Technical note: comparison of manually and automatically evaluated profiles of a PollyXT multiwavelength polarization Raman lidar

Abstract. This technical note presents a comprehensive comparison of manually and automatically analyzed lidar profiles of aerosol optical properties retrieved from an 18-month measurement campaign with a continuously measuring, automated PollyXT multiwavelength polarization Raman lidar in Dushanbe, Tajikistan. The manual analysis was performed using a custom software (known as Verlauf) in multiple analyses steps based on visual inspection of the lidar signals. Its results serve as the reference dataset. The automatic analysis was performed using the PollyNET Processing Chain (PPC) (version 4.0). Retrieval parameters, derived layer-mean values, and seasonal mean and median profiles of aerosol optical properties (backscatter and extinction coefficients, particle depolarization and lidar ratios, and backscatter- and extinction-related Angstrom exponents) of the two datasets were compared. The absolute values of the percentage differences of the seasonal mean and median backscatter coefficient profiles at 1.5 km height are largely below 10 %, except at 532 nm in winter and at 1064 nm in spring and winter. The absolute values of the percentage differences of the seasonal mean and median extinction coefficient profiles at 1.5 km height are largely less than 5 %. The absolute values of the percentage differences of the seasonal mean and median particle depolarization ratios at 1.5 km are largely below 20 % at 355 and below 10 % at 532 nm wavelength. For the most part, these discrepancies are within the measurement uncertainties of the considered quantities despite challenges in the automatic retrieval such as reference height detection, depolarization calibration, and cloud screening. This supports the conclusion that the automatically analyzed profiles of aerosol optical properties are utilizable for common applications, such as extinction statistics, aerosol typing, and retrieval of microphysical and cloud-relevant aerosol properties. However, caution should be exercised when using the particle depolarization ratio at 355 nm and backscatter-related Angstrom exponents, which showed enhanced discrepancies. The results will assist further improvements of the PPC and the implementation of additional features (e.g., optimal estimation methods, retrievals of microphysics). Together with the growing network of PollyXT lidars, the automatic processing chain will remain under constant development with the goal of dissemination of near real-time data on product level with high quality to users and scientific databases.