An analysis of British monthly mean sea level

Abstract

Summary British monthly mean sea level variations over the period 1960-75 are examined by multiple regression analysis. The mean sea level data were recorded by an evenly spaced distribution of 29 tide gauges sited around the British Isles. Over 80 per cent of British monthly mean sea level variance can be related to seasonal changes, the static pressure effect and the influence of winds over the continental shelf. A smaller proportion of the variance can be related to the meteorological influences at stations bordering the southern North Sea. This is due to the cancelling effect of local pressure and wind at these stations. By also modelling the influence of more distant and lagged pressure distributions, it is possible to account for typically 90 per cent of the monthly mean sea level variance. The seasonal variation of mean sea level at each station is approximated by the sum of an annual and semi-annual tide. The amplitude of the mean annual tide is 7 cm. The joint contribution of the equilibrium tide, static pressure effect and local winds is shown to be too small to account for this mean amplitude. The large, unexplained component of the annual tide is probably due to the steric oscillation of the adjacent North Atlantic. Month-by-month changes in the North Atlantic steric levels may also be the cause of the significant correlation between the residual series from the multiple regression models. The monthly mean sea level series for Newlyn, Southend and Aberdeen are subjected to Fourier spectral analysis in an attempt to identify the pole tide. It is shown that the pole tide is more readily detectable in the residual rather than observed mean sea level series. The latitudinal slope of British mean sea level is 5.3 |Mp 0.4 cm per degree of latitude. This northward rise of mean sea level on both east and west coasts is difficult to explain oceanographically. It is suggested that the slope is due to a systematic error in the geodetic levelling.