Research on the influence of satellite instrument index on retrieval accuracy of XCO2 using near infrared bands

With the increasing attention to accurate measurements of the content of atmospheric Carbon dioxide (CO2) and its impact on climate change, satellite observation of XCO2 (the dry air column averaged mixing ratio of CO2) has been primarily used to monitor CO2 source strengths because of its high spatial and temporal resolution, global coverage as well as low cost. The influence of satellite instrument index (including selection of observation bands, spectral resolution and signal to noise ratio (SNR) on XCO2 retrieval accuracy is analyzed for typical atmospheric conditions and imaging geometry using SCIATRAN radiative transfer code. For the selection of observation bands, it is found that the maximum retrieval error appears when the right branch of 1.6 mu m band was used individually (similar to 3.49ppm), while the minimum appears when 1.6 mu m and 2.05 mu m bands were used together(similar to 0.44 ppm). The combination of 2.05 mu m band with whether the left or the right branch of 1.6 mu m band leads to similar retrieval errors. With the decrease of spectral resolution, the retrieval error increases from 0.03ppm to 0.12ppm. While after ten percent uncertainty in Aerosol Optical Depth (AOD) was introduced, the retrieval errors increase for both high and low spectral resolution (similar to 2.73 and 3.42ppm, respectively). For the same SNR error type (negative, positive or random), higher SNR results in better XCO2 retrieval accuracy. Negative systematic error in SNR results in smaller retrieval error as compared to positive systematic error. And random error lies between them. SNR in the range of 300-600 can meet the requirement of the retrieval error smaller than 2ppm.