Above the planetary boundary layer optical turbulence typically occurs on spatial scales of tens of kilometers in the horizontal and only tens of meters or less in the vertical. These “pancake layers” of optical turbulence can be important to applications that require optical viewing along long slant paths at high zenith angles. The small vertical scale of these phenomena means they are almost always sub-grid-scale for mesoscale models and therefore must be parameterized. Unlike the prognostic sub-grid-scale parameterizations that exist in models to account for the energy exchanged between model layers, optical turbulence parameterizations are diagnostic in that they need to quantify a phenomena usually occurring entirely within a model layer without feedback to the adjacent layers. A method to estimate optical turbulence by correlating the turbulence outer scale with vertical wind shear over 300 m intervals has been developed and used extensively by the Air Force Research Laboratory over the past 10 years. While this technique was originally developed for use with high frequency radiosonde data, it has recently been adapted to run with mesoscale model data. Validation studies of the optical turbulence parameterization coupled with mesoscale models have shown that the parameterizaton’s accuracy is highly correlated with the model’s ability to accurately predict the mesoscale wind fields.

An evaluation of the performance of the optical turbulence parameterization using different inputs of mesoscale model data is being carried out. The MM5 will be run in 4 difference horizontal resolutions, 27 km, 9, km, 3 km, and 1 km. The WRF model will be run at 27 km resolution. All five of these configurations will be run with varying number of vertical levels (26, 52, and 78).

The evaluation is based on the accuracy of the resulting optical turbulence predictions and the precision of the mesoscale model output. Accuracy of the optical turbulence predictions is determined by comparison with thermosonde measurements of optical turbulence using integrated variables over simulated slant paths. The precision of the mesoscale model output is assessed by comparing the power spectra of horizontal winds and temperature vertical profiles with that observed by high-resolution radiosonde data. The results from this study will be used to recommend the preferred mesoscale model configuration for forecasting optical turbulence.