pith. sign in

arxiv: 2604.06654 · v1 · submitted 2026-04-08 · 🌌 astro-ph.HE

A Hidden Degeneracy in Two-Spot Models for Thermal X-Ray Pulse-Profile Fitting

Tong Zhao , Mingyu Ge , Renxin Xu This is my paper

Pith reviewed 2026-05-10 18:12 UTC · model grok-4.3

classification 🌌 astro-ph.HE
keywords neutron starsX-ray pulsarspulse profile modelingparameter degeneracyradius measurementNICERPSR J0030+0451multi-modal inference
0
0 comments X

The pith

A hidden degeneracy in two-spot models for thermal X-ray pulse-profile fitting creates multi-modal likelihoods that can bias neutron star radius by 30%.

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

The paper discovers an intrinsic degeneracy in the semi-analytic two-spot model for thermal X-ray pulse profiles from neutron stars. This degeneracy, present in a simplified version with two small circular hot spots and no Doppler effects, leads to multi-modal likelihood surfaces in more complex models that include Doppler effects. With insufficient priors, the posterior can be multi-modal, biasing the inferred neutron star radius by up to 30%. It also explains multi-modal structures in synthetic data recovery for PSR J0030+0451. This has implications for reanalyzing existing NICER data and upcoming eXTP observations.

Core claim

We discover an intrinsic degeneracy in the semi-analytic two-spot model for parameter inference in thermal X-ray pulse-profile modeling. Although this degeneracy exists in our simplified model with two small circular hot spots and without Doppler effects, it causes the likelihood surface of parameter inference based on ST-U and more complex models with Doppler effects to have multi-modal structures. Consequently, the posterior surface may also exhibit multi-modal structures if there is insufficient prior knowledge of parameters. Because of this, the inferred value of the neutron star radius can be biased even by 30%. This finding also provides a promising way to explain the multi-modal ures.

What carries the argument

the intrinsic degeneracy in the semi-analytic two-spot model with two small circular hot spots and without Doppler effects, which produces multiple parameter sets yielding similar pulse profiles

Load-bearing premise

That the degeneracy found in the simplified two-spot model without Doppler effects is the same mechanism producing multi-modal structures in the ST-U and Doppler-inclusive models, and that it directly explains the multi-modal recovery performance in synthetic data for PSR J0030+0451.

What would settle it

A calculation showing the exact parameter trade-offs in the simplified model and checking if they correspond to the locations of multiple modes in the full model's likelihood surface.

Figures

Figures reproduced from arXiv: 2604.06654 by Mingyu Ge, Renxin Xu, Tong Zhao.

Figure 1
Figure 1. Figure 1: The geometry sketch for ST-U and our semi-analytic two￾spot model. ˆk is the unit vector pointing from the center of the neutron star to the observer. ˆk0 is the moment vector of the photon emitted from the hot spot. nˆ is the surface normal at the center of the hot spot. α is the emission angle, the angle between ˆk0 and nˆ. θ is the observer inclination angle, the angle between ˆk and the z￾axis. The z-a… view at source ↗
Figure 2
Figure 2. Figure 2: The corner plot of two results with the same model fitting to the same synthetic data. We run the MCMC twice starting from two initial values to sample and end up with two peaks on the posterior surface. The first best-fit point is close to our assumed value but the second one is completely different. 68-th and 95-th percentile contours for the first run are shown in red and blue, and for the second run in… view at source ↗
Figure 3
Figure 3. Figure 3: Geometry configurations of the two modes corresponding to the two peaks on the posterior surface. The z-axis is the assumed spin axis of the neutron star. Position of two spots are presented by two red points, labeled as S1 and S2, and a vector is pointing from the neutron star center to the observer. algorithm for MCMC sampling, the Markov chains will converge to the wrong best-fit point if the initial va… view at source ↗
Figure 4
Figure 4. Figure 4: The corner plot of two results with the same model fitting to the same synthetic data. We run the MCMC twice starting from two initial values to sample and end up with two peaks on the posterior surface. The two best-fit values for the radius are widely separated. The first best-fit point is close to our assumed value while the second one is close to the predicted degenerate solution. 68-th and 95-th perce… view at source ↗
Figure 5
Figure 5. Figure 5: Contours of 68-th and 95-th percentiles of the posterior distribution for parameter fits to pulse-profiles with different frequencies. Red and blue contours for f = 400 Hz, orange and green contours for f = 500 Hz, and yellow and black contours for f = 600 Hz [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Contours of 68-th and 95-th percentiles of the posterior distribution for parameter fits to pulse-profiles with different total number of photons. Red and blue contours for A = 2, orange and green contours for A = 4 [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
read the original abstract

We discover an intrinsic degeneracy in the semi-analytic two-spot model for parameter inference in thermal X-ray pulse-profile modeling. Although this degeneracy exists in our simplified model with two small circular hot spots and without Doppler effects, it causes the likelihood surface of parameter inference based on ST-U and more complex models with Doppler effects to have multi-modal structures. Consequently, the posterior surface may also exhibit multi-modal structures if there is insufficient prior knowledge of parameters. Because of this, the inferred value of the neutron star radius can be biased even by $30\%$. This finding also provides a promising way to explain the multi-modal structures discovered in the evaluation of recovery performance using synthetic pulse-profiles that mimic the PSR J0030+0451 pulse-profiles~\citep{Vinciguerra2023,Vinciguerra2024}. Our work may have profound implications for the reanalysis of NICER data and the analysis of upcoming eXTP data.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper identifies an intrinsic degeneracy in a semi-analytic two-spot model for thermal X-ray pulse-profile fitting that assumes two small circular hot spots and omits Doppler effects. It asserts that this degeneracy induces multi-modal structures in the likelihood surfaces of the ST-U model and more complex variants that include Doppler boosting, which in turn can produce multi-modal posteriors and bias the inferred neutron star radius by as much as 30 percent; the same mechanism is proposed to explain multi-modal recovery performance in synthetic data mimicking PSR J0030+0451.

Significance. If the claimed continuity of the degeneracy across model complexities is demonstrated, the result would be important for NICER and eXTP analyses because it would flag a model-intrinsic source of systematic bias in radius measurements that could affect neutron-star equation-of-state constraints. The work usefully draws attention to the need for careful prior specification when fitting pulse profiles.

major comments (3)
  1. [Abstract and §4] The central claim that the degeneracy found in the simplified (no-Doppler, circular-spot) model produces multi-modal likelihood surfaces in the ST-U and Doppler-inclusive models is asserted in the abstract and §4 but is not supported by an explicit parameter mapping, analytic continuation, or side-by-side likelihood-surface comparison. Without this bridge the extrapolation remains unverified.
  2. [Abstract and §4] The quantitative statement that the neutron-star radius can be biased by 30% is presented as a direct consequence of the degeneracy, yet no specific numerical example, recovered parameter values, or likelihood-surface slice demonstrating this magnitude of bias is supplied.
  3. [§5] The explanation offered for the multi-modal recovery performance in the Vinciguerra et al. (2023, 2024) synthetic-data tests for PSR J0030+0451 relies on the same unproven continuity between the simplified-model degeneracy and the full ST-U + Doppler model; a direct comparison of the recovered modes would be required to substantiate this link.
minor comments (2)
  1. [§2] Notation for the spot parameters (e.g., colatitude, azimuth, and size) is introduced without a consolidated table; a single reference table would improve readability.
  2. [Figure 3] Figure captions do not state whether the plotted likelihood surfaces are marginalized or conditional; this should be clarified.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments. We agree that making the continuity between the simplified-model degeneracy and the behavior of the ST-U and Doppler-inclusive models more explicit will strengthen the manuscript. We address each major comment below and will incorporate the requested additions in the revised version.

read point-by-point responses

  1. Referee: [Abstract and §4] The central claim that the degeneracy found in the simplified (no-Doppler, circular-spot) model produces multi-modal likelihood surfaces in the ST-U and Doppler-inclusive models is asserted in the abstract and §4 but is not supported by an explicit parameter mapping, analytic continuation, or side-by-side likelihood-surface comparison. Without this bridge the extrapolation remains unverified.

    Authors: We acknowledge that an explicit bridge would improve clarity. The manuscript already demonstrates the degeneracy analytically in the simplified model and separately exhibits multi-modal likelihood surfaces in the ST-U model; however, we will add a dedicated subsection (or appendix) that provides an explicit parameter mapping between the two-spot configurations and the corresponding modes in the ST-U likelihood, together with side-by-side likelihood-surface slices for a representative set of parameters. revision: yes

  2. Referee: [Abstract and §4] The quantitative statement that the neutron-star radius can be biased by 30% is presented as a direct consequence of the degeneracy, yet no specific numerical example, recovered parameter values, or likelihood-surface slice demonstrating this magnitude of bias is supplied.

    Authors: The stated 30% bias follows from the separation of the two posterior modes under broad priors. In the revision we will insert a concrete numerical example in §4, showing the two recovered radius values from a synthetic dataset, the corresponding likelihood slice, and the fractional bias relative to the input radius. revision: yes

  3. Referee: [§5] The explanation offered for the multi-modal recovery performance in the Vinciguerra et al. (2023, 2024) synthetic-data tests for PSR J0030+0451 relies on the same unproven continuity between the simplified-model degeneracy and the full ST-U + Doppler model; a direct comparison of the recovered modes would be required to substantiate this link.

    Authors: We will expand §5 with a direct comparison that maps the parameter values of the multi-modal solutions recovered by Vinciguerra et al. onto the degenerate modes identified in our simplified model, thereby substantiating the proposed connection. revision: yes

Circularity Check

0 steps flagged

No significant circularity; degeneracy reported as intrinsic model property

full rationale

The paper identifies a degeneracy as an intrinsic feature of the semi-analytic two-spot model equations (small circular spots, no Doppler) and asserts that this feature induces multi-modal likelihood surfaces in ST-U and Doppler-inclusive models. No step fits a parameter to data and then renames the fit as a prediction, defines one quantity in terms of another by construction, or relies on a self-citation chain for a uniqueness theorem or ansatz. The central claim about consequences for radius inference and PSR J0030+0451 recoveries is an extrapolation from the simplified case rather than a self-referential loop, leaving the derivation chain self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The abstract introduces no new free parameters, axioms, or invented entities; the work consists of analysis of existing semi-analytic two-spot models.

pith-pipeline@v0.9.0 · 5463 in / 1295 out tokens · 81568 ms · 2026-05-10T18:12:55.635630+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

16 extracted references · 16 canonical work pages

  1. [1]

    2024, ApJL, 971, L20

    Choudhury, D., Salmi, T., Vinciguerra, S., et al. 2024, ApJL, 971, L20

    work page 2024

  2. [2]

    J., Miller, M

    Dittmann, A. J., Miller, M. C., Lamb, F. K., et al. 2024, ApJ, 974, 295

    work page 2024

  3. [3]

    C., Arzoumanian, Z., Adkins, P

    Gendreau, K. C., Arzoumanian, Z., Adkins, P. W., et al. 2016, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, V ol. 9905, Space Telescopes and Instrumentation 2016: Ultraviolet to Gamma Ray, ed. J.-W. A. den Herder, T. Takahashi, & M. Bautz, 99051H

    work page 2016

  4. [4]

    L., Zhang, G., et al

    Li, A., Watts, A. L., Zhang, G., et al. 2025, Science China Physics, Mechanics, and Astronomy, 68, 119503

    work page 2025

  5. [5]

    H., Miller, M

    Lo, K. H., Miller, M. C., Bhattacharyya, S., & Lamb, F. K. 2013, ApJ, 776, 19

    work page 2013

  6. [6]

    C., Lamb, F

    Miller, M. C., Lamb, F. K., Dittmann, A. J., et al. 2019, ApJL, 887, L24 —. 2021, ApJL, 918, L28

    work page 2019

  7. [7]

    Poutanen, J., & Beloborodov, A. M. 2006, MNRAS, 373, 836

    work page 2006

  8. [8]

    2014, ApJ, 787, 136

    Psaltis, D., Özel, F., & Chakrabarty, D. 2014, ApJ, 787, 136

    work page 2014

  9. [9]

    2025, ApJ, 981, 99

    Qi, L., Zheng, S., Zhang, J., et al. 2025, ApJ, 981, 99

    work page 2025

  10. [10]

    E., Watts, A

    Riley, T. E., Watts, A. L., Bogdanov, S., et al. 2019, ApJL, 887, L21

    work page 2019

  11. [11]

    E., Watts, A

    Riley, T. E., Watts, A. L., Ray, P. S., et al. 2021, ApJL, 918, L27

    work page 2021

  12. [12]

    Salmi et al.,The Radius of the High-mass Pulsar PSR J0740+6620 with 3.6 yr of NICER Data,Astrophys

    Salmi, T., Choudhury, D., Kini, Y ., et al. 2024b, arXiv e-prints, arXiv:2406.14466

    work page arXiv

  13. [13]

    O., & Pappas, G

    Sotani, H., Silva, H. O., & Pappas, G. 2019, PhRvD, 100, 043006

    work page 2019

  14. [14]

    L., et al

    Vinciguerra, S., Salmi, T., Watts, A. L., et al. 2023, ApJ, 959, 55 —. 2024, ApJ, 961, 62

    work page 2023

  15. [15]

    2025, ApJ, 982, 112

    Zhao, T., Psaltis, D., & Özel, F. 2025, ApJ, 982, 112

    work page 2025

  16. [16]

    Zhao, T., Psaltis, D., Ozel, F., & Beklen, E. 2024, arXiv:2412.12283 9 R = 13.391+0.101 0.102 66 72 78 84 = 69.206+1.640 1.404 40 60 80 100 120 1 1 = 115.319+1.528 1.515 135 140 145 150 155 2 2 = 140.777+1.521 1.763 31 30 29 28 27 = 27.902+0.492 0.473 1.80 1.95 2.10 2.25 ar ar = 2.054+0.086 0.074 0.135 0.140 0.145 0.150 T1 T1 = 0.150+0.000 0.000 0.125 0.1...

    work page arXiv 2024