A global permanent and seasonal reference water product for improved flood mapping: First results of a Sentinel-1/2 based methodology

Abstract

The inundation extent derived from Earth Observation data is one of the key parameters in successful flood disaster management. This information can be derived with increasing frequency and quality due to a steadily growing number of operative satellite missions and advances in image analysis. In order to accurately distinguish flood inundation from “normal” hydrologic conditions, up-to-date, high-resolution information on the seasonal water cover is crucial. This information is usually neglected in disaster management, which may result in a non-reliable representation of the flood extent, mainly in regions with highly dynamic hydrological conditions. In this study, an automated approach for computing a global reference water product at 10 m spatial resolution is presented, specifically designed for the use in global flood mapping applications. The proposed methodology combines existing processing chains for flood mapping based on Copernicus Sentinel-1 and Sentinel-2 data and calculates permanent as well as monthly seasonal reference water masks over a reference time period of two years. As more detailed mapping of water bodies is possible with Sentinel-2 during clear-sky conditions, this optical sensor is used as primary source of information for the generation of the reference water product. In areas that are continuously cloud-covered complementary information from the Sentinel-1 C-Band radar sensor is used. In order to provide information about the quality of the generated reference water masks, we incorporate an additional quality layer, which gives information on the pixel-wise number of valid Sentinel-2 observations over the derived permanent and seasonal reference water bodies within the selected reference time period. Additionally, the quality layer indicates if a pixel is filled with Sentinel-1 based information in the case that no valid Sentinel-2 observation is available. The reference water product is demonstrated in five study areas in Australia, Germany, India, Mozambique, and Sudan, distributed across different climate zones. Our outcomes are systematically cross-compared with already existing external reference water products. Further, the proposed product is exemplarily applied to three real flood events. The results show that it is possible to generate a consistent reference water product that is suitable for the application in flood disaster response. The proposed multi-sensor approach is capable of producing reasonable results, even if only few or no information from optical data is available. Further, the study shows that the consideration of seasonality of water bodies, especially in regions with highly dynamic hydrological and climatic conditions, is of paramount importance as it reduces potential over-estimations of the inundation extent and gives users a more reliable picture on flood-affected areas.