pith. sign in

arxiv: 2605.23587 · v1 · pith:CRWHUFWFnew · submitted 2026-05-22 · 🌌 astro-ph.IM

A Markov-Chain-Monte-Carlo-based Hybrid Noise Inference for Continuous Wavelet Power Spectra: with Applications to Solar and Stellar Oscillatory Signals

Song Feng , Lin Li , Ding Yuan This is my paper

Pith reviewed 2026-05-25 02:40 UTC · model grok-4.3

classification 🌌 astro-ph.IM
keywords MCMCcontinuous wavelet transformoscillation detectionsolar flaresred noiseBayesian inferencequasi-periodic oscillationstime series analysis
0
0 comments X

The pith

MCMC sampling of a time-varying power-law plus white-noise background enables local significance evaluation in wavelet spectra to detect oscillations without detrending.

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

The paper presents a Bayesian method that combines the continuous wavelet transform with Markov Chain Monte Carlo sampling to infer a time-dependent background spectrum modeled as a power-law plus white-noise component. Parameters of this background are allowed to vary smoothly in time, so significance levels for oscillations can be computed locally from the data. Synthetic tests demonstrate reliable recovery of injected oscillations and suppression of false detections in pure-noise cases. The approach identifies oscillations robustly when the frequency-domain signal-to-noise ratio reaches 2 or higher, and it yields improved temporal localization when applied to solar flare observations.

Core claim

The central claim is that MCMC sampling of parameters for a smoothly time-varying power-law plus white-noise background model, when applied within the continuous wavelet transform power spectrum, permits reliable local significance testing that recovers true oscillations while avoiding spurious detections, with robust performance above a frequency-domain S/N of 2.

What carries the argument

MCMC sampling of time-dependent power-law plus white-noise background parameters inside the continuous wavelet transform framework.

If this is right

  • Injected oscillations are recovered reliably in synthetic data tests.
  • False detections are suppressed in pure-noise cases.
  • Oscillations can be identified robustly when the frequency-domain S/N is greater than or equal to 2 under mixed noise conditions.
  • The detectable period range is limited by wavelet resolution, from about 3-4 sampling intervals up to roughly one-quarter of the total duration.
  • Application to GOES soft X-ray flare observations isolates quasi-periodic oscillations with improved temporal localization compared to standard wavelet and Fourier-based approaches.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The method could extend to other non-stationary time series containing red noise, such as geophysical or biological signals.
  • It implies that many traditional detrending steps in periodicity searches can be replaced by direct background inference.
  • Performance on additional stellar datasets might reveal how the approach affects detection rates in asteroseismology.

Load-bearing premise

The background spectrum is adequately described by a power-law plus white-noise component whose parameters vary smoothly enough in time that MCMC sampling can recover them without explicit detrending.

What would settle it

A large fraction of false positive detections or missed injected signals when the method is tested on synthetic data whose true background deviates strongly from a smooth power-law plus white noise.

Figures

Figures reproduced from arXiv: 2605.23587 by Ding Yuan, Lin Li, Song Feng.

Figure 1
Figure 1. Figure 1: Flowchart of the proposed CWT+MCMC frame￾work for wavelet power spectrum estimation and significance testing. The main steps include wavelet transformation, time-dependent background modeling, and Bayesian infer￾ence of significance levels. An overview of the proposed methodology is illus￾trated in [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Synthetic test with a simple signal. (a)–(d) Individual components: oscillatory signal, flare background, red + white noise. (e) Wavelet power spectrum from the standard CWT+AR(1) method. (f) Result after detrending (80 s) followed by CWT+AR(1) analysis. (g) Result from the CWT+MCMC method. Black contours indicate the 95% confidence level. g) identifies both oscillatory components over their full duration,… view at source ↗
Figure 3
Figure 3. Figure 3: Synthetic signals with multiple non-stationary oscillatory components. Each column corresponds to an independent realization with the same background but different oscillation periods. From top to bottom: (1) background components (red + white noise with flare profile); (2) injected oscillations with time-varying amplitudes and periods; (3) composite signals; (4) global Fourier power spectra (FFT+MCMC); (5… view at source ↗
Figure 4
Figure 4. Figure 4: Detection results for different signal-to-noise ratios (SNRs) and oscillation periods. Three representative periods are considered: 22 s (top), 60 s (middle), and 150 s (bottom). Columns correspond to SNR = 0.5, 1, and 2. Black contours indicate regions above the 95% confidence level. 00:15 00:20 00:25 00:30 00:35 Start time 2020-01-01 00:15 UT 2 1 0 1 2 3 4 Intensity ×10 8 (a) GOES 1-8Å 1024 512 256 128 6… view at source ↗
Figure 5
Figure 5. Figure 5: GOES 1–8 ˚A light curve for a quiet-Sun interval. (a) Observed soft X-ray flux. (b) Fourier power spec￾trum (FFT+MCMC). (c) Wavelet power spectrum from the CWT+AR(1) method. (d) Wavelet power spectrum from the CWT+MCMC method. A localized feature near ∼128 s is present in panel (c) but not in the other representations. Black contours indicate the 95% confidence level [PITH_FULL_IMAGE:figures/full_fig_p007… view at source ↗
Figure 6
Figure 6. Figure 6: GOES 1–8 ˚A light curve for a flare with quasi-periodic pulsations. (a) Observed soft X-ray flux. (b) Fourier power spectrum (FFT+MCMC). (c) Wavelet power spectrum from the CWT+AR(1) method. (d) Wavelet power spectrum from the CWT+MCMC method. Enhanced power near ∼80 s is visible in the Fourier spectrum, and a localized component at a similar period appears in the time–frequency domain. Black contours indi… view at source ↗
Figure 7
Figure 7. Figure 7: Detection results for a GOES soft X-ray flare without evident quasi-periodic pulsations (QPPs). (a) Observed 1–8 ˚A light curve showing a typical flare profile without clear oscillatory signatures. (b) FFT+MCMC: the Fourier spectrum does not exhibit distinct periodic components. (c) CWT+AR(1): no statistically significant oscillatory power is detected, indicating that the method does not produce false posi… view at source ↗
Figure 8
Figure 8. Figure 8: Fourier power spectra from representative GOES 1–8 ˚A time series segments. The spectra are fitted with the model P(f) = Af −α + C. A power-law trend is present at low frequencies, with a gradual flattening at higher frequencies. The fitted spectral indices lie in the range α ∼ 2.5–4.5 [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
read the original abstract

Detecting oscillations in solar and stellar time series is complicated by non-stationary red noise and evolving background emission. Methods based on detrending and AR(1)-based wavelet analysis can introduce spurious periodicities and do not adequately describe time-dependent backgrounds. We develop a Bayesian approach that combines the continuous wavelet transform with MCMC sampling to infer a time-dependent background spectrum. The background is represented by a power-law plus white-noise component, with parameters allowed to vary smoothly in time, so that significance levels can be evaluated locally without explicit detrending. Tests with synthetic data show that injected oscillations are recovered reliably, while false detections are suppressed in pure-noise cases. Using a frequency-domain signal-to-noise ratio (S/N), we find that oscillations can be identified robustly when the S/N is greater than or equal to 2 under mixed noise conditions. The detectable period range is limited by wavelet resolution, from about 3-4 sampling intervals up to roughly one-quarter of the total duration. Application to GOES soft X-ray flare observations shows that the method isolates quasi-periodic oscillations with improved temporal localization compared to standard wavelet and Fourier-based approaches. Meanwhile, this behavior is consistent across a range of noise conditions and signal morphologies.

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

2 major / 2 minor

Summary. The manuscript develops a Bayesian hybrid method that pairs the continuous wavelet transform with MCMC sampling to infer a time-dependent background spectrum modeled as a smoothly varying power-law plus white-noise component. This allows local significance assessment for oscillatory signals without explicit detrending. Synthetic tests are reported to recover injected oscillations reliably while suppressing false detections in pure-noise realizations, with a claimed robust detection threshold at frequency-domain S/N >= 2; the method is then applied to GOES soft X-ray flare data to demonstrate improved temporal localization of quasi-periodic oscillations relative to standard wavelet and Fourier approaches.

Significance. If the validation were robust, the approach would address a recognized limitation of AR(1) and detrending-based wavelet methods for non-stationary red noise in solar and stellar time series. The provision of reproducible synthetic benchmarks and explicit S/N thresholds would be a strength, but the current tests only confirm internal consistency under the exact generative model assumed by the inference.

major comments (2)
  1. [Abstract / synthetic tests] Abstract and synthetic-data tests: the reported recovery of injected signals and suppression of false positives at S/N >= 2 is demonstrated only on data generated from the identical power-law-plus-white-noise model (with the same smoothness hyperparameters) that the MCMC is asked to recover. This establishes internal consistency but supplies no evidence on performance when the true background contains unmodeled structure (broken power laws, additional Lorentzian components, or non-smooth temporal evolution) that is common in solar and stellar observations.
  2. [Abstract] Abstract: the central claim that oscillations can be identified robustly for frequency-domain S/N >= 2 under mixed noise conditions lacks quantitative support (recovery fractions, false-positive rates, dependence on hyperparameter choices, or error propagation). Without these metrics or checks against post-hoc parameter tuning, the S/N threshold cannot be evaluated as a general result.
minor comments (2)
  1. [Methods] The description of the MCMC sampling procedure and the precise definition of the frequency-domain S/N metric should be expanded with explicit equations and pseudocode to allow independent reproduction.
  2. [Application to GOES data] The manuscript should clarify the data-exclusion rules and any preprocessing steps applied to the GOES observations before wavelet analysis.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major point below and indicate the revisions we will make.

read point-by-point responses

  1. Referee: [Abstract / synthetic tests] Abstract and synthetic-data tests: the reported recovery of injected signals and suppression of false positives at S/N >= 2 is demonstrated only on data generated from the identical power-law-plus-white-noise model (with the same smoothness hyperparameters) that the MCMC is asked to recover. This establishes internal consistency but supplies no evidence on performance when the true background contains unmodeled structure (broken power laws, additional Lorentzian components, or non-smooth temporal evolution) that is common in solar and stellar observations.

    Authors: We agree that the synthetic tests establish internal consistency under the assumed generative model but do not directly address robustness to unmodeled background structure. In the revised manuscript we will add a new subsection presenting additional synthetic tests that incorporate broken power-law backgrounds and non-smooth temporal variations. We will also insert a dedicated limitations paragraph that explicitly discusses the model assumptions and the potential effects of unmodeled components on detection performance. The abstract will be updated to qualify the scope of the reported S/N threshold. revision: yes

  2. Referee: [Abstract] Abstract: the central claim that oscillations can be identified robustly for frequency-domain S/N >= 2 under mixed noise conditions lacks quantitative support (recovery fractions, false-positive rates, dependence on hyperparameter choices, or error propagation). Without these metrics or checks against post-hoc parameter tuning, the S/N threshold cannot be evaluated as a general result.

    Authors: The abstract summarizes results obtained from the synthetic tests, yet we acknowledge that it does not supply the supporting quantitative metrics. In the revision we will expand the abstract to report explicit recovery fractions and false-positive rates at the S/N = 2 threshold. We will also move a table of these metrics, together with a brief assessment of hyperparameter sensitivity, into the main text and clarify that the threshold is derived under the tested conditions rather than asserted as universally applicable. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper presents a standard Bayesian MCMC procedure to infer parameters of a power-law plus white-noise background model whose time variation is assumed smooth, then uses those parameters to set local significance thresholds in the CWT power spectrum. Synthetic-data recovery tests generate realizations from the identical functional form that the sampler is asked to recover; this is ordinary internal-consistency validation and does not equate any reported detection threshold or significance level to a quantity defined from the same fitted values by construction. No self-citations, uniqueness theorems, or ansatzes are invoked as load-bearing steps. The central procedure therefore remains self-contained against the external synthetic benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 0 axioms · 0 invented entities

Abstract-only review; the background model implicitly treats the power-law index, white-noise amplitude, and their time-smoothness hyperparameters as quantities to be sampled, but no explicit count or values are given.

free parameters (1)
  • time-smoothness hyperparameters for background parameters
    Parameters are allowed to vary smoothly in time and are inferred via MCMC; their specific functional form or regularization strength is not stated.

pith-pipeline@v0.9.0 · 5757 in / 1233 out tokens · 24779 ms · 2026-05-25T02:40:42.809469+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

15 extracted references · 15 canonical work pages

  1. [1]

    2016, ApJ, 825, 110, doi: 10.3847/0004-637X/825/2/110

    Solomon, J. 2016, ApJ, 825, 110, doi: 10.3847/0004-637X/825/2/110

    work page doi:10.3847/0004-637x/825/2/110 2016

  2. [2]

    2011, Applied and computational harmonic analysis, 30, 243

    Daubechies, I., Lu, J., & Wu, H.-T. 2011, Applied and computational harmonic analysis, 30, 243

    work page 2011

  3. [3]

    H., Zhang, X

    Deng, L. H., Zhang, X. J., Li, G. Y., Deng, H., & Wang, F. 2019, MNRAS, 488, 111, doi: 10.1093/mnras/stz1653

    work page doi:10.1093/mnras/stz1653 2019

  4. [4]

    2018, SoPh, 293, 61, doi: 10.1007/s11207-018-1281-x Garc´ ıa, R

    Lapenta, G. 2018, SoPh, 293, 61, doi: 10.1007/s11207-018-1281-x Garc´ ıa, R. A., & Ballot, J. 2019, Living Reviews in Solar Physics, 16, 4, doi: 10.1007/s41116-019-0020-1

    work page doi:10.1007/s11207-018-1281-x 2018

  5. [5]

    1983, Nature, 304, 689, doi: 10.1038/304689a0

    Gough, D. 1983, Nature, 304, 689, doi: 10.1038/304689a0

    work page doi:10.1038/304689a0 1983

  6. [6]

    2023, ApJ, 944, 16, doi: 10.3847/1538-4357/acb34f

    Guo, Y., Liang, B., Feng, S., et al. 2023, ApJ, 944, 16, doi: 10.3847/1538-4357/acb34f

    work page doi:10.3847/1538-4357/acb34f 2023

  7. [7]

    E., Shen, Z., Long, S

    Huang, N. E., Shen, Z., Long, S. R., et al. 1998, Proceedings of the Royal Society of London. Series A: mathematical, physical and engineering sciences, 454, 903

    work page 1998

  8. [8]

    2016, ApJ, 833, 284, doi: 10.3847/1538-4357/833/2/284

    Gallagher, P. 2016, ApJ, 833, 284, doi: 10.3847/1538-4357/833/2/284

    work page doi:10.3847/1538-4357/833/2/284 2016

  9. [9]

    R., Ireland, J., & Dominique, M

    Inglis, A. R., Ireland, J., & Dominique, M. 2015, ApJ, 798, 108, doi: 10.1088/0004-637X/798/2/108

    work page doi:10.1088/0004-637x/798/2/108 2015

  10. [10]

    2015, Living Reviews in Solar Physics, 12, 6

    Khomenko, E., & Collados, M. 2015, Living Reviews in Solar Physics, 12, 6

    work page 2015

  11. [11]

    2020, Ap&SS, 365, 40, doi: 10.1007/s10509-020-3744-4 L´ opez-Santiago, J

    Liang, B., Meng, Y., Feng, S., & Yang, Y. 2020, Ap&SS, 365, 40, doi: 10.1007/s10509-020-3744-4 L´ opez-Santiago, J. 2018, Philosophical Transactions of the Royal Society of London Series A, 376, 20170253, doi: 10.1098/rsta.2017.0253

    work page doi:10.1007/s10509-020-3744-4 2020

  12. [12]

    Torrence, C., & Compo, G. P. 1998, Bulletin of the American Meteorological Society, 79, 61, doi: 10.1175/1520-0477(1998)079⟨0061: APGTWA⟩2.0.CO;2 Van Doorsselaere, T., Kupriyanova, E. G., & Yuan, D. 2016, SoPh, 291, 3143, doi: 10.1007/s11207-016-0977-z

    work page doi:10.1175/1520-0477(1998)079 1998

  13. [13]

    2018, ApJL, 856, L16, doi: 10.3847/2041-8213/aab4f3

    Wang, F., Deng, H., Li, B., et al. 2018, ApJL, 856, L16, doi: 10.3847/2041-8213/aab4f3

    work page doi:10.3847/2041-8213/aab4f3 2018

  14. [14]

    2024, ApJ, 961, 231, doi: 10.3847/1538-4357/ad10a9 12

    Wang, Y., Fletcher, L., Mulay, S., Ji, H., & Cao, W. 2024, ApJ, 961, 231, doi: 10.3847/1538-4357/ad10a9 12

    work page doi:10.3847/1538-4357/ad10a9 2024

  15. [15]

    2026, A Novel Continuous Wavelet Method with A Markov- Chain-Monte-Carlo-based Hybrid Noise Inference, v1, Zenodo, doi: 10.5281/zenodo.19446731

    Yuan, D., & Feng, S. 2026, A Novel Continuous Wavelet Method with A Markov- Chain-Monte-Carlo-based Hybrid Noise Inference, v1, Zenodo, doi: 10.5281/zenodo.19446731

    work page doi:10.5281/zenodo.19446731 2026