A robust large-scale surface water mapping framework with high spatiotemporal resolution based on the fusion of multi-source remote sensing data

Abstract

Large-scale and dynamic surface water mapping is crucial for understanding the impact of global climate change and human activities on the distribution of surface water resources. Remote sensing imagery has become the primary data source for surface water mapping due to its high spatiotemporal resolution and wide coverage. However, the reliability of current water products during flood seasons is limited due to the influence of clouds on optical remote sensing images. Moreover, annual and seasonal surface water mapping cannot capture intra-month variations of water bodies. To address these challenges, we proposed a high spatiotemporal surface water mapping framework on Google Earth Engine that combines multi-source remote sensing data. Our framework can generate 10 m spatial resolution surface water maps at a 15-day time step. We classified water bodies using Sentinel-2 images and a classification tree algorithm, and then used Sentinel-1 data to compensate for cloudy and missing data areas in Sentinel-2 images, resulting in seamless cloud-unaffected surface water maps. We evaluated the effectiveness of our proposed framework in six floodplains around the world, and experimental results demonstrate that the water maps generated by our framework outperform existing public datasets and our framework has great potential for hydrological applications. Our proposed framework can capture the details of surface water dynamics with higher spatial and temporal resolution and is free from cloud influence, which is necessary for water resources management, flood monitoring, and disaster response.

Keywords

• Surface water
• Google Earth Engine
• Remote sensing
• Sentinel-1
• Sentinel-2