An official website of the United States government
Accurate air pollution monitoring is critical to understand and mitigate the impacts of air pollution on human health and ecosystems. Due to the limited number and geographical coverage of advanced, highly accurate sensors monitoring air pollutants, many low-cost and low-accuracy sensors have been deployed. Calibrating low-cost sensors is essential to fill the geographical gap in sensor coverage. We systematically examined how different machine learning (ML) models and open-source packages could help improve the accuracy of particulate matter (PM) 2.5 data collected by Purple Air sensors. Eleven ML models and five packages were examined. This systematic study found that both models and packages impacted accuracy, while the random training/testing split ratio (e.g., 80/20 vs. 70/30) had minimal impact (0.745% difference for R2). Long Short-Term Memory (LSTM) models trained in RStudio and TensorFlow excelled, with high R2 scores of 0.856 and 0.857 and low Root Mean Squared Errors (RMSEs) of 4.25 µg/m3 and 4.26 µg/m3, respectively. However, LSTM models may be too slow (1.5 h) or computation-intensive for applications with fast response requirements. Tree-boosted models including XGBoost (0.7612, 5.377 µg/m3) in RStudio and Random Forest (RF) (0.7632, 5.366 µg/m3) in TensorFlow offered good performance with shorter training times (<1 min) and may be suitable for such applications. These findings suggest that AI/ML models, particularly LSTM models, can effectively calibrate low-cost sensors to produce precise, localized air quality data. This research is among the most comprehensive studies on AI/ML for air pollutant calibration. We also discussed limitations, applicability to other sensors, and the explanations for good model performances. This research can be adapted to enhance air quality monitoring for public health risk assessments, support broader environmental health initiatives, and inform policy decisions.
more »
« less
Developing a Low-Cost Passive Method for Long-Term Average Levels of Light-Absorbing Carbon Air Pollution in Polluted Indoor Environments
We propose a low-cost passive method for monitoring long-term average levels of light-absorbing carbon air pollution in polluted indoor environments. Building on prior work, the method here estimates the change in reflectance of a passively exposed surface through analysis of digital images. To determine reproducibility and limits of detection, we tested low-cost passive samplers with exposure to kerosene smoke in the laboratory and to environmental pollution in 20 indoor locations. Preliminary results suggest robust reproducibility (r = 0.99) and limits of detection appropriate for longer-term (~1–3 months) monitoring in households that use solid fuels. The results here suggest high precision; further testing involving “gold standard” measurements is needed to investigate accuracy.
more »
« less
- Award ID(s):
- 1737633
- PAR ID:
- 10184552
- Date Published:
- Journal Name:
- Sensors
- Volume:
- 20
- Issue:
- 12
- ISSN:
- 1424-8220
- Page Range / eLocation ID:
- 3417
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Formaldehyde is a known human carcinogen and an important indoor and outdoor air pollutant. However, current strategies for formaldehyde measurement, such as chromatographic and optical techniques, are expensive and labor intensive. Low-cost gas sensors have been emerging to provide effective measurement of air pollutants. In this study, we evaluated eight low-cost electrochemical formaldehyde sensors (SFA30, Sensirion®, Staefa, Switzerland) in the laboratory with a broadband cavity-enhanced absorption spectroscopy as the reference instrument. As a group, the sensors exhibited good linearity of response (R2 > 0.95), low limit of detection (11.3 ± 2.07 ppb), good accuracy (3.96 ± 0.33 ppb and 6.2 ± 0.3% N), acceptable repeatability (3.46% averaged coefficient of variation), reasonably fast response (131–439 s) and moderate inter-sensor variability (0.551 intraclass correlation coefficient) over the formaldehyde concentration range of 0–76 ppb. We also systematically investigated the effects of temperature and relative humidity on sensor response, and the results showed that formaldehyde concentration was the most important contributor to sensor response, followed by temperature, and relative humidity. The results suggest the feasibility of using this low-cost electrochemical sensor to measure formaldehyde concentrations at relevant concentration ranges in indoor and outdoor environments.more » « less
-
Nutrient pollution remains one of the greatest threats to water quality and imposes numerous public health and ecological concerns. Phosphate, the most common form of phosphorus, is one of the key nutrients necessary for plant growth. However, phosphate concentration in water should be carefully monitored for environmental protection requirements. Hence, an easy-to-use, field-deployable, and reliable device is needed to measure phosphate concentrations in the field. In this study, an inexpensive dip strip is developed for the detection of low concentrations of phosphate in water and seawater. In this device, ascorbic acid/antimony reagent was dried on blotting paper, which served as the detection zone, and was followed by a wet chemistry protocol using the molybdenum method. Ammonium molybdate and sulfuric acid were separately stored in liquid form to significantly improve the lifetime of the device and enhance the reproducibility of its performance. The device was tested with deionized water and Sargasso Sea seawater. The limits of detection and quantification for the optimized device using a desktop scanner were 0.134 ppm and 0.472 ppm for phosphate in water and 0.438 ppm and 1.961 ppm in seawater, respectively. The use of the portable infrared lightbox previously developed at our lab improved the limits of detection and quantification by a factor of three and were 0.156 ppm and 0.769 ppm for the Sargasso Sea seawater. The device’s shelf life, storage conditions, and limit of detection are superior to what was previously reported for the paper-based phosphate detection devices.more » « less
-
Abstract The use of air quality monitoring networks to inform urban policies is critical especially where urban populations are exposed to unprecedented levels of air pollution. High costs, however, limit city governments’ ability to deploy reference grade air quality monitors at scale; for instance, only 33 reference grade monitors are available for the entire territory of Delhi, India, spanning 1500 sq km with 15 million residents. In this paper, we describe a high-precision spatio-temporal prediction model that can be used to derive fine-grained pollution maps. We utilize two years of data from a low-cost monitoring network of 28 custom-designed low-cost portable air quality sensors covering a dense region of Delhi. The model uses a combination of message-passing recurrent neural networks combined with conventional spatio-temporal geostatistics models to achieve high predictive accuracy in the face of high data variability and intermittent data availability from low-cost sensors (due to sensor faults, network, and power issues). Using data from reference grade monitors for validation, our spatio-temporal pollution model can make predictions within 1-hour time-windows at 9.4, 10.5, and 9.6% Mean Absolute Percentage Error (MAPE) over our low-cost monitors, reference grade monitors, and the combined monitoring network respectively. These accurate fine-grained pollution sensing maps provide a way forward to build citizen-driven low-cost monitoring systems that detect hazardous urban air quality at fine-grained granularities.more » « less
-
Plastic pollution is a major environmental and health threat due to its widespread presence in ecosystems and food chains. Despite extensive research on microplastics, the detection of submicron plastics remains challenging due to their distinct physical and chemical properties and the limitations of current analytical methods. SERS has attracted significant attention in recent research as an ultra-sensitive approach for detecting nanoplastics compared to other spectroscopy techniques. In this paper, a stable, biodegradable, waste-free novel paper-based SERS substrate is developed for the rapid detection of submicron (200 nm) polystyrene (PS) particles via the controlled deposition of AuNPs onto filter paper using an atmospheric cold plasma jet printing process. The density of AuNPs increases with the number of printing passes, correlating with enhanced SERS results. The resulting SERS substrates are capable of quantifying a broad range of PS concentrations (1–500 μg mL⁻¹) using just 5 μL of analyte. The fabricated SERS substrate enables reliable quantification of PS in water, exhibiting a strong linear correlation (R² = 0.993) between SERS intensity and PS concentration, with a detection limit of 10 μg mL⁻¹ . These substrates demonstrate exceptional stability and reproducibility over a 10-week period, addressing key challenges associated with paper-based SERS substrates and making them suitable for long-term monitoring. Furthermore, analysis of tap water as a representative real-world sample demonstrates the practical applicability of the SERS substrate for environmental monitoring, revealing quantifiable levels of PS contamination.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3390/s20123417
