The dynamics of surface water in the Yangtze Plain is complex, influenced by the coupled impacts of climate change and intensifying human activities. However, remote sensing observations often encounter challenges in this region due to persistent cloud cover, impeding comprehensive studies of water dynamics. This study introduces a novel Monthly Surface Water Mapping (MSWM) approach combining time series Sentinel-1&2 images, resulting in the generation of a Monthly Surface Water Extent (MSWE) dataset. This dataset boasts a spatial resolution of 10 m and a temporal resolution of one month. Validation results indicate the MSWE exhibits a significant improvement of 19.6 % and 8.9 % in F1 score compared to the temporally-aligned Global Surface Water dataset and thresholding results, respectively. The MSWE demonstrates robust spatial precision and temporal tracking capabilities, even in complex scenes and cloudy conditions. The seasonal fluctuation of surface water bodies in the Yangtze Plain was computed using the monthly dataset and a harmonic analysis model. The results characterized distinct monthly change patterns for surface water extent, allowing for the identification and quantification of four lake classes: 6 seasonal lakes, 11 weak seasonal lakes, 21 generally stable lakes, and 46 stable lakes. The MSWM stands out for its capacity to estimate surface water extent regardless of weather conditions, showcasing promising potential for extension to other regions characterized by constant cloud cover. Furthermore, the availability of a monthly water dataset contributes significantly to enhancing our spatiotemporal understanding of surface water dynamics, offering substantial benefits for sustainable water resources management.

中文翻译：

---

根据Sentinel-1和2观测对长江平原地表水范围的估计和动态


受气候变化和人类活动加剧的双重影响，长江平原地表水动态复杂。然而，由于该地区持续的云层覆盖，遥感观测经常遇到挑战，阻碍了水动力学的综合研究。本研究引入了一种新颖的每月地表水测绘 (MSWM) 方法，该方法结合了时间序列 Sentinel-1 和 2 图像，从而生成了每月地表水范围 (MSWE) 数据集。该数据集空间分辨率为10 m，时间分辨率为1个月。验证结果表明，与时间对齐的全球地表水数据集和阈值结果相比，MSWE 的 F1 分数分别显着提高了 19.6% 和 8.9%。即使在复杂场景和多云条件下，MSWE 也表现出强大的空间精度和时间跟踪能力。利用月度数据和调和分析模型计算了长江平原地表水体的季节波动。结果显示了地表水范围明显的每月变化模式，从而可以识别和量化四个湖泊类别：6 个季节性湖泊、11 个弱季节性湖泊、21 个总体稳定的湖泊和 46 个稳定的湖泊。 MSWM 以其在不受天气条件影响的情况下估算地表水范围的能力而脱颖而出，展示了扩展到以持续云层覆盖为特征的其他地区的巨大潜力。此外，每月水数据集的可用性极大地有助于增强我们对地表水动态的时空了解，为可持续水资源管理带来巨大好处。