1Institute of Environmental Physics, University of Bremen, Bremen, Germany
2University of Leicester, Department of Physics and Astronomy, Leicester, UK
3National Institute for Environmental Studies, Tsukuba, Japan
4IMK-ASF, Karlsruhe Institute of Technology, Karlsruhe, Germany
5Colorado State University, Fort Collins, CO, USA
6Netherlands Institute for Space Research, Utrecht, The Netherlands
7Climate Service Center, Helmholtz Zentrum Geesthacht, Hamburg, Germany
Abstract. We analyze an ensemble of seven XCO2 retrieval algorithms for SCIAMACHY (scanning imaging absorption spectrometer of atmospheric chartography) and GOSAT (greenhouse gases observing satellite). The ensemble spread can be interpreted as regional uncertainty and can help to identify locations for new TCCON (total carbon column observing network) validation sites. Additionally, we introduce the ensemble median algorithm EMMA combining individual soundings of the seven algorithms into one new data set. The ensemble takes advantage of the algorithms' independent developments. We find ensemble spreads being often < 1 ppm but rising up to 2 ppm especially in the tropics and East Asia. On the basis of gridded monthly averages, we compare EMMA and all individual algorithms with TCCON and CarbonTracker model results (potential outliers, north/south gradient, seasonal (peak-to-peak) amplitude, standard deviation of the difference). Our findings show that EMMA is a promising candidate for inverse modeling studies. Compared to CarbonTracker, the satellite retrievals find consistently larger north/south gradients (by 0.3–0.9 ppm) and seasonal amplitudes (by 1.5–2.0 ppm).