The long-term evaluation of boundary-layer turbulence in high-resolution numerical weather prediction simulations using Doppler lidar

Turbulence in the atmospheric boundary layer governs the exchange of heat, moisture, and other atmospheric constituents between the surface and the free troposphere, influencing the initiation of moist convection. As numerical weather prediction models advance toward sub-kilometre-scale grid spacing, an increasing fraction of boundary-layer turbulent motions becomes explicitly resolved, motivating a critical reassessment of turbulence parameterisation frameworks at the turbulence grey-zone scale (partially resolved and partially parametrized turbulence). 

 

This study combines long-term Doppler lidar and sonic anemometer observations from Chilbolton, Hampshire (UK) to characterise fundamental turbulence properties of the atmospheric boundary layer, including profiles of vertical velocity variance and skewness, together with surface sensible heat flux. These turbulence statistics are analysed across a range of boundary-layer regimes, identified using cloud and aerosol layer height information, and are used to evaluate the representation of boundary-layer turbulence in the Met Office Unified Model (MetUM). Observational diagnostics are compared with equivalent statistics derived from long-term MetUM forecasts at 1.5 km and 300 m grid spacing using time-step output. Analysis of a case study showed that the sub-km simulation better represents the turbulence than the 1.5 km simulation but still underestimates the peak values and has a different vertical structure compared to observations. Here, emphasis is placed on the long-term statistics of the vertical structure of vertical velocity variance and its sensitivity to boundary-layer regime. Although the analysis focuses on the UK and the MetUM, the methodology is readily transferable to other locations with Doppler lidar observations and high-frequency model output.