S2S2D Conference 2026: Advancing climate predictions from weeks to decades to benefit society

S2S2D Conference 2026: Advancing climate predictions from weeks to decades to benefit society

PML’s Deep S. Banerjee will be presenting a talk at the conference, titled: “Toward Scalable Hybrid Earth System Prediction: Online Machine-Learning Corrections in a Shelf-Sea Testbed”, at 4pm on the 9th of September.

Event description:

‘The science and practice of subseasonal to decadal prediction are at a crossroads. In recent years subseasonal and decadal prediction have matured from exploratory research areas into multi-system operational services, while multi-annual and multi-decadal timescales have emerged as further temporal realms for initialized prediction. However, inherent sources of predictability remain to be fully understood and exploited as predictions aim to address a widening array of Earth system components and to support early warnings of extreme events. Meanwhile, artificial intelligence/machine learning are propelling rapid and revolutionary developments, the outcomes of which are not yet clear.

This event will provide a forum for the current state of Earth system prediction science to be shared and discussed. It will consist of two parallel sub-conferences, respectively focusing on the subseasonal-to-seasonal and seasonal-to-decadal prediction horizons, together with plenary sessions addressing cross-cutting themes.’

Find out more about the event on the Earth System Modelling and Observation (ESMO) website here >>

The event is supported by World Climate Research Programme, and has been endorsed by the World Weather Research Programme (WWRP) of the World Meteorological Organization (WMO). The conference is sponsored by ECMWF, Copernicus Climate Change Service, Climateurope2, Spark Climate and the University of Reading.