-
A data-driven approach for star formation parameterization using symbolic regression
Authors:
Diane M. Salim,
Matthew E. Orr,
Blakesley Burkhart,
Rachel S. Somerville,
Miles Cramner
Abstract:
Star formation (SF) in the interstellar medium (ISM) is fundamental to understanding galaxy evolution and planet formation. However, efforts to develop closed-form analytic expressions that link SF with key influencing physical variables, such as gas density and turbulence, remain challenging. In this work, we leverage recent advancements in machine learning (ML) and use symbolic regression (SR) t…
▽ More
Star formation (SF) in the interstellar medium (ISM) is fundamental to understanding galaxy evolution and planet formation. However, efforts to develop closed-form analytic expressions that link SF with key influencing physical variables, such as gas density and turbulence, remain challenging. In this work, we leverage recent advancements in machine learning (ML) and use symbolic regression (SR) techniques to produce the first data-driven, ML-discovered analytic expressions for SF using the publicly available FIRE-2 simulation suites. Employing a pipeline based on training the genetic algorithm of SR from an open software package called PySR, in tandem with a custom loss function and a model selection technique which compares candidate equations to analytic approaches to describing SF, we produce symbolic representations of a predictive model for the star formation rate surface density ($Σ_\mathrm{SFR}$) averaged over both 10 Myr and 100 Myr based on eight extracted variables from FIRE-2 galaxies. The resulting model that PySR finds best describes SF, on both averaging timescales, features equations that incorporates the surface density of gas, $Σ_\mathrm{gas}$, the velocity dispersion of gas $σ_{\mathrm{gas,~z}}$ and the surface density of stars $Σ_\mathrm{*}$. Furthermore, we find that the equations found for the longer SFR timescale all converge to a scaling-relation-like equation, all of which also closely capture the intrinsic physical scatter of the data within the Kennicutt-Schmidt (KS) plane. This observed convergence to physically interpretable scaling relations at longer SFR timescales demonstrates that our method successfully identifies robust physical relationships rather than fitting to stochastic fluctuations.
△ Less
Submitted 7 May, 2025;
originally announced May 2025.
-
Characterizing the Molecular Gas in Infrared Bright Galaxies with CARMA
Authors:
Katherine Alatalo,
Andreea O. Petric,
Lauranne Lanz,
Kate Rowlands,
Vivian U,
Kirsten L. Larson,
Lee Armus,
Loreto Barcos-Muñoz,
Aaron S. Evans,
Jin Koda,
Yuanze Luo,
Anne M. Medling,
Kristina E. Nyland,
Justin A. Otter,
Pallavi Patil,
Fernando Peñaloza,
Diane Salim,
David B. Sanders,
Elizaveta Sazonova,
Maya Skarbinski,
Yiqing Song,
Ezequiel Treister,
C. Meg Urry
Abstract:
We present the CO(1-0) maps of 28 infrared-bright galaxies from the Great Observatories All-Sky Luminous Infrared Galaxy Survey (GOALS) taken with the Combined Array for Research in Millimeter Astronomy (CARMA). We detect 100GHz continuum in 16 of 28 galaxies, which trace both active galactic nuclei (AGNs) and compact star-forming cores. The GOALS galaxies show a variety of molecular gas morpholog…
▽ More
We present the CO(1-0) maps of 28 infrared-bright galaxies from the Great Observatories All-Sky Luminous Infrared Galaxy Survey (GOALS) taken with the Combined Array for Research in Millimeter Astronomy (CARMA). We detect 100GHz continuum in 16 of 28 galaxies, which trace both active galactic nuclei (AGNs) and compact star-forming cores. The GOALS galaxies show a variety of molecular gas morphologies, though in the majority of cases, the average velocity fields show a gradient consistent with rotation. We fit the full continuum SEDs of each of the source using either MAGPHYS or SED3FIT (if there are signs of an AGN) to derive the total stellar mass, dust mass, and star formation rates of each object. We adopt a value determined from luminous and ultraluminous infrared galaxies (LIRGs and ULIRGs) of $α_{\rm CO}=1.5^{+1.3}_{-0.8}~M_\odot$ (K km s$^{-1}$ pc$^2)^{-1}$, which leads to more physical values for $f_{\rm mol}$ and the gas-to-dust ratio. Mergers tend to have the highest gas-to-dust ratios. We assume the cospatiality of the molecular gas and star formation, and plot the sample on the Schmidt-Kennicutt relation, we find that they preferentially lie above the line set by normal star-forming galaxies. This hyper-efficiency is likely due to the increased turbulence in these systems, which decreases the freefall time compared to star-forming galaxies, leading to "enhanced" star formation efficiency. Line wings are present in a non-negligible subsample (11/28) of the CARMA GOALS sources and are likely due to outflows driven by AGNs or star formation, gas inflows, or additional decoupled gas components.
△ Less
Submitted 13 September, 2024;
originally announced September 2024.
-
Extending a Physics-Informed Machine Learning Network for Superresolution Studies of Rayleigh-Bénard Convection
Authors:
Diane M. Salim,
Blakesley Burkhart,
David Sondak
Abstract:
Advancing our understanding of astrophysical turbulence is bottlenecked by the limited resolution of numerical simulations that may not fully sample scales in the inertial range. Machine learning (ML) techniques have demonstrated promise in up-scaling resolution in both image analysis and numerical simulations (i.e., superresolution). Here we employ and further develop a physics-constrained convol…
▽ More
Advancing our understanding of astrophysical turbulence is bottlenecked by the limited resolution of numerical simulations that may not fully sample scales in the inertial range. Machine learning (ML) techniques have demonstrated promise in up-scaling resolution in both image analysis and numerical simulations (i.e., superresolution). Here we employ and further develop a physics-constrained convolutional neural network (CNN) ML model called "MeshFreeFlowNet'' (MFFN) for superresolution studies of turbulent systems. The model is trained both on the simulation images as well as the evaluated PDEs, making it sensitive to the underlying physics of a particular fluid system. We develop a framework for 2D turbulent Rayleigh-Bénard convection (RBC) generated with the \textsc{Dedalus} code by modifying the MFFN architecture to include the full set of simulation PDEs and the boundary conditions. Our training set includes fully developed turbulence sampling Rayleigh numbers ($Ra$) of $Ra=10^6-10^{10}$. We evaluate the success of the learned simulations by comparing the power spectra of the direct \textsc{Dedalus} simulation to the predicted model output, and compare both ground truth and predicted power spectral inertial range scalings to theoretical predictions. We find that the updated network performs well at all $Ra$ studied here in recovering large-scale information, including the inertial range slopes. The superresolution prediction is overly dissipative at smaller scales than that of the inertial range in all cases, but the smaller-scales are better recovered in more turbulent, than laminar, regimes. This is likely because more turbulent systems have a rich variety of structures at many length scales compared to laminar flows.
△ Less
Submitted 31 January, 2024; v1 submitted 5 July, 2023;
originally announced July 2023.
-
Spinning Bar and a Star-formation Inefficient Repertoire: Turbulence in Hickson Compact Group NGC7674
Authors:
Diane M. Salim,
Katherine Alatalo,
Christoph Federrath,
Brent Groves,
Lisa J. Kewley
Abstract:
The physics regulating star formation (SF) in Hickson Compact Groups (HCG) has thus far been difficult to describe, due to their unique kinematic properties. In this study we expand upon previous works to devise a more physically meaningful SF relation able to better encompass the physics of these unique systems. We combine CO(1--0) data from the Combined Array from Research in Millimeter Astronom…
▽ More
The physics regulating star formation (SF) in Hickson Compact Groups (HCG) has thus far been difficult to describe, due to their unique kinematic properties. In this study we expand upon previous works to devise a more physically meaningful SF relation able to better encompass the physics of these unique systems. We combine CO(1--0) data from the Combined Array from Research in Millimeter Astronomy (CARMA) to trace the column density of molecular gas $Σ_\mathrm{gas}$ and deep H$α$ imaging taken on the Southern Astrophysical Research (SOAR) Telescope tracing $Σ_\mathrm{SFR}$ to investigate star formation efficiency across face-on HCG, NGC7674. We find a lack of universality in star formation, with two distinct sequences present in the $Σ_\mathrm{gas}-Σ_\mathrm{SFR}$ plane; one for inside and one for outside the nucleus. We devise a SF relation based on the multi-freefall nature of gas and the critical density, which itself is dependent on the virial parameter $α_{\mathrm{vir}}$, the ratio of turbulent to gravitational energy. We find that our modified SF relation fits the data and describes the physics of this system well with the introduction of a virial parameter of about 5--10 across the galaxy. This $α_{\mathrm{vir}}$ leads to an order-of-magnitude reduction in SFR compared to $α_{\mathrm{vir}}\approx 1$ systems.
△ Less
Submitted 23 March, 2020;
originally announced March 2020.
-
On the effective turbulence driving mode of molecular clouds formed in disc galaxies
Authors:
Keitaro Jin,
Diane M. Salim,
Christoph Federrath,
Elizabeth J. Tasker,
Asao Habe,
Jouni T. Kainulainen
Abstract:
We determine the physical properties and turbulence driving mode of molecular clouds formed in numerical simulations of a Milky Way-type disc galaxy with parsec-scale resolution. The clouds form through gravitational fragmentation of the gas, leading to average values for mass, radii and velocity dispersion in good agreement with observations of Milky Way clouds. The driving parameter (b) for the…
▽ More
We determine the physical properties and turbulence driving mode of molecular clouds formed in numerical simulations of a Milky Way-type disc galaxy with parsec-scale resolution. The clouds form through gravitational fragmentation of the gas, leading to average values for mass, radii and velocity dispersion in good agreement with observations of Milky Way clouds. The driving parameter (b) for the turbulence within each cloud is characterised by the ratio of the density contrast (sigma_rho) to the average Mach number (Mach) within the cloud, b = sigma_rho/Mach. As shown in previous works, b ~ 1/3 indicates solenoidal (divergence-free) driving and b ~ 1 indicates compressive (curl-free) driving. We find that the average b value of all the clouds formed in the simulations has a lower limit of b > 0.2. Importantly, we find that b has a broad distribution, covering values from purely solenoidal to purely compressive driving. Tracking the evolution of individual clouds reveals that the b value for each cloud does not vary significantly over their lifetime. Finally, we perform a resolution study with minimum cell sizes of 8, 4, 2 and 1 pc and find that the average b value increases with increasing resolution. Therefore, we conclude that our measured b values are strictly lower limits and that a resolution better than 1 pc is required for convergence. However, regardless of the resolution, we find that b varies by factors of a few in all cases, which means that the effective driving mode alters significantly from cloud to cloud.
△ Less
Submitted 28 March, 2017;
originally announced March 2017.
-
The SAMI Galaxy Survey: a new method to estimate molecular gas surface densities from star formation rates
Authors:
Christoph Federrath,
Diane M. Salim,
Anne M. Medling,
Rebecca L. Davies,
Tiantian Yuan,
Fuyan Bian,
Brent A. Groves,
I-Ting Ho,
Robert Sharp,
Lisa J. Kewley,
Sarah M. Sweet,
Samuel N. Richards,
Julia J. Bryant,
Sarah Brough,
Scott Croom,
Nicholas Scott,
Jon Lawrence,
Iraklis Konstantopoulos,
Michael Goodwin
Abstract:
Stars form in cold molecular clouds. However, molecular gas is difficult to observe because the most abundant molecule (H2) lacks a permanent dipole moment. Rotational transitions of CO are often used as a tracer of H2, but CO is much less abundant and the conversion from CO intensity to H2 mass is often highly uncertain. Here we present a new method for estimating the column density of cold molec…
▽ More
Stars form in cold molecular clouds. However, molecular gas is difficult to observe because the most abundant molecule (H2) lacks a permanent dipole moment. Rotational transitions of CO are often used as a tracer of H2, but CO is much less abundant and the conversion from CO intensity to H2 mass is often highly uncertain. Here we present a new method for estimating the column density of cold molecular gas (Sigma_gas) using optical spectroscopy. We utilise the spatially resolved H-alpha maps of flux and velocity dispersion from the Sydney-AAO Multi-object Integral-field spectrograph (SAMI) Galaxy Survey. We derive maps of Sigma_gas by inverting the multi-freefall star formation relation, which connects the star formation rate surface density (Sigma_SFR) with Sigma_gas and the turbulent Mach number (Mach). Based on the measured range of Sigma_SFR = 0.005-1.5 M_sol/yr/kpc^2 and Mach = 18-130, we predict Sigma_gas = 7-200 M_sol/pc^2 in the star-forming regions of our sample of 260 SAMI galaxies. These values are close to previously measured Sigma_gas obtained directly with unresolved CO observations of similar galaxies at low redshift. We classify each galaxy in our sample as 'Star-forming' (219) or 'Composite/AGN/Shock' (41), and find that in Composite/AGN/Shock galaxies the average Sigma_SFR, Mach, and Sigma_gas are enhanced by factors of 2.0, 1.6, and 1.3, respectively, compared to Star-forming galaxies. We compare our predictions of Sigma_gas with those obtained by inverting the Kennicutt-Schmidt relation and find that our new method is a factor of two more accurate in predicting Sigma_gas, with an average deviation of 32% from the actual Sigma_gas.
△ Less
Submitted 27 March, 2017;
originally announced March 2017.
-
A universal, turbulence-regulated star formation law: from Milky Way clouds to high-redshift disk and starburst galaxies
Authors:
Diane M. Salim,
Christoph Federrath,
Lisa J. Kewley
Abstract:
Whilst the star formation rate (SFR) of molecular clouds and galaxies is key in understanding galaxy evolution, the physical processes which determine the SFR remain unclear. This uncertainty about the underlying physics has resulted in various different star formation laws, all having substantial intrinsic scatter. Extending upon previous works that define the column density of star formation (Si…
▽ More
Whilst the star formation rate (SFR) of molecular clouds and galaxies is key in understanding galaxy evolution, the physical processes which determine the SFR remain unclear. This uncertainty about the underlying physics has resulted in various different star formation laws, all having substantial intrinsic scatter. Extending upon previous works that define the column density of star formation (Sigma_SFR) by the gas column density (Sigma_gas), we develop a new universal star formation (SF) law based on the multi-freefall prescription of gas. This new SF law relies predominantly on the probability density function (PDF) and on the sonic Mach number of the turbulence in the star-forming clouds. By doing so we derive a relation where the star formation rate (SFR) correlates with the molecular gas mass per multi-freefall time, whereas previous models had used the average, single-freefall time. We define a new quantity called maximum (multi-freefall) gas consumption rate (MGCR) and show that the actual SFR is only about 0.4% of this maximum possible SFR, confirming the observed low efficiency of star formation. We show that placing observations in this new framework (Sigma_SFR vs. MGCR) yields a significantly improved correlation with 3-4 times reduced scatter compared to previous SF laws and a goodness-of-fit parameter R^2=0.97. By inverting our new relationship, we provide sonic Mach number predictions for kpc-scale observations of Local Group galaxies as well as unresolved observations of local and high-redshift disk and starburst galaxies that do not have independent, reliable estimates for the turbulent cloud Mach number.
△ Less
Submitted 12 May, 2015;
originally announced May 2015.
