We consider an imperfection of real closed-path eddy correlation systems—the decoupling of the water vapour and CO2 concentrations—with respect to the application of the Webb–Pearman–Leuning (WPL) theory. It is described why and how the current application of the WPL theory needs to be adapted to the processes in closed-path sensors. We show the quantitative effects of applying the WPL theory in different ways using CO2 flux measurements taken above the Danish Beech forest CarboEurope site near Sorø, Zealand.
Using the WPL theory in closed-path sensors without taking amplitude damping and decoupling into account, overcorrected the annual flux by 21%, or 31 g m−2 yr−1, to which the decoupling effect contributed with 7%. We suggest either converting the raw data point-by-point to mixing ratios or using the uncorrected covariances of water vapour mole fractions with the vertical wind velocity that were calculated with the same time lag as for the scalar concentration when correcting the dilution effect. We showed that the two approaches yielded equivalent flux results. Correct ways of applying spectral corrections to CO2 fluxes calculated in either way are also shown. The findings reported here do not apply to open-path sensors.