Spectral dependence of the correction for path-smoothing by 3-D anemometers

The use of long-term eddy covariance measurements to accurately assess net ecosystem carbon exchange is critically affected by small biases in the measurement system, particularly during nighttime periods. It is correspondingly important to accurately correct measured fluxes for the effect of limitations of the measuring system ( e.g. Moore, BLM 1986; Massman, in press, 1999).

One of the corrections which is applied is that for the smoothing due to pathlength of the sonic anemometer - any method of spectral correction which involves estimating the instrumental response and integrating over a spectrum uses this correctional curve. There are two problems with this correction as commonly applied

a. the description of this correction is one which has been integrated over the so-called "Kaimal" spectra (Kaimal et al, QJRMS, 1972) which may not be appropriate to the chosen experimental site.

b. in many 3D sonic anemometers the transducer heads are not aligned with the axes on which the velocity components are being measured.

The current paper separates the description of the instrument's geomoetry from that of the turbulence, allowing the examination of both questions. The sensitivity of the form of the path correction to the spectral shape is investigated using preliminary forest spectra, and the error in flux-correction due to inappropriate spectra evaluated at different states of forest canopy development. The effect of the misalignment between axes and velocity components is calculated for the Kaimal spectra, and the errors are found to be significant only at high frequencies.