Background Tests
By Elijah Bernstein-Cooper, July 6, 2015, 0 comments.

These results are completed after reorganizing the parameter estimation code. We should be more confident in these results than previous ones, seeing as each step of the new code was modularized and tested to some extent.

No backgrounds were subtracted in the following results. The data used were Planck $A_V$ derived from the $\tau_{353}$ map.

The following images show the masking progression for each cloud.

## Likelihoods

The following are the resulting likelihood spaces from a course grid search. The values used in the grid search are provided below:

Start End $\Delta$
$HI$ Width [km/s] 1 75 0.32
DGR [$10^{-20}$ cm$^2$ mag] 0.001 0.4 2e-3
Intercept [mag] -2 2 0.1

### Taurus

The $HI$ width is now favored to be much larger than before. 20 km/s now compared to 10 km/s previously. The DGR is similar to previous runs.

### Perseus

Perseus has a much larger $HI$ width than was derived before, and favors a 0 intercept. Judging by the Planck $A_V$ image Perseus should have a background of around $0.6$ mag. The calculation favors a higher $HI$ width and a lower intercept. This means that the $HI$ emission correlates better with the surrounding $A_V$ than a constant offset.

### California

The intercept is now close to $0$ mag. However the $HI$ width favored is quite high. A width this large essentially means including the entire line of sight.