6. September 2019
B. Bose, J. Byun, F. Lacasa, A. Moradinezhad Dizgah, L. Lombriser, arXiv:1909.02504
Future large-scale structure surveys will measure three-point correlations with high statistical significance. This will offer significant improvements on our understanding of gravity, provided we can model these statistics accurately. Here we assess the performance of various theoretical modelling schemes for the matter bispectrum, including perturbative and halo-model based approaches as well as fitting formulae, which we compare to measurements from N-body simulations. We conduct this analysis for general relativity and two alternative theories, f(R) gravity and the DGP braneworld model. The comparison is performed into the highly non-linear regime of structure formation, up to 𝑘=4ℎ/Mpc. We furthermore compute the lensing convergence bispectrum from these theoretical approaches. We find that a halo-model corrected fitting formula achieves the best overall performance. Despite this, we also find that all current modelling prescriptions in modified gravity, in particular for theories with scale-dependent linear growth, fail to attain sufficient accuracy for suitability to lensing in the context of Stage IV surveys such as EUCLID.