Advanced forecasting for Canal & River Trust assets
Since 2020, we have provided bespoke forecasting services to the Canal & River Trust (CRT) to support management of their assets. We designed, built and maintained a system modelling levels at Toddbrook Reservoir, Derbyshire, initially in an emergency response to partial dam collapse in 2019. Since then, we have added a further five reservoirs to the system to provide forecasted rainfall, flows and reservoir levels to help CRT safely operate their assets.
- Client Canal & River Trust
- Location UK
- Services
Integrating high-resolution rainfall data and customised modelling
Our forecasting system integrates the Met Office's highest-resolution rainfall forecasts, utilising RADAR data and advanced numerical weather predictions. We employ the same trusted, industry-standard models - such as the Probability Distributed Moisture (PDM) model and Flood Modeller - used by the Environment Agency, housed within Delft-FEWS. Additionally, we can design and customise models to simulate specific features, including storage reservoirs and other assets that require real-time management.
Addressing forecasting challenges
During the initial emergency response at Toddbrook, the primary challenge was the limited time available to act. Rapidly providing supporting information was critical to understanding reservoir levels and guiding emergency works.
Ongoing challenges include ensuring the accuracy and timeliness of forecasts to support CRT’s operations. These challenges have been addressed through regular calibration and periodic review of the models, ensuring their reliability and effectiveness.
Collaborative innovation: developing bespoke flood forecasting solutions
Collaboration was key as part of this project with our flood forecasting and IT specialists working together to design and calibrate a bespoke system. Externally, we partnered with CRT to develop a tailored solution and provide guidance on its application.
Key developments included enhancements to our Ensemble Flood Forecasting (JEFF) system, which leverages Delft-FEWS - an industry-standard forecasting platform. Additionally, spatial sampling of forecast rainfall using a ‘cookie-cutter’ approach was implemented to account for rainfall uncertainty in predictions.
The project has delivered an intuitive, user-friendly tool for CRT, offering advanced warning of rainfall events and their potential impact on CRT assets and operations.
Our customisable alerting service notifies users via SMS and email whenever pre-defined thresholds are exceeded.
Forecasts are readily accessible through a simple, mobile-friendly web application, ensuring easy access.
Contact Paul Wass or Rebecca Stroud for more information on this project.
