New research has found the UK's most costly weather events are very likely linked rather than independent as previously thought.

Dr John Hillier, from Loughborough University’s Geography Department, analysed time-series of weather and insurance data, and identified intra-annual links between windstorms, shrink-swell damage (associated with drought), and flooding. Dr Neil Macdonald University of Liverpool) and Dr Gregor Leckebusch (University of Birmingham) also contributed to the study.

The data show that the likelihood of a major flood and windstorm event both happening in a year could be up to one-and-a-half times more likely than caused by random chance. This includes not only pairings that are relatively close in time, such as the heavy floods and storms in 2013-14, but also when the events are more separated – for example, the extensive rainfall in winter and summer 2007.

The findings could have implications for the insurance industry and Government agencies involved in planning our resilience to natural hazards, who currently consider these major weather events independently of each other; The single-hazard models currently used by insurers need joining to give an accurate overall assessment of risk.  

Dr Hillier explained: “My analysis demonstrates a systematic, long-term link between major non-synchronous weather events, which means that multi-hazard weather challenges such as the storms and flooding of 2013/14 could occur more often that we might expect.

“This is important because if two serious events happen in the same period, our 'worst case' becomes worse, and agencies and insurers need to have thought about this to put adequate plans in place.

“The next step is to extend the study to include more data and see how this pattern varies in strength and impact in different areas across the country, to understand what processes drive the linkages, and to identify any other correlations.”

The findings are described in the paper Interactions between apparently ‘primary’ weather-driven hazards and their cost published in Volume 10 of IOP Science.