Abstract. Delhi is among the most polluted megacities in the world. Despite a range of interventions, the city’s PM_2.5 and PM₁₀ levels exceed Indian and WHO standards several times. India launched the ambitious National Clean Air Programme (NCAP) to reduce air pollution in its most polluted cities, including Delhi, in 2019. While several studies have looked at the trends of pollutant concentrations in Delhi, very few have adjusted for the effects of meteorology. In this study, we perform weather normalisation or deweathering using a machine learning model to analyse the impact of meteorology and anthropogenic emissions on PM_2.5 and PM₁₀ concentrations in Delhi. The study reveals a statistically insignificant decline in deweathered PM_2.5 and PM₁₀ between 2019 and 2024 across seasons. Also, the average deweathered PM_2.5 and PM₁₀ concentrations almost double (PM_2.5: 64.8 μg/m³ vs 144.3 μg/m³ and 142.4; PM₁₀: 138.1 μg/m³ vs 261.8 μg/m³ and 273.1 μg/m³, respectively) in post-monsoon and winter months (October to November and December to February, respectively) compared to those in monsoon (June to September). Meteorological effects reduce PM_2.5 concentrations in summer (March to May) and monsoon seasons by 3.4 μg/m³ and 25.4 μg/m³, respectively, on average. However, they worsen PM₁₀ concentrations during summer by 7.7 μg/m³, but reduce them by 25.4 μg/m³ during the monsoon season, on average. They also worsen the concentrations of PM_2.5 and PM₁₀ by 10.7 μg/m³ and 18.2 μg/m³ respectively, in post-monsoon. Meteorological effects play a role in reducing PM₁₀ concentrations by 7.8 μg/m³ and increasing PM_2.5 by 17 μg/m³ in winter. Additionally, the effect of meteorology shows no statistically significant trends across years for both pollutants, across seasons. Weekly averaged deweathered PM_2.5 – CO ratio analysis between 2019 and 2024 reveals two distinct spikes in the ratio associated with the post-monsoon stubble burning in the states surrounding Delhi and biomass burning for heating during winter months.