What would life be like if we would know exactly when an earthquake would take place? Or when a heatwave might send thermometers soaring? How about a hurricane or a storm surge? Very different, most probably. But even then, we won’t be able to stop nature, the most we can do is to be better prepared for the inevitable. Obviously, seeing how society acts within this new world is for now the stuff of Hollywood, but it urges us to question what we know about the future and what can we learn about it.
Taking a step back from this hypothetical scenario it is worth focusing on the pressing issue at hand: climate change and global warming. These issues will irreversibly affect the lives of most people around the globe, with the projected effects putting the health and welfare of much of humanity at risk 1, 2. Financial estimations show the future ‘climate value at risk’ of global financial assets to be US$2.5 trillion in the RCP8.5 (‘business-as-usual’) scenario, with the worst outcome projected to up that figure to US$24.2 trillion based on a 2013 estimate of the assets’ value3. Moreover, losses caused by climate change are predicted to reach 23% of the global gross product by 21004, 5.
While global warming is usually described in the media outlets through the mean global temperature increase or by the sea level rise, they often fail to mention some of the most striking effects that are triggered by it - the amplification of extreme weather events. Since 1980, weather-related natural disasters incurred US$5.2 trillion in damages and caused approximately one million excess deaths6. Additionally, the Intergovernmental Panel on Climate Change reports that the impacts of recent extreme-weather events show a lack of preparedness for the current climate variability for countries at all levels of development7.
So, can we better understand the changing effects of future extreme weather events under climate change to better plan for mitigation and adaptation strategies? Work taking place at the University of Cambridge suggests that by using a data and AI-driven approach in conjunction with global climate model simulations, that it may be possible to infer such information about the future. Because coastal flooding and storm surges are usually regarded as the most damaging extreme weather events both for people and assets we have focused on these events.
Taking an overview reveals an efficient solution. By using past records of coastal floods around the UK8 and weather observation data from the past it is possible to train an AI model to understand and classify the days with floods against the days without floods. The trained AI model is combined simulations from global climate models to give a picture of the future frequency of storm surge events.
With a hypothetically perfectly trained AI model which would not make any mistakes on this task, it might be tempting to start evacuating people based on the results. But this would be a mistake, as global climate models do not predict what would happen on a given day in the future, but they show the overall trends during that future period. So, based on the output from the model, it is possible to generate a set of yearly statistics, such as the estimated number of floods, or their intensity and return periods. Those key metrics will be helpful for people, businesses and policymakers to plan better and more accurate adaptation and mitigation strategies. Such an approach can be used in addition to current physics-based models or statistical models that only infer information from past statistics.
Will this project tell us everything we need to know about coastal flooding in the UK in the future? Probably not, especially as we require some assumptions about the physical system involved, and/or we lack all the information required from up-to-date datasets. The success of this project would also enable the proposed framework to be transferred to other extreme weather events, such as heatwaves, droughts or river floods. Additionally, the analysis can be updated as more datasets become available or when the global climate models data improves with each new generation. Finally, the output of this project will add information to what is already known about extreme weather and thus, hopefully, create a benefit for the people at risk.
1 Costello, A., Abbas, M., Allen, A., Ball, S., Bell, S., Bellamy, R., ... & Patterson, C. (2009). Managing the health effects of climate change: lancet and University College London Institute for Global Health Commission. The lancet, 373(9676), 1693-1733.
2 Watts, N., Adger, W. N., Agnolucci, P., Blackstock, J., Byass, P., Cai, W., ... & Costello, A. (2015). Health and climate change: policy responses to protect public health. The lancet, 386(10006), 1861-1914.
3 Dietz, S., Bowen, A., Dixon, C., & Gradwell, P. (2016). ‘Climate value at risk’of global financial assets. Nature Climate Change, 6(7), 676-679.
4 Klusak, P., Agarwala, M., Burke, M., Kraemer, M., & Mohaddes, K. (2021). Rising temperatures, falling ratings: The effect of climate change on sovereign creditworthiness. Australian National University, Crawford School of Public Policy, Centre for Applied Macroeconomic Analysis.
5 Burke, M., Hsiang, S. M., & Miguel, E. (2015). Global non-linear effect of temperature on economic production. Nature, 527(7577), 235-239.
6 MunichRe. (2021). Natural disaster risks: Losses are trending upwards — Munich Re. https://www.munichre.com/en/risks/natural-disasters-losses-are-trending-upwards.html.
7 Field, C. B., Barros, V., Stocker, T. F., & Dahe, Q. (Eds.). (2012). Managing the risks of extreme events and disasters to advance climate change adaptation: special report of the intergovernmental panel on climate change. Cambridge University Press.
8 Haigh, I. D., Ozsoy, O., Wadey, M. P., Nicholls, R. J., Gallop, S. L., Wahl, T., & Brown, J. M. (2017). An improved database of coastal flooding in the United Kingdom from 1915 to 2016. Scientific data, 4(1), 1-10.
