Seminar: Santiago Botia

The contribution of vegetation to the South American carbon balance is critical for understanding the regional dynamics in net carbon exchange. Of particular interest is the role of the Amazon region as a sink or source of carbon to the atmosphere. Recent evidence indicate a weakening of the Amazon carbon sink, and when taking fires into account, the region represents a source of carbon to the atmosphere. In this study we use a regional atmospheric inversion system together with data from the Amazon Tall Tower Observatory (ATTO) and airborne profiles of CO2, to constrain the total land carbon flux in tropical South America over 2010 to 2018. We use a wide set of prior information for Net Ecosystem Exchange, having process-based models, a data-driven product and a diagnostic model. At the domain-wide scale we find that the atmospheric observations can constrain 64% of the land mass, with uncertainty reductions in most of the Amazon region and the adjacent Cerrado and Caatinga biomes. However, the observational constraint is not yet enough for reporting Amazon-wide estimates with confidence and a large uncertainty still remains (+-0.3 PgC/yr), mostly attributed to the west side of the basin (Amazon-Andes foothills). Consequently, we provide a sub-Amazon estimates highlighting the areas that are better constrained by the atmospheric data. At the continental scale, an emerging sink-source gradient between the Amazon region (sink) and the integrated effect of the Cerrado and Caatinga (source) is found, but the source is located in the boundaries and outside the eastern border of the legal Amazon, which could explain the difference between several top-down estimates. Finally, we emphasize on the importance of the neighboring Cerrado biome for estimating carbon exchange in the Amazon and how it can influence atmospheric-inversion estimates.