Galaxy Simulations

Formation and Evolution of galaxies


Simulations have always become an important and complementary tool to test theories on galaxy formation and evolution. Simulations allow the researchers to generate and trace the evolution of a wide range of galactic morphologies. These simulations start from initial conditions generated to resemble the primordial universe, and evolve under well-known physical processes. Results are compared with real observed systems. If results from observations and simulations match, simulations can be used to predict properties of the observed systems that has not been detected, and to understand its formation process. When simulations do not match with observations, new theories need to be developed and implemented in the codes, in order to account for new physical processes. Also, observations have to be checked looking for possible observational biases.

Simulations have become an important piece on the evolution of scientific knowledge. The use of simulations filled a gap between observations and theory as they provide tools to test theories that cannot be tested by current observations. Simulations can also make predictions of what should be observed when new observatories and techniques make it feasible.

The circumgalactic medium

Faint, tenuous gas that surrounds galaxies is one their most unknown components. This gas plays an important role on regulating galaxy formation, and also gives valuable information about recent galaxy evolution.

It is difficult to observe the gas in the circumgalactic medium. CGM emission lines can only be observed from the closest galaxies. On the other hand, by observing background luminous sources (i.e. QSO), it is possible to detect CGM absorption lines, and from them to infer some CGM properties.

Two theories about the origin, properties and distribution of the CGM gas in galaxies, and its dependence on galaxy mass and redshift, are now being considered. CGM gas can be mostly in cold-dense inflowing filaments (Cold scenario), or in a low-density warm-hot gas filling the entire galactic halo (Warm-Hot scenario).

Simulations play an important role on studying the formation of the CGM, as function of galactic mass and redshift. From simulations the evolution of the CGM can be traced. General properties of the CGM are also easily obtained from simulation, thus a comparison with the current CGM models is straightforward. Currently, several predictions on where and how to look for the CGM structures has been made thanks to the new high resolution simulations.

Mass fraction of collisionally ionized OVI as function of redshift. Solid lines show mean values obtained from the NIHAO runs in three di erent Mv bins (see legend). Dashed lines show the same as solid lines but for the VELA runs. Dot-dashed vertical black line indicates the peak on the UV-background intensity. Shadowed regions show the 1 dispersion of the mean in the NIHAO results. Only gas in the CGM (i.e. 0:1Rv to Rv) has been considered.
The role of major and minor mergers

Galaxy formation and evolution has been dominated by interactions and mergers. At the early universe major mergers were mainly the mechanism that drove the growth of galactic systems. In the present days, major mergers became rare and minor merger role the formation of density and kinematic substructures in galaxies.

Simulations of both, isolated systems and cosmological boxes, allow the researchers to test the effects of mergers in the kinematics and structure of galactic systems as function of the redshift.


The Hubble sequence was one of the first classification of galaxies, according to their morphology. When looking to images provided by large galaxy surveys like the SDSS, a full galactic zoo appears. Spiral galaxies, galaxies with one or multiple bars, with rings, warps, flares and a large number of faint structures are common.

One of the big challenges on the study of galaxy formation and evolution is to understand the formation of all these substructures and how it affects the evolution of global parameters like chemical evolution, gas fraction and star formation rates.

Simulations allow the researchers to characterize these structures, investigate their origin and also to induce their formation in unperturbed galactic disks. Results can be compared with observations in order to enrich the general knowledge of how galaxies form and evolve.

Figure 1. Face-on (first column) and edge-on view (last two columns) of stars in our simulated MW-sized galaxy. Each row correspond to a different stellar age. From top to bottom: 11−13.467 Gyr, 10−11 Gyr, 9−10 Gyr, 4−9 Gyr, and 0−4 Gyr. The color code indicates stellar age.
Figure 2. Edge-on (top) and face-on (bottom) views of a young stellar population (0−7 Gyr) of model G.323 at z = 0. Both panels span 50 kpc in the x-axis. The y-axis spans 22 kpc in the top panel and 50 kpc in the bottom panel. The color code indicates qualitative stellar age.


Our Solar System is embedded on the galactic disk of a disk galaxy known as the Milky Way (MW). The MW should be a perfect laboratory to test theories of galaxy formation and evolution, as it is the closest galactic system we have. We can study position and kinematics of faint stellar streams in the Galaxy halo, of ultra-faint dwarf galaxies that are being accreted, and also of a big sample of disk individual stars.

However, it has been challenging to unveil structures in the galactic disk due to extinction by gas and dust. Till now, we do not know the general structure of the MW, i.e. we do not know the number and morphology of the MW spiral arms, bar, warp and/or flare, and not even the real virial mass of the Galaxy.

Recent large surveys like Gaia are providing a large amount of data that is currently being analyzed. Several important discoveries regarding MW structure and kinematics have already been published. Numerical simulations can play an important role to understand these new discoveries and the processes that have driven to the observed situation.

Large scale structures

The formation and properties of large scale structures in external galaxies have been long studied from theory, simulations and observations. However, due to the large distances from these external galactic systems, assumptions need to be made in order to study kinematics and dynamics of large stellar structures like spiral arms. No information on individual stars and on how galactic large scale structures affect them can be obtained from currently available telescopes.

On the other hand, in the MW, we have available information about kinematics of individual stars but not about the morphology of large scale structures.

Simulations can play an important role on connecting information of morphology and dynamics we obtain from external galaxies with its effects on the kinematics of individual stars, since large scale structure dynamics/morphology and also kinematics of individual stars affected by these large scale structures both can be obtained from simulations.

Roca-Fàbrega et al. 2014
Disk density distribution of models TWA1 (top left), B1 (top right), B5 (bottom left), and U5 (bottom right).
Star formation history
Most probable values of the mean SFR for the age bin obtained from the posterior PDF. The vertical error bars indicate the 0.16 and 0.84 quantiles of the posterior PDF. The horizontal error bars indicate the size of the age bin. The grey and black dashed lines are, respectively, an exponential function and a distribution formed by a bounded exponential plus a Gaussian, fitted to the G12NP-S results. The grey solid line is the exponential part of this exponential plus Gaussian fit.

Galaxies are shaped by mergers, interactions and cosmic inflows of low-metallicity gas. All these events are followed by star forming episodes that leave imprints on the galaxy stellar population. In external galaxies, global luminosity is used to compute its stellar population and from it the star formation history. This is done by comparing integrated luminosity and spectra with synthetic spectra obtained from stellar libraries and some theoretical assumptions.

In the Milky Way we have available a large number of individual stellar ages. This makes the computation of the Star Formation History almost direct, with a small number of assumptions. From the SFH it is possible to observe variations on the cosmological star formation history that give information on the acretion/merger events. However, the problem of identifying a star formation event with a merger or with another process is highly degenerated. Simulations can play an important role on breaking this degeneracy.

Kinematic structures in the Galactic disk

The high quality data provided by recent surveys like Gaia unveiled a large number of kinematic structures in the velocity space of individual stars in the solar neighborhood.

Several density structures and disk internal processes can drive to the formation of such structures. Each one of these processes leaves a specific imprint on the velocity space.

Comparison with simulations can help to establish a relation between properties of kinematic structures and its origin.

Density (top) and lv (bottom) polar plots for test particle models TWA1 (left), TWA2 (middle) and TWA3 (right) from Table 1. The thin black lines show density contours of regions with density above the mean. The thick and dashed horizontal black lines show the position of CR and OLR radius, respectively. The thick black lines show the position of the Fourier m = 2 mode locus. The white regions at the bottom panels correspond to regions where the lv relative error is above 50 per cent.
Galactocentric radial VR (top) and azimuthal Vφ (bottom) velocities as a function of cylindrical coordinates formodels TWA0, TWA1, and TWA2 (left, middle, and right). Colours show the median velocity in bins in cylindrical coordinates of size R = 0.5 kpc and φ = 10◦. For Vφ we plot Vφ− < Vφ >R where < Vφ >R is the average median over all bins at the same radius R, that is we subtract the average rotation velocity at each radius. To make the comparison easier, the colour scale is the same for all panels but the scale indicated above each panel shows only the range for that particular model. The theoretical locus of the arms is shown as a thick black line. The solid and dotted curves indicate the over-density of the spiral arms where the density contrast (N− < N >R)/ < N >R is 0 and 0.2 of the maximum value, respectively, where N is the number of particles in each pixel of the grid and < N >R is the azimuthal average. The locations of the main resonances CR, ILR 2:1, ILR 4:1, OLR 4:1, OLR 2:1 are shown with black horizontal lines (solid, dashed, dotted, dashed-dotted, and long-dashed, respectively). The rotation of the Galaxy is towards the left. The black asterisk shows the location of the Sun.
Density structures in the Galactic halo

All galactic systems show a complex stellar density distribution in their galactic halos. This complexity reflects the process involved on the recent galaxy evolution. Streams, shells and other interesting density structures can be observed in galactic halos, providing information on the global mass and density structure of the main galaxy.

The Milky Way halo can be studied in much more detail than the one in external galaxies. Simulations can help to understand the origin of these structures by reproducing observations of the present day halo distribution.

Project AGORA

The AGORA High-resolution Galaxy Simulations Comparison

The Agora project is a worldwide collaboration started back in the 2012. It investigates galaxy formation with high-resolution simulations and compares the results across code platforms and with observations. This is a one-of-a-kind, inter-institutional effort by more than 160 participants from over 60 institutions worldwide (as of 2019).

As mentioned above, the main goal of the Agora project is to compare simulations that have been obtained with the most commonly used numerical codes in astrophysics, and all of them starting from the same initial conditions. This comparison has been designed to allow researchers know how their results are code dependent or not, and also to look for possible bugs in their numerical codes. Several comparisons has been initiated from 2012. The first subproject was to test differences on the gravity solvers used by each one of the participant codes.

The second project revolved around the formation and evolution of an isolated Milky Way like galaxy, including both, gravity and hydrodynamical processes. Currently several new subprojects are ongoing, all of them including a hydro+N-body simulation of the formation of a Milky Way-like galaxy system, in a cosmological context. The largest effort of the collaboration is now on analyzing the formation of the circumgalactic medium as so the generation of gas clumps in the galactic disk. Metallicity evolution and the formation of large scale structures like spiral arms and bars are also being studied.