16cm Primary Beam

Is MIRIAD being a pain? Let us know your experience.

Moderator: Mark.Wieringa

Post Reply
and460
Posts: 79
Joined: Thu Sep 01, 2011 12:23 pm

16cm Primary Beam

Post by and460 »

Hi guys,

I believe that the ATCA primary beam has been newly characterized for 1-- 3 GHz, after reading about an effort to do so on this page: http://www.narrabri.atnf.csiro.au/peopl ... _16cm.html. Just wondering - is this new beam model available in Miriad to tasks correcting for the PB (e.g. linmos)? If so, is it available by default after doing a mirsync, or does it have to be set up somehow?

I ask, because some Stokes I spectra that I have been creating from a set of mosaiced observations seem to be coming out significantly flatter than I expected they would. After some playing around, I ended up finding some unresolved sources present in multiple mosaic pointings, imaging these pointings and then PB-correcting them individually at 2 GHz. The peak flux density of the sources in the different images becomes quite a bit larger the further the source is from the PB centre, making me think that the old PB model is underestimating the true PB at this freq, and is therefore over-correcting. Does this sound like a reasonable conclusion?

Cheers,
Craig.
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: 16cm Primary Beam

Post by Mark.Wieringa »

Hi Craig,

yes, there are new 16cm beam fits available in miriad, I derived these from Jamie's measurements, but refitted them with the 4th order inverse polynomial in 256 MHz intervals (from the 6 order poly), ignoring the orientation effect. They are valid in the 0-50 arcmin*GHz range. To use these fits, you currently have to change the 'telescop' keyword from 'ATCA' to 'ATCA16'. For a single pointing this can be done with puthd, for mosaics you need to do this in the uv data, with uvputhd, before running invert. I was holding off making the new fits the default until some people had actually tried them and compared results.

To answer your second question: yes, the average beam over the 1-3GHz range is wider than the previous 20 and 13 cm beam models, so using the old beam model will overestimate the correction for sources far from the field centre. The new model will interpolate between 7 frequency ranges, so it should get close to the correct number over the entire 16cm band.
A few caveats:
  • there is not really a 'null' in the beam, especially not for wide bandwidths;
    the new beam fits are inaccurate past the first 'null', they keep decreasing whereas the real beam has sidelobes;
    beyond the first 'null' the beam is significantly non-circular and sources will vary in flux with time. Peeling or 2D beam models would be needed to remove the sidelobes of distant sources.
Cheers,

Mark
Post Reply