Comparison of the GMST covariate and the time slice method for probabilistic extreme weather event attribution

The probabilistic attribution has become a valuable tool for analysing the influence of anthropogenic climate change on recent extreme weather events in rapid attribution studies.

For this rapid analysis, the methods of the World Weather Attribution group (WWA) described by Philip et al. (2020) can be used. This method is a straightforward option for evaluation especially in a trend of operationalizing the probabilistic attribution. When estimating the General Extreme Value (GEV) distribution, the global mean surface temperature (GMST) is introduced as covariate such that the GEV shifts or scales with this temperature. This has the advantage that the complete timeseries of every observation and model dataset can be analysed to detect anthropogenic influences. The covariate method provides a trend proportional to the covariate and allows extrapolation of existing datasets to a past or future climate. Depending on the context, this can be seen as an advantage or disadvantage.

To avoid a proportional trend, an alternative method consists in evaluating time slices instead. Two 30-year blocks for a past climate and the current climate or a counterfactual and factual climate are analysed. Optionally, a future scenario from climate models can be included. While only one GEV with one additional covariate is estimated to describe a single model in the previous method, a standard GEV is used for every defined slice. In this case, the single 30-year climate periods are independent from the other time slices. The challenge here is the selection of climate models and scenarios which simulate a similar trend of anthropogenic impacts. Additionally, observation datasets can only be used when the time series is long enough to allow extraction of two independent time slices of 30 years each.

The difference between the two methods and the difference in the results will be analysed and presented. Both methods can be understood as part of the same toolbox and are both equally valid. Here, the main interest is in the ability to understand and explain upcoming differences in extreme weather attribution studies between the two methods.

The research of this project is part of the ClimXtreme Network, funded by the German Federal Ministry of Education and Research (BMBF). Focus of this project are extreme weather events and impacts caused by anthropogenic climate change.

Philip, S., Kew, S., van Oldenborgh, G. J., Otto, F., Vautard, R., van der Wiel, K., King, A., Lott, F., Arrighi, J., Singh, R., and van Aalst, M.: A protocol for probabilistic extreme event attribution analyses, Adv. Stat. Clim. Meteorol. Oceanogr., 6, 177–203, https://doi.org/10.5194/ascmo-6-177-2020, 2020.