Short-term PM2.5 data continuity: Cross-site prediction of portable measurements using a nearby fixed monitoring station
Environmental Pollution · 2026-12 · 411 · 129384
DOI: 10.1016/j.envpol.2026.129384 ↗Kurz gefasst
Fachartikel in der Zeitschrift Environmental Pollution, 2026-12. Verfasst von Tin Saw Pyae und 1 weiteren Autor:innen. Beteiligte Einrichtung: International Program of Hazardous Substances and Environmental Management, College of Interdisciplinary and Integrative Studies, Chulalongkorn Unive…. Hauptthemen laut MEDLINE: Particulate Matter, Environmental Monitoring, Air Pollutants. Die Arbeit nennt 34 Literaturangaben. Gefördert von: Chulalongkorn University Graduate School.
Abstract (Originalsprache)
Short-term portable monitoring can improve the spatial and temporal characterization of urban fine particulate matter (PM2.5), but measurement interruptions may compromise data continuity. This study evaluated whether contemporaneous PM2.5 measurements from a nearby fixed-site monitoring station could provide predictive information for reconstructing isolated missing 6-h DustTrak observations at an urban site in Bangkok, without assuming direct interchangeability between monitoring locations or measurement systems. Six-hour DustTrak monitoring sessions (n = 24) were conducted during the wet and dry seasons from August 2024 to February 2025; complete corresponding observations from the 59T fixed-site station were available for 22 sessions (n = 22). Temporal correspondence and paired concentration differences were evaluated, and direct substitution of 59T concentrations was compared with site-specific regression models incorporating 59T PM2.5 alone or together with meteorological predictors. Predictive performance was assessed primarily using leave-one-session-out cross-validation (LOOCV), with additional temporally structured hold-out validation. DustTrak and 59T concentrations showed very strong temporal correspondence (Spearman's ρ = 0.979, p < 0.001), while the mean paired difference (DustTrak - 59T) was -14.32 μg/m3. Because measurements were obtained at spatially separated sites with different vertical and surrounding microenvironments, this difference was not interpreted as instrument-specific bias. Direct substitution yielded substantial prediction error (LOOCV RMSE = 16.03 μg/m3), whereas the 59T-only regression reduced RMSE to 6.43 μg/m3 (cross-validated R2 = 0.971), a 59.9% reduction. Adding wind direction produced only marginal improvement, while further inclusion of relative humidity and rainfall did not improve LOOCV performance. The 59T-only model also retained good performance in the temporally separated validation subset (R2 = 0.926; RMSE = 5.68 μg/m3). These findings indicate that nearby fixed-site measurements should not be used as direct substitutes for portable observations but may support reconstruction of isolated short-term data gaps through a validated, site-specific predictive relationship. Broader application requires validation across monitoring locations, periods, and independent datasets.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- International Program of Hazardous Substances and Environmental Management, College of Interdisciplinary and Integrative Studies, Chulalongkorn University, Bangkok, 10330, Thailand.
- College of Public Health Sciences, Chulalongkorn University, Bangkok, 10330, Thailand. Electronic address: Kraiwuth.K@chula.ac.th.
Förderung
- Chulalongkorn University Graduate School
Zitieren
APA
Pyae, T. S., & Kallawicha, K. (2026). Short-term PM2.5 data continuity: Cross-site prediction of portable measurements using a nearby fixed monitoring station. Environmental Pollution, 411, 129384. https://doi.org/10.1016/j.envpol.2026.129384
Vancouver
Pyae TS, Kallawicha K. Short-term PM2.5 data continuity: Cross-site prediction of portable measurements using a nearby fixed monitoring station. Environmental Pollution. 2026;411:129384. doi:10.1016/j.envpol.2026.129384
BibTeX
@article{Pyae2026,
author = {Pyae, Tin Saw and Kallawicha, Kraiwuth},
title = {Short-term PM2.5 data continuity: Cross-site prediction of portable measurements using a nearby fixed monitoring station},
journal = {Environmental Pollution},
year = {2026},
volume = {411},
pages = {129384},
doi = {10.1016/j.envpol.2026.129384}
}
EUVIMED