Flood Detection and Cropland Damage Assessment Using Sentinel-1 SAR Time Series Multi-Source EO Data in the Sebou Basin, El Gharb plain, Morocco
This study utilizes Sentinel-1 SAR time series data with VV and VH polarizations to develop a robust, anomaly-based flood detection method that effectively maps inundation and assesses cropland damage in Morocco's Sebou Basin, ultimately informing targeted risk management strategies for the El Gharb plain.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Flooding is a force of nature that has shaped human history, but in the modern world, its impact on agriculture is measured in lost harvests and threatened food security. When heavy rains fall on flat, fertile land, water can spread quickly, drowning crops and turning productive fields into temporary lakes. For farmers and governments, knowing exactly where the water is and how much damage it has caused is critical, yet it is often the hardest thing to determine. Traditional cameras on satellites cannot see through the thick clouds that usually accompany storms, leaving a blind spot precisely when monitoring is most needed. To solve this, scientists have turned to a different kind of eye: radar. Unlike optical cameras that rely on reflected sunlight, radar systems send out their own microwave signals and listen for the echo. These signals can pierce through clouds and darkness, bouncing off the ground to reveal what lies beneath the storm. By tracking how these signals change over time, researchers can distinguish between dry land, wet soil, and standing water, turning a chaotic natural event into a clear, measurable map.
In the Sebou Basin of northern Morocco, a region known as the El Gharb plain, this technology was put to work to understand a severe flooding event that struck in the winter of 2025 and 2026. This area is a low-lying, flat expanse of rich alluvial soil, heavily cultivated with wheat, olives, and citrus, making it a vital engine for the country's food supply. However, its flat topography and extensive irrigation networks make it prone to water accumulation when heavy rains arrive. A team of researchers from the International University of Rabat and Politecnico di Milano set out to map the extent of the flood and, more importantly, to calculate exactly how much farmland was submerged. They utilized data from Sentinel-1, a European satellite equipped with Synthetic Aperture Radar, which captures images of the Earth's surface day and night, regardless of weather.
The researchers did not simply look at a single image taken during the flood. Instead, they built a timeline of the region's conditions. They began by analyzing rainfall data to identify the weeks leading up to the disaster, noting how the soil became increasingly saturated. They then compared radar images taken before the flood with those taken during the peak of the inundation. The core of their method relied on a physical principle: when radar waves hit open water, they bounce away in a specific direction, creating a very dark signal in the image. When they hit rough ground or crops, the signal bounces back more strongly. By measuring the difference in this "backscatter" between the dry period and the wet period, the team could pinpoint where the water had settled. They paid special attention to a specific type of radar signal, known as cross-polarization, which proved to be exceptionally sensitive to the presence of water among vegetation, helping to separate flooded crops from dry fields that might look similar in other types of data.
To ensure their maps were accurate, the team applied a series of logical filters. They knew that water does not flow uphill, so they used a digital map of the terrain to exclude any areas with steep slopes or high elevations where flooding was physically impossible. This step removed false alarms caused by dark patches of bare soil or shadows that might otherwise trick the computer into thinking they were water. Once they had a clean map of the flood's extent, they overlaid it with a detailed land-use map that identified every patch of cropland in the region. This allowed them to calculate the precise area of agricultural land that had been underwater. The results were stark: the flood covered approximately 207,112 hectares of land in total. Of that vast area, nearly 149,000 hectares were active farmland, representing a massive blow to the local agricultural economy.
The study also looked at the aftermath by comparing the radar data with images from an optical satellite that captures visible light and near-infrared light. By measuring the "greenness" of the vegetation before and after the flood, they found a clear drop in plant health exactly where the radar had detected water. This confirmed that the flood had not just covered the fields but had stressed the crops, likely causing significant yield loss. The researchers found that the damage was most severe in the lowest parts of the plain, where water naturally collects and drains slowly. They observed that the flood did not spread randomly; it followed the natural low-lying corridors of the river system, repeatedly hitting the same vulnerable zones.
This work demonstrates that radar technology, when combined with smart computer analysis, can provide a reliable, all-weather view of flood disasters. The team showed that by focusing on how radar signals change over time rather than just looking at a single snapshot, they could map floods with high precision even in complex agricultural landscapes. The findings suggest that the El Gharb plain faces a recurring risk in its low-lying, intensively farmed zones. The researchers argue that this level of detail is essential for planning. They propose that future management should focus on building better water retention systems, improving soil conservation to slow down runoff, and using these flood maps to guide where crops should be planted. By understanding exactly where the water goes and how much land it touches, authorities and farmers can make better decisions to protect the region's food production from the next inevitable storm.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.