Analysis of the horizontal and vertical distribution of a dust weather event in the Tarim Basin based on multi-source observational datasets

Abstract

The study employed multi-source observation data from unmanned aerial vehicles (UAVs), satellites, and LiDAR conduct an observational study on a dust weather event that occurred in the Tarim Basin, China, from May 2 to 4, 2023. The results showed that FY-4A dust storm detection and MODIS aerosol optical depth (AOD) products could effectively observe the horizontal distribution of dust. Dust areas and intensities were identified at the AOD threshold range of 0.54–3.50. The convolutional neural network algorithm dust mask could identify dust structures with more precision compared to traditional FY-4 dust storm detection. Moreover, vertical particulate matter (PM) concentration changes determined by UAVs were analyzed at different altitudes, with low PM concentrations observed at higher altitudes. The dust area obtained through the CALIPSO vertical feature mask product was consistent with the PM concentration changes observed by the UAV. When the visibility value was below 1 km, the ground-based LiDAR 532 nm extinction coefficient (EC), backscatter coefficient (BC), and depolarization ratio (DR) values reached 3.42 km⁻¹, 0.057 km⁻¹sr⁻¹, and 0.47, respectively. The vertical profile changes of EC, BC, and DR were in strong agreement with the vertical profile changes of the PM concentrations by the UAV.