A Finite-Element Model of the Atmospheric Boundary Layer Suitable for Use with Numerical Weather Prediction Models

View More View Less
  • 1 Département de Physique, Université du Québec à Montréal, Montréal, Québec, Canada
  • | 2 Recherche en Prévision Numérique, Service de l'Environnement Atmosphérique, Dorval, Québec, Canada
© Get Permissions
Full access

Abstract

We give a detailed description of an atmospheric boundary layer model capable of simulating the diurnal cycles of wind, temperature and humidity. The model includes a formulation of various physical processes (radiative effects, variation of soil surface temperature and humidity, etc.) and uses a first-order closure for turbulent fluxes that relies upon a time-dependent equation for turbulent kinetic energy and on a mixing length governed by a relaxation process. The exchange processes taking place in the surface layer are dealt with in a separate micrometeorological module.

The one-dimensional model uses a Galerkin technique based on linear finite elements, variable resolution in the vertical, and a time discretization of the Crank-Nicholson type. A simulation test based on day 33 of the Wangara Australian experiment indicates that the model, despite its relative simplicity, gives realistic results that compare favorably with those from higher order models while taking much less space and time on the computer. This could make feasible its use by operational numerical weather prediction models.

Abstract

We give a detailed description of an atmospheric boundary layer model capable of simulating the diurnal cycles of wind, temperature and humidity. The model includes a formulation of various physical processes (radiative effects, variation of soil surface temperature and humidity, etc.) and uses a first-order closure for turbulent fluxes that relies upon a time-dependent equation for turbulent kinetic energy and on a mixing length governed by a relaxation process. The exchange processes taking place in the surface layer are dealt with in a separate micrometeorological module.

The one-dimensional model uses a Galerkin technique based on linear finite elements, variable resolution in the vertical, and a time discretization of the Crank-Nicholson type. A simulation test based on day 33 of the Wangara Australian experiment indicates that the model, despite its relative simplicity, gives realistic results that compare favorably with those from higher order models while taking much less space and time on the computer. This could make feasible its use by operational numerical weather prediction models.

Save