Reduction of Topography related Biases in OCO-2 V9 Data

NASA's Orbiting Carbon Observatory 2 (OCO-2) has been successfully measuring column-averaged dry-air mole fractions of carbon dioxide (X_CO2) in the Earth's atmosphere for almost four years. Validation of the V8 release identified biases in X_CO2 that are spatially correlated with local topography. These systematic biases are introduced by erroneous values of the prior surface pressure taken from meteorological reanalysis and used in the OCO-2 bias correction. The main cause of this error is a slight mispointing of the observatory, leading to false geolocations of the footprints in the geometric calibration algorithm. We found that in areas with large topographic variability and steep slopes, a slight mispointing in the order of 1/3 of a footprint ( 0.06°) might result in biases of up to 3 ppm in X_CO2. The OCO-2 algorithm team utilized the L2FP pre-screeners for the O₂ A-band and CO₂ bands to derive new pointing offsets. Updated geolocations for the eight footprints were determined by minimizing topographic dP (the difference between the retrieved surface pressure and the a priori surface pressure) and associated X_CO2 errors over remote areas with large topographic variability. A revised bias correction was developed that corrects V8 data for geolocation errors, leading to an updated V9 data product. Here we present a detailed description of the quantification and correction of the new geolocations for V9.