A comprehensive study of Surface Water and Ocean Topography (SWOT) Pixel Cloud data for flood extent extraction

Current disaster and emergency management services produce flood maps within hours using satellite data. To handle large-scale events efficiently, a reliable automated method is needed to generate an initial flood extent map, which can then be refined manually. Launched in December 2022, the Surface Water and Ocean Topography (SWOT) satellite, equipped with the Ka-band Radar Interferometer (KaRIn), provides high-resolution radar observations used here for flood detection. While not initially designed for detailed flood mapping in vegetated or urban regions, the performance of SWOT’s Pixel Cloud products was assessed during four major flood events in Greece, France, Brazil, and the USA. Each event is paired with Sentinel-1 or Sentinel-2 imagery within a 3-hour time frame, providing a valuable opportunity to compare and evaluate SWOT’s flood detection capabilities. Three radar variables of the Pixel Cloud products are studied for extracting flood extents: σ 0 , coherent power, and interferometric coherence — which is computed from the two complex interferograms. They are compared to the built-in classification and flood masks computed from Sentinel-1/2. The study demonstrates the capabilities of the SWOT satellite in detecting flooded vegetation, flooded urban areas, and even regions with high snow cover. However, limitations are observed: (1) when high soil moisture is observed, causing signal saturation, (2) SWOT can be sensitive to the incidence angle, both of which lead to less reliable flood extent estimation. These findings highlight the potential of SWOT satellite for improving global flood mapping, as well as the need for further exploration to address current limitations and enhance flood monitoring capabilities in the near future. • SWOT Pixel Cloud studied on four flood events with Sentinel-1 & 2 as comparison. • SWOT built-in classification gives good results for the studied flood event. • Water masks are computed using histogram analysis and thresholds on three variables. • SWOT Pixel Cloud detects water under vegetation and in urban areas. • SWOT radar variables are sensitive to snow cover, soil moisture, and incidence angle.