^1. Key Laboratory of Remote Sensing of Gansu Province, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China ^2. School of Mathematics and Statistics, Lanzhou University, Lanzhou 730000, China ^3. College of Earth and Environmental Sciences, Lanzhou University, Lanzhou 730000, China ^4. Key Laboratory of Ecohydrology of Inland River Basin, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China

About author: HUANG Chunlin (1979-), male, Qingtongxia City, Ningxia Hui Autonomous Region, Professor. Research areas include hydrological remote sensing, multi-source remote sensing data assimilation, SDGs monitoring and evaluation. E-mail: [email protected] Supported by:

Data-driven methods with deep learning as their core have been gradually applied in Earth science; however, challenges remain regarding the interpretability of models and physical consistency. With the background of remote sensing big data, combining deep learning and data assimilation methods to develop new techniques for the simulation and prediction of terrestrial water cycle processes has become an important research direction in Earth science. Τhe progress in deep learning in recent years combines improving the quality of observation data of terrestrial water cycle components and reducing the uncertainty of physical models. Furthermore, the key scientific issues regarding data assimilation in terrestrial hydrology based on deep learning fusing remote sensing big data are classified according to the observations, physical models, and system integration: How can the temporal and spatial representativeness of samples be enhanced when deep learning inverts remote sensing products? How can a new physics-guided deep learning method be developed within the framework of data assimilation? How can the predictability of the terrestrial water cycle be improved through the “data-model” dual drive? Relevant research and exploration should help promote the in-depth application of the “data-model” hybrid modeling method in the field of hydrology and improve the simulation and prediction capacity of the terrestrial water cycle process.

Key words: Deep learning, Remote sensing big data, Data assimilation, Water cycle Fig. 1 Ten ways to apply machine/deep learning in the Earth and Space Sciences （ modified after reference ［
Each application direction is organized by the degree of involvement of physics-based models （horizontal scale） and the degree to which machine/deep learning codes are available and readily applicable （vertical scale） Fig. 1 Ten ways to apply machine/deep learning in the Earth and Space Sciences （ modified after reference ［
Each application direction is organized by the degree of involvement of physics-based models （horizontal scale） and the degree to which machine/deep learning codes are available and readily applicable （vertical scale） 刘元波, 吴桂平, 赵晓松, 范兴旺, 潘鑫, 甘国靖, 刘永伟, 郭瑞芳, 周晗, 王颖, 王若男, 崔逸凡. 流域水文遥感的科学问题与挑战 [J]. 地球科学进展, 2020, 35(5): 488-496. 黄婉彬,鄢春华,张晓楠,邱国玉. 城市化对地下水水量、水质与水热变化的影响及其对策分析 [J]. 地球科学进展, 2020, 35(5): 497-512. 李修仓,姜彤,吴萍. 水分再循环计算模型的研究进展及其展望 [J]. 地球科学进展, 2020, 35(10): 1029-1040. 李浩杰,李弘毅,王建,郝晓华. 河冰遥感监测研究进展 [J]. 地球科学进展, 2020, 35(10): 1041-1051. 谢正辉,陈思,秦佩华,贾炳浩,谢瑾博. 人类用水活动的气候反馈及其对陆地水循环的影响研究——进展与挑战 [J]. 地球科学进展, 2019, 34(8): 801-813. 汤秋鸿,刘星才,李哲,运晓博,张学君,于强,李俊,张永勇,崔惠娟,孙思奥,张弛,唐寅,冷国勇. 陆地水循环过程的综合集成与模拟 [J]. 地球科学进展, 2019, 34(2): 115-123. 马忠, 苏守娟, 龙爱华, 张晓霞. 塔里木河流域社会经济系统水循环分析 [J]. 地球科学进展, 2018, 33(8): 833-841. 刘娜, 王辉, 凌铁军, 祖子清. 全球业务化海洋预报进展与展望 [J]. 地球科学进展, 2018, 33(2): 131-140. 李育, 刘媛. 干旱区内流河流域长时间尺度水循环重建与模拟——以石羊河流域为例 [J]. 地球科学进展, 2017, 32(7): 731-743. 兰鑫宇, 郭子祺, 田野, 雷霞, 王婕. 土壤湿度遥感估算同化研究综述 [J]. 地球科学进展, 2015, 30(6): 668-679. 毛伏平, 张述文, 叶丹, 杨茜茜. 模式时间关联误差对集合平方根滤波估算土壤湿度的影响 [J]. 地球科学进展, 2015, 30(6): 700-708. 汤秋鸿, 黄忠伟, 刘星才, 韩松俊, 冷国勇, 张学君, 穆梦斐. 人类用水活动对大尺度陆地水循环的影响 [J]. 地球科学进展, 2015, 30(10): 1091-1099.