How Much Information Can Be Extracted from Galaxy Clustering at the Field Level?

Area of Science:

• Cosmology
• Astrophysics
• Statistical physics

Background:

• Cosmic large-scale structure provides crucial information about the universe's evolution.
• Constraining cosmological parameters like σ8 is essential for understanding dark matter and dark energy.
• Traditional methods often rely on lower-order statistics like power spectrum and bispectrum.

Purpose of the Study:

• To derive optimal Bayesian cosmological constraints using nonlinear tracers of cosmic large-scale structure.
• To quantify the improvement in σ8 constraints using a field-level approach compared to simulation-based inferences.
• To explore cosmological information encoded in galaxy clustering beyond n-point functions.

Main Methods:

• Utilized nonlinear information from simulated dark matter halos within a large comoving volume.
• Employed a Lagrangian effective field theory-based forward model, leftfield, to sample initial conditions, tracer bias, and noise parameters.
• Compared field-level constraints with simulation-based inferences of the power spectrum and bispectrum.

Main Results:

• Achieved a factor of 3.5 to 5.2 improvement in σ8 constraints using the field-level approach.
• Reduced the σ8 constraint from 20.0% to 5.7% (and 17.0% to 3.3%) by incorporating nonlinear information.
• Demonstrated the power of field-level analysis for extracting cosmological information.

Conclusions:

• Field-level cosmological constraints from nonlinear tracers offer significant advantages over traditional methods.
• This approach provides direct insights into cosmological information within galaxy clustering.
• The leftfield forward model enables precise parameter estimation in cosmology.