MOPITT measurement shows a clear seasonal pattern during the whole year, while the pattern for SCIAMACHY measurement can not be detected. The spatial correlation of annual average between the two measurements is relatively strong in central China among the three regions. The spatial distribution of their annual average is quite similar based on the variograms plotting.
2.Measurements of CO & CO2
Seasonal patterns of MOPITT CO measurement and SCIAMACHY CO2 measurement are quite different, with the former decreasing from spring and latter decreasing from late summer. Spatial correlation of annual average between the two is rather low. However, the variograms pattern of them are relatively similar in both North Africa and central China.
3.Measurement of CO/CO2 & GFED
The seasonal pattern of GFED data (fire emission) and according gas measurement do not agree well in central US and central China. However, the seasonal patterns of the two are quite similar in North Africa, with both showing high value in spring and winter, probably due to biomass burning. Also, relatively strong correlation between MOPITT CO measurement and GFED CO emission is detected in North Africa during spring and winter, but the correlation between CO2 measurement and GFED CO2 emission data is weak.
4.Measurement of CO/CO2 & EDGAR
Although EDGAR data (anthropogenic CO2 emission) provides anthropogenic CO2 emission other than CO, correlation between EDGAR data and measurement of CO is stronger than that between EDGAR and CO2 measurement in all three regions. Meanwhile, the correlation is weaker in North Africa than the other two regions, which are more industrialized. For the spatial distribution detected using variograms and cross-variograms, again, measurement of CO shows closer relationship to EDGAR data than CO2.
Conclusion
Based on the discussions above, CO measurement, especially MOPITT product, has a relatively tighter relationship with GFED (fire emission) data in North Africa (less industrialized area), as well as with EDGAR (anthropogenic CO2 emission) data in central US and central China than CO2, although the correlations found are rather weak. Considering the different ratio of fire emission to anthropogenic emission in three regions (fire emission overweights anthropogenic emission in North Africa while anthropogenic emission is the leading gas emission source in central US and central China), CO measurement, especially MOPITT data, should have potential to better describe the contributions of the emission sources to the atmospheric CO2. In order to validate and explore how CO measurement relates to emission datasets, we should turn to chemical transport models, and analyze on a larger temporal and spatial scale.