Three ways to validate the XCO2 product from the Orbiting Carbon Observatory-2

Carbon dioxide (CO₂) is the primary greenhouse gas emitted into the atmosphere due to anthropogenic activities. While it is naturally present as a part of Earth’s carbon cycle, human activities impact the ability of natural sinks to reduce CO₂ from the atmosphere and thus alter the carbon cycle. It thus becomes pertinent to focus on the long-term monitoring of atmospheric CO₂ and the ability to make precise, accurate, and continuous CO₂ measurements.

The Orbiting Carbon Observatory-2 (OCO-2) was launched in 2014. It is NASA’s first Earth-orbiting satellite dedicated to making observations of CO₂ in the atmosphere and measuring its column-averaged dry-air mole fraction (X_CO2). The primary goal of the OCO-2 mission is to provide X_CO2 measurements with sufficient precision and accuracy alongside quantifying its seasonal and interannual variability. While OCO-2 provides global coverage and consistently measures at high latitudes, the remote sensing measurement of CO₂ from space can be challenging. This is because the goal is to resolve inter-annual CO₂ deviations to subcontinental scales, alongside capturing the known seasonal cycle and trends. Further, the X_CO2 data is susceptible to location- and surface-property-dependent biases that must be corrected. Thus, validation of the X_CO2 data from OCO-2 becomes necessary to ensure a high degree of retrieval accuracy on a global scale.

This study uses the new and improved OCO-2 V11.1 dataset and compares coincident X_CO2 measurements against three independent datasets. The Total Carbon Column Observing Network (TCCON) is a network of solar-viewing ground-based Fourier Transform Spectrometers and the primary validation source for X_CO2 from OCO-2. TCCON measurements are unaffected by surface properties and are minimally sensitive to aerosols. The COllaborative Carbon Column Observing Network (COCCON) is a network of portable ground-based Fourier Transform Infrared spectrometers that are less expensive than full TCCON sites and have lower spectral resolution, but are similarly insensitive to surface properties and aerosols. Comparison of coincident OCO-2 measurements against selected TCCON and COCCON sites indicate that the absolute average bias values are close to 0 ppm for TCCON and less than 0.7 ppm for COCCON in the Land Nadir/Glint and Target mode observations.

Finally, we compare coincident OCO-2 measurements to the airborne Atmospheric Tomography Mission (ATom) when ATom conducted around-the-world flights in each of the four seasons between 2016 and 2018. This study bridges the gap between satellite, ground-based, and airborne X_CO2 measurements and aids the improvement of the OCO-2 X_CO2 data product. Further, it provides the latest validation analysis for OCO-2, providing the most up-to-date information on biases and uncertainty in the OCO-2 data.