The seemingly high number of major floods affecting the UK over the past decade have resulted in questions about whether these floods are ‘caused’ by climate change and new research published in the International Journal of Climatology explores this question.
The stand out finding from the paper is that greenhouse gas emissions (i.e. human induced climate change) are likely to have increased the chance of flooding in winter 2013/14 in many parts of Great Britain, with a stronger influence on longer duration flood events (peaks lasting more than 10 days). This paper strengthens the view that climate change is already happening and having an influence on the risk of flooding today.
Academic and industry collaboration
JBA Risk Management® has continued working alongside climate, hydrological and mathematical modellers following an earlier project published in Nature Climate Change. The contributing researchers include: Dr Alison Kay, Centre for Ecology & Hydrology (CEH); Naomi Booth and Dr Emma Raven, JBA Risk Management; Professor Rob Lamb, JBA Trust; Dr Nathalie Schaller, Centre for International Climate and Environmental Research (CICERO); and Dr Sarah Sparrow, Oxford University.
Climate and hydrological modelling
Whilst no single flood event can be directly attributed to greenhouse gas levels, using a technique called “probabilistic event attribution” (PEA), it is possible to assess how the chance of occurrence of a given flood event has been altered by emissions.
To achieve this, large ensembles of climate model runs are generated to represent the climate as it is now, as well as how it could have been had no anthropogenic emissions occurred. For this study, each ensemble run was a different plausible realisation of the weather between December 2013 and February 2014 (i.e. representing the Winter 2013/14 floods that affected the south of England and Thames catchment).
The climate data from each ensemble simulation was run through a hydrological model and the flow peaks were extracted. The analysis of the flow data across all of the ensembles provides an indication of whether past emissions have increased or decreased the chance of peak river flows.
Impact analysis using catastrophe modelling
One of the novel elements in this project, achieved though the industry-academia collaboration, was the use of catastrophe modelling to explore the impact on society, in this case represented by the count of properties affected. For this part of the study we focused on the Thames, creating a catastrophe model specific to the Thames area by extracting the Thames basin out of the JBA UK Flood Model.
The catastrophe model had all of the features of our commercial model, with the fundamental difference of replacing our normal event set with one representing the peak river flows generated from the climate and hydrological modelling. Instead of an event set made up of thousands of different plausible events, this research event set was a conditional one, made up of thousands of realisations of the same event (i.e. the Winter 2013/14 flood event).
Each event realisation represented a plausible trajectory for the weather over the winter season, starting with the sea surface temperature patterns and atmospheric conditions that prevailed at the start of Winter 2013/14, which were adjusted, using global climate models, to remove the effects of historical greenhouse gas emissions. The large event set enabled the research team to account for the between-model uncertainty (i.e. different climate models giving different results) and the uncertainty inherent in modelling weather over a single season, including the sensitivity of weather predictions to small errors in initial conditions.
A second event set represented the weather of Winter 2013/14 under the present-day climate. By running these research event sets through the Thames catastrophe model, we could explore the count of properties affected by flooding in each ensemble run and compare the results for the present-day climate with what would have occurred in a “natural” world without industrial greenhouse gas emissions.
Across the full range of Winter 2013/14 runs, an average of 457 more properties were found to be affected by flooding due to climate change, compared with the number that would have been affected in the absence of greenhouse gas emissions. These results have an uncertainty band ranging from a reduction of 1334 properties for some climate models to an increase of 4,605 for others.
Although this is a relatively small number compared with the total number of claims in the actual flood event, it represents the first signs of a material contribution to risk caused by climate change, which is expected to have increasingly profound effects over the coming decades.
Wider application
Alongside the paper’s main findings, there are three additional aspects of this work to share:
- It continues to emphasise that climate change is not just something that will impact our future in decades to come and supports the view that recent flood events, and ones that might happen this year, are already being influenced by anthropogenic greenhouse gas emissions.
- It shows the importance of considering impacts at a high resolution because climate changes are spatially variable.
- As noted in the paper’s conclusion: “detailed and proven climate models and hydrological models are required, along with inundation and damage models, to reliably investigate climate-driven changes in floods and their impact on people”. You can find out more about our UK Climate Change Flood Model here.
If you would like to find out more about our research or our Climate Change Model, please get in touch.
This work was supported by the NERC/CEH Water and Pollution Science Programme; JBA Trust and EUCLEIA (EU FP7 grant 607085). The full paper can be accessed here.
References:
(1) Schaller N., Kay AL., Lamb R. et al. Human influence on climate in the 2014 Southern England winter floods and their impacts. Nature Climate Change (2016) 6 (6), 627-634.
(2) Allen M. Liability for climate change. Nature (2003) 421, 891-892.
(3) Jézéques A., Dépoues V., Guillemot H. et al. Behind the veil of extreme event attribution. Climatic Change (2018) 149: 367. DOI: 10.1007/s10584-018-2252-9
