-1 0 1 2 3 4 5 6 7 8 9 10 11

GALEX Diffuse UV Background

3. DIS-08 : a Target with Lots of Dust-Scattered Starlight:

I remind you that some of the images we will look at do require a very large screen to be seen properly.
 3.01
First, let's look at the FUV pipeline diffuse background for DIS.08:
 3.02
What you saw is a smooth image with a very white broad arc in the lower right. "White" means "bright in the UV." Anyway, that arc is dust-scattered ultraviolet starlight, as we can verify by looking at the dust map we created for this target from the Schlegel et al. data.
 3.03
The array of numbers in the dust map is E(B-V); that is, for example, the maximum E(B-V) is 0.04908 which shows in the plot as 490 (red in lower right). Notice the faint gray circle near that 490. For most of our work with the GALEX images, we exclude the part of the image outside this circle, as contaminated with instrumental scattered light. (We will illustrate this in detail, later.)
 3.04
Now let us look at the same dust map, but with a much wider field, to see the context. (Look in the dust map for tiny blue circles that represent TD1 UV-bright stars.)
 3.05
The orientation is now in galactic coordinates, and the target is labelled "dis.08." (Later, we will provide a list of all the Deep Imaging Survey targets, with identification; for now, we focus on this one target.) You will see that several other targets appear in our map, including dis.09 which happens to be the second DIS target that we analyze as our "second example," in the next chapter. Recall that IT is selected because it is almost totally FREE of dust-scattered starlight! Yet look where it is located!
 3.06
By using Mathematica with the Schlegel et al. data, we can get a "3D" view of the dust distribution near dis.08, which is marked with a blue ring.
 3.07
The time has now come to look at our Mathematica GALEX images for DIS-08. Here are the pipeline backgrounds for
FUV    and      NUV    (no dust-scattered starlight removed)
Here are the same two, side by side for comparison. The thing to note is the EXCELLENT correlation between the two images. It is very clear that there is a strong component of dust-scattered starlight present in each. Now, let's look at the correlations with E(B-V)!
 3.08
FUV correlation with E(B-V)
NUV correlation with E(B-V)
Great correlations, but do notice, however, the large scatter (much bigger than the statistical errors).
 3.09
Now let's try our hand at removing the dust-scattered starlight from the images, using our correlations with E(B-V):
FUV    and      NUV    (dust-scattered starlight removed)
Here are the same two, side by side for comparison.
 3.10
What is the bottom line? While we have clearly removed SOME of the dust-scattered starlight, it is also quite clear that we have not removed it all! In fact, there seems to be quite a lot left. So what should we conclude? Our main interest is not in the (relatively boring) starlight scattered from dust. Our main interest is in the mysterious residual! But we have now learned that we cannot reliably identify this residual in targets that have a substantial component of dust-scattered starlight. So our conclusion is simple: ignore such targets!
 3.11
Sorry to have led you through such a long journey, just to come to the conclusion that such targets are pretty worthless! But before we leave, and go on to a target that is extremely valuable, let us look at the correlations between FUV and NUV, before and after the removal of dust:
      NO dust-scattered light removed
      Now, the dust-scattered starlight has been removed! You are not surprised either to see the weaker correlation after dust removal, or the strength of the remaining correlation.

Now we turn to DIS-09, for which there is essentially NO dust-scattered starlight contribution to the images!

-1 0 1 2 3 4 5 6 7 8 9 10 11