Date of Award

Spring 1-1-2019

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Julie J. Lundquist

Second Advisor

Branko Kosovic

Third Advisor

Peter Hamlington

Fourth Advisor

Domingo Munoz-Esparza

Fifth Advisor

John Cassano

Abstract

The accurate simulation of multi-scale atmospheric processes is crucial for applications such as weather forecasting, flow over complex terrain and wind-energy prediction. However, coupling between mesoscale models and microscale models can present many challenges. In this work we explore some of these challenges, and propose a new solution.

First, we evaluate the effects of unrealistic, mesoscale structures affecting nested large eddy simulations (LES) through boundary conditions, in the case of an idealized, convective boundary layer. Such unrealistic structures result from using mesoscale models at resolutions within a range known as the Terra Incognita (TI) (Wyngaard, 2004). From this analysis we find that unrealistic structures delay the generation of turbulence. However, after a proper turbulence-generation fetch, the LES simulations overcome the influence of the unrealistic structures and develop accurate turbulence.

In the second part of this work, we implement and test a stochastic method for turbulence generation based on the cell perturbation method (CPM) (Munoz-Esparza et al., 2015; Munoz- Esparza and Kosovic, 2018) using the Weather Research and Forecasting model (WRF) Skamarock et al. (2008). This new method uses random vertical and horizontal force perturbations (FCPM) to trigger turbulent motions near the inflow boundaries of LES nested within mesoscale simulations. Tests under idealized convective and neutral stability conditions show that the FCPM can produce a comparable turbulence generation fetch to the CPM, pointing to the robustness of family of random perturbation methods. Vertical force perturbations were found to perform better than horizontal force perturbations for both convective, and neutral stability conditions.

Finally, the third part of this work tests the new FCPM for the case of a real, diurnal cycle over flat terrain. Wind-speed measurements from a meteorological tower in the Scaled Wind Farm

Technology (SWiFT) facility are used to validate the model performance. The FCPM produces closer mean flows to the tower measurements. The FCPM was able to produce turbulent motions during stable atmospheric conditions, while such motions were not produced by the unperturbed model. Additionally, the use of the FCPM was found to accelerate the development of turbulence to a fetch of up to 2.2 km, while the unperturbed case required up to 5.9 km to develop the equivalent scales.

Share

COinS