Linus Magnusson, ECMWF, Adjunct Professor at MISU
Linus Magnusson from ECMWF, the European Centre for Medium-Range Weather Forecasts, is appointed Adjunct Professor at MISU on 10 percent, from 1 September. He will e.g. supervise Master degree projects, do Arctic research related to forecast evaluation and model/observation improvements, support planning and forecasting during Arctic campaigns, and some preparation of course material.
Linus Magnusson began his career in the Swedish Armed Forces' weather service but started his PhD at MISU in 2005 on ensemble forecasts. During his PhD studies, Linus worked closely with ECMWF. After completing his PhD in 2009, Linus joined ECMWF, initially working on seasonal forecasts and sea-ice modeling, but he soon returned to the atmosphere and began working on forecast evaluation.
What does forecast evaluation involve?
– My main task at ECMWF is to evaluate the quality of the forecasts and identify areas for improvement. The role requires an understanding of both how the atmosphere works and how the forecast system is designed in order to determine where the shortcomings come from. In terms of research, this means understanding the predictability of the atmosphere and how errors propagate in the forecast. It also involves identifying the limitations of the forecast models in terms of what can be simulated and what is not possible.
– One area of research I have worked extensively with is understanding the predictability of extreme weather. This is, of course, of great importance to society, and it is essential to build trust in forecasts in critical situations. If warnings are issued too often, the public will not take the forecasts seriously, but at the same time, it is important not to miss issuing warnings for weather that could have serious consequences. To find a good balance, it is important to have a good understanding of the quality of the forecasts. Evaluating extremes has also led me to work on how well the models forecast tropical cyclones.
Forecasts based on AI
– In recent years, we have experienced a revolution in AI-based forecasting. But these new forecasting systems also need to be evaluated and knowledge built up about their strengths and weaknesses in different weather situations. Since these forecasts are not directly based on physical principles, it is important to be extra careful to compare the results with our physical understanding of the atmosphere.
Forecast training and weather discussions
– At ECMWF, I am also involved in teaching some of the courses we offer. One course I am involved in is about interpreting our forecasts and understanding different forecast products. Today's forecast models produce such a large amount of data that it is easy for meteorologists to get lost in all the information. That is why we need good forecast products that condense the material into the most essential information. I also teach about predictability on longer time scales, such as monthly and seasonal forecasts. I also hold regular weather discussions for the entire staff at ECMWF, where we discuss forecast quality, but which I also see as an opportunity to improve the level of knowledge for myself and my colleagues.
– In addition to my work at ECMWF, I am also involved in the WMO (World Meteorological Organization) as a member of the Scientific Steering Committee for the World Weather Research Programme (WWRP). In that role, I am active in polar research through the PCAPS project (Polar Coupled Analysis and Prediction for Services) and also in a working group on forecast verification.
How will you contribute to MISU?
– I hope to bring my perspective from an operational centre to MISU, including the challenges we face and the knowledge students need to be able to work in such an environment and use our products. An important product from ECMWF is the global reanalysis ERA5, and hopefully I can help with advice on how to best use that dataset. In terms of teaching, the plan is for me to support courses related to weather and forecast models.
Last updated: October 28, 2025
Source: MISU