Underestimates of sensible heat flux due to vertical velocity measurement errors in non-orthogonal sonic anemometers

Sonic thermometry and anemometry are fundamental to all eddy-covariance studies of surface energy balance. Recent studies have suggested that sonic anemometers with non-orthogonal transducers can underestimate vertical wind velocity (w) and sensible heat flux (H) when compared to orthogonal designs. In this study we tested whether a non-orthogonal sonic anemometer (CSAT3, Campbell Scientific, Inc.) measures lower w and H than an orthogonal sonic anemometer (SATI/3Vx, Applied Technologies, Inc.) and through experimental manipulation we tested if this difference can be attributed to errors in the CSAT3. Four CSAT3s and one SATI/3Vx were mounted symmetrically in a horizontal array on top of the Glacier Lakes Ecosystem Experiments Site (GLEES) AmeriFlux scaffold (southeastern Wyoming, USA) and in close enough proximity to allow covariance measurements between neighboring sonic anemometers. The CSAT3s were paired and measurements of the three orthogonal wind velocities (u, v, and w) were tested by alternatively rotating each sonic anemometer 90° around its u-axis, essentially forcing the sonic v-axis transducer system to measure w. Analysis was performed on data corresponding to gusts of wind located within the 15° cone defined around the u-axis to ensure operation within manufacturer specifications. We found that the CSAT3 measured 8% lower H than the SATI/3Vx and that was associated with a 6–12% lower measurement of w. From the CSAT3 manipulations we found w was underestimated by 6–10% which led directly to an 8–12% underestimate of the kinematic heat flux, the fundamental covariance of H. These results have implications for ecosystem flux research and the energy imbalance problem considering the prevalence of the CSAT3 and the non-orthogonal sonic anemometer design.

► The non-orthogonal CSAT3 measured 8% lower sensible heat flux than the orthogonal SATI/3Vx.
► This was associated with a 6–12% lower measurement of vertical wind velocity.
► We manipulated the CSAT3 and found vertical wind velocity was underestimated by 6–10%.
► This led directly to an 8–12% underestimate of the kinematic heat flux, the fundamental covariance of sensible heat flux.