Leveraging SWOT's water elevation pixel cloud to comprehend analyse the spatial dynamics of flood events

Satellite data have been used for over 40 years, along with airborne and in situ measurements, for monitoring extreme hydrological events, and enabled major progress in our understanding of floods. The available satellite data have long been mostly limited to imagery (SAR, optical, and thermal) providing a map of the flood extent and conventional nadir altimetry providing a 1-dimensional water elevation along the satellite ground track. Since its launch in late 20232, SWOT has opened a new dimension in space altimetry by providing two-dimensional maps of water elevation. Its main instrument is a near-nadir, bistatic, Ka-band SAR altimeter that uses interferometry to measure the elevation of the water pixels (10-60x22m resolution). Although its revisit time (at least twice per 21-day nominal cycle up to 78° latitude) and spatial resolution limits its usability for operational flood monitoring, SWOT opens new perspectives in the understanding of flood dynamics, particularly if used in synergy with high-resolution imagery and real-time in situ measurements. Indeed, water elevation maps can be used to calibrate and validate hydraulic models through their comparison with the elevation of the modeled free surface at the corresponding point in time. In addition, estimations of the river flows are part of the standard SWOT products distributed on the PODAAC and hydroweb.next platforms.

While the early results on recent flood events demonstrated the utility of the SWOT data for understanding the dynamics of floods, research efforts are still needed to fully leverage its scientific and socioeconomic benefits. On the one hand, there is a scope for improvement in the production of the water elevation pixel cloud from the SLC images: the baseline data processing is dedicated to lakes and river monitoring, and custom processing for flood events may improve the quality of the water elevation data in flooded areas. On the other hand, due to their relative novelty, further adaptations will be needed to operationalize their use for key applications (e.g., more accurate modeling of floods to engineer flood-risk infrastructure, assimilation in operational hydraulic models along with other sources of data, improved risk assessment on buildings through better forecasting of water levels,...). Further research works will be able to draw on SWOT's open data, including the calibration and validation phase, which lasted from end of  March to early July 2023 on selected orbits with a 1-day repeat cycle. This phase enabled SWOT acquisitions every 24 hours for multiple flood events, including the flooding caused by the destruction of the Kakhovka Dam in Ukraine. This high temporal revisit allows for fine-scale analysis of the temporal evolution of the water elevation of the flooded area.

In our contribution, we will present early results on selected examples of flood events, and some scientific and technical issues that we believe to be of particular interest.