European Commission logo
English English
CORDIS - EU research results

Towards Operational Supermodel Climate Prediction

Project description

A new supermodel for climate prediction

Climate prediction services are essential to keep societies prepared and in the ready. Their services are even more crucial for preventing possible disasters. Unfortunately, current systems are hindered by systematic errors in the models, which substantially limits their ability to efficiently predict climate changes and makes the process very costly. The ERC-funded TOSCP project will provide a solution. It will use the supermodel developed by the ERC-funded STERCP project for more efficient climate prediction for reduced costs. It will use novel data and assimilation systems to optimise the service.


TOSCP will proof-of-concept a radically new approach to climate prediction based on supermodelling. Climate prediction promises reliable information on climate and its extremes for the coming seasons and years. This information is critical to providing climate services that are needed to build a resilient and sustainable society. Unfortunately, predicting climate in the extra-tropics remains a major challenge. Model systematic error is the major limitation. In the North Atlantic-European sector it leads to the strong under representation of the predictable dynamics, compared to unpredictable atmospheric weather patterns. The current approach to account for such errors is to perform a vast number of independent simulations with different models. This is computationally expensive and impractical in an operational context. The supermodel approach developed in the ERC-STERCP project is aptly suited to improve climate prediction. A supermodel combines a set of different models in runtime so that the individual model errors compensate so as to produce a superior model. The approach is extremely effective in mitigating long-standing model errors, and can control the ratio between predictable and unpredictable dynamics. TOSCP will reconfigure a supermodel developed in the STERCP project for climate prediction. The supermodel is based on three state-of-the-art climate models. We will develop new ensemble generation and data assimilation schemes. We aim to demonstrate that supermodel climate predictions greatly outperform the standard approach to climate prediction that is currently used for climate services. Dialogue with users and providers of climate
services will ensure the development of an optimal configuration of the new prediction system and its use in operational climate services. This will set the stage for the wider exploitation of supermodel climate prediction, leading to improved climate services for the benefit of society.


Host institution

Net EU contribution
€ 150 000,00
5020 Bergen

See on map

Norge Vestlandet Vestland
Activity type
Higher or Secondary Education Establishments
Total cost
No data

Beneficiaries (1)