Agenda/Notes from Systematics WG telecon July 19 2017

on the call: Brendan, Paolo, Colin, Julian, Shaul, Maurizio, Joy

1. S4 tools (Colin)
   • S4 doesn't yet have an instrument design, so at this point, they are considering generic “additive systematics” == something that adds additional B-mode power and could bias a measurment of “r”
   • A generic noise-like systematic (not rolled off by beam) is considered
   • Looked at an “uncorrelated” systematic (i.e. uncorrelated from band to band) such as residual beam mismatch: a white and a 1/ell version
   • Also looked at a correlated systematic which can bias cross-spectra, same amplitude (in CMB units) in all bands (i.e. looks like a CMB fluctuation)
   • Going back to Fischer forecast, including foreground separation in frequency space: can write down what level of systematic corresponds to as a bias in “r” of 1e-4 (target of sigma_r in 5e-4)
   • Colin is now applying this to the S4 data challenge.
   • Longer term this is meant to provide a benchmark to judge systematics against as a real instrument design takes shape.
   • What do we need in order to perform this type of analysis for the CMB probe Imager, given our set of bands and noise levels?
     • Can be done just with map noise and frequency bands: though of course it assumes an analysis method based on BICEP/Kick which may or may not be applicable to a full-sky analysis.
     • Also noise shape for a full-sky mission is likely to be quite different. Atmosphere enters in as an ell_knee
     • Simple parameterization for foreground estimation: probably fine for ~3% sky, but for Probe would use a full-sky template fit.
     • Julian points out that there will be a map-based analysis within S4 that will perhaps test this applicability. The more pixels that are available for a full-sky map, a maximum likelihood starts to become computationally expensive.
2. Planck low ell polarization paper Planck intermediate results XLVI. Reduction of large-scale systematic effects in HFI polarization maps and estimation of the reionization optical depth (Brendan)
   • Summary plots
     • Fig 17: estimates of HFI systematic errors in EE auto-spectra vs. projected noise
     • Fig 18: null tests of HFI detectors
     • Fig 23: LFI systematic errors
   • Specific plots
     • Fig 2: noise PSDs and CSDs: correlated cosmic ray hits; Fig A.1 shows this propagated to TT/EE/BB residuals.
     • Fig 4: HFI far sidelobe pickup: propagated physical optics models. (note: Physical optics models were not incredibly accurate)
     • Fig 9: residual ADC nonlinearity with gain correction leaves low ell polarization residuals: see Fig B.13 for what it looks like on the sky
     • Fig 10: “Empirical transfer function” unknown systematic
     • Fig 13: relative (detector to detector) gain measured from dipole to better than 1e-5 level
     • Fig 14: ground-based vs. sky-based bandpass leakage correction
     • Fig 30: estimated foreground residuals in HFI
     • Fig A.2: HFI warm readout drifts propagated to residual power spectra
3. Systematics List (Ranajoy, Joy) types_of_systematics.pdf
4. CORE systematics paper hit the arxiv last week (Paolo)
5. Any other business