雪水当量（Snow water equivalent，SWE）是积雪储存水量的关键指标，是地表水文模型和气候模型的重要参数。针对现有雪水当量实测成本高、遥感产品精度不稳定，且数据时空不连续等问题。本项目基于机器学习算法，通过融合AMSRE、ESAGB、GlobSnow、GLDAS、ERA5_Land及SWEML等多种雪水当量产品，采用统计重建与数据同化方法，融合生成1980–2024年北极（北纬66°34以北）空间分辨率为25km的日尺度雪水当量产品。通过实测数据验证，表面看该数据集与地面实测数据具有良好一致性，总体均方根误差（RMSE）为18.5mm，平均偏差为-7.45mm，相关系数多数超过0.85。本数据集可为北极水文模拟、陆面过程模型的数据同化提供重要的数据支撑。\n
数据集为Geotiff格式，栅格值为雪水当量乘以10，存储方式为年/月/日，命名方式为Arctic_SWE_YYYYMMDD。", "ds_source": "
AMSRE 是由美国国家航空航天局（NASA）地球观测系统的Aqua卫星和日本宇宙航空研究开发机构（JAXA）的GCOM-W1卫星上的微波扫描辐射计收集的数据 (https://nsidc.org/data/ae_dysno),来自NASA的AMSRE（AMSR-E）提供从2002年6月19日到2011年10月3日的全球每日积雪水当量数据，而来自JAXA的AMSRE（AMSR2）则从2012 年7月2日至今持续提供全球每日积雪水当量数据。GlobSnow 主要受欧空局资助，其积雪水当量产品是一个协同数据集，它利用先进的数据同化方案，将卫星被动微波数据（来自SSM/I和AMSR-E等传感器）与地面气象站的雪深观测数据相结合(https://www.globsnow.info/)。GLDAS 是由NASA戈达德太空飞行中心及其他合作伙伴开发的一个建模系统。它引入卫星和地面观测数据，驱动多个陆面模型（如 Noah、VIC），以生成包括积雪水当量在内的最优陆面状态场。该系统的配置和驱动数据在技术文档中有详细说明。GLDAS的积雪水当量数据从 NASA 戈达德地球科学数据与信息服务中心下载。ERA5-Land是由欧洲中期天气预报中心生产的全球陆面再分析数据集。它是通过重新执行ECMWF的ERA5气候再分析的陆面模块生成(https://cds.climate.copernicus.eu/datasets/derived-era5-land-daily-statistics?tab=download)。此外，SWEML 数据集（版本 3.0）是一个利用机器学习技术并基于原位测量数据训练生成的全球积雪水当量数据集(https://zenodo.org/records/16822772)。", "ds_process_way": "
采用随机森林算法，对AMSRE、ESAGB、GlobSnow、GLDAS、ERA5_Land及SWEML等多种雪水当量进行融合，生成一套时序更完整的雪水当量数据集。", "ds_quality": "
基于实测站点数据集，采用均方根误差（RMSE）、平均绝对误差（MAE）、决定系数（R2）和偏差四个指标评估雪水当量数据的精度。数据融合过程中，为确保时空一致性，将验证范围限定在北纬66°34以上地区。此外，为验证雪水当量数据精度，使用了GLDAS、ESAGB和AMSRE三个雪水当量参考数据集进行交叉验证。", "ds_acq_start_time": "1980-01-01 00:00:00", "ds_acq_end_time": "2024-12-31 00:00:00", "ds_acq_place": "北极", "ds_acq_lon_east": 180.0, "ds_acq_lat_south": 66.53333333333333, "ds_acq_lon_west": null, "ds_acq_lat_north": 90.0, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "login-access", "ds_total_size": 525620, "ds_files_count": 2, "ds_format": "*.tif", "ds_space_res": "25km", "ds_time_res": "日", "ds_coordinate": "WGS84", "ds_projection": "", "ds_thumbnail": "85c795fc-cfd8-43f5-a029-52328c17424c.png", "ds_thumb_from": 2, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "53943799-d453-4bf2-a141-56c205c1355b", "ds_serv_man": "李红星", "ds_serv_phone": "0931-4967592", "ds_serv_mail": "ncdc@lzb.ac.cn", "doi_value": "", "subject_codes": [ "170.4510" ], "quality_level": 0, "publish_time": "2026-05-07 11:29:41", "last_updated": "2026-05-07 11:30:39", "protected": false, "protected_to": null, "lang": "zh", "cstr": "11738.11.NCDC.ARCTIC-CHANGE.DB7145.2026", "i18n": { "en": { "title": "Daily snow water equivalent dataset at 25 km on the Arctic (1980-2024)", "ds_format": "*.tif", "ds_source": "
AMSRE is a satellite-based dataset that includes data collected by microwave scanning radiometers on the Aqua satellite of the National Aeronautics and Space Administration (NASA) Earth Observation System and the GCOM-W1 satellite of the Japan Aerospace Exploration Agency (JAXA) (https://nsidc.org/data/ae_dysno). The AMSRE from NASA (AMSR-E) provides global daily SWE data from 19 June, 2002, to 3 October, 2011, and 130 the AMSRE from JAXA (AMSR2) has been providing global daily SWE data from 2 July, 2012, to the present. The GlonSnow SWE products is a synergistic datasets that combines satellite passive microwave data (from sensors like SSM/I and AMSR-E) with ground-based weather station snow depth observations using an advanced data assimilation scheme (https://www.globsnow.info/). GLDAS is a modeling system developed by NASA Goddard Space Flight Center (GSFC) and other partners. It ingests satellite-based and ground-based observations to force multiple land surface models (e.g., Noah, VIC) to generate optimal fields of land surface states, including SWE. The system configuration and forcing data are described in technical documentation. The GLDAS SWE was downloaded from the NASA Goddard Earth Sciences Data and Information Services Center (GES DISC). ERA5-Land is a global land surface reanalysis dataset produced by the European Centre for Medium-Range Weather Forecasts (ECMWF). It is generated by replaying the land component of the ECMWF's ERA5 climate reanalysis (https://cds.climate.copernicus.eu/datasets/derived-era5-land-daily-statistics?tab=download). Moreover, the SWEML dataset (version3.0) was is a global snow water equivalent dataset using machine learning trained with in-situ measurements. The temporal resolution of the SWEML product is daily, and the spatial resolution is 0.25˚ (approximately 25km). It covers latitudes of 90S to 90N and longitudes of 180W to 180E with global scales, excluding Antarctica. The dataset is provided in NetCDF format, organized by year. Each year contains daily SWE data, including leap days in leap years (https://zenodo.org/records/16822772)", "ds_quality": "
We calculated the error metrics to evaluate the accuracy by comparing SWE dataset and datesets with observations. The root meansquare error (RMSE), mean absolute error (MAE), Pearson correlation coefficient (R) and bias were adopted to assess theaccuracy of the SWE dataset. To ensure spatiotemporal consistency during the data fusion process, the validation scope was confined to regions north of 66°34'N. Moreover, to validate the accuracy in spatial of snow water equivalent data, cross-validation was performed using three reference datasets: GLDAS, ESAGB, and AMSRE.", "ds_ref_way": "", "ds_abstract": "
Snow Water Equivalent (SWE) is a key indicator for the amount of water stored in snow, and also a crucial parameter in surface hydrological models and climate models. Aiming at addressing the existing issues of snow water equivalent, such as high cost of in-situ measurements, unstable accuracy of remote sensing products, and discontinuous spatiotemporal data. Based on machine learning algorithms, this project integrates multiple snow water equivalent products including AMSRE, ESAGB, GlobSnow, GLDAS, ERA5_Land, and SWEML. By adopting statistical reconstruction and data assimilation methods, it fuses and generates a daily-scale snow water equivalent product for the Arctic region (north of 66°34' North Latitude) from 1980 to 2024, with a spatial resolution of 25 km.Verification results show that the dataset has good consistency with observations. The overall Root Mean Square Error (RMSE) is 18.5 mm, the mean bias is -7.45 mm, and most correlation coefficients exceed 0.85. This dataset can provide important data support for hydrological simulation in the Arctic and data assimilation in land surface process models.", "ds_time_res": "", "ds_acq_place": "Arctic", "ds_space_res": "", "ds_projection": "", "ds_process_way": "
By unsing the random forest (RF) model of a machine learning algorithm, multiple snow water equivalent products—including AMSRE, ESAGB, GlobSnow, GLDAS, ERA5_Land, and SWEML—are integrated to generate a snow water equivalent dataset with more complete temporal coverage.", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "license_type": "CC BY 4.0 (Creative Commons Attribution 4.0 International License)", "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": [ "雪水当量", "长序列", "机器学习", "数据融合", "精度评估" ], "ds_subject_tags": [ "自然地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "北纬66°34′以北" ], "ds_time_tags": [ 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, 2023, 2024 ], "ds_contributors": [ { "true_name": "吴通华", "email": "thuawu@lzb.ac.cn", "work_for": "中国科学院西北生态环境资源研究院，青藏高原冰冻圈研究站, 冰冻圈科学国家重点实验室, ", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "吴通华", "email": "thuawu@lzb.ac.cn", "work_for": "中国科学院西北生态环境资源研究院，青藏高原冰冻圈研究站, 冰冻圈科学国家重点实验室, ", "country": "中国" } ], "ds_managers": [ { "true_name": "吴通华", "email": "thuawu@lzb.ac.cn", "work_for": "中国科学院西北生态环境资源研究院，青藏高原冰冻圈研究站, 冰冻圈科学国家重点实验室, ", "country": "中国" } ], "category": "积雪" }