Measuring Scaling Relations: Fitting Technique Matters

Bryanne McDonough, Olivia Curtis, Tereasa G. Brainerd · 2025

Read Measuring Scaling Relations: Fitting Technique Matters by Bryanne McDonough, Olivia Curtis, Tereasa G. Brainerd (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-15702.txt.

Introduction

Bryanne McDonough, Olivia Curtis, and Tereasa Brainerd argue that how astronomers fit a line can change the galaxy story they tell. Scaling relations tie a galaxy's star mass to its star formation and gas, locally and globally. Studies measure them with different samples, resolutions, and fitting habits, which muddies every comparison. The team tests the standard least-squares family against a ridge line method on simulated galaxies from the TNG100 run.

For three resolved relations, the ridge line approach, fitting a bent two-part line, clearly beat least squares. Probing one relation further, they found its low-mass slope stayed put when the images were blurred or rebinned, a relief for comparing surveys. The authors call for the field to refit published relations consistently, and to adopt the ridge line method on physical grounds rather than by habit.

Catalog record

CREATORBryanne McDonough, Olivia Curtis, Tereasa G. Brainerd
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-15702

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Measuring Scaling Relations: Fitting Technique Matters Bryanne McDonough, Olivia Curtis, Tereasa G. Brainerd Abstract Scaling relationships, both integrated and spatially resolved, arise due to the physical processes that govern galaxy evolution and are frequently measured in both observed and simulated data. However, the accuracy and comparability of these measurements are hindered by various differences between studies such as spatial resolution, sample selection criteria, and fitting technique. Here, we compare variations of standard least squares techniques to the ridge line method for identifying spatially resolved scaling relations ( Σ ∗ − Σ SFR \Sigma_{*}-\Sigma_{\rm SFR} , Σ ∗ − Σ gas \Sigma_{*}-\Sigma_{\rm gas} , and Σ gas − Σ SFR \Sigma_{\rm gas}-\Sigma_{\rm SFR} ) for TNG100 galaxies. We find that using the ridge line technique to fit these scaling relations with a double linear function results in significantly better fits than fitting with ordinary least squares. We further illustrate the utility of the ridge line technique with an investigation into the dependence of rSFMS measurements on spatial resolution and smoothing scale. Specifically, we find that the slope of the rSFMS at low- Σ ∗ \Sigma_{*} is independent (within 2 ​ σ 2\sigma ) of spatial resolution and smoothing scale. Finally, we discuss the need for a consistent re-analysis of resolved scaling relations in the literature and physically motivate adoption of the ridge line technique over other fitting methods. SciPy [ 52 ] ; ScaleRPy [ 27 ] ; Plotly [ 19 ] I Introduction Scaling relationships are strong correlations between galaxy properties that arise due to the physical mechanisms that drive galaxy evolution. As ever higher spatial resolutions have become possible, many relations previously identified between integrated (i.e., global) galaxy properties have been found to have spatially resolved analogs, down to \sim 100 s of parsecs [ 36 , e.g.,] . In particular, at low redshifts there exist relationships between the surface densities of gas mass ( \Sigma_{\rm gas} ), stellar mass ( \Sigma_{*} ), and star formation rate ( \Sigma_{\rm SFR} ) on sub-kiloparsec scales [see 42 , for a review] . These relationships are inherently linked to the evolution of galaxies, as they describe the formation of stellar populations. The relationship between \Sigma_{*} and \Sigma_{\rm SFR} is often called the resolved star formation main sequence [ 7 , 25 , 1 , 23 , 51 , 10 , 30 , 22 , 4 , 36 , 3 , rSFMS; e.g.,] and is the main focus of this Letter. However, we also present the resolved gas main sequence (rGMS, i.e., \Sigma_{*}-\Sigma_{\rm gas} relation) and resolved Kennicutt-Schmidt law (rKS, i.e., \Sigma_{\rm gas}-\Sigma_{\rm SFR} relation) for simulated galaxies. Scaling relationships play an important role in cosmological hydrodynamic simulations of galaxy evolution, since unconstrained parameters can be calibrated to reproduce select relationships. For example, the IllustrisTNG simulations [ 32 , 37 , 49 , 31 , 26 , 33 ] were calibrated to reproduce the stellar-to-halo mass, stellar mass – stellar size, and black hole – galaxy mass relations. Relationships that are not used to calibrate a simulation can, instead, be used to validate it. For example, in [ 28 ] we showed that the 100- Mpc box of the IllustrisTNG suite (TNG100) broadly reproduced the rSFMS with a slope in agreement with that found by [ 4 ] for MaNGA [ 6 ] galaxies. In general, scaling relations have broad utility in extragalactic astrophysics. For observational measurements that are challenging to make, scaling relations can be used to identify appropriate proxies, e.g., inferring black hole mass from central velocity dispersion [ 11 , e.g.,] . Semi-analytic models often incorporate scaling relations, for example: populating simulated dark matter halos with galaxies given the halo-to-stellar mass relation [see 8 , for a review] . Further, scaling relations arise due to physical processes, and therefore their shapes encode physically significant information. For example, [ 43 ] demonstrated a power-law relationship between SFR and the surface density of gas. [ 15 ] and [ 16 ] expanded on this work, finding that the SFR was also regulated by the total local density which sets the conditions for gravitational collapse. Given that the timescale for gravitational instabilities to grow in a parcel of gas will scale with local density, a star formation law should follow the form: \rho_{\rm SFR}\propto\rho_{\rm gas}^{1.5} above some threshold density [ 16 ] . More recently, [ 46 ] proposed an extension of the rKS law with \Sigma_{\rm SFR} being dependent on both \Sigma_{\rm gas} and \Sigma_{*} . Many works have found that the rKS holds only for molecular gas [ 17 , e.g.,] . A full review of theoretical predictions for the rKS and similar relations is beyond the scope of this Letter, but in general the slopes of these power-law relations are predicted by theory and can be constrained with observational measurements. As both simulations and observations improve in resolution and complexity, a need for accurate and unbiased quantitative comparisons of scaling relations arises if we are to constrain theory further. Of particular interest is whether the physical properties that drive spatially resolved scaling relations are redshift-dependent, as has been found for some global relations, e.g., the global SFMS [ 47 ] . Unfortunately, differences in galaxy sample selection, spatial resolution, and techniques for measuring scaling relationships have made such comparisons challenging. Measurements of the rSFMS and rKS can be further affected by the criteria for pre-selection of star-forming regions and the indicator used to measure SFR. The slopes of the rSFMS presented in the literature range from \sim 0.5 to \sim 1 , with reported confidence intervals ranging from \sim 0.0002-0.2 [ 7 , 25 , 1 , 23 , 51 , 10 , 4 , 13 , 36 , 28 , e.g.,] . Using spatially resolved maps of simulated galaxies from the FIRE project [ 53 ] , [ 13 ] found that the rSFMS slope depends on the fitting method, star formation timescales, and spatial resolution. The rSFMS may be a byproduct of two more fundamental relations [ 2 , e.g.,] : the molecular gas main sequence (i.e., \Sigma_{H_{2}}-\Sigma_{*} ) and the molecular rKS law (i.e., \Sigma_{H_{2}}-\Sigma_{\rm SFR} ). However, the rSFMS remains useful due to the challenges associated with obtaining molecular gas measurements for large observational samples (i.e., the need for millimeter wavelengths). Even in simulation space, the rSFMS requires only the analysis of stellar particles, while investigations of the other relations requires analysis of both stellar and gas particles. Thus, a consistent measurement of the rSFMS slope and other scaling relations is desirable for validation of simulation results and for probing the physics that drive star formation. Many studies in the literature use some variation of least squares to fit a power law to their samples of spectral pixels (i.e., spaxels) or simulated analogs, usually done in logarithmic space to simplify the problem. Typically, some criteria are used to pre-select star-forming spaxels and a line is then fit to either the full sample or the median of the sample in bins of stellar mass. Mathematically, a least squares fit to a full sample returns a line that traces the conditional mean (e.g., E [\Sigma_{\rm SFR}|\Sigma_{*}] ). This method is not ideal because spaxels are not normally

Read the complete text