Validation of GOSAT SWIR XCO2 and XCH4 retrieved by PPDF-S method

Abstract:

Photon path length probability density function (PPDF)-based method is an effective algorithm for retrieving columnar concentrations of carbon dioxide and methane (XCO₂ and XCH₄) from Greenhouse gases Observing SATellite (GOSAT) data, even under conditions of high aerosol concentrations. Few studies have validated XCO₂ analyzed using a PPDF-based method, and no report describes validation for XCH₄. In this study, we validated both XCO₂ and XCH₄ data analyzed using an advanced version of PPDF, PPDF-simultaneous (PPDF-S), by comparing the retrieved data with the Total Carbon Column Observing Network (TCCON) data (GGG 2012 and 2014 releases) observed at 11 observation sites. As the first step, our validation procedure was applied to the NIES standard products of XCO₂ and XCH₄ (V02.xx) to confirm its performance. Results show that the biases ±their standard devations, defined by the difference from TCCON data (GGG 2012), were evaluated to be -1.94±1.79 ppm and -7.59±11.8 ppb, respectively, for XCO₂ and XCH₄, using radiance data used for NIES V02.11, and to be -1.63±1.88 ppm and -6.45±12.2 ppb using data for NIES V02.21. Absolute values of these biases are slightly larger than those from Yoshida et al. (2013) reported as -1.56±1.88 ppm and -6.10±12.3 ppb for XCO₂ and XCH₄, respectively using radiance data used for NIES V02.00. The validation procedure was then applied to XCO₂ and XCH₄ analyzed using PPDF-S. Results show that biases are 0.16±1.77 ppm and 4.18±14.3 ppb, respectively, for XCO₂ and XCH₄ using radiance used for NIES V02.11. They are 0.27±1.72 ppm and 4.54±14.4 ppb using data for NIES V02.21. We conclude that PPDF-S method exhibits better performance for both XCO₂ and XCH₄ retrieval than for NIES standard products where TCCON (GGG 2012) data are used as validation data, although the standard products might be better for XCH₄ if TCCON (GGG 2014) data are used.