2022 was a completion year for us at ECMWF, when we saw several years of hard work in different fields bear fruit. It was the first full year operating as a multi-site organisation, with staff and systems in place in our three locations and with the Hosting Agreement with Germany formally signed in Berlin on 9 December. It was the year that we started to thoroughly reap the benefits of the changes introduced at ECMWF in the wake of COVID: more flexible working patterns with a generous teleworking policy enabling a better work-life harmony, and more sustainable and wider-ranging events, with all workshops and seminars being either livestreamed or fully hybrid.

2022 was also marked by more diversity in our recruitment, with more diverse pools of candidates showing interest in positions we advertised, leading to a more diverse workplace.

From a scientific perspective, it was the year when Cycle 48r1 was handed over from research to implementation, allowing scientists to focus on the next cycle of the Integrated Forecasting System (IFS). Towards the end of the year, on 18 October, we ran and disseminated our first operational forecast on the new Atos computing system, illustrating what our new technology could deliver.

2022 was also the year when results achieved through the implementation of our Machine Learning Roadmap clearly showed the potential that this technology can and will certainly have on numerical weather prediction.

2022 saw the formal launch by the European Commission of the EU-funded initiative Destination Earth, which ECMWF is jointly implementing with EUMETSAT and ESA.

ECMWF’s role is to develop two digital twins – one on climate change adaptation and one on extreme weather – as well as the digital twin engine. This initiative is a great opportunity for collaboration not just with our partners at EUMETSAT and ESA, but also with our Member States. This is because the two largest tenders were awarded to a coalition led by Météo-France for the extremes twin and to one led by the CSC – IT Center for Science in Finland for the climate change twin.

From an operations perspective, scores in 2022 showed the full extent to which our previous IFS upgrade, Cycle 47r3, has impacted our predictions, with very positive results notably in our upper-air ensemble scores, as shown in the figure. Using the ERA5 reanalysis as a reference reduces the effect of interannual variability on scores and allows a more reliable quantification of performance changes due to individual model cycles.

A lot more happened in 2022, which you can see in the pages of this Annual Report. As usual, thanks go to our staff, whose dedication and talent never fail; our Member States, whose support and contribution are a constant reminder of the importance of working collaboratively; and of course all our partners, users, and delegates of our workshops and seminars – all of you from the meteorological community who are making it all not just possible but inspiring.

Florence Rabier, June 2023