10. Human Contribution to the Record Sunshine of Winter 2014/15 in the United Kingdom

Observational data of sunshine duration since 1930 from the Met Office National Climate Information Centre (NCIC; Perry and Hollis 2005) reveal that winter 2014/15 was the sunniest in the United Kingdom (Fig. 10.1a). The common perception of drab British winters is seemingly challenged by the increasing trend of 2.4 ± 0.7 sunshine hrs decade−1 (mean ± standard deviation) during 1930–2015 (Fig. 10.1a). With winters in the region projected to become warmer and wetter in a changing climate (van Oldenborgh et al. 2013), increasing sunshine would suggest longer sunny spells between heavier rainfall events. Brighter winters may also enhance solar energy production. Annual sunshine over western Europe was found to follow periods of dimming in the 1960–80s and brightening thereafter, while large positive seasonal trends are particularly evident in winter (Sanchez-Lorenzo et al. 2008). Contrary to the changes in Europe, a sunshine decline in recent decades has been observed in parts of the world where aerosol concentrations have been increasing, such as China and the Indian subcontinent (Wang et al. 2012; Liao et al. 2015; Niroula et al. 2015). We attempt to formally establish the role of the overall anthropogenic forcing on the climate based on ensembles of simulations with and without anthropogenic effects produced with an atmospheric model. This well-established methodology (Pall et al. 2011; Christidis et al. 2013) provides distributions of climatic variables in the actual (ALL forcings) and natural (NAT) climate, constructed with the two ensembles. Probabilities P1 and P0 of a threshold exceedance computed with the ALL and NAT simulations help assess the anthropogenic effect in terms of the fraction of attributable risk (FAR; Allen 2003), defined as 1 − (P0 / P1). FAR values close to unity indicate prominent human influence on the event. Changes in the return time of extreme events (estimated from inverse probabilities) can also be examined. As models do not provide a sunshine duration diagnostic, we employ the downward solar (SW) flux at the surface as a proxy (Fig. 10.1a). Observed winter sunshine hours and solar radiation averaged over the United Kingdom have a correlation of 0.9 over the common observational period, though individual years may differ in sign of anomaly (e.g., 2010). Cloud cover (correlation coefficient 0.3 for inverse variable estimated from observations) would be less suitable in our analysis, as it also incorporates a nighttime component. SW winter flux in 2014/15 is a joint record together with 2007/08, though flux observations cover a considerably shorter period than sunshine. We employe d t he Had le y C ent re e vent attribution system (Christidis et al. 2013), built on the HadGEM3–A model, to generate the ALL and NAT simulations. A major upgrade of the model was recently undertaken within the EUCLEIA project (http://eucleia.eu/). As a result, our system now features the highest resolution model used in attribution studies, with 85 vertical levels and about 60-km horizontal resolution. Ensembles of 15 simulations were produced for both the ALL and NAT experiments, which cover the period 1960–2013. Observed sea surface temperatures (SSTs) and sea ice data (Rayner et al. 2003) were used as boundary conditions in the ALL simulations. An estimate of the anthropogenic warming in the SSTs obtained from atmosphere–ocean coupled models (Stone 2013) was subtracted from the SST observations in the NAT simulations and the sea ice was adjusted accordingly (Christidis et al. 2013). Figure 10.1b depicts the modeled time series of the SW winter flux anomaly relative to 1961–90 corresponding to the ensemble Extreme winter sunshine in the United Kingdom, as observed in the record high 2014/15 season, has become more than 1.5 times more likely to occur under the influence of anthropogenic forcings.