3.3 Special features of synthesis imaging

The techniques of synthesis and direct imaging are obviously very different, but many aspects of this are not directly relevant for the observer, especially if he or she is satisfied with the image produced automatically by standard MERLIN data analysis procedures. Nevertheless there are a number of points which are useful to bear in mind when planning or interpreting observations.

Several features of synthesis imaging make life substantially easier for the observer. For instance, polarimetry is provided by default, with no loss in throughput for simple continuum observations. The same applies to spectroscopy except that the range of frequencies covered (the bandwidth) must be reduced to achieve high spectral resolution. At GHz frequencies, the standard MERLIN calibration procedure automatically puts the images onto the ICRF (International Celestial Reference Frame, Ma et al. 1998) with an accuracy of between 10 and 50 milliarcsec. Under most circumstances the flux scale is accurate to about 2%. At 151 and 408 MHz absolute astrometry is rarely possible, and flux density calibration is poor at 151 MHz.

A crucial feature of MERLIN is its high resolution, with a natural beam size from 10 mas at K-band to 550 mas at P-band. High resolution has its downside in the form of reduced surface brightness sensitivity, as discussed in §3.5.

A unique limitation of synthesis images is that the accuracy of the results depends on the angular scale of the structure being imaged. A relatively compact object, or several compact objects separated by large angles, can be imaged with high accuracy, but as the object size increases the image becomes increasingly impressionistic. Very large-scale smooth emission will not be represented at all, even if it should be well above the surface-brightness threshold. These points are discussed in §3.7.