cassini-team-todo/eea-waterbase-cleaned

该数据集是EEA Waterbase – Water Quality ICM (WISE-4) v2018.1的过滤和清理子集，专门为训练Sentinel-2水质模型而准备。数据来源于第11届CASSINI Hackathon – EU Space for Water活动。数据集包含2015年7月1日至2017年12月31日期间欧洲河流的水质测量数据，筛选标准包括：仅限河流水体（参数水体类别为RW）、站点必须具有有效的经纬度坐标、时间窗口与Sentinel-2A卫星可用期对齐，以及五个与Sentinel-2表面反射率关联最强的光学参数（溶解氧、氧饱和度、硝酸盐、铵、磷酸盐-P）。数据集提供四个主要文件：1) station_river_bboxes.geoparquet：包含4,742个监测站点的空间信息，每个站点匹配到最近的HydroRIVERS河段，并生成沿河流形状的约5公里走廊多边形，用于Sentinel-2查询窗口；2) measurements_optics5_long.parquet：长格式测量表，包含251,245行个体测量数据，涉及4,757个站点和五个光学参数；3) station_index_optics5.parquet：每个合格站点的摘要统计信息；4) station_year_coverage_optics5.parquet：审计表，按站点、年份和参数代码统计测量数量。预期用途是通过将站点和日期与Sentinel-2影像结合，训练水质模型，支持环境监测和水文学研究。数据源自欧洲环境管理局（EEA），遵循EEA重用政策。This dataset is a filtered and cleaned subset of EEA Waterbase – Water Quality ICM (WISE-4) v2018.1, specifically developed for training Sentinel-2 water quality models. The data is derived from the 11th CASSINI Hackathon – EU Space for Water event. It includes water quality measurement data of European rivers spanning from July 1, 2015 to December 31, 2017, with the following screening criteria: only river water bodies (water body category parameter marked as RW), stations must possess valid latitude and longitude coordinates, the temporal window is aligned with the operational period of Sentinel-2A satellite, and the five optical parameters that exhibit the strongest correlation with Sentinel-2 surface reflectance (dissolved oxygen, oxygen saturation, nitrate, ammonium, phosphate-P). The dataset provides four core files: 1) station_river_bboxes.geoparquet: Contains spatial information of 4,742 monitoring stations, where each station is matched to the nearest HydroRIVERS river segment, and a ~5 km corridor polygon along the river's course is generated for the Sentinel-2 query window; 2) measurements_optics5_long.parquet: A long-format measurement table with 251,245 rows of individual measurement records, covering 4,757 stations and the five aforementioned optical parameters; 3) station_index_optics5.parquet: Summary statistics for each qualified monitoring station; 4) station_year_coverage_optics5.parquet: An audit table that tallies the number of measurements grouped by station, year and parameter code. The intended application of this dataset is to train water quality models by correlating station locations and measurement dates with Sentinel-2 imagery, thereby supporting environmental monitoring and hydrological research. The data is sourced from the European Environment Agency (EEA) and is subject to the EEA Reuse Policy.