High-resolution profiling of aerosol absorption coefficients using dual-polarization lidar and a PSO–GPR machine learning approach

The aerosol absorption coefficient (AAC) is of great importance to quantify aerosol radiative forcing and associated impacts on the climate. However, reliable vertical profiles of AAC remain limited, as most existing methods are restricted to surface observations or columnar estimates. This study develops a new approach to retrieving vertical profiles of AAC from dual-polarization lidar measurements, which can simultaneously detect polarization ratios both at 355 nm and 532 nm. The approach integrates a particle swarm optimization-Gaussian process regression (PSO-GPR) scheme with in-situ absorption observations collected in Northwest China. The lidar-retrieved AAC values near the surface show good agreement with co-located Aethalometer measurements. Moreover, comparisons with tethered balloon observations indicate a correlation coefficient of 0.74 for vertical profiles, confirming the robustness of the approach. This work presents an effective and practical method for obtaining reliable, high-resolution AAC profiles from remote sensing, offering new opportunities for evaluating aerosol radiative impacts with improved accuracy.