Remaking Stable Boundary Layer Research, From the Ground Up

A recently accepted essay for the Bulletin of the American Meteorological Society by Joe Fernando and Jeff Weil is good background reading for this week’s AMS 19th Symposium on Boundary Layers and Turbulence in Keystone, Colorado.
Fernando and Weil point out that research into the lowest layer of the atmosphere where we all live and breathe will need to evolve to meet needs in numerical weather prediction. While progress is apparent in the modeling of the boundary layer when it is stirred into convection, those models have obvious shortcomings when the low-level air is not buoyant—the stable boundary layer typically encountered at nighttime. The stable boundary layer controls transport of pollution, formation of fog and nocturnal jets in the critical time before the atmosphere “wakes up” in daytime heating. Weil, in his presentation this Thursday at Keystone calls the still-flawed modeling of the stable situation “one of the more outstanding challenges of planetary boundary layer research.”
Fernando and Weil write in BAMS that study of the stable boundary needs to be retooled to embrace interactions of relevant processes from a variety of scales of motion. The weakness and multiplicity of relevant stable boundary processes means that investigations of individual factors will not be fruitful enough to improve numerical prediction. Scientists need to temper their natural tendencies to try to isolate phenomena in their field studies and modeling and instead seek

simultaneous observations over a range of scales, quantifying heat, momentum, and mass flux contributions of myriad processes to augment the typical study of a single scale or phenomenon (or a few) in isolation. Existing practices, which involves painstakingly identifying dominant processes from data, need to be shifted toward aggregating the effects of multiple phenomena. We anticipate development of high fidelity predictive models that largely rely on accurate specification of fluxes (in terms of eddy diffusivities) through computational grid boxes, whereas extant practice is to use phenomenological models that draw upon simplified analytical theories and observations and largely ignore cumulative effects/errors of some processes.

This new perspective, the authors argue, will be a “paradigm shift” in research and modeling.