Estimates of delensing efficiency

From CMB-S4 wiki
Jump to: navigation, search

Delensing with a dedicated telescope at the pole - AL=0.13

This was the assumption made for the CDT report. At the time of writing no new reference configuration is available. The forecast for the delensing efficiency for the CDT configuration was AL=0.13. This does not include the large area survey.

Delensing with the large area telescopes from Chile - AL=0.27

Here I will rely on the recent configuration denoted 'Survey Performance Expectation 01' in the posting Survey Performance Expectations. The post does not contain all required information (e.g. lknee and the exponent for the 1/f contribution). Based on the corresponding pdf, I assume that for polarization lknee=800 and an exponent of -1.5. This leads to the following polarization noise curves

NTT01 pess NL.png

Here the colors go from blue to red with increasing frequency. Assuming a foreground model that contains a 2-component dust model and synchrotron, including dust-synchrotron correlations, I determine the weights for an internal linear combination map. The weights are

NTT01 pess wl.png

and the polarization noise power in the ILC map is

NTT01 pess NL ILC.png

The model does not incorporate spatial variation of spectral parameters so that this is highly optimistic and will underestimate the noise. The size of this effect should be studied.

A Fisher forecast for iterative delensing relying on this noise curve predicts AL=0.27.


Delensing with the large area telescopes from Chile and a dedicated LAT for delensing - AL=0.20

Here I will assume an additional telescope dedicated to delensing with the same characteristics as the LATs described in 'Survey Performance Expectation 01' and observing the Chile mask shown in the posting Sky masks for simulations II. The coadded frequency maps have the following polarization noise curves

NTT01 NL.png

Assuming the same foreground model as above, the weights are

NTT01 wl.png

and the polarization noise power in the resulting ILC map is

NTT01 NL ILC.png

A Fisher forecast for iterative delensing relying on this noise curve predicts AL=0.20.