-
-
-
Archer, L.,
-
Hawker, L.,
-
-
-
-
Rougier, J.
New research led by the University of Bristol Professor Jeff Neal (Fathom’s Chief Scientific Advisor) and Fathom’s Senior Scientific Developer Dr Stephen Chuter and co-authored by several other Fathomers, presents a new model for estimating river channel depth based on data gathered by the Surface Water and Ocean Topography (SWOT) mission.
Accurate flood risk management and mapping needs detailed data on river-floodplain topography, river depth (bathymetry) and measurements of the floodplain topography. This information is critical for model accuracy and the accuracy of models is key for decision-makers in a range of settings, from insurance to banks and event response. However, it is costly to directly measure river bathymetry, leaving only a tiny fraction of floodplains covered by high-accuracy models.
In this UK Space Agency research project, under its enabling technologies program, scientists have developed a way to estimate river bathymetry based on data from the SWOT satellite. The researchers then use the resulting SWOT derived bathymetry estimates to model flooding for the river Severn in the UK, finding their approach to be of similar accuracy to a model based on data from local river surveying.
Fathom and the SWOT mission
The Surface Water and Ocean Topography (SWOT) satellite was launched in December 2022 from Vandenberg Air Force Base in California. A joint development between NASA and Centre National D’Etudes Spatiales (CNES), with contributions from the UK and Canadian space agencies, Fathom’s own Chairman, Professor Paul Bates was one of the lead scientists at the mission’s conception and is still heavily involved with the project.
Read all about this game-changing mission and Fathom’s contribution from the beginning, in our insight:
Explaining SWOT: A satellite survey of oceans and surface waters
The research: A new way to estimate river bathymetry
The SWOT mission launched in December 2022, has provided unparalleled satellite observations of river water surface topography and extent for the first time. Its instruments are able to measure highly variable water surfaces, scanning 120-km swathes of ground at a time and revisiting every 21 days, observing rivers of 50m–100m in width, The research team used data from SWOT to develop a new way to estimate the channel depth in the absence of local observations of river bathymetry. They:
- Use SWOT data on water height to estimate river bathymetry;
- Develop a SWOT-derived inundation model incorporating these bathymetry estimates;
- Apply the method to the test case of the River Severn (a river of 40–70 meters in width and with lots of available validation data);
- Benchmark their results to a terrestrial flood inundation model derived from in situ LiDAR and sonar bathymetry elevation data; and
- Compare inundation extents from these models simulating the autumn 2000 floods to observed flood extents for the same event (observed using airborne synthetic aperture radar, SAR).
Key findings: SWOT can exceed mission requirements for smaller rivers
Overall, the SWOT-based flood inundation model proved accurate and able to improve flood model results, demonstrating how water surface height observations could be used to infer river bathymetry.
The model derived plausible results for riverbed elevations, but generally overestimated water levels compared to measurements from river gauging stations along the Severn (by 0.16–0.89 meters, depending on how strict the SWOT data was filtered ) – something the researchers attribute to the misclassification of higher-elevation adjacent riverbanks as water surfaces by SWOT’s sensor. However, overall coverage of the river system was comprehensive and approaching the accuracy of the original mission design, indicating that SWOT not only meets but can exceed its mission requirements for smaller rivers.
The largest errors occurred in regions of low or complicated flow, but this could be rectified by carefully selecting the bathymetry points around areas such as weirs, say the researchers. The fact that SWOT was able to observe the River Severn’s weir complexes was unexpected given their small scale, and a demonstration of how well the mission was performing. The researchers estimated river bathymetry every kilometer along the Severn, but concluded that surprisingly few locations – only five – were likely needed to achieve most of the model’s accuracy.
To account for noise and uncertainties in the SWOT observations, the bathymetry estimation model trialled a tulip-shaped “misfit function” (a kind of function that determines the difference between SWOT observations and the fitted water surface profile). This treats errors differently depending on the magnitude of their misfit relative to their observation error– and outperformed alternatives.
Considering the flooding on the Severn seen in autumn 2000, a flood inundation model conditioned with the SWOT derived river channel and a LiDAR observed floodplain topography returned remarkably similar results to observations, with a mean absolute difference in water levels across the entire flood event of less than 0.31m, lower errors and a critical success index of 0.88, a measure of statistical accuracy ranging from 0 (no agreement) to 1 (perfect agreement). When benchmarked against a model for the same event, built using observed sonar bathymetry, the SWOT-derived model performed almost equally.
Want to know more about this research? Read the full paper published in Water Resources Research.
Flood inundation modeling using the Surface Water Ocean Topography mission
New research presents a model for estimating river channel depth based on data gathered by the Surface Water and Ocean Topography (SWOT) mission.
Quantifying UK coastal flood exposure under future sea-level rise to 2300
How will sea level rise affect flood risk to populations and critical infrastructure in the UK in the next few centuries? A study by the UK Met Office, Bristol University and Fathom has found that the number of people exposed to flood by 2300 could rise dramatically – but global reductions in greenhouse gas emissions could reduce these numbers.
Automated tail-informed threshold selection for extreme coastal sea levels
A new method, TAILS, sharpens how we identify extreme events by focusing on the most critical data, giving clearer, more reliable estimates of future coastal hazards worldwide.
Introducing BURGER, Fathom’s new extreme rainfall dataset
New research by Fathom’s Senior Hydrologist, Dr Jannis Hoch, and other Fathom scientists describes Fathom’s new dataset, Bottom Up Regionalised Global Extreme Rainfall, or BURGER for short. It’s the first of its kind to be created from rainfall gauge data and the first to contain extreme estimates for durations as short as an hour.
