Jump to Main Content
Global mapping of greenhouse gases retrieved from GOSAT Level 2 products by using a kriging method
- Watanabe, Hiroshi, Hayashi, Kenji, Saeki, Tazu, Maksyutov, Shamil, Nasuno, Isao, Shimono, Yusuke, Hirose, Yoshiyasu, Takaichi, Kazuyosi, Kanekon, Sayaka, Ajiro, Masataka, Matsumoto, Yukio, Yokota, Tatsuya
- International journal of remote sensing 2015 v.36 no.6 pp. 1509-1528
- carbon, carbon dioxide, greenhouse gases, kriging, methane, models, remote sensing
- Because a synoptic overview facilitates understanding of the temporal and spatial changes in the global distribution of greenhouse gases, we developed a statistical spatial estimation method using kriging. Level 3 (L3) data products for the Greenhouse Gases Observing Satellite (GOSAT) Thermal And Near infrared Sensor for Carbon Observation (TANSO) Fourier Transform Spectrometer (FTS) Short Wave Infrared (SWIR) were generated from column-averaged, dry-air mole fractions of carbon dioxide (XCO ₂) and methane (XCH ₄) TANSO-FTS SWIR Level 2 (L2) products using this method. Although there have been some reports on the use of kriging for analysing GOSAT products, the kriging method used in this research was specifically adapted to the statistical characteristics of GOSAT L2 products. In the context of using data for atmospheric research, spatially interpolated data (GOSAT L3 products) cannot be more accurate than model-simulated global distributions of gas concentrations (GOSAT Level 4B (L4B) products), which are generated using an atmospheric tracer transport model. However, the L3 product takes much less time to generate than the L4B. It would take about a year to produce the L4B after generation of an L2 product. The great advantage of the L3 product is that it gives a comprehensive and reasonable monthly global distribution of gas concentrations with little delay. The L3 product using the kriging method can be generated on a monthly basis by estimating global semi-variogram curves from the L2 products for each month and interpolating spatially within a region with a radius of 1000 km from existing L2 data locations. The main purpose of this paper is to describe the methodology and characteristics of kriging used to generate the GOSAT L3 product, not for strictly scientific use of the estimated values, but for a reasonable global map of gas concentrations derived statistically from the sparsely observed L2 products within a short time frame. The characteristics of this method are compared to XCO ₂ products simulated with an atmospheric tracer transport model. The results show that the method proposed in this study is of practical use for generating L3 products from L2 products.