Success of weather models depends on how forecasters use and trust them. We held workshops with Norwegian forecasters to understand how they evaluate model quality and use it in their work. Forecasters rely on experience, peer discussions, and past events. They need tools that show model quality in specific situations and help them make decisions during shifts. Their views should guide tool development. This study highlights the value of collaboration between model developers and forecasters.

Weather forecasts have greatly improved over recent decades, yet errors remain. We studied how people judge different types of forecast errors using two short surveys. People rated errors differently depending on their consequences, with forecasts that underestimated the intensity and persistence of events judged most negatively. These findings show what forecast users value and contribute to the development of verification methods that better reflect public needs.

On 6 July 2023, a severe hailstorm hit Vitoria-Gasteiz, caused by a supercell. Active cells moved southwest to northeast, with some shifting west-east. The storm, with radar reflectivities up to 65 dBz, brought intense showers and hail up to 5 cm in diameter. It caused injuries, vehicle damage and urban flooding. This study analyzes the storm’s conditions and surface effects using different data.

COSMO-REA6 is widely used in energy applications but has known deficits. Its successor, COSMO-R6G2, is evaluated for Germany, focusing on Global Horizontal Irradiance (GHI) as well as wind speed. It shows slightly lower GHI and good wind quality in flat areas, but underestimates night-time wind in mountains. Overall, it is a useful successor, though users should consider remaining limitations.

In this paper we analyse the suitability of different reanalyses and meso-scale models for simulating wind energy in Germany. We found that all datasets overestimate energy production, with errors ranging from 5 % to 45 %. This suggests that the underlying models may not accurately reflect average wind conditions. CERRA and ERA5 performed the best, but they also require regional adjustment. Understanding the cause of these differences is crucial for improving weather and wind energy modelling.