Summary A small error in the computations of Cartwright & Tayler has been corrected, resulting in improved tables of the time harmonics of the principal terms in the tide‐generating potential. The correction removes certain previously anomalous features, and otherwise affects the amplitudes in their 5th decimal places.

The results of a ten-year programme of measurement and analysis of the tides in the open sea west of Europe are described. Instrumental techniques were developed during the period from capacitance-plane sensors used in shelf seas to strain-gauges and quartz crystal sensors used in oceanic depths up to about 4 km. The measuring stations form two chains, one on the edge of the shelf surrounding the British Isles, and the other surrounding an oceanic area bounded by Scotland, Iceland, the Azores and South Portugal, which is subdivided by a line of stations along the 53 1/2 ° parallel from Inishbofin to the mid-Atlantic Ridge. Records from tide gauges at strategic oceanic sites were collected and analysed over long periods to provide a basis for accurate analysis of the relatively short-term pelagic pressure records. Tidal admittances to the generating potential vary smoothly over the region except for a change in the age (phase slope) of the diurnal tides towards Reykjavik. The shelf-edge chain included current meters, enabling a direct estimation of the M 2 power transmission on to the shelf, namely 60 GW between Malin Head (Eire) and Floro (Norway) and 190 GW between Valentia (Eire) and the Brittany coast near Ouessant. When combined with known power fluxes into the Irish Sea and through the Dover Strait, these give figures for tidal dissipation in the North and Scottish seas that considerably exceed direct estimates based on bottom friction, confirming similar results by Robinson (1979) for the Irish Sea. Dissipation in the English Channel and western Celtic Sea is also much greater than was assumed by Miller (1966). Further, when the eastward flux of energy out of the ocean is computed solely on the basis of the cotidal map drawn for M 2 it is found that about 360 GW is fed into the shelf region between Lisbon and Valentia Island, indicating a substantial loss along the Biscay shelf south of Ouessant. Thus, all estimates of energy loss in the region studied greatly exceed previously assumed values. Dissipation within the main oceanic area itself is, however, about as small as expected. The diurnal tides in the shelf zones are considerably unlike Kelvin-waves, having short wavelength variations and a tendency for larger amplitudes on the open shelf than at the coast. This probably results from the presence of shelf-wave modes previously identified only in the region of St Kilda. The semi-diurnal tides however have much more regular behaviour along the shelf, with a Kelvin-wave profile west of Inishbofin. The tidal constants at two representative oceanic sites are compared with estimates from recent computational models, with wide ranges from fair agreement to disagreement, showing that tidal computations for the North Atlantic are not so uncontroversial as is often supposed. Avoiding reliance on computational modelling, the authors present cotidal maps of plausible accuracy for the constituents O 1 , K 1 , M 2 , S 2 and M 3 . Currents and power fluxes calculated from the M 2 map are presented and discussed. The principal oceanic amphidromes for the diurnal and semi-diurnal tides are west of the area studied, but their positions are estimated reliably. A major semi-diurnal amphidrome east of Iceland is confirmed, but there is no evidence for a supposed amphidrome centred on the Faeroe Islands. The cotidal map for M 3 is the first ever produced for an oceanic area. It contains two amphidromes near the mid-Atlantic Ridge.

No abstract