A method to approximate the effects of turbulence by retaining a prognostic equation for mean variables (such as wind or temperature), but where a profile shape of those mean variables is assumed a priori. For example, in the boundary layer, if the daytime profile of potential temperature is assumed to be uniform with height, then only one temperature forecast equation is needed for the whole layer. Similarly at night, if an exponential shape is assumed for the potential temperature profile in the stable boundary layer, then only one temperature forecast equation is needed for the cooling at the surface. This approach is less computationally expensive than solving forecast equations at every height within the mixed layer or stable boundary layer. See closure assumptions, first-order closure, higher-order closure, nonlocal closure.