Investigating drivers and dynamics of hot and dry extremes in Europe by applying machine learning approaches

Felsche, Elizaveta (2024): Investigating drivers and dynamics of hot and dry extremes in Europe by applying machine learning approaches. Dissertation, LMU München: Fakultät für Geowissenschaften

Abstract

Heatwaves and droughts are increasingly impacting our ecosystem, infrastructure and society through their rising intensity and frequency due to the present effect of global warming. A series of recent extreme heatwaves and droughts in Europe, like in 2003, 2006, 2018 and 2022, have revealed the vulnerability by causing various impacts from increased heat stress on humans and the ecosystem to missing cooling water for thermal power plants and increased incidence of climate- sensitive diseases. The frequency and intensity of heatwaves and droughts are expected to increase in most regions of Europe in the upcoming years. This thesis studies the predictability and interrelationship of heatwaves and droughts and their evolution with climate change by applying different machine learning techniques. With the help of artificial intelligence methods such as ANN and hierarchical agglomerative clustering, the study advances the scientific knowledge on the predictability of droughts. The thesis shows that air pressure and teleconnection indices are essential predictors for droughts in Munich and Lisbon, as well as that for Northern Europe number of heatwave days in summer, can serve as a predictor for an agricultural drought in fall. For summer heatwaves in South Europe, the soil moisture deficit in the previous spring can serve as a predictor. The thesis shows that compound hot and dry events will become more probable with rising Global Warming Levels in most European regions. At the same time, there is a difference of up to five to six times in event occurrence when comparing GWL +2K to GWL +3K, underlining the benefits of sticking to a two-degree target. This scientific knowledge is valuable for further studies on the predictability of hot and dry events. The evolution in the future is valuable for decision-makers for implementation of mitigation measures to address the impacts of these events on human health, agriculture, and the environment.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Keywords: heatwave, drought, machine learning, predictability, extreme events
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 550 Geowissenschaften
• Fakultäten: Fakultät für Geowissenschaften
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 28. Mai 2024
• 1. Bericht­erstatter:in: Ludwig, Ralf