Wave predition: progress and applications

TECHNIScHE UNIVERSITErr

[.1 - L6

Mjug 2,

OtI-I

cOIl.7mS

4 WAVE

PREDIcTION: PROGS

MID APPLICATIONS

(4

t2v

Wave prediction is of interest to 1SSC as a means of translating wind, data into wave statistics for estimating loading distributions and to assist retrospective diagnosis of marine accidents. As in the

previous chapter, the question of spectral shape is again

important

'but its significance is now more concerned with un&erstnrng of the phyiscal processes of wave generation 'by wind and particularly of the nature of the Source Function to be used. in wave prediction models,

Outstanding

advances

in such

understanding have recently been achieved

'by the Joint North Sea Wave Project

(Jo1swAP) and

reference has

already

been made to the so

called. Jonswap

spectrum from the point of view of

its influence on wave loading estimates-. The present chapter contributed

by Mr Ewing will begin with a brief statement about the JOESW.PP

project

and associated spectral formulation 'with some comments about its physical validity and general implications for structural design before reviewing

progress - in

development of wave prediction models.1

The Joint North Sea Wave Project (JoNSW&P)

The Joint NorthSea Wave Project originated in 1967 as a co-operative experiment on waves between a number of institutes in the Netherlands,

West Germany, the United States and the United Kingdom.

The two

basic aims of the study were to measure the growth of waves under

limited fetch conditions and. the attenuation of waves coming into

shallow water. The full description of the results from the first

experiments in 1968 and 1969 is contained. in Hasseimarin et al. (4.1) but summaries of the findings of JONSWP of special interest to engineers

have also been given

by Barnett(4.2) and Ewing (4.3).

The principal aim in studying wave spectra under limited fetch conditions in the German Bight was to determine the form of the Source Function of. the enerr balanàe for the wave spectrum during conditions

oI' wave growth. Knowledge of the Source FunctIon is an essential

requirement in numerical -wave m9dels (See later section). it was found.

'that the main featurez cf the observed Source Fnnc'tion could be

accounted for by nonlinear wave-wave interactions. The enerr transfer due to these interactions was found to control the overall balance

-within the wave spectrum to the extent that measured wave spectra -

with

narrow peaks, "overshoot" factors and fetch-dependent high-frequency

tails -

could all be explained in terms of this nonlinear transfer process.

JONSWAP Spectrum

Fetch-limited. wave spectra measured. during JONSWAP were found to be well represented by the function

E(f)

=

Jj_

(j;)_4]Y

eXP[_ (_fm)]

0.-o'bforf)fm

.

-which contains the five parameters f1 a,

'' a

and shown in Figure (4.1).

/4'

P4-&

o,.

Co

i-'

-te -e.

2. /

ci-1

-,-_/

-'fc&cn.

m is the frequency of the maximum of the spectrum and. a the. parameter corresponding to the 'constant' in the Phillips equilibrium

range. _{, a} and °b describe the shape of the spectrum with r defined

as the ratio of the maximum of the enerr speatrum to the maximun of the corresponding Pierson-4/Ioskowitz(4.4) wave spectrum

EPM(f) =

ag2 ()-4 f5

exp

[-with the same values of a and Essentially the JOI'SWAP spectrum is obtained by multiplying EPM by the "peak enhancement"

factor:-yexP

[- (f - fm)21

2 2

2°m

-The scale parameters a and. were found to follow power laws of the

form

'

0.22

a =0.076x

=

0.33

2

where the dimensionless fetch =

_gx/rJ

and _{m = m} U/g. (u is the wind speed at 10 m and x is the fetch). At small fetchesis several

times larger than the PiersonMoskowitz value of 0.0081.

No definite trend with fetch was found. for the shape parameters (, 0a' 0b and average values of '( = 33 0a = 0.07, °b = 0.09 were used to

define a 'mean JONSWAP' spectrum.

Essentially the JONSWAP spectrum represents a generalization of

the PiersonMoskowitz form through the inclusion of fetch as an additional parameter to the wind speed.

Spectral forms very similar to those measured in JONSW.AP have since

been observed by Rye et al. (4.5) and Saetre (4. for high wave conditions

in the North Sea. The spectra analysed by Rye et al. had Y values in

the range 1 to 3 while Saetre obtained. higher values with an average

I

of 4.2. Houmb et al. (4.7) also find spectra with characteristic J0I\TSWAP

features from measurements made off the coast of northern Norway. Other storm wave spectra measurements made in the North Atlantic Ocean are of J01SWAP form due to the generating winds being of finite extent. There seems no doubt that the J0SWP1P spectrum is a better description of. sea conditions in the North Sea and other fetchlimited situations than the PiersonMoskowitz spectrum.

For ISSC purposes an expression which can be entered with wave

height and period as well as wind speed and fetch is needed. The following is an explanation of how this requirement can be met by a recasting of the formula, based on an unpublished note by Ewing.

As in the derivation of the existing ISSC spectrum visual estimates

of height

Hv and period T

are defined by H significant

height = and T average period = me/mi where m0 and _mi are spectral moments of order 0 and 1 respectively). The moments of the

JONSWAP spectrum cannot be determined analy-tically.

Values of mo and

mi have however been calculated numerically for a range of y values

with 5a = 0.07 and b = 0.09. Integration was performed over the finite range f = 0.5 m to 10 _m using the trapezoidal ru.le together with an extrapolation method. (Corresponding calculations for the P-M spectrum showed an error of less than 0.05% in comparison with the known theoretical results). The following results were obtained

* When '= 1 JONSWAP is identical to P-M.

It may be seen that the variation of F2 is small; there is in fact less than 11% difference in the mean periods for the 2 spectra which can be neglected for most practical applications. It may hence be shown

that a J0I']SWAP spectrum Ej-(f) with the same Hv and. approximately the

same Tv as a given P-M spectrum can be defined by

1 (1.296f T

1)23

E(f) _EPM(f).FlrL2

V

where EPM(f) = 0.11

T (Tf5 exp [-o.44(Tf14]ni2/Hz

= 0.07 for f _(1.296

TY!

a=

0.09 for f ) (1.296:T1

Taking ₌ 3.3 this becomes

E3(f) = 0.072 H2T (Tf5 expç-0.44 (Tfi4]

33exp

(1.296r T - 1)2Jm2/Hz.

Implications for Structural Design

The J0NWAP spectrum contains more enerr than the corresponding

-Pierson-Moskowitz spectrum for the same values of cc and f. Ship

structural responses in fetch limited situations or in open ocean conditions where the waves are being actively generated by local winds will therefore

be under-estimated if the Pierson-Moskowitz- spectrum with the same values of '&

and f is used. The difference between calculated responses using the JONSWAP

F1 JONSWAP (m1/m0) JONSWAP F2 = = in0 P-M (m1/m0) P-M

1*

1.0 1.0 2 _{1.24 0.95} 3 1.46 0.93 3.3 1.52 0.92 4 1.66 _0.91 5 1.86 0.90 6 _2.04

089

and PiersonMoskowitz spectrum will be accentuated when the ship transfer function is sharply peaked.

Wave excited vibration and. other phenomena which depend on the ener&y present in the highfrequency tail of the wave spectrum.will also be under-estimated when the PiersonMoskowitz spectrum is used, since ) values in fetchlimited situations are several times larger than those for fully developed. conditions.

Wave group phenomena (Goda(4.8), Ewing

_(4.9))

will be more pronouned.

for the sharply peaked JONSWAP spec.truzn. This effect is of importance

for the forces on moored structures (Hsu and Blenkarn(4.1O), Remery and Hermans (4.11).

Progress on Wave Prediction Models

Earlier reviews of the methods and applications of wave prediction methods have been presented in reports to the 4th and 5th ISSC. Although

considerable progress has been made in both these aspects. further advances in wave prediction depend. on improved specifications of the surface wind field and on knowledge of wave processes to be modelled in the Source Function. (The requirements for predictions of integrated properties of the wave spectrum such as the significant wave height and average period are not so critically dependent either on the aOcuracy of the wind. field or on the basic knowledge of the underlying physical processes of wave generation and dissipation). A recent review by

Cardone (4.12) stresses these points especially in regard to spectral wave predictions but points to the improvement in the specification of the wind. field which will arise from data obtained. from remote sensing measurements from satellites. It seems possible that within the next two decades wave models will 'be able to have as input accurate wind data on a global scale for use in both real time wave forecasting and in wave hindcast studies. In parallel with these developments, improved understanding of the Source Function to be used in wave models is

being obtained from the theoretical and experimental studies of waves being carried out in many countries.

In this section we discuss spectral prediction models which have been developed since our last ISSC report.

(a) Deep water models

Lazanoff et al. (4.13)have extended the use of the New York University model to the Mediterranean Sea with the cooperation,of the U.S. Naval Oceanographic Office and Fleet Numerical Weather Centre. The model is used for real time operations. In Japan, Isozaki and Uji (4.14) have

developed a numerical model which includes terms in the So,u.rce Function accounting for linear and expone,ntial wave growth, wave breaking and dissipation due to opposing winds. Wavewave interactions, as considered

by Barneti (4.15) and Ewing (4.16),are excluded. The model has been

tested against wave measurements in the North Atlantic Ocean and Japan Sea. Isozaki and Uji stress that a mesh size much smaller than 150 kin will be required in the consideration of typhoon winds.

1.1- 20

(b) Shallow water models

Increased attention is being giver to the development of wave

models for shallow seas like the Gulf of Mexico and North Sea. The develop-ment of accurate wave models for shallow seas depends on a correct

representation of the dissipative processes acting on waves in shallow water together with the consideration of refraction.

Collins (4.17)has constructed a finite-difference numerical wave

model which includes both wave generation and bottom dissipation terms in the Source Function. The model is capable of treating parallel and irregular bottom topography and results have been compared with measure-ments taken during two hurricanes. Barnett et ai.(4.18) have shown how waves in the South China Sea can be predicted with a model where the

wave field is q-uantized in a set of discrete ener packets which are

propagated along the wave ray-paths. The Source Function is identical

to that used in (4.15) but with an additional term for bottom friction.

In Holland, J.W. Sanders (in press) has developed a model with the particular application offorecasts for certain coastal stations in the

southern. North Sea. The model is an extension of the method of Haug (4.19)

with the inclusion of dissipation.by bottom friction but without, as yet, incorporating the influence of wave refraction.

A group of scientists, mainly drawn from JONSWAP, have jointly proposed a wave model for the North Sea which aims to include all

important terms in the Source Function including dissipation due to wave

breaking (4.2D), bottom friction and wave refraction. This proposal,

known as NORSWM (North Sea Wave Model), will be initially concerned with hindoasting wave spectral characteristics for the most severe storms

since

1965

but it is hoped the model will, in due course, also be

used in real time forecasting work. Part of the NORSWAM proposal will also consider a more simple parametrical wave model developed by Hasselmann et al. (4.21)which should have computational advantages over the "classical" approach using a Source Function.

Accuracy

The reliability and accuracy of wave model predictions is required. to be known in all applications and studies of their use. As noted already the two main factors governing accuracy are the specification of the wind field and the knowledge of wave processes to be represented in the Source Function.

We first consider the main source of comparisons available from hindcast studies. Feidhausen et al(4.22)have compared hindcasts for the

severe storm of December

1959

from a number of investigators. (The same basic weather analysis was used although individual investigators were permitted a degree of subjectivity in determining the wind field from

the pressure pattern and isolated ship reports). Of the eleven compari-sons which were made those by

Barnett(4.15),

Bretschneider (4.23), Inoue (4.al.)

and Wilson e-t al.

_(4.25)

came closest to a high correlation between

hindcast and measured significant wave height. (The average value of the correlation coefficient was 0.85). Feldhausen et al. note the

inability of all hind.casts to predict the low wave heights recorded

1.1' - 21.

The standard deviation of liindcast wave heights compared to

measure-ments is given in Tab]e(4.1)for a number of wave models. There are

'significant differences in the accuracy of the models with the physically more realistic models givingthe best results with a standard deviation in significant wave height of about 1.0 m. (The standard deviation tends to increase with increasing wave height, as would be expected). Comparisons of hindcasted (one-dimensional) _{wave spectra with measured} wave spectra show considerable variability in all models due mainly to

the' representation of the te±ms in the Source Function. _{Wave models}

which attempt to describe all relevant processes including nonlinear wave-wave interactions usually give the best spectral comparisons.

In the case of realtime wave forecasting the accuracy of the wind field is very important. _{Bunting.30) obtained a standard deviation of} 1.7 m for forecasts of significant wave height _{up to 36hours ahead}

at Argu.s Island using the model of Pierson et al (4.27). The reliability

of North Sea forecasts has been studied by Graham _{and Bell (4.31) using} a "ignificant'wave" mode].. _{According to.their analysis wind forecasts} more than 2 days ahead are not, at present, reliable enough for wave forecasts beyond this period.

Applications

The two main areas for applications of _{wave, models are in}

fore-casting and hindfore-casting. '

Forecasts of wave' spectra are very sensitive to errors in wind speed and for this reason real time forecasts generally use wave models

based on the "sigrifi'cant wave" method. _{Although these models are}

less costly and faster to use on a computer than the spectral models discussed previously they are limited in their ability to predict swell accurately and, of course, they can not forecast details of the spectrum. Nevertheless real time wave forecasts are being used with advantage

in ship routing and other offshore operations. _{Real time wave} fore-casting is a task best carried out by the meteorological _{agencies. For}

example, Ogden (4.32) and Houmb_{(4.29) present the methods used in the}

United Kingdom and Norway respectively. _{North Atlantic forecasts for} ship routing purposes are available through most national _{meteorological}

agencies. _{Private meteorological consultants are also able to offer}

a service _{but their methods and results are not generally published.} Hindcast wave data finds its application mainly in ship design considerations. Generally there are three uses; The first

is in defining the year-round conditions for a given sea area which can be used to investigate the general behaviour of a ship or structure in waves especially in regard to fatigue problems. _{The second use is} in estimating the characteristics of the 25 _{or 50.-year extreme sea} con-dition by extrapolating the results from the analysis of the most

severe storms over a number of years. _{The use of hincasted wave data}

for estimating long-term extreme statistics has been discussed by Ward _(4.33) and. Hsu

_(4.34)

_{for the Gulf of Mexico, Houmb et al. (4,) for the Norwegiah} Sea and in the NORSWAM proposal.

A _{third application of hindcasted wave data is in the consideration} of ship casualty analysis where a loss may be caused by local sea

1.1 - 22

of currents on waves (Mallory(4.37)). _{Under such circumstances it is}

very unlikely that any wave measurements will have been made at the location of the casualty so that the only way of describing conditions in such areas is by using a hindcast wave model (see, for example, Ewing (4.38).

REFERENCEZ

4.1 K. .Hasselmann and others, Measurements of windwave growth and swell decay during the Joint North Sea Wave Project (JONSW.AP), Dt. hydrogr. Z, A8, 12, 1(1973).

4.2 T. P. Barnett, Observation of wind wave generation and

dissipation in the NorthSea : implication from the offshore industry, Offshore Technolor Conference, Houston, GPC 1516 (1972).

4.3 J. A. Ewing, Some results from the Joint North Sea Wave Project of interest to engineers, mt. Symposium on Dynamics of

marine vehicles and structures in waves, London, (1974). 4.4 W. J. Pierson and L. Moskowitz, A proposed spectral form for

fully developed wind seas based on the similarity theory of

S. A. Ki-taigorod.skii, J. Geophys. Res, 5181, (1964).

4.5 H. Rye, R. C. Byrd and A. Torum, Sharply peaked wave enerr

spectra in the North Sea, Offshore Technolor Conference, Houston

OTC 2107, (1974).

-4.6 H. J. Saetre, On high wave conditions in the northern North Sea, Inst. of Ocean. Sciences, Rept. No. 3, (1974).

4.7 0. G. Houmb and H. Rye, Analyses of wave data from the

Norwegian continental shelf. Second Conf. on Port and Ocean Engin. under Lctic conditions. Reykjavik, (1973).

4.8 Y. Goda, Numerical experiments on wave statistics with spectral simulaticn, Rep. Port Harbour Res. Inst., Mm. Transport,

Yokosuka, Japan, , 3, (1970).

4.9 J. A. Ewing, Mean length of runs of high waves, J. Geophys.

Res., , 1933, (1973).

4.10 F. H. Hsu and K. A. Blenkarn, Analysis of peak mooring forces caused by slow vessel drift oscillations in random seas, Offshore Technolo Conference, Houston, YI'C 1159, (1970).

4.11 G. F. M. Remery and A. J. Hermans, The slow drift oscillatios of a moored object in random seas, Offshore Technolor Conference, Houston, OTC 1500, (1971).

4.12 V. J. Cardone, Ocean wave prediction: Two decades of progress and future prospects, Soc. Nay. Arch. Mar. Engrs., "Seakeeping

1.1 - 23

4.13

S. M. Lazanoff, N. M. Stevenson and V. J. Cardone, A Mediterranean Sea wave spectral model, Fifth Offshore Technolor Conference, Paper

crc 1831, (1973).

4.14

I. Isozaki and T. Uji, Numerical prediction of ocean wind waves, Papers Met, and Geophys., Tokyo, ,

207, (1973).

4.15

T. P. Barnett, On the generation, dissipation and prediction of wind waves, J. Geophys. Res.,

ll

_{513, (1968).}

4.16

J. A. Ewing, A numerical wave prediction method for the North

Atlantic Ocean, Dt.. hydrogr. Z., ,

241, (1971).

4.17

J. I. Collins, Prediction of shallow-water spectra, J. Geophys.

Res.,

fl, 2693, (1972).

4.18

T. P. Barnett, C. H. Holland and P. Yager,A general technique for wind-wave prediction, with application to the South China Sea, Westinghouse Elec. Corp. Rport,

(1969).

4.19

0. Haug, A numerical model for prediction of sea and. swell, Meteor. Annaler, Oslo, ,

139, (1968).

4.20

K. Hasselmann, On the spectral dissipation of ocean waves due to white-capping, Boundary-Layer Met.,

_{6, 107, (1974).}

4.21

K. Hasselmann, D. B. Ross, P. Muller and W. Sell, A parametrical wave prediction model (in press).

4.22

P. H. Feldhausen, S. K. Chaicrabarti and B. W. Wilson, Comparison of wave hindcasts of weather station "J" for the North Atlantic Storm of December,

_1959,

Dt. hydrogr. Z.,

_26,

10,

(1973).

4.23

C. La Bretschneid.er, Significant wave-hindcasts for station J North Atlantic Storm, Tech. Rep. Nat. Eng. Sci. Comp. No.

SN-77-1, 46 pp, (1963).

4.24

T. Inoue, On the growth of the spectrum -of a wind generated sea according to a modified Miles-Phillips mechanism and its application to wave forecasting, Geophysical Sciences Lab. Rep.

T-67-5,

New York University,

_(1967).

4.25

B. W. Wilson, S. K. Chakrabarti and P. H. Feldhausen, Numerical prediction of storm waves and wave spectral densities, In:

Fifteenth mt. Conf. on Great Lakes Research, Madison,

Wisconsin,

_(1972).

4.26

J. Darbyshire and 3. H. Simpson, Numerical prediction of' wave spectra over the North Atlantic, DL hydrogr. Z.,

20, 18, (1967).

4.27

W. J. Pierson L. 3. Tick and L. Baer Computer based procedures for preparing global wave forecasts and wind field analysis

capable of using data obtained from a spacecraft Sixth

Li - 24

4. 28

C. L. Bretschneider, H. L. Crutcher, _{3. Darbyshire et al}

_,

Data for high wave conditions _{observed by OWS "Weather Reporter"} in December

_1959,

_{Dt. hyth'ogr. Z.,}

,

243, (1962).

4. 29

0. G. Houmb, Hindcas-Ling wind, _{sea and swell in the North} Sea, From "The Decade ahead _{1970-1980", Mar. Tech. Soc.,}

_281,

(1969).

4.30

_{D. C. Bunting, Evaluating forecasts of oceanwave spectra,}

J. Geophys. Res.,

j, 4131, (1970).

4.31

_{C. Graham and A. 0. Bell, North Sea weather forecasts, Roy.}

Inst. Nay. Arch. Symposium _{on Ocean Engineering,}

(1974).

4.32

R. J. Ogden, Forecasting for North _{Sea..oil and gas rigs,}

Weather,

27, 336, (1972).

4.33

_{E. G. Ward, Gulf of Me:ico meteorolor and oceanography data,}

mt. Symp. on Wave Measurement and _{Analysis, New Orleans,}

(1974).

4.34

F. H. Hsu, Hindcast storm waves for compilation of wave statistics, Soc. Petroleum Engrs. of AIME, _{Paper SPE}

_4323,

(1973).

4.35

_{0. G. Houmb, B. Pedersen and P. Steinbakke, Norwegian wave}

climate study, mt. Symp. _{on Wave Measurement and Analysis,}

New Orleans,

(1974).

4.36

_{W. J. Pierson, The loss of two British trawlers - a study in}

wave refraction, J. of Navigation, ,

291, (1972).

4.37

J. K. Mallory, Abnormal waves on the south east coast of South Africa, mt. Hydrogr. Rev.,

j,

99, (1974).

4.38

_{J. A. Ewing, Environmental. Conditions relevant, to the stability}

of ships in waves, mt. Conf. on Stability of Ships and Ocean Vehicles, Glasgow,

(1975).

0.7

1.1 25

Table 4.1 Statistical Analysis of Hindoasted Significant Wave Heights, Hs from Thimerical

Wave Models Compared to Measurements

Hz

Esduben f .p.um with Iu foe offuho,, winds - ll. Sect t5. 19dB). hnrie refu, to utdsono (of. F. t2

The bes*-ds wiaijiicwi apsa (SAl) we .lso showit The iwid iBaBsc the Osheidon of due Ave trio perwrsre m Ion. (2.4 I)

Figure 4.1

Parameters of JONSWAP Spectrum

and Evolution with Fetch (4.1)

Standard devnatiom

Wave model - Bate end area of computed Hn

from measurement

Rang. of H

(a) Source of data if differentfrom fir.t colmem

Narnett (4.15) 1.1 Feidhaunen ci al. (4.22)

Boetochneider (4.23) 1.0

Derbyshire and Simpoon(4.26)

lnoue (4.24) Pioreon et ci. (4.27) Dec. 1959 StOre (Bretachneider et al. -1.3 1.4 1.0 3-14 m Paidhaucen et ci. Wilson et al. (4.25)

Icozaid and Uji (4.14) North Atlantic 0.91.0 Peldhnnnen et ci. (4.22)

Eng (4.16) Nov. 1966, June 1967.

North Atlantic 0.6 1-6 a

Bang (4.19) Nov. 1966, Oct. and Dec.

1.1 - 72

12 CONCLUSIOI AND RECON11ENDATIONS

This chapter recapitulates the main conclusions of the report and

summarises the recouirnend.ations to which these lea&. Chapter 2 confirms

that the material in this and previous reports of this committee is

relevant to the reui'ements of structural designers and that the primary need continues to be for reliable wave data for use in predicting long term distributions of wave induced bending stresses. It recommends however that in future more attention should be paid to reviewing and

encouraging work on wave recording techniques and data collection in the context of sea trials and normal service voyages, by this. and related

committees. It is suggested in this connection that for the next ISSC

the words 'full scale statistical' should again appear in the title of a committee as in the case of the 3çd ISSC.

This recommendation is reinforced by the conclusions of, chapter.3 which is concerned with choice of spectra for predicting long term loading

distributions. It is found that the reliability of predictions generally

depends more on the'validity of the basic data fed into the spectra

than on theprecise details of spectral shape. It is noted moreover thai the criterion of reliability should be based on comparison with actual

service data for long term loading distributions.

Both chapter 3 and chapter

₄

consider the question of whether the previous recommendations regarding spectra in our report to the. 4th ISSC

should be revised with particular reference to the implications of

experience from the JONW.AP project. The best advice we can give at this stage may be summarised as follows.

a) Investigation should be made of long term distributions of bending stress computed using the existing ISSC spectrum formulation in cornparison.with corresponding results using the JONSWAP spectrum. Similar comparisons for loading of various types of offshore structure should also be made.

Until results from a) above are aVailable and have.been evaluated, it is recommended that the existing ISSC formulation as defined, in our report to the 4th ISSC should in general continue to be:used.

for ship structural loading calculations. Care should be taken however in choice of basic wave statistics used to feed the spectra since

these have a dominant effect on the resuIs..

'For the particular case of calculations involving fetch limited situations such asforexample design. of structures for operatipnin the North Sea, it would seem reasonable to use the JONSWAP spectrum as defined in chapter

₄

of this report. It contains a fetch

parameter and. it is already supported by a substantial amount of evidence to indicate that it is preferable on the grounds of its physical validity as a model of the wave condit.ions.

A form of the JONSWAP spectrum which can be entered 'by wave height

H and period Tv has been derived with ISSO requirements specifically in view, by Ewing, as described in chapter

4.

For a mean JONSW.AP spectrum with peakedness parameter '1= 3.3. this may be written as :

1.1 - 73

E (f) = 0.072

x3.3 exp [_ (1.296f T

-whereo=

0.07

for f'(1.296Ti1

= 0.09 for f ) (1 296

T)1

Directional spreading may of course be applied to this spectrum and recommendations regarding choice of spreading function may be found in our report to the 4th ISSC.

It is hoped that this q-uestion will be extensively discussed during the 6th ISSC in Boston and. readers are strongly urged to study the written reports of these discusion and our reply in the proceedings,

in which we will have the opportunity to make more up to date

recommendations. .

Chapter 4 alto discusses the topic of wave prediction and. draws

attention to its importance for.ISSC.both in deriving long term statistics for cases where direct wave data are not available over long periods

but wind data are, and for retrospective diagnosis of ship structural

damage. ... . .

exp [-0.44

(Tf)4]

)2].

Chapter 5 is concerned with rerements for offshore structure

design. _{This is a very large subject not previously discussed by ISSC}

and it has only been possible inthis report to direct attention to some of the extensive sOurces Of further information in the literature and consider very briefly some of the more important points. _{The concept} of.the 'Design Wave' expressed in terms of nonlinear wave theoryis widely used by offshore structure designers. It is indicated in

chapter 5 however thet at least in deep water wave loads computed using

high order nonlinear theory differ very little from those calculated from linear theory. Also it is noted that there is a growing trend towards specifying use of design spectra for calculating wave loads. This is particularly important in areas such as the North Sea where, due

-to high incidence of rough weather, fatigue rather than extreme _{waves may cause}

failure. _{Other points to which attention is drawn include the need, to}

take account of the influence of the structure itself on the wave especially in the case of large volume gravitystructures which can cause substantial local increase in wave height. Mention is also made of the importance of high local loads due to wave breaking, and the need for more data.on this subject

Chapters 6 and. ₇ are concerned with the acquisition and availability

of reliable wave data. Chapter 6 reviews progress in wave recording. It concludes that for the next few years buoy network development offers the best prospect for accumulation of reliable instrumental data. It is unlikely however ever to provide data covering world shipping routes on a climatological scale as required by ISSC and for this purpose.visual

statistics will still be impor-tait for some time to cOme. In the long run the most promising possibility for accumulating _{eliable measured data} on a global climatological scale is offered _{by. satellite based systems.,} These are already in an advanced stage of development but assessment of

their practical value for ISSC must wait _{until validated working data} is actually derived from an operational system such as the SEASAT A to be launched in

1978.

Chapter _{7 reviews progress in organisational aspects of wave data}

availability. Attention is drawn to the rapid expansion in instrumental

wave climate activity. _{It .is noted that there is a need for an international} body to coordinate this activity and provide an institutional framework

for instrumental wave data management on a global basis. An account is given of progress in encouraging the National Oceanographic Data Centres of the Intergovernmental Oceanographic Commission to play. an increasing role in this area.

Chapter

8

concerned with winds and gusts finds relatively little new material to report. _{Its most significant contents are reference to a} recent American publication giving extensive information on tropical cyclones for all six relevant ocean basins and a list of climatological summaries published by Member Countries of the World Meteorological Organisation.

.Chapter ₉ discusses abnormal waves and notes that there are certain areas where due to special local factors such as refraction due to currents or shoaling, these occur more frequently than elsewhere. Attention is concentrated. on 2 such areas, namely the east coast of South Africa where the Agulhas current can induce very dangerous conditions which have already caused much serious damage to shipping and the Dogger Bank where the

shoaling water can cause very treacherOus steep crossing seas.. 1t is pointed. out moreover that there are many other such areas

and it is recommended that further investigations are needed to understand the physical factors better so that the high risk areas and the conditions

causing danger can be more readily identified and avoided. From the ISSC point of view it is also important to investigate the effects of

these special areas on long term distributiOns of wave loading.

Chapter 10 disoumses the dangers of ice accretion and. ways of avoiding

them with attention to both design and. operational implications. _The

main risks are associated with stability but nonetheless affect _structural

design particularly because of the need. to allow for the _{effect of icing}

in inci'easing the weight and the height of the centre of gravity, and to minimise the degree of irregularity and protrusion which may promote

accretion. Also the adverse effect of very low temperatures on the fracture

properties of the steel must be taken into account. _{Regarding operational}

implications, the importance of good. forecasting is emphasised. _Reviews

of data and-jnformaton sources on icing cliratolor relevant to both design and operational requirements are given in this report and in our report tothe 5th ISSC.

Chapter 11 discusses fouling. It underlines the enormous cost. of damage and loss' of performance due to both fouling and corrosion noting that the 2 effects interact. _{The mechanisms of fouling and available data for} estimating its incidence are reviewed. Attention is drawn to its harmful effects in accelerating corrosion and in the case of offshore structures in increasing the wave loading. _{A .brief account is given of research on}

count ermeasure. mentioning some recent striking successes in, protective coating development. _{It is felt however that the current' research effort is} grossly inadequate for the magnitude' of the problem.''