The most precise way of estimating the dissipation of tidal energy in the oceans is by evaluating the rate at which work is done by the tidal forces and this quantity is completely described by the fundamental harmonic in the ocean tide expansion that has the same degree and order as the forcing function. The contribution of all other harmonics to the work integral must vanish. These harmonics have been estimated for the principal M
2
tide using several available numerical models and despite the often significant difference in the detail of the models, in the treatment of the boundary conditions and in the way dissipating forces are introduced, the results for the rate at which energy is dissipated are in good agreement. Equivalent phase lags, representing the global ocean-solid Earth response to the tidal forces and the rates of energy dissipation have been computed for other tidal frequencies, including the atmospheric tide, by using available tide models, age of tide observations and equilibrium theory. Orbits of close Earth satellites are periodically perturbed by the combined solid Earth and ocean tide and the delay of these perturbations compared with the tide potential defines the same terms as enter into the tidal dissipation problem. They provide, therefore, an independent estimate of dissipation. The results agree with the tide calculations and with the astronomical estimates. The satellite results are independent of dissipation in the Moon and a comparison of astronomical, satellite and tidal estimates of dissipation permits a separation of energy sinks in the solid Earth, the Moon and in the oceans. A precise separation is not yet possible since dissipation in the oceans dominates the other two sinks: dissipation occurs almost exclusively in the oceans and neither the solid Earth nor the Moon are important energy sinks. Lower limits to the
Q
of the solid Earth can be estimated by comparing the satellite results with the ocean calculations and by comparing the astronomical results with the latter. They result in
Q
> 120. The lunar acceleration
n
, the Earth’s tidal acceleration
O
T
and the total rate of energy dissipation
E
estimated by the three methods give astronomical based estimate —1.36 —28±3 —7.2 ± 0.7 4.1±0.4 satellite based estimate —1.03 —24 ±5 — 6.4 ± 1.5 3.6±0.8 numerical tide model — 1.49 —30 ±3 —7.5± 0.8 4.5±0.5 The mean value for
O
T
corresponds to an increase in the length of day of 2.7 ms cy
-1
. The non-tidal acceleration of the Earth is (1.8 ± 1.0) 10
-22
s
~2
, resulting in a decrease in the length of day of 0.7 ± 0.4 ms cy
-1
and is barely significant. This quantity remains the most unsatisfactory of the accelerations. The nature of the dissipating mechanism remains unclear but whatever it is it must also control the phase of the second degree harmonic in the ocean expansion. It is this harmonic that permits the transfer of angular momentum from the Earth to the Moon but the energy dissipation occurs at frequencies at the other end of the tide’s spatial spectrum. The efficacity of the break-up of the second degree term into the higher modes governs the amount of energy that is eventually dissipated. It appears that the break-up is controlled by global ocean characteristics such as the ocean-continent geometry and sea floor topography. Friction in a few shallow seas does not appear to be as important as previously thought: New estimates for dissipation in the Bering Sea being almost an order of magnitude smaller than earlier estimates. If bottom friction is important then it must be more uniformly distributed over the world's continental shelves. Likewise, if turbulence provides an important dissipation mechanism it must be fairly uniformly distributed along, for example, coastlines or along continental margins. Such a global distribution of the dissipation makes it improbable that there has been a change in the rate of dissipation during the last few millennium as there is no evidence of changes in ocean volume, or ocean geometry or sea level beyond a few metres. It also suggests that the time scale problem can be resolved if past ocean-continent geometries led to a less efficient breakdown of the second degree harmonic into higher degree harmonics.
Lunisolar Atmospheric Tides: A New Approach