Leads: Ian McCarthy, Joop Schaye, Matthieu Schaller, Willem Elbers, and the Virgo II Team
A diverse set of new observatories are about to map the large-scale structure in the Universe down to smaller scales than was possible before, which will greatly improve our ability to find inconsistencies in the standard cosmological model. However, the scientific potential of these surveys can only be realised if the model predictions are at least as accurate and precise as the observations.
Although the clustering of the dark matter is well understood, mainly because it is collisionless and therefore only subject to gravity, this is not yet the case for the ordinary, baryonic matter. Unlike the dark matter, baryons are subject to pressure force and are hence redistributed by galactic winds driven by supernova explosions and accreting supermassive black holes. Unless these effects are taken into account, they will cause systematic errors in the measurements of cosmological parameters of tens of per cent for upcoming weak gravitational lensing studies, thus eliminating most of the foreseen improvements on the current constraints.
The FLAMINGO project from the Virgo Consortium addresses this urgent problem by providing a suite of hydrodynamical simulations of volumes large enough to predict the large-scale structure with a resolution that is sufficiently high to model the gas flows caused by feedback processes
associated with galaxy formation. Because these feedback processes cannot yet be predicted from first principles, the free parameters of the model are calibrated to key observables using machine learning techniques.
The need to model large volumes at relatively high resolution necessitates massive simulations using state-of-the-art codes. Indeed, FLAMINGO includes the largest ever hydrodynamical simulation of the entire history of the Universe. The simulations vary the uncertain astrophysical parameters as well as the mass of the elusive neutrino particle, which we expect to be able to measure from upcoming observational surveys.