Coupled Mesoscale-Large-Eddy Modeling of Realistic Stable Boundary Layer Turbulence

TL;DR

This study develops a coupled mesoscale-large-eddy modeling framework using WRF and LES models to generate realistic stable boundary layer turbulence data, validated against diverse observations, highlighting both successes and challenges.

Contribution

It introduces a novel coupled modeling approach for stable boundary layer turbulence, combining WRF and LES models with multi-platform validation.

Findings

Reproduces characteristics of low-level jets.

Captures energy spectrum scaling regimes.

Fails to model intermittent surface fluxes.

Abstract

Site-specific flow and turbulence information are needed for various practical applications, ranging from aerodynamic/aeroelastic modeling for wind turbine design to optical diffraction calculations. Even though highly desirable, collecting on-site meteorological measurements can be an expensive, time-consuming, and sometimes a challenging task. In this work, we propose a coupled mesoscale-large-eddy modeling framework to synthetically generate site-specific flow and turbulence data. The workhorses behind our framework are a state-of-the-art, open-source atmospheric model called the Weather Research and Forecasting (WRF) model and a tuning-free large-eddy simulation (LES) model. Using this coupled framework, we simulate a nighttime stable boundary layer (SBL) case from the well-known CASES-99 field campaign. One of the unique aspects of this work is the usage of a diverse range of observations for characterization and validation. The coupled models reproduce certain characteristics of observed low-level jets. They also capture various scaling regimes of energy spectra, including the so-called spectral gap. However, the coupled models are unable to capture the intermittent nature of the observed surface fluxes. Lastly, we document and discuss: (i) the tremendous spatio-temporal variabilities of observed and modeled SBL flow fields, and (ii) the significant disagreements among different observational platforms. Based on these results, we strongly recommend that future SBL modeling studies consider rigorous validation exercises based on multi-sensor/multi-platform datasets. In summary, we believe that the numerical generation of realistic SBL is not an impossible task. Without any doubt, there remain several computational and fundamental challenges. The present work should be viewed as a first step to confront some of these challenges.