Spatio-temporal statistical assessment of anthropogenic CO2 emissions from satellite data

The analysis of sources and sinks of CO2 is a dominant topic in diverse research fields and in political debates these days. The threat of climate change fosters the research efforts in the natural sciences in order to quantify the carbon sequestration potential of the terrestrial ecosystem and CO2 mitigation negotiations strengthens the need for a transparent, consistent and verifiable Moni- toring, Verification and Reporting infrastructure. This paper provides a spatio-temporal statistical modeling framework, which allows for a quantification of the Net Ecosystem Production and of anthropogenic sources, based on satellite data for surface CO2 concentrations and source and sink connected covariates. Using spatial and temporal latent random effects, that act as space-time varying coefficients, the complex dependence structure can be modeled adequately. Finally, spatio-temporal smoothed estimates for the sources and sinks can be used to provide dynamic maps on 0.5 × 0.5 grid for the Eurasien area in intervals of 16 days between September 2009 and August 2012. Finally, the self-reported CO2 emissions within the UNFCCC can be compared with the model results.