Machine Learning Downscaling Methods for Inversion of Terrestrial Water Storage Anomalies in the Yangtze River Basin by GRACE

LI Qiong; ZHANG Dong; ZHONG Bo; SU Yong; YANG Yiyu

doi:10.13203/j.whugis20240310

您当前的位置：

首页 >

文章列表页 >

Machine Learning Downscaling Methods for Inversion of Terrestrial Water Storage Anomalies in the Yangtze River Basin by GRACE

Geodetic Hydrological Inversion | 更新时间：2026-11-05

- Machine Learning Downscaling Methods for Inversion of Terrestrial Water Storage Anomalies in the Yangtze River Basin by GRACE
- Geomatics and Information Science of Wuhan University Vol. 51, Issue 6, Pages: 1138-1155(2026)
- 作者机构：
  
  1.西南石油大学土木工程与测绘学院，四川成都，610500
  2.武汉大学测绘学院，湖北武汉，430079
  3.地球空间环境与大地测量教育部重点实验室，湖北武汉，430079
  4.华中科技大学国家精密重力测量科学中心，湖北武汉，430074
- 作者简介：
- 基金信息：
- DOI：10.13203/j.whugis20240310
  CLC： P237
- Received：07 March 2025，
  
  Published：05 June 2026
- 稿件说明：
移动端阅览
李琼,张东,钟波，等.GRACE反演长江流域陆地水储量变化的机器学习降尺度方法研究[J].武汉大学学报（信息科学版）,2026,51(6):1138-1155.

LI Qiong,ZHANG Dong,ZHONG Bo,et al.Machine Learning Downscaling Methods for Inversion of Terrestrial Water Storage Anomalies in the Yangtze River Basin by GRACE[J].Geomatics and Information Science of Wuhan University,2026,51(6):1138-1155.
李琼,张东,钟波，等.GRACE反演长江流域陆地水储量变化的机器学习降尺度方法研究[J].武汉大学学报（信息科学版）,2026,51(6):1138-1155. DOI：10.13203/j.whugis20240310

LI Qiong,ZHANG Dong,ZHONG Bo,et al.Machine Learning Downscaling Methods for Inversion of Terrestrial Water Storage Anomalies in the Yangtze River Basin by GRACE[J].Geomatics and Information Science of Wuhan University,2026,51(6):1138-1155. DOI：10.13203/j.whugis20240310

摘要

重力恢复与气候实验（gravity recovery and climate experiment，GRACE）卫星时变重力场反演陆地水储量变化（terrestrial water storage anomaly，TWSA）的空间分辨率有限，制约了其在中小尺度区域水循环和气候变化研究中的应用潜力。当前，基于机器学习的降尺度方法已在提升GRACE TWSA空间分辨率方面取得了显著进展。然而，预测因子的合理选取及其对模型性能的影响机制，以及降尺度结果的精度评估与不确定性分析等问题，仍有待进一步深入研究。以长江流域TWSA的降尺度处理为例，对比传统的水文模型降尺度方法和人工神经网络（artificial neural network，ANN）、随机森林（random forest，RF）、极端梯度提升（extreme gradient boosting，XGBoost）3种机器学习模型，将GRACE反演的长江流域空间分辨率为1°×1°的TWSA数据降尺度到0.25°×0.25°和0.1°×0.1°。为了对不同降尺度方法进行评价，首先使用全球陆面数据同化系统水文模型的TWSA数据进行闭合模拟实验，评估不同降尺度方法的性能；然后对GRACE TWSA数据进行降尺度处理，并采用水文站水位数据评估不同降尺度方法的性能和结果。研究结果表明，机器学习降尺度模型的性能受到预测因子数量的影响，随着预测因子数量的增加，降尺度性能也会持续增强，3种机器学习模型的降尺度性能优劣也会发生改变。采用偏最小二乘回归分析模型得到重要性较高的前6个预测因子，即归一化植被指数、土壤湿度、降水量、气温、径流和U形风，在机器学习降尺度模型中取得较好的效果，增加更多的预测因子对降尺度模型性能的提升相对较小。RF和XGBoost降尺度模型的性能表现较好且非常接近，ANN和水文模型降尺度模型的效果略差。此外，水文模型的降尺度结果依赖水文模型与TWSA数据的相关性，机器学习降尺度模型能够更好地融合水文、气象和植被等多种数据的变化特征，从而恢复中小尺度区域精细的TWSA信号。

Abstract

Objectives

The spatial resolution of the time-varying gravity field model provided by gravity recovery and climate experiment(GRACE) satellite gravity for retrieving terrestrial water storage anomaly (TWSA) is limited

which restricts its application potential in the study of regional water cycle and climate change. The current machine learning downscaling methods have been effective in improving the spatial resolution of GRACE TWSA data

but further exploration is needed on the reasonable selection of predictive factors and their impacts on the performance of machine learning models

as well as the accurate evaluation of downscaling results.

Methods

A hydrological model downscaling method and three machine learning model downscaling methods

including random forest (RF)

extreme gradient boosting (XGBoost)

and artificial neural network (ANN)

were adopted to downscale the TWSA data obtained from GRACE inversion of the Yangtze River Basin with the spatial resolutions from 1°×1° to 0.25°×0.25° and 0.1°×0.1°

respectively. To evaluate different downscaling methods

a closed-loop simulation experiment was conducted using TWSA data from global land data assimilation system hydrological model to evaluate the performance of different downscaling methods. Subsequently

the TWSA data obtained from GRACE was downscaled

and the performance and results of different downscaling methods were comprehensively evaluated using measured water level data.

Results

The performance of machine-learning-models-based downscaling method is sensitive to the number of predictive factors. As the number of predictors increases

the overall downscaling performance improves

although the magnitude of improvement varies among models. Partial least squares regression analysis indicates that the six most important variables

normalized difference vegetation index (NDVI)

soil moisture

precipitation

temperature

runoff

and U-wind

are sufficient to achieve robust performance

while the inclusion of additional predictors yields only marginal gains. Closed-loop simulation experiments further reveal differences among models. Compared with RF and XGBoost models

ANN model shows relatively lower performance metrics; however

it produces the most favorable downscaled spatial patterns. When evaluated against water level observations in the Yangtze River Basin

TWSA remains consistent with observed water levels both before and after downscaling. In particular

RF-based results show a marked improvement in correlation

with all coefficients exceeding 0.7. Using Poyang Lake as a representative case

the long-term trends of the machine learning downscaling results were compared with those of the three predictors with the highest variable importance in projection scores

including NDVI

soil moisture

and precipitation. The results indicate that the downscaled outputs exhibit long-term trends highly consistent with NDVI and soil moisture

and their spatial variations closely correspond to the extent of water-covered areas in Poyang Lake. In contrast

precipitation does not display a similar trend pattern.

Conclusions

The best downscaling methods are RF and XGBoost

while the downscaling methods of ANN and hydrological model perform poorly. Additionally

the results of hydrological model downscaling depend on the correlation between the hydrological model and GRACE data. However

machine learning downscaling methods can better integrate the changing characteristics of different auxiliary data such as hydrology

meteorology

and vegetation (especially important predictive factors)

thus enabling better recovery of detailed TWSA signals in the watershed.

关键词

Keywords

references

HUMPHREY V ， RODELL M ， EICKER A . Using Satellite-Based Terrestrial Water Storage Data： A Review ［J］. Surveys in Geophysics ， 2023 ， 44 （ 5 ）： 1489 - 1517 .

SOLTANI S S ， ATAIE-ASHTIANI B ， SIMMONS C T . Review of Assimilating GRACE Terrestrial Water Storage Data into Hydrological Models： Advances， Challenges and Opportunities ［J］. Earth⁃Science Reviews ， 2021 ， 213 ： 103487 .

刘任莉，佘敦先，李敏，等 . 利用卫星观测数据评估GLDAS与WGHM水文模型的适用性［J］. 武汉大学学报（信息科学版）， 2019 ， 44 （ 11 ）： 1596 - 1604 .

LIU Renli ， SHE Dunxian ， LI Min ， et al . Using Satellite Observations to Assess Applicability of GLDAS and WGHM Hydrological Model ［J］. Geomatics and Information Science of Wuhan University ， 2019 ， 44 （ 11 ）： 1596 - 1604 .

钟敏，段建宾，许厚泽，等 . 利用卫星重力观测研究近5年中国陆地水量中长空间尺度的变化趋势［J］. 科学通报， 2009 ， 54 （ 9 ）： 1290 - 1294 .

ZHONG Min ， DUAN Jianbin ， XU Houze ， et al . Using Satellite Gravity Observation to Study the Variation Trend of Land Water Volume in China in the Middle and Long Spatial Scales in Recent Five Years ［J］. Chinese Science Bulletin ， 2009 ， 54 （ 9 ）： 1290 - 1294 .

瞿伟，陈沛男，张普方，等 . GRACE/GRACE-FO揭示黄河上游2002—2022年GWS及可持续性时空演化［J］. 武汉大学学报（信息科学版）， 2024 ， 49 （ 8 ）： 1287 - 1299 .

QU Wei ， CHEN Peinan ， ZHANG Pufang ， et al . Spatiotemporal Changes of GWS Changes and Their Sustainability in the Upper Reaches of the Yellow River Basin from 2002 to 2022 Revealed by GRACE/GRACE-FO Observations ［J］. Geomatics and Information Science of Wuhan University ， 2024 ， 49 （ 8 ）： 1287 - 1299 .

CHEN J L ， WILSON C R ， TAPLEY B D ， et al . Long-Term Groundwater Storage Change in Victoria， Australia from Satellite Gravity and In-Situ Observations ［J］. Global and Planetary Change ， 2016 ， 139 ： 56 - 65 .

ZHONG B ， LI Q ， LI X P ， et al . Basin-Scale Terrestrial Water Storage Changes Inferred from GRACE-Based Geopotential Differences： A Case Study of the Yangtze River Basin， China ［J］. Geophysical Journal International ， 2023 ， 233 （ 2 ）： 1318 - 1338 .

李琼，罗志才，钟波，等 . 利用GRACE时变重力场探测2010年中国西南干旱陆地水储量变化［J］. 地球物理学报， 2013 ， 56 （ 6 ）： 1843 - 1849 .

LI Qiong ， LUO Zhicai ， ZHONG Bo ， et al . Terrestrial Water Storage Changes of the 2010 Southwest China Drought Detected by GRACE Temporal Gravity Field ［J］. Chinese Journal of Geophysics ， 2013 ， 56 （ 6 ）： 1843 - 1849 .

肖云，李岩，周泽兵，等 . 中国重力卫星进展［J］. 武汉大学学报（信息科学版）， 2026 ， 51 （ 3 ）： 413 - 424 .

XIAO Yun ， LI Yan ， ZHOU Zebing ， et al . Progress on the Chinese Gravimetry Satellite Missions ［J］. Geomatics and Information Science of Wuhan University ， 2026 ， 51 （ 3 ）： 413 - 424 .

TANGDAMRONGSUB N ， DITMAR P G ， STEELE-DUNNE S C ， et al . Assessing Total Water Storage and Identifying Flood Events over Tonlé Sap Basin in Cambodia Using GRACE and MODIS Satellite Observations Combined with Hydrological Models ［J］. Remote Sensing of Environment ， 2016 ， 181 ： 162 - 173 .

CHAO N F ， WAN X W ， ZHONG Y L ， et al . Reconstructing a New Terrestrial Water Storage Deficit Index to Detect and Quantify Drought in the Yangtze River Basin ［J］. Journal of Hydrology ， 2023 ， 625 ： 129972 .

YI S ， SUN W K . Evaluation of Glacier Changes in High-Mountain Asia Based on 10 Year GRACE RL05 Models ［J］. Journal of Geophysical Research： Solid Earth ， 2014 ， 119 （ 3 ）： 2504 - 2517 .

XIE Y ， WANG L S ， BEVIS M ， et al . Joint Inversion of GNSS and GRACE Data for Ice Mass Loads in Greenland ［J］. Earth and Planetary Science Letters ， 2025 ， 658 ： 119329 .

高春春，陆洋，史红岭，等 . 基于GRACE RL06数据监测和分析南极冰盖27个流域质量变化［J］. 地球物理学报， 2019 ， 62 （ 3 ）： 864 - 882 .

GAO Chunchun ， LU Yang ， SHI Hongling ， et al . Detection and Analysis of Ice Sheet Mass Changes over 27 Antarctic Drainage Systems from GRACE RL06 Data ［J］ . Chinese Journal of Geophysics ， 2019 ， 62 （ 3 ）： 864 - 882 .

LIU B S ， ZOU X C ， YI S ， et al . Reconstructing GRACE-Like Time Series of High Mountain Glacier Mass Anomalies ［J］. Remote Sensing of Environment ， 2022 ， 280 ： 113177 .

ZHANG Z Z ， CHAO B F ， CHEN J L ， et al . Terrestrial Water Storage Anomalies of Yangtze River Basin Droughts Observed by GRACE and Connections with ENSO ［J］. Global and Planetary Change ， 2015 ， 126 ： 35 - 45 .

CHEN J L ， TAPLEY B ， RODELL M ， et al . Basin-Scale River Runoff Estimation from GRACE Gravity Satellites， Climate Models， and in Situ Observations： A Case Study in the Amazon Basin ［J］. Water Resources Research ， 2020 ， 56 （ 10 ）： e2020WR028032 .

HUANG Z Y ， YEH P J F ， JIAO J J ， et al . A New Approach for Assessing Groundwater Recharge by Combining GRACE and Baseflow with Case Studies in Karst Areas of Southwest China ［J］. Water Resources Research ， 2023 ， 59 （ 2 ）： e2022WR032091 .

熊景华，郭生练，王俊，等 . 长江流域陆地水储量变化及归因研究［J］. 武汉大学学报（信息科学版）， 2024 ， 49 （ 12 ）： 2241 - 2248 .

XIONG Jinghua ， GUO Shenglian ， WANG Jun ， et al . Variation and Attribution of Terrestrial Water Storage in the Yangtze River Basin ［J］ . Geomatics and Information Science of Wuhan University ， 2024 ， 49 （ 12 ）： 2241 - 2248 .

VON STORCH H ， ZORITA E ， CUBASCH U . Downscaling of Global Climate Change Estimates to Regional Scales： An Application to Iberian Rainfall in Wintertime ［J］. Journal of Climate ， 1993 ， 6 （ 6 ）： 1161 - 1171 .

LI B L ， RODELL M ， ZAITCHIK B F ， et al . Assimilation of GRACE Terrestrial Water Storage into a Land Surface Model： Evaluation and Potential Value for Drought Monitoring in Western and Central Europe ［J］. Journal of Hydrology ， 2012 ， 446 ： 103 - 115 .

TEUTSCHBEIN C ， WETTERHALL F ， SEIBERT J . Evaluation of Different Downscaling Techniques for Hydrological Climate-Change Impact Studies at the Catchment Scale ［J］. Climate Dynamics ， 2011 ， 37 （ 9 ）： 2087 - 2105 .

YIN W J ， ZHANG G Q ， LIU F T ， et al . Improving the Spatial Resolution of GRACE-Based Groundwater Storage Estimates Using a Machine Learning Algorithm and Hydrological Model ［J］. Hydrogeology Journal ， 2022 ， 30 （ 3 ）： 947 - 963 .

WAN Z M ， ZHANG K ， XUE X W ， et al . Water Balance-Based Actual Evapotranspiration Reconstruction from Ground and Satellite Observations over the Conterminous United States ［J］. Water Resources Research ， 2015 ， 51 （ 8 ）： 6485 - 6499 .

YIN W J ， HU L T ， ZHANG M L ， et al . Statistical Downscaling of GRACE-Derived Groundwater Storage Using ET Data in the North China Plain ［J］. Journal of Geophysical Research： Atmospheres ， 2018 ， 123 （ 11 ）： 5973 - 5987 .

ZHANG D ， LIU X M ， BAI P . Assessment of Hydrological Drought and Its Recovery Time for Eight Tributaries of the Yangtze River （China） Based on Downscaled GRACE Data ［J］. Journal of Hydrology ， 2019 ， 568 ： 592 - 603 .

SEYOUM W M ， MILEWSKI A M . Improved Methods for Estimating Local Terrestrial Water Dynamics from GRACE in the Northern High Plains ［J］. Advances in Water Resources ， 2017 ， 110 ： 279 - 290 .

MIRO M E ， FAMIGLIETTI J S . Downscaling GRACE Remote Sensing Datasets to High-Resolution Groundwater Storage Change Maps of California's Central Valley ［J］. Remote Sensing ， 2018 ， 10 （ 1 ）： 143 .

CHEN L ， HE Q S ， LIU K ， et al . Downscaling of GRACE-Derived Groundwater Storage Based on the Random Forest Model ［J］. Remote Sensing ， 2019 ， 11 （ 24 ）： 2979 .

RAHAMAN M M ， THAKUR B ， KALRA A ， et al . Estimating High-Resolution Groundwater Storage from GRACE： A Random Forest Approach ［J］. Environments ， 2019 ， 6 （ 6 ）： 63 .

ZHANG J X ， LIU K ， WANG M . Downscaling Groundwater Storage Data in China to a 1-km Resolution Using Machine Learning Methods ［J］. Remote Sensing ， 2021 ， 13 （ 3 ）： 523 .

ALI S ， KHORRAMI B ， JEHANZAIB M ， et al . Spatial Downscaling of GRACE Data Based on XGBoost Model for Improved Understanding of Hydrological Droughts in the Indus Basin Irrigation System （IBIS）［J］. Remote Sensing ， 2023 ， 15 （ 4 ）： 873 .

JANGA B ， ASAMANI G P ， SUN Z H ， et al . A Review of Practical AI for Remote Sensing in Earth Sciences ［J］. Remote Sensing ， 2023 ， 15 （ 16 ）： 4112 .

KARPATNE A ， EBERT-UPHOFF I ， RAVELA S ， et al . Machine Learning for the Geosciences： Challenges and Opportunities ［J］. IEEE Transactions on Knowledge and Data Engineering ， 2019 ， 31 （ 8 ）： 1544 - 1554 .

CHENG M K ， RIES J C ， TAPLEY B D . Variations of the Earth’s Figure Axis from Satellite Laser Ranging and GRACE ［J］. Journal of Geophysical Research： Solid Earth ， 2011 ， 116 （ B1 ）： B01409 .

SWENSON S ， WAHR J . Post-Processing Removal of Correlated Errors in GRACE Data ［J］. Geophysical Research Letters ， 2006 ， 33 （ 8 ）： 2005GL025285 .

CHEN J L ， WILSON C R ， LI J ， et al . Reducing Leakage Error in GRACE-Observed Long-Term Ice Mass Change： A Case Study in West Antarctica ［J］. Journal of Geodesy ， 2015 ， 89 （ 9 ）： 925 - 940 .

WAHR J ， MOLENAAR M ， BRYAN F . Time Variability of the Earth’s Gravity Field： Hydrological and Oceanic Effects and Their Possible Detection Using GRACE ［J］. Journal of Geophysical Research： Solid Earth ， 1998 ， 103 （ B12 ）： 30205 - 30229 .

WANG Y M ， LI C ， CUI Y J ， et al . Spatial Downscaling of GRACE-Derived Groundwater Storage Changes Across Diverse Climates and Human Interventions with Random Forests ［J］. Journal of Hydrology ， 2024 ， 640 ： 131708 .

CUI L L ， LI Y ， ZHONG B ， et al . Assessing the Impact of 2022 Extreme Drought on the Yangtze River Basin Using Downscaled GRACE/GRACE-FO Data Obtained by Partitioned Random Forest Algorithm ［J］. International Journal of Remote Sensing ， 2025 ， 46 （ 3 ）： 1219 - 1247 .

AGHBASHLO M ， PENG W X ， TABATABAEI M ， et al . Machine Learning Technology in Biodiesel Research： A Review ［J］. Progress in Energy and Combustion Science ， 2021 ， 85 ： 100904 .

OTCHERE D A ， ARBI GANAT T O ， GHOLAMI R ， et al . Application of Supervised Machine Learning Paradigms in the Prediction of Petroleum Reservoir Properties： Comparative Analysis of ANN and SVM Models ［J］. Journal of Petroleum Science and Engineering ， 2021 ， 200 ： 108182 .

DARST B F ， MALECKI K C ， ENGELMAN C D . Using Recursive Feature Elimination in Random Forest to Account for Correlated Variables in High Dimensional Data ［J］. BMC Genetics ， 2018 ， 19 （ 1 ）： 65 .

XU T R ， GUO Z X ， LIU S M ， et al . Evaluating Different Machine Learning Methods for Upscaling Evapotranspiration from Flux Towers to the Regional Scale ［J］. Journal of Geophysical Research： Atmospheres ， 2018 ， 123 （ 16 ）： 8674 - 8690 .

CHEN T Q ， GUESTRIN C . XGBoost： A Scalable Tree Boosting System ［C］// The 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining ， San Francisco， USA ， 2016 .

WEI A H ， YU K N ， DAI F G ， et al . Application of Tree-Based Ensemble Models to Landslide Susceptibility Mapping： A Comparative Study ［J］. Sustainability ， 2022 ， 14 （ 10 ）： 6330 .

PASCAL C ， FERRANT S ， SELLES A ， et al . Evaluating Downscaling Methods of GRACE （Gravity Recovery and Climate Experiment） Data： A Case Study over a Fractured Crystalline Aquifer in Southern India ［J］. Hydrology and Earth System Sciences ， 2022 ， 26 （ 15 ）： 4169 - 4186 .

JIANG Q N ， FERREIRA V G ， CHEN J S . Monitoring Groundwater Changes in the Yangtze River Basin Using Satellite and Model Data ［J］. Arabian Journal of Geosciences ， 2016 ， 9 （ 7 ）： 500 .

禤键豪，陈智伟，张兴福，等 . 联合GRACE与GRACE-FO反演2002—2020年长江流域陆地水储量变化［J］. 大地测量与地球动力学， 2021 （ 9 ）： 961 - 966 .

XUAN Jianhao ， CHEN Zhiwei ， ZHANG Xingfu ， et al . Combining GRACE and GRACE-FO to Derive Terrestrial Water Storage Changes in the Yangtze River Basin from 2002 to 2020 ［J］. Journal of Geodesy and Geodynamics ， 2021 （ 9 ）： 961 - 966 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Spatiotemporal Changes of GWS Changes and Their Sustainability in the Upper Reaches of the Yellow River Basin from 2002 to 2022 Revealed by GRACE/GRACE-FO Observations

Machine Learning-Based Remote Sensing Retrieval of Soil Moisture： Advances and Challenges

A Method of Extracting Building Roof Edge Points from Airborne Point Clouds with Wide-Area Feature Optimization

An Integrated Rule-Based and Machine Learning Method for Trip Purpose Identification

Simulation and Analysis of Short-Period Time-Varying Gravity Field Recovery Using a Very Low Earth Orbit Satellite Constellation

Related Author

QU Wei

CHEN Peinan

ZHANG Pufang

TANG Xingyou

LI Jiuyuan

YUAN Qiangqiang

LI Zhenghao

YANG Qianqian

Related Institution

College of Geology Engineering and Geomatics, Chang'an University

State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University

School of Resources and Environmental Sciences, Wuhan University

Department of Geography, Faculty of Social Sciences, Hong Kong Baptist University

Department of Civil and Environmental Engineering, Norwegian University of Science and Technology, Trondheim

AI问答

⁰