There has been a great deal happening across ECMWF this summer. From open-source software and artificial intelligence to the production of ERA6, we are seeing important progress across many parts of our work. These developments span different areas of science and technology, but they all demonstrate the value of working together across ECMWF, our Member and Co-operating States, and the wider community.
The release of earthkit 1.0 is a particularly important milestone. This new ecosystem of open-source Python packages provides a stable foundation for working with weather and climate data, bringing together tools that can be used individually or combined into complete scientific workflows. It is already supporting ECMWF applications, and its continued development with Member States and external partners will help make our software easier to use, reuse and build on.
Elsewhere, our work on a data-driven Earth-system model under Destination Earth is opening new possibilities for Earth-system prediction. We are developing machine-learning components for the ocean, waves, sea ice, land and hydrology. While these developments remain at an early stage, they point towards a future in which AI complements our physical modelling expertise, helping us represent Earth-system processes more effectively and explore new approaches to prediction. Early results, including the ability of prototype models to represent extreme events absent from their training data, are particularly encouraging.
At the same time, we continue to improve the systems we already rely on. New approaches to land-surface parameter tuning are showing how data assimilation can help address forecast biases, while advances in ocean data assimilation are strengthening the foundations for forecasting and climate applications. The start of ERA6 production is another major achievement, bringing together almost a decade of scientific and technical development to create the next generation of global reanalysis.
The exceptional European heatwaves of this summer also provide an opportunity to reflect on how developments translate into forecasting. Longer-range forecasts captured the broad signal of an unusually warm summer, while predicting individual heatwave events remains challenging. Results from the AIFS show promising performance for some extreme temperatures, highlighting the value of continuing to develop both AI-based and physics-based approaches. Improving our understanding and prediction of extremes will remain one of the most important challenges for our community in the years ahead.
Just as important is the dialogue with those who use our services. The latest cycle of liaison visits with our Member and Co-operating States has given us valuable insight into changing needs and priorities, while UEF2026 brought users together to explore the challenges of forecasting extreme temperatures. As the scientific and technological landscape evolves, maintaining this dialogue is essential.
I am also very pleased to welcome Dr Irina Sandu and Dr Jørn Kristiansen as ECMWF's new Deputy Director-Generals, following their election by Council. Their combined experience across science, forecasting, operations and infrastructure will be a great asset to ECMWF and to the community we serve.
As we look ahead, the pace of change across weather and Earth-system prediction continues to accelerate. By combining scientific excellence, operational reliability, emerging technologies and close collaboration with our Member and Co-operating States, we are well placed to seize these opportunities and deliver even greater value.
Florian Pappenberger
Director-General