Application of Machine Learning for the in-Field Correction of a PM2.5 Low-Cost Sensor Network.

Many low-cost sensors (LCSs) are distributed for air monitoring without any rigorous calibrations. This work applies machine learning with PM2.5 from Taiwan monitoring stations to conduct in-field corrections on a network of 39 PM2.5 LCSs from July 2017 to December 2018. Three candidate models were evaluated: Multiple linear regression (MLR), support vector regression (SVR), and random forest regression (RFR). The model-corrected PM2.5 levels were compared with those of GRIMM-calibrated PM2.5. RFR was superior to MLR and SVR in its correction accuracy and computing efficiency. Compared to SVR, the root mean square errors (RMSEs) of RFR were 35% and 85% lower for the training and validation sets, respectively, and the computational speed was 35 times faster. An RFR with 300 decision trees was chosen as the optimal setting considering both the correction performance and the modeling time. An RFR with a nighttime pattern was established as the optimal correction model, and the RMSEs were 5.9 ± 2.0 µg/m3, reduced from 18.4 ± 6.5 µg/m3 before correction. This is the first work to correct LCSs at locations without monitoring stations, validated using laboratory-calibrated data. Similar models could be established in other countries to greatly enhance the usefulness of their PM2.5 sensor networks.

Laboratory Evaluations of Correction Equations with Multiple Choices for Seed Low-Cost Particle Sensing Devices in Sensor Networks.

To tackle the challenge of the data accuracy issues of low-cost sensors (LCSs), the objective of this work was to obtain robust correction equations to convert LCS signals into data comparable to that of research-grade instruments using side-by-side comparisons. Limited sets of seed LCS devices, after laboratory evaluations, can be installed strategically in areas of interest without official monitoring stations to enable reading adjustments of other uncalibrated LCS devices to enhance the data quality of sensor networks. The robustness of these equations for LCS devices (AS-LUNG with PMS3003 sensor) under a hood and a chamber with two different burnt materials and before and after 1.5 years of field campaigns were evaluated. Correction equations with incense or mosquito coils burning inside a chamber with segmented regressions had a high R2 of 0.999, less than 6.0% variability in the slopes, and a mean RMSE of 1.18 µg/m3 for 0.1-200 µg/m3 of PM2.5, with a slightly higher RMSE for 0.1-400 µg/m3 compared to EDM-180. Similar results were obtained for PM1, with an upper limit of 200 µg/m3. Sensor signals drifted 19-24% after 1.5 years in the field. Practical recommendations are given to obtain equations for Federal-Equivalent-Method-comparable measurements considering variability and cost.

Fusion of Environmental Sensing on PM2.5 and Deep Learning on Vehicle Detecting for Acquiring Roadside PM2.5 Concentration Increments.

Traffic emission is one of the major contributors to urban PM2.5, an important environmental health hazard. Estimating roadside PM2.5 concentration increments (above background levels) due to vehicles would assist in understanding pedestrians' actual exposures. This work combines PM2.5 sensing and vehicle detecting to acquire roadside PM2.5 concentration increments due to vehicles. An automatic traffic analysis system (YOLOv3-tiny-3l) was applied to simultaneously detect and track vehicles with deep learning and traditional optical flow techniques, respectively, from governmental cameras that have low resolutions of only 352 × 240 pixels. Evaluation with 20% of the 2439 manually labeled images from 23 cameras showed that this system has 87% and 84% of the precision and recall rates, respectively, for five types of vehicles, namely, sedan, motorcycle, bus, truck, and trailer. By fusing the research-grade observations from PM2.5 sensors installed at two roadside locations with vehicle counts from the nearby governmental cameras analyzed by YOLOv3-tiny-3l, roadside PM2.5 concentration increments due to on-road sedans were estimated to be 0.0027-0.0050 µg/m3. This practical and low-cost method can be further applied in other countries to assess the impacts of vehicles on roadside PM2.5 concentrations.

Assessing 7-year heat-stress exposures and adaptation strategies for children using a real-time monitoring network in Taiwan.

Wet-bulb globe temperature (WBGT) serves as a suitable heat-stress indicator not only for outdoor workers but also for the general public. However, studies on WBGT exposure among the general population are scarce. This research represents the first attempt to assess WBGT exposure of school-aged children. Utilizing a real-time monitoring network in Taiwan, WBGT exposure of school-aged children (7-15 years) were estimated during May to October from 2016 to 2022. Important determinants and spatiotemporal variability of WBGT levels were explored, with hot spots and peak hours of WBGT identified. Macro- and micro-scale adaptation strategies applicable at schools were also evaluated for their effectiveness in reducing heat stress for students. Results showed that the mean daily maximum WBGT (WBGTmax) was 33.1 ± 3.8 °C at 20 stations across Taiwan but could reach/exceed 36 °C (threshold of the dangerous category) at certain hot spots for 42.3-52.0 % of days between May and October. Local geographic features sometimes outweigh the latitude in explaining the spatial variations. Contrary to temperature, WBGT peaked during 10 am to noon rather than from noon to 1:59 pm in most schools, due to clouds blocking solar radiation in the afternoon. This finding has significant implications for scheduling outdoor physical classes/activities to reduce children's heat-health risks. Setting up on-site WBGT monitoring on surfaces that children mostly encounter at schools or utilizing data from nearby weather stations could provide a near real-time heat-health warning. Moreover, providing shades outdoors, relocating outdoor classes indoors, and using air-conditioning would reduce WBGT by 2.1-5.8, 3.7-7.3, and 2.5-5.9 °C, respectively; and would significantly decrease the percentages of WBGT ≥34 °C, which is associated with increased heat-related emergency visits among children in Taiwan. The methodology applied serves as a useful reference for assessing WBGT exposure and adaptation strategies, providing the scientific foundation for heat-health adaptation measures.

High-sensitivity analysis of six synthetic musks by ultra-performance liquid chromatography-atmospheric pressure photoionization-tandem mass spectrometry.

The occurrence and potential toxicity of synthetic musks in environmental media have aroused concerns for the impacts of these chemicals on ecosystems and human health. Here, we present the first method using ultra-performance liquid chromatography-atmospheric pressure photoionization-tandem mass spectrometry (UPLC-APPI-MS/MS) for analysis of the six most important synthetic musks. Analysis was performed on an API 3000 triple quadrupole equipped with a PhotoSpray APPI source. Two pairs of precursor/product ions are presented that are essential for confirmation. Chromatographic separation is completed in 7 min in the positive mode and 5.1 min in the negative mode. Furthermore, three dopant solutions (0.5% anisole in toluene, 0.5% 2,4-difluoroanisole (DFA) in bromobenzene, and 0.5% DFA in chlorobenzene) are compared in terms of sensitivity, linearity, precision, and accuracy. The best sensitivity is associated with 0.5% anisole in toluene as the dopant; all LODs are below 6 pg. The linear range is 5 to 500 ppb with fairly good precision and accuracy. This analytical method has also proven its applicability by analyzing real air samples. In summary, we present a fast, sensitive, and reliable UPLC-APPI-MS/MS method for six important synthetic musks; it can be readily applied to environmental studies.

A community-based study on associations between PM2.5 and PM1 exposure and heart rate variability using wearable low-cost sensing devices.

Few studies have investigated the effect of personal PM2.5 and PM1 exposures on heart rate variability (HRV) for a community-based population, especially in Asia. This study evaluates the effects of personal PM2.5 and PM1 exposure on HRV during two seasons for 35 healthy adults living in an urban community in Taiwan. The low-cost sensing (LCS) devices were used to monitor the PM levels and HRV, respectively, for two consecutive days. The mean PM2.5 and PM1 concentrations were 13.7 ± 11.4 and 12.7 ± 10.5 µg/m3 (mean ± standard deviation), respectively. Incense burning was the source that contributed most to the PM2.5 and PM1 concentrations, around 9.2 µg/m3, while environmental tobacco smoke exposure had the greatest impacts on HRV indices, being associated with the highest decrease of 20.2% for high-frequency power (HF). The results indicate that an increase in PM2.5 concentrations of one interquartile range (8.7 µg/m3) was associated with a change of -1.92% in HF and 1.60% in ratio of LF to HF power (LF/HF). Impacts on HRV for PM1 were similar to those for PM2.5. An increase in PM1 concentrations of one interquartile range (8.7 µg/m3) was associated with a change of -0.645% in SDNN, -1.82% in HF and 1.54% in LF/HF. Stronger immediate and lag effects of PM2.5 exposure on HRV were observed in overweight/obese subjects (body mass index (BMI) ≥24 kg/m2) compared to the normal-weight group (BMI <24 kg/m2). These results indicate that even low-level PM concentrations can still cause changes in HRV, especially for the overweight/obese population.

A versatile low-cost sensing device for assessing PM2.5 spatiotemporal variation and quantifying source contribution.

This study evaluated a newly developed sensing device, AS-LUNG-O, against a research-grade GRIMM in laboratory and ambient conditions and used AS-LUNG-O to assess PM2.5 spatiotemporal variations at street levels of an Asian mountain community, which represented residents' exposure (at the interface of atmosphere and human bodies leading to potential health impacts). In laboratory, R2 of 1-min AS-LUNG-O and GRIMM was 0.95 ± 0.04 (n = 64,179 for 40 sets). After conversion with individual correction equations, their correlation in ambient tests was 0.93 ± 0.05, with absolute % difference of only 10 ± 9%. Ten AS-LUNG-O sets were installed at street sites with another one at 10 m above ground on July 1-28 and December 2-31, 2017 in Nantou, Taiwan. Important source contributions to PM2.5 were quantified with regression analysis. Temporal variation expressed as the daily max/mean of 5-min PM2.5 reached 13.7 in July and 12.2 in December. Spatial variation expressed as the percent coefficients of variance (%CV) across ten community locations was 22% ± 20% (max: 199%) in July and 19 ± 18% (max: 206%) in December. Incremental contribution from the stop-and-go traffic, market, temple, and fried-chicken vendor to PM2.5 at 3-5 m away were 4.38, 3.90, 2.72, and 1.80 µg/m3, respectively. Significant spatiotemporal variations and community source contributions revealed the importance of assessing neighborhood air quality for public health protection. For long-term air quality monitoring, the percentage of available power and signals of G-sensor provided indicative information of maintenance required. Advantages of low cost (USD 650), small size, light weight, solar power supply, backup data storage, waterproof housing, multiple-sensor flexibility, and high precision and accuracy (after correction) enable AS-LUNG-O to be widely applied in environmental studies.

Correction: Panel study using novel sensing devices to assess associations of PM2.5 with heart rate variability and exposure sources.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Panel study using novel sensing devices to assess associations of PM2.5 with heart rate variability and exposure sources.

BACKGROUND/OBJECTIVE: This work applied a newly developed low-cost sensing (LCS) device (AS-LUNG-P) and a certified medical LCS device (Rooti RX) to assessing PM2.5 impacts on heart rate variability (HRV) and determining important exposure sources, with less inconvenience to subjects. METHODS: Observations using AS-LUNG-P were corrected by side-by-side comparison with GRIMM instruments. Thirty-six nonsmoking healthy subjects aged 20-65 years were wearing AS-LUNG-P and Rooti RX for 2-4 days in both Summer and Winter in Taiwan. RESULTS: PM2.5 exposures were 12.6 ± 8.9 µg/m3. After adjusting for confounding factors using the general additive mixed model, the standard deviations of all normal to normal intervals reduced by 3.68% (95% confidence level (CI) = 3.06-4.29%) and the ratios of low-frequency power to high-frequency power increased by 3.86% (CI = 2.74-4.99%) for an IQR of 10.7 µg/m3 PM2.5, with impacts lasting for 4.5-5 h. The top three exposure sources were environmental tobacco smoke, incense burning, and cooking, contributing PM2.5 increase of 8.53, 5.85, and 3.52 µg/m3, respectively, during 30-min intervals. SIGNIFICANCE: This is a pioneer in demonstrating application of novel LCS devices to assessing close-to-reality PM2.5 exposure and exposure-health relationships. Significant HRV changes were observed in healthy adults even at low PM2.5 levels.

A comparison of PCDD/PCDFs exposure in infants via formula milk or breast milk feeding.

Food is the major source for polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) accumulation in human body. In infant period, breast milk and formula milk are the major food sources. Congener-specific analyses of 17 PCDD/PCDFs were performed on 10 brands of formula milk samples which were milk-based and 37 breast milk samples collected from women living in southern Taiwan. The levels of 17 PCDD/PCDFs in 10 formula milk samples ranged from 0.468 to 0.962 pg WHO-TEQ/g lipid, with a mean value of 0.713+/-0.163 pg WHO-TEQ/g lipid. For the 37 breast milk samples, their PCDD/PCDF levels were 14.7+/-9.36 pg WHO-TEQ/g lipid, with a range between 4.21 and 52.8 pg WHO-TEQ/g lipid. At 12th month of age for infants, average daily intakes (ADI) of PCDD/PCDFs were 2.1 pg WHO-TEQ/kg/day for the formula-feeding infants, and 13 pg WHO-TEQ/kg/day for the breast-feeding infants. The present data may provide useful information for risk-benefit evaluation of formula- and breast-feeding.

Fast analysis of 29 polycyclic aromatic hydrocarbons (PAHs) and nitro-PAHs with ultra-high performance liquid chromatography-atmospheric pressure photoionization-tandem mass spectrometry.

Polycyclic aromatic hydrocarbons (PAHs) and nitro-PAHs are ubiquitous in the environment. Some of them are probable carcinogens and some are source markers. This work presents an ultra-high performance liquid chromatography-atmospheric pressure photoionization-tandem mass spectrometry (UHPLC-APPI-MS/MS) method for simultaneous analysis of 20 PAHs and nine nitro-PAHs. These compounds are separated in 15 minutes in the positive mode and 11 minutes in the negative mode, one half of GC/MS analysis time. Two pairs of precursor/product ions are offered, which is essential for confirmation. This method separates and quantifies benzo[a]pyrene (the most toxic PAHs) and non-priority benzo[e]pyrene (isomers, little toxicity) to avoid overestimation of toxin levels, demonstrating its importance for health-related researches. With 0.5% 2,4-difluoroanisole in chlorobenzene as the dopant, limits of detection of PAHs except acenaphthylene and those of nitro-PAHs except 2-nitrofluoranthene are below 10 pg and 3 pg, respectively, mostly lower than or comparable to those reported using LC-related systems. The responses were linear over two orders of magnitude with fairly good accuracy and precision. Certified reference materials and real aerosol samples were analyzed to demonstrate its applicability. This fast, sensitive, and reliable method is the first UHPLC-APPI-MS/MS method capable of simultaneously analyzing 29 environmentally and toxicologically important PAHs and nitro-PAHs.