Fundamental Cosmology in the Era of Surveys: A Multi-scale Numerical Campaign

We live in a particularly exciting time for cosmology. The wealth of observational data has never been so abundant, with the potential to provide definitive constraints on the nature of dark matter (DM) and dark energy (DE), two fundamental problems in Physics today. Numerous ongoing and upcoming surveys have been designed to make high-precision maps of the large-scale structure of the Universe in an effort to address these problems directly. In order to make sense of these large -- and expensive -- data sets, it is necessary to consider them in the context of equally sophisticated theoretical models. My ongoing FLF project, which lies at the nexus of cosmology, particle physics and high-performance computing, addresses this need, following three broad scientific questions: What are the new frontiers for constraining DM and DE? How does the interplay between dark and luminous matter affect what we infer from future galaxy surveys? What can we learn about the Universe using the smallest “dwarf” galaxies? My research group uses state-of-the-art simulations of galaxy formation to tackle these questions. A significant milestone in my project so far has been the completion of MillenniumTNG (MTNG), which are some of the largest (and most realistic) simulations of galaxy formation ever performed. The results of these data are being used to inform theoretical models connecting dark and luminous matter -- the so-called “galaxy-halo connection” -- that are employed within large galaxy surveys for inferring cosmological parameters, which is one of the fundamental goals of modern cosmology. MTNG data are now being used by research groups across the community. We have also made advances in the study of “modified” gravity (MG), which provides an alternative explanation to DE. My group has produced the first realistic galaxy formation simulations in MG. By creating “mock” samples from the simulated data, we have shown how signatures of MG may be detected in the real Universe. I have also pioneered new methods that use Machine Learning that allow us to capture the complex physics of MG models in a fraction of the time needed to run the full simulations. My theoretical models have also been instructive in devising new tests of DM and galaxy formation in the early universe using the properties of dwarf galaxies, specifically those that orbit the Milky Way. This work suggests that the tightest constraints are possible when using a broad range of observables: from low surface brightness features around nearby galaxies, to the gravitational wave signals from dwarf galaxy mergers in the early Universe. Excitingly, these are exactly the sorts of observations that will targeted by upcoming surveys in the next decade. Over the next several years, I will extend upon the progress I have made in the FLF so far and drive forward advances along each of these three themes. Progress I have made in studying the galaxy-halo connection will be used to inform the next critical phase of analysis for the Dark Energy Spectroscopic Instrument survey (DESI) and Euclid, and to define the mission scope for DESI's successor in the 2030s. Using a new, cutting-edge model of galaxy formation called COLIBRE, I will also drive forward new advances in the study of dwarf galaxies in the early- and late-time Universe to enable robust constraints on the nature of DM and the physics of galaxy formation.