A reduced model for surface wave–current interactions without spatial scale separation

Yohei Onuki 1, Yasushi Fujiwara 2 · 2026

Read A reduced model for surface wave–current interactions without spatial scale separation by Yohei Onuki 1, Yasushi Fujiwara 2 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-03231.txt.

Introduction

Onuki and Fujiwara model how ocean surface waves and currents shape each other. The setting is a weakly nonlinear gravity wave field over a slow, incompressible current in a rotating fluid. The base is Craik-Leibovich theory for the wave-averaged momentum equation. In that theory the Stokes drift is set from outside. Here it is not.

The drift comes from a companion amplitude equation for a narrow-band wave field. The field sits near a wavenumber circle set by a carrier frequency. The derivation mixes a multi-time-scale expansion in wave steepness with a closure. The closure drops fourth-order wave interactions but keeps the third-order Stokes fix. No split of spatial scales is imposed. The wave equation can thus show drift-driven advection, refraction, and scattering in many directions. The coupled system conserves wave action and has closed energy and momentum budgets.

Catalog record

CREATORYohei Onuki 1, Yasushi Fujiwara 2
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.03231

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A reduced model for surface wave–current interactions without spatial scale separation Yohei Onuki 1, Yasushi Fujiwara 2 Abstract We propose a reduced asymptotic model for the mutual interaction between a weakly nonlinear surface gravity wave field and a slowly evolving incompressible current in a homogeneous rotating fluid. The formulation builds on the Craik–Leibovich theory for the wave-averaged momentum equation, but the Stokes drift is not prescribed externally. Instead, it is determined by a companion amplitude equation for a narrow-band wave field concentrated near the wavenumber circle associated with a prescribed carrier frequency. The derivation combines a multiple-time-scale expansion in wave steepness with a phenomenological closure that neglects quartic wave–wave interactions while retaining the third-order Stokes correction. Importantly, no spatial-scale separation is imposed on the wave–current interaction, allowing the wave equation to represent current-induced advection, refraction, and multidirectional scattering. The resulting equations conserve wave action and admit closed energy and momentum budgets for the coupled wave–current system. The model thus provides a tractable bidirectional extension of the classical Craik–Leibovich framework for regimes in which current-induced wave evolution feeds back significantly on the mean flow. † † affiliation: 1 Research Institute for Applied Mechanics, Kyushu University, Kasuga, Fukuoka 816-8580, Japan 2 Graduate School of Maritime Sciences, Kobe University, Kobe, Hyogo 658-0022, Japan 1 Introduction Turbulent mixing in the ocean surface boundary layer plays an essential role in the climate system through its influence on sea-surface temperature, air–sea gas exchange, and tracer transport. One of its principal driving mechanisms is Langmuir circulation, arising from the interaction between surface waves and currents ( 1 ) . The standard theoretical framework for this interaction is the classical Craik–Leibovich (CL) theory ( 2 ; 5 ) , in which the wave effect enters the wave-averaged momentum equation through the Stokes drift and the associated vortex force. Because this formulation captures, in a compact way, how surface waves organize the mean-vorticity field into Langmuir circulations, it has provided the foundation for much of the theoretical and numerical literature on Langmuir turbulence and wave-averaged circulation models. In many applications, the CL equations are closed by prescribing the Stokes drift externally rather than solving for the wave field itself ( 7 , e.g.,) . This is often an effective approximation when wave evolution can be treated independently of the current dynamics of interest. In that prescribed-wave limit, however, the surface wave field is not an explicit dynamical energy reservoir within the reduced model; instead, it shapes the pathways through which the mean flow is destabilized and rearranged. Actual surface waves, by contrast, are advected, refracted, scattered, and amplitude-modulated by spatially inhomogeneous and time-dependent currents. Such mutual wave–current interactions have recently been reported in direct numerical simulations ( 3 ) , ocean current measurements ( 8 ) , and linear stability analyses ( 15 ; 16 ) . Surface waves therefore need not be merely catalytic: they may participate actively in energy exchange with Langmuir-type motions. This viewpoint motivates a wave-resolving closure of the CL framework. Existing two-way coupled wave–current interaction theories provide important precedents. The asymptotic models of 6 , 13 and 10 , together with the wave-action transport model in position–wavevector phase space developed by 14 , all go beyond the standard CL closure with prescribed Stokes drift by allowing current-induced refraction and amplitude modulation. Yet these formulations are based either on leading-order single-phase descriptions or on asymptotically scale-separated (i.e., WKB-type) action/ray descriptions. They are therefore not designed to describe, in general form, multidirectional scattering and directional spectral redistribution induced by fully three-dimensional currents such as Langmuir circulations, whose horizontal scales are comparable to the wavelength. Here we propose a compact model that closes the CL momentum equation by coupling it to an evolution equation for the wave field, without imposing a spatial-scale separation between the waves and the vortical flow. The model is intended for weakly nonlinear, non-breaking surface gravity waves in deep water. Its distinctive spectral assumption is that the wave field is narrowband in frequency around \omega , but not localized in propagation direction: the horizontal spectrum is concentrated near the circle |\bm{k}|=\kappa , rather than near a single carrier wavevector. The vortical flow is assumed to be slow, with velocity of the same order as the Stokes drift, and to evolve on the advective time scale associated with that velocity. Our asymptotic formulation is guided by the reconstituted amplitude equations of 17 , 12 and 11 , which describe wave propagation through prescribed inhomogeneous environments. The closest analogue to the present bidirectional construction is the reduced model of 18 , in which near-inertial waves are coupled to quasi-geostrophic flows. After deriving asymptotically consistent reduced equation systems, we further introduce phenomenological modifications so that the resulting wave–current system has the desired nonlinear dispersion relation and conservation structure. In particular, the wave equation admits a wave-action invariant, while the coupled system has closed energy and momentum budgets. Together, these properties make the model a closed and energetically consistent extension of the CL framework for wave–current interactions that does not rely on WKB-type scale separation. The paper is organized as follows. Section 2 states the final reduced equations in dimensional form together with the assumptions under which they are intended to apply. Sections 3 – 5 then derive the corresponding nondimensional asymptotic system, from which the equations in Section 2 follow. Specifically, Section 3 introduces the dimensional equations and separately scales the wave and current fields, Section 4 derives the wave-amplitude equation and the Stokes drift, and Section 5 modifies the equation system to retain higher-order effects in the wave dispersion relation and improve the energy conservation property. Section 6 establishes the conservation laws for the dimensional reduced system, and Section 7 discusses the limitations of the present closure and suggests possible future work. Appendix A presents a compact derivation for the CL equation used in the main text. 2 Reduced model and asymptotic regime We use Cartesian coordinates (x_{1},x_{2},z) , with z directed upward, and consider an incompressible fluid in -h\leq z\leq 0 with a flat bottom and a free surface. Note that horizontal indices i,j run over 1,2 and are summed when repeated. The averaged flow dynamics involves two kinds of velocities: the Lagrangian-mean velocity \bm{U}^{L}=(U^{L}_{1},U^{L}_{2},W^{L}) and the horizontal Stokes drift \bm{U}^{s}=(U^{s}_{1},U^{s}_{2},0) . We impose the incompressible constraint and the impermeable boundary conditions for the Lagrangian-mean velocity, \displaystyle\nabla\cdot\bm{U}^{L}=0,\qquad W^{L}=0\quad\text{at}\quad z=-h,0. (1) We adopt the traditional approximation for rotation: the Coriolis parameter f is taken constant and only the vertical component of the planetary rotation enters the equations. Here and below, \bm{z} denotes the upward unit vector, and t denotes the time variable. The wave-averaged momentum equation is given by \displaystyle\bm{U}^{L}_{t}+\bm{U}^{L}\cdot\nabla\bm{U}^{L}+(f\bm{z}-\nabla\times\bm{U}^{s})\times\bm{U}^{L}=-\nabla\Pi+\bm{U}^{s}_{t}, (2) where \Pi is a modified pressure enforcing the incompressible constraint. This is a variant of the familiar CL equation ( 9 ) . The wave field is assumed to be narrow-band in intrinsic frequency and concentrated near the circle |\bm{k}|=\kappa in horizontal wavenumber space, without imposing a preferred propagation direction. We propose an evolution equation for the wave field specified by a complex amplitude A(x_{1},x_{2},t) that obeys \displaystyle\left[1+\alpha\left(\partial_{i}^{2}+\kappa^{2}\right)\right]A_{t}-\frac{{\rm i}\omega_{\kappa}}{2\kappa}\left(\partial_{i}^{2}+\kappa^{2}\right)A=\frac{\omega_{\kappa}}{2\kappa\omega}\mathcal{L}(\bm{U}^{L},A), (3) where \displaystyle\mathcal{L}(\bm{U}^{L},A) \displaystyle\equiv\left(\overline{U}^{L}_{i}A_{,ij}\right)_{,j}+\left(\overline{U}^{L}_{i}A_{,j}\right)_{,ij}-\kappa^{2}\tilde{U}^{L}_{i}A_{,i}-\kappa^{2}\left(\tilde{U}^{L}_{i}A\right)_{,i}, (4a) \displaystyle\overline{U}^{L}_{i} \displaystyle\equiv\int_{-h}^{0}\Phi^{2}U^{L}_{i}\,dz,\qquad\tilde{U}^{L}_{i}\equiv\frac{1}{\kappa^{2}}\int_{-h}^{0}\Phi_{z}^{2}U^{L}_{i}\,dz. (4b) The coefficient \displaystyle\alpha=\frac{\kappa\omega_{\kappa\kappa}-\omega_{\kappa}}{4\kappa^{2}\omega_{\kappa}} (5) is the standard reconstitution parameter used to improve the linear dispersion accuracy (

Read the complete text