Rationale: Particulate matter ≤2.5 μm in aerodynamic diameter (PM2.5) is an established cause of lung cancer, but the association with ultrafine particulate matter (UFP; aerodynamic diameter < 0.1 μm) is unclear. Objectives: To investigate the association between UFP and lung cancer overall and by histologic subtype. Methods: The Los Angeles Ultrafines Study includes 45,012 participants aged ≥50 years in southern California at enrollment (1995–1996) followed through 2017 for incident lung cancer (n = 1,770). We estimated historical residential ambient UFP number concentrations via land use regression and back extrapolation using PM2.5. In Cox proportional hazards models adjusted for smoking and other confounders, we estimated associations between 10-year lagged UFP (per 10,000 particles/cm3 and quartiles) and lung cancer overall and by major histologic subtype (adenocarcinoma, squamous cell carcinoma, and small cell carcinoma). We also evaluated relationships by smoking status, birth cohort, and historical duration at the residence. Measurements and Main Results: UFP was modestly associated with lung cancer risk overall (hazard ratio [HR], 1.03 [95% confidence interval (CI), 0.99–1.08]). For adenocarcinoma, we observed a positive trend among men; risk was increased in the highest exposure quartile versus the lowest (HR, 1.39 [95% CI, 1.05–1.85]; P for trend = 0.01) and was also increased in continuous models (HR per 10,000 particles/cm3, 1.09 [95% CI, 1.00–1.18]), but no increased risk was apparent among women (P for interaction = 0.03). Adenocarcinoma risk was elevated among men born between 1925 and 1930 (HR, 1.13 [95% CI, 1.02–1.26] per 10,000) but not for other birth cohorts, and was suggestive for men with ≥10 years of residential duration (HR, 1.11 [95% CI, 0.98–1.26]). We found no consistent associations for women or other histologic subtypes. Conclusions: UFP exposure was modestly associated with lung cancer overall, with stronger associations observed for adenocarcinoma of the lung.

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.

Airborne longwave-infrared (LWIR) hyperspectral imagery acquisitions were coordinated with stationary and mobile ground-based in situ measurements of atmospheric ammonia in regions surrounding California’s Salton Sea, an area of commingled intensive animal husbandry and agriculture operations that is encumbered by exceptionally high levels of persistent ammonia and PM_2.5 pollution. The goal of this study was to validate remotely sensed ammonia retrievals against ground truth measurements as part of a broader effort to elucidate the behavior of the atmospheric ammonia burden in this area of abundant diffuse and point sources. The nominal 2 m pixel size of the airborne data revealed variability in ammonia concentrations at a diversity of scales within the study area. At this pixel resolution, ammonia plumes emitted by individual facilities could be clearly discriminated and their dispersion characteristics inferred. Several factors, including thermal contrast and atmospheric boundary layer depth, contributed to the overall uncertainty of the intercomparison between airborne ammonia quantitative retrievals and the corresponding in situ measurements, for which agreement was in the 16–37% range under the most favorable conditions. Hence, while the findings attest to the viability of airborne LWIR spectral imaging for quantifying atmospheric ammonia concentrations, the accuracy of ground-level estimations depends significantly on precise knowledge of these atmospheric factors.

Accurate air pollution monitoring is critical to understand and mitigate the impacts of air pollu-tion to human health and ecosystems. Due to the limited number and geographical coverage of advanced high accurate sensors to monitor air pollutants, many low-cost but low accuracy sensors have been deployed. Calibrating the low-cost sensors is essential to leverage them for filling this geographical gap in sensor coverage. We examine systematically 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 are examined. This systematic study found that both models and packages impact the accuracy while the choice of random training/testing split ration (e.g., 80/20 vs 70/30) has 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 may be too slow or computation intensive (1.5 hours) in 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, offer good performance at shorter training times (< 1 minute), and may be suitable for such ap-plications. These findings suggest that AI/ML models, particularly LSTM, can effectively calibrate low-cost sensors to produce precise, localized air quality data. This research is the most com-prehensive in comparison to existing literature on AI/ML for air pollutants calibration. We also discussed the limitation, applicability to other sensors and the reason for well performed models. The research can be adapted to enhance air quality monitoring for public health risk assessments and support broader environmental health initiatives and inform policy decisions.

Ammonia (NH₃) plays an important role in atmospheric chemistry and air quality, but its emissions remain poorly constrained due to its short atmospheric lifetime, high spatial heterogeneity, and limited coverage of existing monitoring resources. This study integrates airborne longwave-infrared (LWIR) hyperspectral imaging at ∼2 m spatial resolution with ground-based stationary and mobile in situ measurements to map and characterize NH₃ emissions in two regions near the Salton Sea in southern California: Mecca in the northwest and Imperial in the southeast. Airborne surveys conducted in March and September 2023 with a wide-swath LWIR spectral imager revealed pronounced spatial and seasonal variability. Average NH₃ levels in Imperial were 2.5–8 times higher than those in Mecca, linked primarily to large, concentrated animal feeding operations (CAFOs), geothermal power plants, fumaroles, and intensive agricultural activities. Ground-based mobile monitoring corroborated these findings, showing elevated NH₃ levels near these sources and especially high NH₃ concentrations downwind of CAFOs with large cattle populations. The results underscore the utility of airborne LWIR hyperspectral imaging in detecting and mapping NH₃ at hyperlocal scales, including sources absent from existing inventories. They further highlight the need for routine airborne campaigns and the development of next-generation satellite missions with higher spatial resolution to achieve comprehensive, large-area monitoring. These findings inform air quality management strategies and emphasize the importance of improving emission inventories for effective mitigation of NH₃-driven air pollution.

Ammonia (NH₃) plays an important role in atmospheric chemistry and air quality, but its emissions remain poorly constrained due to its short atmospheric lifetime, high spatial heterogeneity, and limited coverage of existing monitoring resources. This study integrates airborne longwave-infrared (LWIR) hyperspectral imaging at ~2 m spatial resolution with ground-based stationary and mobile in-situ measurements to map and characterize NH₃ emissions in two regions near the Salton Sea in Southern California: Mecca in the northwest and Imperial in the southeast. Airborne surveys conducted in March and September 2023 with a wide-swath LWIR spectra imager revealed pronounced spatial and seasonal variability. Average NH₃ levels in Imperial were 2.5 to 8 times higher than those in Mecca, linked primarily to large, concentrated animal feeding operations (CAFOs), geothermal power plants, fumaroles, and intensive agricultural activities. Ground-based mobile monitoring corroborated these findings, showing elevated NH₃ levels near these sources and especially high NH₃ concentrations downwind of CAFOs with large cattle populations. The results underscore the utility of airborne LWIR hyperspectral imaging in detecting and mapping NH₃ at hyperlocal scales, including sources absent from existing inventories. They further highlight the need for routine airborne campaigns and the development of next-generation satellite missions with higher spatial resolution to achieve comprehensive, large-area monitoring. These findings inform air quality management strategies and emphasize the importance of improving emission inventories for effective mitigation of NH₃-driven air pollution.