作者构建了两种降尺度方案：从像元时间连续性出发采用半监督变分自编码器算法（SSVAER），以及基于像元空间相关性采用多尺度地理加权回归模型（MGWR）。以塔里木河流域为研究区，采用德克萨斯大学奥斯汀分校空间研究中心（CSR-RL06M）和喷气推进实验室（JPL-RL06M）两种GRACE发行机构的Mascon产品均值作为数据源，时间尺度为2002年4月至2022年6月，共210个月。预测因子选择降水、雪水当量、陆地表面温度、蒸发、径流和NDVI。两种方案均实现了GRACE反演地下水储量异常（GWSA）空间分辨率从0.5°提高至0.1°。研究结果表明：在80%以上的研究区域内，降尺度前后的GWSA数据相关性介于0.94-0.98之间；基于SSVAER和MGWR两种方案分别将GWSA与地下水位之间的相关性从0.27提高至0.59和0.52。均方根误差均低于8。本数据集内容包括：（1）基于SSVAER方案的降尺度结果；（2）基于MGWR方案的降尺度结果。该数据集以.tif格式存储，空间分辨率为0.1°。数据集由420个数据文件组成，数据量为62.2 MB（压缩为1个文件，50.1 MB）。基于该数据集的论文发表在Science of the Total Environment，2024年907卷。
", "ds_source": "采用德克萨斯大学奥斯汀分校空间研究中心（CSR-RL06M）和喷气推进实验室（JPL-RL06M）两种GRACE发行机构的Mascon产品均值作为数据源，时间尺度为2002年4月至2022年6月，共210个月。
", "ds_process_way": "", "ds_quality": "", "ds_acq_start_time": null, "ds_acq_end_time": null, "ds_acq_place": "", "ds_acq_lon_east": null, "ds_acq_lat_south": null, "ds_acq_lon_west": null, "ds_acq_lat_north": null, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "login-access", "ds_total_size": 52549745, "ds_files_count": 1, "ds_format": ".tif格式", "ds_space_res": "0.1°", "ds_time_res": "月", "ds_coordinate": "WGS84", "ds_projection": "", "ds_thumbnail": "e9ba7e8a-5cc0-4c17-bd23-98fcba33bf49.jpg", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "a5877b42-96ea-4f13-af7e-246f355413d6", "doi_value": "", "subject_codes": [ "170.45" ], "quality_level": 1, "publish_time": "2025-06-05 12:59:27", "first_publish_time": null, "last_updated": "2025-06-05 12:59:27", "protected": false, "protected_to": null, "lang": "zh", "cstr": null, "license": null, "extra": null, "files_shape": [ { "name": "GWSA_0.1DEG_2002-2022.rar", "size": 52549745, "is_dir": false } ], "features": null, "data_level": 0, "i18n": { "en": { "title": "Experimental data set on monthly changes in 0.1° groundwater reserves in the Tarim River Basin and adjacent areas (2002-2022)", "ds_abstract": "The authors constructed two downscaling schemes: using the semi-supervised variational autoencoder algorithm (SSVAER) based on pixel temporal continuity, and using the multi-scale geographically weighted regression model (MGWR) based on pixel spatial correlation. Taking the Tarim River Basin as the research area, Mascon product averages from two GRACE publishing institutions, the University of Texas at Austin Space Research Center (CSR-RL06M) and the Jet Propulsion Laboratory (JPL-RL06M), are used as data sources. The time scale is from April 2002 to June 2022, a total of 210 months. Prediction factors are precipitation, snow water equivalent, land surface temperature, evaporation, runoff and NDVI. Both schemes have achieved an increase in the spatial resolution of GRACE inversion of groundwater reserve anomalies (GWSA) from 0.5° to 0.1°. The research results show that in more than 80% of the study areas, the correlation between GWSA data before and after downscaling is between 0.94 and 0.98; based on the two schemes of SSVAER and MGWR, the correlation between GWSA and groundwater levels is increased from 0.27 to 0.59 and 0.52 respectively. The root-mean-square errors are all below 8. This dataset includes: (1) downscaling results based on the SSVAER scheme;(2) downscaling results based on the MGWR scheme. The dataset is stored in.tif format with a spatial resolution of 0.1°. The dataset consists of 420 data files, with a data volume of 62.2 MB (compressed into 1 file, 50.1 MB). A paper based on this dataset was published in Science of the Total Environment, Volume 907, 2024.
", "ds_source": "Mascon product averages from two GRACE publishing institutions, the University of Texas at Austin Space Research Center (CSR-RL06M) and the Jet Propulsion Laboratory (JPL-RL06M), were used as data sources. The time scale was from April 2002 to June 2022, a total of 210 months.
", "ds_format": ".tif format", "ds_space_res": "0.1°", "ds_time_res": "months" } }, "license_type": null, "doi_reg_from": "reg_local", "cstr_reg_from": "reg_local", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "ds_topic_tags": [ "GRACE", "降尺度", "地下水储量异常", "机器学习" ], "ds_subject_tags": [ "地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "塔里木河流域" ], "ds_time_tags": [ 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022 ], "ds_contributors": [ "薛冬萍", "桂东伟" ], "ds_meta_authors": [ "薛冬萍" ], "ds_managers": [ "李锦" ], "category": "水资源" }