Estimating the Higher-Order Turbulence Statistics from LES-Generated Atmospheric Boundary Layer Flow Fields

Atmospheric boundary layer (ABL) flows contain a myriad of turbulent eddies with sizes ranging from a few millimeters to several hundred meters. Over the past few decades, research involving large-eddy simulation (LES) of the ABL has been gaining popularity due to its inherent ability to explicitly resolve a significant portion of these eddies (typically sizes greater than a few meters). The accuracy of any LES model largely depends on the performance of its sub-grid scale (SGS) parameterization in representing the energy cascade and energy dissipation processes. Traditionally, if the LES-generated energy spectrum (or second-order structure function) captures the so-called Kolmogorov's inertial-range scaling (K41), the simulation is deemed to be a success. In this work, we raise the bar and investigate if the contemporary SGS parameterizations (including dynamic models) have the ability to capture the higher-order turbulence statistics (e.g., the intermittency exponent). To this end, we utilize surrogate analysis as a rigorous hypothesis testing framework. We extensively compare the LES-generated results with those reported in the literature (primarily based on laboratory experiments and field observations). Lastly, we perform several sensitivity experiments to document the effects of SGS parameterizations and spatial resolution on the simulated higher-order statistics.