The 2015/16 El Niño turned out to be in the same class as the exceptional events in 1982/83 and 1997/98. How well was it predicted? ECMWF’s forecasts of the event as well as the EUROSIP multi-model forecasts produced at ECMWF provide an impressive demonstration of the power of seasonal forecasting. Months in advance they successfully predicted the onset, peak and decline of the 2015/16 El Niño. These forecasts, described in detail in this Newsletter, illustrate the world-class capabilities of ECMWF’s seasonal forecasting system.

But of course there is always room for further improvements. SEAS5, ECMWF’s new seasonal forecasting system, will become operational later this year. It promises to bring even higher-quality seasonal forecasts in the tropics. It will also provide seasonal predictions of sea ice. Another major advance is that it is well aligned with ECMWF’s extended-range system. This marks a significant step on the road towards seamless forecasting across a wide range of timescales, from the medium range out to seasonal forecasts.

A second transition that is under way is the move of multi-model seasonal forecasting at ECMWF from a research environment into operations. The EUROSIP multi-model forecasts combine output from five major forecasting centres: ECMWF, the UK Met Office, Météo-France, the US National Centers for Environmental Prediction (NCEP) and, since March this year, the Japan Meteorological Agency (JMA). Plans are now in place for all these forecasts to be included in the new seasonal forecasting service that is being trialled by the EU-funded Copernicus Climate Change Service (C3S) operated by ECMWF.

The C3S prototype service already brings together ECMWF, the UK Met Office and Météo-France as core providers and will soon include Italy’s Euro-Mediterranean Center on Climate Change (CMCC) and Germany’s national meteorological service, DWD, as additional providers. These are now set to be joined by NCEP and JMA: a shining example of international cooperation for the benefit of users in Europe and beyond. C3S will have dedicated resources to enable the successful operation of this new seasonal service. Continuous product development will happen at a faster pace than was possible in the EUROSIP framework. At the same time, all providers including ECMWF will keep full control of their own seasonal forecasting systems.

Continuously driving improvements in seamless forecasting from the medium range all the way to seasonal forecasts requires constant improvements in ECMWF’s computing resources. As described in this Newsletter, 2016 saw a significant upgrade of ECMWF’s high-performance computing facility. This upgrade will take ECMWF to 2020, at which point a further significant expansion of computational capabilities will be needed, together with a new data centre. We have every confidence that 2017 will be the year in which we lay the foundations for a seamless transition to a new data centre and more powerful computing resources that will take us into the next decade.

Florence Rabier