pith. sign in

arxiv: 1906.11272 · v1 · pith:H5QJG7GDnew · submitted 2019-06-26 · 🌌 astro-ph.GA

Do Reverberation Mapping Analyses Provide an Accurate Picture of the Broad Line Region?

S. W. Mangham , C. Knigge , P. Williams , Keith Horne , A. Pancoast , J. H. Matthews , K. S. Long , S. A. Sim
show 1 more author
N. Higginbottom
This is my paper

Pith reviewed 2026-05-25 15:35 UTC · model grok-4.3

classification 🌌 astro-ph.GA
keywords reverberation mappingbroad line regionactive galactic nucleiaccretion disc windH-alpha emissionvelocity-delay mapsMEMEchoCARAMEL
0
0 comments X

The pith

Reverberation mapping recovers rotation in the broad line region but misses its underlying disc wind structure.

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

The paper tests two standard reverberation mapping techniques, MEMEcho and CARAMEL, on simulated spectroscopic time series generated from a known biconical accretion disc wind model of the broad line region. The simulations use self-consistent ionization and radiative transfer to produce realistic H-alpha responses to an empirical continuum light curve, for both Seyfert and QSO cases. Both methods fail to recover the wind geometry even though the line-forming region is rotation dominated; CARAMEL produces velocity-delay maps and RMS profiles that contradict the input data despite fitting individual spectra well. The Seyfert case yields a negative response that neither method captures, while the QSO case yields only an annular rotation-dominated picture with size overestimated by 50 percent in one analysis.

Core claim

When applied to mock data from a rotating biconical accretion disc wind, neither MEMEcho nor CARAMEL recovers the disc wind nature of the broad line region; CARAMEL velocity-delay maps and RMS line profiles are strongly inconsistent with the input despite good spectral fits, and both methods capture only the rotation-dominated annular character of the H-alpha line-forming region.

What carries the argument

Blind application of MEMEcho and CARAMEL to simulated H-alpha spectroscopic time series generated from a biconical accretion disc wind via self-consistent ionization and radiative transfer.

If this is right

  • For Seyfert-like models producing negative responses, both methods fail gracefully without generating spurious results.
  • For QSO-like models, both methods recover the broadly annular, rotation-dominated character of the line-forming region.
  • MEMEcho overestimates the size of the line-forming region by 50 percent.
  • CARAMEL cannot distinguish additional inflow or outflow components in the QSO model.
  • Since the H-alpha region is rotation dominated, the underlying disc wind geometry remains undetected by either technique.

Where Pith is reading between the lines

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

  • Analyses of real AGN data using these methods may systematically under-represent wind components even when they are present.
  • Extending the test to other emission lines such as C IV could reveal whether the rotation bias is line-specific.
  • Combining RM with independent constraints from microlensing or interferometry might help break the degeneracy between wind and pure rotation geometries.

Load-bearing premise

The simulated time series accurately represent the response that real active galactic nuclei would produce.

What would settle it

A new simulation run in which CARAMEL velocity-delay maps and RMS profiles match the input biconical wind response function would falsify the claim of inconsistency.

Figures

Figures reproduced from arXiv: 1906.11272 by A. Pancoast, C. Knigge, J. H. Matthews, Keith Horne, K. S. Long, N. Higginbottom, P. Williams, S. A. Sim, S. W. Mangham.

Figure 1
Figure 1. Figure 1: Isodelay contour from Peterson et al. (2004). actually produce the observed line emission at a given lag – and which isodelay surfaces actually contribute to a given line – depends on the geometry, density and ionization state of the BLR, as well as on the inclination of the observer with respect to the system. In addition, different parts of a given line – e.g. red and blue line wings – will exhibit dif￾f… view at source ↗
Figure 2
Figure 2. Figure 2: Outline response functions and schematics for Hubble￾type spherical outflow (left), a rotating Keplerian disc viewed at a 20◦ angle (centre), and Hubble-type spherical inflow (right). Winds extend from rmin = 20rg to rmax = 200rg for an AGN of mass 107M . Hubble out/inflows have V(rmin) = ±3 × 103 km s−1 . Solid lines denote the response from the inner and outer edges of the winds, dotted lines from evenly… view at source ↗
Figure 4
Figure 4. Figure 4: Mean line profiles for our the Hα line in our QSO and Seyfert models. 3.1.2 Creating Response Functions We use Python to generate 2-D response functions us￾ing the methodology described in Mangham et al. (2017). Briefly, the response function, ΨR(v, τ), describes how a change in line emission at time t depends upon changes in the continuum across a range of previous times t − τ. This requires making the as… view at source ↗
Figure 5
Figure 5. Figure 5: The ‘true’ velocity-resolved response function (lower) for Hα in the Seyfert model, rescaled to a peak delay of ≈ 3 days. 00  [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Driving continuum and line light-curves for our model and the CARAMEL fit (upper). Spectra associated with three ‘observational’ times (indicated by grey vertical lines on the light￾curves), and their CARAMEL fit (lower). 3.2 Benchmarks: Defining Success To assess the results produced by the RM techniques we are testing, we need to define what constitutes success. At the most basic level, any successful RM… view at source ↗
Figure 8
Figure 8. Figure 8: Trailed spectrograms generated for the continuum-subtracted Hα lines of our QSO (left) and Seyfert (right) models over a simulated observing campaign of 98.9 days [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The emissivity distribution in the QSO model. Dis￾tances have been rescaled to correspond to the rescaled delays (see 3.1.2). X and Y axes are along the disk plane, Z is normal to the disk plane. The red lines indicate the projection of the direc￾tion vector towards the observer in each plot. Note the different (smaller) dynamic range used for the z-axis [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Original and rescaled driving light-curves used in gen￾erating time series of spectra, taken from NGC 5548 (Fausnaugh et al. 2016). provided only with the time-series for the QSO and Seyfert models in their preferred input format, as well as the rescaled continuum light curves used to generate them. Neither were informed that the Seyfert model would exhibit a negative response. The following methods secti… view at source ↗
Figure 12
Figure 12. Figure 12: Model fits to the Seyfert model Hα line profile, inte￾grated Hα flux, and AGN continuum flux. Panel 1: The provided Hα emission-line profile for each epoch. Panel 2: The Hα emission￾line profile for each epoch produced by one sample of the BLR and continuum model. Panel 3: The provided Hα line profile for one randomly chosen epoch (black), and the corresponding profile (red) produced by the model in Panel… view at source ↗
Figure 13
Figure 13. Figure 13: MEMEcho fit to the synthetic Seyfert data. Bottom panel is the driving light-curve, above which are the 1-D echo maps (left) and echo light-curves (right) at selected wavelengths. Note the high target χ 2 /101 = 3 and the poor fits achieved near line center. 4.2 Blind Analysis and Interpretation: MEMEcho Results for the QSO model 4.2.1 1-D delay maps ΨR(τ) for the QSO simulation In [PITH_FULL_IMAGE:figur… view at source ↗
Figure 16
Figure 16. Figure 16: This shows the continuum light-curve and the [PITH_FULL_IMAGE:figures/full_fig_p012_16.png] view at source ↗
Figure 14
Figure 14. Figure 14: Two-dimensional wavelength-delay map ΨR(λ, τ) reconstructed from the MEMEcho fit to the synthetic Seyfert data. Given below the greyscale map are projections of ΨR(λ, τ) giving delay-integrated responses ΨR(λ) for the full delay range (black), and for restricted delay slices 0-5 d (purple), 5-10 d (green), 10-15 d (orange), and 15-20 d (red). To the right of the greyscale map are wavelength-integrated res… view at source ↗
Figure 15
Figure 15. Figure 15: MEMEcho fit to the continuum and integrated Hα light-curve from the synthetic QSO dataset. The fit achieves χ 2 /N = 1 both for the continuum variations (lower panel) and the line variations (upper right panel). Blue curves show the fitted model, including the continuum light-curve C(t) (bottom panel), the delay map ΨR(τ) (upper left panel) and the line light-curve L(t) (upper right panel). Horizontal red… view at source ↗
Figure 17
Figure 17. Figure 17: Two-dimensional wavelength-delay map ΨR(λ, τ) reconstructed from the MEMEcho fit to the synthetic QSO data. Given below the grey-scale map are projections of ΨR(λ, τ) giving delay-integrated responses ΨR(λ) for the full delay range (black), and for restricted delay slices 0-5 d (purple), 5-10 d (green), 10-15 d (orange), and 15-20 d (red). To the right of the grey-scale map are wavelength￾integrated respo… view at source ↗
Figure 6
Figure 6. Figure 6: In our view, the performance of MEMEcho in recov￾ering the input velocity-delay map is quite impressive. To a good approximation, the recovered map is a smoothed ver￾sion of the input map, exactly as one might hope and expect. However, the true peak in the overall delay distribution lies at ' 4 days, whereas the peak in the recovered distribution is estimated to be ' 6 days. This is almost certainly asso￾c… view at source ↗
Figure 18
Figure 18. Figure 18: Model fits to the QSO model Hα line profile, inte￾grated Hα flux, and AGN continuum flux. Panel 1: The provided Hα emission-line profile for each epoch. Panel 2: The Hα emission￾line profile for each epoch produced by one sample of the BLR and continuum model. Panel 3: The provided Hα line profile for one randomly chosen epoch (black), and the corresponding profile (red) produced by the model in Panel 2. … view at source ↗
Figure 19
Figure 19. Figure 19: Posterior distributions of select model parameters for the QSO model. 5 10 15 20 25 30 35 40 Delay (days) 6400 6500 6600 6700 6800 Wavelength (˚A) 0.0 0.5 1.0 [PITH_FULL_IMAGE:figures/full_fig_p017_19.png] view at source ↗
Figure 21
Figure 21. Figure 21: Geometric model of the broad line region that was used to create the transfer function in [PITH_FULL_IMAGE:figures/full_fig_p017_21.png] view at source ↗
Figure 20
Figure 20. Figure 20: Velocity-resolved transfer function for the QSO model, chosen to be representative of the full posterior sample. The right-hand panel shows the velocity-integrated transfer func￾tion and the bottom panel shows the time-averaged line profile. −10 0 10 x (light days) −10 −5 0 10 5 z (light days) −10 −5 0 5 10 y (light days) [PITH_FULL_IMAGE:figures/full_fig_p017_20.png] view at source ↗
Figure 22
Figure 22. Figure 22: RMS residuals for the noisy output time series of spectra and CARAMEL fit to it. ticular annulus, just a smooth distribution across the entire width of the envelope at long delays, and a bright, diagonal “line” at short delays (with a blue-leads-red signature). We have checked whether the particular model shown in [PITH_FULL_IMAGE:figures/full_fig_p018_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: CARAMEL model fits to the QSO model Hα line profile, compared to the model line profiles. Red lines indicate line centres. Panel 1: The mean model and CARAMEL line profiles across all epochs. Panel 2: The model emission-line profile for each epoch. Panel 3: The CARAMEL fit emission-line profile for each epoch. Panel 4: The standardised model - fit residuals for the emission-line profile for each epoch. us… view at source ↗
Figure 24
Figure 24. Figure 24: Diagram showing how a discretised toy response function ΨR(λ, τ) (top left) applies changes in continuum luminosity ∆C to a base spectrum to form a series of spectra at later. Top right 3 panels illustrate how a change in luminosity in a single time step t propagates out over a range of times t + τ. Within a spectrum, columns in blue indicate a wavelength bin whose flux has been decreased by a negative re… view at source ↗
read the original abstract

Reverberation mapping (RM) is a powerful approach for determining the nature of the broad-line region (BLR) in active galactic nuclei. However, inferring physical BLR properties from an observed spectroscopic time series is a difficult inverse problem. Here, we present a blind test of two widely used RM methods: MEMEcho (developed by Horne) and CARAMEL (developed by Pancoast and collaborators). The test data are simulated spectroscopic time series that track the H$\alpha$ emission line response to an empirical continuum light curve. The underlying BLR model is a rotating, biconical accretion disc wind, and the synthetic spectra are generated via self-consistent ionization and radiative transfer simulations. We generate two mock data sets, representing Seyfert galaxies and QSOs. The Seyfert model produces a largely *negative* response, which neither method can recover. However, both fail $``gracefully''$, neither generating spurious results. For the QSO model both CARAMEL and expert interpretation of MEMEcho's output both capture the broadly annular, rotation-dominated nature of the line-forming region, though MEMEcho analysis overestimates its size by 50%, but CARAMEL is unable to distinguish between additional inflow and outflow components. Despite fitting individual spectra well, the CARAMEL velocity-delay maps and RMS line profiles are strongly inconsistent with the input data. Finally, since the H$\alpha$ line-forming region is rotation dominated, neither method recovers the disc wind nature of the underlying BLR model. Thus considerable care is required when interpreting the results of RM analyses in terms of physical models.

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

1 major / 0 minor

Summary. The paper conducts a blind test of MEMEcho and CARAMEL reverberation mapping methods on two mock Hα spectroscopic time series generated from a rotating biconical accretion disc wind BLR model using self-consistent ionization and radiative transfer simulations (Seyfert and QSO cases). It reports that neither method recovers the disc-wind geometry, with the Seyfert case producing an unrecoverable negative response, the QSO case showing a 50% size overestimate by MEMEcho, CARAMEL failing to distinguish inflow/outflow, and CARAMEL maps/profiles inconsistent with input despite good spectral fits; the conclusion is that RM analyses require care when inferring physical BLR models.

Significance. If the simulation is representative, the work supplies concrete, externally grounded evidence of RM method limitations for recovering wind components when the line-forming region is rotation-dominated. The use of known-input mock data with no post-hoc exclusions provides a clear falsifiable test of the methods.

major comments (1)
  1. [Mock data generation] Mock data generation (abstract and associated section): the central claim that failure to recover the disc-wind nature demonstrates an intrinsic limitation of RM methods (rather than a mismatch with this particular forward model) rests on the unvalidated assumption that the chosen biconical wind parameters, ionization balance, and radiative-transfer approximations produce responses that real AGN observations of such a BLR would exhibit; no comparison to observed AGN line profiles or other wind simulations is reported.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We appreciate the referee's positive assessment and recommendation for minor revision. We address the major comment below.

read point-by-point responses

  1. Referee: [Mock data generation] Mock data generation (abstract and associated section): the central claim that failure to recover the disc-wind nature demonstrates an intrinsic limitation of RM methods (rather than a mismatch with this particular forward model) rests on the unvalidated assumption that the chosen biconical wind parameters, ionization balance, and radiative-transfer approximations produce responses that real AGN observations of such a BLR would exhibit; no comparison to observed AGN line profiles or other wind simulations is reported.

    Authors: We note that the manuscript frames its conclusions cautiously, stating that 'considerable care is required when interpreting the results of RM analyses in terms of physical models' rather than asserting an intrinsic limitation of the methods in general. The test is performed on a specific, physically self-consistent disc-wind model with rotation-dominated line emission, chosen to explore whether RM methods can distinguish wind geometries even when rotation is present. The parameters are drawn from established accretion disc wind models in the literature. Although direct comparisons to observed AGN profiles are not included here, the radiative transfer approach has been tested in previous works. This blind test with known input provides evidence that, for such models, the wind signature may not be recovered, supporting the call for careful interpretation. We do not believe a revision is necessary, as the scope of the study is clearly delimited to this forward model. revision: no

Circularity Check

0 steps flagged

No significant circularity: central claim is a direct comparison to known simulation inputs

full rationale

The paper generates mock spectroscopic time series from an explicitly specified rotating biconical accretion disc wind model using self-consistent ionization and radiative transfer, then applies MEMEcho and CARAMEL to test recovery of the input geometry. The conclusion that neither method recovers the disc-wind nature (because the Hα line-forming region is rotation-dominated) follows from comparing method outputs against the known simulation parameters, which is externally grounded by construction rather than derived from the methods themselves. No load-bearing step reduces by the paper's equations or self-citations to a fit or prior result by the same authors; the test is self-contained against the simulation benchmark. Potential concerns about mock-data fidelity to real AGN are validity issues, not circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim depends on the fidelity of the radiative-transfer simulation and the assumption that the chosen wind parameters produce representative negative and positive responses; no additional free parameters are fitted inside the RM analysis itself.

axioms (1)
  • domain assumption The biconical accretion disc wind model with self-consistent ionization produces realistic Hα response time series for both Seyfert and QSO regimes.
    Invoked when generating the mock data sets that serve as ground truth for the blind test.

pith-pipeline@v0.9.0 · 5865 in / 1315 out tokens · 22546 ms · 2026-05-25T15:35:07.511410+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

49 extracted references · 49 canonical work pages · 2 internal anchors

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence 'output.state := if if FUNCTION not #0 #1 if FUNCTION and 'skip pop #0 if FUNCTION or pop #1...

    work page

  2. [2]

    C., et al., 2008, @doi [ ] 10.1086/595719 , http://adsabs.harvard.edu/abs/2008ApJ...689L..21B 689, L21

    Bentz M. C., et al., 2008, @doi [ ] 10.1086/595719 , http://adsabs.harvard.edu/abs/2008ApJ...689L..21B 689, L21

    work page doi:10.1086/595719 2008

  3. [3]

    C., et al., 2010, , 720, L46

    Bentz M. C., et al., 2010, , 720, L46

    work page 2010

  4. [4]

    C., et al., 2013, @doi [ ] 10.1088/0004-637X/767/2/149 , http://adsabs.harvard.edu/abs/2013ApJ...767..149B 767, 149

    Bentz M. C., et al., 2013, @doi [ ] 10.1088/0004-637X/767/2/149 , http://adsabs.harvard.edu/abs/2013ApJ...767..149B 767, 149

    work page internal anchor Pith review doi:10.1088/0004-637x/767/2/149 2013

  5. [5]

    J., P \'a rtay L

    Brewer B. J., P \'a rtay L. B., Cs \'a nyi G., 2011, Statistics and Computing, http://adsabs.harvard.edu/abs/2010ascl.soft10029B 21, 649

    work page 2011

  6. [6]

    A., okas E

    Cackett E. M., Horne K., 2006, @doi [ ] 10.1111/j.1365-2966.2005.09795.x , http://adsabs.harvard.edu/abs/2006MNRAS.365.1180C 365, 1180

    work page doi:10.1111/j.1365-2966.2005.09795.x 2006

  7. [7]

    De Rosa G., 2015, IAU General Assembly, http://adsabs.harvard.edu/abs/2015IAUGA..2257825D 22, 2257825

    work page 2015

  8. [8]

    De Rosa G., et al., 2015, @doi [ ] 10.1088/0004-637X/806/1/128 , http://adsabs.harvard.edu/abs/2015ApJ...806..128D 806, 128

    work page doi:10.1088/0004-637x/806/1/128 2015

  9. [9]

    D., et al., 2009, , 704, L80

    Denney K. D., et al., 2009, , 704, L80

    work page 2009

  10. [10]

    Du P., et al., 2014, @doi [ ] 10.1088/0004-637X/782/1/45 , http://adsabs.harvard.edu/abs/2014ApJ...782...45D 782, 45

    work page doi:10.1088/0004-637x/782/1/45 2014

  11. [11]

    Du P., et al., 2016, @doi [ ] 10.3847/0004-637X/820/1/27 , http://adsabs.harvard.edu/abs/2016ApJ...820...27D 820, 27

    work page doi:10.3847/0004-637x/820/1/27 2016

  12. [12]

    M., et al., 2016, @doi [ ] 10.3847/0004-637X/821/1/56 , http://adsabs.harvard.edu/abs/2016ApJ...821...56F 821, 56

    Fausnaugh M. M., et al., 2016, @doi [ ] 10.3847/0004-637X/821/1/56 , http://adsabs.harvard.edu/abs/2016ApJ...821...56F 821, 56

    work page doi:10.3847/0004-637x/821/1/56 2016

  13. [13]

    M., 1988, @doi [ ] 10.1086/165986 , http://adsabs.harvard.edu/abs/1988ApJ...325..114G 325, 114

    Gaskell C. M., 1988, @doi [ ] 10.1086/165986 , http://adsabs.harvard.edu/abs/1988ApJ...325..114G 325, 114

    work page doi:10.1086/165986 1988

  14. [14]

    R., O'Brien P

    Goad M. R., O'Brien P. T., Gondhalekar P. M., 1993, @doi [ ] 10.1093/mnras/263.1.149 , http://adsabs.harvard.edu/abs/1993MNRAS.263..149G 263, 149

    work page doi:10.1093/mnras/263.1.149 1993

  15. [15]

    J., 2013, PhD thesis, The Ohio State University

    Grier C. J., 2013, PhD thesis, The Ohio State University

    work page 2013

  16. [16]

    S., Sim S

    Higginbottom N., Knigge C., Long K. S., Sim S. A., Matthews J. H., 2013, @doi [ ] 10.1093/mnras/stt1658 , 436, 1390

    work page doi:10.1093/mnras/stt1658 2013

  17. [17]

    S., Matthews J

    Higginbottom N., Proga D., Knigge C., Long K. S., Matthews J. H., Sim S. A., 2014, @doi [ ] 10.1088/0004-637X/789/1/19 , http://adsabs.harvard.edu/abs/2014ApJ...789...19H 789, 19

    work page doi:10.1088/0004-637x/789/1/19 2014

  18. [18]

    M., Horne K., Peterson B

    Horne K., 1994, in Gondhalekar P. M., Horne K., Peterson B. M., eds, Astronomical Society of the Pacific Conference Series Vol. 69, Reverberation Mapping of the Broad-Line Region in Active Galactic Nuclei. p. 23

    work page 1994

  19. [19]

    M., Collier S

    Horne K., Peterson B. M., Collier S. J., Netzer H., 2004, @doi [ ] 10.1086/420755 , http://adsabs.harvard.edu/abs/2004PASP..116..465H 116, 465

    work page doi:10.1086/420755 2004

  20. [20]

    D., 2007, @doi [Computing in Science and Engineering] 10.1109/MCSE.2007.55 , 9, 90

    Hunter J. D., 2007, @doi [Computing In Science & Engineering] 10.1109/MCSE.2007.55 , 9, 90

    work page doi:10.1109/mcse.2007.55 2007

  21. [21]

    Kaspi S., Netzer H., 1999, , 524, 71

    work page 1999

  22. [22]

    M., Vestergaard M., Jannuzi B

    Kaspi S., Maoz D., Netzer H., Peterson B. M., Vestergaard M., Jannuzi B. T., 2005, , 629, 61

    work page 2005

  23. [23]

    P., Gaskell C

    Koratkar A. P., Gaskell C. M., 1989, @doi [ ] 10.1086/167937 , http://adsabs.harvard.edu/abs/1989ApJ...345..637K 345, 637

    work page doi:10.1086/167937 1989

  24. [24]

    H., Done C., 1995, @doi [ ] 10.1086/175258 , http://adsabs.harvard.edu/abs/1995ApJ...440..166K 440, 166

    Krolik J. H., Done C., 1995, @doi [ ] 10.1086/175258 , http://adsabs.harvard.edu/abs/1995ApJ...440..166K 440, 166

    work page doi:10.1086/175258 1995

  25. [25]

    H., Kriss G

    Krolik J. H., Kriss G. A., 1995, @doi [ ] 10.1086/175896 , http://adsabs.harvard.edu/abs/1995ApJ...447..512K 447, 512

    work page doi:10.1086/175896 1995

  26. [26]

    Li Y.-R., Wang J.-M., Bai J.-M., 2016, @doi [ ] 10.3847/0004-637X/831/2/206 , https://ui.adsabs.harvard.edu/#abs/2016ApJ...831..206L 831, 206

    work page doi:10.3847/0004-637x/831/2/206 2016

  27. [27]

    S., Knigge C., 2002, @doi [ ] 10.1086/342879 , 579, 725

    Long K. S., Knigge C., 2002, @doi [ ] 10.1086/342879 , 579, 725

    work page doi:10.1086/342879 2002

  28. [28]

    W., Knigge C., Matthews J

    Mangham S. W., Knigge C., Matthews J. H., Long K. S., Sim S. A., Higginbottom N., 2017, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stx1863 , 471, 4788

    work page doi:10.1093/mnras/stx1863 2017

  29. [29]

    H., Knigge C., Long K

    Matthews J. H., Knigge C., Long K. S., Sim S. A., Higginbottom N., 2015, @doi [ ] 10.1093/mnras/stv867 , http://adsabs.harvard.edu/abs/2015MNRAS.450.3331M 450, 3331

    work page doi:10.1093/mnras/stv867 2015

  30. [30]

    H., Knigge C., Long K

    Matthews J. H., Knigge C., Long K. S., Sim S. A., Higginbottom N., Mangham S. W., 2016, @doi [ ] 10.1093/mnras/stw323 , http://adsabs.harvard.edu/abs/2016MNRAS.458..293M 458, 293

    work page doi:10.1093/mnras/stw323 2016

  31. [31]

    Murray N., Chiang J., 1996, , 382, 789

    work page 1996

  32. [32]

    M., Long K

    Noebauer U. M., Long K. S., Sim S. A., Knigge C., 2010, @doi [ ] 10.1088/0004-637X/719/2/1932 , http://adsabs.harvard.edu/abs/2010ApJ...719.1932N 719, 1932

    work page doi:10.1088/0004-637x/719/2/1932 2010

  33. [33]

    E., 2015, Guide to NumPy, 2nd edn

    Oliphant T. E., 2015, Guide to NumPy, 2nd edn. CreateSpace Independent Publishing Platform, USA

    work page 2015

  34. [34]

    A., Ferrarese L., Merritt D., Peterson B

    Onken C. A., Ferrarese L., Merritt D., Peterson B. M., Pogge R. W., Vestergaard M., Wandel A., 2004, , 615, 645

    work page 2004

  35. [35]

    O., 2018, @doi [Practical Free Alternatives to Commercial Software] https://doi.org/10.1007/978-1-4842-3075-6_10

    Oualline S., G. O., 2018, @doi [Practical Free Alternatives to Commercial Software] https://doi.org/10.1007/978-1-4842-3075-6_10

    work page doi:10.1007/978-1-4842-3075-6_10 2018

  36. [36]

    J., Treu T., 2011, @doi [ ] 10.1088/0004-637X/730/2/139 , http://adsabs.harvard.edu/abs/2011ApJ...730..139P 730, 139

    Pancoast A., Brewer B. J., Treu T., 2011, @doi [ ] 10.1088/0004-637X/730/2/139 , http://adsabs.harvard.edu/abs/2011ApJ...730..139P 730, 139

    work page doi:10.1088/0004-637x/730/2/139 2011

  37. [37]

    J., Treu T., 2014a, @doi [ ] 10.1093/mnras/stu1809 , http://adsabs.harvard.edu/abs/2014MNRAS.445.3055P 445, 3055

    Pancoast A., Brewer B. J., Treu T., 2014a, @doi [ ] 10.1093/mnras/stu1809 , http://adsabs.harvard.edu/abs/2014MNRAS.445.3055P 445, 3055

    work page doi:10.1093/mnras/stu1809

  38. [38]

    J., Treu T., Park D., Barth A

    Pancoast A., Brewer B. J., Treu T., Park D., Barth A. J., Bentz M. C., Woo J.-H., 2014b, @doi [ ] 10.1093/mnras/stu1419 , http://adsabs.harvard.edu/abs/2014MNRAS.445.3073P 445, 3073

    work page doi:10.1093/mnras/stu1419

  39. [39]

    Pearson T., , PGPLOT Graphics Subroutine Library, http://www.astro.caltech.edu/ tjp/pgplot/

    work page

  40. [40]

    Pei L., et al., 2017, @doi [ ] 10.3847/1538-4357/aa5eb1 , http://adsabs.harvard.edu/abs/2017ApJ...837..131P 837, 131

    work page doi:10.3847/1538-4357/aa5eb1 2017

  41. [41]

    M., 1993, @doi [ ] 10.1086/133140 , http://cdsads.u-strasbg.fr/abs/1993PASP..105..247P 105, 247

    Peterson B. M., 1993, @doi [ ] 10.1086/133140 , http://cdsads.u-strasbg.fr/abs/1993PASP..105..247P 105, 247

    work page doi:10.1086/133140 1993

  42. [42]

    M., et al., 2004, , 613, 682

    Peterson B. M., et al., 2004, , 613, 682

    work page 2004

  43. [43]

    P., Wanders I., 1994, @doi [ ] 10.1093/mnras/271.1.183 , https://ui.adsabs.harvard.edu/#abs/1994MNRAS.271..183P 271, 183

    Pijpers F. P., Wanders I., 1994, @doi [ ] 10.1093/mnras/271.1.183 , https://ui.adsabs.harvard.edu/#abs/1994MNRAS.271..183P 271, 183

    work page doi:10.1093/mnras/271.1.183 1994

  44. [44]

    Shlosman I., Vitello P., 1993, @doi [ ] 10.1086/172670 , http://adsabs.harvard.edu/abs/1993ApJ...409..372S 409, 372

    work page doi:10.1086/172670 1993

  45. [45]

    A., okas E

    Sim S. A., Drew J. E., Long K. S., 2005, @doi [ ] 10.1111/j.1365-2966.2005.09472.x , 363, 615

    work page doi:10.1111/j.1365-2966.2005.09472.x 2005

  46. [46]

    J., Bentz M

    Skielboe A., Pancoast A., Treu T., Park D., Barth A. J., Bentz M. C., 2015, @doi [ ] 10.1093/mnras/stv1917 , http://adsabs.harvard.edu/abs/2015MNRAS.454..144S 454, 144

    work page doi:10.1093/mnras/stv1917 2015

  47. [47]

    The Astropy Collaboration et al., 2018, preprint, http://adsabs.harvard.edu/abs/2018arXiv180102634T ( @eprint arXiv 1801.02634 )

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  48. [48]

    H., Horne K., 1996, @doi [ ] 10.1093/mnras/283.3.748 , http://adsabs.harvard.edu/abs/1996MNRAS.283..748U 283, 748

    Ulrich M. H., Horne K., 1996, @doi [ ] 10.1093/mnras/283.3.748 , http://adsabs.harvard.edu/abs/1996MNRAS.283..748U 283, 748

    work page doi:10.1093/mnras/283.3.748 1996

  49. [49]

    F., Horne K., 1991, @doi [ ] 10.1086/170530 , 379, 586

    Welsh W. F., Horne K., 1991, @doi [ ] 10.1086/170530 , 379, 586

    work page doi:10.1086/170530 1991