<title>Measurements of temperature and velocity structure parameters from sonic anemometers</title>

Abstract

A study of two methods for computing temperature and velocity structure parameters C<SUB>T</SUB><SUP>2</SUP> and C<SUB>v</SUB><SUP>2</SUP> is presented. Fluctuating temperature and wind data obtained from sonic anemometers were processed utilizing time and frequency analyses to derive C<SUB>T</SUB><SUP>2</SUP> and C<SUB>v</SUB><SUP>2</SUP>. The time-domain method is less rigorous and simplifies processing of the sonic anemometer data for deriving the turbulence parameters. The frequency-domain method has the advantage of providing the spectral characteristics of turbulence. This feature greatly facilitates in identifying data that contain spike noise. Thus, corrections may be made to adjust the derived values accordingly. The comparison shows a very good correlation between the two methods. However, the frequency-domain method yield results that are 2 2/3 times greater than the corresponding values from the time-domain method. The difference was determined to be a bias in the frequency-domain method, resulting from the coefficient used in the inertial subrange form of the one-dimensional spectrum, (Phi) <SUB>theta</SUB>((kappa) ) equals (alpha) C<SUB>theta</SUB><SUP>2</SUP>(kappa) <SUP>-5/3</SUP>. In order to remove the bias, the coefficient (alpha) must be changed from the original value of 0.25 to 0.50.