NIKA2 Cosmological Legacy Survey. First measurement of the confusion noise at the IRAM 30 m telescope

TL;DR

This study directly measures the confusion noise for the IRAM 30 m telescope in a deep GOODS-North field using the dual-band NIKA2 camera. A novel cross-variance estimator leverages independent maps across 1.2 mm and 2 mm to extract both solo and cross-band confusion while accounting for the map-transfer function via end-to-end SIDES-UCHUU simulations. The reported confusion levels are consistent with SIDES predictions, with $139.1^{+15.9}_{-19.2} ext{(stat)} \,\pm 11.9 ext{(cosmic variance)}\,\mu$Jy/beam at 1.2 mm, $38.6^{+9.6}_{-13.1}\text{(stat)}\,\pm 3.7\text{(cosmic variance)}\,\mu$Jy/beam at 2 mm, and a cross-band value of $49.6^{+15.9}_{-24.8}\text{(stat)}\,\pm 6.4\text{(cosmic variance)}\,\mu$Jy/beam. The methodology robustly separates instrumental/processing noise from genuine sky fluctuations and highlights how the faint source population just below detection limits shapes the confusion in deep fields. This work strengthens confidence in SIDES while providing a practical framework for confusion characterization in upcoming mm surveys.

Abstract

The NIKA2 Cosmological Legacy Survey (N2CLS) is a large programme using the NIKA2 dual-band camera on the IRAM 30\,m telescope. Its goal is to improve our understanding of the physics of distant Dusty Star Forming Galaxies (DSFGs) by carrying out deep surveys of two fields, GOODS-North and COSMOS. This work is focussed on GOODS-North, which was observed for 78.2 hours, simultaneously at 1.2 and 2\,mm, with a field of view of $\sim$240\,arcmin$^2$. With such a deep integration, we were able to measure, for the first time, the confusion noise limits at the 30\,m telescope using the best sampled $\sim 62$\,arcmin$^2$ and masking sources with a flux greater than 0.54 or 0.17\,mJy at 1.2 or 2\,mm, respectively. We found a confusion noise of $139.1^{+ 15.9}_{- 19.2}\pm11.9$\,$μ$Jy/beam at 1.2\,mm and $38.6^{+ 9.6}_{- 13.1} \pm3.7$\,$μ$Jy/beam at 2\,mm (the first uncertainty is statistical, the second is the cosmic variance). In this region, this corresponds to half the instrumental noise. To derive these estimates, we devised a novel estimator, referred to as the cross variance, which also enabled us to estimate the correlated confusion noise between the two bands. Thus, we obtained a result of $49.6^{+ 15.9}_{- 24.8}\pm 6.4$\,$μ$Jy/beam. These values are consistent with the state of the art Simulated Infrared Dusty Extragalactic Sky (SIDES) model.

Figures (8)

• Figure 1: Maps of GOODS-North used in this work (top 1.2 mm, bottom 2 mm). The black contours define the region (i.e. mask) that we use to compute the confusion.
• Figure 2: Variance of the N2CLS maps vs the out of atmosphere integration time on a region in the center equivalent to one NIKA2 field of view (6.5 arcmin diameter). To determine these data points, we simply take the variance of the unmasked pixels (see Fig. \ref{['fig:maps']}) in the central area. For each NIKA2 band, the lower plot shows the residuals compared to the $t^{-1/2}$ instrumental noise integration that is fitted here for times of integration smaller than 20 000 s. The error bars are the square roots of the diagonal elements of the covariance matrix of the measures (Fig. \ref{['fig:rho_covar_matrix']}), derived from 100 Monte-Carlo simulations of map-based realistic noise (per scan). There is a clear excess of signal at high integration time due to the confusion, but as explained in the main text, this plot is for illustrative purposes only and is not used to derive our final results.
• Figure 3: Likelihoods (normalised to their maximum) of the confusion at 1 mm, 2 mm, and the cross-confusion 1$\times$2 mm for different subsets of our data. The central points and their 1-$\sigma$ equivalent error bars are shown for convenience and have arbitrary ordinates. 'Half 1' refers to scans oriented at $40^\circ$. 'Half 2' refers to scans oriented at $-50^\circ$. 'All' refers to the total sum of the two.
• Figure 4: Relative contribution to the background (blue curve) and to the fluctuations causing the confusion (red curve) as a function of the flux density derived with the SIDES simulation at 1.2 mm (top) and 2 mm (bottom) (see Sect. \ref{['se:confusion_sources']}). The same quantities derived from the source counts measurements in GOODS-N by bing are shown as blue circles and red squares, respectively. The difference between the prediction of galaxy counts from SIDES and the source counts from N2CLS is well understood and caused by the blending of galaxies into the NIKA2 beam (see the analysis in bing). The grey area shows the flux density regime where the sources are masked and do not contribute to the confusion.
• Figure 5: Estimates of the cross-variance at 1.2 mm for all pairs of independent observation runs in $(\mu$Jy/beam)$^2$. Error bars are the square root of the diagonal elements of the covariance matrix of $\rho_i$ shown on Fig. \ref{['fig:rho_covar_matrix']}. The dashed line marks zero and the solid line is the combined cross-variance (see Eq. (\ref{['eq:opt_rho']})).