小振幅海洋内波的演变、破碎和所致混合

A THREE-DIMENSIONAL NUMERICAL STUDY OF EVOLUTION, BREAKING AND MIXING OF INTERNAL WAVES IN STRATIFIED OCEAN

利用基于谱方法和MPI并行运算的数值模式SpectralModel,直接数值模拟了三维小振幅海洋内波的演变、破碎和所致湍流混合,指出导致其不稳定而破碎的为PSI(parametric subharmonic instability)机制;对于内波破碎所致的湍流混合过程,分析了跨等密度面扩散系数kρ、混合效率γ、浮力通量谱、动能谱以及势能谱等统计性质:内波破碎前,kρ和γ保持低值水平,浮力通量谱值为负,且集中在低波数段;内波破碎后,kρ和γ迅速增大,最大值分别约为0.9×103m2/s和0.18,浮力通量谱值在低波数段为负值,在高波数段为正值,这是因为层化湍流中势能向小尺度运动传递和动能向小尺度运动传递相比更为有效。在内波破碎、强湍流混合阶段,势能谱存在一谱段满足kz3律,P(kz)=0.2N2kz3。此外,与二维模拟结果相比较,导致内波不稳定而破碎的均为PSI机制,kρ、浮力通量谱、势能谱变化趋势大体一致;但三维数值实验中,内波破碎时间提前,湍流衰减加快;KE谱在高波数部分下降速度相对减小,更接近于kz3。

Utilizing spectral method and MPI parallel computing, three dimensional ocean internal waves were direct numerically simulated by SpectralModel to study its evolution, breaking, and turbulent mixing. It is the PSI mechanism that induced the instable and breaking of internal waves. For the turbulent mixing process, the diapycnal diffusion coefficient, mixing efficiency, buoyancy flux spectra, kinetic energy spectra and potential energy spectra were analyzed. Before wave breaking, kρ and γ remained low-value levels, and the buoyancy flux spectra were negative, and concentrated in the low wavenumber segment. After wave breaking, kρ and γ increased rapidly, maximum value reached 0.9×10？3m2/s and 0.18, respectively. The buoyancy flux spectra values were negative in the low wavenumber segment and positive in the high wavenumber segment. This is because the potential energy is transferred more efficiently toward small scales than kinetic energy. During the stage of wave breaking and strong turbulence, The PE spectra fitted the kz3 law： P （kz） = 0.2N2kz3. In addition, the results are compared to that of the two-dimensional direct numerical simulations, and the similarities and differences of them were drawn. Three-dimensional stimulation has the wave breaking occurred ahead of time and the turbulence decay was with accelerated rate, and the KE spectra have smaller decrease rate in high wavenumber segment. However, kρ, buoyancy flux and PE spectra have almost the same variation trends between the twoand three-dimensional simulations.