Summary/Abstract
Main findings
● Extreme flooding occurred as a direct consequence of a 2-day heavy rainfall event at the coast of Eastern South Africa (see Fig. 1). We therefore analyse the annual maximum 2-day total rainfall in the affected area.
● The South African Weather Service and eThekwini municipality issued early warnings. There are indications, however, that the warnings had limited reach and that the people who did receive them may not have known what to do based on them.
● While the full profile of the impacts on human life and livelihoods has yet to be analysed, initial assessments show that the floods disproportionately affected marginalised communities, with particular devastation in informal settlements. Thus, the magnitude of this disaster on these groups has been exacerbated by pre-existing structural vulnerability in the region.
● The magnitude of the event is given by maximum 2-day rainfall, averaged over the homogenous area to make observations and model output comparable. The defined event has a return time of about 20 years in today’s climate in the ERA5 observational data set. An event of this magnitude would have been rarer in a 1.2°C cooler world, with a return time of about 40 years.
● At individual stations that had highest rainfall amounts, return times are much higher, e.g. 1 in 200 years at Mount Edgecombe.
● To determine the role of climate change in these observed changes we combine observations with climate models. We conclude that greenhouse gas and aerosol emissions are (at least in part) responsible for the observed increases.
● Furthermore, when taking models into account as well the changes indicate a clear increase in likelihood and intensity. We conclude that the probability of an event such as the rainfall that resulted in this disaster has approximately doubled due to human-induced climate change. The intensity of the current event has increased by 4-8%.
● Heavy rainfall events are projected to increase in frequency and magnitude in the future with additional global warming levels.