Sustainability

Environment Agency using AI to preserve data vital to understanding flooding risk

Updated:No Comments2 Mins Read
Flooding at the River Ouse in the historic centre of York, UK. Photo credit: Getty Images Signature
Flooding at the River Ouse in the historic centre of York, UK. Photo credit: Getty Images Signature

The government body charged with protecting and enhancing the environment, including mitigating the growing risk of flooding, has come up with a solution to preserving thousands of perishable data assets vital to help model flood risk through the use of artificial intelligence techniques.

The Department for Environment, Food & Rural Affairs approached the Accelerated Capability Environment (ACE) on behalf of the Environment Agency to explore the feasibility of using artificial intelligence and machine learning technology to digitise, read and interpret the physical data significantly faster while maintaining accuracy. 

The Environment Agency has around 10,000 years’ worth of hydrological data on river levels and flows, but it is stored on materials that are fast degrading and it says that it could take 40 years to digitise it using manual processes. This data could be used to build more accurate climate and flood modelling and help forecast and minimise the impact of future adverse weather events. 

“The importance of collating and analysing good quality historic data and records to better understand climatic trends and management of river catchments cannot be underestimated as it is key to protecting the UK’s security in the face of emergencies.”

EA Case Study from the Accelerated Capability Environment.

The work to build and test a proof of concept ‘data-rescue tool’ was carried out by the London Data Company – who also worked on a machine learning project for HMICFRS last year. The Environment Agency project looked at how to deal with handwritten information and how well it can be preserved using fully automated techniques versus those where the technology was supported by some human intervention.

The proof of concept concluded that the ‘human-in-the-loop tool for AI-assisted data rescue’ produced better results. The project also recommended that with some adaptation, the approach could increase the effectiveness of this tool on live datasets as well as historical data.

Read the full case study from the Accelerated Capability Environment.

Find out about the HMICFRS data project and read our article, Using machine learning to flag issues with police forces sooner.

You can learn more about what happened next as the project team will be speaking at The Emergency Tech Show on 18 September at 1000. Registration is open now.

Share.