Posters

Abstract

Meso-microscale coupling has the potential to offer a "best-of-the-both-worlds" approach to wind resource modeling by combining the high-resolution topography of a microscale model with the capability of a weather model to simulate the mesoscale flow. The linear WAsP model and Large-eddy simulations represent two extremes of microscale models which have been used in a coupled framework. The former, while computationally efficient, struggles in complex terrain, whereas the latter remains too computationally heavy to be used within the cost and time constraints of a typical wind farm developer. Reynolds-Averaged Navier Stokes (RANS) solutions present an attractive compromise between simulation fidelity and cost, but have received less attention in the meso-microscale coupling community. In this work, we propose a method to couple a RANS solver to a mesoscale model through an addition of a term in the horizontal momentum equations that nudges (or relaxes) the RANS solution closer to the time-averaged mesoscale flow. It is known that the error in an uncoupled microscale model grows with increasing spatial scale, hence instead of relaying on a spatially constant nudging strength, we propose a new method where the nudging is scale-aware. The purpose of the nudging is to account for the large differences in the way the two models simulate the boundary layer, and nudging should be applied cautiously such that the coupled model only approaches the mesoscale solutions on spatial scales where those phenomena dominate. In this presentation, we will show how the coupled model behavior varies with the strength of the nudging for a site with strong katabatic winds. Finally, we discuss cross-prediction errors and whether the use of the nudging may lead to an improvement in model accuracy.