Data-driven approaches to link climate drivers to energy impacts

The transition from fossil fuels to renewable energy sources in Europe is accelerating under climate policy commitments (Kapica et al., 2024), including the Paris Agreement (2015) and European Green deal (2019). This shift increasingly relies on weather-dependent generation from wind and solar power, while climate change is simultaneously intensifying summer heatwaves and associated cooling demand (Filahi et al., 2024). As a result, electricity systems are becoming more exposed to weather-driven variability, highlighting the need for more research in regions historically characterised by winter peak loads.

Despite the growing importance of climate–energy interactions, progress in understanding these dynamics is still limited by a disconnect between climate and energy modelling communities. Climate modellers often do not provide outputs that are directly usable for energy system applications, while energy system models tend to overlook climate related uncertainty (Craig et al., 2022). This mismatch makes it difficult to properly assess system reliability and economic stress under future climate conditions.

Some recent studies have started to bridge this gap by linking large-scale weather patterns to energy system impacts. For example, targeted circulation types have been developed to better capture weather sensitivity in electricity systems (Bloomfield et al., 2020), while other studies connect weather regimes directly to metrics such as Energy Not Served (ENS), reflecting system reliability from a grid operator perspective (Biewald et al., 2025; Wuijts et al., 2023). However, these approaches still struggle to fully account for climate uncertainty and economic signals such as price variability.

This paper proposes a data-driven framework to better connect climate variability with energy system impacts, focusing on indicators such as energy shortfall, ENS and high-price events. Data-driven methods are explored to identify climate drivers behind these energy impacts, allowing a more system-relevant characterization of climate risks. These drivers can then be used as input for AI-based forecasting approaches, improving the prediction of system stress under uncertain future conditions. Bridging this gap is essential for ensuring reliable, affordable, and resilient low-carbon energy systems in a changing climate.