Radio continuum surveys: lessons learned

Abstract

While making and using radio continuum surveys, radio astronomers have had many learning experiences, as in “Oh, no! Not another learning experience!” They include problems with confusion from multiple sources in each beam area, insensitivity to low-brightness sources, correcting source counts for partially resolved or missing sources, and ignoring systematic errors, particularly the consequences of limited dynamic range. The principal lessons that I have learned from these experiences are: (1) The universe is not a vacuum. The design of surveys (and survey telescopes) should take into account known source properties, not just purely instrumental parameters such as “survey speed.” (2) The radiometer equation and Gaussian statistics are insensitive to the most insidious problems that plague radio continuum surveys. More sophisticated mathematical analyses are needed to optimize telescope and survey design. (3) Never forget the difference between flux density (Jy) and peak flux density (Jy beam−1 or K), which is a surface brightness. Ideally a survey is complete to some flux density (a source property), but all survey images are sensitivity limited by surface brightness (Jy beam−1), which depends on the instrumental beam solid angle. (4) Resolution corrections to counts of faint sources are so large and poorly understood that they limit the usefulness of sub-mJy surveys. These lessons are applied to design a possible Jansky Very Large Array (VLA) “wedding cake” survey that could provide a strong extinction-free constraint on the cosmological evolution of the star formation rate over the redshift range 0≤ z . 3 in which most stars were formed.