Records of sea level for several North Sea ports for the winter of 1953-4 have been in vestigated. They were split into 14-day intervals, and each 14-day record was Fourieranalyzed to determine if any non-astronomical periods were present. There was evidence of some activity between 40 and 50 h period, and a determination of the phase angles at different ports showed that the activity could be due to a disturbance travelling southwards from the north of the North Sea. The disturbance was partly

more »

... reflected somewhere near the line from Lowestoft to Flushing, so that one part returned past Flushing and Esbjerg to wards Bergen while the other part travelled towards Dover, and there was evidence of its existence on the sea-current records taken near St Margaret's Bay. These results were con firmed by subtracting the predicted astronomical tidal levels from the observed values of sea level and cross-correlating the residuals so obtained for each port with those found at Lowestoft. The residuals at Lowestoft and Aberdeen were compared with the meteorological con ditions, and it was found that, although they could be attributed to a large extent to con ditions within the North Sea, there was an additional effect due to a travelling surge which was of the same order of magnitude at both Lowestoft and Aberdeen and which was closely related to the rate of change with time of the atmospheric pressure difference between Wick and Bergen. [ 260 ] many smaller surges were discovered. As a first measure in the analysis, it was decided to use the same technique as has been used for the investigation of sea waves, using the wave analyzer described by Barber, Ursell, Darbyshire & Tucker (1946). The period October 1953 to March 1954 was chosen because the flood warning system was in operation during this time, and thus hourly values of both observed sea levels and predicted tides were available for a number of ports, in particular Aberdeen, Leith, Tyne entrance and Lowestoft. Records were also available from-Dover, Flushing, Esbjerg and Bergen. The predicted values were not required with the wave analyzer, which dealt only with the observed values. The period October 1953 to January 1954 was split into 14-day intervals and the tidal records for the ports mentioned above for these times were analyzed. There were also available records of the current through the Straits of Dover, which were obtained by measuring the potential difference induced across a tele phone cable lying between St Margaret's Bay and Sangatte (Bowden 1956). These records were analyzed for the same times as the tidal records.