Meteorological Modulation of Diurnal PM2.5 Variability: Performance of Low-Cost Sensors at an Urban Background Site in Mumbai

Mobile monitoring platforms equipped with low-cost sensors (LCS) are increasingly used to enhance the spatial coverage of air quality monitoring networks. In this study, we systematically evaluated the performance of ATMOS sensors for measuring PM₂.₅ by assessing their accuracy, precision and agreement with a reference-grade Beta Attenuation Monitor (BAM). Six identical ATMOS units were co-located and operated continuously for 27 days during the winter season at a Continuous Ambient Air Quality Monitoring Station (CAAQMS) situated at an urban background site within the IIT Bombay campus. The site is influenced by nearby traffic emissions and a lake, representing complex urban micro-environment. This study investigated the role of meteorological conditions in modulating PM₂.₅ concentration and its measurement by LCS relative to BAM observations. Diurnal variations in temperature and relative humidity recorded by ATMOS sensors showed strong agreement with BAM, yielding Pearson correlation coefficients of 0.89 and 0.96, respectively. In contrast, PM₂.₅ measurements from the LCS exhibited systematic biases with temperature–humidity regimes and between daytime (06:00–18:00 local time) and nighttime (18:00–06:00 local time). During daytime conditions characterized by relative humidity ≤70% and temperatures >20°C, the LCS consistently underestimated PM₂.₅ concentrations compared to BAM. Conversely, nighttime conditions with elevated relative humidity (>70%) and lower temperatures (<20°C) led to overestimation by the LCS. Optimal agreement between the LCS and BAM was observed within a temperature range and relative humidity range of 15–25°C and 30%–60%,respectively, indicating favorable operating conditions for the sensors. Hourly PM₂.₅ distributions from LCS revealed enhanced particulates (100–130 µg/m³) during daytime hours at certain days, coinciding with high relative humidity (>80%). These observations underscore the influence of humidity on PM₂.₅ measurements relative to temperature, likely through hygroscopic particle growth. Overall, the findings demonstrate that low-cost PM₂.₅ sensors can provide robust and consistent measurements under a range of meteorological conditions when their environmental sensitivities are explicitly characterized. These results support the application of LCS for air quality monitoring and exposure assessment, particularly when combined with regime-specific corrections or calibration strategies.