Closed mafavaron closed 6 years ago
Cause found in the very definition of wind speed. In SonicLib, the correspondent of pbl_met is named "resultant.vel".
This closes the current issue, as far as the "bug" is considered.
The question remains, as whether extracting wind speed and other common quantities a posteriori is better than computing them directly on eddy covariance processing, as made for example with u*.
Surely, a-posteriori extraction may limit the possibilities, and could produce misleading results more easily than computation-within.
The version under test tends to underestimate SonicLib wind speed.