Using satellites to monitor river ice for flood early warning
Published: 6 May 2020
Early warning of ice jams and flooding can reduce the risk to local population and allow hazard mitigation. However, these events are very difficult to accurately model and forecast. Monitoring of river ice therefore plays a very important role in early warning.
River ice monitoring relies on multiple data sources, including flying planes over the rivers get an overview. Satellite data offers flood forecasters a lower-cost, wider-coverage and higher-frequency way to monitor river ice. Using near real-time satellite data and new methods of river ice detection helps to detect and classify river ice in an early stage.
Using radar data
The River Ice Teams for the Canadian provinces of Alberta and Quebec are working together with Deltares to set-up operational access to near real-time satellite data and develop methods to detect and classify river ice. Synthetic Aperture Radar (SAR) is one type of data being used to monitor and detect river ice. SAR satellites offer sensitivity to the physical structure of river ice and can be used to differentiate between types of ice formations. River ice can be detected using the Sentinel-1 SAR imagery which when classified can show the progression and accumulation of river ice.
The two images show the path of the Athabasca River through Fort McMurray. Two SAR images were processed to show which parts of the river are covered by ice (white), and which parts are open water (blue). Between April 23rd and April 28th, ice consolidated and jammed in the river bend at Fort McMurray. This caused river levels to rise behind and around the ice, causing ongoing flooding. On 28th and 29th of April were two cloudless days, which coincided with an overpass of Sentinel 2A and 2B satellites. Using a false composite image, the extent of the flooded area becomes clearly visible.
New opportunities for river ice monitoring
The Canadian RADARSAT Constellation Mission (RCM), launched in June 2019, offers new opportunities for river ice monitoring. This constellation of satellites can also cover large areas and has more frequent overpass compared to previous RADARSAT missions. Given that RCM was only launched last year, little is known about its potential in ice classification. This research aims to develop ice classification methods, with a focus on how the orientation of the radar influences the detection of different types of ice.
Integrated in Delft-FEWS
The research project is supported by Deltares and supervised by other experts from Natural Resources Canada, TU Delft. River ice classification methods from this initiative will be integrated into operational flood early warning systems designed by Deltares, providing a customized overview of satellite data over the rivers of interest.
Advanced modelling of river ice
This project provides better monitoring of river ice for flood forecasters and reservoir operators in cold regions and demonstrates a near real-time application of satellite data for decision-making. Satellite-derived information will improve the quality of information available and can eventually be used for advanced modelling of river ice.