Fusion of multi-source near-surface CO2 concentration data based on high accuracy surface modeling

Abstract

Under the background of growing greenhouse gas emissions and the resulting global warming, researches about the spatial-temporal variation analysis of the concentration of carbon dioxide in the regional and global scale has become one of the most important topics in the scientific community. Simulating and analyzing the spatial-temporal variation of the carbon dioxide concentration on a global scale under limited observation data has become one of the key problems to be solved in the research field of spatial analysis technology. A new research approach based on high accuracy surface modeling data fusion (HASM-DF) method was proposed in this paper, in which the output of the CO2 concentration of the GEOS-Chem model were taken as driving field, and the observation values of CO2 concentration at ground observation station were taken as accuracy control conditions. The new approach's objective is to fulfill the fusion of the two kinds of CO2 data, and obtain the distribution of CO2 on a global scale with a higher accuracy than the results of GEOS-Chem. Root mean square error (RMSE) was chosen as the basic accuracy index, and the experimental analysis shows that the RMSE of the result of the proposed approach is 1.886 ppm, which is significantly lower than that of the GEOS-Chem's 2.239 ppm. Furthermore, compared with the results created by the interpolation methods used the observation values at stations; the fusion results keep a good spatial heterogeneity similar to the results of GEOS-Chem. This research analyzed the spatial distribution and time series variation of the near-surface CO2 based on the fusion result on a global scale. And it can found that areas such as East Asia, Western North American, Central South America and Central Africa and other region show a relatively high value of the near-surface CO2 concentration. And we also found that the near-surface CO2 concentration changes with season, especially in North America and Eurasia, the near-surface CO2 in summer was significantly lower than winter in these areas.