One area often overlooked in mesoscale modeling is proper treatment of turbulence in a saturated environment. Turbulence parameterizations which do not accurately represent mixing within cloudy layers can produce saturated unstable layers with unrealistic temperature and moisture profiles. This in turn affects the boundary layer cloud and radiative forcings. We investigate the sensitivity of mesoscale model solutions to the treatment of condensate in a 1.5-order turbulent kinetic energy (TKE) scheme within the Penn State University / National Center for Atmospheric Research mesoscale model MM5.

Two different case studies are used to assess the role of turbulence in saturated environments: one case features marine boundary layer stratus in the California coastal zone and the other involves radiation fog in the San Joaquin Valley. Nested model simulations with 36-km, 12-km and 4-km grid resolutions are used. The model uses the variables ice-liquid water potential temperature and total water mixing ratio within its turbulence scheme because they are conservative in both saturated and unsaturated environments. Expressions for important terms present in the parameterized TKE equation (e.g., the buoyancy production and the Brunt-Vaisala frequency) depend on the saturation state of the grid cell. In saturated conditions, it is shown that they depend more strongly on the vertical moisture gradients than the thermal gradients, and that turbulent mixing may be too weak in cloudy layers when using the classical dry formulations for these parameters. Results show how the TKE, boundary layer structure, cloud and fog fields may be adversely affected when not properly accounting for mixing in saturated layers.