Fig. 1. Kalman smoothing algorithm to process XCO₂ data observed by spaceborne IPDA lidar

Fig. 2. Schematic diagram of extracting XCO₂ pseudo truth sequence. (a) XCO₂ results calculated based on WRF-GHG simulated CO₂ volume fraction; (b) comparison of XCO₂ results before and after interpolation under possible simulated satellite-borne IPDA lidar observation trajectory

Fig. 3. Pseudo observation sequences with different levels of observation errors superimposed

Fig. 4. Schematic diagram of filtering performance of Kalman smoothing algorithm using different state transfer matrices for pseudo-observation sequences with three observation error levels

Fig. 5. Filtering performance of Kalman smoothing algorithms for three state equations. (a) MAE; (b) RMSE; (c) correlation coefficient

Fig. 6. Comparisons of filtering effects between sliding average algorithm and Kalman smoothing algorithm. (a) Standard deviation of observed noise is 6×10^-6; (b) standard deviation of observed noise is 9×10^-6

Fig. 7. Kalman smoothing algorithm and sliding average algorithm to reconstruct XCO₂ observation sequence for point source emission monitoring. (a)(c) Schematic diagrams of simulation results of XCO₂ enhancement caused by point source emissions and location of observed data; (b)(d) comparisons of XCO₂ enhancement calculated based on filtering results and XCO₂ enhancement caused by matched emission rate