Modeling wave attenuation induced by the vertical density variations of vegetation

Abstract A phase-averaged model, SWAN, and a phase-resolving RANS-type numerical model, NHWAVE, were compared to previously reported physical model data to evaluate the effectiveness of these models in simulating the attenuation of irregular waves propagating over emergent vegetation with variations in stem heights. The physical model was conducted with two treatments of vegetation, one with a uniform stem height and the other with different heights approximating a linear distribution in stem density over the vertical. The number of stems in the second treatment was doubled so that the total projected area over the vertical was the same between the two tests. The drag coefficients C D used by SWAN-SL (single-layer in SWAN), SWAN-ML (multi-layer in SWAN) and NHWAVE to model the presence of the vegetation were calibrated separately against the physical model data after removing the effects of bottom and sidewall friction. Although it was expected that every model would have to be recalibrated for a given wave condition and water depth due to the dependence of C D on the Keulegan–Carpenter number, this paper showed that it was also necessary to recalibrate C D in SWAN-SL for a given wave condition and water depth when the two different vegetation treatments were applied. Conversely, it was shown that the C D values in NHWAVE and SWAN-ML changed very little between the two treatments, highlighting the utility of a layered SWAN and RANS model for estimating wave attenuation when the aboveground biomass is no longer uniform over the vertical. The vertical structure of vegetation-induced turbulent kinetic energy, eddy viscosity and dissipation rate were investigated numerically using NHWAVE.