Accuracy of Surface Drag Coefficient Predictions using Monin-Obukhov Similarity Theory

The surface drag coefficient is a critical parameter in weather and climate forecasting models over both land and ocean surfaces. It is often used as a transfer coefficient in the parameterization of fluxes at the surface (over land) and the air-sea interface (over the ocean). This transfer coefficient contains the majority of the unknown physics that is not directly modeled. Therefore, the capability to accurately predict the surface drag coefficient is crucial in the determination of turbulent momentum fluxes impacting forecasting models.

The surface drag coefficient is typically determined using Monin-Obukhov Similarity Theory (MOST). The main outcome from MOST is that the horizontal wind, corrected for atmospheric stability, scales with the surface momentum flux. Predictions based on MOST have been shown to work well over homogeneous terrain with stationary flow conditions for a stable and moderately unstable atmosphere. MOST does not generally perform well in highly unstable conditions. It is hypothesized in the present study that, in the absence of free convection, differences between experimental measurements of the surface drag coefficient and predictions from MOST can be attributed to (i) uncertainties in the measurements, (ii) nonstationarity effects, and (iii) a nonconstant stress layer.

The hypothesis was tested using experimental data acquired at the Surface Layer Turbulence and Environmental Facility (SLTEST) in Dugway, Utah. Data were collected from nine CSAT3 sonic anemometers sampled at 20Hz and mounted at varying heights ranging from 1.42 meters to 25.69 meters. Three data sets were analyzed for the present study representing 42, 38, and 27 hours of continuous measurements. The Utah Salt Flats provide a flat surface with little vegetation, which eliminates the complexities that are introduced by roughness elements such as hills, trees, and ocean waves. It was found that a 5% uncertainty in the mean wind and the friction velocity yield a 11.2% relative uncertainty in the drag coefficient. Nonstationarities in the heat flux and wind direction were used to account for the majority of the inconsistencies between MOST and the experimental data; while the presence of a nonconstant stress layer had a less significant effect on the ability of MOST to predict the data.