Deeper SAT from Chile II

From CMB-S4 wiki
Revision as of 14:06, 1 April 2019 by Keskitalo (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

March 22, 2019 - Reijo Keskitalo and Julian Borrill


We continue the work on maximizing the depth of Chile SAT observations started in Deeper SAT from Chile. In this post we have made the following changes to the proposed scan strategy:

  • lower the minimum observing elevation from 50 to 45 degrees
  • promote the Southern deep field over the Northern field: when ever the two are both available choose the Southern field
  • move the Southern field to the center of the Southern hole with minimal foreground polarization

Here are the new locations of the tiles:

Priority by pixel fwhm01 nbin10.full.v2.png

As before, red tiles are top priority and are always observed when available. The scheduler maintains an even number of observations in each of the Southern red tiles. Black and grey tiles are second and third tier targets that are observed when the higher tiers are not available.

The hitmaps and noise variance maps shown in this post are available at NERSC under /global/cscratch1/sd/keskital/cmbs4_scanning_simulations

You can find a movie of the Chile observable sky in Chile Scanning Movie.


The new observing schedule spends a total of 77.7 days over the year observing the Southern deep field (previous schedule spent 50.3 days). Northern patch is observed for 31.8 days (previously 21.9 days). Both observing times have increased because of the relaxed observing elevation constraint.

To be more concrete, we ran a scanning simulation with a mock hexagonal focal plane with a 35-degree FOV and 217 pixels on a regular grid. Here is the resulting hit map scaled to match 100Hz observations with 1000 pixels:

Total hits.png

The fsky numbers in the title correspond to raw, noise and signal-dominated versions of the same metric:


Due to the wide range of observing elevations, it is necessary to consider also the elevation-dependent noise in the detectors. We use the analytical fits to the 95GHz elevation profiles from S4 NET forecasts III and modulate the individual detector noise levels for each scan. The resulting depth map scaled to 1000 pixels is here:

Total wcov.png

Further work

Relaxing the observing elevation constraint by just 5 degrees allowed us to focus significantly more time over the year to observing the deep patches. With the elevation noise profiles we may ask what would happen to the map depth if we went even lower. In the extreme, we would point the bore sight so low that the bottom of the focal plane would be almost useless but we would still accumulate integration time in the deep field from the top of the focal plane.

We created another observing schedule with minimum observing elevation of just 35 degrees but with a progressively higher scheduling penalty at lower elevations. This means that only the Tier 1 tiles are ever targeted at low elevation and only when there are no Tier 1 tiles at higher elevation. The schedule now includes 101.2 days of the Southern deep field (instead of 77.7) and 43.4 days in the Northern deep field (instead of 31.8).

The resulting depth map is visually very similar to the earlier version:

Total wcov.low el test.png

but is 10% deeper in the deep field.