Dependence of the clustering power spectrum variance (0.13 < k < 1 arcmin⁻¹) on the observing frequency (ν) and the survey size (Ω). Top row: the black points are the CO data while the high transparency points are the ones excluded due to the CO(2–1) contamination as explained in Sect. 5.2. The solid orange line is the best-fit model. Bottom row: the black points are the [CII] data and the orange solid line is the best-fit model.