本研究提出了全球陆地表面 TWS 异常值的长期（即 1940-2022 年）和高分辨率（即 0.25°）月度时间序列。重建是通过一套机器学习模型实现的，该模型包含大量预测因子，包括气候和水文变量、土地利用/土地覆被数据以及植被指标（如叶面积指数）。此外，我们的重建成功地再现了气候多变性的影响，如强烈的厄尔尼诺现象。
\nGTWS-MLrec 数据集包括三个基于 JPL、CSR 和 GSFC mascons 的重建，三个去趋势和去季节化重建，以及六个陆地区域的全球平均 TWS 序列（包括格陵兰和南极洲）。GTWS_MLrec 具有广泛的属性，可以支持广泛的应用，如更好地了解全球水预算、约束和评估水文模型、气候-碳耦合和水资源管理。
", "ds_source": "模型训练数据：GRACE/GRACE-FO TES.
\n机器学习模型的输入：来自第五代欧洲中期天气中心的11个气象要素；ERA5的两个水文变量；土地利用和覆盖数据；植被指标，即LAI和太阳诱导荧光；11个气象变量等。
\n作为比较，使用了两个最广泛使用的全球TWS重建数据集（0.5°分辨率），即GRACE-Humphrey和Gudmundsson的REC数据集以及最近的类似于GRACE的重建TWS数据集。", "ds_process_way": "
重建是通过一套机器学习模型实现的，该模型使用了广泛的输入驱动因素，包括气候和水文变量、土地利用/土地覆盖数据以及植被指标。机器学习模型由 GRACE/GRACE-FO 测量数据训练而成，
", "ds_quality": "机器学习重构的 TWS 估计值（即 GTWS-MLrec）与 GRACE/GRACE-FO 测量结果非常吻合，在 GRACE 时代显示出高相关系数和低偏差。还利用其他独立数据集对 GTWS-MLrec 进行了评估，如陆地-海洋质量预算、341 个大河流域的大尺度水量平衡以及 10,168 个测站的溪流测量数据。我们发现，所提出的方法在整体上与之前的 TWS 数据集相比，性能更佳或更可靠。
", "ds_acq_start_time": "1940-01-01 00:00:00", "ds_acq_end_time": "2022-01-01 00:00:00", "ds_acq_place": "全球", "ds_acq_lon_east": 180.0, "ds_acq_lat_south": -90.0, "ds_acq_lon_west": -180.0, "ds_acq_lat_north": 90.0, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "login-access", "ds_total_size": 49636781181, "ds_files_count": 11, "ds_format": "nc、xlsx", "ds_space_res": "", "ds_time_res": "", "ds_coordinate": "无", "ds_projection": "", "ds_thumbnail": "6b23bed9-627c-4653-8157-2ddd5b323888.png", "ds_thumb_from": 2, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "0a4269e1-65f4-45f1-aeba-88ea3068eebf", "ds_serv_man": "敏玉芳", "ds_serv_phone": "0931-4967596", "ds_serv_mail": "ncdc@lzb.ac.cn", "doi_value": "", "subject_codes": [ "170.45" ], "quality_level": 3, "publish_time": "2024-06-18 10:08:11", "last_updated": "2026-01-14 10:54:16", "protected": false, "protected_to": null, "lang": "zh", "cstr": "11738.11.NCDC.ZENODO.DB6521.2024", "i18n": { "en": { "title": "Reconstruction of Global Land Water Storage Dataset Based on Machine Learning (1940 present)", "ds_format": "nc、xlsx", "ds_source": "Model training data: GRACE/GRACE-FO TES
\nInput for machine learning model: 11 meteorological elements from the 5th Generation European Mid Range Weather Center; Two hydrological variables of ERA5; Land use and cover data; Vegetation indicators, namely LAI and solar induced fluorescence; 11 meteorological variables, etc.
\nAs a comparison, two of the most widely used global TWS reconstruction datasets (0.5 ° resolution) were used, namely GRACE Humphrey and Gudmundsson's REC dataset, as well as the recent GRACE like reconstructed TWS dataset.", "ds_quality": "
The TWS estimation value reconstructed by machine learning (i.e. GTWS MLrec) is highly consistent with the GRACE/GRACE-FO measurement results, showing high correlation and low bias in the GRACE era. GTWS MLrec was also evaluated using other independent datasets, such as land ocean quality budget, large-scale water balance of 341 major river basins, and stream measurement data from 10168 stations. We found that the proposed method performs better or more reliably overall compared to the previous TWS dataset.
", "ds_ref_way": "", "ds_abstract": "This study presents a long-term (1940-2022) and high-resolution (0.25 °) monthly time series of global land surface TWS anomalies. Reconstruction is achieved through a set of machine learning models that include a large number of predictive factors, including climate and hydrological variables, land use/land cover data, and vegetation indicators such as leaf area index. In addition, our reconstruction successfully reproduced the effects of climate variability, such as the strong El Ni ñ o phenomenon.
\nThe GTWS MLrec dataset includes three reconstructions based on JPL, CSR, and GSFC mascons, three de trending and de seasonal reconstructions, and global average TWS sequences for six land regions (including Greenland and Antarctica). GTWS.MRec has a wide range of attributes that can support a wide range of applications, such as better understanding global water budgets, constraining and evaluating hydrological models, climate carbon coupling, and water resource management.
", "ds_time_res": "", "ds_acq_place": "Global", "ds_space_res": "", "ds_projection": "", "ds_process_way": "Reconstruction is achieved through a set of machine learning models that utilize a wide range of input drivers, including climate and hydrological variables, land use/land cover data, and vegetation indicators. The machine learning model is trained on GRACE/GRACE-FO measurement data
", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "doi_reg_from": "reg_outside", "cstr_reg_from": "reg_outside", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "ds_topic_tags": [ "陆地储水", "气候变化", "全球" ], "ds_subject_tags": [ "地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "全球" ], "ds_time_tags": [ 1940, 1941, 1942, 1943, 1944, 1945, 1946, 1947, 1948, 1949, 1950, 1951, 1952, 1953, 1954, 1955, 1956, 1957, 1958, 1959, 1960, 1961, 1962, 1963, 1964, 1965, 1966, 1967, 1968, 1969, 1970, 1971, 1972, 1973, 1974, 1975, 1976, 1977, 1978, 1979, 1980, 1981, 1982, 1983, 1984, 1985, 1986, 1987, 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022 ], "ds_contributors": [ { "true_name": "尹家波", "email": "jboyn@whu.edu.cn", "work_for": "武汉大学", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "尹家波", "email": "jboyn@whu.edu.cn", "work_for": "武汉大学", "country": "中国" } ], "ds_managers": [ { "true_name": "尹家波", "email": "jboyn@whu.edu.cn", "work_for": "武汉大学", "country": "中国" } ], "category": "水文" }