The challenges of vegetation - atmosphere coupling in the roughness sublayer of a numerical weather prediction system

Many studies have shown that the standard flux-gradient relations based on the similarity theory (MOST) are not fully applicable close to a surface covered with high roughness elements, due to the canopy-induced turbulent mixing caused by their interaction with the atmospheric flow. The effects of the so called roughness sublayer (RSL) on the surface-atmosphere coupling became more important since the lowest atmospheric level in NWP systems has been placed closer to the (vegetation or urban) canopies acting as roughness elements, entering the RSL. This in turn affects the effective surface fluxes of momentum, energy and gasses between the canopy and the atmosphere.

Different corrections to the traditional MOST attempting to account for the RSL effects on the surface layer theory are available. In this work we integrate Harman and Finnigan’s RSL theory (HF2008) into SURFEX8.1, a land surface model which is extensively used both offline and coupled to different NWP atmospheric models. HF2008 stands out as the most advanced theory of the RSL. It has been tested over forests and urban areas and implemented in land surface and NWP models, adding increased physical details in the classical similarity theory over a vegetated surface, and thus improving the surface-atmosphere coupling in NWP systems above tall vegetation. 

Our implementation is tested both through offline SURFEX experiments and online simulations in the Harmonie-Arome operational NWP system, and validated against observations in forest stations from the ICOS network. We address the differences in performance of the parameterisation, caused mainly by assumptions and simplifications usually done in the vegetation-atmosphere coupling in NWP systems.