Characterization of Near-surface Wind Profiles Based on Atmospheric Thermal Stability and Turbulence Characteristics

Characterizing the atmospheric boundary layer (ABL) wind is vital for wind turbine power generation, air pollution control and wind loading design on civil structures. However, the behavior of winds in the ABL is heavily influenced by turbulence, something which has not been adequately considered for the prediction of mean wind flow speeds. Vertical wind speed profiles predicted by the power-law (PL) model are commonly used for wind speed extrapolation in wind power forecasting. We investigated how the power-law shear exponent varies based on different quantifications of thermal stability and turbulence using ABL weather data collected by both cup anemometers and high-resolution ultrasonic sensors from a 106m meteorological tower at Cedar Rapids, Iowa. Bulk and flux Richardson numbers were compared as indicators of stability. Autocorrelation analysis was performed to determine integral time and length scales of surface-layer turbulence. With such tools in hand, we employ a basic machine learning (ML) approach and compare the results of traditional PL models and ML methodology to better understand alpha value distributions across various stability and turbulence stratum.