Classifying observations of stable boundary layers for model validation

The stable boundary layer (SBL) forms the interface between the earth surface and the free troposphere during night time. The development of the SBL directly influences surface temperature, which in turn feeds back on the surface radiation balance. The large wind shear including turning of the wind with height in the SBL causes synoptical scale convergence and divergence. The height to which the SBL develops is of importance in environmental applications like dispersion of atmospheric contaminants. A correct representation of the SBL in atmospheric models is of great importance both for short term weather forecast as for climate change studies.

The atmospheric modellers community is interested in evaluating atmospheric boundary layer schemes with observations not only for more or less ideal SBL cases, but for all the important boundary conditions encountered over the globe. Several interesting observational sites with different boundary conditions are identified within the GEWEX GABLS initiative. Model evaluation is often done on a per case basis. Such an evaluation appears to be quite difficult for all but the most ideal cases because important SBL processes may not be incorporated in the model or may not be measured in the observational data set at hand. Here we investigate ways to present observations from non-ideal sites to the model community. The study is based on long term SBL observations from the Cabauw and Lindenberg meteorological towers. An attempt is made to classify SBL’s at the two sites in a statistical manner exploiting the long time series that exists on both sites. Observations of the structure of the SBL like rate of change of surface temperature, temperature difference and wind veering over the height of the SBL are linked to classes of the relevant forcing variables like long wave cooling and geostrophic wind.

By applying in the future the same methodology to observations at other observational sites valuable information is gathered about the relation between boundary conditions and SBL behaviour which can be used as input for model evaluation.