We utilise ground-based, balloon-borne and satellite climatology data to reconstruct site and season-dependent vertical profiles of precipitable water vapour (PWV). We use these profiles to solve radiative transfer through the atmosphere, and derive atmospheric brightness temperature ($T_{\rm atm}$) and optical depth ($\tau$) at centimetre wavelengths. We validate the reconstruction by comparing the model column PWV with photometric measurements of PWV, performed in clear sky conditions pointed towards the Sun. Based on the measurements, we devise a selection criteria to filter the climatology data to match the PWV levels to the expectations of the clear sky conditions. We apply the reconstruction to the location of a Polish 32-metre radio telescope, and characterise $T_{\rm atm}$ and $\tau$ year-round, at selected frequencies. We also derive the zenith distance dependence for these parameters, and discuss the shortcomings of using planar, single-layer, and optically thin atmospheric models in continuum radio-source flux-density measurement calibrations. We obtain PWV-$T_{\rm atm}$ and PWV-$\tau$ scaling relations in clear sky conditions, and constrain limits to which the actual $T_{\rm atm}$ and $\tau$ can deviate from those derived solely from the climatological data. Finally, we suggest a statistical method to detect clear sky that involves ground-level measurements of relative humidity. Accuracy is tested using local climatological data. The method may be useful to constrain cloud cover in cases when no other (and more robust) climatological data are available.