On the sensitivity of heat waves to physical drivers and climate change

Abstract

Extreme events have always been a threat to societies and ecosystems but they are natural to the climate system as a result of its internal variability. A specific extreme event is caused by the unique interplay of a number of different driving factors. The atmosphere, the ocean and the land surface are major drivers of climate variability and participate in feedback processes that may amplify or reduce the risk of an extreme event. As all of these natural drivers work together to create a specific extreme event, it is difficult to disentangle the individual contributions. Furthermore, by burning fossil fuels and converting land, humans have started to affect the climate system. At present, this has led to a global warming of about 1°C above pre-industrial levels. This change in the background global climate is altering natural processes that can be relevant for extreme events. However, the impact of this climatic change is particularly difficult to detect for extreme events as they are rare. Nonetheless, given the devastating impacts of extreme events, it is essential to study them to provide a basis for mitigation and adaptation planning. Understanding the underlying processes and involved uncertainties helps to provide reliable short-term forecasts for extreme events and projections for the future. This thesis focuses on heat waves and, in some cases, concurrent droughts. A main goal is to disentangle the role of the drivers involved. Global climate model simulations are conducted to separate the contributions from thermodynamic and dynamic processes by forcing parts of the climate system toward observations. Atmospheric circulation constitutes the dynamic part of the climate system and can be controlled by constraining its variability using atmospheric nudging. The thermodynamics can be controlled by constraining for example the land surface conditions, which is done here using soil moisture prescription. In the first part of this thesis, an evaluation of climatological model biases is provided. Global climate models often show consistent biases in the surface climate, for example a hot and dry bias in the Northern Hemisphere midlatitudes during summer, which affects the representation of hot extremes. To identify the origin of such biases, atmospheric nudging is used to control the large-scale atmospheric circulation in a global climate model. The results show that biases related to the atmospheric circulation are often of minor relevance for many regions of the world. This highlights that a large part of the biases is not related to circulation, but originates from incomplete or erroneous representation of the thermodynamics. The focus shifts to heat waves in the second part of the thesis. The driving processes that contributed to five heat waves occurring between 2010-2016 are studied individually. By nudging the atmosphere toward observed circulation and prescribing soil moisture, the role of three key physical drivers is quantified: (1) atmospheric circulation, (2) land surface conditions, and (3) sea surface temperatures. In addition, the role of recent climate change since the mid-1990s is evaluated. The contribution of the natural and anthropogenic drivers to daily maximum temperatures can be assessed quantitatively using conditional event attribution. The results show that the ocean conditions play a minor role for most of the investigated events, while the atmospheric circulation and land surface can each contribute up to 70% of the events' anomalies. Recent climate change amplified all of the investigated events. It played the largest role for the 2015 European heat wave, contributing about 40% to the temperature anomaly. The third part of this thesis combines climate change scenario information with atmospheric nudging to create storylines for a recent heat wave. Storylines are a tool to create analogues of an event for alternative levels of global background warming. Here these analogues are created by nudging the large-scale atmospheric circulation toward the same observed circulation for each storyline. The method is applied to the 2018 Northern Hemisphere heat wave, which affected a large fraction of populated and agriculturally used land. In an alternative world without global warming, the impact of this event would have been strongly reduced. Looking into a possible future with a higher level of global warming the event puts an increasing fraction of land and population at risk: the area experiencing extreme temperatures increases from 9% in 2018 to 13% (34%) at 2°C (4°C) global warming. This thesis contributes to the process understanding of heat waves in present-day and possible alternative future climates. The results demonstrate that thermodynamic processes related to the land surface can be as important as dynamic processes for driving heat waves. This highlights the value of assessing the separate contributions of these processes and points to the need for a joint effort in improving their representation in global climate models. The results further reveal an influence of recent climate change on aggravating heat waves and provide an outlook to future events. The presented scientific methods and concepts were proven useful in this thesis and will help to shed light on the role of thermodynamics and atmospheric dynamics in future studies on heat waves and other extreme events.