Direct advection measurements do not help to solve the night-time CO2 closure problem : evidence from three different forests

The ADVEX project involved conducting extensive advection measurements at three sites, each with a different topography. One goal of the project was to measure the [CO2] balance under night-time conditions, in an attempt to improve NEE estimates. Four towers were arranged in a square around a main tower, with the sides of the square about 100 m long. Equipped with 16 sonic anemometers and [CO2] sampling points, the towers were installed to measure vertical and horizontal advection of [CO2]. Vertical turbulent fluxes were measured by an eddy covariance system at the top of the main tower. The results showed that horizontal advection varied greatly from site to site and from one wind sector to another, the highest values being reached when there were large friction velocities and fairly unstable conditions. There was less variation in vertical advection, the highest values being reached when there were low friction velocities and stable conditions. The night-time NEE estimates deduced from the mass balance were found to be incompatible with biologically driven fluxes because (i) they varied strongly from one wind sector to another and this variation could not be explained in terms of a response of the biologic flux to climate, (ii) their order of magnitude was not realistic and (iii) they still showed a trend vs. friction velocity. From a critical analysis of the measurement and data treatment we concluded that the causes of the problem are related to the representativeness of the measurement (control volume size, sampling resolution) or the hypotheses underlying the derivation of the [CO2] mass balance (ignoring the horizontal turbulent flux divergence). This suggests that the improvement of eddy flux measurements by developing an advection completed [CO2] mass balance at night would be practically difficult.