ML Demonstrator –Enhancing Data Usability with AI-based Multi-Model Data Fusion for DestinE

This Machine Learning Demonstrator implements an AI-based system (SynCast) that uses forecasts from multiple numerical weather prediction models into a single, optimised product. By leveraging the DestinE Digital Twin infrastructure and EuroHPC systems, SynCast provides high-resolution, near-real-time forecasts for critical near-surface parameters.

In a nutshell

This Machine Learning Demonstrator implements an AI-based system (SynCast) that uses forecasts from multiple numerical weather prediction models into a single, optimised product. By leveraging the DestinE Digital Twin infrastructure and EuroHPC systems, SynCast provides high-resolution, near-real-time forecasts for critical near-surface parameters. This approach improves forecast accuracy and reliability, helping decision-makers in sectors like renewable energy, agriculture, and disaster management respond more effectively to extreme weather.

Technical Overview

AgricultureData ManagementDisaster assessmentExtreme WeatherML DemonstratorWeather
Digital Twins
LocalNationalRegional
AgricultureCivil protection agenciesEnergy Operators
Civil protectionClimate adaptationDisaster responseWeather impact

Challenge

Even the most advanced numerical weather prediction (NWP) models face limitations in representing small-scale and near-surface processes. These gaps reduce the precision of local forecasts, which are essential for managing risks from extreme weather and for supporting sectors such as disaster management, renewable energy, and agriculture.

DestinE Solution

DestinE addresses this challenge by applying AI-based methods to optimise and combine outputs from multiple NWP models. Deployed on EuroHPC supercomputers SynCast utilises DestinE Digital Twin data, and delivers improved, high-resolution forecasts in near real time. The results are made accessible through the Core Service Platform, ensuring interoperability with other DestinE components.

Syncast scheme. Credit: DWD.

Impact

Provide more accurate and locally relevant forecasts, strengthening the ability to anticipate and respond to extreme events. Outputs enhance decision-making in sectors such as wildfire and disaster management, energy system planning, and agriculture, and directly support the objectives of the EU Green Deal, Climate Adaptation, and Disaster Risk Reduction policies.

Contributions

Providers

Deutscher Wetterdienst