On how BSRN data reveal poorly understood features of land-atmosphere interaction during the morning transition. A tribute to Cabauw Ruisdael Observatory

Analysing 1 minute data gathered at the Baseline Surface Radiation Network (BSRN) station at Cabauw, the Netherlands, for approximately 50 cloudless sky days, it is found that  the apparent effective emission coefficient ε of the atmosphere, defined via R_in = εσT₂⁴ , with R_in the incoming longwave radiation, T₂ the air temperature at 2 m (in K) and  the Stephan-Boltzmann constant, dips just after sunrise. The objective our study is to show that this feature is a diagnostic for various relatively little-described aspects of land-atmosphere interactions during the morning transition. The observed dip is caused by the fact that just after sunrise T₂rises faster than R_in .This occurs during two so-called cross-over periods. Firstly, a stable cross-over period, earlier found by Bosveld et al. (2014) and Angevine et al. (2001) when the atmosphere just above the surface is stably stratified, and, secondly, an unstable cross-over in which this feature continues to exist, when a convective boundary layer (CBL) develops. These findings imply that during the morning transition a decoupling takes place between the air layer near the ground, in which T₂is increasingand an air layer higher in the atmosphere determining R_in. By rewriting Eq. 1 in terms of an effective temperature, T_eff, to which an effective height, z_eff, can be assigned, it is made plausible that the height of latter air layer is higher than 200 m. The consequence of this finding is that, during the morning transition of clear days,all empirical formulas for ε found in the literature can hold only under the locally dependent environmental conditions for which these are trained. These formulas cannot be applied universally. Furthermore, because the conditions just after sunrise are non-stationary, it is to be expected that radiation algorithms in numerical models in which time steps of several hours are used, will fail to determine accurately R_in,during the morning transition. This might explains the relatively large systematic errors in R_in in climate models as reported in the literature.It will be shown that formulas for ε derived for stationary conditions using sophisticated radiation transfer models such a MODTRAN, also do not work during the highly non-stationary circumstances in the morning transition.