Precipitation diagram from Jack Rogers' dissertation

A diagram from Jack's dissertation showing the three-day rainfall total from 11 to 13 August 2002 over central Europe. The bold black line represents the path of the low-pressure system, named Ilse 2, that brought this heavy rain. The circles represent the position of the low at 12-hourly intervals and are labelled with a date and time (GMT). Thin black lines demarcate country borders. The red circle indicates the location of the Zinnwald weather station.

Jack Rogers (2016, St Edmund Hall) who graduated this summer, has won one of two runners-up prizes in the 2019 British Hydrological Society competition for his dissertation entitled ‘An investigation into the influence of climate change on the extreme precipitation event that occurred over central Europe in August 2002’.

The Society offers annual awards for undergraduate students who have demonstrated excellence by exhibiting personal initiative and innovation in research/design projects in hydrology.

Jack’s dissertation used climate models to investigate how climate change could have altered the probability of the 2002 central European floods. It then went on to look at the extent to which climate models are able to simulate both the dynamics and intensity of such an event.

You can read his full abstract below:

Understanding the effect of anthropogenic warming on extreme weather events is a crucial challenge in climate science. This study uses a multi-method attribution approach to investigate the human contribution to an extreme precipitation event that occurred over central Europe in August 2002. The event caused severe flooding that led to a record-breaking 15 billion Euros of economic damage and the highest flood levels on the river Elbe since 1275.

To evaluate the role of anthropogenic warming on the event, this study uses large ensembles of a high-resolution regional circulation model generated by the weather@home project. The model generates thousands of simulations under both actual climate conditions and under conditions where anthropogenic influences have been removed. Consistent with thermodynamic expectations, a comparison of the two different ensembles shows that the risk of extreme rainfall of a similar magnitude to the 2002 event has doubled as a result of anthropogenic warming. The model validation results show a high model competency in simulating observed three-day rainfall levels and thus confidence in the results.

The study then takes the evaluation one stage further by extracting circulation analogues of the 2002 precipitation event to focus the analysis on isolated dynamically similar rainfall events. A second model evaluation is conducted for the circulation analogue approach and shows that the model struggles to fully simulate the dynamics associated with the 2002 event. This means that the circulation analogue results are associated with a greater level of uncertainty than the results of the standard attribution analysis. The circulation analogue approach results show no clear indication that the frequency or magnitude of dynamically similar storms has changed as a result of anthropogenic warming. While uncertain, this result indicates that the effect of anthropogenic climate change is less clear-cut than suggested by the standard attribution result.

Overall, this study can confidently state that in general extreme rainfall over the event region has been increasing as a result of anthropogenic warming. However, the circulation analogue approach suggests there is no clear evidence that the likelihood or magnitude of storms dynamically similar to the 2002 event have been altered by anthropogenic warming. These results highlight the value of using the standard attribution methodology in combination with the circulation analogue approach. This study highlights the need for an explicit focus on event dynamics in precipitation attribution studies and discusses why this remains challenging given current modelling capabilities.