Spherical collapse, formation hysteresis and the deeply non-linear cosmological power spectrum

I examine differences in non-linear structure formation between cosmological models that share a z=0 linear power spectrum in both shape and amplitude, but that differ via their growth history. N-body simulations of these models display an approximately identical large-scale-structure skeleton, but reveal deeply non-linear differences in the demographics and properties of haloes. I investigate to what extent the spherical-collapse model can help in understanding these differences, in both real and redshift space. I discuss how this is difficult to do if one attempts to identify haloes directly, because in that case one is subject to the vagaries of halo-finding algorithms. However, I demonstrate that the halo model of structure formation provides an accurate non-linear response in the power spectrum, but only if results from spherical collapse that include formation hysteresis are properly incorporated. I comment on how this fact can be used to provide per cent level accurate matter power-spectrum predictions for dark energy models for k <= 5 h Mpc(-1) by using the halo model as a correction to accurate Lambda CDM simulations. In the Appendix, I provide some fitting functions for the linear-collapse threshold (delta(c)) and virialized overdensity (Delta(v)) that are valid for a wide range of dark energy models. I also make my spherical-collapse code available at https://github.com/alexander-mead/collapse.