The impact of data gaps and quality control filtering on the balances of energy and carbon for a Southwest Amazon forest

Fluxes of CO2, water vapor, sensible heat and several heat storage terms were used to characterize the dependence of the energy balance closure and carbon balance on gaps introduced by screening for high quality or turbulent conditions. The current work is the first study where a sensitivity analysis was applied to the carbon balance and energy balance closure for an Amazonian forest site. The measurements were part of the RACCI/DRY-TO-WET Experiment, within the framework of the Large Scale Biosphere-Atmosphere Experiment in Amazonia (LBA). The energy balance closure varied from 88% to 98%, with an average intercept of 5 W m(-2). The level of closure was dependent on the amount of data filtered out according to the quality control flag. A compromise between moderate data quality and good sampling of daytime periods was determined to be the best choice for this site. Different values of friction velocity threshold and data quality used in the screening of data resulted in a carbon balance ranging from -0.10 +/- 0.15 to 0.31 +/- 0.25 t C ha(-1), where positive values indicate a net source of carbon in the period. Applying the screening based on friction velocity and quality control changed this ecosystem in the period analyzed from a sink into a source of carbon to the atmosphere. Year-round continuous measurements would be required to verify if the net release of carbon during this transition period is part of the natural cycle for this forest or if that was caused by disturbances in the ecosystem or local climatology. Our results show that the carbon sink strength associated with this forest site could have been overestimated in previous works due to underestimation of nocturnal respiration, caused by lack of filtering of low turbulence conditions. (C) 2010 Elsevier B.V. All rights reserved.