Pith. sign in

REVIEW 2 major objections 4 minor 16 references

Finite slab first passage statistics of Henyey Greenstein scattering

T0 review · 2 major / 4 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read This paper establishes that the reflectance, transmittance, and absorptance of a finite scattering slab factor into three independent random-walk statistics: the half-space first-return probability, a slab-thickness survival factor, and an

desk verdict A useful conceptual factorization, well-validated numerically, but the central proof is missing and the angle dependence is left ambiguous. read the letter →

arxiv 2607.00290 v2 pith:I5WMSFTE submitted 2026-07-01 physics.optics

Robert Cordery , Claude Zeller This is my paper · ORCID
classification physics.optics MSC 60J6560G5060J4582C70
keywords randomwalkfirstpassageHenyey–GreensteinradiativetransferslabreflectanceBrownianexcursionabsorptionsurvivalfactor
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

The paper tries to establish that the classical slab transmission problem—light entering a plane-parallel scattering slab, then reflecting, transmitting, or absorbing—factors into three independent random-walk statistics. Its central identity writes slab reflectance R(a,tau) as a sum over scattering order n of the half-space first-return probability P_inf(n), a slab-thickness survival factor S(n,tau), and an absorption weight a^n. If true, all thickness dependence of the slab sits in S(n,tau), and the seemingly two-boundary transport problem becomes a half-space return problem plus a one-dimensional survival factor. The authors support the identity with two independent numerical methods—a direct radiative-transfer operator integrated order by order, and a Monte Carlo scheme that extracts slab excursions from very long walks—which agree to about 1e-3 over a range of asymmetry, thickness, albedo, and incidence angle. The result matters because it makes slab optics reducible to universal random-walk laws, with absorption entering only as a per-collision weight.

What carries the argument

The factorization identity R(a,tau)=sum P_inf(n) S(n,tau) a^n is the load-bearing object. It separates half-space first-return statistics P_inf(n), the slab survival factor S(n,tau)=Pr(zmax<tau|n), and the absorption weight a^n. The RT operator divides the slab into thin layers, treats scattering to first order per layer, and stacks them with adding equations that resum inter-layer reflections; the MC method generates extremely long walks and cuts out excursions through slabs, relying on the memoryless exponential step length so partial boundary-crossing steps have the same distribution. The universal n^{-3/2} first-return tail of Sparre-Andersen is what gives the square-root edge in reflect

What would settle it

Compute the order-resolved reflection probabilities P_R(n,tau) from the RT operator at two incidence angles (say normal and theta0=60) at fixed tau and g. If the ratio P_R(n,tau)/P_inf(n) differs between the two angles, then S(n,tau) depends on mu0 and the single-sum factorization (14) is not valid for oblique incidence. A direct version: use Eq. (14) with S extracted at normal incidence to predict R at theta0=60 for tau=4, g=0.5, a=1, and compare with the operator value 0.6610 from Table 2; a discrepancy beyond the quoted 1e-3 would falsify the angle-independent form.

Watch

Extended reading notes

Core claim

The central claim is Eq. (14): R(a,tau) = sum_{n>=1} P_inf(n) S(n,tau) a^n, with P_inf(n) the order-resolved first-return probability of a half-space walk and S(n,tau)=Pr(zmax<tau|n) the probability that an n-step walk stays below the upper boundary. The paper argues that reflectance, transmittance, and absorptance of a slab all follow from this decomposition, with absorption entering only as the factor a^n. The same walk confined to 0<z<tau and killed at the faces is the transfer operator whose order-resolved escape probabilities give R and T; the Monte Carlo method harvests many slab excursions from a few long walks and applies the same weight. As tau grows, S approaches 1 order by order a

Load-bearing premise

The factorization assumes the survival factor S(n,tau)=Pr(zmax<tau|n) is a well-defined conditional probability that does not depend on the entry direction; the text never specifies whether 'n' means an arbitrary n-step walk, an n-step first-return walk, or one with a fixed incident angle, and R(a,tau) in Eq. (14) carries no mu0 index even though the operator's reflectance clearly depends on incidence angle.

Editorial extensions

If this is right

  • Given P_inf(n) and S(n,tau), slab reflectance, transmittance, and absorptance follow by a single weighted sum; no full two-boundary integration is needed.
  • All thickness dependence of reflectance is carried by S(n,tau); once S is known for a given tau, all albedos a are obtained by the same sum.
  • As tau -> infinity, S(n,tau) -> 1 order by order, so the thick-slab reflection law coincides with the half-space first-return law up to a scattering order n*(tau) that grows with thickness.
  • The universal n^{-3/2} tail implies the conservative reflectance approaches 1 as 1-R(a) ~ 2 sqrt(pi) C(g) sqrt(1-a), a non-analytic square-root cusp at a=1.
  • A single Monte Carlo ensemble of long walks can be re-used for many slab thicknesses and many albedos, because excursions are recorded order-by-order and absorption is applied as a post-hoc weight.

Reading between the lines

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

  • Because Eq. (14) carries no incidence-angle index, the factorization as written likely holds either for normal incidence or for angle-averaged illumination; for oblique incidence a mu0-dependent survival factor S(n,tau|mu0) would be needed, and the full bidirectional reflectance distribution would require an extra integral over entry angle.
  • The scaling of S(n,tau) with tau/(sigma_g sqrt n) points to the Brownian-excursion maximum law as a potential closed-form replacement for numerical S, which would make Eq. (14) fully analytic and extend it to arbitrary thickness without further computation.
  • A practical extension would treat a layered or graded medium as a composition of slab survival factors, using the same factorization repeatedly; that could turn multilayer radiative transfer into a product of half-space kernels.
  • The parameter-free agreement between the operator and Monte Carlo below 1e-3 across g, tau, and a suggests the factorization could serve as a fast forward model for estimating asymmetry or thickness from reflectance measurements.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper develops two complementary methods for computing reflectance, transmittance, and absorptance of a plane-parallel scattering slab under Henyey–Greenstein scattering with exponential free paths. The central result is Eq. (14), which writes the slab reflectance as R(a,τ) = ∑_{n≥1} P∞(n) S(n,τ) a^n, where P∞(n) is a half-space order-resolved first-return probability and S(n,τ) is described as the probability that an n-step trajectory remains below the upper boundary. The authors implement a radiative-transfer (RT) operator via thin-layer stacking and adding- / successive-orders-type composition, and a long-walk Monte Carlo (MC) method that extracts slab excursions from boundary crossings. They report agreement between RT and MC of about 1e-3 absolute, and also compare with adding-doubling, successive-orders, the Chandrasekhar H-function, and a semi-infinite BTF/Motzkin formula.

Significance. If the factorization in Eq. (14) is correct, the paper offers a conceptually appealing reduction of finite-slab transport to half-space first-return statistics plus a thickness survival factor, with absorption entering as a^n. The numerical engine appears credible: the RT operator matches a full 3D MC over a nontrivial range of g, τ, a, and incidence angle, and the additional comparisons to independent deterministic solvers and the exact g=0 H-function edge go beyond mere internal consistency. However, the central factorization is asserted rather than proved, and the incidence-angle dependence of P∞ and S is not stated. Because Eq. (14) is the main claim of the paper, these issues are load-bearing and require a major revision.

major comments (2)
  1. [Sec. 4.3, Eqs. (12)–(14)] The conditioning of S(n,τ) is undefined. For a reflected photon, the event 'n' must mean a first return to z=0 at scattering order n in the half-space, and S must be the conditional probability Pr(zmax<τ | first return at order n). If S is instead the maximum of an arbitrary n-step walk, Eq. (12) is false because most n-step walks from the boundary do not return to z=0. In addition, Eq. (14) carries no μ0 index, yet Table 2 and Fig. 5 show that R depends on incidence angle (e.g., at g=0.5, τ=4, R goes from 0.509 at θ0=0° to 0.661 at θ0=60°). The factorization must either be restricted to normal incidence (or a specified angle-averaged ensemble), or all quantities in Eqs. (12)–(14) must be written with μ0 indices. A proof by conditioning on the first-return event is needed; 'admits the decomposition' is not sufficient.
  2. [Sec. 4.3, Eq. (15)] The text states 'Steps are independent' immediately after writing 'HG memory E[μ_{i+k}|μ_i] = g^k μ_i'. These statements are contradictory: the direction cosines are Markov-correlated under the HG kernel. The variance formula itself is correct provided one uses E[μ_i μ_{i+k}] = g^k E[μ_i^2] and independence of step lengths from directions, but the sentence as written gives an incorrect justification for a formula that feeds the Brownian-excursion scaling of S(n,τ). Please correct the wording and show the intermediate step.
minor comments (4)
  1. [Eq. (11)] The absorptance expression contains a typo: the second term should be PT(n,τ), not PR(n,τ). Please also specify the n=0 term in the transmittance sum.
  2. [Table 1 caption / Sec. 4.4] The formula T≈1.68/(τ(1−g)+2z0) is introduced without derivation or a statement of whether the prefactor 1.68 is a fit parameter or a derived constant. Since this is presented as an approximation, please clarify its status.
  3. [Sec. 7] The sentence 'their agreement rules out any bias common to the Monte Carlo' is too strong. Adding-doubling and successive-orders are algorithmically distinct but solve the same scalar RT equation; the agreement is strong evidence against a common MC/RT bias, but it does not logically rule out every possible shared modeling assumption. Please soften the claim.
  4. [Eq. (15) notation] The notation σ_g^2 is used for a per-step variance but the displayed summation is over the correlation of successive increments; please define the quantity precisely (e.g., long-time variance per step).

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; minor self-citations and an under-specified conditioning in Eq. (12) are correctness concerns, not circular reductions.

full rationale

The central factorization, Eq. (14), is not circular: it is a path decomposition, not a definition. The paper defines S(n,τ) as Pr(zmax<τ|n), a depth-survival probability, and connects it to the Kennedy/Chung excursion-maximum law; it does not define S as PR/P∞. If S were fitted from the same PR(n,τ) data, Eq. (12) would be circular, but the text gives an independent probabilistic meaning and the RT and MC implementations both record excursion geometry. The MC and RT methods are derived from the same HG random-walk model, so their mutual agreement is an internal consistency check; the paper also validates against independent adding-doubling, successive-orders, and the Chandrasekhar H-function, so the numerical support is not solely self-referential. The self-citations [10,11] are present, and [11] is cited for the half-space first-return law and used in the thick-slab validation, but the factorization has independent content and is checked against external benchmarks. The main genuine issue is that Eq. (12)-(13) do not specify the conditioning event for S(n,τ) and drop the incidence cosine μ0; as written, S may be unconditional (making Eq. (12) generally false) or conditional on first-return order n (in which case it should carry a μ0 index). That is a correctness/proof gap, not circularity, because the claimed result is not equivalent to its inputs by construction.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

No parameters are fit to external measured data, but C(g) is measured from the same operator being tested, and the 1.68 prefactor appears fitted to the table. The central claim also rests on model assumptions (HG/Exp), on the azimuthal-averaging reduction, and on asymptotic Brownian-excursion theory.

free parameters (2)
  • C(g) — half-space first-return tail amplitude = C(0)=0.820 (exact); C(0.5)≈1.15, C(0.8)≈1.75 (from Fig. 2)
    Appears in Eq. (17) as the amplitude of the sqrt(1-a) edge. For g>0 it is not derived; it is read off the operator's large-n tail, so the 'prediction' 1-R(a) carries a numerically extracted coefficient.
  • 1.68 prefactor in T≈1.68/(tau(1−g)+2z0) = 1.68
    Stated in the Table 1 caption without derivation; appears fitted to the table entries rather than derived from the model.
assumptions (5)
  • domain assumption The Henyey-Greenstein phase function and exponential step-length model adequately describe slab photon transport.
    The entire paper solves this model; no comparison to measured physical reflectances is made.
  • domain assumption Averaging HG over azimuth yields a closed Markov process in (z, mu), with m=0 modes sufficient for R/T.
    Sec. 3.1 and 4.1; higher azimuthal modes are deferred to 'future work'.
  • standard math Sparre Andersen n^{-3/2} tail and Kennedy/Chung Brownian excursion maximum law apply to the discrete HG depth walk.
    Used in Sec. 4.3 to justify the universal edge and the form of S(n,tau); asymptotic and not proven for finite n.
  • domain assumption Thin-layer first-order scattering plus adding-doubling resummation converges to the exact transport solution as Delta_tau -> 0.
    Standard numerical method [4,14], but the paper does not report Delta_tau or convergence checks.
  • domain assumption At boundary crossings the direction-cosine distribution is the equilibrium flux law p(mu)=2|mu|.
    Used in Sec. 6 for angular distributions; assumes fully developed isotropic directions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Finite slab first passage statistics of Henyey Greenstein scattering." pith.science (2026). https://pith.science/paper/I5WMSFTE

@misc{pith2026260700290,
  author       = {Pith},
  title        = {Pith review of: Finite slab first passage statistics of Henyey Greenstein scattering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/I5WMSFTE}},
  note         = {Machine review of arXiv:2607.00290}
}
read the original abstract

A photon entering a plane parallel scattering slab performs a random walk and eventually escapes through one of the two faces or is absorbed. The standard model employs a Henyey Greenstein phase function (HG) and an exponential step length distribution (Exp). Slab reflectance, transmittance, absorptance, and emergent angular distributions can be calculated in terms of random walk statistics. A central result is that the slab calculations factor into the order resolved first passage statistics of a half space combined with the a factor for the slab thickness. Absorptance is derived from order resolved walk statistics using the absorption rate. Two approaches are used. In the Monte Carlo (MC) approach, an extremely long random walk with many steps is efficiently generated without regard to any boundaries. The intersection of this walk with a large collection of target objects creates an ensemble of excursions of the objects. The MC approach relies explicitly on the memoryless property of Exp so that the portion of the first and last steps inside the object follow the same length distribution as the walk steps. The details of each excursion are recorded and any statistics can be extracted from the database of excursions. In particular, first passage statistics are extracted from this ensemble. In this work the objects are slabs with different positions and thicknesses. In the radiative transfer (RT) approach the slab is divided into thin layers with scattering treated to first order in each layer. The RT equations are then directly integrated over the slab to give the desired first passage statistics, reflectance, transmittance, and absorptance. The two methods agree to the Monte Carlo precision over the tested range of random walk parameters.

Figures

Figures reproduced from arXiv: 2607.00290 by the authors.

Figure 1
Figure 1. R, T against single-scattering absorption at τ = 4. pz(µi , µi+1) = (1 − g 2 ) E(k) π (α − β) √ α + β , (5) where E(k) is the complete elliptic integral of the second kind and α = 1 + g 2 − 2g µiµi+1, β = 2g q (1 − µ 2 i )(1 − µ 2 i+1), k2 = 2β α + β . (6) For a layer so thin that a photon scatters at most once inside it, the reflection and transmission operators are linear in ∆τ . Writing the angular redistribution… view at source ↗
Figure 2
Figure 2. The conservative edge and the first-return tail. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Left: R(τ ) → 1 (conservative). Right: PR(n, τ ) → P∞(n) (g = 0), peeling off at high order. The half-space reflectance R(a) = P n≥1 P∞(n) a n rises to its conservative value R(1) = 1 as a → 1 −, and the manner of approach is fixed by the large-n tail of the first-return law. That tail carries the universal Sparre–Andersen exponent [8], P∞(n) ∼ C(g) n −3/2 . (15) Write the conservative deficit as 1 − R(a) = P n≥1 P∞… view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: R, T against optical thickness τ (a = 1, normal incidence). 5 Monte Carlo random walk method Our method uses a Monte Carlo random walk to estimate the statistics of excursions on objects in 3D space. We generate one or a few long random walks with exponentially distrib…
Figure 5
Figure 5. Figure 5: R, T against incidence cosine µ0 at τ = 4, a = 1. g = 0 g = 0.5 g = 0.8 τ R T R T R T 1 0.3413 0.6587 0.1761 0.8239 0.0600 0.9400 2 0.5175 0.4825 0.3203 0.6797 0.1272 0.8728 4 0.6909 0.3091 0.5090 0.4910 0.2547 0.7453 8 0.8218 0.1782 0.6890 0.3110 0.4416 0.5584 16 0.90…
Figure 6
Figure 6. Figure 6: Order-resolved reflection law PR(n, τ ) at τ = 4, a = 1. azimuthal angle. A quaternion that produces it is q(Θ, Φ) =  cos Θ 2 , sin Θ 2 ,(− sin Φ, cos Φ, 0) . The quaternion representing a direction is not unique. There is an additional “roll” about the direction of …
Figure 7
Figure 7. Figure 7: g=0.80 Spot position and displacement in reflection for a thick slab with short, medium [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 7
Figure 7. Figure 7: g=0.80 Spot position and displacement in reflection for a thick slab with short, medium [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: g=0.80 Point-spread-function in reflection showing the collapse to the power law for [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Radiation transfer and MC random walk comparison [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Left: thick-slab angular reflectance R(µ0) at g = 2/3 from the RT operator (lines) against the semi-infinite BTF/Motzkin formula of [11] (circles), for a = 0.95, 0.8, 0.6. Right: relative deviation at a = 0.8. 8 Discussion A single depth walk underlies the whole calcu…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

16 extracted references

  1. [1]

    J.9370–83

    Henyey L G and Greenstein J L 1941 Diffuse radiation in the galaxyAstrophys. J.9370–83

    1941

  2. [2]

    Chandrasekhar S 1960Radiative Transfer(New York: Dover)

  3. [3]

    Chandrasekhar S 1943 Stochastic problems in physics and astronomyRev. Mod. Phys.151–89

    1943

  4. [4]

    van de Hulst H C 1980Multiple Light Scattering: Tables, Formulas and Applicationsvol 1–2 (New York: Academic)

  5. [5]

    Opt.272502–9

    Stamnes K, Tsay S-C, Wiscombe W and Jayaweera K 1988 Numerically stable algorithm for discrete-ordinate-method radiative transfer in multiple scattering and emitting layered media Appl. Opt.272502–9

    1988

  6. [6]

    Opt.394195–204

    Melnikova I N, Dlugach Zh M, Nakajima T and Kawamoto K 2000 Calculation of the reflection function of an optically thick scattering layer for a Henyey–Greenstein phase functionAppl. Opt.394195–204

    2000

  7. [7]

    Express 3040968–90

    Libois Q and Davis A B 2022 Photon path distributions in optically thin slabsOpt. Express 3040968–90

    2022

  8. [8]

    Scand.1 263–85; 1954Math

    Sparre Andersen E 1953 On the fluctuations of sums of random variablesMath. Scand.1 263–85; 1954Math. Scand.2195–223

    1953

Show all 16 references
  1. [9]

    Redner S 2001A Guide to First-Passage Processes(Cambridge: Cambridge University Press)

  2. [10]

    Zeller C and Cordery R 2020 Light scattering as a Poisson process and first-passage probability J. Stat. Mech.2020063404

    2020

  3. [11]

    Zeller C and Cordery R 2026 First-return statistics in Henyey–Greenstein scattering: Motzkin polynomials and the Cauchy kernelJ. Stat. Mech.2026043206

    2026

  4. [12]

    Kennedy D P 1976 The distribution of the maximum Brownian excursionJ. Appl. Probab.13 371–6

    1976

  5. [13]

    Mat.14155–77

    Chung K L 1976 Excursions in Brownian motionArk. Mat.14155–77

    1976

  6. [14]

    Astrophys.183371–91

    de Haan J F, Bosma P B and Hovenier J W 1987 The adding method for multiple scattering calculations of polarized lightAstron. Astrophys.183371–91

    1987

  7. [15]

    Theory Stat

    Garcia R D M and Siewert C E 1985 Benchmark results in radiative transferTransp. Theory Stat. Phys.14437–83

    1985

  8. [16]

    Engler H 2015 Computation of scattering kernels in radiative transferJ. Quant. Spectrosc. Radiat. Transfer16538–42 13

    2015

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.