Fine-scale turbulent bursts in stable atmospheric boundary layer in complex terrain

Turbulence in the atmospheric boundary layer (ABL) is usually measured using sonic anemometers (sonics), but coarse spatial (~10 cm) and temporal (~32 Hz) resolutions of sonics preclude direct measurement of fine-scale parameters such as the turbulent kinetic energy (TKE) dissipation rate. Instead, is estimated using techniques based on Kolmogorov theory. Fine-scale measurements of ABL turbulence down to Kolmogorov scale were made with a sonic and hot-film anemometer dyad (a 'combo' probe) during the field campaigns of the Mountain Terrain Atmospheric Modeling and Observations (MATERHORN) programme. The hot-film probe was located on a gimbal within the sonic probe volume, and was automated to rotate in the horizontal plane to align with the mean flow measured by sonic. This procedure not only helped satisfy the requirement of hot-film alignment with the mean flow, but also allowed in situ calibration of hot-film probes. This paper analyses a period of nocturnal flow that was similar to a stratified parallel shear flow. The combo-probe measurements showed an interesting phenomenon - the occurrence of strong bursts, characterized by short-term increase of velocity fluctuations and simultaneous increase of TKE dissipation rate by orders of magnitude. These bursts were indicative of unusual turbulence activity at finer (~0.1-0.4 m) scales that are not captured by sonics since the smallest scales resolved by the latter are greater than 0.6 m. With bursting present, the spectra exhibited bumps at scales intermediate to inertial and dissipation subranges, resembling a bottleneck phenomenon. Its manifestation, although unequivocally related to bursts, may not convincingly fit into the framework of previous bottleneck-effect theories that allude to either viscous effects or buoyancy effects modifying the local energy cascade via non-local effects. The origins of burst are yet to be identified. Stratified ABL with bursts exhibits non-Kolmogorov behaviour, and hence should be modelled with caution.