天山是世界上距离海洋最远的山系，世界七大山系之一。该区域属于我国典型的高寒山区，被誉为“中亚水塔”，对于新疆乃至中亚地区均具有重要的战略意义。随着遥感技术的发展，卫星反演降水成为估算山区降水的重要手段，但是山区地形复杂且不均匀，致使山区降水反演产品精度不高。针对此问题，本研究开展天山山区多源降水融合数据集研制，以GSMaP卫星降水数据为初始场，同期区域1065个台站的实况降水数据，发展一种基于最优插值的星地降水产品融合方法，最终生成2000-2022年天山山区逐日降水产品集｡本数据集在研制过程中对实况数据进行了严格质控，对逐日融合降水数据进行了质量评估，可望为复杂地形区域水资源管理与高效利用提供数据支撑。
", "ds_source": "本数据集生产主要基于GSMaP卫星降水以及雨量站实况降水数据。
\n(1) GSMaP卫星降水：全球卫星降水图（GSMaP）提供了分辨率为0.1 x 0.1度的全球小时降雨量。GSMaP是全球降水测量任务的产物，该任务使用GPM核心卫星的多波段无源微波和红外辐射计，并在其他卫星星座的协助下估计得到每隔三小时全球降水观测。GSMaP 使用低轨卫星观测提供的微波数据集和地球同步卫星观测提供的可见光/红外数据集作为反演算法输入源，主要采用了云移动矢量法和卡尔曼滤波方法对源数据进行处理，生产了 3 种不同的遥感降水数据产品：GSMaP_NRT、GSMaP_MVK、GSMaP_Gauge，其中近实时数据产品 GSMaP_NRT 处理过程中采用的是前向云矢量运动方案，而标准产品 GSMaP_MVK 采用了双向（前向和后向）云矢量运动方案，GSMaP_Gauge 则是 基 于 GSMaP_MVK 与 CPC（Climate Prediction Center）全球地面雨量站观测资料的校正版本。3 套产品的时空分辨率均为 1 h 和 0.1°×0.1°[14]。对比评估显示，GSMaP日降水在众多卫星降水产品中准确率最高[5] ，因此本研究选取经过雨量站校正的GSMaP_Gauge作为初始场开展融合降水数据集研制。
\n(2) 实况降水：选取天山山区104个固态降水站（其中包含 57 个国家站，并已剔除 8 个国际交换站）、961个区域自动站逐时降水数据，台站分布情况如图1所示，实况降水时间范围与GSMaP降水数据一致。该数据由新疆气象局信息中心整理并通过气候极值检验、单站极值检验和数据一致性检验等质量控制。需要指出的是，固态降水站安装的是称重式降水测量仪器，既可测量降雨也可测量降雪，而区域自动站安装的是翻斗式雨量计，只能测量降雨。由于961个区域自动站雨量计在冷季停止观测，因此本研究选取暖季的5-9月开展研究。按照10折交叉验证，将实况站点按不同海拔分为10组，每次选取每组中的9份，共计90%用于建模，剩余每组中的1份共计10%组成独立数据集进行融合产品的精度验证，以此保证训练样本及验证样本的代表性。
", "ds_process_way": "本研究开展天山山区多源降水融合数据集研制，以GSMaP卫星降水数据为初始场，同期区域1065个台站的实况降水数据，发展一种基于最优插值的星地降水产品融合方法，最终生成2000-2022年天山山区逐日降水产品集｡本研究中最优插值分析以GSMaP降水作为初估场，以站点实况降水为真值，每个格点上的最终降水分析值Ak等于该点的初估值Fk加上该格点上实况观测值与初估值的偏差，而这个偏差由一定范围内n个格点上已知的实况观测值Oi与初估值Fi的偏差加权估计得到。
", "ds_quality": "本数据集在研制过程中对实况数据进行了严格质控，对逐日融合降水数据进行了质量评估，可望为复杂地形区域水资源管理与高效利用提供数据支撑。选取天山山区 104 个固态降水站、以及 961 个区域自动站逐时降水数据，实况降水时间范围与 GSMaP降水数据一致，该数据由新疆气象局信息中心整理并通过气候极值检验、单站极值检验和数据一致性检验等质量控制。此外，基于本数据集的融合方法对比研究成果已通过同行专家评审，发表在业内权威期刊《Journal of Hydrology》，表明数据具有较高的可信度。
", "ds_acq_start_time": "2000-01-01 00:00:00", "ds_acq_end_time": "2022-12-31 00:00:00", "ds_acq_place": "新疆天山山区", "ds_acq_lon_east": 96.0, "ds_acq_lat_south": 38.5, "ds_acq_lon_west": 73.5, "ds_acq_lat_north": 45.5, "ds_acq_alt_low": null, "ds_acq_alt_high": null, "ds_share_type": "apply-access", "ds_total_size": 7143, "ds_files_count": 4, "ds_format": "电子表格", "ds_space_res": "10km", "ds_time_res": "逐日", "ds_coordinate": "WGS84", "ds_projection": "", "ds_thumbnail": "468f31f6-9214-4d7d-acfa-965586b38ec7.jpg", "ds_thumb_from": 2, "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": "2024-08-13 15:03:58", "first_publish_time": "2024-08-13 15:03:58", "last_updated": "2025-03-07 15:45:03", "protected": false, "protected_to": null, "lang": "zh", "cstr": "33110.11.ariddc.00013", "license": null, "extra": null, "files_shape": [ { "name": "anaconda-ks.cfg", "size": 2349, "is_dir": false }, { "name": "initial-setup-ks.cfg", "size": 2397, "is_dir": false }, { "name": "微信截图_20221219174455.png", "size": 0, "is_dir": false }, { "name": "AAA", "size": null, "is_dir": true } ], "features": null, "data_level": 0, "i18n": { "en": { "title": "Multi-source precipitation fusion dataset in Tianshan region, Xinjiang (2000-2022)", "ds_abstract": "Tianshan Mountain is the farthest mountain system from the ocean in the world and one of the seven major mountain systems in the world. This region belongs to my country's typical alpine mountainous area and is known as the \"Central Asia Water Tower\". It is of great strategic significance to Xinjiang and even Central Asia. With the development of remote sensing technology, satellite precipitation retrieval has become an important means to estimate precipitation in mountainous areas. However, the mountainous terrain is complex and uneven, resulting in low accuracy of precipitation retrieval products in mountainous areas. In response to this problem, this study carried out the development of a multi-source precipitation fusion data set in the Tianshan Mountains. Using GSMaP satellite precipitation data as the initial field and the live precipitation data from 1065 stations in the region during the same period, we developed a satellite-ground precipitation product fusion method based on optimal interpolation, and finally generated a daily precipitation product set in the Tianshan Mountains from 2000 to 2022. During the development process, this dataset has strictly controlled the actual data and evaluated the quality of daily integrated precipitation data. It is expected to provide data support for water resources management and efficient utilization in complex terrain areas.
", "ds_source": "The production of this dataset is mainly based on GSMaP satellite precipitation and actual precipitation data from rainfall stations.
\n(1) GSMaP satellite precipitation: The Global Satellite Precipitation Map (GSMaP) provides global hourly rainfall with a resolution of 0.1 x 0.1 degrees. GSMaP is a product of the Global Precipitation Measurement Mission, which uses multi-band passive microwave and infrared radiometers from the GPM core satellite and estimates global precipitation observations every three hours with the assistance of other satellite constellations. GSMaP uses microwave data sets provided by low-orbit satellite observations and visible/infrared data sets provided by geosynchronous satellite observations as input sources of the inversion algorithm. It mainly uses cloud movement vector method and Kalman filtering method to process the source data, and produces 3 different remote sensing precipitation data products: GSMaP_NRT, GSMaP_MVK, and GSMaP_Gauge. Among them, the near-real-time data product GSMaP_NRT uses a forward cloud vector motion scheme in the processing process, while the standard product GSMaP_MVK uses a two-way (forward and backward) cloud vector motion scheme. GSMaP_Gauge is a corrected version based on GSMaP_MVK and CPC (Climate Prediction Center) global surface rainfall station observation data. The spatio-temporal resolutions of the three sets of products are all 1 h and 0.1°×0.1°[14]. Comparative evaluation shows that GSMaP daily precipitation has the highest accuracy among many satellite precipitation products [5]. Therefore, this study selects GSMaP_Gauge corrected by rainfall stations as the initial field to develop a fused precipitation data set.
\n(2) Live precipitation: Hour-by-hour precipitation data from 104 solid precipitation stations in the Tianshan Mountains (including 57 national stations and 8 international exchange stations have been excluded) and 961 regional automatic stations are selected. The distribution of stations is as shown in Figure 1. The live precipitation time range is consistent with the GSMaP precipitation data. The data was compiled by the Information Center of the Xinjiang Meteorological Administration and passed quality controls such as climate extreme value test, single station extreme value test and data consistency test. It should be pointed out that solid precipitation stations are equipped with weighing precipitation measuring instruments, which can measure both rainfall and snowfall, while regional automatic stations are equipped with tipping bucket rain gauges, which can only measure rainfall. Since 961 regional automatic stations stop observing rain gauges during the cold season, this study selects May to September during the warm season for research. According to the 10-fold cross-verification, the live stations are divided into 10 groups according to different altitudes. Nine samples from each group are selected each time, with a total of 90% used for modeling. The remaining one sample from each group totaling 10% constitutes an independent data set. Accuracy verification of fusion products to ensure the representativeness of training samples and verification samples.
", "ds_process_way": "This study carried out the development of a multi-source precipitation fusion data set in the Tianshan Mountains. Using GSMaP satellite precipitation data as the initial field and the live precipitation data from 1065 stations in the region during the same period, a satellite-ground precipitation product fusion method based on optimal interpolation was developed, and finally generated A daily precipitation product set in the Tianshan Mountains from 2000 to 2022. The optimal interpolation analysis in this study uses the GSMaP precipitation as the preliminary estimation field, and the actual precipitation at the station as the true value. The final precipitation analysis value Ak at each grid point is equal to the initial estimation Fk at that point plus the deviation between the actual observation value and the initial estimation at that grid point. This deviation is weighted estimated from the deviation between the known actual observation value Oi and the initial estimation Fi at n grid points within a certain range.
", "ds_quality": "During the development process, this dataset has strictly controlled the actual data and evaluated the quality of daily integrated precipitation data. It is expected to provide data support for water resources management and efficient utilization in complex terrain areas. Hourly precipitation data from 104 solid precipitation stations and 961 regional automatic stations in the Tianshan Mountains are selected. The actual precipitation time range is consistent with the GSMaP precipitation data. The data was compiled by the Information Center of the Xinjiang Meteorological Administration and passed the climate extreme value test, single station extreme value test and data consistency test. In addition, the comparative research results of fusion methods based on this dataset have passed peer expert review and were published in the industry's authoritative journal \"Journal of Hydrology\", indicating that the data has high credibility.
", "ds_acq_place": "Tianshan Mountains of Xinjiang", "ds_format": "spreadsheet", "ds_space_res": "10km", "ds_time_res": "daily" } }, "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": [ "降水" ], "ds_subject_tags": [ "地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "新疆", "天山" ], "ds_time_tags": [ 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": [ "卢新玉", "王秀琴", "刘艳", "伏晓慧", "火红" ], "ds_meta_authors": [ "卢新玉" ], "ds_managers": [ "卢新玉" ], "category": "气象" }