Northumberland County Council is undertaking a pilot project, 'FloodAI', to innovatively enhance flood and coastal resilience. This project is funded by Defra as part of the £200 million Flood and Coastal Innovation Programme
Overview
Northumberland County Council is undertaking a pilot project, 'FloodAI', to innovatively enhance flood and coastal resilience. This project is funded by Defra as part of the £200 million Flood and Coastal Innovation Programmes which is managed by the Environment Agency. The programmes will drive innovation in flood and coastal resilience and adaptation to a changing climate.
The research project aims to trial a flood prediction service for six catchments within the county that are particularly vulnerable to flash floods. Regions susceptible to these rapid inundations often lack suitable flood warning systems, as existing methods struggle with insufficient data, the inherent unpredictability, and the rapid response times demanded. Traditional models are often inadequate for areas prone to flash flooding, and relying on them can hinder the ability to mitigate harm to residents and their properties. Consequently, the exploration of alternative technologies is essential for creating effective flood warnings.
In collaboration with Northumbria University, Arup, Isle Utilities, and the Environment Agency Northumberland County Council has secured funding from Defra as part of the Flood and Coastal Innovation Programme, which is managed by the Environment Agency. This programme aims to address both current and future flood risks originating from the sea, rivers, groundwater, and surface water. The council has assumed the role of project manager, and each partner organisation has a specific function within the project, ensuring all elements work together for its successful delivery. Funding for the project is secured until 2027, and the council is currently investigating alternative funding models to ensure its continuation.
AI flood prediction model tools are being developed by Northumbria University, offering several advantages over traditional hydraulic models. These models can learn intricate patterns and relationships from extensive datasets, enabling them to process more diverse and dynamic data inputs as required by sensors, such as real-time sensor data and weather forecasts. The technology is designed with transparency in mind to help build confidence in the model's predictions.
The technology
Through a collaborative partnership, Northumberland County Council has harnessed regional expertise by joining forces with academics at Northumbria University. Professors specialising in machine learning have developed the technology, in collaboration with PhD students. The disciplines of machine learning, flood risk management, User Experience design, and digital service development have been combined to create an AI-driven flood forecasting service. To train the models, data has been gathered from various sources, including the Met Office, the Environment Agency, and hydrological models. This combination of data sources simulates conditions in areas where sensors were previously unavailable, enabling the models to learn and improve their forecasting accuracy.
As the AI-based model can utilise machine learning to identify complex patterns and relationships within large datasets, the technology is capable of handling diverse data formats, such as real-time sensor readings and climate projections, which will facilitate flash flood identification. This capability was not achievable using traditional hydraulic methods. The technology has been designed to ensure stakeholders can understand and trust the model’s predictions by transparently illustrating how different factors influence the forecast.
The technology is currently at the alpha stage and will soon be ready for limited piloting. A small group of individuals within the six catchment areas will be provided with private login credentials to access the technology. This initial phase will then inform further iterations of the technology based on testing and feedback.
Partnership approach
A close partnership approach has been central to this project. Defra and the Environment Agency manage the Flood and Coastal Resilience Innovation Programme, ensuring findings are disseminated and lessons can be learned across the public sector. Northumberland's "FloodAI" pilot is one of 25 local projects demonstrating how practical, innovative actions can improve resilience to flooding and coastal erosion in a changing climate. Each of these projects is managed by a council, fostering a supportive cohort of local authorities. The involved councils collaborate, support each other, and share learning from their respective projects.
Northumberland County Council has taken on the role of project sponsor and manager, bringing together the various partner organisations to progress the work. Core partner organisations include:
- Northumbria University – developing the AI technology
- Arup – working with local communities to co-create the service and providing support with data collection
- Isle Utilities – responsible for information governance, outreach, and securing the project's longevity beyond the 2027 funding round
These organisations maintain regular contact to facilitate their collaboration, manage data sharing agreements, and discuss their interconnected roles and responsibilities to achieve the best possible outcomes for the project.
Engagement
Engaging with local communities has been a central focus of this project, ensuring residents are involved and their perspectives are integrated throughout. Communities lacking formal flood warning systems have often developed their own methods of protecting lives and property. Volunteer flood warden groups play a vital role in bolstering community flood resilience by providing warnings to residents. These volunteers diligently monitor water levels, disseminate warnings, and support residents and their communities during flood events. Their dedicated efforts help minimise damage, protect lives, and cultivate a sense of collective responsibility.
Therefore, it is crucial that newly developed technologies and council interventions support this existing work and contribute to enhanced community preparedness for flash flood events. With the project's aim to increase community resilience, it's essential that the technology is fully embraced by the communities, rather than being perceived as a separate entity. To capitalise on this valuable, locally-grounded knowledge, Arup is co-creating the flood warning messages with local stakeholders. They possess detailed insights that must be incorporated into any new flood prediction system,
Lessons learnt and next steps
Embarking on the FloodAI pilot marked an exciting step into AI for the council. While there were initial concerns regarding the project's name and potential resident perceptions. However, communities have embraced the use of AI with greater enthusiasm than anticipated. By maintaining transparency about how AI could positively benefit residents, the project has gathered strong support from local groups.
Looking ahead, a key priority for the council is ensuring the long-term success of FloodAI. With current funding secured through Defra until 2027, the council is proactively exploring various adoption pathways and sustainable commercial models to guarantee the project's continuation, resulting in a positive impact on the community.