Constraining CO2 fluxes in the Amazon Basin with a regional aircraft campaign

Background/Question/Methods

The Balanço Atmosférico Regional de Carbono na Amazônia (BARCA) aircraft program consisted of two field campaigns in November 2008 and May 2009. It encompassed the dry to wet and wet to dry transition seasons respectively and provided approximately 150 vertical profiles covering the Basin. The goals of this experiment were to estimate a carbon budget for the Basin and to determine if regional aircraft experiments can provide strong constraints for a budget.

We use a Lagrangian Partical Dispersion Model to integrate satellite data, aircraft data, surface observations, and mesoscale meteorological fields to link bottom-up and top-down diagnostic models to provide constraints and error bounds for regional fluxes. Specifically, the Stochastic Time-Inverted Lagrangian Transport (STILT), model driven by meteorological fields from Brazilian Regional Atmospheric Modeling System (BRAMS) is coupled to a biosphere model, the Vegetation Photosynthesis Respiration Model (VPRM) to determine regional CO₂ fluxes for the Basin. The VPRM is a prognostic biosphere model driven by MODIS 8-day EVI and LSWI indices, shortwave radiation and temperature from tower measurements, and mesoscale meteorological data. The VPRM is tuned using 2002-2006 eddy flux tower data from the Large Scale Biosphere Atmosphere experiment. It computes hourly CO₂fluxes for 8 different vegetation types. A priori fluxes modeled by STILT-VPRM are optimized using data from BARCA, eddy covariance sites, and flask measurements. The aircraft mixing ratios are applied as a top down constraint in an inversion that solves for parameters that control the flux calculation from which we make an estimate of the carbon budget of the Basin.

Results/Conclusions

Preliminary results show that the STILT-VPRM model does a good job of simulating the net emission of CO₂ from both transition periods. There is significant contribution from biomass burning during the November 2008 flight profiles. A CO/CO₂ ratio derived from flask measurements is used to remove the biomass burning contribution bringing the model results inline with the aircraft data. Comparing column calculations for each of the vertical profiles shows our model represents the variability in the diurnal cycle. Our initial inversion shows a simple scalar can be used to adjust the model parameters but a more robust Bayesian inversion is being conducted. The high altitude CO₂ values from above 2750-m are very similar to the lateral boundary conditions indicating little influence from surface fluxes at these levels. Finally, our initial results suggest the Basin has a small net source of CO₂ during the BARCA intensives.