Comparison between Time and Frequency-domain Processing of Anemometer Data

Turbulence strength can be estimated from sonic anemometer data using both time-domain and frequency-domain approaches; these techniques are compared here. Atmospheric turbulence has a strong impact on both laser propagation and long-range imaging. Measurement of turbulence is thus important for predicting optical system performance. The ultrasonic anemometer is an attractive instrument for making point measurements of turbulence in support of optical experiments. It makes rapid measurements of air temperature and wind speed. These measurements are then processed to produce estimates of turbulence strength as well as the turbulence outer scale. There are two main routes for estimating turbulence from these measurements. In a frequency-domain approach, the temporal power spectrum of the temperature measurements is fit to a -5/3 power-law, since turbulence is expected to obey this Kolmogorov power spectrum at least within some inertial range. In a time-domain approach, the structure function of the temperature measurement is computed, and this structure function is fit to a 2/3 power-law. This is the structure function predicted for the Kolmogorov power spectrum. Because these different approaches use the temperature measurements quite differently, there are some differences in the results obtained. With the frequency domain approach, the desired power law expected to be observed for frequencies in the middle of the range used, while with the time-domain approach, the desired functional form is only seen for the shortest separations in time. These two approaches are used to analyze experimental data collected in field experiments at Wright-Patterson AFB at heights of 2.64 and 5 meters above the ground. Because these two approaches use the same data in different ways some differences in the results obtained can be d. The reasons for these differences in the results will be discussed.