Parameterization of the Atmospheric Boundary Layer A

tion models, in spite of some advances, the boundary layer is still not represented realistically. Figure 1 illustrates an aspect of this problem by showing the opposing response of boundary layer (low) clouds to perturbation experiments (e.g., double CO2) in current climate models. The general problem of parameterization in fluids dates back to the first modern studies of turbulence during the nineteenth and early twentieth centuries. By then, it was already clear that for turbulent flows such as the atmosphere, it was not feasible (or even relevant) to try and follow every parcel of f luid in its turbulent trajectories. Instead, research should concentrate on trying to understand the statistical properties of turbulent f lows. With the advent of computers came the possibility of developing numerical models for weather and climate prediction. Numerical discretizations imply a limit for the temporal/spatial scales below which the flow cannot O ne of the main components of the climate system is the atmospheric boundary layer, which mediates the interactions between the ocean/land surface and the free atmosphere. Several boundary Parameterization of the Atmospheric Boundary Layer A View from Just Above the Inversion