Long Term Observations of Longwave Radiative Flux Divergence in the Stable Boundary Layer over Land

During clear and calm nights the atmospheric stability is often too strong to maintain turbulence. Then longwave radiation divergence plays a relatively important role in governing the temperature structure under these conditions in the stable boundary layer. However, in the literature there is no agreement on the sign and the magnitude of near surface longwave radiation divergence under low wind conditions. At the same time we realize that large-scale weather forecast and climate models poorly represent (and mostly underestimate) longwave radiation divergence in the stable boundary layer. This is because radiation schemes are a) computationally expensive, b) resolution dependent and c) are usually calibrated on the mean cooling in the full atmospheric column. It appears that the radiative cooling rate in the shallow stable boundary layer can significantly differ from the tropospheric cooling rate. Since the atmospheric state of the SBL over land can change quickly and is often non-stationary, radiation divergence needs to be represented properly and with high temporal resolution.

Direct observations of radiation divergence close to the surface are scarce. In this study we present a new half-year dataset of observed radiation divergence on a 20 m tower over grassland in Wageningen, The Netherlands. We find that the radiative cooling is strongest after the day-night transition, and amounts then on average 2 K/h, and gradually diminishes during night. In addition, substantial radiative heating during the day is found. Furthermore, the radiative cooling is dominated by the divergence of the upwelling longwave component.

Our aim is to develop a robust and simple and computationally cheap model for parameterizing long wave radiation divergence in large-scale models. As such we examine the performance of a statistical (linear regression) model and a physically based model.