Negative nanomaggies? ---> Cannot convert to AB mag

Question

I'm downloading some images from the legacy survey and the units are a bit challenging to understand. I've already figured out how to get my fits and header, but I cannot go from nanomaggies (the units it comes in) to AB mag using:

AB mag = 22.5 - 2.5*log10(nanomaggy)

Because some of the counts in the downloaded fits are negative. I'm assuming there was some sort of background subtraction done, but I do not see any zeropoint in the header or anything that can help me make the entire data non-negative, such that I can convert the units from nanomaggies to magnitude if the physics units are negative.

Answer 1

Could you please give a specific example?  In particular, are these negative values in pixels or in a catalog value?

If this is in pixels, are you trying to convert individual pixels to magnitudes, or are you integrating over an aperture, & the total flux is still negative?

Comments

This is in pixels and I am trying to convert individual pixels to another unit. I'm assuming I am missing some piece of logic to go from nanomaggies to magnitude

---

What is the specific use case for converting pixels to magnitude units?  That conversion is not particularly meaningful for individual pixels.  If you are trying to measure the flux of objects in an aperture, you should sum the pixels in flux units over the aperture first, then convert to magnitude units.

Answer 2

If you look at the description pages (eg https://www.legacysurvey.org/dr10/description/) you will find the following paragraphs:

The Community Pipeline removes a sky level that includes a sky pattern, an illumination correction, and a single, scaled fringe pattern. These steps are described on the NOIRLab Community Pipeline page. These corrections are intended to make the sky level in the processed images near zero, and to remove most pattern artifacts. A constant sky level, that is the mean of what was removed, is then added back to the image.

The brightnesses of objects are all stored as linear fluxes in units of nanomaggies. The conversion from linear fluxes to magnitudes is m=22.5−2.5log10(flux)�=22.5−2.5log10⁡(flux). These linear fluxes are well-defined even at the faint end, and the errors on the linear fluxes should be very close to a normal distribution. The fluxes can be negative for faint objects, and indeed we expect many such cases for the faintest objects.

Do these answer your question?