The prediction of changes in the climate mean state, variability and extremes remains a key challenge on decadal to centennial timescales. Recent advances in climate modelling, downscaling, artificial intelligence (AI) and post-processing techniques and ensemble techniques provide the basis for generating climate information on local to regional and global scales. To make such information actionable for users, relevant information needs to be derived and provided in a way that can support decision-making processes. This requires a close dialogue between the producers and wide-ranging users of such a climate service.

National climate change assessments and scenarios have become an essential requirement for decision-making at international, national and sub-national levels. Over recent years, many European countries have set up quasi-operational climate services informing on the current and future state of the climate in the respective country on a regular basis (e.g. KNMI'23 in the Netherlands, UKCP in the UK, CH2018 and CH2025 in Switzerland, ÖKS15 and ÖKS26 in Austria, National and federal states Climate Reports in Germany). However, the underpinning science to generate actionable climate information in a user-tailored approach differs from country to country. This session aims at an international exchange on these challenges focusing on:

- Practical challenges and best practices in developing national, regional and global climate projections and predictions to support adaptation action and impact assessments.

- Developments in dynamical and statistical downscaling techniques, process-based model evaluations, AI techniques and quality assessments.

- Methods to quantify uncertainties from climate model ensembles, combination of climate predictions and projections to provide seamless user information.

- Examples of tailoring information for climate impacts and risk assessments to support decision-making and demonstration on evaluation steps taken to monitor the uptake of climate information.