New Tool Utilizes AI to Predict Flooding Risks in Jammerbugt Municipality
As climate change continues to have a profound impact on various regions across the globe, the threat of flooding looms large for many residents, including those in Jammerbugt Municipality. In the coming years, the risk of flooding in areas like summer houses and agricultural land is expected to increase significantly, particularly following intense and sustained periods of rainfall.
Concerns About Flooding
Residents like Susanne Nielsen are well aware of the potential dangers posed by flooding, especially in low-lying areas close to bodies of water. Susanne, who looks after her parents’ summer house near Jammerbugt Bay, acknowledges the looming threat of rainwater seeping into their property due to high groundwater levels.
“The possibility of flooding is always at the back of our minds, especially during heavy downpours,” she remarks.
Early Warning System
To combat this growing risk, researchers from DTU have collaborated with Jammerbugt Municipality to develop an innovative early warning tool. This tool utilizes artificial intelligence to analyze various data sources, such as satellite imagery, weather forecasts, and water levels, to predict potential flooding events up to 48 hours in advance. It is the first of its kind to offer localized flood warnings, providing residents and decision-makers with valuable time to prepare and respond accordingly.
Artificial Intelligence in Action
The heart of this early warning tool lies in a sophisticated model known as the “wet index,” which leverages artificial intelligence to make precise calculations regarding flood risks. By integrating AI algorithms with comprehensive data on hydrological dynamics, the tool can accurately simulate the movement and accumulation of water in the region’s open landscapes.
Roland Löwe, an Associate Professor at DTU specializing in hydrology, explains that the AI model incorporates a deep understanding of how water interacts with its surroundings, allowing for more nuanced flood predictions.
Challenges and Successes
Following a trial run in 2023, Jammerbugt Municipality witnessed both successes and setbacks with the early warning tool. While the tool excelled in forecasting flooding during the wet spring months, it struggled to make accurate predictions during the drier summer period. This discrepancy was attributed to the lack of summer-specific data in the training set, as satellites have difficulty capturing water beneath dense vegetation.
Despite these challenges, the developers remain optimistic about the tool’s potential, emphasizing the importance of continual improvement and refinement to enhance its reliability and accuracy.
By staying vigilant and proactive in addressing flood risks, communities like Jammerbugt Municipality can better safeguard their residents and assets against the unpredictable forces of nature.