Underestimation of Significant Wave Height (Hs) Using FVCOM-SWAVE

26 views

Skip to first unread message

Mostafa Beyramzadeh

unread,

Aug 28, 2026, 11:42:02 AM (3 days ago) Aug 28

to FVCOM User Forum

Hi Forum Members,

Over the past few months, I have been trying to simulate wave–current interaction using FVCOM-SWAVE. I decided to prescribe the bulk wave parameters (Hs, Tp, and mean wave direction) along the open boundary. Only a single computational domain was considered in this study, and no nested domain was used. The required bulk wave parameters were extracted from the WAVEWATCH III model.

A three-day warm-up period was considered for the model, starting from 2010-05-24 00:00:00. The JANSSEN formulation was applied for wind input as well as whitecapping dissipation. Due to the zero initial condition, AGROW, which represents linear growth during the early stage of wave development, was activated in the INPUT file. GEN3_JANS_CDS1 = 2.5 and GEN3_JANS_DELTA = 0.7 were used for the wave model.

The domain-wide Hs distributions from WAVEWATCH III and FVCOM-SWAVE are shown for 2010-05-28 10:00:00. It is clear that SWAVE underestimates Hs, particularly in the Persian Gulf, where the wave field is less affected by distant swells propagating from the Arabian Sea and the Indian Ocean.

It is noteworthy that the tuning parameters are identical in SWAVE and WAVEWATCH III. Therefore, we expected the Hs distributions produced by the two models to be relatively similar; however, the results do not show the expected agreement.

FVCOM version 5.0 was used for this simulation.

The results, along with the relevant configuration files, including INPUT, namelist, and make.inc, are attached to provide further details about the model setup

Has anyone encountered a similar issue or could provide some guidance on how to identify and resolve the cause of this Hs underestimation?

Any suggestions would be greatly appreciated.

Best regards,
Mostafa

make.inc

PGGO2010_run.nml

INPUT

Hs.png

Siqi Li

unread,

9:28 AM (1 minute ago) 9:28 AM

to FVCOM User Forum

Hi Mostafa,

Using the same Janssen tuning parameters in SWAVE and WAVEWATCH III does not necessarily guarantee similar Hs distributions, since the results can also be sensitive to the spectral boundary forcing, wind forcing, spectral discretization, and shallow-water dissipation.

One point I would suggest checking first is the open-boundary forcing. You mentioned that only bulk parameters (Hs, Tp, and mean wave direction) were prescribed. In this case, SWAVE needs to reconstruct the boundary spectrum from these parameters, and that reconstructed spectrum may differ substantially from the original WAVEWATCH III directional spectrum. Therefore, identical boundary Hs does not necessarily imply identical wave-energy propagation into the domain.

Since the largest underestimation appears in the Persian Gulf, where locally generated wind waves may be important, I would also compare the wind forcing used by SWAVE and WAVEWATCH III carefully, including wind magnitude, temporal resolution, interpolation, and land masking. It would also be useful to check whether bottom friction and other shallow-water dissipation settings are consistent between the two models.

I would also examine whether the three-day spin-up is sufficient by plotting Hs time series at several representative locations. AGROW helps initialize wave growth from a zero-wave condition, but it does not necessarily guarantee that the wave field has reached equilibrium after a prescribed period.

A useful diagnostic test may be to separate the contributions from boundary waves and locally generated waves—for example, running one case without wind input and another without incoming boundary waves. This could help determine whether the underestimation originates mainly from the boundary treatment or from the local source/sink terms.