Downscaling climate models with physics-informed machine learning to tackle the energy transition
The escalating climate crisis poses a growing threat to human society and infrastructure, making climate mitigation an urgent global priority. The European Green Deal aims for climate neutrality by 2050, while Austria has set an even more ambitious target of 2040. A key driver of these efforts is the transition from fossil fuels to renewable energy, with significant research dedicated to assessing Austria’s potential for solar, wind, and hydropower.
Previous studies have evaluated renewable energy availability based on observed and projected climate conditions, considering factors such as solar radiation, wind patterns, and available land. Additionally, the impact of climate change on energy production has been examined, highlighting risks such as heatwaves with high energy demand but low wind speeds and reduced solar panel efficiency. However, the influence of extreme weather events on energy infrastructure remains largely unexplored.
EnergAIze addresses this gap by integrating physics-informed AI with high-resolution climate modeling to assess renewable energy availability under different climate scenarios (SSP2.6, SSP4.5, SSP8.5). The project reduces uncertainty related to extreme events—flooding, landslides, hail, storms, and extreme heat—by improving input data quality and advancing climate model verification through AI-driven approaches.
Facts
- Project start: 01.10.2023
- Project duration (in months): 18
- Budget: 317 kEUR
- Funding: 200 kEUR
- Project partners: GeoSphere Austria
