Evaluation of eddy dissipation rate within a regional atmospheric model (MetUM) using radar retrievals

Turbulent processes are important for cloud evolution and their morphology. Turbulent mixing is partly parametrised in sub-km numerical weather prediction models, whose simulations of convection are sensitive to the configuration of sub-grid turbulence schemes. Past studies have used large-eddy simulations and aircraft observations to characterise turbulence generated by clouds and thunderstorms in Germany and Australia. However, characteristics of in-cloud turbulence and especially its spatial distribution remain poorly understood.

Here, we present an evaluation of sub-km simulations of convective storms with the Met Office Unified Model against turbulence estimates derived from radar measurements collected as part of the 2023 Wessex convection (WesCon) field campaign over the southern United Kingdom. Turbulence intensity is expressed as an eddy dissipation rate, ε and retrieved by isolating the turbulent component of the Doppler velocity spectrum width observed by the Chilbolton Advanced Meteorological Radar.

In a WesCon deep convection case, median in-cloud values of retrieved ε range from 3 × 10^-3 to 2 × 10^-2 m²s^-3, with values increasing with height. Results are compared with equivalent statistics derived from 300-m, 100-m and 55-m grid-length Met Office Unified Model simulations of the observed cases to evaluate the model’s blended sub-filter mixing scheme. More intense turbulence was found near the tops of simulated reflectivity cores with regions of high ε co-located with regions of strong horizontal shear around updrafts. The 95th and 99th percentiles of 300-m grid length model ε are comparable with observations, while simulated ε values within finer grid resolutions are up to half an order of magnitude lower. In contrast to observations, turbulence intensity within simulations peaks in the mid-levels of the convective clouds before decreasing with height.