The DREAMS Project: A New Suite of 1,024 Simulations to Contextualize the Milky Way and Assess Physics Uncertainties

Abstract

We introduce a new suite of 1,024 cosmological and hydrodynamical zoom-in simulations of Milky Way-mass halos, run with Cold Dark Matter, as part of the DREAMS Project. Each simulation in the suite has a unique set of initial conditions and combination of cosmological and astrophysical parameters. The suite is designed to quantify theoretical uncertainties from halo-to-halo variance, as well as stellar and black hole feedback. We develop a novel weighting scheme that prioritizes regions of the input parameter space, yielding galaxies consistent with the observed present-day stellar mass--halo mass relation. The resulting galaxy population exhibits a wide diversity in structural properties that encompasses those of the actual Milky Way, providing a powerful statistical sample for galactic archaeology. To demonstrate the suite's scientific utility, we investigate the connection between a galaxy's merger history, focusing on Gaia-Sausage-Enceladus~(GSE) analogs, and its present-day properties. We find that galaxies with a GSE analog have lower star formation rates, more compact disks, and more spherical stellar halos. Crucially, significant halo-to-halo scatter remains, demonstrating that matching more than the most significant events in the Milky Way's past is necessary to recover its present-day properties. Our results highlight the necessity for large statistical samples to disentangle the stochastic nature of galaxy formation and robustly model the Milky Way's unique history.

Figures (17)

• Figure 1: Corner plot showing the weights associated with the astrophysical parameters in the emulated DREAMS CDM dataset. Each 2D panel shows the median weight, as defined in Section \ref{['sec:weight_method']}, for a pair of parameters, marginalized over the other one. The highest weights are indicated by yellow/orange, while the lowest weights are indicated by blue/black. The fiducial TNG values are shown with a cyan 'o'. The top panel of each column shows the marginalized weight distribution for a given parameter, with the black line corresponding to the median and the gray band to the 16--84% range. This procedure down-weights regions of parameter space that do not produce the data-driven SMHM relation from SAGAV. Notably, there are significant degeneracies between some of the parameters, which could be further constrained by including other observations into the weighting procedure.
• Figure 2: The SMHM for the emulated DREAMS CDM galaxy population, used to constrain the astrophysical parameters $\kappa_w$, $\widebar{e}_w$, and $\epsilon_{f,{\rm high}}$. The left panel shows the full, unweighted distribution of emulated galaxies as a 2D histogram, with color indicating the density of points. The center panel shows the weighted distribution, after applying the scheme from Section \ref{['sec:weights']} to down-weight unphysical regions of parameter space. The right panel directly compares the unweighted (red) and weighted (blue) populations, showing their mean (solid lines) and 1$\sigma$ intrinsic halo-to-halo scatter (shaded bands). In all three panels, the UM-SAGA SMHM relation from SAGAV is shown for comparison as a dashed grey line (mean) and solid gray bands (1$\sigma$ scatter). Applying the weights brings the mean of the simulated population into closer agreement with observations.
• Figure 3: A summary of the key physical properties of the MW-mass galaxies in the DREAMS CDM suite. The value for each property is normalized and clipped to the $1^{\rm st}$ and $99^{\rm th}$ percentile of the simulated values, see Table \ref{['tab:props']} for the unnormalized values. The weighted distributions for the 803 DREAMS disk galaxies are shown as colored violins, grouped by galactic (gray), bulge (blue), disk (yellow), and halo (pink) properties. Note that these results correspond to the actual simulated galaxies, not the emulated set shown earlier. Detailed descriptions for each property can be found in Section \ref{['sec:props']}. The black points with 1$\sigma$ error bars show the corresponding observed values and their uncertainties for the MW taken from 2016Bland. These results put our Galaxy in context, providing a direct comparison with the broader simulated population in the DREAMS suite.
• Figure 4: The fraction of emulated DREAMS hosts that undergo an RA event (left) and a GSE-like event (right), plotted as a function of the five simulation parameters ($\Omega_{\rm m}$, $\sigma_8$, $\widebar{e}_w$, $\kappa_w$, $\epsilon_{f,{\rm high}}$), normalized to unity. The lines are the means of the emulated samples with the $1\sigma$ error bands corresponding to the aleatoric and epistemic uncertainties. Note the difference in scales between the y-axes of each figure panel. We adopt the definition of an RA event from 2025Folsom where the debris from the merging galaxy makes up 50% of the $z=0$ inner stellar halo and has a significant radial anisotropy, $\beta>0.5$. A GSE merger is an RA event, with additional criteria on the mass and accretion of the progenitor, as well as the host's accretion history (see Section \ref{['sec:gse_rarity']}). The occurrence of RA- and GSE-like events is most strongly correlated with the energy of SN winds, $\widebar{e}_w$.
• Figure 5: A comparison of key structural properties between the general emulated MW-mass galaxy population (black) and the subset of galaxies with GSE analogs (red). The six panels show distributions for the star formation rate (SFR; top left), disk radial scale length ($R_\mathrm{disk}$; top middle), stellar halo mass ($M_{*, \mathrm{halo}}$; top right), stellar halo break radius ($r_s$; bottom left), inner halo shape ($q_\mathrm{in}$; bottom middle, and outer halo shape ($q_\mathrm{out}$; bottom right). The population of galaxies with a GSE analog shows systematically different distributions for most properties, including a strong tendency toward lower SFRs, smaller $R_\mathrm{disk}$, and higher $q_\mathrm{in}/q_\mathrm{out}$. However, we find no significant systematic difference in $M_{*, \mathrm{halo}}$ between the two populations.