CNSW 1.0: Prefectural Reconstruction of China's Surface Water Resources Using Machine Learning Methods

A comprehensive and long-term dataset of prefectural surface water resources is crucial for effective water resource management in China. However, there has been a significant gap in the availability of such datasets, with no existing datasets providing comprehensive long-term coverage. To address this gap, we have developed CNSW 1.0, the first long-term (2000-2020) dataset of prefectural surface water resources in China. Utilizing surface water resource data from official water resource bulletins, we employed 14 advanced machine learning models to reconstruct the CNSW 1.0 dataset. The resulting dataset exhibits high accuracy, with an R² of 0.978 for total surface water resources and acceptable level of bias across China. CNSW 1.0 not only outperforms existing datasets like CNRD v1.0, GRUN, and ISIMIP in terms of simulation accuracy and spatial distribution but also fills a critical gap in water resource data for China. This dataset is expected to be an invaluable tool for developing more informed water resource management strategies at the administrative level in China, particularly in the context of climate change. The CNSW 1.0 encompasses surface water resource data for 341 prefectural-level administrative units in mainland China over a period from 2000 to 2020, comprising a total of 16 surface water datasets. The data are stored in both CSV and shapefile (SHP) formats. The naming convention for annual spatial distribution shapefiles follows the structure "CNSW_1.0_YYY_ZZZZ.shp," where "YYY" represents the model’s name, and "ZZZZ" denotes the year. The naming convention for multi-year average spatial distribution shapefiles is "CNSW_1.0_mean_YYY.shp," and for the spatial evolution trend shapefiles, it is "CNSW_1.0_trend_YYY.shp." A total of 368 shapefiles are provided, which approximately take up 4.26GB of disk space. Additional information, such as region, province, and prefecture area, is also stored in the files.

全面且长期的地级地表水资源数据集，是中国开展高效水资源管理的核心支撑。然而当前此类数据集存在显著缺口，尚无公开数据集可实现全面的长期覆盖。为填补这一研究空白，我们研发了中国首个地级地表水资源长期数据集CNSW 1.0，其时间覆盖范围为2000年至2020年。本研究依托官方水资源公报发布的地表水资源数据，采用14种先进机器学习模型重构得到CNSW 1.0数据集。该数据集具备高精度表现：全国尺度下地表水资源总量的决定系数（R²）达0.978，偏差水平处于合理区间。CNSW 1.0不仅在模拟精度与空间分布表现上优于CNRD v1.0、GRUN及ISIMIP等现有数据集，更填补了中国水资源数据领域的一项关键空白。该数据集有望成为中国各级行政区制定更科学合理的水资源管理策略的宝贵工具，尤其在气候变化研究背景下极具应用价值。CNSW 1.0涵盖了2000年至2020年间中国大陆341个地级行政区的地表水资源数据，共计包含16类地表水资源数据集。数据集采用CSV与矢量形状文件（SHP）两种格式存储。年度空间分布矢量形状文件的命名规则遵循"CNSW_1.0_YYY_ZZZZ.shp"格式，其中"YYY"代表所用机器学习模型的名称，"ZZZZ"代表数据年份。多年平均空间分布矢量形状文件的命名规则为"CNSW_1.0_mean_YYY.shp"，空间演化趋势矢量形状文件的命名规则则为"CNSW_1.0_trend_YYY.shp"。本次共提供368个矢量形状文件，总存储空间约为4.26GB。数据文件中还附带存储了区域、省份及地级行政区面积等补充信息。