---
title: 'Magneticum Pathfinder: Cosmological Simulations'
url: https://www.emergentmind.com/topics/magneticum-pathfinder
type: topic
---

# Magneticum Pathfinder: Cosmological Simulations

Magneticum Pathfinder is a suite of hydrodynamical cosmological simulations designed to follow the joint evolution of dark matter, gas, stars, and supermassive black holes from early times to the present day. In the 2025 synthesis of the project, it is described as the suite that “self-consistently covers the largest range in box volumes and resolutions” and as the only cosmological simulation suite “tuned on the hot gas content of galaxy clusters instead of the stellar mass function,” with galaxy properties and scaling relations then emerging within the same framework [2504.01061].

## 1. Programme architecture and numerical scope

The main Magneticum runs adopt a WMAP-7 flat $\Lambda$CDM cosmology with
$$
\Omega_{\mathrm{m}} = 0.272,\quad
\Omega_{\mathrm{b}} = 0.0456,\quad
\Omega_{\Lambda} = 0.728,\quad
h = 0.704,\quad
\sigma_8 = 0.809,\quad
n_s = 0.963.
$$
This common cosmological setup underlies applications ranging from satellite systems and galaxy stellar halos to groups, clusters, and thermal SZ light-cones [2504.01061].

A distinctive feature of the programme is its use of multiple periodic volumes at fixed resolution tiers, conventionally denoted mr, hr, uhr, and xhr. The 2025 overview gives the following box structure.

| Box | Side length | Resolution level |
|---|---:|---|
| Box0 | 3820 Mpc | mr |
| Box1a | 1300 Mpc | mr |
| Box2b | 910 Mpc | hr |
| Box2 | 500 Mpc | hr |
| Box3 | 180 Mpc | hr / uhr |
| Box4 | 68 Mpc | uhr |
| Box5 | 26 Mpc | xhr |

For these tiers, representative particle masses and softenings are explicitly tabulated. In mr,
$$
m_{\rm DM} = 1.3\times 10^{10}\,h^{-1}M_\odot,\quad
m_{\rm gas}^{\rm init} = 2.6\times 10^{9}\,h^{-1}M_\odot,
$$
with $\epsilon_{\rm DM/gas} = 10~h^{-1}{\rm kpc}$ and $\epsilon_* = 5~h^{-1}{\rm kpc}$. In hr,
$$
m_{\rm DM} = 6.9\times 10^{8}\,h^{-1}M_\odot,\quad
m_{\rm gas}^{\rm init} = 1.4\times 10^{8}\,h^{-1}M_\odot,
$$
with $\epsilon_{\rm DM/gas} = 3.75~h^{-1}{\rm kpc}$ and $\epsilon_* = 2~h^{-1}{\rm kpc}$. In uhr,
$$
m_{\rm DM} = 3.6\times 10^{7}\,h^{-1}M_\odot,\quad
m_{\rm gas}^{\rm init} = 7.3\times 10^{6}\,h^{-1}M_\odot,
$$
with $\epsilon_{\rm DM/gas} = 1.4~h^{-1}{\rm kpc}$ and $\epsilon_* = 0.7~h^{-1}{\rm kpc}$. In xhr,
$$
m_{\rm DM} = 1.9\times 10^{6}\,h^{-1}M_\odot,\quad
m_{\rm gas}^{\rm init} = 3.9\times 10^{5}\,h^{-1}M_\odot,
$$
with $\epsilon_{\rm DM/gas} = 0.45~h^{-1}{\rm kpc}$ and $\epsilon_* = 0.25~h^{-1}{\rm kpc}$ [2504.01061].

Published applications illustrate how these boxes are used. Box2b/hr is employed for environmental statistics of AGN and star formation in cluster populations over a $(640\,\mathrm{Mpc}/h)^3$ comoving volume, while Box4(uhr) is used for high-resolution studies of planes of satellite galaxies and of tidal shells and streams [2307.06374] [2208.05496] [2509.25307]. This combination of very large cosmological volumes and smaller ultra-high-resolution runs is central to the project’s division of labor across scales.

## 2. Numerical framework and baryonic physics

Magneticum is built on GADGET-3 or closely related P-GADGET3/P-Gadget3-XXL branches. The hydrodynamics is SPH-based, with later programme summaries emphasizing an entropy-conserving formulation, a Wendland $C^6$ kernel with 295 neighbours, time-dependent viscosity, time-dependent artificial conductivity, a wake-up scheme for time stepping, and isotropic thermal conduction at $\kappa = 1/20$ of Spitzer [2504.01061]. Other papers in the suite describe the implementation as an improved SPH formulation with high-order kernels, improved artificial viscosity, and artificial conduction, specifically for realistic ICM and galaxy-gas dynamics [2106.14542].

The baryonic model includes radiative cooling, UV/X-ray background heating, star formation in a multiphase ISM, stellar feedback, chemical enrichment, black-hole growth, and AGN feedback. Across the published descriptions, the recurring ingredients are radiative cooling following Wiersma et al. (2009), the Springel & Hernquist (2003) multiphase ISM star-formation model, stellar evolution with a Chabrier IMF, enrichment from SNII, SNIa, and AGB stars, and quasar/radio-mode AGN feedback [2307.06374] [1708.00027] [1509.05134].

The star-formation law is implemented volumetrically, but the suite summary states that for resolved disks it reproduces a Kennicutt–Schmidt-type surface-density scaling,
$$
\Sigma_{\rm SFR} \propto \Sigma_{\rm gas}^N,\quad N\sim 1.4.
$$
The same overview notes that each gas particle can spawn up to four star particles, so the effective stellar mass resolution exceeds the gas-particle resolution [2504.01061].

The black-hole model is a major organizing feature. In the fiducial implementation, seed black holes of
$$
M_{\rm seed} = 2\times 10^5\,h^{-1}M_\odot
$$
are inserted in halos whose stellar mass exceeds $\sim 10^{10}\,h^{-1}M_\odot$, and their gas accretion follows a Bondi–Hoyle–Lyttleton form,
$$
\dot{M}_\bullet = \frac{4\pi G^2 M_\bullet^2 \alpha \rho}{(c_s^2 + v^2)^{3/2}},
$$
capped at the Eddington rate [2504.01061]. In the SZ analysis the same family of prescriptions is written as
$$
\dot{M}_\bullet = \frac{4\pi G^2 M_\bullet^2 f_\mathrm{boost} \rho}{(c_s^2 + v^2)^{3/2}},
$$
with $f_\mathrm{boost}=100$, radiative efficiency $\epsilon_r=0.1$, and a two-mode AGN feedback model in which the radio mode is activated below $f_{\rm edd}<10^{-2}$ [1509.05134]. An advanced black-hole model, used in some Box3/uhr runs, separates hot and cold gas accretion and yields faster BH growth and earlier quenching at high redshift [2504.01061].

Haloes and galaxies are identified with FoF and SUBFIND, and merger trees are constructed with L-BaseTree in studies that require explicit progenitor tracking [2509.25307]. This common infrastructure makes it possible to move between halo-scale thermodynamics, galaxy-scale kinematics, and stellar-halo archaeology within a single simulation family.

## 3. Calibration strategy, scaling relations, and validation

The most explicit statement of Magneticum’s calibration philosophy appears in the 2025 overview: the suite is tuned to the hot gas content of galaxy clusters rather than to the galaxy stellar mass function [2504.01061]. The same paper then evaluates 28 scaling relations from $z=4$ to $z=0$, including the halo-to-stellar-mass relation, the Kennicutt–Schmidt relation, the cosmic star formation rate density, the Fundamental Plane, the mass–size relation, the mass–metallicity relation, the Magorrian relation, and the temperature–mass relation [2504.01061].

Several of the reported successes are unusually broad in scope. The suite is stated to match “a remarkable number of the observed scaling relations from $z=4$ to $z=0$,” including “the number density of quiescent galaxies at cosmic dawn, the mass--size evolution, the mass--metallicity relation, the Magorrian relation, and the temperature--mass relation” [2504.01061]. In the same synthesis, the cosmic star formation rate density peaks at $z\sim 2.5$–3 in uhr boxes and at $z\sim 2$ in hr boxes, while the advanced BH model in Box3/uhr produces massive quenched galaxies at $z\gtrsim 3$–5 with number densities consistent with the cited JWST-era measurements [2504.01061].

Cluster thermodynamics and SZ observables are among the best quantified parts of the programme. Using $\sim 50{,}000$ clusters with $M_{\rm 500c}>1.4\times10^{14}\,M_\odot$ out to $z=2$, the pressure-profile study finds significant mass and redshift dependence in profile shape, and fits a GNFW model with
$$
P_0 = 0.1701,\quad c_{500} = 1.21,\quad \gamma = 0.37,\quad \alpha = 1.23,\quad \beta = 5.06
$$
for the global sample [1612.05266]. The same analysis finds that thermal pressure at $R_{\rm 500c}$ accounts for only 80 percent of the pressure required to maintain hydrostatic equilibrium, implying that even idealized hydrostatic mass estimates are biased at the 20 percent level [1612.05266]. For spherical $Y_{500\mathrm{c}}$ the quoted intrinsic scatter is
$$
\sigma_{\ln Y}\simeq 0.0875,
$$
whereas full light-cone measurements yield
$$
\sigma_{\ln Y}\simeq 0.159
$$
because of line-of-sight structure [1612.05266].

The large-volume SZ programme extends these cluster results to sky statistics. From a deep $8.8^\circ\times 8.8^\circ$ light-cone out to $z=5.2$, the thermal SZ one-point PDF has a high-$y$ tail following a power law with index $-3.2$, and the mean fluctuating Compton-$Y$ is
$$
\bar Y = 1.18\times 10^{-6}
$$
for the native cosmology, or
$$
\bar Y = 1.57\times 10^{-6}
$$
after rescaling to the Planck 2015 cosmology [1509.05134]. The same study states that nearly half of this mean signal, about $5\times10^{-7}$, comes from halos below a virial mass of $10^{13}M_\odot/h$ [1509.05134]. This SZ work is methodologically consistent with the hot-gas tuning strategy and is one reason the suite occupies a distinctive position relative to stellar-mass-function-tuned programmes.

## 4. Galaxy clusters, the cosmic web, and dynamical state

Magneticum has been used not only for statistical scaling relations but also for detailed analog studies of complex cluster systems. In the A3391/95 case study, an analogue is extracted from the $(352\,h^{-1}\,\mathrm{Mpc})^3$ Box2/hr volume at $z=0.07$. The selected pair has
$$
M_{500}^1 = 1.96\times 10^{14}\,M_\odot,\qquad
M_{500}^2 = 1.67\times 10^{14}\,M_\odot,
$$
with a 3D separation of $4.54\,\mathrm{Mpc}$ and projected separations of $2.6$, $4.5$, and $3.8\,\mathrm{Mpc}$ in the three principal projections [2106.14542]. The two clusters have overlapping $3R_{200}$ spheres of influence but disjoint virial radii, and the interconnecting bridge is found to be diffuse WHIM-like gas with typical temperature $\sim 1$ keV, overdensity $\sim 100$, and iron abundance $\sim 0.1\,Z_{\rm Fe,\odot}$, while 90–95 percent of the present-day bridge gas has never been inside $R_{200}$ of any progenitor of the two main clusters [2106.14542]. The published interpretation is therefore explicitly pre-merger and filamentary rather than one of tidally stripped cluster atmospheres.

A different cluster-scale application concerns the relation between projected substructure and recent assembly. In the Abell 2744 study, the authors use Box2b/hr and a projection-based mass measurement to mimic lensing-derived substructure masses. They identify a cluster projection with eight substructures whose mass fractions,
$$
3.6,\ 3.5,\ 2.6,\ 2.6,\ 2.6,\ 2.3,\ 2.2,\ 2.2\%,
$$
are strikingly similar to the Abell 2744 benchmark, and they show that projected substructure masses are generally 2–3 times larger than the corresponding bound SUBFIND masses [2209.09916]. The same work argues that both the total projected substructure mass fraction and the mass fraction of the eighth substructure are excellent tracers of recent dynamical assembly, with high fractions indicating merger activity within the last 2 Gyr [2209.09916].

These two applications illustrate a broader pattern in Magneticum usage. Cluster pairs, bridges, infalling groups, strong shock morphologies, projected substructure overdensities, and large-scale gas filaments are all treated as outcomes of the same cosmological hydrodynamical framework, rather than as isolated zoomed phenomena. The combination of large volumes and explicit baryonic physics is therefore not incidental; it is the condition under which these systems can be interpreted in terms of assembly history rather than only morphology.

## 5. Galaxy evolution, stellar populations, and low-surface-brightness structure

A substantial fraction of Magneticum’s scientific output concerns galaxies in dense environments and the coupling between star formation, AGN activity, and structure. In box2b/hr, the cluster-environment study selects galaxies with $\log M_\ast(M_\odot)>10.15$ in halos with $M_{500}>10^{13}M_\odot$ at $z=0.25$ and $z=0.90$. It finds that low-mass galaxies show monotonic declines of both star-forming ratio and AGN fraction toward the cluster center, whereas massive galaxies show a non-monotonic response: the star-forming ratio peaks between $0.5$ and $1\,R_{500}$, while the AGN fraction reaches a minimum between $1$ and $2\,R_{500}$ and then rises rapidly for $r\lesssim R_{500}$, with an additional AGN excess around $\sim 3R_{500}$ that persists even in isolated regions [2307.06374]. This result is used there to argue that AGN fueling and star formation are not synchronized in massive cluster galaxies.

The post-starburst analysis extends this environmental program to quenching channels. In Box2, 647 PSBs are identified at $z\sim 0$ with $M_\ast \geq 5\times10^{10}\,M_\odot$ and tracked over 3.6 Gyr. Of the galaxies with merger trees extending back 2.5 Gyr, 88.9 percent experienced at least one merger and 64.7 percent at least one major merger; 23 percent of the $z=0$ PSBs were rejuvenated during their starburst [2011.06602]. The paper further states that field PSBs are generally shut down via a strong increase in AGN feedback, with
$$
P_{\rm AGN,PSB} \geq 10^{56}\ {\rm erg/Myr},
$$
whereas cluster PSBs at $z\lesssim 0.5$ are predominantly infalling and show no signs of enhanced AGN activity, supporting ram-pressure stripping as the dominant environmental quenching channel [2011.06602].

On smaller halo scales, the suite has been used to study anisotropic satellite systems. In Box4(uhr), the “Momentum in Thinnest Plane” method finds that thin planes of satellites with thickness $0.2\,R_{\rm halo}$ containing at least 50 percent of the satellites can be found in almost all systems, and that in Milky Way-mass-like hosts around 86 percent of such planes have at least 90 percent of the total satellite momentum aligned within the plane [2208.05496]. The same study reports no correlation between the existence of a plane and central-galaxy morphology, but a clear preference for the minor axes of the satellite plane and the host galaxy to be aligned [2208.05496].

The stellar-halo programme uses the same Box4(uhr) for shells and streams. In LSST-like mock $r$-band images with limiting surface brightness
$$
\mu_\mathrm{lim} = 30.3\,\mathrm{mag\ arcsec^{-2}},
$$
shells have a mean radius peaking at $\sim 20\,\mathrm{kpc}$ and a median width $\approx 6.55\,\mathrm{kpc}$, while streams have median radius $\sim 27\,\mathrm{kpc}$ and median width $\approx 5.55\,\mathrm{kpc}$ [2509.25307]. The paper finds that both shells and streams are associated with localized depressions in stellar velocity dispersion, that shells are commonly produced by radial major mergers but also by minor mergers, that streams usually form from minor mergers on circular orbits, and that the visible shell or stream mass corresponds on average to approximately 20 percent of the progenitor mass [2509.25307]. It also introduces a class of star-forming streams formed in situ from gas structures rather than from disrupted satellites [2509.25307].

Chemical enrichment studies provide a complementary perspective on the baryon cycle. The metals paper states that Magneticum reproduces the normalization and slope of the gas-phase mass–metallicity relation at both $z=0$ and $z=2.3$, reproduces stellar Fe gradients, and broadly matches ICM Fe abundances and multi-element abundance patterns, while also identifying too-steep gas-phase metallicity gradients and nearly flat stellar $\alpha$/Fe radial profiles as signatures of insufficient mixing or absent explicit metal diffusion [1708.00027]. This chemical programme is closely connected to the environmental, post-starburst, and stellar-halo applications, because all of them rely on the same integrated treatment of star formation, feedback, outflows, and stripping.

## 6. Methodological debates, limitations, data access, and distinct usage of the name

Magneticum’s published results are frequently framed against apparent tensions between observations and $\Lambda$CDM. Two of the clearest cases are substructure masses in Abell 2744 and planes of satellites. In the Abell 2744 study, the claimed tension is explicitly traced to a methodological mismatch between observationally inferred projected masses and 3D bound subhalo masses from SUBFIND, with the projection-based estimator yielding masses typically 2–3 times larger [2209.09916]. In the satellite-plane study, the authors argue that thin, kinematically coherent planes are common in a full hydrodynamical $\Lambda$CDM simulation, and that the Local Group “planes of satellites” problem is therefore strongly sensitive to plane-finding method, baryonic physics, and selection effects [2208.05496].

The suite’s own synthesis also emphasizes unresolved issues. The 2025 overview notes over-prediction of the stellar mass function at $M_\ast\lesssim 10^{10}M_\odot$, somewhat high BCG+ICL stellar masses when very large apertures are used, and central stellar velocity dispersions that are generally lower than in IFU surveys [2504.01061]. The metals study separately identifies too-steep gas metallicity gradients, a shallower stellar mass–metallicity relation than observed, and flat stellar $\alpha$/Fe profiles as limitations attributable to insufficient mixing, lack of explicit metal diffusion, resolution, and uncertainties in stellar yields and SNIa delay-time modeling [1708.00027]. The shells-and-streams analysis adds that sub-kpc features are beyond the nominal reach of the $0.7\,\mathrm{kpc}/h$ stellar softening in Box4(uhr), and that low-surface-brightness feature selection remains partly visual [2509.25307].

These limitations coexist with unusually rich public-facing infrastructure. The 2025 overview states that a subset of Magneticum data products is accessible through the Cosmosim web portal at `https://c2papcosmosim.uc.lrz.de/`, including interactive visualization, on-demand X-ray mocks via PHOX, map-making via SMAC, and SimCut extraction tools; additional material is available via `www.magneticum.org/Data` [2504.01061]. Earlier project papers also refer to a public web portal for cluster catalogues and derived products [2209.09916] [2011.06602].

A final terminological point is necessary because the name has acquired a second, unrelated usage in the LISA Pathfinder literature. Two 2024 papers use “Magneticum Pathfinder” informally to describe how LISA Pathfinder functioned as a magnetic pathfinder for future space-based gravitational-wave missions: one provides the first complete estimate of magnetic-induced acceleration noise in the February 2017 run, finding
$$
0.25_{-0.08}^{+0.15}\,\mathrm{fm\,s^{-2}/\sqrt{Hz}}
$$
at 1 mHz and
$$
1.01_{-0.24}^{+0.73}\,\mathrm{fm\,s^{-2}/\sqrt{Hz}}
$$
at 0.1 mHz, while the companion paper reports precision measurements of the test-mass magnetic parameters, including
$$
|M_r| = (0.245\pm0.081)\,\mathrm{nA\,m^2},\quad
B_{\text{back},x} = (414 \pm 74)\,\mathrm{nT},\quad
\partial_x B_{\text{back},x} = (-7.4\pm 2.1)\,\mu\mathrm{T/m},
$$
and
$$
\chi = (-3.35\pm0.15)\times 10^{-5}
$$
in the low-frequency regime [2407.04427] [2407.04431]. In astrophysics, however, “Magneticum Pathfinder” overwhelmingly denotes the cosmological hydrodynamical simulation suite described above.

Source: https://www.emergentmind.com/topics/magneticum-pathfinder