Shear calibration LSST
(Emmanuel Schaan writing) Here are some preliminary forecasts for calibrating the shear bias from LSST using CMB lensing from CMB S4 (paper in preparation with Tim Eifler, Elisabeth Krause, Hironao Miyatake, Olivier Doré, Jason Rhodes and David Spergel). We use all the 2-point correlations of galaxy positions, convergence from galaxies and convergence from the CMB. In particular, we use only the convergence from CMB S4 (e.g. not the temperature and polarization power spectrum). We constrain 10 shear biases for each of the 10 source redshift bins of LSST (File:Zbins LSST.pdf), while marginalizing over cosmological parameters, galaxy biases, and photo-z means and variances in each bin. We use the full non-Gaussian covariances for all the 2-point functions, and explore the likelihood with MCMC.
We consider different combinations of observables (File:Observables.pdf) and find the following 68% confidence constraints on the shear biases: File:LSST summary m.pdf. The main result is that the shear biases can be constrained to close to the LSST requirements, to about 0.5%. The method works best at higher z (where the correlation with CMB lensing is best), which is also where it is hardest to calibrate the shear otherwise. The calibration is robust to Gaussian photo-z errors and intrinsic alignments (not shown here).
We assume that LSST and CMB S4 fully overlap on 18,000deg^2 (fsky=44%). The fiducial run assumes 1muK' and 1' for CMB S4, lmin=30 and lmax=3000 for T and 5000 for E and B. We show that the calibration is sensitive to the CMB S4 sensitivity File:LSST varying noise.pdf and not to the resolution File:LSST varying beam.pdf. This is mostly because of the fixed lmax imposed.