ENVIRONMENT AND ENERGY
Detection and Attribution of Anthropogenic Drivers in Extreme Events
Principal Investigator:
Dr. Petra Friederichs
Affiliation:
Collaborative Research Center 1502 DETECT, University of Bonn – Institute of Geosciences, Bonn
Local Project ID:
detectdaee
HPC Platform used:
JUWWELS CPU of JSC
Date published:
Introduction
In July 2021, a devastating flood hit Central and Western Europe, causing severe damage, especially in the Ahr region in Germany. Researchers at the University of Bonn investigated the role of soil moisture in intensifying this extreme event. Using the JUWELS supercomputer at Forschungszentrum Jülich, they simulated varying soil moisture conditions to assess its impact on precipitation. The findings suggest that land surfaces contributed significantly to the heavy rainfall, with potential for even more precipitation under wetter soil moisture conditions. These insights can help to improve understanding of land-atmosphere interactions and disentangle drivers of extreme events.
Report
The July 2021 flood in Central and Western Europe caused widespread damage, with the Ahr region being one of the hardest hit. While much of the research around this event has focused on hydrological factors and the influence of climate change, our project aimed to explore the less understood land-atmosphere interactions – specifically, the role of soil moisture in the heavy precipitation that contributed to the flood.
The question for this study was whether and how soil moisture influenced the precipitation during the flood. At the time of the event, the slow-moving low-pressure system "Bernd" settled over Central Europe, bringing with it very warm and moist air masses, which led to persistent heavy rainfall. This large scale driver was locally amplified by the terrain features around the Ahr valley, meaning the surrounding hills forced the moist air upwards, intensifying the rain. It was reported that the soils were already saturated due to earlier rain events, causing fast run-off of the rain. This combination overwhelmed smaller and larger rivers alike.
In this context, they wanted to determine if soil moisture was a significant factor in fueling the rainfall. Earlier studies suggested differing sources of moisture – some indicated that the North and Baltic Seas played a dominant role, while others proposed that the surrounding land was a significant contributor.
Numerical weather prediction models are extremely complex, requiring significant computational resources. These models simulate a large variety of interactions between land, ocean, and atmospheric processes. This is why national weather centers typically use their own high-performance computing resources for their operational models. To be able to use state of the art models at high resolutions, the JUWELS supercomputer at Forschungszentrum Jülich was used. Its CPU resources were leveraged to run multiple simulations with different soil moisture conditions and with this the impact on the precipitation patterns leading up to the flood could be analyzed.
Figure 1: Time series of precipitation summed over our focus area. The un-altered simulation is shown in blue, while the modified simulations feature drier conditions (green) and wetter soil conditions (orange).
By running the ICON model, which is the current model of the German weather service DWD, with altered initial soil moisture conditions, a clear positive relationship between soil moisture and precipitation intensity was found. As can be seen in figure 1, it takes a while for the modifications to take effect. At the start of the event around the 13th of July, the scenarios start to diverge. In extremely dry conditions (green), the rainfall amounts were significantly reduced, while with higher-than-normal soil moisture (orange), there was potential for even more extreme precipitation than what was predicted at the time (blue).
However, further simulations are required to fully assess the relative contributions of land moisture versus ocean moisture sources, and future research could include moisture tracking to better understand the origins of the moisture, as well as comparisons with climate change influences. Nevertheless, these findings suggest that soil moisture played a substantial role in fueling the flood event and that wetter-than-usual soils likely exacerbated the heavy rainfall and not only the severity of the flood.
The insights from this research contribute to the broader goals of the collaborative research center DETECT, which aims to better understand the interactions between land use and weather patterns. By focusing on how land changes influence extreme weather events, DETECT seeks to improve predictive capabilities for future climate-related disasters.
The findings will benefit a wide range of stakeholders, including environmental scientists, weather forecasters, and policymakers. Specifically, our results provide valuable input for flood risk assessments, as they highlight the importance of land surface conditions in amplifying extreme precipitation events. In the future, this research could help refine disaster preparedness and inform land management practices that minimize the risk of severe flooding.