LBNL-2016: CMB-Lensing

From CMB-S4 wiki
Jump to: navigation, search

Session Organizer: Neelima Sehgal

CMB Lensing Parallel Session: Monday, March 7

Status of CMB Lensing Chapter (Gil Holder) [10+5 min]

Important questions to address to properly forecast parameters when including CMB lensing

  • Non-gaussian covariance between 2pt and 4pt lensing signals (Marcel Schmittfull) [10+5 min]
  • Iterative delensing (Kendrick Smith/Uros Seljak) [15+5 min]
  • Including astrophysical systematics in lensing forecasts (Alexandar van Engelen) [5+5 min]

Summary of CMB Lensing chapter (Gil): File:S4 berkeley lensing chapter summary.pdf

Covariance between 2pt and 4pt (Marcel): File:S4 lens covs.pdf


Overview of Chapter:

  • 3 arcmin resolution seems to be fine to get most of lensing S/N from EB
  • Need to know As and tau. Can we get around tau limit with LSST shear cross correlations?
  • Putting in current Planck low ell as opposed to Fisher Planck low ell increases error bars by a factor of 2. What should we use for forecasts? Can we assume adding CLASS etc.?
  • Unless S4 will have lmin~5, we will need to delens to constrain r

2pt/4pt lens covariance:

  • Analytic covariances are in hand but need to be precisely checked with simulations. However, we can use these analytic covmats now in science book, and quantify to what percent they are accurate.

Iterative delensing:

  • Forecasts should be ok, but should monte carlo error bars
  • Allison et al said get only 3% improvement adding iterative delensing - tau was probably limiting this, and that is why didn't see gain in delensing
  • Useful numbers: at 5uK-arcmin, delensing starts to help / at 2uK-arcmin, iterative delensing starts to help
  • r forecasts are super sensitive to low ell assumption
  • In limit of no noise, Kendrick claims can delens perfectly - Uros says curl imposed a floor - Kendrick is not sure
  • 3 arcmin vs 1 arcmin resolution not so important for iterative delensing - more important thing is low noise
  • Factor of 3 lower noise in EB reconstruction with iterative delensing / Factor of 2 lower noise in ClBB when get to 1uK-arcmin noise levels

Summary of Chapter and Parallel Session: File:S4 lens summary.pdf

Wiki navigation

Return to main workshop page