Pith. sign in

REVIEW 4 major objections 3 minor 48 references

Low frequency X-ray timing with Gaussian processes and reverberation in the radio-loud AGN 3C 120

T0 review · 4 major / 3 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Gaussian-process timing recovers a 3,800-second X-ray echo from the inner disc of the radio-loud AGN 3C 120.

desk verdict Solid GP-based timing method with a plausible but not fully secured reverberation claim in 3C 120; worth refereeing with a requested hard-lag comparison. read the letter →

arxiv 1908.06099 v1 pith:WECM4PCD submitted 2019-08-16 astro-ph.HE

D.R. Wilkins This is my paper · ORCID
classification astro-ph.HE
keywords X-rayreverberationGaussianprocessesactivegalacticnucleiaccretiondiscironKlinetimelags3C120corona
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 claims that X-ray timing analysis can be pushed below the frequency floor set by orbital gaps by modelling each energy-band light curve as a Gaussian process, drawing continuous sample light curves, and computing Fourier-domain lags on those samples. It applies the method to two XMM-Newton orbits of the radio-loud AGN 3C 120 and reports a lag-energy spectrum between $5\times10^{-6}$ and $3\times10^{-5}$ Hz whose shape matches X-ray reverberation from the inner accretion disc. The 5–7 keV iron K line peak lags the 1.6–2 keV continuum-dominated band by $(3800\pm1500)$ s, interpreted as light-travel delay between a corona and the disc, placing the corona at $(13\pm8)\,r_g$ above the disc. If the interpretation holds, this is the first reverberation measurement in a radio-loud AGN, and it opens low-frequency reverberation studies to low-Earth-orbit missions and higher-mass black holes.

What carries the argument

The central mechanism is a Gaussian process model of each energy-band light curve: the data are drawn from a multivariate Gaussian whose covariance is set by a kernel function, and after optimising the kernel hyperparameters, continuous light-curve realisations are drawn from the conditional distribution of unobserved times given the observed points. The paper uses the rational quadratic kernel, a mixture of squared-exponential correlations on many timescales, which reproduces the broken power-law power spectral density of AGN variability better than a single-scale kernel. It then computes the cross-spectrum, coherence, lag-frequency spectrum, and lag-energy spectrum on each pair of sample light curves and averages over thousands of samples, so the gaps are handled probabilistically rather than by interpolation.

What would settle it

Fit a quantitative hard-lag model in which the lag increases monotonically with energy to the measured $0.3$–$10$ keV lag-energy spectrum between $5\times10^{-6}$ and $3\times10^{-5}$ Hz; if such a model matches the data without an iron K peak, the reverberation interpretation is not unique. A longer observation that resolves the sharp drop above 7 keV and the rise below 1.6 keV would also test whether the claimed iron K line profile in the lag is real.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that a Gaussian process fitted independently to each X-ray energy band can preserve the phase relationship between bands across gaps, so that standard Fourier-domain lag measurements can be made at frequencies as low as $5\times10^{-6}$ Hz from concatenated orbit segments. Applied to the 2013 XMM-Newton observations of 3C 120, the method yields a lag-frequency spectrum in which the 4–7 keV band lags the 1.2–4 keV band by $(6300\pm2700)$ s at $7.5\times10^{-6}$ Hz, and a lag-energy spectrum over $5\times10^{-6}$ to $3\times10^{-5}$ Hz with the characteristic iron K profile: the continuum-dominated 1.6–2 keV band responds earliest, the 5–7 keV core of the line responds latest, and the redshifted wing at 3–5 keV responds between them. The author interprets the $(3800\pm1500)$ s lag between the 5–7 keV and 1.6–2 keV bands as X-ray reverberation from the inner disc, detected at 99.5 per cent confidence, and converts it to a coronal scale height of $(13\pm8)\,r_g$ above the disc.

Load-bearing premise

The identification of the lag-energy profile as disc reverberation rather than a continuum hard lag rests on its shape alone; the paper does not fit a quantitative hard-lag model to rule that alternative out.

Editorial extensions

If this is right

  • Low-frequency reverberation can be measured from light curves with gaps, so satellites in low-Earth orbit such as NICER, NuSTAR, and the proposed STROBE-X can perform reverberation experiments without continuous visibility.
  • Combining successive orbits extends reverberation studies to more massive AGN whose lags appear at lower Fourier frequencies than a single orbit can sample.
  • In 3C 120, the coronal height of $(13\pm8)\,r_g$ sits at the upper end of the iron K lag sample for Seyferts, suggesting that a jet-associated corona may be more extended than in radio-quiet AGN.
  • Comparing lag-energy profiles between radio-loud and radio-quiet AGN can constrain whether a collimated coronal core can coexist with a jet.

Reading between the lines

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

  • A natural extension the paper does not pursue is applying the same Gaussian-process sampling procedure to covariance spectra and bispectra, which the paper notes are straightforward products of the same sample draws; this would let gap-ridden data probe non-linear variability and the log-normal flux distribution.
  • If the 86.3 per cent-confidence hint that the redshifted 3–4 keV wing does not lead the 1–2 keV continuum is confirmed by longer exposures, the lag-energy morphology would favour an extended disc-hugging corona over a slowly propagating jet-base fluctuation source in 3C 120.
  • The simulation results imply that at STROBE-X count rates the systematic bias near the orbital frequency (1/5700 s) becomes the limiting error, so a kernel matched more exactly to the power spectral density should be developed before the method is used on very high signal-to-noise data.
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

4 major / 3 minor

Summary. This paper introduces a Gaussian-process framework for Fourier-domain X-ray timing on light curves with orbital gaps. The method fits a kernel (rational quadratic or Matern-1/2) to each energy-band light curve, draws continuous sample light curves, and computes lag-frequency and lag-energy spectra from the samples. Simulations with injected lags validate that phase relationships and lag amplitudes are recovered with known systematic and statistical errors, including for low-Earth-orbit geometries and for long XMM-Newton orbits. The framework is applied to three 2013 XMM-Newton segments of the radio-loud AGN 3C 120. The paper reports a lag-frequency spectrum in which the 4-7 keV band lags the 1.2-4 keV band at low frequencies, and a lag-energy spectrum over 5x10^-6 to 3x10^-5 Hz whose shape is interpreted as reverberation from the inner disc. The 5-7 keV iron-K band is found to lag the 1.6-2 keV continuum by (3800 +/- 1500) s, implying a coronal height of (13 +/- 8) rg. The claimed physical result is a first reverberation measurement in a radio-loud AGN.

Significance. The Gaussian-process timing method is a potentially valuable contribution: it is implemented in a public package (pylag), validated on simulations with known lags, includes a null test for zero lag, and marginalizes over hyperparameter uncertainties with MCMC. If the 3C 120 result is correct, it extends reverberation measurements to higher-mass radio-loud AGN and provides a direct probe of the corona-jet connection. However, the astrophysical claim currently rests on a qualitative exclusion of the hard-lag continuum interpretation, a post-hoc frequency-band choice, and a significance measure that is internal to the GP posterior rather than calibrated against a null hypothesis. These issues are fixable but are load-bearing for the central conclusion.

major comments (4)
  1. [Section 5.3, Fig. 11] The dismissal of the continuum hard-lag alternative is qualitative rather than quantitative. The sentence 'A hard lag within the continuum emission itself would have produced a lag-energy spectrum smoothly increasing with energy which is not seen here' is not supported by a fit of any hard-lag model to the measured lag-energy points. A model such as tau(E) = a ln(E/E0) + b should be fitted to the points in Fig. 11, with the 7-10 keV bin treated via band contamination, and the goodness of fit compared with the reverberation interpretation. Without this test, the (3800 +/- 1500) s lag is not uniquely attributable to light travel time, and the derived coronal height of (13 +/- 8) rg is not secured.
  2. [Sections 5.2-5.3, Figs 9 and 11] The frequency band (5x10^-6 to 3x10^-5 Hz) over which the lag-energy spectrum is computed appears to be selected after inspecting the same data: the caption of Fig. 11 describes this range as 'where the iron K band is seen to lag behind the energy band dominated by the X-ray continuum'. This post-hoc selection is not corrected for in the reported significance. The paper should apply the identical selection procedure to null and alternative simulations and report the resulting false-alarm rate, or otherwise account for the trials factor.
  3. [Section 5.3, significance paragraph] The 99.5 per cent significance is defined as the percentage of GP posterior samples in which the 5-7 keV bin responds later than the 1.6-2 keV bin. Because the GP hyperparameters are fit to the same observed light curves and the samples are conditional on the observed data, this is not a frequentist detection probability. The significance should be calibrated with simulations of lag-free light curves processed through the same GP pipeline, and the false-alarm probability should be quoted.
  4. [Section 4.1.2, Fig. 7; Section 5.3] The validation simulations show systematic errors of order 10-20 per cent in the lowest-frequency bins for the 3C 120-like observing geometry (Fig. 7), yet the quoted (3800 +/- 1500) s error bar is only the spread of the GP samples. The systematic uncertainty from the GP method should be propagated into the reported lag and coronal-height estimates, or shown to be negligible for the specific 2013 observation geometry, before the quantitative value of the lag can be taken at face value.
minor comments (3)
  1. [Throughout] There are numerous typographical errors (e.g., 'frequnecy', 'predicitons', 'requried', 'funtions', 'segnents', 'obervations') that should be corrected in a revised manuscript.
  2. [Fig. 7 caption] The caption says 'a 500 s lag' while the text and the simulation description refer to a 5000 s lag; this should be harmonized.
  3. [Section 4.1.2] The text states that the GP statistical uncertainty is 'around three times' that of continuous light curves, then later states that the Poisson-resampling test shows the GP uncertainty is 'around 10 times' that of continuous light curves; these statements should be reconciled and clarified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the GP-derived lag is not imposed by the model, the method is validated on simulated light curves, and the coronal-height conversion is direct light-travel-time arithmetic.

full rationale

The paper's central measurement is not circular. The Gaussian process is fit independently to each energy band's light curve with a kernel that models only the single-series autocorrelation; no lag or cross-spectral phase parameter is introduced in the fit, so the (3800 +/- 1500) s lag is not an input. The method is validated externally on simulated light curves with known injected lags (200 s and 5000 s), and a null test confirms that zero lag is recovered when none is present. The conversion from the measured lag to coronal height is direct arithmetic using the black-hole mass and light-travel time, not a fitted parameter renamed as a prediction. Self-citations (e.g., Wilkins & Fabian 2011, 2012, 2013, 2016; Wilkins et al. 2017) are used for interpreting the lag-energy profile shape and for the scale-height framework, but the detection itself is not justified by those citations. The main weaknesses are statistical and model-selection issues: the low-frequency band (5e-6 to 3e-5 Hz) is chosen after seeing the lag-frequency spectrum, the reported 99.5 percent significance is computed from GP posterior samples, and the hard-lag alternative is excluded by shape rather than by fitting a quantitative model. These are correctness risks, not circularity by construction.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central measurement is not a closed-form derivation; it rests on fitted GP kernels, standard astrophysical assumptions, and published mass and geometry relations. No new physical entities are introduced. The most important uncharged items are the kernel family choice, the stationarity assumption, and the reverberation interpretation of the lag-energy spectrum.

free parameters (1)
  • RQ kernel hyperparameters per energy band = not reported in text
    Section 5.2 and 5.3: amplitude, scale mixture, length scale and white-noise level are fitted by maximum likelihood for each band and then sampled by MCMC. The lag-frequency and lag-energy results depend on these fitted values, so the measurement is conditional on the chosen kernel family.
assumptions (5)
  • standard math The conditional distribution of a multivariate Gaussian gives the correct interpolation of the latent light curve.
    Invoked in Section 2, Equation 5, to draw sample light curves in the gaps.
  • domain assumption AGN X-ray variability is stationary in log count rate and its covariance is adequately described by a rational quadratic kernel.
    Section 3 and 5.2: the kernel choice is justified by broken power-law PSDs and applied to real 3C 120 data.
  • domain assumption Poisson counting noise is well approximated by additive Gaussian white noise.
    Section 3.2: valid in the high-count limit, which is approximate for the faintest segments of the 3C 120 observations.
  • domain assumption The lag-energy spectrum shape can be interpreted as disc reverberation, while a continuum hard lag would rise smoothly with energy.
    Section 5.3: this interpretive premise is stated qualitatively and is not tested by fitting a hard-lag model to the measured spectrum.
  • domain assumption The published black hole mass and the coronal-height conversion of Wilkins and Fabian 2013 apply to 3C 120.
    Sections 5 and 6: used to convert the 3800 s lag into a corona scale height of 13 +/- 8 rg.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Low frequency X-ray timing with Gaussian processes and reverberation in the radio-loud AGN 3C 120." pith.science (2026). https://pith.science/paper/WECM4PCD

@misc{pith2026190806099,
  author       = {Pith},
  title        = {Pith review of: Low frequency X-ray timing with Gaussian processes and reverberation in the radio-loud AGN 3C 120},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WECM4PCD}},
  note         = {Machine review of arXiv:1908.06099}
}
read the original abstract

A framework is developed to perform Fourier-domain timing analysis on X-ray light curves with gaps, employing Gaussian processes to model the probability distribution underlying the observed time series from which continuous samples can be drawn. A technique is developed to measure X-ray reverberation from the inner regions of accretion discs around black holes in the low frequency components of the variability, on timescales longer than can be probed employing standard Fourier techniques. This enables X-ray reverberation experiments to be performed using data from satellites in low-Earth orbit such as NICER, NuSTAR and the proposed X-ray timing mission STROBE-X, and enables long timescale reverberation around higher mass AGN to be measured by combining multiple observations. Gaussian processes are applied to observations of the broad line radio galaxy 3C120 spanning two orbits with XMM-Newton to measure the relative time lags of successive X-ray energy bands. The lag-energy spectrum between 5E-6 and 3E-5Hz, estimated using Gaussian processes, reveals X-ray reverberation from the inner accretion disc for the first time in this radio-loud AGN. Time lags in the relativistically broadened iron K line are significantly detected. The core of the line lags behind the continuum by (3800 +/- 1500)s, suggesting a scale height of the corona of (13 +/- 8)rg above the disc. The ability to compare the structure of coronae in radio loud AGN to their radio quiet counterparts will yield important insight into the mechanisms by which black holes are able to launch jets.

Figures

Figures reproduced from arXiv: 1908.06099 by the authors.

Figure 1
Figure 1. Assessment of the squared exponential kernel in describing the variability in an X-ray light curve of an accreting black hole. (a) The simulated light curve with a f−2 power spectrum and gaps corresponding to observation by a satellite in low-Earth orbit, with a single sample drawn from the Gaussian process shown by the purple line and the prediction from averaging 1000 samples shown by the blue band. (b) Comparison… view at source ↗
Figure 2
Figure 2. As [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. As [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: The coherence, quantifying the phase relationship between pairs of sample light curves drawn from Gaussian processes, compared to the coherence between the original time series with no gaps for a simulated observation of an AGN like Ark 564 with a 200 s lag between the…
Figure 5
Figure 5. Figure 5: As [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: The lag-frequency spectrum estimated by Gaussian processes fit to pairs of simulated light curves representing observations of the nearby, bright Seyfert galaxy Ark 564 with a 200 s lag using (a) NICER and (b) the large collecting area XRCA on board the proposed STROBE…
Figure 7
Figure 7. Figure 7: As [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: X-ray light curve of the BLRG 3C 120 observed with the EPIC pn camera on board XMM-Newton in the 0.3-10 keV energy band in 2013 February. Shaded regions show the predictions marginalised across 1000 samples drawn from a Gaussian process, with a rational quadratic kerne…
Figure 9
Figure 9. Figure 9: The lag-frequency spectrum of 3C 120 between the 1.2-4 keV energy band, dominated by the X-ray continuum, and 4-7 keV energy band dominated by the iron K line from the accre￾tion disc, estimated from sample light curves drawn from Gaus￾sian processes fit to the observe…
Figure 10
Figure 10. Figure 10: The discrete correlation function (DCF) between the observed light curves in the 1.2-4 keV and 4-7 keV energy bands. Positive lags indicate variability in the hard band lagging behind that in the soft. as shown in [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 16 canonical work pages

  1. [1]

    Ar \'e valo P., Uttley P., 2006, @doi [ ] 10.1111/j.1365-2966.2006.09989.x , http://adsabs.harvard.edu/abs/2006MNRAS.367..801A 367, 801

    work page arXiv 2006

  2. [2]

    R., Jaffe A

    Bond J. R., Jaffe A. H., Knox L., 1998, @doi [ ] 10.1103/PhysRevD.57.2117 , https://ui.adsabs.harvard.edu/abs/1998PhRvD..57.2117B 57, 2117

    doi:10.1103/physrevd.57.2117 1998

  3. [3]

    M., Zoghbi A., Reynolds C., Fabian A

    Cackett E. M., Zoghbi A., Reynolds C., Fabian A. C., Kara E., Uttley P., Wilkins D. R., 2014, @doi [ ] 10.1093/mnras/stt2424 , http://adsabs.harvard.edu/abs/2014MNRAS.438.2980C 438, 2980

    doi:10.1093/mnras/stt2424 2014

  4. [4]

    S., Reynolds C

    Cowperthwaite P. S., Reynolds C. S., 2012, @doi [ ] 10.1088/2041-8205/752/2/L21 , https://ui.adsabs.harvard.edu/abs/2012ApJ...752L..21C 752, L21

    doi:10.1088/2041-8205/752/2/l21 2012

  5. [5]

    S., Andrews S

    Czekala I., Mandel K. S., Andrews S. M., Dittmann J. A., Ghosh S. K., Montet B. T., Newton E. R., 2017, @doi [ ] 10.3847/1538-4357/aa6aab , https://ui.adsabs.harvard.edu/abs/2017ApJ...840...49C 840, 49

    work page doi:10.3847/1538-4357/aa6aab 2017

  6. [6]

    M., Fabian A

    De Marco B., Ponti G., Cappi M., Dadina M., Uttley P., Cackett E. M., Fabian A. C., Miniutti G., 2013, @doi [ ] 10.1093/mnras/stt339 , http://adsabs.harvard.edu/abs/2013MNRAS.431.2441D 431, 2441

    doi:10.1093/mnras/stt339 2013

  7. [7]

    A., Krolik J

    Edelson R. A., Krolik J. H., 1988, @doi [ ] 10.1086/166773 , https://ui.adsabs.harvard.edu/abs/1988ApJ...333..646E 333, 646

    doi:10.1086/166773 1988

  8. [8]

    J., Pounds K., Vaughan S., Markowitz A., Marshall H., Dobbie P., Warwick R., 2002, @doi [ ] 10.1086/323779 , http://adsabs.harvard.edu/abs/2002ApJ...568..610E 568, 610

    Edelson R., Turner T. J., Pounds K., Vaughan S., Markowitz A., Marshall H., Dobbie P., Warwick R., 2002, @doi [ ] 10.1086/323779 , http://adsabs.harvard.edu/abs/2002ApJ...568..610E 568, 610

    doi:10.1086/323779 2002

Show all 48 references
  1. [9]

    E., 2016, @doi [ ] 10.1051/0004-6361/201527665 , https://ui.adsabs.harvard.edu/abs/2016A&A...591A.113E 591, A113

    Epitropakis A., Papadakis I. E., 2016, @doi [ ] 10.1051/0004-6361/201527665 , https://ui.adsabs.harvard.edu/abs/2016A&A...591A.113E 591, A113

    2016 doi

  2. [10]

    C., Rees M

    Fabian A. C., Rees M. J., Stella L., White N. E., 1989, , http://ukads.nottingham.ac.uk/abs/1989MNRAS.238..729F 238, 729

    1989

  3. [11]

    C., et al., 2009, @doi [ ] 10.1038/nature08007 , http://adsabs.harvard.edu/abs/2009Natur.459..540F 459, 540

    Fabian A. C., et al., 2009, @doi [ ] 10.1038/nature08007 , http://adsabs.harvard.edu/abs/2009Natur.459..540F 459, 540

    2009 doi

  4. [12]

    C., Arzoumanian Z., Okajima T., 2012, in

    Gendreau K. C., Arzoumanian Z., Okajima T., 2012, in . p. 844313, @doi 10.1117/12.926396

    2012 doi

  5. [13]

    M., Fabian A

    George I. M., Fabian A. C., 1991, , http://ukads.nottingham.ac.uk/abs/1991MNRAS.249..352G 249, 352

    1991

  6. [14]

    C., Cackett E

    Kara E., Fabian A. C., Cackett E. M., Uttley P., Wilkins D. R., Zoghbi A., 2013, @doi [ ] 10.1093/mnras/stt1055 , http://adsabs.harvard.edu/abs/2013MNRAS.434.1129K 434, 1129

    2013 doi

  7. [15]

    Kara E., et al., 2015, @doi [ ] 10.1093/mnras/stu2136 , http://adsabs.harvard.edu/abs/2015MNRAS.446..737K 446, 737

    2015 doi

  8. [16]

    N., Fabian A

    Kara E., Alston W. N., Fabian A. C., Cackett E. M., Uttley P., Reynolds C. S., Zoghbi A., 2016, @doi [ ] 10.1093/mnras/stw1695 , https://ui.adsabs.harvard.edu/\#abs/2016MNRAS.462..511K 462, 511

    2016 doi

  9. [17]

    Kara E., et al., 2019, @doi [ ] 10.1038/s41586-018-0803-x , https://ui.adsabs.harvard.edu/abs/2019Natur.565..198K 565, 198

    2019 doi

  10. [18]

    Kataoka J., et al., 2007, @doi [ ] 10.1093/pasj/59.2.279 , https://ui.adsabs.harvard.edu/abs/2007PASJ...59..279K 59, 279

    2007 doi

  11. [19]

    M., et al., 2013, @doi [ ] 10.1088/0004-637X/772/2/83 , https://ui.adsabs.harvard.edu/abs/2013ApJ...772...83L 772, 83

    Lohfink A. M., et al., 2013, @doi [ ] 10.1088/0004-637X/772/2/83 , https://ui.adsabs.harvard.edu/abs/2013ApJ...772...83L 772, 83

    2013 doi

  12. [20]

    P., Jorstad S

    Marscher A. P., Jorstad S. G., G \'o mez J.-L., Aller M. F., Ter \"a sranta H., Lister M. L., Stirling A. M., 2002, @doi [ ] 10.1038/nature00772 , https://ui.adsabs.harvard.edu/abs/2002Natur.417..625M 417, 625

    2002 doi

  13. [21]

    C., Reynolds C

    Matt G., Fabian A. C., Reynolds C. S., 1997, , http://adsabs.harvard.edu/abs/1997MNRAS.289..175M 289, 175

    1997

  14. [22]

    M., Papadakis I

    McHardy I. M., Papadakis I. E., Uttley P., Page M. J., Mason K. O., 2004, @doi [ ] 10.1111/j.1365-2966.2004.07376.x , http://adsabs.harvard.edu/abs/2004MNRAS.348..783M 348, 783

    2004

  15. [23]

    M., Koerding E., Knigge C., Uttley P., Fender R

    McHardy I. M., Koerding E., Knigge C., Uttley P., Fender R. P., 2006, @doi [ ] 10.1038/nature05389 , http://adsabs.harvard.edu/abs/2006Natur.444..730M 444, 730

    2006 doi

  16. [24]

    J., Reeves J

    Miller L., Turner T. J., Reeves J. N., Lobban A., Kraemer S. B., Crenshaw D. M., 2010a, @doi [ ] 10.1111/j.1365-2966.2009.16149.x , https://ui.adsabs.harvard.edu/abs/2010MNRAS.403..196M 403, 196

    2009

  17. [25]

    J., Reeves J

    Miller L., Turner T. J., Reeves J. N., Braito V., 2010b, @doi [ ] 10.1111/j.1365-2966.2010.17261.x , http://adsabs.harvard.edu/abs/2010MNRAS.408.1928M 408, 1928

    2010

  18. [26]

    Miyamoto S., Kitamoto S., 1989, @doi [ ] 10.1038/342773a0 , http://adsabs.harvard.edu/abs/1989Natur.342..773M 342, 773

    1989 doi

  19. [27]

    Mizumoto M., Done C., Hagino K., Ebisawa K., Tsujimoto M., Odaka H., 2018, @doi [ ] 10.1093/mnras/sty1114 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.478..971M 478, 971

    2018 doi

  20. [28]

    A., Vaughan B

    Nowak M. A., Vaughan B. A., Wilms J., Dove J. B., Begelman M. C., 1999, @doi [ ] 10.1086/306610 , http://adsabs.harvard.edu/abs/1999ApJ...510..874N 510, 874

    1999 doi

  21. [29]

    L., Markoff S., 2010, @doi [ ] 10.1088/0004-637X/723/2/1343 , https://ui.adsabs.harvard.edu/abs/2010ApJ...723.1343P 723, 1343

    Polko P., Meier D. L., Markoff S., 2010, @doi [ ] 10.1088/0004-637X/723/2/1343 , https://ui.adsabs.harvard.edu/abs/2010ApJ...723.1343P 723, 1343

    2010 doi

  22. [30]

    Pozo Nu \ n ez F., Ramolla M., Westhues C., Bruckmann C., Haas M., Chini R., Steenbrugge K., Murphy M., 2012, @doi [ ] 10.1051/0004-6361/201219107 , https://ui.adsabs.harvard.edu/abs/2012A&A...545A..84P 545, A84

    2012 doi

  23. [31]

    E., Williams C., 2006, Gaussian Processes for Machine Learning

    Rasmussen C. E., Williams C., 2006, Gaussian Processes for Machine Learning . MIT Press

    2006

  24. [32]

    S., et al., 2018, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series

    Ray P. S., et al., 2018, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series . p. 1069919, @doi 10.1117/12.2312257

    2018 doi

  25. [33]

    S., 2000, @doi [ ] 10.1086/308697 , https://ui.adsabs.harvard.edu/abs/2000ApJ...533..811R 533, 811

    Reynolds C. S., 2000, @doi [ ] 10.1086/308697 , https://ui.adsabs.harvard.edu/abs/2000ApJ...533..811R 533, 811

    2000 doi

  26. [35]

    Str \"u der L., et al., 2001, @doi [ ] 10.1051/0004-6361:20000066 , http://adsabs.harvard.edu/abs/2001A\

    2001 doi

  27. [36]

    Timmer J., Koenig M., 1995, , http://adsabs.harvard.edu/abs/1995A\

    1995

  28. [38]

    M., Vaughan S., 2005, @doi [ ] 10.1111/j.1365-2966.2005.08886.x , https://ui.adsabs.harvard.edu/\#abs/2005MNRAS.359..345U 359, 345

    Uttley P., McHardy I. M., Vaughan S., 2005, @doi [ ] 10.1111/j.1365-2966.2005.08886.x , https://ui.adsabs.harvard.edu/\#abs/2005MNRAS.359..345U 359, 345

    2005

  29. [39]

    M., Fabian A

    Uttley P., Cackett E. M., Fabian A. C., Kara E., Wilkins D. R., 2014,

    2014

  30. [40]

    J., et al., 2013, @doi [ ] 10.1088/2041-8205/777/2/L23 , http://adsabs.harvard.edu/abs/2013ApJ...777L..23W 777, L23

    Walton D. J., et al., 2013, @doi [ ] 10.1088/2041-8205/777/2/L23 , http://adsabs.harvard.edu/abs/2013ApJ...777L..23W 777, L23

    2013 doi

  31. [41]

    R., Fabian A

    Wilkins D. R., Fabian A. C., 2011, @doi [ ] 10.1111/j.1365-2966.2011.18458.x , http://adsabs.harvard.edu/abs/2011MNRAS.414.1269W 414, 1269

    2011

  32. [43]

    R., Fabian A

    Wilkins D. R., Fabian A. C., 2013, @doi [ ] 10.1093/mnras/sts591 , http://adsabs.harvard.edu/abs/2013MNRAS.430..247W 430, 247

    2013 doi

  33. [44]

    R., Cackett E

    Wilkins D. R., Cackett E. M., Fabian A. C., Reynolds C. S., 2016, @doi [ ] 10.1093/mnras/stw276 , http://adsabs.harvard.edu/abs/2016MNRAS.458..200W 458, 200

    2016 doi

  34. [45]

    R., Gallo L

    Wilkins D. R., Gallo L. C., Silva C. V., Costantini E., Brandt W. N., Kriss G. A., 2017, @doi [ ] 10.1093/mnras/stx1814 , https://ui.adsabs.harvard.edu/\#abs/2017MNRAS.471.4436W 471, 4436

    2017 doi

  35. [46]

    D., Wilkins D

    Yuan Y., Blandford R. D., Wilkins D. R., 2019a, @doi [ ] 10.1093/mnras/stz332 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.484.4920Y 484, 4920

    doi

  36. [47]

    D., Wilkins D

    Yuan Y., Spitkovsky A., Blandford R. D., Wilkins D. R., 2019b, @doi [ ] 10.1093/mnras/stz1599 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.487.4114Y 487, 4114

    doi

  37. [48]

    C., Reynolds C

    Zoghbi A., Fabian A. C., Reynolds C. S., Cackett E. M., 2012, @doi [ ] 10.1111/j.1365-2966.2012.20587.x , http://adsabs.harvard.edu/abs/2012MNRAS.422..129Z 422, 129

    2012

  38. [49]

    M., Miniutti G., Kara E., Fabian A

    Zoghbi A., Reynolds C., Cackett E. M., Miniutti G., Kara E., Fabian A. C., 2013a, @doi [ ] 10.1088/0004-637X/767/2/121 , http://adsabs.harvard.edu/abs/2013ApJ...767..121Z 767, 121

    doi

  39. [50]

    M., 2013b, @doi [ ] 10.1088/0004-637X/777/1/24 , https://ui.adsabs.harvard.edu/abs/2013ApJ...777...24Z 777, 24

    Zoghbi A., Reynolds C., Cackett E. M., 2013b, @doi [ ] 10.1088/0004-637X/777/1/24 , https://ui.adsabs.harvard.edu/abs/2013ApJ...777...24Z 777, 24

    doi

  40. [51]

    Zoghbi A., et al., 2014, @doi [ ] 10.1088/0004-637X/789/1/56 , http://adsabs.harvard.edu/abs/2014ApJ...789...56Z 789, 56

    2014 doi

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.