The Role of Different Meteorological Regimes on Surface-Layer Turbulence Characteristics Observed with a Wind Lidar Installed at a Forested Ridgetop

The accurate representation of turbulent exchange processes occurring between the land surface and overlying atmosphere is essential for numerical weather prediction (NWP) models. As the horizontal and vertical resolution of NWP models continues to increase and NWP models are better able to resolve terrain complexities, the importance of improved characterizations of turbulent processes becomes increasingly evident. In the present work we characterize the temporal and vertical evolution of turbulence characteristics above a forested mountain ridge and evaluate how these characteristics vary with different meteorological regimes. We perform this evaluation using one year (i.e., 1 June 2023 through 31 May 2024) of observations obtained from a WindCube V2.1 wind lidar installed near Oak Ridge, Tennessee in the Southeast U.S. The wind lidar has 20 m range gates spanning from 40 m above ground level (AGL), or about 1.5 times the canopy height, up to 300 m AGL. We use high-frequency lidar observations to calculate vertical velocity variance (σ_w²), turbulent kinetic energy (TKE), and higher-order moments (i.e., vertical velocity skewness, S, and kurtosis, K). These observations are supplemented by within- and above canopy turbulence observations, as well as a suite of other ancillary meteorological observations, obtained from a nearby 60 m micrometeorological tower located along the same ridgetop as the wind lidar.

We found strong agreement between the lidar-derived and tower-derived σ_w² and TKE over the entire diurnal cycle; the root-mean square error (RMSE) and best-fit slope (m_b) between the lidar- and tower-derived values of σ_w² were 0.18 m² s^-2 and 0.98, respectively. For TKE, RMSE and m_b were 0.64 m² s^-2 and 1.02, respectively, and good agreement was found when distinguishing between daytime and nighttime. The good agreement between the sampling platforms provided us with confidence in using the tower datasets to distinguish and interpret turbulence characteristics across four distinct meteorological conditions: 1) clear and weak winds, 2) clear and strong winds, 3) cloudy and weak winds, and 4) cloudy and strong winds.

Throughout the study period, consistent with previous research at other ridgetop sites, mean winds were larger during the nighttime than during the daytime and the magnitude of this difference was largest at the uppermost sampling heights. Near-surface wind directions were typically from the northwest during the nighttime but were westerly during the daytime, whereas mean winds 300 m AGL were from the west and exhibited little time-of-day dependence. σ_w² at 40 m AGL (300 m AGL) ranged from 0.25 m² s^-2 (0.50 m² s^-2) during the nighttime to a maximum of 0.75 m² s^-2 (1.25 m² s^-2) during the early afternoon, whereas TKE sampled 40 m AGL (300 m AGL) ranged from ~ 1.0 m² s^-2 (1.25 m² s^-2) to 2.5 m² s^-2 (3.0 m² s^-2). When evaluating different meteorological regimes, we observed that the greatest decrease in the diurnal wind speed occurred on clear, windy days. These days were also characterized by the largest σ_w² and TKE, which reached peak values around 2 m² s^-2 and 5 m² s^-2, respectively, in the early afternoon. Larger S and K were observed during the nighttime than during the daytime, and the magnitude of the variability was quite similar across all sampling heights and meteorological regimes. The results from this work provide valuable insights into turbulence processes over forested, complex terrain and can contribute to refining turbulence parameterizations in NWP models.