Efforts towards a more complex and realistic representation of the exchanges between the surface and the atmospheric boundary layer in Canada's prediction systems

Even though one of the most important objectives of numerical weather prediction is to provide information and guidance on meteorological conditions at the surface, accurate prediction of screen-level and near-surface temperature, humidity, and winds remain a challenge for all national prediction centres. Success in this endeavour depends on several factors including process modeling and specification of initial conditions for surface as well as upper-air modeling. But it also depends on the manner in which the surface and atmosphere are made to interact in atmospheric models. In most systems currently operational at national prediction centres, the surface and atmosphere interacts at the same horizontal resolution through turbulent exchanges of heat, water, and momentum based on stability functions that are used to make the connection between the surface canopy and the first atmospheric level, typically located 10-20 meters above the surface canopy.

The main objective of this presentation is to describe recent efforts at Environment Canada to represent in a more complex manner the link between the surface and the amosphere. This work includes impact evaluation of substantially increasing vertical resolution of the atmospheric model near the surface. It also examines results from the implementation of a multi-layer version of the Town Energy Balance (TEB) scheme, i.e., a version in which the urban canopy interacts with the atmosphere over a certain number of levels, tested for a near-neutral nocturnal case over Oklahoma City. It will be shown at the conference that both of these developments led to significant improvements of near-surface atmospheric conditions. The possibility of using higher-resolution land surface modeling (with respect to the atmospheric modeling) will also be discussed, along with a possible harmonization with the previous developments related to the vertical structure of the model.