Bias between uSonic3- and CSAT3-fluxes explained?

The bias of measured eddy-correlation-fluxes of energy and matter became a topic of concern since they were collected on a larger scale in the frame of earth system research.
A recent study by Horst et al. provides new insights into the flow distortion of sonic anemometers and its impact on EC-fluxes. The conclusion of the paper is that scalar fluxes are underestimated by few percent, if non-orthogonal sensor heads are used and if no flow distortion correction is applied. Consequently the application of flow correction is  recommended for non-orthogonal sensor heads.
This study considers particularly the Campbell Scientific CSAT3 sensor head, and it seems that the results help to resolve an old unexplained discrepancy between METEK uSonic3- (former designation: USAT3) and CSAT3-fluxes:
Although a full 3D flow distortion correction is available for the uSonic3 sensor head, some customers are not comfortable with the application of this correction, because the resulting EC fluxes show a positive bias of few percent with respect to CSAT3 results, which are often considered as reference. After the above mentioned study there is strong evidence that the observed bias could be largely removed by applying an appropriate flow correction also to the CSAT3 results.