水文时间序列分析方法研究进展  

Research Advances on Hydrologic Time Series Analysis Methods

作　　者：李路 宫辉力[1,2,3,4,5] 郭琳[1,2,3,4,5] 朱琳 陈蓓蓓[2,3,4,5] LI Lu;GONG Huili;GUO Lin;ZHU Lin;CHEN Beibei(Beijing Laboratory of Water Resources Security,Capital Normal University,Beijing 100048,China;The Key Lab of Resource Environment and GIS of Beijing,Capital Normal University,Beijing 100048,China;Base of the State Key Laboratory of Urban Environmental Process and Digital Modeling,Capital Normal University,Beijing 100048,China;Key Laboratory of 3D Information Acquisition and Application,MOE,Capital Normal University,Beijing 100048,China;Key Laboratory of mechanism,Prevention and Mitigation of Land Subsidence,MOE,Capital Normal University,Beijing 100048,China)

机构地区：[1]首都师范大学水资源安全北京实验室,北京100048 [2]首都师范大学资源环境与地理信息系统北京市重点实验室,北京100048 [3]首都师范大学城市环境过程与数字模拟国家重点实验室培育基地,北京100048 [4]首都师范大学三维信息获取与应用教育部重点实验室,北京100048 [5]首都师范大学地面沉降机理与防控教育部重点实验室,北京100048

出　　处：《地球信息科学学报》2024年第4期927-945,共19页Journal of Geo-information Science

基　　金：国家自然科学基金项目(41930109,41771455,42371081);临沂市城市地质调查项目(SDGP371300202102000468);北京卓越青年科学家项目(BJJWZYJH01201910028032)。

摘　　要：水文时间序列分析领域的发展对于有效管理和利用水资源至关重要。本文基于WoS核心合集数据库和CNKI数据库,采用文献计量学方法和CiteSpace软件,揭示国内外水文时间序列分析领域的发展脉络、研究热点及发展方向。首先从水文时间序列的随机性、非线性、不确定性等出发,结合机器学习、神经网络等新兴方法,将水文时间序列分析领域的相关进展分为6个方面。然后,对各方面开展了详细介绍,并与传统方法作对比,总结了传统方法的缺陷;最后,指出提高水文时间序列分析结果准确性的方向,主要包括:(1)在时空尺度上进行建模,并融合多元数据进行分析;(2)将物理机制融入机器学习模型,提高模型的可解释性和泛化能力;(3)在研究过程中考虑气候变化(极端天气事件)和水文过程的耦合;(4)将多个复杂特性综合进行研究,并提高每个复杂特性的研究水平。通过揭示国内外水文时间序列分析领域的发展脉络、研究热点及发展方向,我们能够更好地理解和应对气候变化、极端天气事件以及人类活动对水资源的影响,提高我们对水文过程的认识,为水资源规划、水灾风险管理和可持续发展提供科学依据。The development of hydrologic time series analysis is crucial for the effective management and utilization of water resources.Based on the WoS Core Collection database and the CNKI database,this paper employs bibliometrics and CiteSpace software to reveal the development trends,research hotspots,and future directions in the field of hydrologic time series analysis both domestically and internationally.Firstly,starting with the randomness,nonlinearity,and uncertainty of hydrologic time series,as well as emerging methods such as machine learning and neural networks,this paper divides the recent advances in the field of hydrologic time series analysis into six aspects.Then,a detailed introduction for each advance is provided,and a comparison with traditional methods is also made to summarize the shortcomings of traditional methods.Finally,the directions for improving the accuracy of hydrologic time series analysis are pointed out,including:1)modeling at spatiotemporal scales and integrating multi-source data for analysis;2)incorporating physical mechanisms into machine learning models to enhance interpretability and generalization capabilities;3)considering the coupling of climate change(extreme weather events)and hydrologic processes in research advances;4)conducting comprehensive research on multiple complex characteristics and improving the research level of each complex characteristic.By revealing the development trends,research hotspots,and future directions of hydrologic time series analysis both domestically and internationally,we can better understand and respond to the impacts of climate change,extreme weather events,and human activities on water resources,enhance our understanding of hydrologic processes,and provide scientific basis for water resources planning,flood risk management,and sustainable development.

关 键 词：水文时间序列分析 CITESPACE 机器学习 非线性 不确定性 非平稳性 耦合 

分 类 号：P332[天文地球—水文科学]