Chicago-2016: Coordinating r and non-r Forecasting

From CMB-S4 wiki
Jump to: navigation, search

Next-generation forecast planning: coordinating "r" and "non-r"

Moderator: Tom Crawford

File:R nonr.pdf

Questions to consider:

  • Do we even really need to?
    • If not, what to do about de-lensing?
  • If we do, what are the first couple of steps?
    • suggestion: N_l and foregrounds
  • If so, how far do we go?

Notes

Do we need to do it?

  • No because of different foreground sky issues
    • But delensing
  • Consistent noise
  • Similar foreground simulation needs
  • Commonality in delensing
    • Hardest thing to simulate => common will to solve ...

First steps

  • Beyond white noise
    • N_ell
    • Based on re-scaled BICEP noise curves (from 3 of 4 atmospheric windows)
    • No N_ell(TT) curves yet
      • Could use atmosphere models
      • No, Use observed PSDs (ACT, SPT)
        • Does this include atmospheric cross-correlation? (Cross-detector, cross-band)
        • In the band-power covariances, but based on non-simultaneous observations
        • Can you fit this out (cf. SUSIE), and at what cost/benefit.
        • Provide a matrix at each ell ...
    • "Stay the course for large scale; implement N_ell for small scale; push for more expts to release data"
    • Noise maps?
  • Foregrounds
    • Existing temperature simulations & component separation codes
      • Are they sufficient?
    • Common CS codes between r/non-r?
      • Maybe not, but common foreground inputs, compare codes.
      • Provides standard outputs to non-CS teams.
    • Many separate issues; lensing/delensing is always common
      • ACTION ITEM: How fast does small-scale dust polarization fall off and is this a total non-issue for 2-point?
    • Common repository for simulations & results?
      • PSM & pySM for gal foreg
      • Nick's fast extra-galactic (n-body based) sims
    • Products rather than (just) codes!
    • Link on the wiki under sky-modeling already exists: USE IT
    • Push towards map-based forecasting
      • Agree on a set of maps spanning a useful parameter range
      • Agree on a set of survey cut-outs
      • Agree on a set of noise models
      • Incorporates all correlations/biases etc
      • NERSC resources as a central repository with storage and cycles available to all.
        • sign up! httpd://crd.lbl.gov/cmb

Other points made (not r/non-r coordination-specific)

  • Future audience (beyond just DOE science)
  • Feeding into CDT