ARCH I.EF
Nü'J. 1976NORWEGI METEOROLOGICAL SUPPORT
TO COASTAL ZONE ACTIVITY.
Lars H1and x)
S6:CS
The Norwegian
Meteorological Institute
Lab. v.
Sc1eepsb0UTechnische Hogeschoo1
NORWEGIAN METEOROLOGICAL SUPPORT TO COASTAL ZONE ACTIVITY.
1. Introduction.
When the planning of drilling operations in the Norwegian part of the North Sea started in 1965, requests for a number of new types of services were made to the Norwegian Meteorological
Institute. The requests may be grouped in two main categories:
1.1 _grecasting service which in addition to the usual
fore-casts for shipping and aviation also included forefore-casts of wave heights and periods. Sea as well as swell should be included in
the seaway forecast.
1.2 Data for o2erational and critical design criteria. Some of these data could be extracted from the climatological archives
present. Separate and carefully controlled magnetic tape files
were established for 15 selected lighthouses and the two weather ships available, Polarfront at 66°N 2°E and Fainita at 57°N 3E.
Fig. 1. The series from Polarfront and the l.ighthouses
cover the
period from l99. Famita has been in operation since 1959.
From the files adequate longterm Statistics of different climatological parameters such as temperatures, humidity, cloud and visibility or combinations of these, could be established. Re-garding wind and waves, the two most important elements, however, the standard meteorological observation proved insufficient. Avail-able were 10 minutes mean of wind at 10 mtrs. height, state of sea subjectively observed according to Table 1 at the lighthouses and visually observed wave height and periods at the weather ships. Requested were longterm statistics of wind at different heights, of gustiness and of corresponding wave heights and periods. The most
advanced design criteria required power spectrum of wind and waves. Further, the data should be extrapolated to cover the en-tire continental shelf. Particularely the extrapolation of the wave data represented a difficult problem.
Recently also forecast and hindcast of likely drift of oil spill have been requested.
1.3 A research program was established in order to solve the
problems. The program included:
A model computing wave data from surface pressure maps. The model should be used for forecasting as well as
hind-casting.
A project based on the wave model and historical
weather maps. By hindcast techniques the seaway data from
lighthouses and weather ships should be transformed to ade-quate wave data, and these data should be extrapolated to other areas of the continental shelf.
A new data acquisition system including data for power spectrum analysis of wind and waves.
A model for forecast and hindcast of oil spill drift.
The wave model.
.1 The basic concept of the model /1/ is the assumptions that :he winddriven sea at any time is uniquely determined by the Pier-;ori-Moskowitz energy spectrum a cosine square directional energy
Listrjbutjon, and that the velocity of energy propagation is the ;roup velocity of gravity waves of given frequencies.
The calculation of wave energy are performed for the inter-ection points of a regular square grid with a mesh-width of 150 km. he area covered is shown in Fig. 2.
The calculations are carried out in time-steps of 3 hrs, nd the major dynamic effects which cause energy to change in a
rid-point are propagation of energy, loss of energy due to change f wind direction in time and space, and transfer of energy from ir to sea. The values used for the growth rate of energy are
aken from the works of Neumann /2,3/.
The primary input of data is sea level pressure at all ridpoints for every 6 hrs. The wind components are computed in .ccordance with the physical effects which determined the
relation-hip between sea level pressure and wind, taking into consideration .eformation, curvature, friction and the stability of the air mass.
The calculations of sea energy are continuous from day to ay in the sense that initial values for energy in grid-points are he last computed values the previous day which are based upon
ob-erved pressure data. The calculations up to date are then per-ormed upon new Bets of observed input data, and are, for fore-asting purposes, continued for another 214 hrs based upon prog-ostic pressure data.
As a separate part of the program the swell conditions energy and direction) are calculated for selected locations based pon the computed winddriven sea energy for a period of 148 hrs revious to the actual time.
.2 The first version of the model was put in operation in 1967
n 1971 the model was reprogrammed. Since ist October 1972 the omputed wave energy together with the wind- and pressure-data
very 6 hrs at all grid-points have been permanently stored on
agnetic tape. Only wave data computed from observed pressure data ave been stored. For practical reasons one selected to store the
3
quantity S
jse
dso k 4.29
T period of wave component in seconds y = wind velocity in knots
From this quantity in combination with wind velocity all relevant wave data as significant wave height, wave periods etc. can be derived by simple calculation
/145/.
The permanent archives of pressure, wind- and wave-data are updated every day, and have already proved very useful when new methods are being tested.
2.3 Representativity of the computed wave data. In a series of cases the computed wave data have been compared with actual
obser-vations /6/. These comparisons have led to the conclusion that there are no apparent systematic errors in the way the model
developes sea energy. Small scale variations, however, seem to be somewhat smoothed due to the resolution of the grid.
Some examples of observed and computed significant wave height from extreme storms during the winter season 1973/1974 are
shown in Fig. 3-5.
2.4 The fiord model. The modele used for hindcast and forecast
of ocean waves with a resolution of 150 km between grid-points is not well suited to give a good description of the sea in waters
close to the coast. In order to study in greater detail the varia-tions in wave condivaria-tions due to the topography of a rugged coast, a special version has been developed where the area of integration
and the mesh-width can be chosen according to needs.
The "fine mesh" version is based upon the same concepts and numerical techniques as the routine version, but is for the reason of computational economy restricted to single input values of wind speed and direction which are considered constant in time
and espace over the integration period. The integration starts from
zero wave energy, and output of the developed wave conditions are given for every 6 hrs.
Trie model is not a shallow water model, but is prepared for extension to a version which takes into consideration the effects of depth upon the wave conditions in a simple way. The model has Leen u.ìed forcertain waters, inlets and fiords, apparently with reasonable results.
k_______
kTwheres
, SkT max
I. The hindcast project.
1.1 Reliable longterm statistics have to be based on rather .ong time series. As the wave model has proved, at least for sta-:istical purposes, to produce adequate wave data from surface
ressure maps, and as historical weather maps are present for a long eriod, it is possible to establish such statistics by hindcast.
1-ie preparation of input data from weather maps involves, however,
L considerable amount of manual work, and for practical reasons a
election of storms has to be made. If hindcasts are undertaken for
xtrenie storms only, likely maximum wave heights may be investigated,
'ut useful information for operational criteria which can be derived rom frequency distributions of wave heights, is lost. For this season the problems were approached in the following
manner.
.2 Since October 1972 the wave data computed by the model very day were present as well as observations of state of sea at she lighthouses and visually observed wave heights at the weather
hips. The distribution of
significant
wave height computed by theLodel in all cases
within
the same class interval of state of searas examined. The distribution proved to be a normal one of the
:ype:
(aV)1exp
202
a = standard deviation
d = the difference between the central value of the class interval and the mean of the normal distribution.
The two parameters, o and d, varied with the class interval, 'ut were similar at the weather ships and also at most of the light-ouses when all cases of offshore wind were removed from the data,
Using data for the test period only, the parameters could e determined in the class intervals of highest frequencies.
In
the igher intervaiB, S 7, 8, 9, however, the amount of data wasin-ufficient. hindcast of appr. 650 storms, i.e. all storms during
'hich a state of sea of at lea8t 7 has occurred anywhere on the
:oast, were undertaken. The distribution of the computed significant 'ave heights in each of the class intervals proved also a close fit
o normal distributions, and the parameters for all class intervals
:ould thus be determined,
Applying these parameters a set of matrisses could be
stablished by which the frequency distributions of visual observed 'ave height and frequency distributions of state of sea observations
ould be transformed to comparable and realistic frequency
distri-utions of
significant
wave height.The results are shown in Table 2. At the lighthouses the .istributions are valid for a point 20 N.M. off the coast. The sea. .eveloped by off shore winds have been calculated by the "fine mesh"
-5-model. The distribution for FAMITA is valid for the winter
months
October to March only.
3.3 The results of the investigation could be summarized
as
follows: Although high waves may occur rather far to the south in
the North Sea, the wave conditions bacome more severe moving north-ward tonorth-wards 62°N, i.e. into waters with higher storm frequensy and which are open not only to seaway coining from the Norwegian Sea but also from the Atlantic. From 62°N to 68°N the conditions are more or less similar, and north of 689N a general decrease of the wave
heights occurs.
In coastal waters variations due to local winds occur. The most pronounced of these effects is the increase of local winds
and
sea in areas where the general direction of the coast sharply
deflects. Lista near the Naze, KMkenes near Cape Stadt and
Fru-holmen near North Cape are examples.
3.4 The examination of the data from the hindcast investigation
is still is progress. The probable maximum wave height in 50 years
and 100 years return periods and wave periods for design
purposes
are still to be determined. The work is expected to be finished in
1977.
4. The Environmental Data Center.
The data filed in the center are of two different types.
4.1 On all locations where observations are made, an ordinary
SHIP observation is undertaken every third hour.
These data will be processed in accordance with WMO regulations and will also be trans-ferred to the Maritime Data Center in Bracknell. Selected observa-tions will also be transmitted in real time. Together with similar data from the weather ships and other ships in the area, the data form general reference files for temperature humudity, icing and
weather in general.
4.2 On selected fixed and floating platforms special
observa-tions will be carried out. The elements, sampling frequencies, resolution etc. are shown in Table 3. For the time being observa-tions at 4 sites are being planned, three in the North Sea and one at appr. 72N 23°E.
Observations of this kind require rather expensive instrumentation and data processing. New techniques have to be
tested. The first reliable data are expected during the winter season 1976-1977.
The expenses are distributed between
compánjes granted production licence in Norwegian. Waters,
the Norwegian Petroljum Directorate and the Norwegian Meteorological Institute,
5.. Drift of oil, spill.
A procedure for predicting drift of oil on the sea surface
has been developed.. The. input data
to thé model are surface wind components derived from the
surface pressure in a manner similar to the one used in the wave model.
The basic concept is the
assurnp-tions that the surface covered by oil is small compared with the ocean surface and that the wind factors and the surface current are known or can be estimated within reasonable ranges.
Fig. 6 shows trajectories
for anoil spill at 61915'N,
l55'E for the nine days following
23th March 1976. Trajectories
for different wind factors
are plotted as well as trajectories assuming no surface current and a likely northward bound coastal
current of 0,5 kts..
The model can also be used for hindcast.
purposes, and some preliminary studies have
been made of the probability of presence of oil in different locatjöns at different
time intervals after an
oil spill. The study is based upon simulated trajectories
for the
period October 1973 to March 1976.
W.ind factors derived from the drift of oi.l escaping the wreck of "Torrey
Canyont' have been used.
Fig. 7 shows the probability of.
presence of oil 9 days after an oil spill at 56°35'N, 3°15'E.,
The procedure can also be used for
prediction of drift of
,ther objects as life boats and rafts
provided reasonable wind actors can be estimated.
References.
/1/
1-laug, O.:
"A Numerical Model for prediction
of sea and swell"
Meteorologiske Annaler,. vol. 5,
no. 4, 1968
/2/
Neumann, G.:
"Ober die komplexe Natur
desSeeganges.
Teil lund 2'
Deutsche Hydrographisce
Zeitschrift., Band 5, 1952
/3/
Neumann, G.:
"Zur Charakteristik des Seeganges"
Arch. für Meteorol. Geophys. and Biochiimatologie,
Serie A 7 352-377, 1954
/4/
Longuet-Higgins, M.S.:
"On the Statistical Distribution
of Height of Sea Waves"
Journal of Maritime Research,
vol. XI, no. 3, 1952
/5/
Longuet-Higgins, M.S.:
"The Statistical Analysis
of a Radom, Mcving surface"
Phil. Trans. Roy. Soc., vol.
249 A, 966, 1957
/6/
Srnáland, E.:
."Verification of Numerical Wave
Computation"
P74.
Reconstruction of extreme
storms.
Preliminary ReportNo. 1, Norw.
Met. Inst., 1975
/7/
Hála.nd, L. and Srnáland, E.:
"On the Connection between
Observed and Computed Wave
Heights"
.P74.
Reconstruction of extreme storms.
Table 1.
S State of sea
Class
interval Descriptive terms Height in metres
O Calm (glassy) O i Calm (rippled) 0 0.1 2 Smooth (wavelets) 0.1 - 0.5 3 Slight 0.5 - 1.25 Moderate 1.25 - 2.5 5 Rough 2.5 - '4 6 Very rough '4 - 6 7 High 6
-9
8 Very high 9 -l'e
Table 2.
Percentage sea frequencies at the weather ship Polarfront, the rescue vessel Famita and from the 15 lighthouses 20 NM off the
coast.
1/3 (meter) 0-0.5 0.5-1.25 1.25-2.5 2.5-t4 'e-6 6-9 9-1k ) iLe
SEA STATE S 0+1+ 2 3 4 5 6 7 8 9 FLRDLR 22.4 26.8 38.2 9.3 2.8 .4 .02 0.000 LYNGØR 21.7 28.]. 36.9 10.]. 2.8 .4 .02 0.000 LiSTA 11.4 22.3 44.2 16.7 '4.6 .8 .014 0.001 UTSIRA 14.1 24.5 32.5 19.8 7.3 1.7 .16 0.004 HELLISØY 18.2 27.0 37.9 12.4 3.8 .7 .05 0.000 KRAKENES 9.1 19.6 40.8 19.3 8.3 2.4 .28 0.006 UNA 1'4.6 25.3 35.2 16.7 6.5 1.5 .15 0.003 SULA 18.8 24.2 36.0 13.3 6.0 1.5 .13 0.001 NORDØYAN 10.2 17.9 46.8 17.1 6.14 1.5 .12 0.001 MYKEN 14.9 23.1 39.0 14.8 6.2 1.7 .17 0.002 SK0MVIR 20.3 29.1 29.2 14.7 5.5 1.]. .08 0.001 ANDENES 23.3 28.4 34.2 10.1 3.3 .7 .06 0.001 TORSVAG. 12.9 23.2 '48.0 12.4 3.2 14 .03 0.000 FRUHOLM.EN 14.0 23.2 42.7 14.0 4.9 1.1 .14 0.002 VARDØ 27.1 29.8 33.9 7.4 1.6 .2 .01 0.000 POLARFRONT 10.9 20.2 27.6 23.9 13.3 3.6 .44 0.021 FAMITA 14.7 22.5 28.6 21.8 10.2 2.2 .22 0.009
Table 3.
Range and accuracy of instruments for recording of type E-data.
PARAMETER RANGE INSTRUMENT ACCURACY RESOLUTION REMARKS Wind speed o - 60 rn/sec _ 0.5 rn/sec up to 5 rn/sec ± 10 % above 5 rn/sec Wind direction O - 360 5°
Water (mud- line) depth
O - 600 ni
:o - 2000 ft)
Could be manually recorded
Wave height
O - 30 m :o - 105 ft)
+
0.2 m
Current speed 0.02 - 2.5 rn/sec
- 8 ft/sec)
+
3 - 5 %
-0.02 rn/sec minimum Could be selfcon- tamed instruments with their own recording system
Current direction
O - 360°
+ - 5'
Current depths O - 600 ni O - 2000 ft)
Could be manually recorded
Water level
±
p
ama
IIIP!II
rr
aQrif ivavAmar'
se i
Llr
4-Ly---
-.4
a
I-Íit
aaaii
-
n imí
4"!IJIIIO
-..i
gL1
mimaaE
L
ammauar-ammum a
i miii
ni.
r
-4 --41+iodeI" coast (ie.
-L-F.2
-T'
I
rrl
I j I I\H.
ni
r4\r+
-1
_Jl__\
Ir
I L...._... JS,9.wve (iei9kt (vneters)
12 11lo
2 9 g 71-
/ / / 5-, 4 3 / 09 o I f 00 00 06 2 8 00121.74
13.1.74 Co pt4ted--- -Obse.ryec froi OCEAN VIKING (Vi$i1)
Wiiid speed (kKofs)
00 06 12
1 00
Wind direcop (dekderees)
s s
00 06 12 18
36
27
Sii. wave !ieikt (eer)
/ / I / / / / I / / / / i 2/
/
'1g /r---Coipfed
- - - -
-
Oserved Çrow FAM ITA
(viuI)
Wid speed C kots)
so 40 -nA
'
LV s' fo s'/
.5,
cg 4 00 06 12 1 00 0 12 18 005.f2.73
OC? 06 12 ig 00 06 laWiid directo (dtkderees)
06 12 00 06 t2. / / 12 10 5 4 3 -2 o 00
wive ke9ht (iiieters)
12-
1f-lo -7.-
6-5.
4-3-. a-I/ // / O?-t---,
i 00 0 12 18 00 0 12 i 1&11.73 19.11.73 36 27 1 09Conputed
---
Observed froii FAM
ItA (vi6uI)
Wud sped(kHofs)
60 50 140 30 2a lo o / / J\
I / V /WiHd direcfio,
(Dekaderees)
P5
s . I I / / O 12 oc, 0G 12 00 0G IZ 18 00 06 18 00TrcjectorieS for drift of cv.
o( spd! ¡rovii 61°15't1
j°5'E or the iiie
ys towiZh. Mrci 1976
r
I Fi
FiS 'T
Probihyof
prese'tce of.oiI '9d