Future Particulate Matter Research Articles

EvalPM: a framework for evaluating machine learning models for particulate matter prediction

Air pollution through particulate matter (PM) is one of the largest threats to human health. To understand the causes of PM pollution and enact suitable countermeasures, reliable predictions of future PM concentrations are required. In the scientific literature, many methods exist for machine learning (ML)-based PM prediction, though their quality is difficult to compare because, among other things, they use different data sets and evaluate the resulting predictions differently. For a new data set, it is not apparent which of the existing prediction methods is best suited. In order to ease the assessment of said models, we present evalPM, a framework to easily create, evaluate, and compare different ML models for immission-based PM prediction. To achieve this, the framework provides flexibility regarding data sets, input features, target variables, model types, hyperparameters, and model evaluation. It has a modular design consisting of several components, each providing at least one required flexibility. The individual capabilities of the framework are demonstrated using 16 different models from the related literature by means of temporal prediction of PM concentrations for four European data sets, showing the capabilities and advantages of the evalPM framework. In doing so, it is shown that the framework allows fast creation and evaluation of ML-based PM prediction models.

Airborne particulate matter in Southeast Asia: a review on variation, chemical compositions and source apportionment

Environmental context Airborne particulate matter (PM) is a major public health risk in Southeast Asia. The annual average concentration of fine PM (PM2.5) in the region is significantly higher than the WHO air quality guidelines, and higher PM2.5 levels were recorded during dry seasons, primarily due to biomass burning. In this paper, we provide an overview of the seasonal variations in concentrations, chemical compositions, and sources of PM in Southeast Asian countries. Abstract Airborne particulate matter (PM) in Southeast Asia is the most important air pollutant, causing millions of premature deaths. This review provides an overview of the levels, chemical compositions and sources of PM and compared these with studies from megacities in other regions. Daily average PM2.5 concentrations were lower than polluted megacities such as Delhi but substantially higher than 24-h mean air quality guideline of the WHO. Levels of PM2.5 in maritime continental Southeast Asia (Indonesia, Brunei, Malaysia, Philippines and Singapore) were lower than in mainland continental countries (Cambodia, Myanmar, Thailand, Vietnam, and Laos). PM levels are usually two times higher during dry than wet seasons. Organic carbon is a key chemical component, contributing 9–52% of PM mass. SO42−EN22044_IE1.gif, NO3−EN22044_IE2.gif, and NH4+EN22044_IE3.gif are major ions and NO3−EN22044_IE4.gif/SO42−EN22044_IE5.gif ratios were notably lower than in megacities in other regions, implying lower emissions from mobile relative to stationary sources. Source apportionment indicates biomass burning is one of the most important sources, particularly during the haze (dry) seasons, followed by road traffic emission in dense-traffic cities such as Bangkok, Kuala Lumpur and Hanoi. The secondary aerosols contribution to PM mass is usually lower than that in cities from other regions. We suggest future PM source apportionment studies in Southeast Asia to include both inorganic and organic tracers and apply both chemical mass balance and multivariate receptor models.

A Particulate Matter Concentration Prediction Model Based on Long Short-Term Memory and an Artificial Neural Network.

Many countries are concerned about high particulate matter (PM) concentrations caused by rapid industrial development, which can harm both human health and the environment. To manage PM, the prediction of PM concentrations based on historical data is actively being conducted. Existing technologies for predicting PM mostly assess the model performance for the prediction of existing PM concentrations; however, PM must be forecast in advance, before it becomes highly concentrated and causes damage to the citizens living in the affected regions. Thus, it is necessary to conduct research on an index that can illustrate whether the PM concentration will increase or decrease. We developed a model that can predict whether the PM concentration might increase or decrease after a certain time, specifically for PM (fine PM) generated by anthropogenic volatile organic compounds. An algorithm that can select a model on an hourly basis, based on the long short-term memory (LSTM) and artificial neural network (ANN) models, was developed. The proposed algorithm exhibited a higher F1-score than the LSTM, ANN, or random forest models alone. The model developed in this study could be used to predict future regional PM concentration levels more effectively.

Satellite remote sensing of atmospheric particulate matter mass concentration: Advances, challenges, and perspectives

Mapping the mass concentration of near-surface atmospheric particulate matter (PM) using satellite observations has become a popular research niche, leading to the development of a variety of instruments, algorithms, and datasets over the past two decades. In this study, we conducted a holistic review of the major advances and challenges in quantifying PM, with a specific focus on instruments, algorithms, datasets, and modeling methods that have been developed over the past 20 years. The aim of this study is to provide a general guide for future satellite-based PM concentration mapping practices and to better support air quality monitoring and management of environmental health. Specifically, we review the evolution of satellite platforms, sensors, inversion algorithms, and datasets that can be used for monitoring aerosol properties. We then compare various practical methods and techniques that have been used to estimate PM mass concentrations and group them into four primary categories: (1) univariate regression, (2) chemical transport models (CTM), (3) multivariate regression, and (4) empirical physical approaches. Considering the main challenges encountered in PM mapping practices, for example, data gaps and discontinuity, a hybrid method is proposed with the aim of generating PM concentration maps that are both spatially continuous and have high precision.

Can liquefied petroleum gas vehicles join the fleet of alternative fuel vehicles? Implications of transportation policy based on market forecast and environmental impact

Given the growing severity of particulate matter (PM) and greenhouse gas pollution, the Korean government recently began supporting the use of eco-friendly vehicles, such as electric vehicles and hydrogen fuel cell vehicles. However, the uptake of these vehicles has been slower than anticipated because of the high purchase price and lack of infrastructure. Therefore, the Korean government has eased restrictions on liquefied petroleum gas (LPG) vehicles to encourage the vast majority of consumers to purchase LPG vehicles, expecting that they will serve as a stepping stone to switch from internal combustion engine vehicles to their eco-friendly counterparts. However, there have been only a few studies on consumer preferences for LPG vehicles in consideration of this amendment. To analyze consumer preferences for all fuel types, including LPG, this study used a discrete choice experiment combined with a reference-dependent model. The results showed a loss aversion tendency and asymmetric preferences among consumers with regard to vehicle choice. Based on these results, a simulation analysis was conducted to predict the future market share and PM pollution level. The findings indicated that the government's LPG vehicle deregulation policy could achieve positive PM reduction effects in the short term but none in the long term.

Numerical assessment of PM2.5 and O3 air quality in Continental Southeast Asia: Impacts of future projected anthropogenic emission change and its impacts in combination with potential future climate change impacts

Air quality in future will change in response to anthropogenic emission change as well as climate change. In this study, an advanced online coupled Weather Research and Forecasting (WRF) and Community Multiscale Air Quality (CMAQ) model was applied to the Continental Southeast Asia. The targets were to investigate the impacts of projected change in future anthropogenic emissions and its impacts in combination with climate change impacts on future particulate matter with an aerodynamic diameter of 2.5 μm or less (PM2.5) and ozone (O3) air quality. Future anthropogenic emissions were produced by scaling Hemispheric Transport of Air Pollution (HTAP) emissions in year 2010 using ratios taken from the Evaluating the Climate and Air Quality Impacts of Short-Lived Pollutants (ECLIPSE) emission database in year 2050 and 2010 under current legislation (CLE) scenario. Future climate projections were obtained by implementing a dynamical downscaling method based on pseudo global warming (PGW) technique using future (2046-2055) climate conditions and present (2006-2015) climate conditions provided by the Coupled Model Intercomparison Experiment (CMIP5) Community Earth System Model (CESM) data and the 2014 meteorological conditions. Two Representative Concentration Pathway (RCP) climate scenarios, RCP4.5 and RCP8.5, were considered. Overall, anthropogenic emission change alone appears to cause higher PM2.5 and O3 levels in future period. Across four target countries namely Cambodia, Laos, Thailand, and Vietnam, on average, PM2.5 and O3 concentrations increase by +4.56 μg/m3 (+13.76%) and +5.12 ppb (+13.20%) during the dry season, +3.03 μg/m3 (+27.72%) and +5.89 ppb (+23.97%) during the wet season, respectively. PM2.5 increase is attributed to the emission growth of primary PM2.5 and PM2.5 precursors. Among PM2.5 species, while raising trend was mainly predicted for anthropogenic secondary organic aerosol (PM2.5 ASOA), sulfate ion (PM2.5 SO42−), nitrate ion (PM2.5 NO3−), and ammonium ion (PM2.5 NH4+) concentrations, reducing trend was mainly predicted for element carbon (PM2.5 EC) and primary organic aerosol (PM2.5 POA) concentrations due to the projected primary PM2.5 emission decrease in China. O3 increase is driven by high nitrogen oxides (NOx) emission increase rather than non-methane volatile organic compounds (NMVOCs) emission increase since NOx-limited regime mostly dominates over the region. Both future PM2.5 and O3 increases in Continental Southeast Asia are also contributed by pollutant emission growth in India. Driven by climate change and emission change in combination, the rise in PM2.5 and O3 concentrations is larger in RCP8.5 scenario than that in RCP4.5 scenario. Averagely, under RCP4.5 scenario, PM2.5 and O3 concentrations, across four target countries, were projected to increase by +3.43 μg/m3 (+10.36%) and +4.37 ppb (+11.27%) during the dry season, +2.20 μg/m3 (+20.15%) and +5.20 ppb (+21.16%) during the wet season. Under RCP8.5 scenario, they increase by +7.10 μg/m3 (+21.43%) and +6.35 ppb (+16.38%) during the dry season, +3.08 μg/m3 (+28.15%) and +6.20 ppb (+25.20%) during the wet season, respectively. The simulation results indicated that emission trend is a major factor in the variation in PM2.5 and O3 concentrations, while the climate change also plays an important role. Current legislations and pollution controls as in ECLIPSE CLE scenario are inadequate for better PM2.5 and O3 air quality in the region. A stronger emission control strategy will be required to cope with this PM2.5 and O3 air quality degradation, particularly under RCP8.5 scenario.

Determination of PM mass emissions from an aircraft turbine engine using particle effective density

Inventories of particulate matter (PM) emissions from civil aviation and air quality models need to be validated using up-to-date measurement data corrected for sampling artifacts. We compared the measured black carbon (BC) mass and the total PM mass determined from particle size distributions (PSD) and effective density for a commercial turbofan engine CFM56-7B26/3. The effective density was then used to calculate the PM mass losses in the sampling system. The effective density was determined using a differential mobility analyzer and a centrifugal particle mass analyzer, and increased from engine idle to take-off by up to 60%. The determined mass-mobility exponents ranged from 2.37 to 2.64. The mean effective density determined by weighting the effective density distributions by PM volume was within 10% of the unit density (1000 kg/m3) that is widely assumed in aircraft PM studies. We found ratios close to unity between the PM mass determined by the integrated PSD method and the real-time BC mass measurements. The integrated PSD method achieved higher precision at ultra-low PM concentrations at which current mass instruments reach their detection limit. The line loss model predicted ∼60% PM mass loss at engine idle, decreasing to ∼27% at high thrust. Replacing the effective density distributions with unit density lead to comparable estimates that were within 20% and 5% at engine idle and high thrust, respectively. These results could be used for the development of a robust method for sampling loss correction of the future PM emissions database from commercial aircraft engines.

The role of light duty vehicles in future air pollution: a case study of Sacramento

On-road light-duty vehicles (LDVs) play an important role in contributing to urban air pollution. Although vehicles are getting cleaner, regional growth in vehicle population and vehicle miles traveled would somewhat offset California's efforts in transportation pollution reduction. To better understand the role of LDVs in future air pollution, we conduct a case study for Sacramento, California, and investigate future trends in urban air pollution attributable to the light-duty fleet. Results indicate that ambient concentrations of CO, NO x , and total organic gases (TOGs) caused by future light-duty fleets would dramatically decrease over coming years. The resulting concentrations in 2030 might be as low as approximately 20% of the 2005 concentrations. These reflect the improvements in vehicle/fuel technologies and standards in California. However, the future particulate matter (PM10) pollution could be slightly worse than that caused by the 2005 fleet. This is a result of the growing fleet-average emission factors of particulates from 2005 to 2030. For purposes of future particulate control, more attention needs to be paid to LDVs, besides heavy-duty vehicles.

Key Scientific Findings and Policy- and Health-Relevant Insights from the U.S. Environmental Protection Agency’s Particulate Matter Supersites Program and Related Studies: An Integration and Synthesis of Results

In 1998, the U.S. Environmental Protection Agency (EPA) initiated a major air quality program known as the Particulate Matter (PM) Supersites Program. The Supersites Program was a multiyear, $27 million air quality monitoring program consisting of eight regional air quality projects located throughout the United States, each with differing atmospheric pollution conditions resulting from variations in source emissions and meteorology. The overall goal of the program was to elucidate source-receptor relationships and atmospheric processes leading to PM accumulation on urban and regional scales; thus providing the scientific underpinning for modeling and data analysis efforts to support State Implementation Plans and more effective risk management approaches for PM. The program had three main objectives: (1) conduct methods development and evaluation, (2) characterize ambient PM, and (3) support health effects and exposure research. This paper provides a synthesis of key scientific findings from the Supersites Program and related studies. EPA developed 16 science/policy-relevant questions in conjunction with state and other federal agencies, Regional Planning Organizations, and the private sector. These questions were addressed to the extent possible, even given the vast amount of new information available from the Supersites Program, in a series of papers published as a special issue of the Journal of Air & Waste Management Association (February 2008). This synthesis also includes discussions of: (1) initial Supersites Program support for air quality management efforts in specific locations throughout the United States; (2) selected policy-relevant insights, based on atmospheric sciences findings, useful to air quality managers and decision makers planning emissions management strategies to address current and future PM National Ambient Air Quality Standards (NAAQS) and network planning and implementation; (3) selected health-relevant insights interpreted from atmospheric sciences findings in light of future directions for health and exposure scientists planning studies of the effects of PM on human health; and (4) selected knowledge gaps to guide future research. Finally, given the scope and depth of research and findings from the Supersites Program, this paper provides a reference source so readers can glean a general understanding of the overall research conducted and its policy-relevant insights. Supporting details for the results presented are available through the cited references. An annotated table of contents allows readers to easily find specific subject matter within the text.

New Innovations in Exposure Assessment: Mapping Personal Exposure to Airborne Particles With GPS, GIS, and Real-Time PM Monitors

TAA3-O-03 Introduction: Tracking human exposure to airborne particulate matter (PM) has become increasingly more sophisticated in recent decades. The use of real-time, personal-exposure monitors allows for the temporal resolution of an individual's exposure on a minute basis. The development of small, portable global positioning systems (GPS) allows for the resolution of spatial location on a similar time scale. By integrating these 2 technologies, we may generate a highly resolved exposure map for an individual through space and time. Such information could provide detailed information on human exposure-activity patters as well as establish when and where the most serious exposures to PM occur. Methods: Personal exposures were monitored in occupational (8-hour) and home/school (24-hour) settings. Aerosol concentrations were estimated with personal, real-time aerosol monitors (pDR 1200, Thermo-Electron Inc.) carried by the subject and weighing approximately 0.7 kg. The pDR-1200 is a light-scattering photometer that, in conjunction with a pump, can perform integrated filter based sampling as well as real-time sampling. Size selective sampling ensured that either respirable or inhalble PM samples were collected. Data from the pDR and GPS were combined in a spreadsheet file by matching the time/date data from GPS-location file to the time/date data from the pDR-concentration file. In this way, both data streams were merged and imported into a GIS file. Results: Sampling resutls indicate preliminary success and show spatio/temporally dependent PM concentrations assocaited with different regional locations within a major city (P < 0.001). In addition, correlation of the real-time signal with spatial location allows us to define geographic regions of specificaly high or low PM concentrations from a given exposure track. Periodic drift from the real-time instrument signal indicates that calibration with a filter sample (gravimetric calibration) is warranted. The GPS units also suffer from signal weakening indoors and shortened battery lifetimes. Discussion: This study demonstrates that real-time, personal PM monitors may be coupled with GPS tracking technology to generate spatial-temporal exposure maps on a personal level. In the future, workplace or community-based PM concentration maps may be derived from multiple tracks of individuals without the use of a cost-intensive sampling campaign; such maps may provide enhanced exposure data as an alternative to traditional, fixed-point monitors.

An informative Bayesian structural equation model to assess source-specific health effects of air pollution

A primary objective of current air pollution research is the assessment of health effects related to specific sources of air particles or particulate matter (PM). Quantifying source-specific risk is a challenge because most PM health studies do not directly observe the contributions of the pollution sources themselves. Instead, given knowledge of the chemical characteristics of known sources, investigators infer pollution source contributions via a source apportionment or multivariate receptor analysis applied to a large number of observed elemental concentrations. Although source apportionment methods are well established for exposure assessment, little work has been done to evaluate the appropriateness of characterizing unobservable sources thus in health effects analyses. In this article, we propose a structural equation framework to assess source-specific health effects using speciated elemental data. This approach corresponds to fitting a receptor model and the health outcome model jointly, such that inferences on the health effects account for the fact that uncertainty is associated with the source contributions. Since the structural equation model (SEM) typically involves a large number of parameters, for small-sample settings, we propose a fully Bayesian estimation approach that leverages historical exposure data from previous related exposure studies. We compare via simulation the performance of our approach in estimating source-specific health effects to that of 2 existing approaches, a tracer approach and a 2-stage approach. Simulation results suggest that the proposed informative Bayesian SEM is effective in eliminating the bias incurred by the 2 existing approaches, even when the number of exposures is limited. We employ the proposed methods in the analysis of a concentrator study investigating the association between ST-segment, a cardiovascular outcome, and major sources of Boston PM and discuss the implications of our findings with respect to the design of future PM concentrator studies.

An electrostatic trap for control of ultrafine particle emissions from gasoline-engined vehicles

Health concerns over ultrafine (&lt; 100 nm) particles in the urban atmosphere have focused attention on measurement and control of particle number as well as mass. Gasoline-engined as well as diesel-engined vehicles are likely to be within the scope of future particulate matter (PM) emission regulations. As a potential option for after-treatment of PM emissions from gasoline engines, the trapping performance of a catalysed wire-cylinder electrostatic trap has been investigated, first in a laboratory rig with simulated PM and then in the exhaust of a direct injection spark ignition engine. In the simulation experiments, the trap achieved capture efficiencies by total particle number exceeding 90 per cent at wire voltages of 7–10 kV, gas temperatures up to 400°C, and operating durations up to one hour, with no adverse effects from a catalyst coating on the collecting electrode. In the engine tests, at moderate speeds and loads, capture efficiency was 60–85 per cent in the homogeneous combustion mode and 50–60 per cent, of a much larger number of engine-out particles, in the stratified (overall-lean) mode. Gas residence time in the trap appeared to be a major factor in determining efficiency. The electrical power requirement and the effect on engine back-pressure were both minimal.

The U.S. Environmental Protection Agency Particulate Matter Health Effects Research Centers Program: a midcourse report of status, progress, and plans.

In 1998 Congress mandated expanded U.S. Environmental Protection Agency (U.S. EPA) health effects research on ambient air particulate matter (PM) and a National Research Council (NRC) committee to provide research oversight. The U.S. EPA currently supports intramural and extramural PM research, including five academically based PM centers. The PM centers in their first 2.5 years have initiated research directed at critical issues identified by the NRC committee, including collaborative activities, and sponsored scientific workshops in key research areas. Through these activities, there is a better understanding of PM health effects and scientific uncertainties. Future PM centers research will focus on long-term effects associated with chronic PM exposures. This report provides a synopsis of accomplishments to date, short-term goals (during the next 2.5 years) and longer-term goals. It consists of six sections: biological mechanisms, acute effects, chronic effects, dosimetry, exposure assessment, and the specific attributes of a coordinated PM centers program.

Monitoring of 1-Min Personal Particulate Matter Exposures in Relation to Voice-Recorded Time–Activity Data

Recent studies on the association between exposures to airborne particulate matter (PM) and disease have identified short-term peaks in PM exposures as posing especial health threats. Lightweight personal instruments are needed to characterize short-term exposures to PM and to identify the most important sources of high PM excursions. In this study, we measured exposure to fine PM using a small personal nephelometer (pDR; MIE, Inc) to investigate the utility of this instrument in identifying activities and microenvironments most associated with high PM exposures and the magnitude and duration of peaks in PM exposures. Ten adult volunteers wore a pDR recording PM concentrations at 1-min time intervals for 1 week each. PM concentrations were measured by the pDR in units of μg/m3 based on light scatter. The use of a time-stamped voice recorder enabled activity and location to be continuously documented in real time. In addition, a small, inexpensive light intensity logger was affixed to the pDR to evaluate the potential of this instrument to assist in verifying wearer- recorded data. For each person, patterns of PM exposure were remarkably consistent over daily activities and showed large excursions associated with specific indoor and outdoor microenvironments and activities, such as cooking. When the magnitude and duration of excursions in PM were analyzed, we found that high PM levels occurred in relatively few of the minutes measured but comprised a substantial fraction of the total exposure to PM. Fifteen-minute averaged PM levels were found to be as much as 10 times the daily average. When the data were analyzed with a generalized estimating equation model to account for effects of autocorrelation and clustering, PM exposure was significantly higher during subject-reported events including barbeque, yard work, being near pets or construction activities, cooking, and environmental tobacco smoke exposure, as compared with periods with no pollution events. When light intensity data were explored to determine whether these loggers could be of potential use in establishing or verifying indoor vs outdoor location for future PM studies, we found that personal light intensity measurements differed among indoor, outdoor, and in-car environments (P<0.001). Overlap between measured values implies that light intensity cannot be used to absolutely predict location; however, a sudden increase or decrease in light intensity was highly associated with participant report of location change between indoors and outdoors. This study demonstrates the utility of the pDR in identifying patterns of personal exposures to particulate matter and especially in registering the magnitude and duration of excursions in PM in relation to location and activity.

Particulate Matter Physicochemistry and Toxicology: In Search of Causality—A Critical Perspective

Results from epidemiology studies have shown statistical associations between ambient air particulate matter (PM) concentrations and mortality/morbidity, particularly among susceptible subpopulations. Although the epidemiology of ambient air PM is compelling, there remains considerable uncertainty in PM risk assessment, particularly with regard to identification of PM properties and mechanisms that are responsible for its observed adverse health effects. In addition, the biological mechanisms by which specific PM properties mediate their adverse health effects are currently not known. In vivo and in vitro toxicological studies have examined particles derived from a number of sources such as ambient air, combustion, and natural sources, as well as laboratory-derived surrogate particles, in order to identify the characteristics of particles that are responsible for their adverse health effects. These studies have identified a number of potential causal biogenic, physical, and chemical properties of PM. A number of PM physical and chemical properties have been found to elicit biological responses in animal models of disease, which has enhanced their plausibility as being properties responsible for PM-associated health effects. In vitro mechanistic studies have shown that PM derived from a variety of sources mediate their adverse biological effects by inducing an oxidative stress, which may alter intracellular signal transduction pathways that regulate a number of biological responses such as cell growth, stress response, apoptosis, and inflammation. The challenge of future PM toxicology research will be to demonstrate “biological plausibility and coherence” for potential causal PM properties.