One of the traditional uses of scintillation measurements is to infer path-averaged values of the turbulent surface fluxes of momentum (tau) and sensible (H_s) and latent (H_L) heat.  Many scintillometer set-ups, however, measure only the path-averaged refractive index structure parameter, C_n²; the wind information necessary for inferring the fluxes comes from point measurements or is entirely absent.  In this latter case, H_s and H_L can be obtained only under the assumption of free convection; and a path-averaged tau is not available.  The Scintec SLS20 surface-layer scintillometer system, on the other hand, measures both C_n² and the inner scale of turbulence, l₀, where l₀ is related to the dissipation rate of turbulent kinetic energy, epsilon.  As a result, the SLS20 is presumed to provide path-averaged estimates of both tau and H_s.

            I have used an SLS20 in two experiments:  one over Arctic sea ice during SHEBA (the experiment to study the Surface Heat Budget of the Arctic Ocean), and a second over a mid-latitude land site during spring (the so-called Rapid Forcing Experiment).  In each deployment, we also made eddy-covariance measurements of tau, H_s, and H_L near the scintillometer path.  In this presentation, I will compare the scintillometer and eddy-covariance measurements of tau and H_s for these two diverse sites.  For both tau and H_s, the correlation between eddy-covariance and scintillometer-derived values is reasonable, but the scintillometer results tend to be biased low.  Moreover, estimates of both tau and H_s degrade as C_n² decreases.  Essentially, the signal-to-noise ratio for the scintillometer-derived fluxes decreases as C_n² decreases.

            An essential question that arises during these comparisons is what similarity functions to use for nondimensionalizing C_n² and epsilon.  From the combined scintillation and eddy-covariance measurements, I will evaluate which of several candidate similarity functions are consistent with the data.