We use the Stochastic Time Inverted Lagrangian Transport (STILT) model, a Lagrangian Particle Dispersion Model (LPDM), driven by BRAMS and ECMWF mesoscale meteorological fields to simulate carbon dioxide (CO2) measurements throughout the Brazilian Amazon basin. Flask measurements of CO and SF6 are used, in conjunction with NOAA station data to create boundary conditions for our STILT simulations. The STILT-BRAMS/ECMWF simulations are convolved with the Vegetation Photosynthesis Respiration Model (VPRM) to determine a priori CO2 fluxes for the Basin. The VPRM is a prognostic biosphere model driven by BRAMS or ECMWF shortwave radiation and temperature along with MODIS 8-day EVI and LSWI indices to calculate hourly CO2 fluxes across the Basin. The VPRM is optimized for 8 different vegetation types using eddy covariance tower observations from the Large-scale Biosphere Atmosphere-Ecology (LBA-Eco) program. We then used airborne measurements from the Balanco Atmosferico Regional de Carbono na Amazonia (BARCA) campaigns in November 2008 and May 2009 to constrain a Bayesian inversion from which we can estimate a basin-wide CO2 budget for the time period of the BARCA campaign.