LBNL-2016: Inflation

From CMB-S4 wiki
Jump to navigationJump to search

Session Organizers: Lloyd Knox, Raphael Flauger, Sarah Shandera

Break-out sessions:

First 50 minutes: State of the Inflation Chapter (Raphael Flauger and Sarah Shandera)


S4 and CMB temperature anomalies

We discuss how future polarization measurements can help to further the study of known statistical anomalies of CMB temperature fluctuations. As a proof of concept we focus on one example: the hemispherical power asymmetry. Using temperature data and the TE correlation given by the theory, we can extract an E-mode polarization map in which the piece that is correlated with temperature has been removed. The theory predicts that what remains is Gaussian random statistically isotropic, thus an observation of an anomalous asymmetry in this map (especially one aligned with the T asymmetry) would be a good suggestion that the feature observed in T is not just a fluke. We study how the sensitivity of a measurement would depend on sky coverage by analyzing four simplified test cases: a sky in which only galaxy and point sources are masked (using Planck's Common mask), two in which we additionally mask polar caps (Celestial North or South) and finally one in which the entire celestial northern hemisphere is masked together with foreground.

Applying the method of choice to temperature data, we find a dipolar asymmetry with a p-value of 0.1% on a sky masked with the Common mask. Assuming a measurement with same significance in the uncorrelated E-mode signal masked identically, we find through best-fit realizations what would be the corresponding p-value for the three other configurations. The result is shown in the figures via the cumulative distribution function. The vertical line marks the amplitude corresponding to p=0.1% for the case where only the Common mask is used. As the labels show, the p-value increases to approximately 1% when one of the celestial polar caps is also masked and then to roughly 20% when only the southern celestial sky is observed. This argues for significant northern and southern celestial sky coverage. We expect the case for northern sky coverage to be strengthened if one more closely mirrors the temperature anomaly by comparing only the suppression of power in the northern sky. Our ability to similarly probe another anomalies should also be investigated.

CDF: Media:M_ODwyer_CDF.png
p-value: Media:M_ODwyer_pval_zoom_log.png

Figures: O'Dwyer, Knox, Starkman, Copi (to appear)

Final 10 minutes: Toward inflation-science-driven experimental design (Kovac and Dunkley)

  • Importance of realistic forecasting.
    • What are the inputs from S1/S2 experiments needed to ground S4 forecasts in reality?
    • How/when to incorporate additional feedback from S3?
  • For a given expt set-up, what are our forecast r limits?
    • Is our forecast robust to Galactic foreground simulation?
    • Does our planned expt include sufficient null tests (including seeing a result at different frequencies, region of sky, etc)?
    • Does our forecast capture realistic systematics?
    • Is our forecast robust to different methods of foreground cleaning?
    • Does our forecast include realistic residual lensing noise?
  • Can we then optimize the experimental set-up for r?
    • Plan for r=0 and ultra-deep limits, with fallback plan in case r is detected? See SB Fig 3 flowchart.
  • Are design requirements (resolution, freq, sky area) same for 'high-ell' inflation parameters?
  • To this end, now have set of codes that implement different r-forecast methods including Victor Buza et al's spectrum-based cleaning code based on BK14 analysis, Josquin Errard/Stephen Feeney's forecast code, David Alonso's map-based cleaning code developed for ACT (plus more).
    • All must be validated for unbiased r recovery using map-level sims with realistic, common inputs: map noise and FG models grounded in current observations

Wiki navigation

Return to main workshop page