Created on 8/5/2003 12:35
1
FREAK WAVE EVENT AT DRAUPNER JACKET JANUAR 1 1995
Sverre Haver, Statoil, PTT-KU-MA
Introductory remarks
As a response to a request for some info on this subject during the summer 2003, a brief
description of the wave event and the conditions January 1 1995 at the Draupner field is given
below. The description has been updated and modified several times during the last year.
The platform in the front (the platform with living quarter) is Draupner S, which was installed
in the mid-eighties, while the aft platform, Draupner E, was installed during the summer
1994. Draupner E is the first platform build with a bucket foundation and in order to verify
the bucket loading and bucket behaviour, the platform was instrumented for that purpose
during the first winter. The instrumentation involved a down looking laserbased wave sensor.
The wave sensor is said to be located at a platform corner, Haver and Karunakaran (1998).
Presently I do not remember which corner, but it is most likely the northern platform corner.
The well-known Draupner wave was measured at the wave sensor on the Draupner E
platform, i.e. the unmanned platform. It is seen that the platform is rather transparent, i.e.
wave measurements are not likely to be much disturbed by the platform legs.
Fig. 0 The Draupner S and Draupner E platforms in the North Sea,
Photo: Øyvind Hagen, Statoil.
Created on 8/5/2003 12:35
2
The Weather Situation January 1
The weather situation in the North Sea Januar 1 at 12UTC is indicated in Fig 1. The weather
pattern is dominated by a major low with center in Southern Sweden causing a strong
northerly wind field over the whole North Sea and Norwegian Sea. In addition to the major
low, a smaller low has been moving southwards in the North Sea during the morning hours
and in Fig. 1 it is indicated to be located close to Skagerak. The effect of this smaller low is to
strengthen the wind field in the western North Sea. As the smaller low moved southwards, the
area with the strongest wind field moved southwards as indicated in Fig. 2. The wind
conditions seemed to peak at hurricane level wind between 12UTC and 18UTC. The
Draupner platform is located more or less straight west of the southern tip of Norway midway
between Norway and Scotland. This means that the platform was outside the area where the
wind conditions seem to have been most extreme.
The England – Norway ferry “Color Viking” was hit by this storm event on its way over the
North Sea. In an interview the ship master told he had never experienced such a weather
during his years as a ship master of Color Line ferries. The bridge of this ferry is about 30m
above sea level when floating on even keel. In the same interview the ship master said that in
connection with the highest waves, those on the bridge looked upwards on the wave crests
approaching the vessel. Although it may be somewhat difficult relate observations to a
horizontal base line in this sort of weather conditions due to ship pitching and rolling, the
statement clearly suggests that a number of rather extreme wave crests seem to have occurred
in a certain area of the storm event.
One interesting aspect of the overall situation is the local enhancement of an underlying
strong wind field by the secondary low moving southwards. It may well be that this will result
a significant increase of sea state steepness over certain areas. In turn this may represent
conditions being favourable regarding non- linear wave phenomena.
A more detailed discussion of the weather pattern that particular day can be obtained by
contacting e.g. The Norwegian Meteorological Institute in Bergen.
Wave Conditions at Draupner
At Draupner the wave conditions started increasing late in morning. The significant wave
height reached 11 –12 m just after noon and stayed at that level until early evening, Haver and
Andersen (2000). For a significant wave height we will expect the maximum wave height
(height from trough to crest) to be about 20, but occasionally it may well be 10-15% higher.
The maximum crest height (height from still water level until wave crest) would be expected
to be about 15m. We have a number of 20- minute observations of the wave condition during
peak level of the storm. Most of these show nothing particular , see e.g. Fig. 3, which is rather
representative for most of the 20- min. time histories. (It should be noted that we merely have
one 20- min record every hour, i.e. no measurements are made for 2/3 of the time.)
The 20- min. wave history collected at about 15:20, however, did include a very extreme wave
event bearing in mind the sea state conditions. The time history is shown in Fig 4, and a closer
look at the extreme wave is given in Fig. 5. The maximum wave height is close to 26m. For
the actual area, the wave with an annual probability of 10
-2
(i.e. the so-called 100-year wave
height) was estimated to be around 27m. This means that the wave height in itself is not that
extreme, but we did not expect the wave height to be realized in a sea with a 12m significant
wave height. It is more likely to be experienced in a sea state with 13-14m significant wave
height.
Created on 8/5/2003 12:35
3
The wave spectrum for the time history shown in Fig. 4 is shown in Fig.6. Although the
estimated spectral ordinates are associated with a considerable statistical uncertainty, the
estimated spectrum clearly suggests the existence of two sea systems. This is in agreement
with the previous description of the weather scenario January 1 1995.
The extreme pattern of this wave event is its asymmetry about the still water level. The crest
height reaches about 18.5m above still water leve l. For the actual area, the crest height
occurring with an annual probability of 10
-4
(i.e. 10000-year crest height) was at that time
estimated to be 19.5m. This means that the crest event was not beyond design parameters, but
we did not expect such a wave to occur for a significant wave height of 12m. None of our
basic engineering approaches suggest the possibility of such a crest height for a 12m sea state.
We think the measured event is of good quality. Of course some uncertainties may be
associated with the crest height, but the uncertainties or possible biases cannot bring the wave
event into the class of typical (slightly non-Gaussian) wave events. The measured structural
loads do also suggest that a big wave hit the structure, but the loads is relatively speaking not
as bad as the observed crest height. Although the crest height is well beyond the 10
-2
probability crest height, the maximum platform loads do not exceed that level. This suggests
that the shape of the wave do differ from our typical design waves. As the wave passed the
structure, it was rather close to a temporary deck and as I remember being told some minor
damage was reported to equipment located on that deck.
To my knowledge, people onboard the platform did not particularly note the wave event as it
happened. But of course during those weather conditions nobody will be out on the open deck
area. It is said in the platform daily journal that all work on deck was stopped at 15:00 due to
very strong wind.
Freak waves
A number of people and institutions have been given access to the Draupner time history. I
think there is a general agreement that the wave event can be referred to as a freak event, i.e.
an event that would not be expected under our typical engineering models for extreme wave
predictions. I think there is competing ideas regarding explaining the event and not least
regarding modelling it mathematically, but my favourite theory involves self- focusing of
energy of a wave group into one majestic wave event, i.e. the big event steal energy from its
neighbours. The major challenge from my point of view is to understand when and why this
strongly non- linear self- focusing process is turned on.
References:
Sunde, Alv (1995): “Kjempebølger i Nordsjøen” (Extreme Waves in the North Sea), Paper is
written in Norwegian, Vær &Klima, Nr.1, 1995.
Haver, S. and Andersen, O. J. (2000): “ Freak waves – Rare Realizations of a Typical
Population or Typical Realizations of a Rare Population?”, ISOPE’2000, Seattle.
Haver, S. and Karunakaran, D. (1998): “Probabilistic description of crest heights of ocean
waves”, Proceedings 5
th
International Workshop of Wave Hindcasting and Forecasting,
Melbourne, Florida, January 1998.
Created on 8/5/2003 12:35
4
Fig. 1 Weather situation in the North Sea, Januar 1 at 12 UTC, Sunde (1995)
Created on 8/5/2003 12:35
5
Fig 2 Area with most severe wind during January 1, Sunde (1995).
Created on 8/5/2003 12:35
6
Wave Record at 16:20
-10,00
-6,00
-2,00
2,00
6,00
10,00
14,00
0
200
400
600
800
1000
1200
Time (s)
Wave elevation (m)
Fig. 3 A typical 20- min. time history of surface elevation during the storm.
Draupner wave record
January 1 1995 at 15:20, hs = 11.9m
-8
-4
0
4
8
12
16
20
0
200
400
600
800
1000
1200
Time (s)
Surface elevation (m)
Fig. 4 Time history including an possible freak wave event
Created on 8/5/2003 12:35
7
Fig. 5 A closer look at the possible freak wave event.
Fig. 6 Estimated wave spectrum for the 15:20 wave time history, Haver and Karunakaran
(1998)
Draupner wave record
January 1 1995 at 15:20, hs = 11.9m
-8
-4
0
4
8
12
16
20
180
200
220
240
260
280
300
320
340
360
380
Time (s)
Surface elevation (m)