Towards Earth system modelling: coupled ocean forecasting

Forecasting across different Earth system components has initially been achieved independently, but increasing computer power, increasing model accuracy, increasing connectivity between experts, and increasing need for multi-hazard weather warning is changing the scene. Coupling methods, which involve exchanging information between discrete modelling systems, enable us to gain accuracy and consistency across Earth system components. This paper explains the principles of two-way coupling, where models run simultaneously and exchange information both ways. As individual models reach better accuracy, coupling becomes a key factor to improve forecasting capability because it reproduces the natural complexity of the environment: a wealth of literature shows the benefits of coupling. However, coupling is still limited in operational oceanography by its large demands on computational resources, by data assimilation techniques (currently not very well harmonised between the different models), and by administrative separation of forecasts across different Earth system components. Overcoming these barriers will support ocean predictions towards a multi-hazard approach and a more accurate representation of the Earth system component interactions and improve collaborations between multi-disciplinary forecasting communities.