Visibility Processing

• Flagging: Flagging differs only in that, when the instrument is continually changing to a new pointing centre, it can be more difficult to spot the outliers which indicate bad data.
• XY Phase Correction and Splitting: XY phase correction is identical with a conventional reduction. Again, as with conventional reduction, for calibration and imaging purposes, it is easiest in MIRIAD \ to split the data into datasets containing a `single source' and single frequency band. For a mosaic experiment, all the pointings from the object of interest should be considered a `single source'. As the ATCA on-line system labels each pointing with a different field name, uvsplit would normally break the dataset into one file per pointing. To avoid this, use uvsplit 's mosaic option. This causes uvsplit to use the source name up to any underscore character. Thus, assuming you have used the field naming convention of Section 20.3, uvsplit

  will not split apart all the separate pointings.

  If you failed to follow the naming convention, the select keyword of uvaver and uvsplit can be used to split the dataset up. This is left as an exercise for the reader.

  For example, assuming a dataset containing calibrators and an observation of the LMC (with appropriately named fields), the following inputs will generate datasets for the calibrators and a single dataset for all the LMC data.

  

• Data Selection: With mosaic visibility datasets, the select=ra and select=dec selection sub-commands can be useful to extract or manipulate a subset of the visibility data.
• Calibration: Determining the calibration solutions does not differ from that described in Chapter 11. You will then use gpcopy

  to copy the calibration tables to the multi-pointing dataset.