The wave that proved rogue waves are real Image: Office of NOAA Corps Operations. (Public domain)

The wave that proved rogue waves are real

At 15:20 on 1 January 1995 a downward-looking laser on the Draupner E gas platform, 160 km off the coast of Norway, recorded a wave whose trough-to-crest height was 25.6 m. The crest alone stood 18.5 m above the mean sea level. The storm that afternoon was heavy but not unusual for the North Sea in winter, with a significant wave height of about 12 m. Nothing on the platform was damaged beyond some equipment on a lower deck, and the record sat in the instrument's memory until engineers at Statoil looked at it. Until that day the giant waves in sailors' stories had no measurement behind them. After it, they did.

What the laser saw

The instrument measured the distance to the sea surface several times a second for twenty minutes at a time. The record of that afternoon shows an ordinary storm sea, waves of 8 to 12 m rolling past, and then one crest more than half again as high as anything around it, with a shallow trough before it and a shallow trough after it. The wave was steep, sharp at the top, and gone in a few seconds. The engineer who first analysed the record, Sverre Haver, checked whether the laser could have been fooled by spray or by a glitch and found nothing wrong with it. The platform's lower deck, 18 m above the sea, had been hit, which agreed with the crest height. Both checks held, and the Draupner wave became the first rogue wave accepted by the scientific community as real.

How rare should it have been?

Oceanographers describe a sea by its significant wave height, written \(H_s\), which is the mean height of the highest third of the waves. It is close to what an experienced observer would report as "the wave height" by eye. In a sea that has settled into a statistical steady state, the heights of individual waves follow, to a good approximation, the Rayleigh distribution, and the probability that a given wave is higher than some height \(h\) is

\[ P(H > h) = \exp\!\left(-\frac{2h^2}{H_s^2}\right). \]

Put in the Draupner numbers. With \(h = 25.6\) m and \(H_s = 12\) m the exponent is \(2 \times (25.6/12)^2 = 9.1\), and the probability is \(e^{-9.1}\), about one in 9,000. A storm sea with a period of about 12 s produces 300 waves an hour, so a wave twice the significant height, which the law puts at one in 3,000, should pass a fixed point every ten hours or so in a long storm, and one as high as Draupner's every thirty hours. That is rare, but it is not the kind of number that overturns a theory. The trouble lay in the crest.

In a linear sea the crest and the trough share the height equally, so a 25.6 m wave would have a crest of 12.8 m. The Draupner crest was 18.5 m, and the Rayleigh probability for a crest that high, \(\exp(-8 \times (18.5/12)^2)\), is about one in 200 million waves. A platform would wait several thousand years of continuous storm for one. Linear theory could produce the height; it could not produce the shape. Something was lifting the crest and flattening the troughs.

Two explanations

Part of the answer was already in the textbooks. Real waves are not sinusoids. When a wave is steep, the second-order correction to the linear theory sharpens its crest and broadens its trough, and the crest rises above half the height by an amount that grows with the steepness. For the Draupner wave this correction accounts for a few metres but not for all of it.

The rest was argued over for twenty years. One school pointed to the modulational instability, a process known from optics and from laboratory wave tanks in which a train of waves of nearly equal wavelength spontaneously concentrates its energy into one or two waves at the expense of its neighbours. It produces spectacular rogue waves in a narrow flume. The difficulty is that the open sea is not a narrow flume: its waves come from many directions at once, and theory and experiment both show that directional spread switches the instability off.

The other school argued that nothing exotic was needed. In 2016 Francesco Fedele of the Georgia Institute of Technology and his colleagues took the Draupner record and two other measured rogue waves, the Andrea wave of 2007 and the Killard wave of 2014, and reproduced them with a model in which waves of different lengths and directions simply added up, with the second-order sharpening included, and no modulational instability at all. Their paper in Scientific Reports showed that the piling-up of a directionally spread sea is enough to produce the measured crests at the measured rate. The rogue wave, in this picture, is the ordinary tail of the distribution once the real shape of waves is taken into account.

In 2019 a group at the University of Oxford and the University of Edinburgh recreated the Draupner wave in a circular wave tank by sending two wave groups at each other at an angle of about 120°. The crossing allowed the crest to grow without breaking, and the reproduced wave had the measured height and the measured shape. The Draupner storm, it turned out, had been a crossing sea, with swell from one direction and wind waves from another, which fits.

What has changed since

Rogue waves are now counted rather than doubted. Buoys, platform lasers, and satellite radar have recorded thousands of waves higher than twice the significant height. A review of buoy records from around the world finds that they occur at roughly the rate the corrected statistics predict, with some seas, especially crossing seas and seas over strong opposing currents such as the Agulhas Current off South Africa, producing more than their share. Ship design rules, which long assumed a maximum wave of about 15 m, have been revised, and the forecast models of several weather services now issue a rogue wave risk alongside the significant wave height.

The Draupner platform still stands, and the laser still measures. On a quiet day the record shows waves of a metre or two passing beneath a steel deck that was once, for a few seconds, below the top of the sea.

Unit 6 of the Ocean Masterclass derives the dispersion relation of water waves, defines the significant wave height and the Rayleigh statistics used above, and works through the Draupner probability as one of its examples, before going on to tsunamis and the tides. The full list of units is in the curriculum.

Study the ocean properly

The Ocean Masterclass: weekly units, quizzes, a final exam, and a certificate.
About the program