[daip] Flux Question

[Eric Greisen]

Cara Battersby wrote:
> HI there,
> 
> I've been getting some final numbers from VLA data that I reduced awhile 
> ago (Project AJ345), and have come up with some puzzling questions about 
> the flux.  I have been getting differences for point source fluxes of 
> the order of a factor of two.  My first method was somewhat standard 
> (simple uv clipping of very high points, simple cleaning) and achieved a 
> point source flux (for the central source) of about 0.35 mJy/beam.  My 
> friend, who is trying to learn CASA, also reduced the data and achieved 
> a point source flux (for the central source) of about 0.2 mJy/beam.  The 
> main qualitative difference in the images are the large scale ripples in 
> my original map.  So, I went back and removed the short UV spacings 
> (though there ARE extended sources in the map, so I don't believe that 
> this is a real solution).  This achieved a point source flux (for the 
> central soruce) of about 0.45 mJy/beam.  Both this map, and the CASA map 
> appear qualitatively better, due to the lack of ripples.  However, I 
> don't want to just exclude data (because there IS large scale 
> structure), but I am very confused about the source of the ripples, and 
> more importantly, the wildly different point source fluxes that these 
> different methods achieve.
> 
> Do you have any idea why this might be happening, or have a suggestion 
> for something else to try?
>

The large scale ripples suggest bad data at short spacings.  There are 
ways to find that - UVFND or FFT the clean images and look for seriously 
high values.

The MS clean options in IMAGR may let you deal better with large scale 
structure but that will not solve bad data.  Of course, your data set 
may simply not have sampled the large-scale structure enough to image
it.

Let me know how this goes.  Try measuring the flux with IMFIT including 
a DC bias "component".

Eric Greisen