Energy dissipation across the wave spectrum - Breaking and Dissipation of Ocean Surface Waves - page 276

breaking of swell. This was done within the SWAN model ( Booij et al. , 1999 ). Over-
all, however, most models still lack these new insights into the physics of wave breaking.
Since there appears to be some confusion in this regard, we would like to point out, that,
for example, the analytical model of Hasselmann ( 1974 )(see Section 7.1 ) is neither in
contrast nor in support of the above-mentioned threshold behaviour, but is unrelated to it.
It predicts the behaviour of the whitecapping dissipation provided that whitecaps already
exist and conduct the negative work on the forward face of the wave, i.e. what happens if
the waves are above the steepness/spectral threshold and are already breaking.
These observed features need to be accommodated in modern dissipation terms, other-
wise the models do not reflect the correct physics and do not adequately describe the reality.
It is fast becoming clear that, without incorporating these new features, the models cannot
properly forecast complex or non-standard circumstances. Ardhuin et al. ( 2007 )showed
two such situations: wave growth in the presence of swell and over slanting fetches.
The most apparent non-standard circumstances where failure of the standard-tuned terms
is to be expected are extreme and complex wind-wave conditions - which are also of
utmost interest from practical points of view ( Greenslade , 2001 ; Babanin et al. , 2010d ). As
discussed in Section 7.3.5 above, the dissipation function is altered under such conditions,
and so is the wind input. No amount of good tuning and statistical fitting, as opposed
to employing correct physics, will be able to extrapolate source terms, tuned to standard
conditions, into such extreme situations (see Cavaleri , 2009 , for the most recent update on
spectral-modelling limitations in this regard).
Babanin et al. ( 2010d ) specifically investigated the modelling capacity of modern third-
generation models by means of hindcasting the wave conditions in Typhoon Krosa prior to
landfall on Taiwan in October 2007. Two models were used, SWAN ( Booij et al. , 1999 ) and
WWM ( Hsu et al. , 2005 ; Roland , 2009 ). During this typhoon, the highest wind-generated
waves ever measured were recorded, with the trough-to-crest elevation for an individual
wave of

H max = 32 . 3m (7.34)

and the maximal significant wave height, measured over 10 min duration, of

H s max = 23 . 9m (7.35)

( Liu et al. , 2007 , 2009 ; Doong et al. , 2008 ).
After careful analysis, Babanin et al. ( 2010d ) decided that the measurement did not
appear faulty and is physically realistic. Numerous numerical tests were then performed in
order to reproduce the observed conditions, and it was concluded that neither SWAN nor
WWMII are able to hindcast the extreme observation. Thus, an urgent need for introduction
of new and adequate physics and numerics into wave-forecast models is apparent.
One of the first dedicated efforts to incorporate some of the new experimental observa-
tions (i.e. the breaking-threshold behaviour, see Section 5.2 ) into a nonlinear dissipation
function was conducted by Alves & Banner ( 2003 ). Babanin & van der Westhuysen ( 2008 )
later showed that the statement that the form of the Alves & Banner ( 2003 ) dissipation