Uncertainty Analysis And Algorithm Optimization Of Aerosol Measurements Using Twin Scanning Lidars

Aerosol detection method based on twin scanning lidars can be used to finely detect the optical properties of the horizontal inhomogeneous aerosol in the lower atmosphere, which uses scanning lidar as the remote sensing tool and combines with cross scan detection mode. The method needs to coordinate and mesh the cross scanning area, and finally the two unknown parameters of extinction and backscatter coefficient can be solved by the way of scanning lidar equations. The proposed novel method can achieve the full-field profile cross-probing in the same airspace near surface aerosol.In this paper, the wok is carried out from the uncertain sources of aerosol detection results using twin scanning lidars, and the main contents have two aspects, which one aspect analyzes system errors caused by the lidar system configuration and detection mode, and another aspect analyzes the influences of inversion algorithm for aerosol optical properties,however the emphasis of the paper focus on the uncertainties of the algorithm optimization. Aiming at the influence of the grid resolution on the accuracy of the data retrieval, the reasonable method is proposed to determine the grid resolution by considering the elements of the pixel element and the spot diameter. The high accuracy relative system constant is known in the process of solving the unknown quantities, and then the optimization algorithm is proposed to solve relative system constant by the way of the multi-angle method and multiple lidar ratios. On the integral formula of atmospheric optical depths, the idea of optical path library is proposed,which the new method requires the extinction coefficient matrix and the optical path matrix,finally it is achieved that computing efficiency can be improved. To obtain two unknowns in the S-function, the Monte Carlo method is put forward to estimate the atmospheric extinction and backscatter coefficients in the whole scanning region.To verify the feasibility of the optimized algorithms,the article applies the data of long-range scanning lidar at the same time and the different times, and then the method determines that the optimal resolution of the grid pixel unit is 50m and the relative system constant is 29.92±0.43. The retrieved results show that the new method can obtain aerosol optical parameters and aerosol concentration is stratified in the cross region, and the results show that the Monte Carlo method and multi-angle method are comparable, and the new method shows that atmosphere is inhomogeneous in the horizontal.Finally, this paper analyzes the uncertainties of twin scanning lidars detection results which the uncertainties include aerosol extinction coefficients and backscatter coefficients using of the scanning data at the different times. The number of simulation has a great influence on the uncertainties of atmospheric optical depth and backscatter coefficient, and the uncertainties of extinction coefficient are greatly affected by the uncertainties of relative system constant and lidar ratio. Statistical results of lidar ratio in the cross region show that the lidar ratios are mainly distributed between 30Sr and 50Sr, which are in agreement with those results by local lidar ratio observations.