Wide-Field Imaging and Spectral Index Correction

Got an image problem? Let us help!

Moderator: Mark.Wieringa

Post Reply
julie
Posts: 12
Joined: Tue Mar 22, 2011 2:53 pm

Wide-Field Imaging and Spectral Index Correction

Post by julie »

Hi All,

I have been reading over the forum and found a few interesting posts regarding the correction for the spectral index and primary beam over the 2GHz band.

Just to sum up and make sure I understand:

When creating a wide-field image with the 2GHz band one would follow these steps:

(0) Data has been cleaned and self-calibrated across the full band if required.
(1) Split your band up into subbands, i.e. make the data narrow band instead of wide band.
(2) Image the subbands separately and correct for the different resolution (convolving or using different robust parameter value) across the band.
(4) Use the task MFSPIN to determine the spectral index in each subband.
(5) Correct each subband image for the spectral index, i.e. adjust the flux to the centre of the subband.
(6) Put all subband images into LINMOS and correct for the primary beam.

Does this look reasonable?

Condon+ (2012) (http://adsabs.harvard.edu/abs/2012ApJ...758...23C) do something similar with EVLA data. They have one pointing in the 2 - 4 GHz band. Using equations 8 and 9 they correct for the spectral index and primary beam across the field of view. They use an average spectral index of -0.7 instead of using a result like what MIRIAD gives for MFCLEAN and MFSPIN.

Anyone have any thoughts or additional comments on wide-band wide-field imaging with CABB?

Cheers,
Julie
julie
Posts: 12
Joined: Tue Mar 22, 2011 2:53 pm

Re: Wide-Field Imaging and Spectral Index Correction

Post by julie »

Previous posts regarding primary beam correction and spectral index correction.

Nick Seymour: http://atcaforum.atnf.csiro.au/viewtopic.php?f=10&t=102
Emil Lenc: http://atcaforum.atnf.csiro.au/viewtopic.php?f=9&t=54
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: Wide-Field Imaging and Spectral Index Correction

Post by Mark.Wieringa »

Hi Julie,

I'm not sure how you would do step 4 and 5.
In step 4 mfspin would not give a very accurate result over a narrow subband, and the result it gives is an image of the spectral index, not a single value.
How would you use this in step 5?

If the subbands are narrow enough, the average flux should be close to the flux at the central frequency of the subband. If there are a lot of flagged channels in a subband you'd need to use the mean frequency instead of the centre frequency. I think invert already does this with the mfs option.

The Condon et al scheme tries to optimize the signal to noise for sources of a particular (average) spectral index. Skipping step 4 & 5 would give the optimum for alpha=0.
As Condon et al noted, the effective frequency of the image after correction for primary beam will vary (decrease) with distance from the field centre.

I'll have another look to see if we can use the spectral index plane mfclean produces to do a better job at wide band primary beam correction.

Cheers,

Mark
Post Reply