Using cosmological simulations to develop large-scale structure emulators to constrain dark sector physics

Student: Alberto Acuto
Supervisor: Ian McCarthy

The predicted large-scale distribution of matter in the Universe depends sensitively on the assumed nature of dark matter and dark energy.  Therefore by combining observations of large-scale structure (such as gravitational lensing) with theoretical predictions, we can infer the properties of these mysterious components and test fundamental physics.  The computational challenge is to produce precise theoretical predictions for a wide range of cosmological scenarios whilst also realistically capturing the effects of important astrophysical processes (such as negative feedback from supermassive black holes) that can also modify the distribution of matter.  This requires the use of large-scale self-consistent hydrodynamical simulations, that start from cosmological initial conditions and follow the evolution of matter into the non-linear regime, solving simultaneously for the gas, stellar, black hole, and dark matter evolution in the presence of an evolving cosmological background.

The proposed project entails the production of a new suite of custom state-of-the-art simulations and the use of advanced statistical methods (e.g., Gaussian process modelling) to develop so-called "emulators", which will allow one to precisely predict the matter distribution for arbitrary sets of cosmology parameters.  The development of such tools will represent a step change in large-scale structure cosmology which promises to yield exciting advances in our understanding of dark sector physics.