Weather forecasts are of great value to society, especially for high-impact events. We consider forecasts of low-pressure systems (cyclones) which can cause strong winds and heavy rain. Today, weather forecasters have access to a vast amount of data, too much to analyse in the time available. In our previous research, we created a dataset of cyclone intensity measures and identified key metrics to concisely describe cyclone intensity. This project uses these findings to benefit society by co-designing real-time forecast products with weather forecasters and making them operational. These products will show cyclone intensity forecasts for Finland and their uncertainties, compare them to climatology, and automatically alert forecasters about unusual events. We use forecast data from traditional weather forecast models and from a new machine learning based model. The work is performed in INAR at the University of Helsinki.

2026

2027