Development of match-up tool of satellite and ground-based greenhouse gases data for GOSAT–2 data validation.

Abstract:

Many satellite data have been compared to ground-based data, which have higher accuracy and precision. Satellite and ground-based instruments generally observe target quantity with different spatial scale/position and different observation time, therefore, it is important to find more plausible coincidence criteria between them. In the case of validating GOSAT data with Total Carbon Column Observing Network (TCCON), various methods in order to increase the number of match-up data have been used: “Geophysical” (e.g., Morino et al., 2011), “Meteorological” (Wunch et al., 2011), “Model” (Guerlet et al., 2013), and “Geostatistical” (Nguyen et al., 2014).Geophysical, Meteorological, and Model methods use match-up data within a geometric distance centered at each TCCON site, the same potential temperature field at 700-hPa as a proxy for equivalent latitude for CO₂ gradients (Keppel-Aleks et al., 2011), and the same concentration field predicted or assimilated with atmospheric transport model (e.g., TM5, CarbonTracker, MACC-II), respectively. Geostatistical method decides a range of semivariogram based on kriging. The range indicates implicated resemblance by the covariance function with GOSAT and TCCON site of interest. We developed a match-up tool that implemented above-mentioned four types of co-locating methods for GOSAT-2, to be launched in early 2018, and applied to the GOSAT data. We present comparisons of areas and numbers of match-up data, and bias/correlation of GOSAT data at each TCCON site using our tool. In Meteorological method, we demonstrate the differences among atmospheric reanalysis data of NCEP, ERA Interim, MERRA, and JRA55.