Inverse modeling of CH4 fluxes based on GOSAT and ground-based observations

• Main Authors: Maksyutov S., S. Kim H., Saeki T., Belikov D.A., Ito A., Yoshida Y., Morino I., Yokota T.
• Format: Conference Object
• Language: English
• Published: 2016
• Subjects: Arctic
• Online Access: https://nipr.repo.nii.ac.jp/?action=repository_uri&item_id=11672; http://id.nii.ac.jp/1291/00011619/; https://nipr.repo.nii.ac.jp/?action=repository_action_common_download&item_id=11672&item_no=1&attribute_id=16&file_no=1

Expectations on changes in methane emissions in Arctic due to climate warming are high. In order to better quantify the possible climate feedback, present era CH4 emissions should be understood better as well as their interannual variability. As the data obtained with ground based atmospheric CH4 observations are sparse, it is important to test utility of the observations by satellites, such as GOSAT, AIRS, TES and others. Monthly CH4 fluxes for 43 regions in 2009-2010 are estimated by an inverse model using the GOSAT SWIR Level 2 XCH4 data and ground-based CH4 observations archived at WDCGG. The flux estimates and global distribution of methane concentrations in the atmosphere are prepared for a distribution as the GOSAT Level 4 research product. We used interannually varying CH4 emissions by the GFED and VISIT ecosystem model (Ito and Inatomi, 2012) and interannually repeating the EDGAR CH4 emissions and chemical sink fields prepared by the TransComCH4 project in a forward simulation by the NIES transport model. Monthly scale adjustments were applied to 4 categories of fluxes independently for each region (as in Kim et al., 2011). The inverse problem of optimizing the fluxes was solved with a fixed-lag Kalman smoother (Saeki et al., 2013). We compared the inversion results using the two different datasets to assess the utility of GOSAT XCH4 data in flux estimates and found good fit to the data. Mean residual misfit between simulations and GOSAT data is 5 ppb, which is smaller than difference with TCCON and GOSAT observations. The inversion using groundbased data only estimated larger uncertainty of fluxes over tropical regions, South America and Temperate Asia where the data are sparse. Adding large number of the GOSAT data to the inversion leads to decreasing the uncertainty in Temperate Asia (by 41%), northern South America (26%), Tropical Asia (24%), Europe (23%) and other regions. Monthly mean XCH4 simulated with fluxes estimated using the ground-based data is close to the GOSAT observations in the north ...