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

Lung, Shih-Chun Candice; Wang, Wen-Cheng Vincent; Wen, Tzai‐Hung; Chun-hu, Liu; Hu, Shu-Chuan

doi:10.1016/j.scitotenv.2020.137145

Cited by 40 publications

(75 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…LCS devices integrate LCS, power, and data transmission components. AS-LUNG-O is an LCS device integrated by our team and designed for long-term outdoor monitoring for scientific research [ 10 ]. AS stands for Academia Sinica (the research institute that supports its development), while LUNG indicates the human organ most commonly affected by air pollutants, and O indicates the “outdoor” version.…”

Section: Methodsmentioning

confidence: 99%

“…Therefore, such monitors are situated at a height of 10–15 m above the ground. However, the intensive emissions of community pollution sources in Asia, such as restaurants and temples, result in high PM 2.5 levels in the immediate living environments of citizens at street level [ 7 , 8 , 9 , 10 ]. Even living in the same airshed, residents from different communities with different emission sources are exposed to different PM 2.5 levels, demonstrating the need for community air monitoring.…”

Section: Introductionmentioning

confidence: 99%

“…However, these sensors installed by citizens or NGOs are not calibrated, and data accuracy is not assured. Thus, their readings may be 2–3 times the actual concentrations [ 10 ]. Such limitations in data quality have restrained the application of these sensors in research and policy development.…”

Section: Introductionmentioning

confidence: 99%

“…LCSs need to be calibrated, since their readings usually deviate from those of research-grade instruments [ 10 ]. The actual cut-off points of particle sizes for LCSs are not consistent with the statements declared by the manufactures [ 16 ]; individual correction equations are, thus, needed for each sensor, even those from the same batch [ 10 ]. Most environmental scientists evaluate the performance of LCSs with side-by-side comparisons against research-grade instruments in the laboratory or in the field.…”

Section: Introductionmentioning

confidence: 99%

“…Most environmental scientists evaluate the performance of LCSs with side-by-side comparisons against research-grade instruments in the laboratory or in the field. Research-grade instruments, which are considered the golden standard for such comparisons, include instruments with measuring principles based on light scattering, oscillating tapered elements, β-attenuation, and filter-weighing, such as GRIMM instruments [ 10 , 17 , 18 , 19 , 20 ], TSI SidePake AM510 [ 21 , 22 ], tapered element oscillating microbalance [ 18 , 23 ], β-attenuation monitor [ 17 , 18 , 19 , 24 , 25 ], and ultrasonic personal aerosol sampler [ 26 ]. Correction equations are established to adjust the LCS readings to those of research-grade measurements.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

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

Wang

Lung

Liu

2020

Sensors

Self Cite

View full text Add to dashboard Cite

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.

show abstract

Section: Methodsmentioning

confidence: 99%