ACS ES&T Air最新文献

Few studies examine health effects of metals in ambient fine particulate matter (PM_2.5), as measurements of elemental composition are sparse. To facilitate intraurban studies in Denver, Colorado, we developed land use regression models for seven speciescopper (Cu), iron (Fe), titanium (Ti), zinc (Zn), potassium (K), calcium (Ca), and magnesium (Mg). As part of the Healthy Start Cohort study, we collected filter-based PM_2.5 using personal air samplers at 67 locations across Denver. Sample collection occurred from May 2018 through March 2019, accounting for all meteorological seasons. Exposure models were informed by 83 geospatial covariates, with traffic-related predictors as the strongest and most consistent across models. Model performance was evaluated using 10-fold cross validation and overall, varied by sampling campaign and season, with R ² values ranging from 0 to 0.63. At best, our model predicts Cu and Fe during fall (R ² = 0.56 and 0.63, respectively); whereas it fails to capture species unrelated to traffic year-round (R ² < 0.40). This highlights the influence of missing predictors (e.g., wildfire smoke, atmospheric transport and other meteorological factors) on PM_2.5 concentrations and spatial gradients. Despite limitations, resulting models enable estimation of intraurban metal exposures and support future analyses of long-term health impacts in Denver.

Despite considerable reductions in mobile source emissions, annual average aerosol concentrations measured in Los Angeles using Federal Reference Methods (FRM) have not appreciably declined over the last decade. Here, we use submicron aerosol measurements and zero-dimensional modeling to quantify the impacts of these emissions reductions on aerosol formation in Pasadena, CA during the late spring and summer of 2022. Reductions in secondary organic aerosol (SOA) concentrations expected from reduced mobile source emissions appear to have been largely offset by increases in hydroxyl radical concentrations, an indirect effect of reduced nitrogen oxide (NO_x) emissions. As a result, while the predicted contribution of mobile sources to the SOA burden has declined from ~50% in 2010 to only ~25% in 2022, concentrations of locally-formed SOA have remained relatively constant. In contrast, reductions in mobile source NO_x emissions have likely reduced overnight production of nitric acid and ammonium nitrate (AN) aerosol. We provide indirect evidence that FRM measurements may have failed to capture the reduction in AN since 2010 due to evaporation of semi-volatile species from FRM filter samples. Our results suggest that given the effectiveness of historical regulatory efforts aimed at mobile sources, and on-road sources in particular, additional reductions in submicron aerosol concentrations in Los Angeles will likely require increased focus on abating emissions from non-road and area sources.

Biomass burning (BB) is a major source of atmospheric particles and trace gases, influencing climate change, air quality, and human health. During the Georgia Wildland-Fire Simulation Experiment, we measured the hygroscopicity (κ) and size-resolved cloud condensation nuclei (CCN) activity of BB particles from controlled burns of fuel beds representative of three ecoregions in Georgia, United States. Primary BB particles were predominantly organic, and photooxidation in an oxidation flow reactor produced secondary organic aerosol (SOA) in a new nucleation mode while transforming primary organic aerosol (POA) into oxidized POA (OPOA) in the pre-existing accumulation mode. We measured hygroscopic growth from 20% to 90% relative humidity using a quartz crystal microbalance and assessed size-resolved CCN activity for particles from 30 to 350 nm at supersaturation between 0.13% and 0.99%. We found that the hygroscopicity parameter of OPOA (κ_OPOA = 0.10-0.19) was higher than that of POA (0.04-0.10), reflecting the influence of heterogeneous oxidation, whereas the hygroscopicity parameter of SOA (κ_SOA = 0.07-0.14) fell between the two. Both fresh and aged BB particles displayed size-dependent κ values and evidence of external mixing, likely because of complex emission characteristics of fuel beds and size-dependent deposition processes. Growth factor-derived and CCN-derived κ values were consistent when accounting for such heterogeneity. A strong positive correlation was found between the mass-averaged κ and O/C ratio, described by the regression κ = 0.31 ± 0.02-(O/C) - 0.05 ± 0.02, which broadly agrees with previous findings for a wide range of laboratory SOA and ambient oxidized organic aerosols. This suggests the potential applicability of a generalized hygroscopicity parameterization across organic aerosols within acceptable uncertainty. Our results highlight the role of BB particles as significant CCN sources during atmospheric aging and emphasize the importance of heterogeneous oxidation in physicochemical evolution of BB particles.

Large scale sporting and cultural events attract many spectators to a single site, leading to changed emissions and potentially creating local air pollution hot spots. Here, we monitored the air quality during the Birmingham 2022 Commonwealth Games, held from July 28th to August 8th, 2022, with 323,000 spectators attending the athletics events, including during the opening and closing ceremonies at the (open air) Alexander Stadium in Birmingham, UK. Particulate (PM_2.5 and PM₁₀) concentrations in fan areas around the stadium peaked ahead of the athletics events and opening and closing ceremonies with PM_2.5 concentrations up to 10 times higher than at nearby urban background monitoring stations. For a spectator attending a full day of events at Alexander Stadium, this would represent a 125% increase in their exposure to PM_2.5 relative to the urban background. Nonrefractory particulate composition in these periods was dominated by organics. Four factors were identified from Positive Matrix Factorization (PMF) analysis of particle composition data recorded using a Quadrupole Aerosol Chemical Speciation Monitor (Q-ACSM): two representing cooking aerosol accounting for 71% of the total PM mass during the athletic sessions demonstrating that cooking sources were responsible for the majority of particulate pollution at the venue. The high particulate concentrations at this venue were driven by fast food production at temporary concession stands, common across many large events, leading to a large increase in particulate matter exposure for staff and visitors.

Exposure to fine particulate matter (PM_2.5) in ambient air is recognized as the leading environmental risk factor for mortality. A more comprehensive characterization of its chemical composition is needed for its management and health effects research. We improve estimates of total PM_2.5 mass concentration and its chemical composition across North America by developing, optimizing, and applying convolutional neural networks (CNN) with information from satellite-, simulation-, and monitor-based sources to estimate the local bias in monthly geophysical a priori PM_2.5 and component concentrations over 2000-2023. Significant long-term agreement is found with traditional 10-fold spatial cross-validation for total PM_2.5 (R ² = 0.82), sulfate (R ² = 0.98), nitrate (R ² = 0.93), ammonium (R ² = 0.94), organic matter (R ² = 0.83), black carbon (R ² = 0.78), dust (R ² = 0.71), and seasalt (R ² = 0.37). We introduce Buffered Leave Isolated Sites and Clusters Out (BLISCO) spatial cross-validation to evaluate the model extrapolation ability over remote regions, and find that traditional spatial cross-validation may overestimate performance and underrepresent uncertainty due to the spatial autocorrelation of ground monitors. The use of geophysical information from a chemical transport model (GEOS-Chem) significantly increases CNN performance in BLISCO cross-validation, for example, increasing R ² for NO₃ ^- (0.51 to 0.81) and NH₄ ⁺ (0.27 to 0.67). We represent spatial uncertainty for PM_2.5 and its components based on the statistical results of BLISCO cross-validation by integrating information from both the spatial distribution of ground observations and the variability in predictors space representation, and find that distance from monitor is a key predictor of uncertainty.

Wildfire smoke is an increasingly significant contributor to air pollution in the western United States, posing serious health risks and complicating efforts to assess personal exposure, particularly indoors. The January 2025 Palisades and Eaton Fires in Los Angeles County caused elevated levels of PM_2.5 in the downwind cities. This study leverages a high-resolution network of crowdsourced PurpleAir sensors to evaluate indoor and outdoor PM_2.5 levels before, during, and after the wildfire smoke events. We matched indoor-outdoor sensors and analyzed disparities in smoke exposure across communities with different CalEnviroScreen (CES) vulnerability scores, ventilation types, and home values. Results indicate that outdoor PM_2.5 increased substantially during smoke days, with the highest CES-burdened communities experiencing the greatest ambient concentrations. Indoor PM_2.5 also increased across all neighborhoods but indoor/outdoor (I/O) ratios declined during the smoke period, indicating partial indoor protection and likely occupant behavior changes. Infiltrated PM_2.5 increased during the smoke period and varied across the CES groups. Building attributes showed limited predictive power. These findings highlight the interplay between behavioral actions and neighborhood factors in shaping wildfire smoke exposure and underscore the need for targeted interventions to improve indoor air quality in vulnerable communities.

Indoor cooking generates intense, short-duration fine particulate matter (PM_2.5) peaks with acute health risks. To quantify the efficacy of natural ventilation configurations, we conducted approximately two months of continuous monitoring in a modern UK one-bedroom apartment, comparing three ventilation scenarios during cooking: fully opened (all windows and internal doors open), door-opened only (internal doors open but windows closed), and fully closed (all windows and internal doors closed). Air quality sensors were calibrated against a reference instrument (Fidas 200E) both before and after the field deployment. During the study period, outdoor PM_2.5 mass concentrations ranged from 0.4 to 31.0 μg m^-3, averaging 6.3 μg m^-3. Indoor concentrations were substantially higher than average outdoor levels, with the fully opened scenario yielding the lowest exposure at 14.9 μg m^-3 in the living room/kitchen and 15.4 μg m^-3 in the bedroom. Relative to the fully opened scenario, PM_2.5 concentrations increased by 58.4% (living room/kitchen) and 55.8% (bedroom) under door-opened only conditions, and under fully closed conditions by 28.9% and 27.9%, respectively. These findings demonstrate that simultaneous opening of windows and internal doors during cooking can substantially reduce acute PM_2.5 exposure, offering a simple, low-energy strategy to mitigate short-term health risks in naturally ventilated apartments.

Air frying has emerged as a popular low-oil cooking method, yet its impact on indoor air pollutant emissions remains insufficiently understood. In our study, emissions of volatile organic compounds (VOCs), nitrogen oxides (NO _x ), and ultrafine particles (UFPs) were measured during the air frying of 12 different dishes within a ca. 0.15 m³ Perspex chamber. Pollutant emissions varied significantly depending on the food type, with rates in the ranges of 17.8-184.0 μg min^-1 for total cooking VOCs, 24.6-37.9 μg min^-1 for NO _x , and 0.1-17.4 × 10¹² # min^-1 for UFPs, primarily due to Maillard reactions and lipid thermal decomposition. While pollutant concentrations and ozone formation potentials were elevated within the chamber, scaling to the volume of a small kitchen indicated substantially lower levels compared to conventional frying methods. Notably, only high-fat foods produced UFP concentrations comparable to those of deep frying. No NO _x emissions were found during blank (empty appliance) runs, and NO _x was only detectable while cooking certain types of foods. However, residues accumulating within inaccessible areas of the air fryer following over 70 uses led to increases of 23% in VOC and 236% in UFP concentrations while not cooking food.

The Los Angeles (LA) metropolitan region remains in nonattainment for ozone despite decades of reductions of ozone precursors, nitrogen oxides (NO _x ) and volatile organic compounds (VOCs). NO _x emissions from freight vehicles (ships, heavy duty trucks, trains, and airplanes) are expected to exceed emissions from passenger vehicles in southern California by 2030. Here, we use random forest machine learning to estimate the impact of freight activity on hourly NO _x concentrations and determine summertime ozone production regimes across the LA basin. We find that freight activity contributes over half of weekday NO _x impacts relative to non-truck traffic. During peak ozone hours, coastal areas, south LA, areas downwind (east) of downtown LA, and downtown San Bernardino are VOC-limited. Our results suggest that as of 2021, the Los Angeles urban core and nearby downwind areas have not transitioned to a NO _x -limited regime on most days during the May to September ozone season. This study shows the applicability of machine learning to estimate concentration impacts from specific sources in the face of uncertain emission inventories and to analyze current ozone production regimes in areas with hourly ground observations.

Monte Carlo (MC) simulations are a common way to estimate the methane emissions and cost-effectiveness of leak detection and repair (LDAR) programs in the upstream oil and gas industry. In this paper we consider a simplified version of a LDAR program and demonstrate how to simulate the underlying system in an efficient manner, resulting in estimators of the quantities of interest that have a smaller variance than is possible using contemporary techniques. The proposed method relies on two ideas: the first is to leverage the underlying stochastic models to perform an event-driven simulation rather than a daily one; and the second is to use low-discrepancy sampling rather than plain random sampling to generate samples of potential scenarios for the underlying system, which results in a more systematic and balanced exploration of the scenario space. We show that in the context of a sensitivity analysis example, the proposed approach reduces the error of the estimates by factors of about 3 to 4 for the same computation time. This increased precision can provide conclusive statistical evidence that an LDAR program significantly reduces emissions compared to another, while the naive method's error is often too large to draw any conclusion.