Micrometeorology of the Southern Great Plains: Legacy Relationships and Short-Term Weather Forecasting for Wind Energy

A central problem for micrometeorologists has been the relationship of air-surface exchange rates of momentum, heat, water vapor, carbon dioxide and various pollutants to quantities that can be predicted with confidence or extracted from available data sources. Flux-gradient formulations are intended to facilitate this process. Regardless of what form of these relationships might be used, their significance over any short period of time is questionable since all of these relationships between fluxes and gradients apply to averages that might rarely occur. It is well accepted that the assumption of stationarity and homogeneity do not reflect the true chaotic nature of the processes that control the variables considered in these relationships, with the net consequence that the levels of predictability theoretically attainable might never be realized in practice. This matter is of direct relevance to modern prognostic models which construct forecasts by assuming the universal applicability of relationships among averages for the lower atmosphere, which rarely maintains an average state. As an outcome of discussions between Duke Energy Generation and NOAA/ARL following the 2009 AMS Summer Community Meeting, in Norman Oklahoma, ARL and Duke Energy Generation (Duke) signed a Cooperative Research and Development Agreement (CRADA) which allows NOAA to conduct atmospheric boundary layer (ABL) research using Duke renewable energy sites as research testbeds. One aspect of this research has been the evaluation of legacy flux-gradient formulations (the ϕ functions, see Monin and Obukhov, 1954) for the exchange of heat and momentum. Sonic anemometers reported winds and temperature fluctuations at 10Hz for heights above the ground of 2.99m, 8.50m, 10.0m, 14.75m, 17.2m, 20.99m, and 27.39. From these observations, ϕm and ϕh were derived. The overall dependence of ϕm and ϕh on Z/L is characterized by considerable scatter with the familiar relationships best describing the averages; however it is the scatter that largely defines the attainable levels of predictability.