Exact dispersion relation for linear surface waves on arbitrary vertical shear
Kjell S. Heinrich, Simen Å. Ellingsen · 2026
Read Exact dispersion relation for linear surface waves on arbitrary vertical shear by Kjell S. Heinrich, Simen Å. Ellingsen (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-2604-24484.txt.
Introduction
Heinrich and Ellingsen derive an exact rule linking wave speed, frequency, and wavelength for surface waves riding on a current that changes with depth. Their approach uses a Green's function frame for the governing wave equation, sets viscosity aside, and makes no other simplifications about the current profile. By isolating bend effects in one ordered factor, they recover earlier deep-water and special-case results while giving a clean base for future approximations.
Catalog record
| CREATOR | Kjell S. Heinrich, Simen Å. Ellingsen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49863 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.24484
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Exact dispersion relation for linear surface waves on arbitrary vertical shear Kjell S. Heinrich, Simen Å. Ellingsen Abstract We derive the formal solution to the dispersion relation for linear surface waves on a horizontal mean current with arbitrary vertical dependence. The problem is cast in a Green’s function framework for the Rayleigh equation, neglecting viscosity but making no further approximations about the mean velocity profile. The solution is the dispersion relation in the form of a single, implicit equation relating – and containing only – the velocity profile, wave frequency, and wavenumber. By isolating curvature effects in a path-ordered exponential, we obtain a solution that serves as a natural starting point for systematic approximations. We demonstrate that our solution reduces to the expression found by Shrira (1993, J. Fluid Mech. 252 , 565–584) in the deep-water limit, yields known asymptotic approximations, and recovers known analytical solutions in special cases. I Introduction Surface gravity waves propagating on vertically sheared currents are central to a wide range of oceanographic and engineering problems, from wave forecasting and remote sensing to air–sea interaction and coastal transport. In a range of important questions involving waves and currents together, the vertical structure of a current plays an important role, not just its depth-average or surface value. Examples where vertical shear often needs to be taken into account are ocean transport of plastics and pollutants ( 15 ; 20 ) , wind-stress modelling ( 23 ) , wave-breaking predictions ( 35 ) , wave statistics ( 34 ) , rogue-wave formation and kinematics ( 16 ; 12 ) and wave forces on vessels ( 18 ) and structures ( 32 ) . The problem we consider is a classical one and has been treated lucidly by 24 , 26 , and many others since. We assume waves are propagating on the surface, taken to be at z=0 on average. Here, z is the vertical axis pointing up. The waves travel on a horizontal mean current with arbitrary vertical shear, \bm{U}(z)=[U(z),V(z),0] . The bottom is at a constant depth z=-h . We linearise the incompressible Euler and continuity equations, along with the boundary conditions at the free surface and bottom. We may then assume solutions of Fourier form \varphi(\bm{x},z,t)\sim\hat{\varphi}(z)\exp[\mathrm{i}\psi(\bm{x},t)] , with the phase \psi(x,t)=\bm{k}\cdot\bm{x}_{h}-\omega\,t . The hat represents a Fourier transformed quantity, and \varphi represents a wave-induced velocity \bm{u}=[u,v,w] (the x,y and z components), pressure perturbation, or surface elevation \zeta (where \hat{\zeta} is independent of z ). The horizontal position is \bm{x}_{h}=[x,y,0] ; \bm{k}=[k_{x},k_{y},0] is the wave vector, and \omega is the wave frequency. Only the component of the current parallel to \bm{k} enters the problem. To simplify, we define the shorthand U_{\gamma}\equiv\bm{k}\cdot\bm{U}/k=U\cos\gamma , so that \gamma is the angle between \bm{k} and \bm{U} , and k=|\bm{k}| . Eliminating pressure and horizontal wave velocities, one is left with a second-order differential equation for the vertical wave velocity \hat{w}(z) . This is the Rayleigh, or inviscid Orr-Sommerfeld equation, in the form \displaystyle\hat{w}^{\prime\prime}-k^{2}[1+q(z)]\,\hat{w} \displaystyle=0, (1) with shorthands q(z)\equiv\frac{-U_{\gamma}^{\prime\prime}(z)}{k\Omega(z)};\qquad\Omega(z)\equiv\omega-kU_{\gamma}(z), (2) where a prime denotes a derivative with respect to z . Note that \Omega(z) is the ‘locally Doppler-shifted’ frequency and that the intrinsic wave frequency, i.e. measured in the surface-following reference system, is \sigma(\bm{k})=\Omega(0) . We now turn to the linearised boundary conditions at the bottom and surface, which can be written ( 26 ) \hat{w}(-h)=0;\qquad\sigma^{2}\,\hat{w}^{\prime}(0)=[(g+\Upsilon k^{2})k^{2}-\sigma\,k\,U_{\gamma}^{\prime}(0)]\,\hat{w}(0). (3) The gravitational acceleration is g , and \Upsilon is the kinematic surface tension coefficient. In the absence of current, the solution is \sigma=\omega=\omega_{0} with \omega_{0}\equiv\sqrt{(gk+\Upsilon k^{3})\tanh kh}. (4) Assuming \bm{k} is given, the eigenvalue problem defined by ( 1 ) and ( 3 ) has \hat{w}(z) as an unknown function. There are two permitted eigenvalues for frequency, \omega=\omega_{\pm}(\bm{k}) , corresponding to waves propagating parallel and anti-parallel to \bm{k} . Thus, \omega_{\pm}(-\bm{k})=-\omega_{\mp}(\bm{k}) , so there is only one value of \omega for each wave \bm{k} , with \sigma(\bm{k})>0 . All approaches to the linear dispersion relation – exact, approximate, or numerical – take this classical eigenvalue problem as their starting point. For practical purposes, it is common to derive explicit approximate expressions, such as \omega\approx\omega(\bm{k};\bm{U}) . This method assumes that either a suitably depth-averaged mean-current shear or curvature is small. The classical and most common asymptotic form was derived for h=\infty by 29 and generalised to finite depth by 27 . It reads {\omega(\bm{k};U_{\gamma})\approx\omega_{0}(k)+k\langle U_{\gamma}\rangle} with the second term acting as an “effective Doppler shift”: \displaystyle\langle U_{\gamma}\rangle \displaystyle=2k\int_{-h}^{0}\mathrm{d}z\,U_{\gamma}(z)\frac{\cosh 2k(z+h)}{\sinh 2kh}=U_{\gamma}(0)-s(\bm{k})\frac{\omega_{0}(k)}{k}; (5) \displaystyle s(\bm{k}) \displaystyle\equiv\frac{k}{\omega}_{0}\int_{-h}^{0}\mathrm{d}z\,U_{\gamma}^{\prime}(z)\frac{\sinh 2k(z+h)}{\sinh 2kh}\overset{kh\to\infty}{\longrightarrow}\frac{k}{\omega}_{0}\int_{-h}^{0}\mathrm{d}z\,U_{\gamma}^{\prime}(z)\mathrm{e}^{2kz}. (6) The final form of ( 5 ) is obtained by partial integration. The quantity s(\bm{k}) is a non-dimensional number (for gravity driven waves it might be called a shear–Froude number) comparing the depth-averaged shear to the shear-free wave frequency as a measure of the effect of the shear on wave dispersion; we refer to it as the wave–shear-strength number. In the last form of ( 6 ), \tanh kh\to 1 is understood. Equation 5 was derived under the assumption of a weak current, but 10 showed the actual sufficient criterion is s\ll 1 . Approximations are discussed further in Section III.2 . Alternatively, various iterative numerical schemes have been used to solve the eigenvalue problem. The most classical approach is the piecewise-linear approximation, which approximates U_{\gamma}(z) by partitioning it into a finite number of linear segments ( 31 ; 33 ) . The dispersion relation becomes a solution of a linear system requiring removal of a large number of spurious solutions ( 28 ) . 6 developed a shooting algorithm for solving the Rayleigh equation, while 21 developed a powerful path-following approach suitable when thousands of calculations with very high precision are needed. The ‘Direct integration method’ (DIM) of 17 has gained some popularity due to its rapid convergence and arbitrary accuracy. Explicit, analytical dispersion relations in the form \omega=\omega(\bm{k};U_{\gamma}) have been found only for very special forms of U_{\gamma}(z) . Beyond this, the existing solutions in practical use are either approximations valid when U_{\gamma} satisfies certain k -dependent conditions (these are common), or require solving the eigenvalue problem numerically. Here, we instead derive a dispersion relation equation of implicit form D(\bm{k},\sigma;\,\bm{U})=0 (7) which is fully self-contained and does not involve any unknown eigenfunction of z and allows (numerical or in special cases, analytical) evaluation of \sigma(\bm{k}) . The present work obtains an exact analytical dispersion relation of this form by casting the Rayleigh problem in a Green’s function framework, writing the vertical-velocity Green’s function as a path-ordered exponential that isolates the curvature effect. By constructing two fundamental solutions that separately satisfy the bottom and free-surface boundary conditions and then matching them, we obtain the formal solution; an implicit one-equation dispersion relation that depends only on the prescribed shear profile. This form recovers the classical no-shear and constant-vorticity limits, connects directly to the near-potentiality approximation of 10 , and provides a natural starting point for systematic approximations and numerical evaluation. II Approach and analytical solution Our starting point is the eigenvalue problem in equations ( 1 )–( 3 ). In our derivation, we assume that the surface disturbance comes from a submerged point source. In a linearised theory, a solution for a delta-function disturbance can be integrated to obtain solutions for any distributed source, which is the