摘要
数据密集研究范式主导的数据水文学正在成为水文研究的一个重要方向,而善于从大量数据中挖掘规律的深度学习理论推动了近年来数据驱动水文预报的研究热潮,并不断和水文学科融合,逐步成为数据水文学的重要研究方法体系。从深度学习与水文预报学科交叉的角度,简要介绍水文领域常用深度学习模型的原理与结构及其应用于水文预报中的一般建模方法,在此基础上进一步介绍深度学习与水文物理机制整合的基本方法,以期为相关研究人员开展深度学习水文预报研究提供有益参考。
Data hydrology,led by the data-intensive research paradigm,is becoming an important research direction of hydrology.Deep learning,which performs well in extracting information from big data,has contributed significantly to the advancement of data-driven hydrological forecasting in recent years.It is continuously integrated with hydrology and becomes an indispensable tool in data hydrology.From a transdisciplinary view,the principle and structure of deep learning models commonly used in hydrology and the general workflow when applying them in hydrological forecasting are introduced.In addition,several approaches for integrating deep learning with domain knowledge in hydrology are described.A useful reference for researchers who are interested in exploring deep learning on hydrological forecasting provided.
作者
欧阳文宇
叶磊
王梦云
孟子文
张弛
OUYANG Wenyu;YE Lei;WANG Mengyun;MENG Ziwen;ZHANG Chi(School of Hydraulic Engineering,Dalian University of Technology,Dalian 116024,China)
出处
《南水北调与水利科技（中英文）》
CAS
北大核心
2022年第4期650-659,共10页
South-to-North Water Transfers and Water Science & Technology
基金
国家自然科学基金项目(51925902,52179006,92047302)。
关键词
数据水文学
深度学习
水文预报
模型概念
应用方法
物理机制
data hydrology
deep learning
hydrological forecast
model concept
applying method
domain knowledge
