Mobility-driven changes in urban air quality during the COVID-19 lockdown: Evidence from the Baridhara monitoring station, Dhaka, Bangladesh
PLOS Climate · 20.11.2026 · 5 (11) · e0001033
DOI: 10.1371/journal.pclm.0001033 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS Climate, online erschienen am 20.11.2026. Verfasst von Maheen Zaman und 2 weiteren Autor:innen. Die Arbeit nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Themen: Mobilisation
Abstract (Originalsprache)
The COVID-19 lockdown created a natural experiment for examining how large changes in human activity coincided with urban air-quality change. This study evaluated daily Air Quality Index (AQI) and PM2.5 measured at the U.S. Embassy Baridhara monitoring station in northern Dhaka, Bangladesh, using a 2016–2022 air-quality record, Bangladesh-level Google Community Mobility Reports, and NASA POWER meteorological variables. Because the mobility series were available only at national scale, they were treated as contextual proxies for broad activity change rather than as direct measurements of mobility within Dhaka. Four complementary approaches were applied: a SARIMAX interrupted time-series model, partial wavelet coherence, artificial-neural-network weather normalization selected from eight machine-learning models by five-fold time-series cross-validation, and eight mobility-prediction models interpreted with SHAP. The interrupted time series identified a large immediate decline at lockdown onset (AQI −90.3 units; PM2.5 −110.0 µg m − 3; p < 0.001; adjusted R2 = 0.80 and 0.87). After meteorological adjustment, mobility–pollution coherence remained moderate, with the strongest bands at 8–16 days for AQI and 2–4 days for PM2.5. A direct correlation analysis, performed to complement the normalization, showed that the meteorology–pollution relationships of the lockdown months were indistinguishable in sign and magnitude from those of the pre-pandemic and post-lockdown March–June windows. Random Forest best predicted weather-normalized AQI (R2 = 0.60), while XGBoost best predicted weather-normalized PM2.5 (R2 = 0.66); confidence intervals overlapped among leading ensembles. SHAP indicated that residential, grocery and pharmacy, workplace, and transit mobility were the most influential predictors in the fitted models. The findings show a temporary, weather-independent association between broad mobility change and air quality at one monitoring site, but they do not establish a city-wide or causal effect. The results support transport-demand and emissions-management strategies that can deliver local air-quality benefits while complementing climate-mitigation goals.
Zitieren
APA
Zaman, M., Khan, M. J., & Zaman, M. A. (2026). Mobility-driven changes in urban air quality during the COVID-19 lockdown: Evidence from the Baridhara monitoring station, Dhaka, Bangladesh. PLOS Climate, 5(11), e0001033. https://doi.org/10.1371/journal.pclm.0001033
Vancouver
Zaman M, Khan MJ, Zaman MA. Mobility-driven changes in urban air quality during the COVID-19 lockdown: Evidence from the Baridhara monitoring station, Dhaka, Bangladesh. PLOS Climate. 2026;5(11):e0001033. doi:10.1371/journal.pclm.0001033
BibTeX
@article{Zaman2026,
author = {Zaman, Maheen and Khan, Md. Jobayer and Zaman, Md. Ahnaf},
title = {Mobility-driven changes in urban air quality during the COVID-19 lockdown: Evidence from the Baridhara monitoring station, Dhaka, Bangladesh},
journal = {PLOS Climate},
year = {2026},
volume = {5},
number = {11},
pages = {e0001033},
doi = {10.1371/journal.pclm.0001033}
}
EUVIMED