Impact of CO2 storage flux sampling uncertainty on net ecosystem exchange measured by eddy covariance
Research output: Contribution to journal › Article
Complying with several assumption and simplifications, most of the carbon budget studies based on eddy covariance (EC) measurements quantify the net ecosystem exchange (NEE) by summing the flux obtained by EC (FC) and the storage flux (SC). SC is the rate of change of a scalar, CO2 molar fraction in this case, within the control volume underneath the EC measurement level. It is given by the difference in the quasi-instantaneous profiles of concentration at the beginning and end of the EC averaging period, divided by the averaging period. The approaches used to estimate SC largely vary, from measurements based on a single sampling point usually located at the EC measurement height, to measurements based on profile sampling. Generally a single profile is used, although multiple profiles can be positioned within the control volume. Measurement accuracy reasonably increases with the spatial sampling intensity, however limited resources often prevent more elaborated measurement systems. In this study we use the experimental dataset collected during the ADVEX campaign in which turbulent and non-turbulent fluxes were measured in three forest sites by the simultaneous use of five towers/profiles. Our main objectives are to evaluate both the uncertainty of SC that derives from an insufficient sampling of CO2 variability, and its impact on concurrent NEE estimates.Results show that different measurement methods may produce substantially different SC flux estimates which in some cases involve a significant underestimation of the actual SC at a half-hourly time scales. A proper measuring system, that uses a single vertical profile of which the CO2 sampled at 3 points (the two closest to the ground and the one at the lower fringe of the canopy layer) is averaged with CO2 sampled at a certain distance and at the same height, improves the horizontal representativeness and reduces this (proportional) bias to 2–10% in such ecosystems. While the effect of this error is minor on long term NEE estimates, it can produce significant uncertainty on half-hourly NEE fluxes.
|Research areas and keywords||
Subject classification (UKÄ) – MANDATORY
|Number of pages||12|
|Journal||Agricultural and Forest Meteorology|
|State||Published - 2018 Jan 15|