How to optimize single-dish observations?

One of the main difficulty of the short-spacing problematic is the need of observations from a single-dish telescope at least as big as the interferometer antennas¹⁰. In this respect, the IRAM-30m and NOEMA are very complementary. Nevertheless, when observing with the single-dish telescope, a few precautions are needed to avoid contaminating the interferometric data with possible artifacts of single-dish data.

• The field-of-view of the single-dish map must be twice the field-of-view covered by the mosaic. The only exception to this rule happens when the source intensity decreases to zero in a smaller field-of-view. Indeed, there is no point in observing an empty sky.
• The observing strategy must enforce Nyquist sampling (or better) of the source at the resolution of the single-dish telescope.
• A particular care should be taken of the pointing, tracking and amplitude calibration and baseline removal as those are critical issues in obtaining a high quality single-dish map to produce short-spacing information. For instance, data with too large tracking errors should be discarded.
• Among “baseline” issues, the presence of continuum sources is to be treated with care. Continuum is difficult to measure with single-dish telescopes, and a (linear or polynomial) spectral baseline is often fitted to avoid atmospheric contamination. In such cases, the combination should be made with interferometer data where the continuum has been removed, and added back later...
• We advise to make many On-The-Fly coverages of the observed field-of-view to get homogeneous observing conditions. Scanning in perpendicular directions is needed to decrease stripping.

Sometimes, single-dish telescope time is scarce and some of the above criteria can not be fulfilled. In those cases, you can still try to use your single-dish observations and our algorithm will try to make its best to get a sensible result. However, any artifact in the combination may directly come from wrong single-dish observations. In other words, do not blame the software unless you are sure of the quality of your single-dish (and interferometric) observations...