Usage

General usage of WSClean

WSClean is a command line program. The "wsclean" executable accepts many parameters. The "wsclean" program can be run without parameters to get a list of allowed parameters together with a short description. The general syntax of wsclean is as follows:

wsclean [-options] <obs1.ms> [<obs2.ms> ..]

Here's an example of a typical MWA run:

# Make a Stokes I image of size 3072x3072 with 0.7' pixels, do 10000 iterations,
# use CS cleaning and stop when the peak flux has reached 3σ level.
wsclean -size 3072 3072 -scale 0.7amin -niter 10000 \
  -mgain 0.8 -auto-threshold 3 obs.ms

This performs a Cotton-Schwab clean. The Cotton-Schwab algorithm is enabled with the "-mgain 0.8" parameter, which means that the peak flux is reduced by 80% until a new major iteration is started. Multiple measurement sets can be specified on the command line to image the integration of those observations.

For fast imaging with somewhat less accuracy, you can perform a Högbom clean and disable padding:

# Similar to above statement, but now only Högbom cleaning and no padding
wsclean -size 3072 3072 -scale 0.7amin -niter 10000 \
  -auto-threshold 3 -padding 1 obs.ms

Because the 'mgain' parameter was left out, WSClean will not iteratively go back to the visibilities. This also implies that the MODEL_DATA column will not be filled (see self-calibration with WSClean for more info).

A description of the basic cleaning parameters is given on the manual page for basic cleaning.

An advanced MWA example

As a more enhanced example, Natasha Hurley-Walker has been imaging GLEAM data with the following command:

wsclean -name ${obsname} -size 4000 4000 -niter 40000 -threshold 0.3 -pol xx,yy \
  -weight briggs -1.0 -scale 0.0125 -absmem 128 -join-polarizations \
  -j 12 -mgain 0.95 1068210256.ms

A little explanation of this command:

• Briggs' weighting with robustness of -1 is used. For the MWA, this decreases the noise in single snapshots slightly. See image weighting for more info on supported weightings.
• The (instrumental) XX and YY polarizations are imaged separately and cleaned together. This allows more accurate primary beam correction for the MWA. See polarized cleaning for more info.
• A larger mgain value is used because the MWA synthesized beam is well behaved.
• A larger image is made because GLEAM includes lower frequencies, at which the primary beam is larger.
• For GLEAM, the W-snapshot algorithm is used by executing the chgcentre command prior to imaging, as described on the W-snapshot page.

Next chapter: basic cleaning