Cosmological Simulations for Large-Scale Sky Surveys

PI Name: 
Salman Habib
PI Email: 
habib@anl.gov
Institution: 
Argonne National Laboratory
Allocation Program: 
INCITE
Allocation Hours at ALCF: 
100 Million
Year: 
2014
Research Domain: 
Physics

The next generation of large-scale sky surveys aims to establish a new regime of cosmic discovery through fundamental measurements of the universe’s geometry and the growth of structure. The aim of this project is to accurately characterize key quantities, such as the spatial statistics of the distribution of mass, and the sum of neutrino masses and how they impact cosmic evolution and structure formation. By looking at the signatures of dynamical dark energy models on the large-scale structure of the universe, the project will also address fundamental questions at the junction of cosmology and particle physics.

This INCITE project will use the Hardware/Hybrid Accelerated Cosmology Code (HACC) computational cosmology framework to run two sets of simulations: one to build precision emulators to attack the problem of inference, the other to generate realistic synthetic sky catalogs and simulated observations. Additionally, this project will rely on the Cosmic Calibration Framework to build precision prediction tools, or emulators, for survey observables; and pipelines to generate synthetic galaxy catalogs.

Emulators, which will span multiple cosmological probes, are now widely recognized as an essential component of next-generation cosmological analyses, and the synthetic sky catalogs are an invaluable resource for modeling the observation chain, testing the inference pipeline, and investigating sources of possible systematic errors. The desired improvements for simulations over the next decade are measured in orders of magnitude; high accuracy and robustness are central requirements that must be satisfied by this program.

The proposed simulation campaign will support a large number of follow-on projects. Therefore, to fully exploit its scientific power, the simulations and data products will be shared with the community at large.