X-Shooting ULLYSES: Massive stars at low metallicity XV. On the metallicity dependence of B-supergiant mass-loss rates
Pith reviewed 2026-05-08 14:18 UTC · model grok-4.3
The pith
B-supergiant mass-loss rates exhibit weak metallicity dependence and lack the predicted bi-stability jump, providing new empirical constraints on line-driven wind models.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The metallicity trends present in the mass-loss prescriptions (Z^(0.41-1.4)) explored here overestimate the empirical metallicity dependence in the B-star regime, which appears very weak. We do not find an increase in mass-loss rate at approximately spectral type B1.
Load-bearing premise
That simultaneous fitting of UV and optical spectra with fastwind and Kiwi-GA yields unbiased wind parameters, including clumping factors, across the 13-29 kK temperature range without significant systematic errors from model assumptions or data quality.
read the original abstract
Context. For stellar evolution models we rely on mass-loss rate prescriptions that show features that lack empirical confirmation, such as the bi-stability jump. This jump is an increase in mass loss in the decreasing temperature regime Teff 28-21 kK. Although papers compared empirical results to prescriptions,a large observational sample of B stars for which the wind has been scrutinised over different metallicities is still lacking. Aims. By modelling of both ultraviolet (ULLYSES) and optical (XShootU) spectra, we determined the stellar and wind parameters, of 24 SMC B stars ranging in Teff from 13 to 29 kK. By combining this sample with LMC studies, we compared the wind behaviour of B stars in two different metallicity regimes. We compared our results to common mass-loss rate prescriptions to test features present in these models and their metallicity dependence. Methods. We have used the model atmosphere code fastwind and the genetic algorithm code Kiwi-GA to fit the UV and optical spectra simultaneously. This allows us to determine wind properties including clumping behaviour. Results. The metallicity trends present in the mass-loss prescriptions (Z^(0.41-1.4)) explored here overestimate the empirical metallicity dependence in the B-star regime, which appears very weak. We do not find an increase in mass-loss rate at approximately spectral type B1. We show that on average 40% of the wind mass is located in the wind medium between the clumps. We compiled a sample of 80+ O and B stars in the SMC and LMC. From a comparison we find a clear difference in O- and B-type metallicity dependence. Conclusions. The lack of a bi-stability jump in the B-star regime and a weak metallicity dependence for the same stars offers new empirical constraints to models of line-driven winds. As differences between these models are large (reaching orders of magnitude) such constraints are much needed.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper analyzes UV (ULLYSES) and optical (XShootU) spectra of 24 SMC B-supergiants spanning Teff = 13-29 kK using the fastwind atmosphere code and Kiwi-GA genetic algorithm for simultaneous fitting. This yields stellar and wind parameters including clumping factors. Combined with LMC data, the authors compile a sample of 80+ O and B stars and conclude that mass-loss prescriptions with Z^(0.41-1.4) scaling overestimate the empirical metallicity dependence (observed to be very weak in B-stars), that there is no bi-stability jump near spectral type B1, and that ~40% of the wind mass resides in the inter-clump medium on average. They highlight a difference in O- versus B-type metallicity dependence.
Significance. If the derived mass-loss rates and clumping parameters are free of Teff-dependent systematics, the work supplies important empirical constraints on line-driven wind theory in the B-supergiant regime at low metallicity. The reported weak Z-dependence and lack of a bi-stability jump challenge standard prescriptions and offer falsifiable inputs for stellar evolution calculations, strengthened by the combined SMC+LMC sample size and the explicit inter-clump mass fraction result.
major comments (2)
- [Methods section] Methods section (spectral fitting description): The central claim of weak empirical metallicity dependence and absence of the bi-stability jump rests on the assumption that fastwind + Kiwi-GA simultaneous UV/optical fits produce unbiased Mdot and clumping factors across 13-29 kK. The manuscript does not provide explicit validation tests (e.g., recovery of injected parameters in synthetic spectra or sensitivity to velocity-law and line-blanketing assumptions) for potential Teff-dependent biases in the two-component clumping prescription, which directly affects the reliability of the SMC-LMC comparison.
- [Results section] Results section (metallicity trend and bi-stability discussion): The statement that prescriptions overestimate the Z-scaling and that no jump occurs at B1 is presented without tabulated uncertainties on individual Mdot values or a quantitative assessment of the trend's statistical significance for the 24 SMC stars, undermining evaluation of whether the 'very weak' dependence is robust against sample variance or fitting degeneracies.
minor comments (2)
- [Abstract] The abstract would be clearer if it explicitly stated the number of LMC stars included in the 80+ sample and the precise temperature range over which the bi-stability jump is tested.
- [Results section] Notation for the inter-clump density parameter and the exact formula used to compute the 40% inter-clump mass fraction should be defined once in the text or a table to aid reproducibility.
Simulated Author's Rebuttal
We thank the referee for their constructive and detailed report. Their comments highlight important aspects of our methodology and presentation that we address below. We have revised the manuscript to incorporate clarifications and additional analysis where feasible.
read point-by-point responses
-
Referee: [Methods section] Methods section (spectral fitting description): The central claim of weak empirical metallicity dependence and absence of the bi-stability jump rests on the assumption that fastwind + Kiwi-GA simultaneous UV/optical fits produce unbiased Mdot and clumping factors across 13-29 kK. The manuscript does not provide explicit validation tests (e.g., recovery of injected parameters in synthetic spectra or sensitivity to velocity-law and line-blanketing assumptions) for potential Teff-dependent biases in the two-component clumping prescription, which directly affects the reliability of the SMC-LMC comparison.
Authors: We agree that dedicated validation tests for Teff-dependent biases would further strengthen confidence in the results. The fastwind code combined with the Kiwi-GA optimizer has been validated through recovery tests and sensitivity analyses in prior publications using similar UV+optical datasets and parameter ranges (e.g., our earlier XShootU papers on O stars). In the present work, we performed internal consistency checks by examining fit quality across the full Teff range and by testing variations in the velocity law and clumping parameters during the GA runs. To directly address the referee's point, we will add a new subsection in the Methods section summarizing these checks, referencing the literature validations, and discussing potential remaining systematics in the two-component clumping approach. This revision does not require new synthetic spectrum injections but clarifies the existing analysis. revision: yes
-
Referee: [Results section] Results section (metallicity trend and bi-stability discussion): The statement that prescriptions overestimate the Z-scaling and that no jump occurs at B1 is presented without tabulated uncertainties on individual Mdot values or a quantitative assessment of the trend's statistical significance for the 24 SMC stars, undermining evaluation of whether the 'very weak' dependence is robust against sample variance or fitting degeneracies.
Authors: The full parameter table (Table 2) already reports 1-sigma uncertainties on each Mdot value, derived directly from the Kiwi-GA posterior distributions. These uncertainties were used internally to assess the trends but were not explicitly highlighted in the metallicity discussion or plotted with error bars in the key figures. We will revise the Results section to (i) explicitly reference and display the tabulated Mdot uncertainties, (ii) add error bars to the relevant metallicity trend plots, and (iii) include a quantitative statistical assessment, such as a weighted linear regression of log Mdot versus log Z with bootstrap-derived significance, to evaluate robustness against sample variance and degeneracies. This will allow readers to judge the strength of the 'very weak' dependence claim directly. revision: yes
Circularity Check
No significant circularity in empirical B-supergiant mass-loss analysis
full rationale
The paper derives stellar and wind parameters (including mass-loss rates and clumping) for 24 SMC B stars by simultaneous fastwind + Kiwi-GA fitting of ULLYSES UV and XShootU optical spectra, then compares the resulting empirical Mdot values to LMC data and external theoretical prescriptions (Z^0.41-1.4 scaling, bi-stability features). No step reduces a claimed result to an input by construction: the weak empirical Z-dependence and absence of a B1 jump are direct observational outcomes from the fitted sample, not predictions of the same quantities or self-referential definitions. Self-citations to prior ULLYSES/XShootU papers exist but are not load-bearing; the central claims rest on new spectral modeling and cross-metallicity comparison rather than any fitted parameter or ansatz imported from the authors' own prior work. The analysis is self-contained against external benchmarks.
Axiom & Free-Parameter Ledger
free parameters (1)
- clumping factor and inter-clump density
axioms (1)
- domain assumption Line-driven wind theory governs mass loss in B-supergiants across the sampled temperature range
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.