Search for: All records

Abstract Existing motor vehicle pollutant measurement techniques, including those that employ ground-based and multirotor small uncrewed aircraft system (sUAS) methods, can accurately measure traffic-related air pollution (TRAP) concentrations at a single location. However, these techniques often lack the mobility to assess pollutant trends across a large horizontal area. Fixed-wing sUAS represents an alternative instrument platform compared to ground-based systems and multirotor sUAS, as fixed-wing sUASs are able to carry air pollutant monitor payloads, have extended endurance, and offer expansive three-dimensional ranges across a field site. To demonstrate the utility of fixed-wing sUAS for urban TRAP assessment, we conducted two flights using a Super Robust Autonomous Aerial Vehicle–Endurant Nimble (RAAVEN) sUAS [University of Colorado (CU) Boulder] at a large field site adjacent to a major highway in Erie, Colorado. Concentrations of solid particulate matter (PM₁₀) and gas-phase (carbon monoxide) pollutants displayed decay as a function of altitude. During the morning flight, PM₁₀concentrations decreased from 19.0μg m⁻³at ground level to a minimum concentration of 14.3μg m⁻³at 90 m above ground level. During the afternoon flight, concentrations of PM₁₀displayed minimal vertical stratification, ranging from 8.9 at ground level to 10.0μg m⁻³at 45 m above ground level. Similarly, pollutants displayed decreasing concentrations as the horizontal distance from the roadway increased. Concentrations of TRAP may be significantly elevated in the area both above and beyond roadways, which contribute to additional pollutant exposure from on-road pollution sources. This study demonstrated that the general behavior of TRAP in near-road environments and that the use of fixed-wing sUAS are viable option for urban air quality measurements. Significance StatementThis study represents one of the first uses of a fixed-wing small uncrewed aircraft system (sUAS) to assess near-roadside concentrations of traffic-related air pollution (TRAP) in urbanized areas. We found that local meteorology, including local wind and solar radiation, had a substantial influence on the concentrations of common air pollutants, including particulate matter, black carbon, carbon monoxide, and carbon dioxide. Furthermore, we found large-scale spatiotemporal variation in pollutant concentrations as a function of the vertical and horizontal distance from the highway, indicating that diminished spatial variation employed in multirotor sUAS studies may not be sufficient to fully assess TRAP in roadside environments. 

The impacts of wildfires along the wildland urban interface (WUI) on atmospheric particulate concentrations and composition are an understudied source of air pollution exposure. To assess the residual impacts of the 2021 Marshall Fire (Colorado), a wildfire that predominantly burned homes and other human-made materials, on homes within the fire perimeter that escaped the fire, we performed a combination of fine particulate matter (PM2.5) filter sampling and chemical analysis, indoor dust collection and chemical analysis, community scale PurpleAir PM2.5 analysis, and indoor particle number concentration measurements. Following the fire, the chemical speciation of dust collected in smoke-affected homes in the burned zone showed elevated concentrations of the biomass burning marker levoglucosan (medianlevo = 4147 ng g−1), EPA priority toxic polycyclic aromatic hydrocarbons (median Σ16PAH = 1859.3 ng g−1), and metals (median Σ20Metals = 34.6 mg g−1) when compared to samples collected in homes outside of the burn zone 6 months after the fire. As indoor dust particles are often resuspended and can become airborne, the enhanced concentration of hazardous metals and organics within dust samples may pose a threat to human health. Indoor airborne particulate organic carbon (median = 1.91 μg m−3), particulate elemental carbon (median = .02 μg m−3), and quantified semi-volatile organic species in PM2.5 were found in concentrations comparable to ambient air in urban areas across the USA. Particle number and size distribution analysis at a heavily instrumented supersite home located immediately next to the burned area showed indoor particulates in low concentrations (below 10 μg m−3) across various sizes of PM (12 nm–20 μm), but were elevated by resuspension from human activity, including cleaning. 

Abstract. Advances in ambient environmental monitoring technologies are enabling concerned communities and citizens to collect data to better understand their local environment and potential exposures. These mobile, low-cost tools make it possible to collect data with increased temporal and spatial resolution, providing data on a large scale with unprecedented levels of detail. This type of data has the potential to empower people to make personal decisions about their exposure and support the development of local strategies for reducing pollution and improving health outcomes. However, calibration of these low-cost instruments has been a challenge. Often, a sensor package is calibrated via field calibration. This involves colocating the sensor package with a high-quality reference instrument for an extended period and then applying machine learning or other model fitting technique such as multiple linear regression to develop a calibration model for converting raw sensor signals to pollutant concentrations. Although this method helps to correct for the effects of ambient conditions (e.g., temperature) and cross sensitivities with nontarget pollutants, there is a growing body of evidence that calibration models can overfit to a given location or set of environmental conditions on account of the incidental correlation between pollutant levels and environmental conditions, including diurnal cycles. As a result, a sensor package trained at a field site may provide less reliable data when moved, or transferred, to a different location. This is a potential concern for applications seeking to perform monitoring away from regulatory monitoring sites, such as personal mobile monitoring or high-resolution monitoring of a neighborhood. We performed experiments confirming that transferability is indeed a problem and show that it can be improved by collecting data from multiple regulatory sites and building a calibration model that leverages data from a more diverse data set. We deployed three sensor packages to each of three sites with reference monitors (nine packages total) and then rotated the sensor packages through the sites over time. Two sites were in San Diego, CA, with a third outside of Bakersfield, CA, offering varying environmental conditions, general air quality composition, and pollutant concentrations. When compared to prior single-site calibration, the multisite approach exhibits better model transferability for a range of modeling approaches. Our experiments also reveal that random forest is especially prone to overfitting and confirm prior results that transfer is a significant source of both bias and standard error. Linear regression, on the other hand, although it exhibits relatively high error, does not degrade much in transfer. Bias dominated in our experiments, suggesting that transferability might be easily increased by detecting and correcting for bias. Also, given that many monitoring applications involve the deployment of many sensor packages based on the same sensing technology, there is an opportunity to leverage the availability of multiple sensors at multiple sites during calibration to lower the cost of training and better tolerate transfer. We contribute a new neural network architecture model termed split-NN that splits the model into two stages, in which the first stage corrects for sensor-to-sensor variation and the second stage uses the combined data of all the sensors to build a model for a single sensor package. The split-NN modeling approach outperforms multiple linear regression, traditional two- and four-layer neural networks, and random forest models. Depending on the training configuration, compared to random forest the split-NN method reduced error 0 %–11 % for NO2 and 6 %–13 % for O3.