This project seeks to build a "Real Time, Large-Scale Parallel, Intelligent, Carbon dioxide (CO2) Data Assimilation (DA) System". As part of this system we will integrate three critical software components (data acquisition, modeling, and intelligent analysis) into an operational framework that will be designed to assimilate satellite and ground based observations of atmospheric CO2 concentrations, and provide global estimates of carbon flux at fine spatial and temporal resolutions. The data acquisition component will be able to assimilate large amount of CO2 concentration and related ancillary information from publicly available sources (e.g. web-based resources), ground-based measurement networks, and a multitude of satellites. The modeling component will incorporate options for performing ensemble and/or variational DA within both Bayesian and geostatistical frameworks. Lastly, the intelligent analysis component will examine the spatial patterns and time series of estimated global carbon fluxes to identify anomalies and features of interest. The availability of such an operational, large scale DA system will help in considerably advancing carbon/climate science through access to better and more timely estimates of global carbon flux, which will improve understanding of the processes governing carbon flux, allow for monitoring of carbon emissions, and provide timely policy guidance. The proposed system will be (a) the first scalable, parallel real time system producing grid scale (e.g. 1o×1o) maps of global carbon flux, (b) the first modular DA system with multiple atmospheric transport models and DA methods, and (c) the first DA system with autonomous intelligent agents for identifying anomalies in the CO2 flux maps.