Orbiting Carbon Observatory-2 (OCO-2): A Progress Report at the End of 2 Years of Operations

The Orbiting Carbon Observatory-2 (OCO-2) is the first NASA satellite designed to measure atmospheric carbon dioxide (CO₂) with the precision, accuracy, resolution, and coverage needed to identify surface sources and sinks on regional scales. Its 3-channel imaging grating spectrometer has been returning almost one million soundings over the sunlit hemisphere each day since 6 September 2014. On monthly time scales over the first two years of operations, between 7 and 21% of these soundings were sufficiently cloud free to yield spatially-resolved, full-column estimates of the column-averaged CO₂ dry air mole fraction, X_CO2 along its narrow (< 10 km) swath. Nadir soundings over land yield X_CO2 estimates with single-sounding random errors that increase from 0.5 ppm to 1 ppm between the sub-solar latitude and solar zenith angles near 60 degrees. Observations of the glint spot over the ocean yield X_CO2 estimates with single sounding random errors near 0.5 ppm at solar zenith angles below 70 degrees.

The OCO-2 glint/nadir observing strategy has been refined over the first two years of operations to improve the measurement coverage and yield.  The initial observing strategy recorded only glint or nadir observations over the entire sunlit hemisphere for a complete, 16-day, ground-track repeat cycle, and then used the other observing mode in the next 16-day cycle. This approach provided adequate coverage of oceans and continents on monthly time scales, but produced 16-day long gaps in the coverage of the ocean while in nadir mode, and limited coverage of high latitude continents while in glint mode. In early July of 2015, this observation strategy was modified to alternate between glint and nadir observations on alternate orbits to yield more continuous coverage of the entire sunlit hemisphere every day.  In November of 2015, the observation strategy was refined further to always collect glint data on orbits that were primarily over ocean.

The OCO-2 team started delivering Version 7 products to the Goddard Earth Sciences Data and Information Services Center (GES-DISC) in early June 2015.  These products included calibrated, spectral radiances (Level 1 products), and retrieved geophysical quantities, including spatially resolved estimates of XCO2, surface pressure, and solar-induced chlorophyll fluorescence (Level 2 products). These X_CO2 estimates have been validated against results from the Total Carbon Column Observing Network (TCCON) and other standards to assess their accuracy and correct regional scale biases. After correction, the median bias between OCO-2 and TCCON X_CO2 estimates is less than 0.5 ppm and root-mean-square (RMS) differences are typically less than 1.5 ppm.  OCO-2 results are also being cross-calibrated and cross-validated with measurements and data products from the Japanese Greenhouse gases Observing SATellite (GOSAT, nicknamed "Ibuki"), so that these two satellite datasets can be combined to produce a uniform X_CO2 climate data record for use by the carbon cycle science community.

Preliminary, global maps of X_CO2 compiled from the OCO-2 Version 7 data product clearly reveal the most robust features of the atmospheric carbon cycle, such as the large buildup of CO₂ in the northern hemisphere during the winter and early spring, followed by the strong CO₂ drawdown in the late spring and early summer. Regions of enhanced X_CO2 that are co-located with intense fossil fuel emission sources in the eastern US and eastern China were most obvious in the fall and early winter, when the north-south gradient in X_CO2 was small. X_CO2 enhancements coincident with intense biomass burning in the Amazon, central Africa, and the Indonesian Archipelago were also most obvious during this season. This dataset also provides a high resolution description of the atmospheric carbon cycle’s response to the strong 2014-2015 El Niño Southern Oscillation (ENSO). As the carbon cycle science community continues to analyze these OCO-2 data, quantitative estimates of regional-scale emission sources and natural sinks (absorbers) are expected to emerge.