Seminar by Thomas Bakx, Utrecht
Current and upcoming surveys of the large-scale structure (LSS) of the Universe will deliver unprecedented constraints on the clustering properties of galaxies, which are luminous tracers of the underlying dark matter distribution and therefore contain rich cosmological information. In order to harness the full potential of such data, it is crucial to understand the clustering of galaxies under the influence of gravity beyond the linear regime. At large separations or small wavenumbers, these nonlinearities can be described from first principles with perturbation theory. However, sufficiently efficient computation of the resulting corrections to general N-point statistics (i.e. beyond just the power spectrum at next-to-leading order) has remained challenging due to the complicated nature of the associated loop integrals. In this talk, I will describe a new method called COBRA which allows for efficient computation of N-point statistics at any order by linearly separating the cosmology dependence in the linear power spectrum from the scale dependence. As such, loop integrals reduce to very small matrix multiplications and do not require direct emulation. This results in a substantial advantage over conventionally used methods (e.g. FFTLog) already for the power spectrum at one loop, and likely even more so for more complicated cases. These results may also be relevant for other cosmological observables in e.g. weak lensing or CMB studies. [Based on https://arxiv.org/abs/2407.04660]
Host: Colin Hill