Force-Based Perturbations for Turbulence Generation in Mesoscale-to-LES Grid-Nesting Applications

Atmospheric phenomena are characterized by a very wide range of scales of motion, making it difficult to simultaneously model all the factors that may influence an atmospheric process. Fine scales require high resolution, while coarse scales require large domains. Very large, finely resolved domains are computationally prohibiting. As a solution to this issue, multi-scale modeling methods are being explored for concurrent modeling of coarse and fine scales. One such method is grid-nesting, which uses the output of large-domain, coarsely resolved simulations as boundary conditions for small-domain, highly resolved simulations. This method can even be used to couple between mesoscale, weather-prediction models and microscale models such as Large Eddy Simulations (LES).

A recurring issue for the transition between mesoscale and microscale models is the generation of turbulence in the microscale simulations. When grid-nesting between mesoscale and LES, large fetches are necessary for turbulence to develop in the nested LES domain. The highly resolved, long distance needed for turbulence generation significantly increases computational costs. For years, turbulence generation methods have been explored to trigger turbulent motions and reduce this distance. One widely used technique is the cell perturbation method [Muñoz-Esparza et al., 2014, 2015], which adds random potential temperature perturbations within a region near the nested domain’s inflow boundaries that significantly reduce the fetch to generate fully developed turbulence.

Perturbation methods have shown advantages over other types of inflow generators, including rapid development of turbulence and computational efficiency. In addition, previous studies have suggested the potential of a combination of temperature, and velocity perturbations as a way to accelerate the generation of turbulence [Mirocha et al., 2013]. In the present study, we implement a new variant of the CP method, which uses perturbations in the form of horizontal and vertical forces instead of potential temperature perturbations. Implementation of this new approach has been carried out within WRF, and a set of comparisons of the different variants of perturbation methods for stable, convective and neutral atmospheric conditions will be presented.

Figure: Horizontal contours of x-velocity, u, show the development of turbulence using different perturbation techniques.