OCO-2 v9 Data Overview

In late 2017 and early 2018 the maturity of the Orbiting Carbon Observatory-2 (OCO-2) total column CO2 (XCO2) measurements with the 2017 v8 release and increased scrutiny from a larger user base returned the attention of the algorithm team to a previously seen but underappreciated aspect of the OCO-2 XCO2 product: an apparent correlation of XCO2 and the local topography. Subsequent analysis by the OCO-2 algorithm team found that the correlation was being introduced primarily via the delta pressure (dP, the retrieved surface pressure minus the apriori surface pressure) term of the bias correction. The erroneous dP values were traced to a small miss-specification of the geographic location of the footprints via instrument-to-spacecraft pointing offsets, which map into dP when there are local topographic features. The OCO-2 algorithm team derived new pointing offsets to minimize topographic dP and associated XCO2 errors. Using the surface pressure apriori from the revised footprint locations allows v8 to be corrected without rerunning the entire retrieval. The hybrid product which uses v8 retrieval results with a revised bias correction derived using the updated surface pressure apriori is labelled v9. Here we show the differences between v8 and v9 on global and regional scales along with some examples of specific locations where v9 shows large improvements. The primary effect of v9 is to reduce the scatter of the XCO2 values over areas of land that feature steep topography.