Long-term foehn reconstruction combining unsupervised and supervised learning

Reto Stauffer,Achim Zeileis,Georg J. Mayr
2024-06-04
Abstract:Foehn winds, characterized by abrupt temperature increases and wind speed changes, significantly impact regions on the leeward side of mountain ranges, e.g., by spreading wildfires. Understanding how foehn occurrences change under climate change is crucial. Unfortunately, foehn cannot be measured directly but has to be inferred from meteorological measurements employing suitable classification schemes. Hence, this approach is typically limited to specific periods for which the necessary data are available. We present a novel approach for reconstructing historical foehn occurrences using a combination of unsupervised and supervised probabilistic statistical learning methods. We utilize in-situ measurements (available for recent decades) to train an unsupervised learner (finite mixture model) for automatic foehn classification. These labeled data are then linked to reanalysis data (covering longer periods) using a supervised learner (lasso or boosting). This allows to reconstruct past foehn probabilities based solely on reanalysis data. Applying this method to ERA5 reanalysis data for six stations across Switzerland and Austria achieves accurate hourly reconstructions of north and south foehn occurrence, respectively, dating back to 1940. This paves the way for investigating how seasonal foehn patterns have evolved over the past 83 years, providing valuable insights into climate change impacts on these critical wind events.
Applications,Atmospheric and Oceanic Physics,Machine Learning
What problem does this paper attempt to address?
This paper focuses on the reconstruction of long-term Foehn events. Foehn is a wind that occurs on the leeward side of a mountain range, usually accompanied by a rapid increase in temperature and changes in wind speed, which significantly affect the local climate and environment. Due to the difficulty in directly measuring Foehn, researchers often rely on meteorological station data and classification methods to infer Foehn events. The paper proposes a novel approach that combines unsupervised and supervised learning statistical probability models to reconstruct historical Foehn events. Firstly, a unsupervised learning model (finite mixture model) is trained using recent decades of field observation data (automatic weather station data) to automatically classify Foehn. Then, these labeled data are combined with long-term reanalysis data (such as ERA5) and a supervised learning model (such as Lasso or Boosting) is used to reconstruct the occurrence probability of past Foehn events based solely on the reanalysis data. The application of this method at six sites in Switzerland and Austria demonstrates the ability to accurately trace the hourly occurrence probability of Foehn back to 1940, thus studying the variations in seasonal Foehn patterns over the past 83 years and providing valuable insights into the impact of climate change on this crucial wind event. The key words used in the research include: Foehn, reconstruction, unsupervised learning, mixture model, supervised learning, climate, and trend.