Implications of a contracted dark matter halo for the Milky Way's inferred virial mass

arXiv:2603.13516 · astro-ph.GA · Submitted 2026-03-13 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: We're looking at the paper: "Implications of a contracted dark matter halo for the Milky Way's inferred virial mass".

Jocelyn: Here's the gist of the paper. The paper investigates how baryonic processes modify dark matter halo structure and how this impacts the reliability of inferring global halo properties from dynamical data restricted to inner radii.

Vera: The key finding is that explicitly modeling baryon-induced contraction, using the Contracted NFW (CNFW) model, provides the most unbiased inferences for halo mass and concentration when observational constraints are limited to inner regions Baryonic Effects on Inner Halo Structure The study examines how baryonic physics alters the central dark matter distribution within Milky Way-mass haloes using ARTEMIS simulations The most striking consequence of baryonic physics in Milky Way–mass systems is the so-called contraction of the dark matter distribution, i.e.

Jocelyn: an enhancement in the central DM density compared to the DMO counterpart of a halo This effect is illustrated in Fig.

Vera: 2, which shows the DM density profiles for three representative ARTEMIS haloes (G40, G18 and G14) Baryonic condensation enhances the central dark matter density typically for r ≲ 20 kpc Effectively, it steepens the profile slope and brings it closer to an isothermal value This is evident for G40 (pink), where the hydrodynamical profile maintains a near-constant slope of-2 over r ≈ 2–20 kpc Model Performance and Radial Coverage The authors test the performance of four widely used theoretical profiles—Einasto, NFW, gNFW, and Contracted NFW (CNFW)—to spatially restricted regions of simulated haloes They find that excellent local fits do not guarantee accurate extrapolation when constrained only by radii ≲ 20–30 kpc Specifically, NFW, gNFW, and Einasto models infer systematically biased halo mass and concentrations even when their residuals within the fitted domain are minimal The CNFW model exhibits substantially more stable behaviour, returning extrapolated halo masses within ∼ 0.06 dex (15%) of the true value on average even when fitted only to ∼ 20 kpc Systematic Biases in Extrapolated Properties When using restricted radial ranges, NFW fits systematically underestimate halo mass and overestimate concentrations The biases arise due to baryon-induced contraction creating a broad isothermal shell that strongly influences local curvature but carries little information about the outer halo Models without explicit contraction physics compensate by distorting their fitted parameters, typically producing over-concentrated and under-massive haloes The Contracted NFW (CNFW) model is the only model that avoids the strong systematic biases observed for the NFW, gNFW, and Einasto profiles Implications for Mass Inference The results suggest that kinematic probes residing in the outer halo should be considered of higher fidelity when attempting a precision estimate of the Milky Way’s halo mass Furthermore, increasing the maximum fitting radius toward the typical virial radius substantially improves the recovery of halo mass and, to a lesser extent, concentration for all profile families The authors conclude that baryon-induced contraction is both universal and dynamically significant, and accounting for it is essential for achieving unbiased inferences of the global properties of Milky Way–mass dark matter haloes In summary, the Contracted NFW (CNFW) model provides the least biased inferences of global halo properties, M200c and c, when observational constraints are limited to inner regions This work advocates for the routine incorporation of contraction-aware models when interpreting mass tracers confined to the inner ∼ 30 kpc The findings underscore the need for contraction-aware models in dynamical studies of the Milky Way halo, even more so in efforts to interpret the concentration–mass relation in real galaxies SOFTWARE The following software packages were used in this work: numpy (Harris et al.

Paper summary: Vera: So to summarize this paper "Implications of a contracted dark matter halo for the Milky Way's inferred virial mass", their core argument is that baryonic processes significantly modify the central structure of dark matter haloes, leading to contraction.

Jocelyn: They use ARTEMIS simulations to show how this contraction makes the central dark matter density higher than what you would see in a dark matter only simulation.

Subrahmanyan: The paper claims that this effect complicates our task of inferring the host halo properties from galactic dynamics because the galaxy itself influences the halo, and this is expected to be most significant where we get our best dynamical data, which is often in the disk.

Vera: So, they are saying that without a reliable theoretical model to account for this baryonic contraction, we run into systematic errors when trying to figure out how massive these Milky Way-mass haloes actually are from observations confined to inner radii.

Jocelyn: The key takeaway is that using the Contracted NFW model provides the most unbiased inferences for halo mass and concentration when observational constraints are limited to those inner regions, because it accounts for that contraction.

Conclusion: Vera: So, looking at the title, "Implications of a contracted dark matter halo for the Milky Way's inferred virial mass," it really boils down to how much our measurements are skewed by ignoring these baryonic effects when we try to calculate the total mass of our host halo.

Jocelyn: Essentially, they are arguing that if you only look at the inner parts of a galaxy, you need a model like Contracted NFW because without it, standard models give you systematically biased results for both mass and concentration.

Subrahmanyan: The implication is that for any cosmological study relying on Milky Way-sized haloes, especially when using dynamical data from the inner regions, we really need to incorporate contraction-aware models to get a realistic picture of the halo's properties.

Vera: And they emphasize that increasing the fitting radius up toward the virial radius substantially helps recover better mass and concentration estimates for all profile families, which suggests that getting more data out there is always beneficial.

Jocelyn: So, even though there's this scatter in the results, it strongly implies that we need to move past simple assumptions about dark matter profiles and start modeling how the galaxy interacts with the halo structure itself.

Institute for Computational Cosmology, Durham University · Centre for Extragalactic Astronomy, Durham University Centre for Extragalactic Astronomy, Stockholm University Astrophysics Research Institute, Liverpool John Moores University

astro-ph.GA

Submitted: 2026-03-13

Updated: 2026-11-05

Comments: Accepted for publication in MNRAS

Journal ref: MNRAS, 552, stag1865 (2026)

DOI: 10.1093/mnras/stag1865

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 68/100

The gist: The paper investigates how baryonic processes modify dark matter halo structure and how this impacts the reliability of inferring global halo properties from dynamical data restricted to inner radii.

Key concepts

Baryonic Contraction
This refers to how normal matter, like gas and stars, pulls the surrounding dark matter inward in the center of a galaxy. In Milky Way-mass systems, this effect enhances the central dark matter density compared to what would be expected from a purely dark matter halo model.
NFW Profile
The Navarro-Frenk-White profile is a commonly used mathematical function to describe the density distribution of dark matter halos. While widely adopted, this simple model often fails to accurately represent the true inner structure when baryonic effects are present.
Contracted NFW (CNFW) Model
This is an improved mathematical model that explicitly accounts for the inward pull caused by baryonic physics. When fitting data from only the inner regions, this model yields stable and unbiased estimates for the halo's mass and concentration, unlike simpler models.
Radial Coverage Bias
When analyzing dynamical data restricted to small radii (like inner 30 kpc), standard dark matter profiles introduce systematic errors. The authors show that these biases are significant, meaning conclusions about the total halo mass can be inaccurate unless contraction effects are explicitly included.

Summary

The paper investigates how baryonic processes modify dark matter halo structure and how this impacts the reliability of inferring global halo properties from dynamical data restricted to inner radii. p. 0

Where the paper says it

Received YYY; in original form ZZZ ABSTRACT We investigate how reliably the global properties of Milky Way–mass dark matter haloes can be recovered from dynamical data from a limited range of radii, particular ≲ 30 kpc where observations are most sensitive but baryonic processes modify the halo structure. Using the ARTEMIS simulations, which produce varying degrees of baryon-induced contraction, we fit dark matter profiles over restricted radial ranges using commonly adopted parametric models.

The key finding is that explicitly modeling baryon-induced contraction, using the Contracted NFW (CNFW) model, provides the most unbiased inferences for halo mass and concentration when observational constraints are limited to inner regions

Baryonic Effects on Inner Halo Structure

The study examines how baryonic physics alters the central dark matter distribution within Milky Way-mass haloes using ARTEMIS simulations The most striking consequence of baryonic physics in Milky Way–mass systems is the so-called contraction of the dark matter distribution, i.e. an enhancement in the central DM density compared to the DMO counterpart of a halo This effect is illustrated in Fig. p. 4

Where the paper says it

an enhancement in the central DM density compared to the DMO counterpart of a halo. This effect is illustrated in Fig.

2, which shows the DM density profiles for three representative ARTEMIS haloes (G40, G18 and G14) Baryonic condensation enhances the central dark matter density typically for r ≲ 20 kpc Effectively, it steepens the profile slope and brings it closer to an isothermal value This is evident for G40 (pink), where the hydrodynamical profile maintains a near-constant slope of-2 over r ≈ 2–20 kpc

Model Performance and Radial Coverage

The authors test the performance of four widely used theoretical profiles—Einasto, NFW, gNFW, and Contracted NFW (CNFW)—to spatially restricted regions of simulated haloes They find that excellent local fits do not guarantee accurate extrapolation when constrained only by radii ≲ 20–30 kpc Specifically, NFW, gNFW, and Einasto models infer systematically biased halo mass and concentrations even when their residuals within the fitted domain are minimal The CNFW model exhibits substantially more stable behaviour, returning extrapolated halo masses within ∼ 0.06 dex (15%) of the true value on average even when fitted only to ∼ 20 kpc

Systematic Biases in Extrapolated Properties

When using restricted radial ranges, NFW fits systematically underestimate halo mass and overestimate concentrations The biases arise due to baryon-induced contraction creating a broad isothermal shell that strongly influences local curvature but carries little information about the outer halo Models without explicit contraction physics compensate by distorting their fitted parameters, typically producing over-concentrated and under-massive haloes The Contracted NFW (CNFW) model is the only model that avoids the strong systematic biases observed for the NFW, gNFW, and Einasto profiles

Implications for Mass Inference

The results suggest that kinematic probes residing in the outer halo should be considered of higher fidelity when attempting a precision estimate of the Milky Way’s halo mass Furthermore, increasing the maximum fitting radius toward the typical virial radius substantially improves the recovery of halo mass and, to a lesser extent, concentration for all profile families The authors conclude that baryon-induced contraction is both universal and dynamically significant, and accounting for it is essential for achieving unbiased inferences of the global properties of Milky Way–mass dark matter haloes

Conclusion

In summary, the Contracted NFW (CNFW) model provides the least biased inferences of global halo properties, M200c and c, when observational constraints are limited to inner regions This work advocates for the routine incorporation of contraction-aware models when interpreting mass tracers confined to the inner ∼ 30 kpc The findings underscore the need for contraction-aware models in dynamical studies of the Milky Way halo, even more so in efforts to interpret the concentration–mass relation in real galaxies

SOFTWARE

The following software packages were used in this work: numpy (Harris et al. 2020), scipy3, pandas (The pandas development team 2020), astropy (Astropy Collaboration et al. p. 0

Where the paper says it

In contrast, the contracted halo prescription from Cautun et al. (2020) yields stable extrapolations and recovers unbiased halo mass estimates over all radii.

2022), h5py (Collette et al. 2022)

ACKNOWLEDGEMENTS

DD acknowledges support by the Royal Society through a Dorothy Hodgkin Fellowship (DHF/R1/231105). S.T.B. and A.F. acknowledge support by a UK Research and Innovation (UKRI) Future Leaders Fellowship (grant no MR/T042362/1) and Sweden’s Wallenberg Academy Fellowship IGM acknowledges support from STFC grant ST/Y002733/1

DATA AVAILABILITY

The data underlying the plots in this article are available from the corresponding author upon request The raw data from the ARTEMIS simulation suite are not publicly available; researchers wishing to access the simulations may contact the ARTEMIS collaboration

REFERENCES

Abadi M. G., Navarro J. p. 0

Where the paper says it

Reliable theoretical models for the distribution of mass with DM are necessary to be able to robustly interpret kinematic data. In the absence of baryonic effects it is well established that haloes follow on average a Navarro–Frenk–White (NFW; Navarro et al.

p. 0
Where the paper says it

Reliable theoretical models for the distribution of mass with DM are necessary to be able to robustly interpret kinematic data. In the absence of baryonic effects it is well established that haloes follow on average a Navarro–Frenk–White (NFW; Navarro et al.

p. 0
Where the paper says it

Reliable theoretical models for the distribution of mass with DM are necessary to be able to robustly interpret kinematic data. In the absence of baryonic effects it is well established that haloes follow on average a Navarro–Frenk–White (NFW; Navarro et al.

F., Fardal M., Babul A., Steinmetz M., 2010, MNRAS, 407, 435 Astropy Collaboration et al. p. 1
Where the paper says it

Recent studies using large precise data sets from APOGEE DR17 (Majewski, S. , 2017) and Gaia DR3 (Gaia Collaboration, et al.

, 2022, ApJ, 935, 167 Binney J. p. 0
Where the paper says it

Key words: cosmology: dark matter – galaxies: kinematics and dynamics – galaxies: evolution 1 INTRODUCTION In the ΛCDM cosmogony, galaxies form at the centres of virialised dark matter (DM) haloes, with their subsequent evolution shaped by the interplay between hierarchical accretion, gas cooling, and feedback from stars and active galactic nuclei (e. Press & Schechter 1974; Binney 1977; Frenk et al.

, 1977, ApJ, 215, 483 Brown S. p. 0
Where the paper says it

Key words: cosmology: dark matter – galaxies: kinematics and dynamics – galaxies: evolution 1 INTRODUCTION In the ΛCDM cosmogony, galaxies form at the centres of virialised dark matter (DM) haloes, with their subsequent evolution shaped by the interplay between hierarchical accretion, gas cooling, and feedback from stars and active galactic nuclei (e. Press & Schechter 1974; Binney 1977; Frenk et al.

T., McCarthy I. G., Diemer B. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

, Font A. S., Stafford S. G., Pfeifer S., 2020, MNRAS, 495, 4994 Bullock J. p. 0
Where the paper says it

Even within this context, observations within the Milky Way (MW) offer unparalleled detail to map the DM distribution within our host halo (e. reviews by Bullock & Boylan-Kolchin 2017; Zavala & Frenk 2019).

S., Boylan-Kolchin M. p. 0
Where the paper says it

Even within this context, observations within the Milky Way (MW) offer unparalleled detail to map the DM distribution within our host halo (e. reviews by Bullock & Boylan-Kolchin 2017; Zavala & Frenk 2019).

, 2017, ARA&A, 55, 343 Bullock J. p. 0
Where the paper says it

Even within this context, observations within the Milky Way (MW) offer unparalleled detail to map the DM distribution within our host halo (e. reviews by Bullock & Boylan-Kolchin 2017; Zavala & Frenk 2019).

S., Kolatt T. S., Sigad Y., Somerville R. S., Kravtsov A. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

V., Klypin A. A., Primack J. R., Dekel A., 2001, MNRAS, 321, 559 Callingham T. p. 0
Where the paper says it

2017; Patel et al. 2018; Callingham et al.

M., et al., 2019, MNRAS, 484, 5453 Callingham T. p. 1
Where the paper says it

Watkins et al. 2019; Callingham et al.

M., Cautun M. p. 0
Where the paper says it

Einasto and generalised-NFW models provide excellent local fits but retain similar global biases. In contrast, the contracted halo prescription from Cautun et al.

p. 0
Where the paper says it

Einasto and generalised-NFW models provide excellent local fits but retain similar global biases. In contrast, the contracted halo prescription from Cautun et al.

, Deason A. p. 0
Where the paper says it

uk beyond the central disk, including the stellar halo (e. 2008; Deason et al.

p. 0
Where the paper says it

uk beyond the central disk, including the stellar halo (e. 2008; Deason et al.

J., Frenk C. S., Grand R. J. J., Marinacci F., Pakmor R., 2020a, MNRAS, 495, 12 Callingham T. p. 0
Where the paper says it

2017; Patel et al. 2018; Callingham et al.

M., Cautun M., Deason A. J., Frenk C. S., Grand R. J. J., Marinacci F., Pakmor R., 2020b, MNRAS, 495, 12 Cautun M. p. 0
Where the paper says it

Einasto and generalised-NFW models provide excellent local fits but retain similar global biases. In contrast, the contracted halo prescription from Cautun et al.

, et al., 2020, MNRAS, 494, 4291 Cimatti A. p. 0
Where the paper says it

Press & Schechter 1974; Binney 1977; Frenk et al. 1988; Cimatti et al.

, Fraternali F., Nipoti C., 2019, arXiv e-prints, p. p. 0
Where the paper says it

, the radius enclosing a mean matter density 200 times the critical density at 𝑧 = 0. While the internal structure of DM haloes in the absence of baryons is relatively well understood, it has been shown that the inclusion of © 2026 The Authors arXiv:2603.

arXiv:1912. p. 0
Where the paper says it

, the radius enclosing a mean matter density 200 times the critical density at 𝑧 = 0. While the internal structure of DM haloes in the absence of baryons is relatively well understood, it has been shown that the inclusion of © 2026 The Authors arXiv:2603.

06216 Collette A., et al., 2022, h5py/h5py: 3.7.0, doi:10.5281/zenodo.6575970 Davis M. p. 1
Where the paper says it

This yields a sample of 42 central haloes. 2 Halo finding & virial parameters Haloes are identified using the friends-of-friends (FOF) algorithm (Davis et al.

, Efstathiou G., Frenk C. S., White S. D. M., 1985, ApJ, 292, 371 Deason A. p. 0
Where the paper says it

uk beyond the central disk, including the stellar halo (e. 2008; Deason et al.

J., et al., 2012, MNRAS, 425, 2840 Di Cintio A. p. 1
Where the paper says it

In some cases strong feedback driven gaseous outflows can lead to the formation of a DM core due to the energy transfer from the fluctuating potential (e. Pontzen & Governato 2012; Di Cintio et al.

, Brook C. B., Macciò A. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

V., Stinson G. S., Knebe A., Dutton A. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

A., Wadsley J., 2014a, MNRAS, 437, 415 Di Cintio A. p. 1
Where the paper says it

Pontzen & Governato 2012; Di Cintio et al. 2014a), with this behaviour typically observed in the dwarf galaxy regime.

, Brook C. B., Dutton A. A., Macciò A. V., Stinson G. S., Knebe A., 2014b, MNRAS, 441, 2986 Diemer B. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

, Kravtsov A. V., 2015, ApJ, 799, 108 Drukier A. p. 1
Where the paper says it

2019; Callingham et al. Reconciling such discrepancies is key for a broad range of applications, from predictions of Local Group satellite abundances to the local DM density relevant for directdetection experiments (Goodman & Witten 1985; Drukier et al.

K., Freese K., Spergel D. N., 1986, Phys. p. 1
Where the paper says it

Reconciling such discrepancies is key for a broad range of applications, from predictions of Local Group satellite abundances to the local DM density relevant for directdetection experiments (Goodman & Witten 1985; Drukier et al. 1986; Jungman et al.

Rev. D, 33, 3495 Dutton A. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

A., Macciò A. V., 2014, MNRAS, 441, 3359 Eadie G. p. 0
Where the paper says it

2012), globular clusters (e. Eadie et al.

M., Springford A., Harris W. E., 2017, ApJ, 835, 167 Eilers A. p. 1
Where the paper says it

, 2023), but restricted to the disk, have inferred the mass to be ≈ 1011–5 × 1011 M⊙, depending on the assumed profile (e. Eilers et al.

-C., Hogg D. W., Rix H.-W., Ness M. K., 2019, ApJ, 871, 120 Einasto J. p. 0
Where the paper says it

1988; Cimatti et al. 2019; Marasco et al.

, 1965, Trudy Astrofizicheskogo Instituta Alma-Ata, 5, 87 Font A. p. 3
Where the paper says it

(7) To ensure physically plausible fits, we constrain the slope parameter to the range 𝛾 ∈ [0, 2]. 3 Einasto The Einasto profile (Einasto 1965) is another widely adopted model for describing dark matter halo density profiles, both in simulations and observational studies.

S., et al., 2020, MNRAS, 498, 1765 Frenk C. p. 0
Where the paper says it

In contrast, the contracted halo prescription from Cautun et al. (2020) yields stable extrapolations and recovers unbiased halo mass estimates over all radii.

S., White S. D. M., Davis M., Efstathiou G., 1988, ApJ, 327, 507 Gaia Collaboration, et al. p. 1
Where the paper says it

Recent studies using large precise data sets from APOGEE DR17 (Majewski, S. , 2017) and Gaia DR3 (Gaia Collaboration, et al.

, 2023, A&A, 674, A1 Gnedin O. p. 1
Where the paper says it

, 2017) and Gaia DR3 (Gaia Collaboration, et al. , 2023), but restricted to the disk, have inferred the mass to be ≈ 1011–5 × 1011 M⊙, depending on the assumed profile (e.

Y., Kravtsov A. V., Klypin A. A., Nagai D., 2004, ApJ, 616, 16 Goodman M. p. 1
Where the paper says it

2019; Callingham et al. Reconciling such discrepancies is key for a broad range of applications, from predictions of Local Group satellite abundances to the local DM density relevant for directdetection experiments (Goodman & Witten 1985; Drukier et al.

W., Witten E. p. 1
Where the paper says it

2019; Callingham et al. Reconciling such discrepancies is key for a broad range of applications, from predictions of Local Group satellite abundances to the local DM density relevant for directdetection experiments (Goodman & Witten 1985; Drukier et al.

, 1985, Phys. p. 1
Where the paper says it

2019; Callingham et al. Reconciling such discrepancies is key for a broad range of applications, from predictions of Local Group satellite abundances to the local DM density relevant for directdetection experiments (Goodman & Witten 1985; Drukier et al.

Rev. D, 31, 3059 Grand R. p. 4
Where the paper says it

Schaller et al. 2015; Grand et al.

J. J., et al., 2017, MNRAS, 467, 179 Harris C. p. 0
Where the paper says it

Even within this context, observations within the Milky Way (MW) offer unparalleled detail to map the DM distribution within our host halo (e. reviews by Bullock & Boylan-Kolchin 2017; Zavala & Frenk 2019).

R., et al., 2020, Nature, 585, 357 Hinshaw G. p. 1
Where the paper says it

As such, ARTEMIS is well suited for studying generic baryonic effects on dark matter haloes – including baryonic contraction – and for quantifying systematic biases in halo property inference from kinematic data confined to the inner regions of galaxies. The ARTEMIS simulations adopt a flat ΛCDM cosmology consistent with WMAP9 (Hinshaw et al.

, et al., 2013, ApJS, 208, 19 Hunter D. p. 0
Where the paper says it

1996; Walter et al. 2008; Hunter et al.

A., et al., 2012, AJ, 144, 134 Jungman G. p. 1
Where the paper says it

Reconciling such discrepancies is key for a broad range of applications, from predictions of Local Group satellite abundances to the local DM density relevant for directdetection experiments (Goodman & Witten 1985; Drukier et al. 1986; Jungman et al.

, Kamionkowski M., Griest K., 1996, Phys. p. 0
Where the paper says it

Persic et al. 1996; Walter et al.

Rep., 267, 195 Kafle P. p. 0
Where the paper says it

2008; Deason et al. 2012; Kafle et al.

R., Sharma S., Lewis G. F., Bland-Hawthorn J., 2012, ApJ, 761, 98 Karukes E. p. 8
Where the paper says it

Degeneracies between 𝑟𝑠 and 𝛾 for the gNFW model have already been highlighted when fitting the Milky Way’s rotation curve (e. Karukes et al.

V., Benito M., Iocco F., Trotta R., Geringer-Sameth A., 2019, J. Cosmology Astropart. p. 0
Where the paper says it

Font4 and Ian G. McCarthy4 1 Institute for Computational Cosmology, Department of Physics, Durham University, South Road, Durham DH1 3LE, UK 2 Centre for Extragalactic Astronomy, Department of Physics, Durham University, South Road, Durham DH1 3LE, UK 3 The Oskar Klein Centre, Department of Physics, Stockholm University, Albanova University Center, 106 91 Stockholm, Sweden 4 Astrophysics Research Institute, Liverpool John Moores University, 146 Brownlow Hill, Liverpool L3 5RF, UK Accepted XXX.

p. 0
Where the paper says it

Font4 and Ian G. McCarthy4 1 Institute for Computational Cosmology, Department of Physics, Durham University, South Road, Durham DH1 3LE, UK 2 Centre for Extragalactic Astronomy, Department of Physics, Durham University, South Road, Durham DH1 3LE, UK 3 The Oskar Klein Centre, Department of Physics, Stockholm University, Albanova University Center, 106 91 Stockholm, Sweden 4 Astrophysics Research Institute, Liverpool John Moores University, 146 Brownlow Hill, Liverpool L3 5RF, UK Accepted XXX.

Phys., 2019, 046 Li Z. p. 0
Where the paper says it

1988; Cimatti et al. 2019; Marasco et al.

-Z., Jing Y. P., Qian Y.-Z., Yuan Z., Zhao D.-H., 2017, ApJ, 850, 116 Ludlow A. p. 0
Where the paper says it

Cosmological simulations predict these to be strongly correlated leading to a well defined concentration-mass relation that varies with redshift and cosmology (e. Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al.

D., Bose S., Angulo R. E., Wang L., Hellwing W. A., Navarro J. F., Cole S., Frenk C. S., 2016, MNRAS, 460, 1214 Ludlow A. p. 0
Where the paper says it

Dutton & Macciò 2014; Diemer & Kravtsov 2015; Ludlow et al. 2016; Brown et al.

D., Schaye J. p. 1
Where the paper says it

2020) is a suite of 42 cosmological zoom-in hydrodynamical simulations of Milky Way-mass haloes. These simulations are run using the Gadget-3 code (last described in Springel 2005) with the same hydrodynamics solver and galaxy formation (subgrid) physics as the EAGLE project (Schaye et al.

, Bower R., 2019, MNRAS, 488, 3663 Majewski, S. p. 1
Where the paper says it

In recent years a tension in the inferred Milky Way mass has arisen between observations from dynamics in the central Galaxy (5 < 𝑟/kpc ≲ 30) and those using from the outer halo (𝑟 ≳ 30 kpc). Recent studies using large precise data sets from APOGEE DR17 (Majewski, S.

R. et al., 2017, AJ, 154, 94 Marasco A. p. 0
Where the paper says it

1988; Cimatti et al. 2019; Marasco et al.

, Cresci G., Posti L., Fraternali F., Mannucci F., Marconi A., Belfiore F., Fall S. M., 2021, MNRAS, 507, 4274 Navarro J. p. 0
Where the paper says it

Reliable theoretical models for the distribution of mass with DM are necessary to be able to robustly interpret kinematic data. In the absence of baryonic effects it is well established that haloes follow on average a Navarro–Frenk–White (NFW; Navarro et al.

F., Frenk C. S., White S. D. M., 1997, ApJ, 490, 493 Ou X. p. 0
Where the paper says it

In the absence of baryonic effects it is well established that haloes follow on average a Navarro–Frenk–White (NFW; Navarro et al. 1997) profile, which exhibits an inner density slope of 𝜌 ∝ 𝑟 −1 and an outer slope of 𝜌 ∝ 𝑟 −3 , with a characteristic scale radius, 𝑟s , describing the transitions between the two limits.

, Eilers A. p. 1
Where the paper says it

, 2023), but restricted to the disk, have inferred the mass to be ≈ 1011–5 × 1011 M⊙, depending on the assumed profile (e. Eilers et al.

-C., Necib L., Frebel A., 2024, MNRAS, 528, 693 Ou X. p. 1
Where the paper says it

2019; Ou et al. 2024), while observations using tracers beyond the disk imply a more common mass of ≈ 1012M⊙ (e.

, Necib L., Wetzel A., Frebel A., Bailin J., Oeur M., 2025, ApJ, 994, 128 Patel E. p. 0
Where the paper says it

2010; Li et al. 2017; Patel et al.

, Besla G., Mandel K., Sohn S. T., 2018, ApJ, 857, 78 Persic M. p. 0
Where the paper says it

With dedicated surveys we now have robust kinematic data for large samples of galaxies in the nearby Universe (e. Persic et al.

, Salucci P., Stel F., 1996, MNRAS, 281, 27 Pillepich A. p. 0
Where the paper says it

Persic et al. 1996; Walter et al.

, et al., 2018, MNRAS, 473, 4077 Pontzen A. p. 1
Where the paper says it

In some cases strong feedback driven gaseous outflows can lead to the formation of a DM core due to the energy transfer from the fluctuating potential (e. Pontzen & Governato 2012; Di Cintio et al.

, Governato F. p. 1
Where the paper says it

In some cases strong feedback driven gaseous outflows can lead to the formation of a DM core due to the energy transfer from the fluctuating potential (e. Pontzen & Governato 2012; Di Cintio et al.

, 2012, MNRAS, 421, 3464 Poole-McKenzie R. p. 1
Where the paper says it

2004; Schaller et al. 2015; Poole-McKenzie et al.

, Font A. S., Boxer B., McCarthy I. G., Burdin S., Stafford S. G., Brown S. T., 2020, J. Cosmology Astropart. Phys., 2020, 016 Posti L. p. 0
Where the paper says it

Eadie et al. 2017; Posti & Helmi 2019; Watkins et al.

, Helmi A., 2019, A&A, 621, A56 Power C. p. 0
Where the paper says it

1988; Cimatti et al. 2019; Marasco et al.

, Navarro J. F., Jenkins A., Frenk C. S., White S. D. M., Springel V. p. 1
Where the paper says it

2020) is a suite of 42 cosmological zoom-in hydrodynamical simulations of Milky Way-mass haloes. These simulations are run using the Gadget-3 code (last described in Springel 2005) with the same hydrodynamics solver and galaxy formation (subgrid) physics as the EAGLE project (Schaye et al.

, Stadel J., Quinn T., 2003, MNRAS, 338, 14 Press W. p. 0
Where the paper says it

Key words: cosmology: dark matter – galaxies: kinematics and dynamics – galaxies: evolution 1 INTRODUCTION In the ΛCDM cosmogony, galaxies form at the centres of virialised dark matter (DM) haloes, with their subsequent evolution shaped by the interplay between hierarchical accretion, gas cooling, and feedback from stars and active galactic nuclei (e. Press & Schechter 1974; Binney 1977; Frenk et al.

H., Schechter P. p. 0
Where the paper says it

Key words: cosmology: dark matter – galaxies: kinematics and dynamics – galaxies: evolution 1 INTRODUCTION In the ΛCDM cosmogony, galaxies form at the centres of virialised dark matter (DM) haloes, with their subsequent evolution shaped by the interplay between hierarchical accretion, gas cooling, and feedback from stars and active galactic nuclei (e. Press & Schechter 1974; Binney 1977; Frenk et al.

, 1974, ApJ, 187, 425 Rubin V. p. 0
Where the paper says it

Key words: cosmology: dark matter – galaxies: kinematics and dynamics – galaxies: evolution 1 INTRODUCTION In the ΛCDM cosmogony, galaxies form at the centres of virialised dark matter (DM) haloes, with their subsequent evolution shaped by the interplay between hierarchical accretion, gas cooling, and feedback from stars and active galactic nuclei (e. Press & Schechter 1974; Binney 1977; Frenk et al.

C., Ford Jr. W. K., Thonnard N., 1980, ApJ, 238, 471 Sawala T. p. 2
Where the paper says it

For example it has been shown that for dwarf galaxies there is notable change to the DM halo mass (e. Sawala et al.

, et al., 2015, MNRAS, 448, 2941 Schaller M. p. 1
Where the paper says it

2004; Schaller et al. 2015; Poole-McKenzie et al.

, et al., 2015, MNRAS, 451, 1247 Schaye J. p. 1
Where the paper says it

These simulations are run using the Gadget-3 code (last described in Springel 2005) with the same hydrodynamics solver and galaxy formation (subgrid) physics as the EAGLE project (Schaye et al. 2015), but at a significantly increased resolution.

, et al., 2015, MNRAS, 446, 521 Sellwood J. p. 4
Where the paper says it

Gnedin et al. 2004; Sellwood & McGaugh 2005; Callingham et al.

A., McGaugh S. p. 4
Where the paper says it

Gnedin et al. 2004; Sellwood & McGaugh 2005; Callingham et al.

S., 2005, ApJ, 634, 70 Springel V. p. 1
Where the paper says it

2020) is a suite of 42 cosmological zoom-in hydrodynamical simulations of Milky Way-mass haloes. These simulations are run using the Gadget-3 code (last described in Springel 2005) with the same hydrodynamics solver and galaxy formation (subgrid) physics as the EAGLE project (Schaye et al.

, 2005, MNRAS, 364, 1105 Springel V. p. 1
Where the paper says it

2020) is a suite of 42 cosmological zoom-in hydrodynamical simulations of Milky Way-mass haloes. These simulations are run using the Gadget-3 code (last described in Springel 2005) with the same hydrodynamics solver and galaxy formation (subgrid) physics as the EAGLE project (Schaye et al.

, Yoshida N., White S. D. M., 2001, New Astron., 6, 79 The pandas development team 2020, pandas-dev/pandas: Pandas, doi:10. p. 3
Where the paper says it

Later in Section 3. 1 we will explicitly test the model predictions against the ARTEMIS simulations, not used in the original development and calibration of the.

5281/zenodo.3509134 Walter F. p. 0
Where the paper says it

Persic et al. 1996; Walter et al.

, Brinks E., de Blok W. J. G., Bigiel F., Kennicutt Jr. R. C., Thornley M. D., Leroy A., 2008, AJ, 136, 2563 Watkins L. p. 0
Where the paper says it

1996; Walter et al. 2008; Hunter et al.

L., Evans N. W., An J. H., 2010, MNRAS, 406, 264 Watkins L. p. 0
Where the paper says it

Watkins et al. 2010; Li et al.

L., van der Marel R. P., Sohn S. T., Evans N. W., 2019, ApJ, 873, 118 Xue X. p. 0
Where the paper says it

1988; Cimatti et al. 2019; Marasco et al.

X., et al., 2008, ApJ, 684, 1143 Zavala J. p. 0
Where the paper says it

Even within this context, observations within the Milky Way (MW) offer unparalleled detail to map the DM distribution within our host halo (e. reviews by Bullock & Boylan-Kolchin 2017; Zavala & Frenk 2019).

, Frenk C. S.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI system can robustly infer global halo properties like virial mass and concentration by explicitly incorporating baryon-induced contraction using a model like Contracted NFW (CNFW). This prevents systematic biases observed when using standard models like NFW, gNFW, and Einasto, especially when observational constraints are limited to inner regions.

  2. Improved AI system can distinguish between the good and bad extrapolations of halo properties by recognizing that internal diagnostics like small residuals do not guarantee accurate extrapolation. It can be trained to recognize that wellbehaved fits — i.e., small residuals, smooth profiles, stable parameter posteriors — do not distinguish ‘good’ from ‘bad’ when extrapolating to larger radii.

  3. Improved AI system can prioritize observational data sources based on the physical region they probe. The system can learn that kinematic probes residing in the outer halo should be considered of higher fidelity when attempting a precision estimate of the Milky Way’s halo mass compared to inner-halo kinematics for constraining mass, and inner-halo kinematics are more informative for constraining concentration provided a contraction-aware model is used.

  4. Improved AI system can predict how different profile models will affect the concentration mass relation based on radial coverage. The system can learn that NFW fits in the MW-mass regime are unreliable for precision estimates of M200c and c, especially when extrapolating from parameters constrained using the inner, contracted galactic regions.

  5. Improved AI system can provide a quantitative pathway for re-interpreting rotation curve-based mass estimates. This involves recognizing that the biases we identify are not random; they follow systematic trends that depend on how contraction steepens the inner density profile and on how much of the halo is radially sampled.

Abstract

We investigate how reliably the global properties of Milky Way-mass dark matter haloes can be recovered from dynamical data from a limited range of radii, particularly 30 kpc where observations are most sensitive but baryonic processes modify the halo structure. Using the ARTEMIS (Assembly of high-ResoluTion Eagle simulations of MIlkyWay-type galaxieS) simulations, which produce varying degrees of baryon-induced contraction, we fit dark matter profiles over restricted radial ranges using commonly adopted parametric models. Assuming negligible observational uncertainties allows the systematic errors from these choices to be isolated. When fits are confined to inner radii, a Navarro-Frenk-White (NFW) profile underestimates the virial mass by a factor of about 2 on average (about 4 for some systems), and the concentration by a factor of about 2. Einasto and generalized-NFW models provide excellent local fits but retain similar global biases. In contrast, a contracted halo prescription yields stable extrapolations and recovers unbiased halo mass estimates over all radii. The inferred mass improves systematically with increasing radial coverage, and tracers beyond 50 kpc largely eliminate the mean bias for all models. The local dark matter density at the Solar radius is recovered to within 5% for all profiles other than NFW. These biases are sufficient to reconcile recent low Milky Way-mass estimates derived from inner rotation-curve analyses with the canonical about 10 12 M. We additionally find a halo-to-halo scatter of 0.1 dex (about 25%) persists even under idealized conditions, setting a likely lower limit for the precision of halo mass estimates.

Sources

Related papers