9C spectral-index distributions and source-count estimates from 15 to 93 GHz - a re-assessment

TL;DR

This paper re-evaluates earlier spectral-index and source-count estimates at high frequencies, correcting a significant measurement error at 43 GHz and discussing its impact on previous results and comparisons with other surveys.

Contribution

It identifies and corrects a major measurement error in VLA flux densities at 43 GHz, revising earlier source count predictions and providing a more accurate high-frequency source population estimate.

Findings

VLA 43 GHz flux densities were about 70% of their true values.

Revised source count estimates at 22-93 GHz are lower than previous predictions.

Comparison with South Pole Telescope data supports the revised estimates.

Abstract

In an earlier paper (2007), we used follow-up observations of a sample of sources from the 9C survey at 15.2 GHz to derive a set of spectral-index distributions up to a frequency of 90 GHz. These were based on simultaneous measurements made at 15.2 GHz with the Ryle telescope and at 22 and 43 GHz with the Karl G. Jansky Very Large Array (VLA). We used these distributions to make empirical estimates of source counts at 22, 30, 43, 70 and 90 GHz. In a later paper (2013), we took data at 15.7 GHz from the Arcminute Microkelvin Imager (AMI) and data at 93.2 GHz from the Combined Array for Research in Millimetre-wave Astronomy (CARMA) and estimated the source count at 93.2 GHz. In this paper, we re-examine the data used in both papers and now believe that the VLA flux densities we measured at 43 GHz were significantly in error, being on average only about 70 per cent of their correct values. Here, we present strong evidence for this conclusion and discuss the effect on the source-count estimates made in the 2007 paper. The source-count prediction in the 2013 paper is also revised. We make comparisons with spectral-index distributions and source counts from other telescopes, in particular with a recent deep 95 GHz source count measured by the South Pole Telescope. We investigate reasons for the problem of the low VLA 43-GHz values and find a number of possible contributory factors, but none is sufficient on its own to account for such a large deficit.