Modification of the moving point test method for nighttime eddy CO flux filtering on hilly and complex terrains.

The measurement of carbon dioxide (CO) fluxes using the eddy covariance technique is difficult in forests in complex terrain because of the horizontal advection of CO at night. This results the under- or overestimation of net ecosystem exchanges of CO. We propose a technique for nighttime filtering (and correction) of CO fluxes to eliminate (and replace) those significantly affected by horizontal advection: the modified moving point test method. This was developed by merging the friction velocity filtering and van Gorsel methods. It is based on an approach using moving windows for time and friction velocity, allowing a nighttime CO flux correction that includes an assessment of CO drainage at midnight. We tested the method using datasets from two flux towers in forests in hilly and complex terrains, where the application of generic nighttime filtering methods is difficult because CO drainage is generated earlier than the time assumed by the generic methods. The method produced carbon budgets consistent with previous research results, while showing improved applicability. •We propose a nighttime CO flux filtering method for hilly and complex terrain that combines the friction velocity filtering and van Gorsel methods.•This method determines the friction velocity threshold and the significance of CO drainage at midnight based on an approach using moving windows for time and friction velocity.•The method produced consistent results and shows improved applicability.