Predicting Resolved Dust Attenuation from Local Galaxy Properties Using MaNGA
Pith reviewed 2026-05-15 19:14 UTC · model grok-4.3
The pith
Local star formation rate surface density predicts dust attenuation in galaxies to within a factor of 1.3 on kiloparsec scales.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
Using a sample of 5155 galaxies, A_V maps are derived from the Balmer decrement for nearly 1.9 million star-forming spaxels. The relation with Σ_SFR predicts A_V with R² = 0.69 and RMSE = 0.22 mag. The model reproduces A_V maps for various morphologies and recovers radial A_V profiles that depend on stellar mass and star formation activity.
What carries the argument
The empirical relation linking local star formation rate surface density (Σ_SFR) to visual dust attenuation A_V, fitted from MaNGA integral field data.
If this is right
- The predictions match observed A_V within a factor of about 1.3 on kpc scales.
- Iterative application corrects star formation rate surface density values and converges after four steps with residual bias near zero.
- Radial attenuation profiles are recovered for galaxies spanning ranges of stellar mass and star formation activity.
- The approach reproduces A_V maps for diverse morphologies including edge-on systems.
Where Pith is reading between the lines
- Surveys lacking full spectral coverage could approximate dust corrections if they can estimate local star formation rate surface density from other data.
- Applying the same relation at higher redshifts would test whether the link between star formation density and dust remains constant over cosmic time.
- Because residuals are Gaussian and centered at zero the model supplies statistically reliable corrections but may require galaxy-specific adjustments for precise work.
- Other local properties such as gas surface density could be tested as additional predictors to reduce the remaining scatter.
Load-bearing premise
The fitted relation between star formation rate surface density and dust attenuation holds universally enough to apply beyond the MaNGA sample without large systematic bias from the iterative correction.
What would settle it
A new sample of galaxies with independent Balmer decrement measurements where A_V predictions from Σ_SFR produce errors much larger than 0.22 mag or clearly non-Gaussian residuals would disprove the model's reliability.
read the original abstract
Accurate spatially resolved dust corrections are critical for interpreting the structure and evolution of star-forming galaxies (SFGs). We present an empirical model for predicting spatially resolved dust attenuation ($A_V$) in SFGs using integral field spectroscopy from the Mapping Nearby Galaxies at Apache Point Observatory (MaNGA) survey. Using a sample of 5,155 galaxies over $7.20<M_\ast<11.14$ and $0.0002 < z < 0.1444$, we derive $A_V$ maps from the Balmer decrement across more than 1,898,954 star-forming spaxels. Using local star formation rate surface density ($\Sigma_{\text{SFR}}$) as a predictor, the model achieves $R^2 = 0.69$ and RMSE $=0.22$ mag, with residuals that are approximately Gaussian and centred near zero. It predicts $A_V$ within a factor of $\sim$1.3 on kpc scales. We also demonstrate that the relation can be applied iteratively to recover dust-corrected $\Sigma_{\mathrm{SFR}}$ from uncorrected values, converging by the fourth iteration with minimal residual bias ($-0.01$ mag) and low RMSE ($0.42$ mag). The model accurately reproduces $A_V$ maps across diverse morphologies and orientations, including edge-on systems. It also recovers the observed radial $A_V$ profiles, capturing their dependence on stellar mass and relative star formation activity, with more massive and more strongly star-forming galaxies showing steeper gradients.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper presents an empirical model to predict spatially resolved dust attenuation A_V in star-forming galaxies using MaNGA integral-field spectroscopy. From Balmer-decrement A_V maps across >1.9 million spaxels in 5155 galaxies (7.20 < M* < 11.14, 0.0002 < z < 0.1444), it fits a relation to local star-formation-rate surface density Σ_SFR that yields R² = 0.69 and RMSE = 0.22 mag. The model is shown to recover observed A_V maps and radial profiles across morphologies; an iterative procedure is introduced to obtain dust-corrected Σ_SFR from uncorrected Hα-based values, reported to converge by the fourth iteration with −0.01 mag residual bias and 0.42 mag RMSE.
Significance. If the Σ_SFR–A_V mapping proves sufficiently universal, the work supplies a practical, observationally accessible correction for dust attenuation on kpc scales that does not require full Balmer-decrement coverage. The reproduction of radial gradients as a function of stellar mass and specific star-formation rate would be a useful addition to the toolkit for interpreting large IFS datasets, provided the iterative recovery step is shown to be stable and unbiased.
major comments (3)
- [Abstract / iterative method] Abstract and iterative-recovery section: the claim that the procedure converges by the fourth iteration with −0.01 mag bias is not accompanied by a demonstration that the fixed point remains unique and unbiased when the initial (uncorrected) Σ_SFR differs from the true value by the factor ∼1.3–2 expected for A_V ∼ 1–2 mag. No test is shown for cases in which scatter or secondary dependencies (metallicity, ionization parameter) correlate with A_V, which could shift the converged solution systematically.
- [Methods / sample and fitting] Methods on sample construction and model fitting: the manuscript provides no explicit description of the precise sample-selection cuts beyond the quoted mass and redshift ranges, the functional form and error treatment of the Σ_SFR–A_V fit, or any cross-validation on held-out spaxels or galaxies. Consequently the quoted R² = 0.69 and RMSE = 0.22 mag cannot be assessed for generalization beyond the training sample.
- [Results / radial profiles] Results on radial-profile recovery: while the model reproduces average radial A_V trends, the paper does not quantify how residuals vary with galactocentric radius or with galaxy inclination, leaving open whether the relation remains accurate in the high-attenuation central regions where the iterative correction is most needed.
minor comments (2)
- [Abstract] The abstract states that residuals are “approximately Gaussian,” but no quantitative test (e.g., Kolmogorov–Smirnov statistic or quantile–quantile plot) is referenced.
- [Methods] Notation for Σ_SFR is introduced without an explicit statement of the IMF or the exact conversion from Hα luminosity used to compute it.
Simulated Author's Rebuttal
We thank the referee for their constructive comments, which have helped us strengthen the presentation and validation of our empirical model. We address each major comment below and will revise the manuscript to incorporate the suggested improvements.
read point-by-point responses
-
Referee: [Abstract / iterative method] Abstract and iterative-recovery section: the claim that the procedure converges by the fourth iteration with −0.01 mag bias is not accompanied by a demonstration that the fixed point remains unique and unbiased when the initial (uncorrected) Σ_SFR differs from the true value by the factor ∼1.3–2 expected for A_V ∼ 1–2 mag. No test is shown for cases in which scatter or secondary dependencies (metallicity, ionization parameter) correlate with A_V, which could shift the converged solution systematically.
Authors: We agree that explicit tests of uniqueness and bias under realistic initial offsets are needed. In the revised manuscript we will add a new figure and accompanying text showing convergence when the iteration is initialized with Σ_SFR values scaled upward by factors of 1.3 and 2.0 relative to the true dust-corrected values. We will also report results from Monte-Carlo tests that inject additional scatter and introduce correlations between A_V and secondary parameters (metallicity, ionization parameter) to confirm that the fixed point remains stable with negligible systematic bias. revision: yes
-
Referee: [Methods / sample and fitting] Methods on sample construction and model fitting: the manuscript provides no explicit description of the precise sample-selection cuts beyond the quoted mass and redshift ranges, the functional form and error treatment of the Σ_SFR–A_V fit, or any cross-validation on held-out spaxels or galaxies. Consequently the quoted R² = 0.69 and RMSE = 0.22 mag cannot be assessed for generalization beyond the training sample.
Authors: We will expand the Methods section with a complete description of all sample-selection criteria (including S/N thresholds, BPT classification boundaries, and spaxel quality cuts). The functional form of the fit will be stated explicitly as an orthogonal-distance-regression linear relation between log Σ_SFR and A_V, with full details on error propagation. We will also add a cross-validation subsection reporting R² and RMSE obtained from 5-fold splits on both held-out spaxels and held-out galaxies to demonstrate generalization. revision: yes
-
Referee: [Results / radial profiles] Results on radial-profile recovery: while the model reproduces average radial A_V trends, the paper does not quantify how residuals vary with galactocentric radius or with galaxy inclination, leaving open whether the relation remains accurate in the high-attenuation central regions where the iterative correction is most needed.
Authors: We will augment the radial-profile analysis with quantitative residual statistics. The revised manuscript will include binned mean, median, and scatter of A_V residuals versus normalized radius R/R_e, shown separately for different stellar-mass and sSFR bins. We will further stratify the residuals by galaxy inclination (using b/a axis ratio) and will highlight performance specifically in the central (R < 0.5 R_e) regions to address concerns about high-attenuation zones. revision: yes
Circularity Check
In-sample empirical fit to MaNGA spaxels presented as out-of-sample prediction with iterative recovery on identical data
specific steps
-
fitted input called prediction
[Abstract]
"Using local star formation rate surface density (Σ_SFR) as a predictor, the model achieves R² = 0.69 and RMSE =0.22 mag, with residuals that are approximately Gaussian and centred near zero. It predicts A_V within a factor of ∼1.3 on kpc scales."
The R², RMSE and factor-of-1.3 claims are the direct goodness-of-fit statistics of the empirical relation fitted to the identical MaNGA star-forming spaxels whose Balmer-derived A_V and dust-corrected Σ_SFR were used to construct the predictor; no train/test split or external sample is invoked.
-
fitted input called prediction
[Abstract]
"We also demonstrate that the relation can be applied iteratively to recover dust-corrected Σ_SFR from uncorrected values, converging by the fourth iteration with minimal residual bias (−0.01 mag) and low RMSE (0.42 mag)."
The iterative procedure re-applies the identical fitted Σ_SFR–A_V mapping to attenuated Hα-based Σ_SFR values derived from the same MaNGA observations; the reported convergence bias and RMSE are evaluated against the original Balmer-derived A_V used in the fit, making the fixed-point result a direct consequence of the in-sample relation.
full rationale
The paper derives A_V from Balmer decrement on >1.8M MaNGA spaxels, fits a direct Σ_SFR–A_V relation, and reports R²/RMSE as model performance. It then applies the same fitted relation iteratively to recover corrected Σ_SFR starting from attenuated values on the same spaxels, quoting convergence metrics against the original Balmer A_V. Both the quoted predictive accuracy and the iteration's low bias are therefore in-sample fit properties rather than independent tests, producing moderate circularity without external validation or held-out data.
Axiom & Free-Parameter Ledger
free parameters (1)
- coefficients of the A_V versus Σ_SFR relation
axioms (1)
- domain assumption The Balmer decrement provides an accurate measure of visual extinction A_V under case-B recombination conditions in star-forming regions
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.