LINMOS using CAR

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

Moderator: Mark.Wieringa

Post Reply
naomimg
Posts: 3
Joined: Fri Jul 19, 2013 11:04 am

LINMOS using CAR

Post by naomimg »

I am having enormous trouble linmos-ing many pointings together and think there must be an odd quirk in the code.

Here's the situation: I am trying to mosaic together a large number of pointings (~950) using limos. I have several different types of images that I am working with, Stokes Q 8 MHz channel images, and Stokes I MFS (128 MHz) images. For each pointing I have imaged the two products very similarly, using a common offset position for each pointing and options=mosaic in invert. Examples of the invert history and resultant headers for the two types are given below. I then proceed to clean each pointing and each channel and then mosaic with linmos. This is where the trouble starts.

For the Stokes I MFS image linmos works beautifully and I end up with a very nice image with the following header:

prthd: Revision 1.10, 2012/03/02 06:13:02 UTC

****************************************************************
Filename: GC.mfs.icln.linmos/
Telescope: ATCA
Object: gcl_001 Observer: nmm-g
Image type: intensity
Map flux units: JY/BEAM
Beam Size: 61.65 by 22.19 arcsec.
Position angle: -3.9 degrees.
This image has 4 axes.
----------------------------------------------------------------
Type Pixels Coord Value at Pixel Coord Incr Units
RA---NCP 13104 17:45:37.200 6589.00 -4.000000E+00 arcsec
DEC--NCP 13404 -28:56:10.200 7287.00 4.000000E+00 arcsec
FREQ-LSR 1 1.380904E+00 1.00 8.000000E-03 GHz
STOKES 1 I
----------------------------------------------------------------
Average time of observation: 03SEP24:21:37:31.4
Equinox: J2000.0
Doppler reference frame: TOPOCENT
Observatory radial velocity: -21.30 km/s
Primary beam type: SINGLE
Number of iterations: 500
Mask item is present ... some data are blanked
History item is present
Mosaicing information table is present

For the Stokes Q images, linmos produces an image in a CAR projection, which has clearly caused problems, sources aren't at their correct positions and the edges of the mosaic are chopped off. The header for the image is:
prthd: Revision 1.10, 2012/03/02 06:13:02 UTC

****************************************************************
Filename: GC.poli.linmos/
Telescope: ATCA
Object: gcl_001 Observer: nmm-g
Image type: intensity
Maximum: 1.05939552E-01 Minimum: -1.28651503E-03 JY/BEAM
Beam Size: 61.65 by 20.85 arcsec.
Position angle: -2.6 degrees.
This image has 4 axes.
----------------------------------------------------------------
Type Pixels Coord Value at Pixel Coord Incr Units
RA---CAR 13102 17:45:37.200 6588.00 -4.000000E+00 arcsec
DEC--CAR 12534 -28:56:10.200 6418.00 4.000000E+00 arcsec
FREQ-LSR 12 1.432000E+00 1.00 8.000000E-03 GHz
STOKES 1 Q
----------------------------------------------------------------
Average time of observation: 03SEP13:22:54:18.4
Equinox: J2000.0
Doppler reference frame: TOPOCENT
Observatory radial velocity: -20.46 km/s
Primary beam type: SINGLE
Number of iterations: 136
Mask item is present ... some data are blanked
History item is present
Mosaicing information table is present

As you can see the Declination direction of the Stokes Q image is much smaller than the Stokes I MFS image.

Does anyone have any idea what's going on here? I like the Stokes I image, but I need the Stokes Q and U images!

Thanks, Naomi



History for gcl_001.mfs.icln
INVERT: Miriad invert: Revision 1.20, 2014/03/03 04:49:43 UTC
INVERT: Executed on: 2015-06-02T17:08:05.0
INVERT: Command line inputs follow:
INVERT: map=gcl_001.mfs.imap
INVERT: beam=gcl_001.mfs.beam
INVERT: cell=4
INVERT: vis=gcl_001.1384
INVERT: offset=17:45:37.20,-28:56:10.2
INVERT: robust=2
INVERT: stokes=i
INVERT: fwhm=10
INVERT: options=double,mfs,mosaic
INVERT: select=-uvrange(0.01,1.8)
INVERT: imsize=3,3,beam

Resulting in the following cleaned image header:
prthd: Revision 1.10, 2012/03/02 06:13:02 UTC
****************************************************************
Filename: gcl/gcl_001.mfs.icln/
Telescope: ATCA
Object: gcl_001 Observer: nmm-g
Image type: intensity
Map flux units: JY/BEAM
Beam Size: 61.65 by 22.19 arcsec.
Position angle: -3.9 degrees.
This image has 4 axes.
----------------------------------------------------------------
Type Pixels Coord Value at Pixel Coord Incr Units
RA---NCP 1578 17:45:37.200 1062.00 -4.000000E+00 arcsec
DEC--NCP 1578 -28:56:10.200 1860.00 4.000000E+00 arcsec
FREQ-LSR 1 1.381146E+00 1.00 8.000000E-03 GHz
STOKES 1 I
----------------------------------------------------------------
Average time of observation: 03SEP24:21:37:31.4
Equinox: J2000.0
Doppler reference frame: TOPOCENT
Observatory radial velocity: -21.30 km/s
Nominal Theoretical Rms: 5.939E-04
Number of iterations: 500
History item is present
Mosaicing information table is present

History for GCL_001.c01.qcln:
INVERT: Miriad invert: Revision 1.20, 2014/03/03 04:49:43 UTC
INVERT: Executed on: 2015-06-02T17:08:13.0
INVERT: Command line inputs follow:
INVERT: map=gcl_001.c01.imap,gcl_001.c01.qmap,gcl_001.c01.umap,gcl_001.c01.vmap
INVERT: beam=gcl_001.c01.beam
INVERT: cell=-4.0
INVERT: vis=gcl_001.1384
INVERT: offset=17:45:37.20,-28:56:10.2
INVERT: robust=2
INVERT: line=channel,1,1
INVERT: stokes=i,q,u,v
INVERT: fwhm=10
INVERT: options=double,mosaic
INVERT: select=-uvrange(0.01,1.8)
INVERT: imsize=1578

resulting in the following cleaned image header
prthd: Revision 1.10, 2012/03/02 06:13:02 UTC

****************************************************************
Filename: gcl/gcl_001.c01.qcln/
Telescope: ATCA
Object: gcl_001 Observer: nmm-g
Image type: intensity
Map flux units: JY/BEAM
Beam Size: 61.65 by 20.85 arcsec.
Position angle: -2.6 degrees.
This image has 4 axes.
----------------------------------------------------------------
Type Pixels Coord Value at Pixel Coord Incr Units
RA---NCP 1576 17:45:37.200 1061.00 -4.000000E+00 arcsec
DEC--NCP 1576 -28:56:10.200 1859.00 4.000000E+00 arcsec
FREQ-LSR 1 1.432000E+00 1.00 8.000000E-03 GHz
STOKES 1 Q
----------------------------------------------------------------
Average time of observation: 03SEP13:22:54:18.4
Equinox: J2000.0
Doppler reference frame: TOPOCENT
Observatory radial velocity: -20.46 km/s
Nominal Theoretical Rms: 2.196E-03
Number of iterations: 136
History item is present
Mosaicing information table is present

Note: I actually requested 1578 pixels in the invert step, but I am consistently receiving 2 less in each dimension. Another quirk?
naomimg
Posts: 3
Joined: Fri Jul 19, 2013 11:04 am

Re: LINMOS using CAR

Post by naomimg »

Follow-up. I think I have traced the problem to one pointing with a mis-matched header which forces linmos to use CAR projection. Thank goodness for well-commented miriad fortran code.
Mark.Wieringa
ATCA Expert
Posts: 297
Joined: Mon Feb 08, 2010 1:37 pm

Re: LINMOS using CAR

Post by Mark.Wieringa »

Hi Naomi,

it sounds like you solved your main problem.

To answer your other question: the image sizes are adjusted slightly (1-2 pixels) when mosaicing due to the shift to a new reference pixel.

Cheers,

Mark
Post Reply