Mesoscale Model Response to Random, Surface-based Perturbations - a Sea-breeze Experiment

The introduction into a mesoscale model of random (in space) variations in roughness length, or random (in space and time) surface perturbations of temperature and friction velocity, produces a measurable, but barely significant, response in the simulated flow dynamics of the lower atmosphere. The perturbations are an attempt to include the effects of sub-grid variability into the ensemble-mean parameterization schemes used in many numerical models. Their magnitude is set in our experiments by appeal to real-world observations of the spatial variations in roughness length and daytime surface temperature over the land on horizontal scales of one to several tens of kilometers. With sea-breeze simulations, comparisons of a number of realizations forced by roughness-length and surface-temperature perturbations with the standard simulation reveal no significant change in ensemble mean statistics, and only small changes in the sea-breeze vertical velocity. Changes in the updraft velocity for individual runs, of up to several cm s-l (compared to a mean of 14 cm s-l), are directly the result of prefrontal temperature changes of 0.1 to 0.2 K, produced by the random surface forcing. The correlation and magnitude of the changes are entirely consistent with a gravity-current interpretation of the sea breeze.