A new background method for greenhouse gases flux calculation based in back-trajectories over the Amazon

The large amount of carbon stored in trees and soils of the Amazon rain forest is under pressure from land use as well as climate change. Therefore, various efforts to monitor greenhouse gas exchange between the Amazon forest and the atmosphere are now ongoing, including regular vertical profile (surface to 4.5 km) greenhouse gas measurements across the Amazon. These profile measurements can be used to calculate fluxes to and from the rain forest to the atmosphere at large spatial scales by considering the enhancement or depletion relative to the mole fraction of air entering the Amazon basin from the Atlantic, providing an important diagnostic of the state, changes and sensitivities of the forests. Previous studies have estimated greenhouse gas mole fractions of incoming air ('background') as a weighted mean of mole fractions measured at two background sites, Barbados (Northern Hemisphere) and Ascension (Southern hemisphere) in the Tropical Atlantic, where the weights were based on sulphur hexafluoride (SF6) measured locally (in the Amazon vertical profiles) and at the two background sites. However, this method requires the accuracy and precision of SF6 measurements to be significantly better than 0.1 parts per trillion (picomole mole-1), which is near the limit for the best SF6 measurements and assumes that there are no SF6 sources in the Amazon basin. We therefore present here an alternative method. Instead of using SF6, we use the geographical position of each air-mass back-trajectory when it intersects the limit connecting these two sites to estimate contributions from Barbados versus Ascension. We furthermore extend the approach to include an observation site further south, Cape Point, South Africa. We evaluate our method using CO2 vertical profile measurements at a coastal site in Brazil comparing with values obtained using this method where we find a high correlation (r2 = 0.77). Similarly, we obtain good agreement for CO2 background when comparing our results with those based on SF6, for the period 2010-2011 when the SF6 measurements had excellent precision and accuracy. We also found high correspondence between the methods for background values of CO, N2O and CH4. Finally, flux estimates based on our new method agree well with the CO2 flux estimates for 2010 and 2011 estimated using the SF6-based method. Together, our findings suggest that our trajectory-based method is a robust new way to derive background air concentrations for the purpose of greenhouse gas flux estimation using vertical profile data.