Drone measurements reveal high near-surface urban haze

Abstract Post-winter haze events in Delhi, India, comprise great air quality challenges, yet remain poorly understood due to limited measurements of vertical profiles of particulate matter (PM) concentrations. This study employs a drone-mounted PM low-cost sensor (PM-LCS) with an optimized sampling system to capture vertical PM 2.5 profiles during March 2021. Elevated PM 2.5 concentrations (160 µg/m 3 ) were observed at an altitude of 100 m, being 60% higher than ground level. Vertical profiles of the PM 1 /PM 2.5 ratio under humid conditions (RH > 70%), showed that haze formation is likely driven by hygroscopic inorganic aerosols. Comparison with model simulations showed significant underestimation of PM 2.5 (−52.6 ± 5.5%) during morning haze episodes, coinciding with a dry bias in modeled RH (−30.1 ± 8.3%). During non-hazy episodes, PM 2.5 underestimation decreased to 10.8 ± 1.2% with a minimal RH bias. This suggests that the dry bias of the model limits its ability to simulate aerosol hygroscopic growth. Overall, our findings demonstrate that drone-mounted PM-LCS provides a valuable vertical air quality assessment tool.