Klaus Dolag’s research while affiliated with Ludwig-Maximilians-Universität in Munich and other places

Cosmological simulations are powerful tools in the context of structure formation. They allow us to explore the assembly and clustering of dark matter halos, to validate or reject possible scenarios of structure formation, and to investigate the physical properties of evolving galaxies across time. Cosmological hydrodynamical simulations are especially key to study how the complex interstellar medium of forming galaxies responds to the most energetic processes during galaxy evolution, such as stellar feedback ensuing supernova explosions and feedback from AGN. Given the huge dynamical range of physical scales spanned by the astrophysical processes involved in cosmic structure formation and evolution, cosmological simulations resort to sub-resolution models to capture processes occurring below their resolution limit. The impact of different sub-grid prescriptions accounting for the same process is striking, though often overlooked. Some among the main aforementioned processes include: hot gas cooling, star formation and stellar feedback, stellar evolution and chemical enrichment, black hole growth and feedback. Producing simulations of cosmic structure formation and galaxy evolution in large computational volumes is key to shed light on what drives the formation of the first structures in the Universe, and their subsequent evolution. Not only are predictions from simulations crucial to compare with data from ongoing observational instruments, but they can also guide future observational campaigns. Besides, since we have entered the era of high-performance computing, it is fundamental to have numerical codes which are very efficient from the computational point of view. In this chapter, we review the main hydrodynamic methods used in cosmological simulations and the most common techniques adopted to include the astrophysical processes which drive galaxy formation and evolution (abridged).

Cosmic Voids are a promising probe of cosmology for spectroscopic galaxy surveys due to their unique response to cosmological parameters. Their combination with other probes promises to break parameter degeneracies. Due to simplifying assumptions, analytical models for void statistics are only representative of a subset of the full void population. We present a set of neural-based emulators for void summary statistics of watershed voids, which retain more information about the full void population than simplified analytical models. We build emulators for the void size function and void density profiles traced by the halo number density using the Quijote suite of simulations for a broad range of the $\Lambda\mathrm{CDM}$ parameter space. The emulators replace the computation of these statistics from computationally expensive cosmological simulations. We demonstrate the cosmological constraining power of voids using our emulators, which offer orders-of-magnitude acceleration in parameter estimation, capture more cosmological information compared to analytic models, and produce more realistic posteriors compared to Fisher forecasts. We find that the parameters $\Omega_m$ and $\sigma_8$ in this Quijote setup can be recovered to $14.4\%$ and $8.4\%$ accuracy respectively using void density profiles; including the additional information in the void size function improves the accuracy on $\sigma_8$ to $6.8\%$. We demonstrate the robustness of our approach to two important variables in the underlying simulations, the resolution, and the inclusion of baryons. We find that our pipeline is robust to variations in resolution, and we show that the posteriors derived from the emulated void statistics are unaffected by the inclusion of baryons with the Magneticum hydrodynamic simulations. This opens up the possibility of a baryon-independent probe of the large-scale structure.

It has been recently shown that a powerful way to constrain cosmological parameters from galaxy redshift surveys is to train graph neural networks to perform field-level likelihood-free inference without imposing cuts on scale. In particular, de Santi et al. [58] developed models that could accurately infer the value of Ωm from catalogs that only contain the positions and radial velocities of galaxies that are robust to different astrophysics and subgrid models. However, observations are affected by many effects, including (1) masking, (2) uncertainties in peculiar velocities and radial distances, and (3) different galaxy population selections. Moreover, observations only allow us to measure redshift, which entangles the galaxy radial positions and velocities. In this paper we train and test our models on galaxy catalogs, created from thousands of state-of-the-art hydrodynamic simulations run with different codes from the CAMELS project, that incorporate these observational effects. We find that while such effects degrade the precision and accuracy of the models, the fraction of galaxy catalogs for which the models retain high performance and robustness is over 90%, demonstrating the potential for applying them to real data.

Aims. Our aim is to understand how the interplay between black hole (BH) feedback and merge processes can effectively turn cool-core galaxy clusters into hot-core clusters in the modern universe. Additionally, we also aim to clarify which parameters of the BH feedback model used in simulations can cause an excess of feedback at the scale of galaxy groups while not efficiently suppressing star formation at the scale of galaxy clusters. Methods. To obtain robust statistics of the cool-core population, we compare the modern Universe snapshot ( z = 0.25) of the largest Magneticum simulation ( Box2b/hr ) with the eROSITA eFEDS survey and Planck SZ-selected clusters observed with XMM-Newton. Additionally, we compare the BH feedback injected by the simulation in radio mode with Chandra observations of X-ray cavities, and LOFAR observations of radio emission. Results. We confirm a decreasing trend in cool-core fractions towards the most massive galaxy clusters, which is well reproduced by the Magneticum simulations. This evolution is connected with an increased merge activity that injects high-energy particles into the core region, but it also requires thermalisation and conductivity to enhance mixing through the intra-cluster medium core, where both factors are increasingly efficient towards the high mass end. On the other hand, BH feedback remains as the dominant factor at the scale of galaxy groups, while its relative impact decreases towards the most massive clusters. Conclusions. The problems suppressing star formation in simulations are not caused by low BH feedback efficiencies. They root in the definition of the black hole sphere of influence used to distribute the feedback, which decreases as density and accretion rate increase. Actually, a decreasing BH feedback efficiency towards low-mass galaxy groups is required to prevent overheating. These problems can be addressed in simulations by using relations between accretion rate, cavity power, and cavity reach derived from X-ray observations.

Recent observations with JWST and the Atacama Large Millimeter/submillimeter Array have revealed extremely massive quiescent galaxies at redshifts of z = 3 and higher, indicating both rapid onset and quenching of star formation. Using the cosmological simulation suite Magneticum Pathfinder, we reproduce the observed number densities and stellar masses, with 36 quenched galaxies of stellar mass larger than 3 × 10 ¹⁰ M ⊙ at z = 3.42. We find that these galaxies are quenched through a rapid burst of star formation and subsequent active galactic nucleus (AGN) feedback caused by a particularly isotropic collapse of surrounding gas, occurring on timescales of around 200 Myr or shorter. The resulting quenched galaxies host stellar components that are kinematically fast rotating and alpha-enhanced, while exhibiting a steeper metallicity and flatter age gradient compared to galaxies of similar stellar mass. The gas of the galaxies has been metal enriched and ejected. We find that quenched galaxies do not inhabit the densest nodes, but rather sit in local underdensities. We analyze observable metrics to predict future quenching at high redshifts, finding that on shorter timescales <500 Myr, the ratio M bh / M * is the best predictor, followed by the burstiness of the preceding star formation, t 50 – t 90 (time to go from 50% to 90% stellar mass). On longer timescales, >1 Gyr, the environment becomes the strongest predictor, followed by t 50 – t 90 , indicating that at high redshifts the consumption of old gas and lack of new gas are more relevant for long-term prevention of star formation than the presence of a massive AGN. We predict that relics of such high- z quenched galaxies should best be characterized by a strong alpha enhancement.

Context. The amount of turbulent pressure in galaxy clusters is still debated, especially in relation to the impact of the dynamical state and the hydro-method used for simulations. Aims. We study the turbulent pressure fraction in the intracluster medium of massive galaxy clusters. We aim to understand the impact of the hydrodynamical scheme, analysis method, and dynamical state on the final properties of galaxy clusters from cosmological simulations. Methods. We performed non-radiative simulations of a set of zoom-in regions of seven galaxy clusters with meshless finite mass (MFM) and smoothed particle hydrodynamics (SPH). We used three different analysis methods based on: (i) the deviation from hydrostatic equilibrium, (ii) the solenoidal velocity component obtained by a Helmholtz-Hodge decomposition, and (iii) the small-scale velocity obtained through a multi-scale filtering approach. We split the sample of simulated clusters into active and relaxed clusters. Results. Our simulations predict an increased turbulent pressure fraction for active clusters compared to relaxed ones. This is especially visible for the velocity-based methods. For these, we also find increased turbulence for the MFM simulations compared to SPH, consistent with findings from more idealized simulations. The predicted nonthermal pressure fraction varies between a few percent for relaxed clusters and ≈13% for active ones within the cluster center and increases toward the outskirts. No clear trend with redshift is visible. Conclusions. Our analysis quantitatively assesses the importance played by the hydrodynamical scheme and the analysis method to determine the nonthermal or turbulent pressure fraction. While our setup is relatively simple (non-radiative runs), our simulations show agreement with previous, more idealized simulations, and represent a step closer to an understanding of turbulence.

Our aim is to understand how the interplay between AGN feedback and merge processes can effectively turn cool-core galaxy clusters into hot-core clusters in the modern universe. Additionally, we also aim to clarify which parameters of the AGN feedback model used in simulations can cause an excess of feedback at the scale of galaxy groups while not efficiently suppressing star formation at the scale of galaxy clusters. To obtain robust statistics of the cool-core population, we compare the modern Universe snapshot (z=0.25) of the largest Magneticum simulation (Box2b/hr) with the eROSITA eFEDS survey and Planck SZ-selected clusters observed with XMM-Newton. Additionally, we compare the AGN feedback injected by the simulation in radio mode with Chandra observations of X-ray cavities, and LOFAR observations of radio emission. We confirm a decreasing trend in cool-core fractions towards the most massive galaxy clusters, which is well reproduced by the Magneticum simulations. This evolution is connected with an increased merge activity that injects high-energy particles into the core region, but it also requires thermalization and conductivity to enhance mixing through the ICM core, where both factors are increasingly efficient towards the high mass end. On the other hand, AGN feedback remains as the dominant factor at the scale of galaxy groups, while its relative impact decreases towards the most massive clusters. The problems suppressing star formation in simulations are not caused by low AGN feedback efficiencies. They root in the definition of the black hole sphere of influence used to distribute the feedback, which decreases as density and accretion rate increase. Actually, a decreasing AGN feedback efficiency towards low-mass galaxy groups is required to prevent overheating.

The full-sky measurements of the cosmic microwave background (CMB) temperature anisotropies by WMAP and Planck have highlighted several unexpected isotropy-breaking features on the largest angular scales. We investigate the impact of the local large-scale structure on these anomalies through the thermal and kinetic Sunyaev-Zeldovich effects. We used a constrained hydrodynamical simulation that reproduced the local Universe in a box of 500 h ⁻¹ Mpc to construct full-sky maps of the temperature anisotropies produced by these two secondary effects of the CMB, and we discuss their statistical properties on large angular scales. We show the significant role played by the Virgo cluster on these scales, and we compare it to theoretical predictions and random patches of the Universe obtained from the hydrodynamical simulation Magneticum . We explored three of the main CMB large-scale anomalies, that is, the lack of a correlation, the quadrupole-octopole alignment, and the hemispherical asymmetry, in the latest Planck data (PR4), where they are detected at a level similar to the previous releases. We also use the simulated secondaries from the local Universe to verify that their impact is negligible.

Aims. Detecting diffuse synchrotron emission from the cosmic web is still a challenge for current radio telescopes. We aim to make predictions about the detectability of cosmic web filaments from simulations. Methods. We present the first cosmological magnetohydrodynamic simulation of a 500 h ⁻¹ c Mpc volume with an on-the-fly spectral cosmic ray (CR) model. This allows us to follow the evolution of populations of CR electrons and protons within every resolution element of the simulation. We modeled CR injection at shocks, while accounting for adiabatic changes to the CR population and high-energy-loss processes of electrons. The synchrotron emission was then calculated from the aged electron population, using the simulated magnetic field, as well as different models for the origin and amplification of magnetic fields. We used constrained initial conditions, which closely resemble the local Universe, and compared the results of the cosmological volume to a zoom-in simulation of the Coma cluster, to study the impact of resolution and turbulent reacceleration of CRs on the results. Results. We find a consistent injection of CRs at accretion shocks onto cosmic web filaments and galaxy clusters. This leads to diffuse emission from filaments of the order S ν ≈ 0.1 μJy beam ⁻¹ for a potential LOFAR observation at 144 MHz, when assuming the most optimistic magnetic field model. The flux can be increased by up to two orders of magnitude for different choices of CR injection parameters. This can bring the flux within a factor of ten of the current limits for direct detection. We find a spectral index of the simulated synchrotron emission from filaments of α ≈ −1.0 to –1.5 in the LOFAR band.

How much cosmological information can we reliably extract from existing and upcoming large-scale structure observations? Many summary statistics fall short in describing the non-Gaussian nature of the late-time Universe in comparison to existing and upcoming measurements. In this article we demonstrate that we can identify optimal summary statistics and that we can link them with existing summary statistics. Using simulation based inference (SBI) with automatic data-compression, we learn summary statistics for galaxy catalogs in the context of cosmological parameter estimation. By construction these summary statistics do not require the ability to write down an explicit likelihood. We demonstrate that they can be used for efficient parameter inference. These summary statistics offer a new avenue for analyzing different simulation models for baryonic physics with respect to their relevance for the resulting cosmological features. The learned summary statistics are low-dimensional, feature the underlying simulation parameters, and are similar across different network architectures. To link our models, we identify the relevant scales associated to our summary statistics (e.g. in the range of modes between $k= 5 - 30 h/\mathrm{Mpc}$) and we are able to match the summary statistics to underlying simulation parameters across various simulation models.