The Science Traceability Matrix of the CDT report (Section 3.2) -- report linked HERE -- contained Q/U map noise levels for the "delensing instrument," which were calculated by Tom Crawford. Some months later Clem and I noticed that the 270 GHz map depth looked too good, which led to a long correspondence with Tom to understand how these numbers were computed and to me attempting to reproduce them. In this process I arrived at different numbers than the ones in the STM table. Finally, Tom reproduced my calculations and agreed that the updated numbers were the correct ones. For posterity, I attach my code and some details of our correspondence.

The code used to produce the new numbers is as follows. (The 1.3 muK-arcmin SPT performance number comes from Tom.)This code yields the following (rounded) Q/U map depths, in muK-arcmin, for [20, 30, 40, 95, 145, 220, 270] GHz:

Compared to the STM numbers:

Tom's recalculation for [40, 95, 145, 220, 270] GHz yielded:

which agree very well with the first row.

To get to the numbers that Raphael Flauger posted under the "Hi-res Ultra-deep Field" in this summary posting, one simply needs to take the first row, scale down by