Comparison of Atmospheric Carbon Dioxide Concentrations Based on GOSAT, OCO-2 Observations and Ground-Based TCCON Data

Abstract

Carbon dioxide (CO2) is one of the most significant greenhouse gases, and its concentration and distribution in the atmosphere have always been a research hotspot. To study the temporal and spatial characteristics of atmospheric CO2 globally, it is crucial to evaluate the consistency of observation data from different carbon observation satellites. This study utilizes data from the Total Carbon Column Observing Network (TCCON) to verify the column-averaged dry air mole fractions of atmospheric CO2 (XCO2) retrieved by satellites from October 2014 to May 2016, specifically comparing the XCO2 distributions from the Greenhouse Gases Observing Satellite (GOSAT) and Orbiting Carbon Observatory 2 (OCO-2). Our analysis indicates a strong correlation between the TCCON and both the GOSAT (correlation coefficient of 0.85) and OCO-2 (correlation coefficient of 0.91). Cross-validation further reveals that the measurements of the GOSAT and OCO-2 are highly consistent, with an average deviation and standard deviation of 0.92 ± 1.16 ppm and a correlation coefficient of 0.92. These differences remain stable over time, indicating that the calibration in the data set is reliable. Moreover, monthly averaged time-series and seasonal climatology comparisons were also performed separately over the six continents, i.e., Asia, North America, Europe, Africa, South America, and Oceania. The investigation of monthly XCO2 values across continents highlights greater consistency in Asia, North America, and Oceania (standard deviation from 0.15 to 0.27 ppm) as compared to Europe, South America, and Africa (standard deviation from 0.45 to 0.84 ppm). A seasonal analysis exhibited a high level of consistency in spring (correlation coefficient of 0.97), but lower agreement in summer (correlation coefficient of 0.78), potentially due to cloud cover and aerosol interference. Although some differences exist among the datasets, the overall findings demonstrate a strong correlation between the satellite measurements of XCO2. These results emphasize the importance of continued monitoring and calibration efforts to ensure the accurate assessment and understanding of atmospheric CO2 levels.

Reconstructing global daily XCO2 at 1° × 1° spatial resolution from 2016 to 2019 with multisource satellite observation data

“…This adjustment can effectively increase the consistency of the data and help eliminate the noise and error caused by the small fluctuations of the data. Zheng et al 37 found that the standard error of XCO 2 between TCCON and GOSAT lite products was found to be 1.20 ppm, and the standard error between TCCON and OCO-2 OCO2_L2_Lite_FP_10r was 0.97 ppm. Because the OCO-3 satellite was recently launched and there is no corresponding validation study for the OCO3_L2_Lite_FP_10.4r data, we assume that its standard error is also 0.97 ppm.…”

Validation of Remotely Sensed XCO2 Products With TCCON Observations in East Asia

“…They indicated that Tansat XCO 2 had the highest accuracy with an R 2 of 0.88 and an RMSE of 1.06 ppm, followed by OCO-2 XCO 2 with an R 2 of 0.82 and an RMSE of 1.07 ppm, and ACOS-GOSAT XCO 2 had the lowest accuracy with an R 2 of 0.81 and an RMSE of 1.23 ppm. Zheng et al [25] validated the OCO-2 XCO 2 product from September 2014 to February 2022 and the ACOS-GOSAT XCO 2 product from April 2009 to May 2016 using 20 TCCON stations and found that OCO-2 and ACOS-GOSAT XCO 2 products showed strong correlation with TCCON XCO 2 with R 2 of 0.91 and 0.85, respectively. Karbasi et al [26] validated the GOSAT XCO 2 product retrieved by ACOS, NIES, and SRFP algorithms from 2009 to 2021 using 8 TCCON stations.…”