We claim that the transition towards a highly renewable and strongly interconnected pan‑European energy system requires weather and climate datasets, and derived variable renewable energy (VRE) generation time series, that go beyond today’s standards. While recent pan‑European datasets have significantly improved the availability and quality of wind and solar generation time series, further methodological developments are needed to support robust system expansion and adequacy assessments towards 2050 and beyond.

This paper reviews the state of the art and discusses key future development needs for continental‑scale weather and climate data and VRE generation time series in increasingly weather‑dependent energy systems. The conclusions are derived from multiple Nordic and European research projects, close collaboration with industry and system operators and a synthesis of the scientific literature.

First, although climate projections are increasingly included in large‑scale datasets for energy system analysis, such as the Pan‑European Climate Database (PECD) used by ENTSO‑E, the underlying global climate models typically have coarse spatial resolution and validated high‑resolution downscaled data remain limited. Temporal resolution is also often insufficient for studies requiring at least hourly modelling. We discuss recent advances in statistical, AI‑ and machine‑learning‑based downscaling methods capable of handling large collections of climate models, scenarios and multi‑decadal time horizons, while highlighting remaining challenges related to validation for energy system applications.

Second, the representation of large‑scale wind farm wakes is a critical challenge as offshore wind deployment accelerates and multi‑GW clusters are developed in close proximity. While intra‑farm wakes are commonly considered, farm‑to‑farm wake effects are typically neglected in large‑scale system studies, leading to overestimated energy yields and misrepresentation of offshore grid and hub investments. We discuss the need for computationally efficient yet physically credible wake modelling approaches, including surrogate models calibrated against mesoscale simulations.

Third, explicit consideration of wind and solar forecast errors in long‑term system expansion and adequacy studies is increasingly important. Current approaches often assume perfect foresight or rely on stylised reserve requirements, neglecting the weather‑dependent structure and spatial correlation of forecast errors. Developing large‑scale, long‑term datasets of VRE forecasts and forecast errors would enable more realistic assessment of balancing needs and system resilience.

Finally, recognising that future studies may involve multi‑decadal or even multi‑century weather datasets, we discuss approaches for compressing weather‑driven information into reduced datasets suitable for optimisation under computational constraints. Focusing on VRE, we highlight differences between project‑level and system‑level treatments of long‑term weather variability and discuss metrics for identifying critical conditions for adequacy and cost‑efficient system reinforcement.