Attribution of floods in the Okavango basin, Southern Africa

• Published in: Journal of hydrology (Amsterdam) 2014-04, Vol.511, p.350-358
• Main Authors: Wolski, Piotr, Stone, Dáithí, Tadross, Mark, Wehner, Michael, Hewitson, Bruce
• Format: Article
• Language: English
• Subjects: Air pollution; Anthropogenic climate change; Attribution; Climate; Computer simulation; Economics; Flood; Floods; Greenhouse effect; Greenhouse gases; Hydrology; Okavango; Risk
• ISSN: 0022-1694; 1879-2707
• DOI: 10.1016/j.jhydrol.2014.01.055

•We assess role of anthropogenic climate change in recent floods in the Okavango basin.•Hydrological model is forced with simulated current and pre-industrial climates.•Anthropogenic climate change has reduced probability of high floods in the Okavango.•Reduction is due to increased evaporation; rainfall changes are minor. In the charismatic wetlands of the Okavango Delta, Botswana, the annual floods of 2009–2011 reached magnitudes last seen 20–30years ago, considerably affecting life of local populations and the economically important tourism industry. In this study, we analyse results from an attribution modelling system designed to examine how anthropogenic greenhouse gas emissions have contributed to weather and flood risk in our current climate. The system is based on comparison of real world climate and hydrological simulations with parallel counterfactual simulations of the climate and hydrological responses under conditions that might have been had human activities not emitted greenhouse gases. The analyses allow us to address the question of whether anthropogenic climate change contributed to increasing the risk of these high flood events in the Okavango system. Results show that the probability of occurrence of high floods during 2009–2011 in the current climate is likely lower than it would have been in a climate without anthropogenic greenhouse gases. This result is robust across the two climate models and various data processing procedures, although the exact figures for the associated decrease in risk differ. Results also differ between the three years examined, indicating that the “time-slice” method used here needs to be applied to multiple years in order to accurately estimate the contribution of emissions to current risk. Simple sensitivity analyses indicate that the reduction in flood risk is attributed to higher temperatures (and thus evaporation) in the current world, with little difference in the analysed domain’s rainfall simulated in the two scenarios.