applets.

Tide Machine

← All apps

About Tide Machine

Not for navigation: use official tide tables. Every harbour here is an example of a type, with rounded figures. None is the real constants of a real port.

The rise and fall of the tide at a harbour can be written as a sum of cosine waves, each with a speed fixed by astronomy and a height and phase measured from that harbour's own records. Add the waves up and you have the tide for any time you like. This page is a machine that does that addition with a wire.

THE MACHINE: eight waves, each a brass wheel with a crank pin. The pin's distance from the centre is the wave's height, its starting angle the phase. Each crank moves a pulley up and down, one wire runs under and over all the pulleys, and the pen on the end of the wire draws whatever length is left. Set the sliders, then RUN, or drag the crank handle by hand. Watch the fortnight appear as M2 and S2 drift in and out of step. Pick 365 to see a whole year at once: the springs themselves swell and shrink twice a year, because K2 comes into step with S2 every half year. In real harbours the biggest springs fall near the equinoxes in March and September. The machine has no seasonal change in mean sea level (the Sa and Ssa waves), so it shows only the astronomical part of that.

HARBOURS: six kinds of tide, and the form number that sorts them. Picking one sets up the machine.

WHY: the Sun, the Earth and the Moon seen from above the north pole. Drag the Moon or change the date. The arrows show the true tide-raising force, which is the difference between the pull at a place and the pull at the Earth's centre.

TIDE TABLE: highs and lows for a week, found by root-finding on the curve itself. The Rule of Twelfths is checked against the tide you pick, and a sill calculator finds when there's enough water.

THE STORY

Newton explained in the Principia (1687) that tides come from the difference in the pull of the Moon and the Sun across the Earth. This gives the two bulges of the equilibrium theory. In the 1770s Laplace wrote the equations for a real ocean in motion, and showed that the tide is a response to the forcing at certain fixed speeds, which is why adding up waves can work at all.

From 1867 William Thomson, later Lord Kelvin, led a British Association committee on harmonic analysis of tides. Kelvin then designed a machine to do the adding up. The first was built in 1872-73 with Edward Roberts, Alexander Légé making it. It summed ten waves and survives in the Science Museum in London. Machines of the kind went on working for about a century, in several countries.

In 1921 Arthur Doodson published a full harmonic development of the tide-raising forces, with hundreds of constituents. The Liverpool Tidal Institute, where he worked and which moved to Bidston Observatory in 1929, became a world centre of tide prediction.

For D-Day, the Admiralty's tidal expert, Commander William Farquharson, asked Doodson in 1943 for predictions for a place identified only by a code, Position Z. Doodson and his colleagues ran them at Bidston on their machines, the Roberts-Légé and a Kelvin machine. The Doodson-Légé machine you may read about came later, in the late 1940s.

HOW IT WORKS

Each wave is a cos(speed × t − phase). The speeds are built from the rates of the Earth's turn, the Moon's orbit, the Sun's apparent orbit and the Moon's perigee, and they match the published ones. M2 and S2, at 28.9841° and 30° an hour, slip a full turn apart every 14.77 days, which is half the 29.53-day month between one new moon and the next. That's springs and neaps.

The wire in the drawing is exact. Its straight runs are all vertical, and a pulley going down by d takes up 2d of wire, so the pen rises by twice the sum of the pulleys. The drawn pen's height is worked out from the wire's length, not from the tide formula, and the two agree.

Highs and lows are where the slope of the sum crosses zero, found by bisection. Turning points less than 1 cm apart are taken out in pairs, since a table wouldn't show them. The form number F = (K1 + O1) / (M2 + S2) uses Courtier's classes: below 0.25 semidiurnal, up to 1.5 mixed and mainly semidiurnal, up to 3 mixed and mainly diurnal, above that diurnal.

The Rule of Twelfths says that over the six hours of a tide the water moves 1, 2, 3, 3, 2 and 1 twelfths of the range. A perfect cosine is within 1.7% of the range of that. Here the rule's "hours" are one sixth of the tide's real length; the fit is called good within 5% if the tide takes about six hours, poor if it's over 10% out or the tide is far from six hours long.

WHAT IT LEAVES OUT

FURTHER WATCHING AND READING

Not for navigation: use official tide tables.
DAYS
CRANK
Or drag the crank handle.
WAVEPERIODHEIGHTPHASE

The record: highs and lows on the roll