Flooding in Oxfordshire, February 2014.  Photo: Julia Lawrence.

Potential influences on the United Kingdom's floods of winter 2013-14

Dr Chris Huntingford, Centre for Ecology & Hydrology

Last winter, severe flooding affected large parts of the UK.  In a paper published in Nature Climate Change, scientists at the Centre for Ecology & Hydrology, working with colleagues from the Met Office and a number of universities, looked at the possible drivers behind the floods.  Chris Huntingford, a climate modeller based at the Centre for Ecology & Hydrology in Wallingford, Oxfordshire, was the lead author of the paper.

None of the individual rainfall events in the UK in recent months was unprecedented, but the weather patterns behind them persisted for three months causing a near-continuous succession of westerly storms.  This had the cumulative effect that for much of the southern UK, the total winter rainfall was record-breaking.  Preliminary analysis suggests that particularly warm ocean conditions and heavy rainfall in and around Indonesia triggered wind patterns across the Pacific that travelled northwards before ultimately drawing cold air down across the USA.  This in turn forced a particularly strong and persistent jet stream across the Atlantic and towards the UK.  The Met Office is now studying this sequence of events in significantly more detail.

Questions arise as to whether fossil fuel burning could have a role.  We have reviewed existing research literature for Earth system factors that may be both changing through global warming, and additionally are identified as influences on storm features for the UK.  As expected, this confirms how complex and interconnected the climate system is.  Multiple possible UK rainfall drivers are identified that link to the state of the oceans, the atmosphere and sea-ice extent.  Interestingly the recent rapid decrease in Arctic sea-ice that is widely attributed to global warming, for the UK at least is often portrayed as likely to bring more easterly winds and colder conditions.  The previous three winters had these features for some of the time, in marked contrast to winter 2013-14.  Although the precise details of linkages between changing large-scale features of the climate system and UK rainfall intensity are still not fully understood, we hope our review article is a complete list of such connections.  To apply that frequently used expression, we trust there are no ‘unknown unknowns’ lurking out there we have yet to consider.

Assuming that we do have a pretty good idea of all drivers expected to affect rainfall, and that require on-going computer modelling, three challenges are noted in how to proceed.  These are: (1) the need for continued enhancement of physical process representation via ever better parameterized differential equations of the oceans, atmosphere and ice-sheets; (2) increase further the numerical grid resolution of climate models, on which these equations are calculated; and (3) undertake significantly higher numbers of simulations, all with slightly different initial conditions, creating a large ensemble of projections.  The call for better resolution is because some characteristics of storms occur on fine spatial detail, thus needing small spacings between grid-points on which calculations are updated.  The request for large ensembles is because extremes, by definition, are rare events, and so we need to ensure that all heavy rainfall ‘return times’ are fully sampled.  This is both for pre-industrial and for raised levels of atmospheric greenhouse gases.

During the major flood events affecting much of southern England from December 2013 to February 2014, it was inevitable that questions would be asked as to whether fossil burning could have a role.  It is always (and correctly) stated that no single observed extreme event can be formally attributed to human-induced changes to atmospheric composition.  But a statistic can be derived that assesses any changing probability of a particular extreme event occurring, a quantity sometimes referred to as ‘Fractional Attributable Risk’.  By satisfying the three challenges we listed above, we will get near to stating if humans are increasing, decreasing or leaving invariant the chances of rainfall events of the type witnessed.  However, even now limitations remain on computer speed and resource, and expenditure on climate research can only ever be finite.  Hence an especially lively debate will now occur as to what constitutes the optimal balance between pursuing these three challenges, in order to get us most quickly towards the required answers.

Anyone studying meteorological systems, or the full Earth system, soon realizes of course how tightly coupled all features are of the climate system.  In this review, by trying to collate in to a single paper the main factors affecting UK rainfall, this did though provide a timely reminder of such comprehensive interconnections.  Understanding these further suggests a very interesting time lies ahead for climate change research.