the single OSELM model. |
| |
| Keywords: | saltwater intrusion Pearl River Estuary error correction online learning real-time forecast |
| 本文献已被 维普 等数据库收录！ |
| 点击此处可从《水文》浏览原始摘要信息 |
| 点击此处可从《水文》下载免费的PDF全文 |
|
