Toward a Local Climate Zone-based drag and mixing length parametrization for the urban environment

To prepare future urban climate modelling and numerical weather prediction at the hectometer scale in cities with heterogeneous morphology and high-rise buildings, urban climate models have to be coupled at multiple levels with atmospheric models. Vertical profiles of the building drag coefficient and the urban mixing length need to be specified to parametrize the effect of the buildings on the flow. Building-resolving micro-scale simulations can be employed to derive these quantities.

In the present contribution, micro-scale large-eddy simulations of eleven Local Climate Zone (LCZ) based urban morphologies with various building plan and frontal density are used to provide velocity, sectional drag coefficient and mixing length reference vertical profiles for the urban environment. The micro-scale simulations, which are of 1-m resolution in both horizontal and vertical directions, are performed with the MesoNH meteorological research model. This model represents explicitly the obstacles with the Immersed Boundary Method and accounts for the impact of the large-scale turbulence structures on the urban canopy thanks to dynamical downscaling and embedded numerical domains using the grid nesting method. The micro-scale results show that, contrary to traditional assumptions, the velocity profile is generally not exponential and the mixing length is not constant in the urban canopy. This is in agreement with more recent research. The results also show that the building frontal density seems to be a key parameter for the shape of the velocity profile, within and directly above the urban canopy.

The sectional drag and mixing length profiles are then used to propose a new LCZ-based parametrization for the wind dynamics in the urban environment when using the Meso-NH model at the hectometer scale. The results show that the proposed parametrization is more efficient than the current one, consisting in a constant drag coefficient and no specific modification of the turbulent mixing length scale in the urban environment. These results open new perspectives to better parametrize the dynamic effects of real urban areas at the hectometer scale.