3.11 Spectral line mapping

The principles of post-correlation spectral line data reduction are no different from multi-channel continuum processing. If the target is to be self-calibrated, a single channel (or a few adjacent channels) are used and the solutions are applied to all channels. In some situations additional techniques are needed:

(i)

To allow more than one line or a broad line to be observed in a single run;

(ii)

To compensate for the increase in noise in narrow channels;

(iii)

When continuum subtraction is required. %endenumerate

The present 15 MHz maximum bandwidth, or more often a narrower range imposed by the required spectral resolution, restricts the maximum instantaneous velocity range which can be observed. However in case (i) frequency switching on a cycle of a few minutes allows, for example, both OH maser mainlines at 1665 and 1667 MHz to be imaged in a single run. This is more efficient than spending the same amount of time on target in separate observations as the phase-cal is observed at a single frequency using the technique described for case (ii) to compensate for frequency-dependent phase offsets.

In case (ii), if a very narrow bandwidth, e.g. 0.125 MHz in 1 kHz channels, is used for the target, most point-cals are too weak in a single channel to be used as a bp-cal. Appendix C indicates those which are suitable. The bp-cal is used to establish the initial flux scale channel by channel using dprograms as well as to provide phase and amplitude bandpass calibration (usually done in AIPS). However if the target is relatively weak (less than a few Jy) the most accurate flux scale will be established by also observing a point-cal of similar flux density and using this to refine the flux scale e.g. in AIPS.

In addition most phase-cals do not provide adequate signal-to-noise for good solutions in a few MHz or less, and in such cases they are observed in the full 16-MHz bandwidth. There can be a slight instrumental phase offset between wide and narrow band data, or between nearby separate narrow bands as in case (i), and corrections for this are derived using observations of the bp-cal in all configurations.

Case (iii) covers observations of absorption or emission lines in the presense of continuum emission with a high enough surface brightness to be imaged by MERLIN^3.4 The experiment must be designed so that on either side of the line emission there are sufficient continuum-only channels which will give a high enough signal to noise ratio (when added together) to subtract the continuum accurately.

The general procedures are similar to §3.8.8 but the bp-cal and the target and any other sources to be processed in line mode must be carefully inspected for interference at an early stage in data processing. §5.1.9 gives more details of line data processing.

Footnotes

... MERLIN^3.4

If there is bright extended emission resolved-out by MERLIN e.g. in star-forming regions it may be necessary to allow for this in the initial stages of data processing (§5.1.5), and it is even possible to add data from smaller arrays to improve the chances of detecting compact absorption against a diffuse continuum!