Journal of Fluid Mechanics

Papers

Direct numerical simulation of wind-wave generation processes

MEI-YING LINa1a3, CHIN-HOH MOENGa2 c1, WU-TING TSAIa1a4, PETER P. SULLIVANa2 and STEPHEN E. BELCHERa5

a1 Department of Civil Engineering, National Chiao Tung University, Hsinchu 30050, Taiwan

a2 National Center for Atmospheric Research, Boulder, Colorado 80307, USA

a3 Taiwan Typhoon and Flood Research Institute, Taichung 40763, Taiwan

a4 Institute of Hydrological Sciences, National Central University, Taoyuan 32001, Taiwan

a5 Department of Meteorology, University of Reading, Earley Gate, PO Box 243, Reading RG6 6BB, UK

Abstract

An air–water coupled model is developed to investigate wind-wave generation processes at low wind speed where the surface wind stress is about 0.089 dyn cm−2 and the associated surface friction velocities of the air and the water are u*a~8.6 cms−1 and u*w~0.3 cms−1, respectively. The air–water coupled model satisfies continuity of velocity and stress at the interface simultaneously, and hence can capture the interaction between air and water motions. Our simulations show that the wavelength of the fastest growing waves agrees with laboratory measurements (λ~8–12 cm) and the wave growth consists of linear and exponential growth stages as suggested by theoretical and experimental studies. Constrained by the linearization of the interfacial boundary conditions, we perform simulations only for a short time period, about 70s; the maximum wave slope of our simulated waves is ak~0.01 and the associated wave age is c/u*a~5, which is a slow-moving wave. The effects of waves on turbulence statistics above and below the interface are examined. Sensitivity tests are carried out to investigate the effects of turbulence in the water, surface tension, and the numerical depth of the air domain. The growth rates of the simulated waves are compared to a previous theory for linear growth and to experimental data and previous simulations that used a prescribed wavy surface for exponential growth. In the exponential growth stage, some of the simulated wave growth rates are comparable to previous studies, but some are about 2–3 times larger than previous studies. In the linear growth stage, the simulated wave growth rates for these four simulation runs are about 1–2 times larger than previously predicted. In qualitative agreement with previous theories for slow-moving waves, the mechanisms for the energy transfer from wind to waves in our simulations are mainly from turbulence-induced pressure fluctuations in the linear growth stage and due to the in-phase relationship between wave slope and wave-induced pressure fluctuations in the exponential growth stage.

(Received June 06 2007)

(Revised August 20 2008)

Correspondence:

c1 Email address for correspondence: moeng@ucar.edu

Metrics