Detection and attribution of anthropogenic drivers in extreme events
PD Dr. Petra Friederichs
University of Bonn | +49 228 73-5187 |
The collaborative research center’s integrated modeling system enables physically consistent simulations of the whole water and energy cycle. Our aim in this project is to investigate to what extent this improves the representation of extreme events. Our focus here is on meteorological droughts and heat waves. We compare the characteristics of extremes in the integrated modeling system with observations and reanalyses. Additionally, the impacts of different scenarios on simulation results will be addressed to attribute the extreme behavior to different forcings. Using the comprehensive representation of the water and energy cycle in the modeling system, we are especially interested in human influence, specifically changes in human land and water use.
Contribution to the CRC
Apart from providing a basic catalog of events, our work expands the other analyses done on the outputs of the CRC by a thorough statistical investigation of extreme events. Based on this, we can inform the project partners about the quality of the simulation with respect to extremes and changes in extremal patterns and their possible causes. This information can be used to improve the modeling efforts, e.g., check the effect of the addition of irrigation in D03 on extremes like heat waves and droughts. The high socio-economic impact of such events makes regime shifts, e.g., in intensity or frequency, of meteorological extremes important to economic modeling efforts in the CRC.
The first step in this project is to compile a catalog of extreme events. Based on this, specific events can be chosen for closer inspection. To make statistical analyses feasible, we expand the existing simulations by additional ensemble members and alternative scenarios. The latter is the key step to test the effects of different forcings. We look for changes by extracting compact descriptions of the extremes to limit the degrees of freedom of the comparison problem. If a large enough difference is visible, we can then compute the likelihood for the scenario being the cause of the changes. One example are the TerrSysMP simulations provided to us, which are run with and without human water use. In that case we can look for a change in extreme behavior, which could be explained by the water use, but might require an ensemble if one wants to look at specific events.
Main Results in 2022
A collection of extreme heat waves and droughts has been assembled. It compares the results of analyzing the time frame between 1995 and 2019 using the “Extremal Pattern Index” with the collection of reports done in the ClimXtreme project. Times spans, region and special characteristics of the events are available.
Main Results in 2023
In our second project, we look at the impact of heat waves onto the boundary layer height. The hypothesis is that the high temperatures should cause exceptionally high boundary layers. This could have important consequences for the onset and persistence of heat waves. Our goal is to perform a comprehensive investigation, which is why we chose to use reanalysis products as our main data source. The first challenge here was to choose and implement an appropriate heat wave definition. We again use the „Extremal Pattern Index“, because the approach also allows us to compare extremal patterns for temperature and boundary layer height. As established methods did not prove effective for estimating the boundary layer height for COSMO-REA6 reanalysis data, a machine learning based approach was developed and applied. A comparison of the development in maximum daily boundary layer heights in COSMO-REA6 for a heat wave in August 2014 against normal conditions can be seen in Figure 1. Although an increase in boundary layer height can be seen in many cases on a heat wave scale, differences were mostly not significant for COSMO-REA6 on a grid point and model scale. Meanwhile, we found a clear signal in CERRA data. A publication is currently being prepared.
Figure 1. Empirical distributions of daily maximum boundary layer heights during a heat wave in August 2014. The blue bars represent the reference, i.e., boundary layer heights taken from the same region on time steps without heat waves. Both sets are plotted with transparency, s.t., the overlapping region appears darker.