[daip] [!6263]: aips - SAD: Questions

[Eric Greisen]

Eric Greisen updated #6263
--------------------------

       Staff (Owner): Eric Greisen (was: -- Unassigned --)
                 Due: - Cleared - (was: 02 March 2015 08:45 PM)

SAD: Questions
--------------

           Ticket ID: 6263
                 URL: https://help.nrao.edu/staff/index.php?/Tickets/Ticket/View/6263
           Full Name: Robert H Gray
               Email: roberthansengray at gmail.com
             Creator: User
          Department: AIPS Data Processing
       Staff (Owner): Eric Greisen
                Type: Issue
              Status: Open
            Priority: Default
                 SLA: NRAO E2E
      Template Group: Default
             Created: 26 February 2015 08:45 PM
             Updated: 26 February 2015 09:40 PM
      Resolution Due: 06 March 2015 08:45 PM (7d 23h 5m)



1. The rms is normally in Jy/beam which I suppose could be printed with a metric scaling

2. The explain file is in error - the 1000000 scale mark is an exclamation point

3. Flux is the integral and peak is the FIT peak of the Gaussian.  It has been found that in 
resolved sources, the flux is a better measure of the total power of the source.  But
when a source is unresolved or barely resolved, the peak is a better measure of the total
power.

4. SAD searches down to a level such as your 0.218 Jy/beam but that does not mean that
what is fit will have a peak value > 0.218.  That will depend on the quality of the fit, whether 2 
Gaussians were fit in that island, etc.   I agree that 67 millis does seem low....

5. Your procedure stops on channel 2037 not 2036.  IMSTAT returns an error code if there
are no valid pixels.   Note that IMSTAT does a simple rms including all pixels.  If there is a
significant source this will be way higher than the actual off-source noise.  This may not
be a serious concern for you.  Task IMEAN gets a better rms but also sets an error code
when there are no valid pixels.

On your procedure, CUTOFF is not an adverb to SAD - CPARM(i) sets potentially multiple
cutoff levels, starting high and working down.  Does the rms vary a lot with channel?  You
could use IMEAN over the full cube to determine a robust RMS and put that in the header
or use it to control the CPARMS(s) and not have to do IMSTAT on each plane.  If the RMS
is not in the header, then SAD will find a robust RMS in each plane and use that to determine 
the error bars.

I expect for fully flagged channels you will just have to run your loop over multiple ranges
 of channel.

Eric Greisen




------------------------------------------------------
Staff CP:  https://help.nrao.edu/staff