Changing storminess? An analysis of long-term sea level data sets

This paper considers the analysis of long-term observational sea level data sets within the context of a possible change in storminess over North-West Europe. Sea level variations are studied as a proxy for storminess, i.e, i n the synoptic frequency band. In that band, a proxy correspondence is assu med between sea levels and wind stress, although with some damping of the h ighest frequencies. Trends and variability over the past 100 yr are studied . However, a complicating factor is that many data sets include astronomica l tidal influences (e.g. spring tide -neap tide cycle), which mask the sign al that is of most interest in this study. Sea level data sets from station s in the coastal zones of North-West Europe were collected, homogenised and made free of auto-correlation. To analyse these data sets, the 'quantile a nalysis method' is presented. This method involves an advanced analysis tec hnique which, on the basis of a frequency analysis, determines the roughnes s/smoothness of succeeding decades in relation to the complete time period of a data set. Possible storm-related trends and variations in this decadal quantity are defined by means of linear regression. Besides this main meth od, a time-shift variant of the quantile analysis method is applied in orde r to discuss the sensitivity of the obtained results. This sensitivity is a lso investigated with respect to the length of the time period in which a c omplete data set is split up. Although the analysis results show considerab le natural variability on relatively short (decadal) time scales, no sign o f a significant increase in storminess over North-West Europe is detected o ver the complete time period of the data sets. The results indicate a disti nction between stations in the German Eight and stations in the southern pa rt of the North Sea. In the latter area, the natural variability is more mo derate and there appears to be a tendency towards a weakening of the storm activity over the past 100 yr (not significant). Stations in the German Eig ht show a more enhanced natural variability on relatively short (decadal) t ime scales, with no indication of a weakening of the storm climate.