Relationship between erythema effective UV radiant exposure, total ozone, cloud cover and aerosols in southern England, UK

Evidence of an underlying trend in the dependence of erythema effective ultraviolet (UV) radiant exposure (H-er) on changes in the total ozone, cloud cover and aerosol optical depth (AOD) has been studied using solar ultraviolet radiation measurements collected over a 25-year period (1991-2015) at Chilton in the south of England in the UK. The monthly mean datasets of these measures corrected for underlying seasonal variation were analysed. When a single linear trend was fitted over the whole study period between 1991 and 2015, the analyses revealed that the longterm variability of H-er can be best characterised in two subperiods (1991-2004 and 2004-2015), where the estimated linear trend was upward in the first period (1991-2004) but downward in the second period (2004-2015). Both cloud cover (CC) and total ozone (TO) were found to have a highly statistically significant influence on H-er, but the influence of the AOD measure was very small. The radiation amplification factor (RAF) for the erythema action spectrum due to TO was -1.03 at constant levels of CC over the whole study period; that is, for a 1.0% increase in TO, H-er decreases by 1.03 %. Over the first period (1991-2004), the RAF related to CC was slightly higher at 0.97 compared to that for TO at 0.79. The proportion of the change in H-er explained by the change in CC (47 %) was much greater than the proportion explained by changes in TO (8 %). For the second period (2004-2015), the pattern reversed, with the observed RAF related to TO being -1.25, almost double that of CC (-0.65). Furthermore, in this period the proportion of variation in H-er explained by TO variation was 33 %, double that of CC at 16 %, while AOD changes had a negligible effect (1 %). When the data were examined separately for each season, for the first period (1991-2004) the greatest effect of TO and CC on H-er (i.e. the largest RAF value) was found during spring. Spring was also the season during which TO and CC variation explained the greatest proportion of variability in H-er (82 %). In the later period (2004-2015), the RAF and greatest influence of TO and CC were observed in winter (67 %) and the AOD effect explained a further 5% variability in H-er. This study provides evidence that both the increasing trend in H-er for 1991-2004 and the decreasing trend in H-er for 2004-2015 occur in response to variation in TO, which exhibits a small increasing tendency over these periods. CC plays a more important role in the increasing trend in H-er for 1991-2004 than TO, whereas for 2004-2015, the decreasing trend in H-er is less associated with changes in CC and AOD.