Large Mosaics: the SAULT method

The default Mosaic processing uses the GUETH method described in Gueth et al. (1995). The processing involves imaging each field on the same grid as the full mosaic. This has two drawbacks for large mosaics: 1) it requires lot of memory 2) beam smearing effect at edges may become significant.

To circumvent this, one can use the SAULT method (Sault et al. (1996)), where only a limited area is imaged around each field, with a phase center near the pointing direction, and a correction to handle the proper common projection center. They are then added at the appropriate place to the larger Mosaic. This method limits beam-smearing effects to those of a single field. The method is faster and uses less memory than the GUETH method for mosaics larger than about 16 fields.

In IMAGER, this mode can be activated by using command MOSAIC. It applies to the UV_MAP and UV_RESTORE commands. Deconvolution is identical in both methods.⁸