This chapter focuses on ways to do the initial calibration of interferometric fringe-visibility data in . The sections which follow concentrate primarily on continuum calibration for connected-element interferometers, especially the VLA. However, the information in these sections is useful to spectral-line, solar, and VLBI observers as well. For specific advice on the new EVLA, consult Appendix E. For additional advice on spectral-line calibration, see §4.7; for advice on calibrating observations of the Sun, see §4.8; and for the gory details of VLBI, read Chapter 9. After the initial calibration has been completed, data for sources with good signal-to-noise are often taken through a number of cycles of imaging with self-calibration. See §5.4 for information on these later stages of the reduction process. For accurate calibration, you must have accurate a priori positions and structural information for all your calibration sources and accurate flux densities for at least one of them. It is best if the calibration sources are unresolved “point” sources, but it is not required.
For the basic calibrations, visibility (“uv”) data are kept in “multi-source data sets,” each of which contains, in time order, visibility data for one or more “unknown” sources and one or more calibration sources. Associated with these data are “extension” files containing tables describing these data. When VLA archive data are first read into a number of basic tables are created and filled with information describing the data set. These are
An initial CL table contains gains due to known antenna functions of elevation and measured atmospheric opacities. VLBI, and especially VLBA, data sets will end up with even more table files. Calibration and editing tasks then create, as needed, other tables including
All of these tables can be written to, and read back from, FITS files along with the visibility data. These, and any other, tables can be manipulated and examined using the general tasks PRTAB, TACOP, TABED, TAMRG, TASRT, TAFLG and TAPPE.
The visibility data within the multi-source data set are not normally altered by the calibration tasks. Instead, these tasks manipulate the tabular information to describe the calibration corrections to be applied to the data and any flagging (deletion) of the data.
The programs discussed in this chapter are part of a package that has been developed to calibrate interferometer data from a wide range of connected-element and VLB arrays, especially the VLA and VLBA. These programs therefore support many functions (and inputs) that are not required when calibrating normal VLA data. The examples given below show only the essential parameters for the operation being described, but, to get the results described, it is essential that you check all the input parameters before running any task. Remember that adverbs are global and will be “remembered” as you proceed. A list of calibration-related symbols is given in §13.6, but a possibly more up-to-date list can be obtained by typing ABOUT CALIBRAT in your session. More general information on calibration can be routed to your printer by typing DOCRT FALSE ; EXPLAIN CALIBRAT C R, while deeper information on a specific task is obtained with EXPLAIN taskname C R.
When you are satisfied with the calibration and editing (or are simply exhausted), the task SPLIT is used to apply the calibration and editing tables and to write uv files, each containing the data for only one source. These “single-source” uv files are used by imaging and deconvolution tasks that work with only one source at a time. Many of the tasks described in this chapter will also work on single-source files. For VLA calibration, there are several useful procedures described in this chapter and contained in the RUN file called VLAPROCS. Each of these procedures has an associated HELP file and inputs. Before any of these procedures can be used, this RUN file must be invoked with:
There is a “pipeline” procedure designed to do a preliminary calibration and imaging of ordinary VLA data sets. This provides a good first look at the data. Nonetheless, the results are still not likely to be of publishable quality. To run the pipeline, enter
> VLARUN C R
to execute the pipeline.