针对青海省地形复杂、地面观测站稀疏导致降水数据精度不足问题，融合 GPM、ERA5、DEM 等多源数据，基于 BP - LSTM 和随机森林模型，生成 0.01° 空间分辨率、1h/3h 时间分辨率高精度降水数据集，验证显示 RMSE≤2.5mm，降水事件捕捉率≥90%，为区域生态评估、水文模拟等提供数据支撑。
", "ds_source": "融合 GPM（NASA 提供，0.5h/10km 分辨率 ，https://gpm.nasa.gov/data/imerg ）、ERA5（ECMWF 提供，1h/25km 分辨率 ，https://www.ecmwf.int ）、NDVI（16d/30m ）、DEM（30m ）及地面观测站数据（青海省气象局，1h 分辨率）。
", "ds_process_way": "1. 数据预处理：采用 bilinear 插值将 GPM、ERA5 降尺度至 0.01°，最邻近插值将 NDVI、DEM 重采样至 0.01°，过滤小于 0.1mm/h 降水数据；
\n2. 模型融合：BP - LSTM 模型（前后 BPNN 模块捕捉空间相关性，核心 LSTM 模块捕捉时间相关性 ）、随机森林模型（100 棵决策树集成学习 ）；
\n3. 精度验证：采用 RMSE、RB、MAE 评估准确性，POD、FAR、MISS 评估降水事件捕捉能力
", "ds_quality": "融合数据 RMSE≤2.5mm，RB（绝对值）≤15%，POD≥0.9，FAR≤0.6；数据经过多源验证，与地面观测数据匹配度高，误差在可接受范围。
", "ds_acq_start_time": "1990-01-01 00:00:00", "ds_acq_end_time": "2020-12-31 00:00:00", "ds_acq_place": "青海省", "ds_acq_lon_east": 105.0, "ds_acq_lat_south": 30.0, "ds_acq_lon_west": 89.0, "ds_acq_lat_north": 40.0, "ds_acq_alt_low": 1500.0, "ds_acq_alt_high": 6500.0, "ds_share_type": "apply-access", "ds_total_size": 6929371194, "ds_files_count": 0, "ds_format": "*.nc", "ds_space_res": "0.01°", "ds_time_res": "3小时", "ds_coordinate": "WGS84", "ds_projection": "", "ds_thumbnail": "75308376-f417-457b-ace1-0f5a5f043919.png", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "支持 GeoTIFF 格式的软件（如 ArcGIS、QGIS ）处理空间数据，CSV 数据可用 Python/R 等处理；数据为栅格均值，站点精确值需结合观测数据校正；用于非商业研究需注明来源", "ds_from_station": "", "organization_id": "5b99d600-008a-4069-8fc3-7adb9c3f2f8b", "ds_serv_man": "权晨", "ds_serv_phone": "18397119373", "ds_serv_mail": "quanchen007@sina.com", "doi_value": "", "subject_codes": [ "170.15", "170.45" ], "quality_level": 0, "publish_time": "2026-06-10 10:03:09", "last_updated": "2026-06-10 10:03:09", "protected": false, "protected_to": "2027-08-20 00:00:00", "lang": "zh", "cstr": "", "i18n": { "en": { "title": "Three-hour precipitation fusion dataset based on random forest model in Muli Kuangqu, Qinghai Province (1990-2020)", "ds_format": "*.nc", "ds_source": "Fusion GPM (provided by NASA, 0.5h/10km resolution, https://gpm.nasa.gov/data/imerg ）ERA5 (provided by ECMWF, with a resolution of 1h/25km), https://www.ecmwf.int ）NDVI (16d/30m), DEM (30m), and ground observation station data (Qinghai Provincial Meteorological Bureau, 1-hour resolution)
", "ds_quality": "Fusion data RMSE ≤ 2.5mm, RB (absolute value) ≤ 15%, POD≥0.9，FAR≤0.6； The data has been verified by multiple sources and has a high degree of matching with ground observation data, with errors within an acceptable range
", "ds_ref_way": "", "ds_abstract": "In response to the problem of insufficient accuracy of precipitation data due to complex terrain and sparse ground observation stations in Qinghai Province, multi-source data such as GPM, ERA5, and DEM were integrated, based on BP-LSTM and random forest model, a high-precision precipitation dataset with 0.01° spatial resolution and 1h/3h temporal resolution was generated. The verification showed that RMSE was ≤2.5mm, and the precipitation event capture rate was ≥90%, providing data support for regional ecological assessment, hydrological simulation, etc.
", "ds_time_res": "", "ds_acq_place": "Qinghai Province", "ds_space_res": "", "ds_projection": "", "ds_process_way": "1. Data preprocessing: Bilinear interpolation is used to downscale GPM and ERA5 to 0.01 °, nearest neighbor interpolation is used to resample NDVI and DEM to 0.01 °, and precipitation data less than 0.1mm/h is filtered;
", "ds_ref_instruction": "Software that supports GeoTIFF format (such as ArcGIS, QGIS) can process spatial data, and CSV data can be processed using Python/R, etc; The data is the grid mean, and the precise values of the stations need to be corrected in conjunction with the observed data; For non-commercial research, the source must be indicated" } }, "submit_center_id": "ncdc", "data_level": 0, "recommendation_value": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "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, "belong_to_nieer": false, "ds_topic_tags": [ "降水数据融合", "BP - LSTM", "随机森林", "高海拔地区" ], "ds_subject_tags": [ "大气科学", "地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "青海省", "木里矿区" ], "ds_time_tags": [ 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 ], "ds_contributors": [ { "true_name": "权晨", "email": "quanchen007@sina.com", "work_for": "青海省气象科学研究所", "country": "中国" }, { "true_name": "张小丹", "email": "xdzhang@qhu.edu.cn", "work_for": "青海大学", "country": "中国" }, { "true_name": "赵彤", "email": "taytaycma@163.com", "work_for": "青海省气象科学研究所", "country": "中国" }, { "true_name": "刘畅", "email": "644496605@qq.com", "work_for": "青海大学", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "权晨", "email": "quanchen007@sina.com", "work_for": "青海省气象科学研究所", "country": "中国" }, { "true_name": "张小丹", "email": "xdzhang@qhu.edu.cn", "work_for": "青海大学", "country": "中国" }, { "true_name": "刘畅", "email": "644496605@qq.com", "work_for": "青海大学", "country": "中国" }, { "true_name": "赵彤", "email": "taytaycma@163.com", "work_for": "青海省气象科学研究所", "country": "中国" } ], "ds_managers": [ { "true_name": "权晨", "email": "quanchen007@sina.com", "work_for": "青海省气象科学研究所", "country": "中国" }, { "true_name": "张小丹", "email": "xdzhang@qhu.edu.cn", "work_for": "青海大学", "country": "中国" } ], "category": "气象" }