IV-G. Spikes

At the end, we had a nice catalog of >350,000 stars in M31, and >150,000 stars in M33. But, we found that there were some false detections of diffraction spike artifacts near bright stars. We needed to get rid of these to spare ourselves some embarassment later on. But, how?

Following up a suggestion by Knut, I found that the ellipticity of these false detections were high. Furthermore, the position angle of these "stars" had to always be oriented towards the bright stars that created the defraction spikes. The spikes were located at 45 degree angles.

So, we used the USNO-B1 to find all stars with an magnitude less than 14. (Access through Vizier). We downloaded ";" separated ASCI files centered on each of the nominal field centers (radius=50'). We had to do it that way due to a bug in Vizier. (Incidently, the survey extends 2.8 degrees in dec and 2.2 deg in RA....) I cat'd them all into a single file: m31usno14.dat. The program "viztoradec.f" translates this to something sensible--->m31usnord.cat

Now the fun begins: We use the script spike.cl (with the FORTRAN routine self.f) to do the hard work:

• First, make a list "allimages" of all of the stacked images (including the full path names), i.e., /Volumes/lg1/massey/m31/m31f1/stacked/M31F1m.fits
• Then we run "spike.cl". This produces a catalog called "m31spikes". We have visually checked these, and nearly all the false detections are gone, with very few real stars spuriously removed.
• Run "mrclean". This goes through the M31Mos.cat and produces M31Mosed.cat, with the spikes removed.

In the end, for M31 we eliminated 491 stars and kept 371,178. But, man, they were bugging me!