Mathematical aspects of space-time horizontal ray method

Aleksandr Kaplun; Boris Katsnelson

REVIEW 4 major objections 4 minor 21 references

Mathematical aspects of space-time horizontal ray method

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Predict a pulse's shape from one waveguide eigenvalue

desk verdict The variation system and Jacobian caustic criterion are real extensions of the vertical-mode horizontal-ray framework, but the transport equation rests on a self-adjointness identity that fails for realistic density contrasts, so the amplitude formula and signal parameter predictions are not justified. read the letter →

arxiv 2411.14178 v1 pith:NX47XTTG submitted 2024-11-21 math-ph math.MP

Aleksandr Kaplun , Boris Katsnelson This is my paper · ORCID

classification math-phmath.MP MSC 35A1735L05

keywords space-timeraysverticalmodeshorizontalunderwateracousticseikonalequationtransportcausticsfrequencymodulation

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper develops the space-time ray version of the standard 'vertical modes and horizontal rays' method for sound propagation in shallow water. It works out the full set of equations—an eikonal equation for the phase, a transport equation for the amplitude, and the corresponding Hamiltonian ray equations—using only the squared vertical-mode eigenvalue $q^2(\mathbf r,k_0)$. The central claim is that once this function is known, one can predict a signal's amplitude, frequency modulation, phase and pulse fronts, and space-time caustic positions at an observation point without solving the wave field for many frequencies separately. The reason this matters is practical: it turns a computationally heavy multi-frequency mode calculation into a ray-geometry reading of one dispersion-related function.

What carries the argument

The central object is the eigenvalue function $q(\mathbf r,k_0)$, the positive square root of the eigenvalue of the vertical Sturm-Liouville problem $\psi''+n^2(\mathbf r,z)k_0^2\psi=q^2\psi$ with boundary conditions at the surface and bottom. This one function carries all information about dispersion and horizontal variability: it defines the eikonal equation, the Hamiltonian $H=|\mathbf k|^2-q^2$, the space-time ray velocity $\hat{\boldsymbol\kappa}=(-\partial q/\partial k_0,\mathbf k/q)$, and the group velocity $v=-(\partial q/\partial k_0)^{-1}$. The paper's distinctive machinery is the transport equation written as a three-dimensional divergence $\widehat{\mathrm{div}}(qA^2\hat{\boldsymbol\kappa})=0$, together with the fundamental variation matrix $M$ that supplies the Jacobian $D(\tau,\mu,\nu)$ used for amplitudes and for the caustic condition $D=0$.

What would settle it

A decisive check would be a numerical experiment in a range-dependent shallow-water waveguide: compute the exact vertical-mode inner product $\langle n^2\psi,\psi\rangle$ as a function of $k_0$, compare it with $(q/k_0)\partial q/\partial k_0$, and then propagate a broadband pulse with the ray formulas and with a full-wave solver. If the relative error in the inner product is not small, or if the ray-predicted amplitude and frequency modulation diverge from the full-wave result near a known caustic, the transport equation (2.36) would be falsified.

Watch

Extended reading notes

Core claim

The paper establishes that a time-dependent acoustic field in a shallow-water waveguide, written as a sum of adiabatic vertical modes times a slowly varying amplitude and a fast phase $e^{i\varepsilon^{-1}\varphi}$, obeys a Hamiltonian ray system in the four-dimensional phase space $(\tau,\mathbf r,k_0,\mathbf k)$. The phase satisfies the eikonal equation $q^2(\mathbf r,\partial\varphi/\partial\tau) - |\nabla\varphi|^2 = 0$, where $q(\mathbf r,k_0)$ is the square root of the vertical-mode eigenvalue, and the leading amplitude satisfies the transport equation $\widehat{\mathrm{div}}(q A^2 \hat{\boldsymbol\kappa}) = 0$ in the three-dimensional space-time of $\tau,\mathbf r$. From these two equations the paper derives explicit formulas: the amplitude along a ray as $A(\tau)=A(0)\sqrt{g(0)D(0)/(g(\tau)D(\tau))}$, the phase advance $\varphi(\tau)=\varphi(0)+k_0\tau+\int_0^\tau q v\,d\tau'$, and the condition $D(\tau,\mu,\nu)=0$ that locates space-time caustics. The paper's central claim is that the observable signal parameters at a given point—frequency modulation, amplitude, phase-front and pulse-front tilts—are all read off these ray objects, so a single function $q$ controls the propagation.

Load-bearing premise

The load-bearing premise is that a certain weighted average of the squared refractive index over a vertical mode can be replaced by the simple expression $(q/k_0)\partial q/\partial k_0$ with no error bound. That approximation is what converts the transport equation into its final divergence form, and therefore everything the paper says about amplitude and frequency modulation inherits it. If vertical modes change noticeably with frequency, the replacement is not controlled and the predictions can be wrong.

Editorial extensions

If this is right

In any waveguide where the eigenvalue function $q(\mathbf r,k_0)$ is available, the method gives closed-form predictions for the frequency modulation $k_0(\rho)$ and the phase-front normals at a specified observation point.
The vanishing of the Jacobian $D(\tau,\mu,\nu)$ marks space-time caustics, so caustic surfaces and their projections can be located directly from the variation matrix $M$.
When several rays reach the observation point, the total field is a coherent sum of their contributions; instantaneous frequency and wave vector are no longer unique, and the model accounts for the resulting interference pattern.
In the constant-group-velocity limit the $\tau$-front and the $s$-front coincide, and the amplitude along a ray decays as $1/\sqrt{s}$, giving a simple characterization of the amplitude front.
All of these predictions depend on the model only through $q$, so the same equations apply to any shallow-water environment once the vertical eigenvalue problem is solved.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

This formulation suggests a natural extension beyond non-caustic points: a Maslov-type or uniform asymptotic continuation of the amplitude formula would remove the $D^{-1/2}$ singularities at space-time caustics, which the paper does not attempt.
Because the derivation uses only the dispersion relation $q(\mathbf r,k_0)$, the same space-time ray apparatus should transfer to electromagnetic pulses in dispersive media, where pulse-front tilt and angular dispersion are already known effects.
A direct numerical check would be feasible: compare the predicted $k_0(\rho)$ and amplitude $A(\tau)$ against a full-wave solution for a range-dependent wedge with a strongly frequency-dependent mode; the paper does not report such a comparison.
The weakest numerical step, the approximation (2.32), could be tested in isolation by comparing exact eigenfunction inner products with $(q/k_0)\partial q/\partial k_0$ over the frequency band of interest; if the discrepancy is small, the amplitude and frequency predictions are on solid ground.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The variation system (3.30)-(3.33) and the Jacobian-based caustic criterion are the genuine additions here, and they are worked out carefully. The broader framework — vertical modes plus space-time rays — is known from Babich and from Burridge-Weinberg, so the novelty is incremental but legitimate. The eikonal and ray equations are standard but cleanly derived, and the front analysis in Section 4 is clear. Credit where due: the paper is honest about omitting mode coupling, and the bibliography is appropriate.

The problem is the transport equation. The inner product (2.22) uses rho as a weight, but the interface condition (2.19) makes L = d^2/dz^2 + n^2 k0^2 non-self-adjoint in that inner product. A direct integration by parts gives a boundary term at z = h proportional to (rho_+ - rho_-^2/rho_+) psi'(h-) psi(h). That term is missing from eq. (2.31). Consequently eq. (2.32) is not the identity they claim, and the simplified transport equation (2.36) — and the amplitude formula (2.53) — do not follow. If the inner product were weighted with 1/rho, the identity would hold; as written it does not. The authors might be able to fix this, but the paper does not.

This is a load-bearing flaw, not a cosmetic one. The abstract promises a simple method for predicting signal amplitude and frequency modulation; that promise depends on (2.36). There is also no numerical test or worked example, and the approximation in (2.32) is uncontrolled even apart from the missing boundary term. So the paper currently supports the ray geometry and the caustic criterion, but not the amplitude predictions.

Who gets value from this: a specialist in underwater acoustics or asymptotic methods who is interested in the variation system and the caustic Jacobian, and who is willing to do some algebra to repair the transport equation. The paper is not ready for use as a predictive tool as written.

Recommendation: send it to peer review, but flag the self-adjointness issue prominently. It is the sort of specific, fixable mathematical error that referees can work through. With the transport equation corrected and at least one illustrative example or error estimate, the paper would be a solid contribution. Without that, it remains a promising sketch.

Referee Report

4 major / 4 minor

Summary. The paper develops a space-time horizontal ray method for acoustic wave propagation in a shallow-water waveguide. It starts from the wave equation (2.10) and an ansatz (2.11) separating a vertical mode ψ from a two-dimensional phase φ. Substitution yields an eikonal equation (2.23) in which the vertical-mode eigenvalue q plays the role of a refractive index, and a transport equation (2.36) for the amplitude A. The associated Hamiltonian system (2.29) defines space-time rays; the paper then derives a linearized system (3.30)-(3.33) for ray-coordinate variations, uses it to compute the Jacobian of the transition to ray coordinates, and characterizes space-time caustics as zeros of that Jacobian. It also treats phase fronts, τ-fronts, s-fronts, and multiple-ray interference. The derivation is formal: no numerical examples, no error estimates, and no comparison with full-wave solutions are given.

Significance. If the derived equations are correct, the method provides a compact description of pulsed signal propagation in a dispersive waveguide, with amplitude, frequency modulation, and front geometry expressed through the single function q(⃗r,k0). The derivation is self-contained and does not introduce fitted parameters; I also find no circularity, and the potential inconsistency between the inner product (2.22) and the interface condition (2.19) is resolved by the interface condition itself. However, the transport equation rests on several unquantified approximations, and the absence of numerical validation means the claimed predictive power is not demonstrated. The paper is a useful formal framework that could become a practical tool after further justification and testing.

major comments (4)

[§2.4, Eqs. (2.32)-(2.36)] The approximation ⟨n^2ψ,ψ⟩ ≈ (q/k0)∂q/∂k0 is used without derivation or error estimate. Under the normalization (2.22), this identity is exact if one differentiates the eigenvalue equation Lψ = q^2ψ with respect to k0 and uses self-adjointness; the '≈' is therefore misleading. More importantly, Eq. (2.34) omits the term (1/2)⟨∂n^2/∂τ ψ,ψ⟩, which is generally nonzero and contributes to the transport equation. The authors should prove the simplified transport equation (2.36) under explicit scale-separation assumptions or provide a remainder estimate.
[§2.4, Eq. (2.33)] The equality ⟨(∇φ,∇ψ),ψ⟩ = ½(∇φ,∇⟨ψ,ψ⟩) is not generally valid, because the inner product (2.22) depends on the horizontal coordinates through ρ± and h(⃗r). Differentiation with respect to x introduces boundary terms from ∇h and terms from ∇ρ that are not negligible in the asymptotic ordering of §2.1. These missing terms affect the divergence form (2.36) and the amplitude formula (2.53). The transport equation should be derived from the exact modal relations, keeping all parametric derivatives of the inner product.
[Throughout (no numerical section)] The paper contains no numerical example or comparison with a direct solution. The abstract promises a simple method for predicting signal form and parameters, but no demonstration is provided. I request at least one benchmark (for instance, a range-dependent Pekeris waveguide) in which the amplitude predicted by (2.53) is compared with a full-wave solution, to assess the quantitative accuracy of the approximations and to establish the method's practical range of validity.
[§3.4, Eqs. (3.58)-(3.59)] The coherence condition on the initial data is stated as a first-order PDE for φ0, but the existence of solutions is not analyzed. For a prescribed family of rays (ρ0,r0,k0,α0), the compatibility of (3.59) is not obvious. The paper should give conditions on the initial functions that guarantee solvability, or explain how φ0 is constructed in practice, since this condition is essential for the ray-coordinate framework.

minor comments (4)

[§3.1, Eq. (3.13)] The definition of Δα(τ) subtracts the same expression from itself, making it identically zero; this is a typo and should read α(τ,μ,ν) - α(τ,μ',ν') or the reverse. The same issue appears in the initial-condition formula that follows.
[§3.1-§3.2] The notation 'Δ_|||' (with three vertical bars) appears in equations (3.7), (3.11), (3.18), (3.30), and elsewhere; it should be 'Δ_||' (two bars).
[Abstract and footnote 1] There are minor typos: 'vertical mode s' in the abstract should be 'vertical modes', and 'ommit' in footnote 1 should be 'omit'.
[§2.4, Eq. (2.32)] The bare '≈' in (2.32) and (2.34) should be replaced by an explicit statement of the approximation regime (e.g., weak frequency dependence of the vertical modes) and an indication of the order of the neglected terms.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ray predictions follow from the Sturm–Liouville eikonal and transport equations without fitted inputs or load-bearing self-citation.

full rationale

The paper's claimed predictions (signal form, frequency modulation, front angles, caustics) are derived from the eikonal equation (2.23), q^2 - |∇φ|^2 = 0, and the transport equation (2.36)-(2.37), with q defined as the positive square root of the Sturm-Liouville eigenvalue from (2.17)-(2.21). No parameter appearing in the predicted amplitude or phase is fitted to the quantity being predicted; initial amplitudes and phases are arbitrary data of the asymptotic scheme, and the propagation law (2.53) follows from the divergence-form transport equation. The approximation (2.32), ⟨n^2ψ,ψ⟩ ≈ (q/k0) ∂q/∂k0, is an analytic simplification inside the derivation, and any failure of that approximation would be a correctness issue, not a circular reduction. References to the authors' prior work ([13], [14], [15]) supply background on shallow-water modes and interference; they are not used to forbid alternatives or as the sole justification of the central ray equations. Thus no circular step can be exhibited, and the appropriate finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central derivation rests on standard asymptotic ray theory, the assumption of a single adiabatic mode, and an uncontrolled approximation for the transport equation. No fitted parameters are introduced. The eigenvalue field q(r,k0) is an input obtained from the medium's sound-speed profile and bottom properties via a Sturm-Liouville problem; ε is an asymptotic bookkeeping parameter, and c0 is a reference speed.

assumptions (6)

domain assumption The discrete positive spectrum of the Sturm-Liouville operators L[r,k0] is simple and finite for the whole range of parameters.
Section 2.2, equation (2.20). The theory relies on well-defined isolated eigenvalues q_l^2.
domain assumption The multi-scale estimate ∂/∂(εX) ~ ∂/∂(εY) ~ ∂/∂(εc0T) ~ ∂/∂Z with ε<0.1, i.e. slow horizontal/temporal variation compared to vertical.
Section 2.1, equations (2.7)-(2.9). Central to the ansatz and the ε-asymptotics.
ad hoc to paper The approximation ⟨n^2ψ,ψ⟩ ≈ (q/k0)(∂q/∂k0), used to simplify the transport equation.
Section 2.4, equation (2.32). This is an approximation without an error bound; the precise form (2.38) is also given, but the main amplitude formula uses the approximation.
domain assumption Only one vertical mode propagates; mode coupling is omitted.
Footnote 1, Section 2.1. The ansatz assumes a single vertical mode.
standard math Radiation conditions and ψ ∈ L^2(R_+) for vertical modes.
Section 2.2, footnote 2.
domain assumption Smoothness of q(r,k0) and no caustics near observation point for the expansions used.
Sections 3.3-3.5. The inversion of J and front normals require nondegeneracy except at caustics.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Mathematical aspects of space-time horizontal ray method." pith.science (2026). https://pith.science/paper/NX47XTTG

@misc{pith2026241114178,
  author       = {Pith},
  title        = {Pith review of: Mathematical aspects of space-time horizontal ray method},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NX47XTTG}},
  note         = {Machine review of arXiv:2411.14178}
}

read the original abstract

The following development of the well-known "vertical modes and horizontal rays" approach for acoustic waves propagation in shallow water, introduced in different works, is studied. In this approach we study so-called space-time horizontal rays, constructed on the base of decomposition of the sound field, depending on time, over adiabatic vertical modes (solutions of the Sturm-Liouville problem). Using this technique we obtain different properties of signals, propagating in underwater waveguide, such as space-time caustics, and provide rather simple method for the prediction of the form of the signal and all its parameters (amplitude and frequency modulation, different front angles, etc.) at some point of observation.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

[1]

Babich, Ivan A

Vasili M. Babich, Ivan A. Molotkov, and Vladimir S. Buldy rev. The space- time ray method . Cambridge Univ. Press, 1998

work page 1998
[2]

Space-time ray method of the calculation of waves in slightly inhomogen eous layered medium

Vasilii Mikhailovich Babich and Natal’ya Sergeevna Gri gor’eva. Space-time ray method of the calculation of waves in slightly inhomogen eous layered medium. Zapiski Nauchnykh Seminarov POMI , 99:5–18, 1980

work page 1980
[3]

quasiphotons

V.M. Babich and V.V. Ulin. Complex space-time ray method and “quasiphotons”. Journal of Soviet Mathematics , 24:269–273, 1984

work page 1984
[4]

D.R. Bergman. Generalized space-time paraxial acousti c ray tracing. Waves in Random and Complex Media , 15(4):417–435, 2005. 19

work page 2005
[5]

Bor and B

Zs. Bor and B. Racz. Group velocity dispersion in prisms a nd its applica- tion to pulse compression and travelling-wave excitation. Optics communi- cations, 54(3):165–170, 1985

work page 1985
[6]

Horizontal rays and vertical modes

Robert Burridge and Henry Weinberg. Horizontal rays and vertical modes. Wave propagation and underwater acoustics , pages 86–152, 2005

work page 2005
[7]

Molotkov, and Ivan Pˇ senˇ c ´ ık

Vlastislav ˇCerven` y, I.A. Molotkov, and Ivan Pˇ senˇ c ´ ık. Space-time r ay method for seismic wave ﬁelds. Studia Geophysica et Geodaetica, 26(4):342– 351, 1982

work page 1982
[8]

Kenneth Allen Connor and L.B. Felsen. Complex space-tim e rays and their application to pulse propagation in lossy dispersive media . Proceedings of the IEEE , 62(11):1586–1598, 1974

work page 1974

Show all 21 references

[9]

Paul C. Etter. Underwater acoustic modeling and simulation . CRC press, 2018

2018
[10]

L.B. Felsen. Hybrid ray-mode ﬁelds in inhomogeneous wa veguides and ducts. The Journal of the Acoustical Society of America , 69(2):352–361, 1981

1981
[11]

J. Hebling. Derivation of the pulse front tilt caused by angular dispersion. Optical and Quantum Electronics , 28:1759–1763, 1996

1996
[12]

Generation of high-power terahertz pulses b y tilted-pulse- front excitation and their application possibilities

J´ anos Hebling, Ka-Lo Yeh, Matthias C Hoﬀmann, Bal´ azs Bartal, and Keith A Nelson. Generation of high-power terahertz pulses b y tilted-pulse- front excitation and their application possibilities. JOSA B , 25(7):B6–B19, 2008

2008
[13]

Fundamentals of shallow water acoustics

Boris Katsnelson, Valery Petnikov, and James Lynch. Fundamentals of shallow water acoustics . Springer Science & Business Media, 2012

2012
[14]

Katsnelson, Valery A

Boris G. Katsnelson, Valery A. Grigorev, and James F. Ly nch. Variability of phase and amplitude fronts due to horizontal refraction i n shallow water. The Journal of the Acoustical Society of America , 143(1):193–201, 2018

2018
[15]

Katsnel’Son and A

B.G. Katsnel’Son and A. Yu. Malykhin. Space-time sound ﬁeld interference in the horizontal plane in a coastal slope region. Acoustical Physics, 58:301– 307, 2012

2012
[16]

Re- ﬂection of space-time ray amplitudes from arbitrary moving boundary

Natal’ya Yakovlevna Kirpichnikova and Mikhail Mikhai lovich Popov. Re- ﬂection of space-time ray amplitudes from arbitrary moving boundary. Za- piski Nauchnykh Seminarov POMI , 128:72–88, 1983

1983
[17]

Yu. A. Kravtsov and Yu. I. Orlov. Caustics, catastrophes and wave ﬁelds , volume 15. Springer Science & Business Media, 2012

2012
[18]

Kravtsov and Yuri Ilich Orlov

Yury A. Kravtsov and Yuri Ilich Orlov. Geometrical optics of inhomoge- neous media, volume 38. Springer, 1990. 20

1990
[19]

Kov´ acs, Zsuzsanna Heiner, G´ abor Kurdi, J´ ozsef Klebniczki, and M´ arta Csat´ ari

K´ aroly Osvay, Attila P. Kov´ acs, Zsuzsanna Heiner, G´ abor Kurdi, J´ ozsef Klebniczki, and M´ arta Csat´ ari. Angular dispersion and te mporal change of femtosecond pulses from misaligned pulse compressors. IEEE Journal of selected topics in quantum electronics , 10(1):213–...

2004
[20]

Topp and Gail C

Michael R. Topp and Gail C. Orner. Group dispersion eﬀec ts in picosecond spectroscopy. Optics Communications, 13(3):276–281, 1975

1975
[21]

Horizontal ray the ory for ocean acoustics

Henry Weinberg and Robert Burridge. Horizontal ray the ory for ocean acoustics. The Journal of the Acoustical Society of America , 55(1):63–79, 1974. 21

1974

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Babich, Ivan A

Vasili M. Babich, Ivan A. Molotkov, and Vladimir S. Buldy rev. The space- time ray method . Cambridge Univ. Press, 1998

work page 1998

[2] [2]

Space-time ray method of the calculation of waves in slightly inhomogen eous layered medium

Vasilii Mikhailovich Babich and Natal’ya Sergeevna Gri gor’eva. Space-time ray method of the calculation of waves in slightly inhomogen eous layered medium. Zapiski Nauchnykh Seminarov POMI , 99:5–18, 1980

work page 1980

[3] [3]

quasiphotons

V.M. Babich and V.V. Ulin. Complex space-time ray method and “quasiphotons”. Journal of Soviet Mathematics , 24:269–273, 1984

work page 1984

[4] [4]

D.R. Bergman. Generalized space-time paraxial acousti c ray tracing. Waves in Random and Complex Media , 15(4):417–435, 2005. 19

work page 2005

[5] [5]

Bor and B

Zs. Bor and B. Racz. Group velocity dispersion in prisms a nd its applica- tion to pulse compression and travelling-wave excitation. Optics communi- cations, 54(3):165–170, 1985

work page 1985

[6] [6]

Horizontal rays and vertical modes

Robert Burridge and Henry Weinberg. Horizontal rays and vertical modes. Wave propagation and underwater acoustics , pages 86–152, 2005

work page 2005

[7] [7]

Molotkov, and Ivan Pˇ senˇ c ´ ık

Vlastislav ˇCerven` y, I.A. Molotkov, and Ivan Pˇ senˇ c ´ ık. Space-time r ay method for seismic wave ﬁelds. Studia Geophysica et Geodaetica, 26(4):342– 351, 1982

work page 1982

[8] [8]

Kenneth Allen Connor and L.B. Felsen. Complex space-tim e rays and their application to pulse propagation in lossy dispersive media . Proceedings of the IEEE , 62(11):1586–1598, 1974

work page 1974