Evaluating the JAXA GOSAT CO2 retrieval product using NASA CO2 retrieval products and NOAA Carbon Tracker

Abstract

Space-based Greenhouse Gas (GHG) observations done by Japan’s Greenhouse Gas Observing Satellite (GOSAT) and NASA’s Orbiting Carbon Observatory (OCO) missions have collected long term and spatially dense CO2 data globally. The satellite GHG data have contributed to the monitoring of global CO2 concentrations and the detection of their regional and local changes. Given the high stake of the climate and environment applications, the evaluation of the space-based GHG data is a critical task. Space-based GHG data are often compared to data collected at ~ 30 Total Carbon Column Observing Network (TCCON) sites for examining potential biases and errors. Bias-correction methods are often developed based on the comparison to the TCCON data. While satellite GHG data products often show good agreement with TCCON data, we still see regional disagreements among different GHG data. We argue that the TCCON-based evaluation is powerful, but limited, and thus further evaluations of satellite GHG products are necessary. Recently, JAXA developed a new GOSAT GHG product named JAXA/GHG product. Our retrieval product includes total and partial column concentration values of CO2, CH4, H2O as well as solar-induced chlorophyll fluorescence (SIF). The JAXA/GHG GOSAT CO2 product is compared to NASA’s Atmospheric Carbon Observations from Space (ACOS)-GOSAT L2 full physics retrieval data, NASA’s OCO-2 satellite-based L2 full physics retrieval data, as well as simulated CO2 from the Carbon Tracker global atmospheric inversion system developed by NOAA Global Monitoring Laboratory. Our XCO2 comparisons with other satellite products show regional discrepancies over the Pacific Ocean, central Africa, south-east Asia (land), and Amazon areas. The results suggest that these discrepancies could be attributable to retrieved surface pressure and aerosol properties. We also compare near surface and upper tropospheric partial CO2 retrieved values to Carbon Tracker simulated values. These comparisons show the systematic positive discrepancy (~2-3 ppm) in JAXA’s near surface (surface to ~ 4km) CO2 concentration over the oceans against Carbon Tracker.