Periodic Reporting for period 1 - TOSCP (Towards Operational Supermodel Climate Prediction)
Período documentado: 2023-01-01 hasta 2024-06-30
Supermodelling is a novel technique for combining existing models to reduce errors and produce a superior model. The technique was developed in the previous EU FP7 SUMO and ERC CoG STERCP grants. Supermodelling was shown to be able to dramatically reduce long-standing errors in climate models. The objective of the ERC PoC TOSCP project was thus to prove that these developments can be used to improve seasonal predictions. A secondary objective was to facilitate the translation of the supermodel approach to operational climate prediction.
We have performed the first ever climate predictions with a supermodel. Our supermodel consisted of three state-of-the-art Earth System Models, connected by their ocean components. Promising results were obtained that indicate that the supermodel can mitigate the spring predictability barrier—a major challenge in seasonal prediction. However, skill degradation was identified and related to the configuration of the supermodel. To overcome these issues, we have developed a capability to efficiently connect atmospheric components in a supermodel. Further work is required to apply this new configuration in climate prediction.
To facilitate the translation of supermodelling to climate prediction we actively engaged academic and non-academic users. We show cased the ability of supermodels to improve climate predictions and services, through presentations at workshop and meetings and by developing innovative outreach material—such as Lego and video games. We also clarified IPR and knowledge transfer strategies, but our supermodelling technology is still at an early stage.
To facilitate the translation of supermodelling to climate prediction we actively engaged academic and non-academic users. We show cased the ability of supermodels to improve climate predictions and services, through presentations at workshop and meetings and by developing innovative outreach material—such as Lego and video games. We also clarified IPR and knowledge transfer strategies, but our supermodelling technology is still at an early stage.
We have taken supermodelling one important step closer to operational climate prediction, by performing the first ever set of seasonal predictions and by engaging the academic and non-academic users of such predictions. The skill improvements from supermodelling can enhance climate services, through providing more accurate information for decision making, for example, in energy, agriculture, and health sectors.
We have also identified and developed strategies to mitigate deficiencies in our prototype predictions. These are being further developed in the EU Impetus4Change project.
We have also identified and developed strategies to mitigate deficiencies in our prototype predictions. These are being further developed in the EU Impetus4Change project.