Skip to content

Power spectrum tests

David Alonso edited this page Jan 30, 2017 · 16 revisions

Validation of CoLoRe sims

Simulations suite:

100 simulations with:

  • N_grid=1536
  • z_max=1.1 (Lbox=4962.7)
  • R_smooth=5 Mpc/h
  • ~7 million objects with positions, redshifts and ellipticities

Analysis

  • Added Gaussian photo-z to sources with \sigma_z=0.03
  • Sample divided into two overlapping redshift bins (see above).
  • Interpolated sources to Nside=128 maps to create maps of the galaxy overdensity and cosmic shear (E and B modes).
  • Computed all (galaxy-galaxy, galaxy-lensing and lensing-lensing) full-sky power spectra with NaMaster for each simulation and pair of bins.
  • Averaged power spectra over all simulations.
  • Estimated theoretical power spectra with LimberJack, accounting for lognormal transformation, and using non-Limber for l<100.
  • The theory must be corrected (analytically) for several smoothing effects:
  • Simulation resolution.
  • Interpolation into spherical shells.
  • HealPix resolution.
  • Compared theory and simulations.

Results

Main resolved issues

  • Accounting for lognormal transformation.
  • Accounting for interpolation into spherical shells.
  • Shot-noise correction.
Clone this wiki locally