Carbon Balance in the Amazon Basin: Factors Influencing the Accuracy of Eddy Covariance Measurements

To assess the role of Amazonia as a source or a sink for atmospheric CO2, we are using ground-based biometry measurements together with whole-system CO2 fluxes (via eddy covariance) to explore the ecological and climatic controls on the carbon balance at an old-growth Amazon forest (Tapajos National Forest, Santarem, Para, Brazil). Here, we focus on the initial 6 months of eddy covariance results from a new system that measures eddy covariance at two levels (47 m and 58 m) and measures canopy CO2 storage in two ways: by cycling through tower-mounted sample inlets at 8 levels (over 16 minutes), and by making a single integrated measurement of all 8 levels simultaneously. The integrated canopy storage measurement gives a storage flux estimate with much lower noise than the more conventional method of interpolating through separate measurements staggered in time, allowing for improved estimates of net ecosystem exchange (eddy flux + storage flux). This provides a cleaner way to identify nighttime periods when turbulent mixing was weak. We show that such periods of weak mixing clearly cause underestimates in flux, but that a continuous low-noise storage flux measurement can make adjustments for the effects of such periods more robust. After adjusting for periods of weak mixing, this tropical forest site appears to be in approximate carbon balance during this 6-month period -- in agreement with the biometry observations over the first two years that woody biomass increment is in approximate balance with tree mortality.