Variations in Mixed-Layer Depths Arising from Inhomogeneous Surface Conditions

Current approaches to parameterizations of sub-grid-scale variability in surface sensible heat fluxes in general circulation models normally neglect the associated variability in mixed-layer depths. Observations and a numerical mesoscale model are used to show that the magnitude of such variability can be significant. Over a domain of (41 km)2, the standard deviation of simulated mixed-layer depths was found to be 21%–24% of the mean noontime values on three days, and the mean depths were not simply related to the mean sensible heat fluxes. Results obtained with two-dimensional simulations over idealized distributions of warm, dry and cool, or moist surfaces show that as the characteristic sizes of individual patches increase, the distributions of mixed-layer depths tend to assume a bimodal nature. under these conditions, the mean mixed-layer depth may have little physical relevance. Finally, the use of domain-averaged values of wind and temperature to compute surface fluxes is shown to be another potential source of error in flux parameterizations.