- 1Sri Ramachandra Institute of Higher Education and Research, Faculty of Public Health, Environmental Health Engineering, Chennai, India (naveen@ehe.org.in)
- 2Department of Biostatistics, St.John's Medical College, Bengaluru, India (santu.g@stjohns.in)
- 3Berkeley Public Health, University of California, Berkeley, USA (ajaypillarisetti@berkeley.edu)
Personal exposure to fine particulate matter (i.e., PM2.5) istypically measured for 24 or 48 hours in health effects research. Real-time, low-cost sensors (LCS) offer long-term PM monitoring solutions with potentially high spatiotemporal resolution that can facilitate better exposure – response analysis. We assessed long-term exposures to PM2.5 among pregnant women in the Tamil Nadu Air Pollution and Health Effects-II (TAPHE-II) and Reproductive effects from Exposure to Airborne Chemicals in urban Homes (REACH) cohorts using such real-time LCS.
Battery-operated real-time PM sensors equipped with PMS7003™ (Plantower Inc., China) were used to monitor living-room PM2.5 levels for a period of 21-days and 7-months in the TAPHE-II (n=80) and REACH (n=15) cohort homes, respectively. Further, pregnant women wore a portable ultrasonic personal air sampler (UPAS™ v2.1) for 24 hours, equipped with a 37-mm PTFE filter, to measure PM mass concentration. The LCS recorded PM, temperature, and relative humidity at 1-minute time intervals and transmitted data in real-time to the cloud. Sensors were collocated with gravimetric samplers for a period of 24-h on three consecutive days in 25 homes to develop indoor-specific calibration equations. In addition, all sensors were collocated with a reference-grade beta attenuation monitor pre- and post-monitoring period; linear models were used to derive ambient calibration coefficients.
Continuous PM data was monitored on average 21 (SD 3) days; data availability ranged between 97 to 100% across rural and urban homes in the TAPHE-II cohort. Precision across all sensors was satisfactory, with a standard deviation of 2.6 µg/m3 and a coefficient of variation of 15.6%. The normalized root mean square error (NRMSE) for indoor and ambient collocation was 31.6% (r=0.86) and 43.2% (r=0.80), respectively. Correlation (NRMSE) between measured personal daily exposures and 21-day real-time PM2.5 measures was 0.62 (47.9%). Long-term averages (min–max) of indoor PM2.5 levels were high among biomass users (n=20, 49.8 µg/m3 (35.8 – 80.6)), followed by mixed-fuel (n=14, 28.7 µg/m3 (29.4 – 61.9)), and liquefied petroleum gas (LPG) (n=46, 25.7 µg/m3 (21.3 – 39.5)) users.
We demonstrate the applicability of LCS for long-term indoor PM monitoring to assess health risks associated with indoor air pollution. Indoor-specific calibrations capture the true range of PM exposures and temporal variability, minimizing uncertainty in exposure – response relationships in health effects research. To more accurately assess the exposure of urban pregnant women, we are using LCS to measure indoor PM throughout most of the gestational period (i.e., up to 7 months). This data will be used to evaluate the representativeness of the 24-hour and 21-day average PM2.5 levels as a proxy of gestational exposure.
How to cite: Puttaswamy, N., Vackacherla, S., Ghosh, S., Saidam, S., Sendhil, S., Jayakumar, D., Patil, S., Pillarisetti, A., and Balakrishnan, K.: Quantifying long-term exposures to fine particulate matter (PM2.5) using real-time, low-cost sensors to assess the impact of household air pollution on birthweight in two cohort studies in southern India , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3891, https://doi.org/10.5194/egusphere-egu25-3891, 2025.
