The role of cyclone clustering during the stormy winter of 2013/2014:

Extratropical cyclones are the primary natural hazard affecting western Europe. They are associated with strong winds and rainfall (Lamb, 1991) which can result in significant societal impacts. For example, windstorms Anatol, Lothar and Martin in 1999 resulted in approximately €16 billion of total insured losses (Swiss Re, 2016). Cyclones contribute to more than 70% of the precipitation that falls over northwest and central Europe in winter (Hawcroft et al., 2012). Extratropical cyclones in the North Atlantic tend to travel in a northeasterly direction, forming off the east coast of North America and dissipating over the northeastern Atlantic Ocean or Nordic Sea. Clustering of extratropical cyclones is often defined as the passage of multiple high intensity cyclones through one geographical region within a relatively short period of time (Pinto et al., 2014). It has been statistically shown that extratropical cyclones cluster in the exit region of the storm track and in the vicinity of northwestern Europe (Mailier et al., 2006). The physical mechanisms behind the observed clustering that affected western Europe for several highimpact months across various winter seasons was initially explored by Pinto et al. (2014) and expanded on by Priestley et al. (2017). The latter study looked at all clustered events for 36 winter seasons from 1979/1980 to 2014/2015 and found that a typical clustering event affecting the British Isles was associated with a very strong and straight jet stream in the North Atlantic, with the jet being flanked to the north and south by Rossby wave breaking (RWB; more details on RWB in Box 1). It is hypothesised that the physical mechanisms leading to clustering include steering by the large-scale flow, which orientates storms in the same direction and provides the conditions necessary for rapid intensification (Mailier et al., 2006; Hanley and Caballero, 2012). An additional mechanism is secondary cyclogenesis, in which new cyclones form on the cold fronts of more mature cyclones (Parker, 1998); this allows for the occurrence of many storms in a short period of time. The combination of these mechanisms results in clustering (Pinto et al., 2014), that is the largescale flow allows for all cyclones that form in a region to follow the same direction; in addition, the occurrence of secondary cyclogenesis ensures that the time gap between the passage of these cyclones is reduced. December, January and February of 2013/2014 (DJF1314) was one of the most extreme DJF periods in the British Isles and the stormiest in 143 years (Matthews et al., 2014). The season was known for its persistent wet weather. The average precipitation accumulation across England and Wales was 456mm for the DJF1314 season according to the England–Wales precipitation (EWP) series (Alexander and Jones, 2001), which corresponds to 175% of the seasonal average and is the wettest such period since the record began in 1766 (Kendon and McCarthy, 2015). Several intense storms affected the British Isles during this period (such as Bernd, Dirk, Erich and Tini; all named winter storms herein were named by Freie Universität Berlin: http://www.met.fu-berlin. de/adopt-a-vortex/historie/), which brought high winds and intense rainfall to large parts of the country. The storminess persisted from early December through to late January and early–mid February (Kendon and McCarthy, 2015). This continuously unsettled and severe weather made coastal areas particularly vulnerable (Figure 1(c)), with the main rail line from Exeter to Plymouth collapsing at Dawlish (Figure 1(a)) and the flooding of the Somerset levels being a persistent feature of this season (Figure 1(b)). Since highly clustered seasons such as DJF1314, 1990/1991 and 1999/2000 are clearly able to cause large-scale environmental and socio-economic impacts, it is important to understand the atmospheric drivers causing them. The aim of this study is to investigate a short period in the DJF1314 season to determine if one anomalously stormy period follows the dynamical framework