Robust AGN and host-galaxy decomposition in optical spectral fitting
Pith reviewed 2026-05-07 06:40 UTC · model grok-4.3
The pith
A combined spectral fitting technique reliably separates AGN and stellar components, yielding consistent physical properties across different activity states.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The paper claims that combining pPXF for stellar population modeling with PyQSOFit for AGN components allows robust decomposition of optical spectra. In samples of changing-look AGN observed in bright and dim states, as well as repeated observations of quasars and galaxies, host properties remain consistent for AGN continuum fractions below 0.8. Black-hole masses from broad Hα and Hβ lines agree across epochs once measurements satisfy 3-sigma confidence on flux and width, and these masses follow established scaling relations. Many of the changing-look objects show breathing broad-line regions.
What carries the argument
The joint pPXF-PyQSOFit fitting procedure that simultaneously models stellar templates and AGN power-law continuum plus emission lines.
If this is right
- Stellar mass and velocity dispersion measurements are trustworthy provided the AGN contributes no more than 80 percent of the total light.
- Single-epoch black-hole mass estimates can be vetted using 3-sigma significance on broad-line flux and FWHM to ensure they are stable over time.
- Changing-look AGN commonly display coordinated changes in broad-line width and strength, consistent with breathing behavior.
- After quality filtering, the masses match those predicted by known relations between black-hole mass and host properties.
Where Pith is reading between the lines
- The quality criteria could be adopted as standard practice for selecting reliable black-hole masses from large spectroscopic surveys.
- Non-virialized systems identified in the sample may require separate dynamical models to understand their line profiles during state changes.
- Applying the same decomposition to multi-epoch data from upcoming surveys would allow tracking of individual black-hole growth episodes.
- The method might help resolve debates about whether apparent changes in black-hole mass are real or artifacts of poor decomposition.
Load-bearing premise
The fitting codes correctly capture the AGN continuum shape and broad emission lines without significant template mismatches or unaccounted residuals, even when the AGN brightness varies substantially between observations.
What would settle it
If repeated observations of the same changing-look AGN, after decomposition, yield stellar masses or black-hole masses that differ by more than their uncertainties between bright and faint states, the robustness claim would be refuted.
Figures
read the original abstract
Unraveling the growth of supermassive black holes and their connection to host galaxies requires disentangling the Active Galactic Nuclei (AGN) emission from that of the stellar populations. When an AGN spectrum is observed at different activity phases, if the spectral decomposition properly recognizes the nuclear and stellar components, key physical properties - such as black-hole mass, stellar mass, and stellar velocity dispersion - should remain consistent. We present a novel optical spectral-fitting approach that combines pPXF and PyQSOFit to robustly decompose spectra into stellar and AGN components. We apply this technique to three SDSS samples with repeated optical spectra of the same objects at z<0.55: 32 changing-look AGN in bright and dim states, and 15 quasars and 15 galaxies with three single-epoch and one stacked spectrum each. To compare with the literature, we use SDSS spectra and photometric data from AGN in the eFEDS field, as well as Gemini and VLT observations of some of our changing-look AGN. We evaluate the reliability of stellar mass, velocity dispersion, and black-hole mass measurements, especially in relation to the AGN-to-total continuum contribution (fAGN). For host-derived properties, especially when fAGN<0.8, our method yields consistent results. For single-epoch black-hole mass estimates from Ha and Hb, 3-sigma confidence in the broad-line flux and FWHM provides effective criteria for selecting reliable measurements. After applying these quality cuts, measurements across different epochs agree within uncertainties, and their reliability is confirmed by the alignment with previously established scaling relations. Many changing-look AGN in our sample exhibit "breathing" broad-line regions, as determined from Ha analysis, while some deviate significantly, suggesting non-virialized systems across the spectral transition.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript presents a spectral decomposition technique combining pPXF (for stellar continuum and velocity dispersion) with PyQSOFit (for AGN continuum, power-law, and broad-line components). It is applied to repeated SDSS spectra of 32 changing-look AGN (bright-to-dim transitions), plus control samples of 15 quasars and 15 galaxies at z<0.55. The central claims are that host-galaxy properties (M* and σ*) remain consistent across epochs when fAGN<0.8, that 3σ cuts on broad Hα/Hβ flux and FWHM yield reliable single-epoch black-hole masses that agree within uncertainties and follow established scaling relations, and that many changing-look objects exhibit “breathing” broad-line regions while a subset appear non-virialized.
Significance. If the decomposition is shown to be free of state-dependent template biases, the method would provide a practical, publicly implementable route to extract host and black-hole properties from single-epoch spectra even when the AGN contributes substantially to the optical continuum. The multi-epoch internal-consistency test on real changing-look objects is a useful validation step, and the breathing-BLR result adds to the literature on BLR dynamics. However, the absence of quantitative residual analysis versus fAGN, mock-spectrum tests, or external cross-validation against independent photometry or higher-resolution data limits the immediate impact and leaves open the possibility that reported consistencies partly reflect fitting degeneracies rather than physical stability.
major comments (3)
- [§4] §4 (changing-look AGN results): The validation rests entirely on epoch-to-epoch agreement for the 32 objects whose AGN state changes markedly. Because bright-to-dim transitions alter the AGN continuum shape and line strengths, any systematic residuals from PyQSOFit’s AGN templates or pPXF’s stellar library that correlate with fAGN would bias M* and σ* in a manner that the internal consistency test cannot detect. The manuscript should add (i) a plot of fit residuals versus fAGN and (ii) tests on simulated spectra with known input parameters to demonstrate that no such leakage occurs.
- [Methods] Methods (pPXF + PyQSOFit interface): The description of how the two codes are sequenced or coupled is insufficient. It is unclear whether the AGN power-law and broad lines are masked or modeled during the pPXF stellar fit, or whether stellar templates are subtracted before PyQSOFit. This detail is load-bearing for the claim of “robust” decomposition, because unaccounted degeneracies between the AGN continuum and stellar features could systematically affect the reported host properties at fAGN ≈ 0.6–0.8.
- [§5] §5 (black-hole mass section): The 3σ significance cut on broad-line flux and FWHM is introduced as an effective selection criterion, yet no quantitative assessment is given of (a) the fraction of measurements rejected, (b) the resulting change in scatter relative to literature values, or (c) a full error budget that includes template choice and continuum subtraction uncertainties. Without these, it is difficult to judge whether the post-cut agreement across epochs is meaningfully tighter than expected from measurement noise alone.
minor comments (3)
- [§2 and figure captions] The definition of fAGN (AGN-to-total continuum ratio at a specific rest-frame wavelength) should be stated explicitly in the text and in all figure captions where it is used as a threshold.
- [Figures] Decomposition figures would be clearer if they included the residual spectrum (data minus model) in a separate panel and indicated the wavelength regions used for the fAGN measurement.
- [§4] A short table summarizing the number of objects and spectra retained after each quality cut (fAGN<0.8, 3σ line cuts) would help readers assess sample completeness.
Simulated Author's Rebuttal
Thank you for the referee's insightful comments on our manuscript. We appreciate the suggestions for strengthening the validation and clarity of our methods. We have addressed each point below and will make the corresponding revisions to the manuscript.
read point-by-point responses
-
Referee: [§4] §4 (changing-look AGN results): The validation rests entirely on epoch-to-epoch agreement for the 32 objects whose AGN state changes markedly. Because bright-to-dim transitions alter the AGN continuum shape and line strengths, any systematic residuals from PyQSOFit’s AGN templates or pPXF’s stellar library that correlate with fAGN would bias M* and σ* in a manner that the internal consistency test cannot detect. The manuscript should add (i) a plot of fit residuals versus fAGN and (ii) tests on simulated spectra with known input parameters to demonstrate that no such leakage occurs.
Authors: We thank the referee for this comment. We argue that the epoch-to-epoch consistency test is in fact sensitive to fAGN-dependent biases: if systematic residuals in the decomposition correlated with fAGN, the derived host properties would vary between the bright and dim states, yet we observe no such variation for fAGN < 0.8. Nevertheless, we will add a plot of the fit residuals as a function of fAGN to further demonstrate the fit quality. For mock spectrum tests, while valuable, they are beyond the current scope as simulating realistic changing-look AGN spectra involves many assumptions; we will instead expand on our use of control samples and external data (eFEDS photometry and Gemini/VLT observations) to support the robustness. We will update the discussion accordingly. revision: partial
-
Referee: [Methods] Methods (pPXF + PyQSOFit interface): The description of how the two codes are sequenced or coupled is insufficient. It is unclear whether the AGN power-law and broad lines are masked or modeled during the pPXF stellar fit, or whether stellar templates are subtracted before PyQSOFit. This detail is load-bearing for the claim of “robust” decomposition, because unaccounted degeneracies between the AGN continuum and stellar features could systematically affect the reported host properties at fAGN ≈ 0.6–0.8.
Authors: We apologize for the lack of detail in the original submission. In the revised manuscript, we will expand the Methods section to clearly describe the coupling: The procedure begins with PyQSOFit applied to the observed spectrum, where the AGN power-law continuum, broad emission lines, and narrow lines are modeled while masking regions dominated by stellar absorption features to avoid degeneracies. The best-fit AGN model is then subtracted from the spectrum, and pPXF is applied to the residual to determine the stellar population and velocity dispersion. This order was chosen after testing alternatives, as it yields more stable results. A workflow diagram will be added to illustrate the process. revision: yes
-
Referee: [§5] §5 (black-hole mass section): The 3σ significance cut on broad-line flux and FWHM is introduced as an effective selection criterion, yet no quantitative assessment is given of (a) the fraction of measurements rejected, (b) the resulting change in scatter relative to literature values, or (c) a full error budget that includes template choice and continuum subtraction uncertainties. Without these, it is difficult to judge whether the post-cut agreement across epochs is meaningfully tighter than expected from measurement noise alone.
Authors: We agree that providing these quantitative details will improve the paper. In the revised version, we will include: (a) the fraction of broad-line measurements rejected by the 3σ cut on flux and FWHM, (b) the resulting change in scatter in black-hole mass estimates relative to literature values, and (c) an expanded error budget that includes uncertainties from AGN template choice and continuum subtraction. These will demonstrate that the post-cut agreement is tighter than expected from noise alone. revision: yes
Circularity Check
No significant circularity: empirical consistency checks on public data with open tools
full rationale
The paper applies the established, publicly available tools pPXF and PyQSOFit to SDSS spectra of changing-look AGN, quasars, and galaxies. It reports direct measurements of stellar mass, velocity dispersion, and single-epoch black-hole masses, followed by empirical epoch-to-epoch consistency tests and comparisons to literature scaling relations. No equations, predictions, or first-principles derivations are presented that reduce reported quantities to the same fit parameters by construction. No load-bearing self-citation chains or ansatzes imported from prior author work are invoked to justify the central claims; the results are falsifiable external checks on independent data.
Axiom & Free-Parameter Ledger
free parameters (2)
- fAGN threshold of 0.8
- 3-sigma significance cut on broad-line flux and FWHM
axioms (2)
- domain assumption Stellar population synthesis templates in pPXF adequately represent the host galaxy light without AGN contamination
- domain assumption Broad-line region gas is virialized so that line width and luminosity yield black hole mass
Reference graph
Works this paper leans on
-
[1]
N., Adelman-McCarthy, J
Abazajian, K. N., Adelman-McCarthy, J. K., Agüeros, M. A., et al. 2009, ApJS, 182, 543 Abdurro’uf, Accetta, K., Aerts, C., et al. 2022, ApJS, 259, 35 Almeida, A., Anderson, S. F., Argudo-Fernández, M., et al. 2023, ApJS, 267, 44 Angthopo, J., Granett, B. R., La Barbera, F., et al. 2024, A&A, 690, A198 Aydar, C., Merloni, A., Dwelly, T., et al. 2025, A&A, ...
2009
-
[2]
validation is more realistic than relying on the statistical uncertainty
The bottom panel displays observations with𝑓 AGN >0.8, marked with empty markers; these are excluded from the Clean statistics estimate. validation is more realistic than relying on the statistical uncertainty. The choice to use 0.14 dex as the uncertainty is also supported byexternalvalidationtests(seeFig.7),whichyielded 0.3-0.4dexsystematicdifferencesac...
1994
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.