Comparative Study of Performance of Wind Wave Model: Wavewatch—Modified By New Source Function

Abstract

ABSTRACT:With the aim of assessing the merits of the new source function proposed earlier in Polnikov (2005), it was tested and validated by means of the modification of the well known model WAVEWATCH-III. Assessment was done on the basis of comparing the wave simulation results found from both models against the buoy data obtained in three oceanic regions: eastern and western parts of the North Atlantic, and the Barenz Sea. First of all, incorporation of the new source function into the numerical codes of WAVEWATCH was done, and this modified version of WAVEWATCH was fitted and tested by standard tests. On the basis of these results some physical conclusions were drawn. Then, sophisticated fitting of the modified model was carried out, using the observation data of 19 buoys obtained in two regions of the North Atlantic for a period of 30 days at 1-hour time intervals. After this, the standard validation of both models was continued on the basis of the 8-month historical data, corresponding to 12 buoys located in the western part of the North Atlantic. Finally, comparative validation of the models was done against the wave observations obtained at one buoy in the Barenz Sea for a period of 3 years at 6-hour time intervals. Estimations of simulation accuracy were carried out for the three parameters of wind waves: significant wave height, H s , peak wave period, T p , and mean wave period, T m . Comparative analysis of these estimations was carried out for the original and modified model WAVEWATCH. The advantage of the modified model was revealed, with an increase of simulation accuracy for H s by 20 to 50% for more than 70% of the buoys considered. Additionally, it was found that the speed of calculation was increased by 15%.

Spectral Description of the Dissipation Mechanism for Wind Waves. Eddy Viscosity Model

“…Herewith, the change of terms NL and IN does not practically change the physics of SF, whilst the change of DIS, in opposite, does it radically. Variation of the models modification, realized by means of a consecutive and separate change of the SF-terms, permitted us to find that all principal differences in accuracy of wavefield calculations are caused by the change of the DIS-term [31].…”

“…Unfortunately, the results were published partially [28], whilst the main derivations were retained in a manuscript of the author's doctoral dissertation [29]. Many years later, the author has sophisticated the final phenomenological result [4] and verified them successfully by means of their implementation in the mathematical shells of the world-wide known models mentioned above: WAM [30,31]. Nevertheless, the turbulent viscosity model is not widely recognized yet, as far as its full and regular justification does not exist.…”

“…In fact, existence of this coefficient corresponds to the empirical data [16] that dissipation intensity at the spectrum tale is defined by the intensity of the dominant waves breaking. For example, our simulations show [30,31] that varying c σ allows, at some extent, to change the mean frequency, m σ , which is one of the integrated characteristics of the spectrum, given by the ratio…”

“…In all the mentioned papers, it was shown for a great data base that the change of DIS-term resulting in increasing the simulation accuracy for significant wave height s H on 15-20% and for mean period m T up to 50% (see in [31]. For more convincingness, we plot here the time history ( ) s H t (Fig.…”

“…First of all, we note that the only possibility to proof these preferences is the procedure of comparative verification. Such kind verification was performed in papers [30,31,34] where the models WAM and WW were accepted as the reference ones. The comparative verification procedure was performed for the integrated characteristics of wave field: the significant wave height H S , and the mean period T m , only.…”