Combining Machine Learning and Satellite Observations to Predict Spatial and Temporal Variation of near Surface OH in North American Cities

Interannual changes in atmospheric oxidation over forests determined from space

The hydroxyl radical (OH) is the central oxidant in Earth's troposphere, but its temporal variability is poorly understood. We combine 2012-2020 satellite-based isoprene and formaldehyde measurements to identify coherent OH changes over temperate and tropical forests with attribution to emission trends, biotic stressors, and climate. We identify a multiyear OH decrease over the Southeast United States and show that with increasingly hot/dry summers the regional chemistry could become even less oxidizing depending on competing temperature/drought impacts on isoprene. Furthermore, while global mean OH decreases during El Niño, we show that near-field effects over tropical rainforests can alternate between high/low OH anomalies due to opposing fire and biogenic emission impacts. Results provide insights into how atmospheric oxidation will evolve with changing emissions and climate.

New Insights on the Formation of Nucleation Mode Particles in a Coastal City Based on a Machine Learning Approach

Atmospheric particles have profound implications for the global climate and human health. Among them, ultrafine particles dominate in terms of the number concentration and exhibit enhanced toxic effects as a result of their large total surface area. Therefore, understanding the driving factors behind ultrafine particle behavior is crucial. Machine learning (ML) provides a promising approach for handling complex relationships. In this study, three ML models were constructed on the basis of field observations to simulate the particle number concentration of nucleation mode (PNCN). All three models exhibited robust PNCN reproduction (R2 > 0.80), with the random forest (RF) model excelling on the test data (R2 = 0.89). Multiple methods of feature importance analysis revealed that ultraviolet (UV), H2SO4, low-volatility oxygenated organic molecules (LOOMs), temperature, and O3 were the primary factors influencing PNCN. Bivariate partial dependency plots (PDPs) indicated that during nighttime and overcast conditions, the presence of H2SO4 and LOOMs may play a crucial role in influencing PNCN. Additionally, integrating additional detailed information related to emissions or meteorology would further enhance the model performance. This pilot study shows that ML can be a novel approach for simulating atmospheric pollutants and contributes to a better understanding of the formation and growth mechanisms of nucleation mode particles.

Location-Specific Control of Precursor Emissions to Mitigate Photochemical Air Pollution

The effects of precursor emission controls on air quality can vary greatly depending on where emission reductions occur. We use the adjoint of the Community Multiscale Air Quality (CMAQ) model to evaluate impacts of spatially targeted NO_x emission reductions on odd oxygen (O_x = O₃ + NO₂). The air quality responses studied here include one population-weighted regionwide and three city-level receptors in Central California. We map high-priority locations for NO_x control and their changes over decadal time scales. The desirability of NO_x-focused emission control programs has increased between 2000 and 2022. We find for present-day conditions that reducing NO_x emissions by 28% from targeted high-priority locations can achieve 60% of the air quality benefits of uniform NO_x reductions at all locations. High-priority source locations are found to differ for individual city-level versus regionwide receptors of interest. While high-impact emission hotspots for improving city-level metrics are found within the city itself or closely adjacent, the spatial pattern of emission hotspots for improving regionwide air quality is more complex and requires comprehensive consideration of upwind sources. Results of this study can help to inform strategic decision-making at local and regional levels about where to prioritize emission control efforts.

Spatiotemporal variations of NO2 and its driving factors in the coastal ports of China

Nitrogen Dioxide (NO₂) is one of the major air pollutants in coastal ports of China. Understanding the spatiotemporal varying effects of driving factors of NO₂ is vital for the implementation of differentiated air pollution control measures for different port areas. Based on the Ozone Monitoring Instrument (OMI) satellite data, we adopted a Geographically and Temporally Weighted Regression (GTWR) model to explore the influences of meteorological and socioeconomic factors on the NO₂ Vertical Column Concentrations (VCDs) in coastal ports of China from 2015 to 2021. The results indicate that NO₂ VCD in most ports has decreased since 2016 and the ports with serious NO₂ pollution are mainly distributed in northern China. The associations between NO₂ VCD levels and their drivers exhibit obvious spatiotemporal heterogeneity. Higher wind speed and relative humidity are more helpful to alleviate NO₂ pollution in ports of the Bohai Rim and the Pearl River Delta. Cargo throughput has more closely associated with NO₂ pollution in Beibu Gulf in recent years, yet there is no significant association found for Shanghai ports. The positive relationship between transportation emissions and NO₂ VCD is more significant in southern ports. This work provides some implications for the formulation of targeted emission reduction policies for different ports along the Chinese coast.

Elucidating Contributions of Anthropogenic Volatile Organic Compounds and Particulate Matter to Ozone Trends over China

In China, emissions of ozone (O₃)-producing pollutants have been targeted for mitigation to reduce O₃ pollution. However, the observed O₃ decrease is slower than/opposite to expectations affecting the health of millions of people. For a better understanding of this failure and its connection with anthropogenic emissions, we quantify the summer O₃ trends that would have occurred had the weather stayed constant by applying a numerical tool that "de-weathers" observations across 31 urban regions (123 cities and 392 sites) over 8 years. O₃ trends are significant (p < 0.05) over 234 sites after de-weathering, contrary to the directly observed trends (only 39 significant due to high meteorology-induced variability). The de-weathered data allow categorizing cities in China into four different groups regarding O₃ mitigation, with group 1 exhibiting steady O₃ reductions, while group 4 showing significant (p < 0.05) O₃ increases. Analysis of the relationships between de-weathered odd oxygen and nitrogen oxides illustrates how the changes in NO_x, in anthropogenic volatile organic compounds (VOCs), and reductions in fine particulate matter (PM_2.5) affect the O₃ trends differently in these groups. While this analysis suggests that VOC reductions are the main driver of O₃ decreases in group 1, groups 3 and 4 are primarily affected by decreasing PM_2.5, which results in enhanced O₃ formation. Our analysis demonstrates both the importance of and possibility for isolating emission-driven changes from climate and weather for interpreting short-term air quality observations.

Estimate of OH trends over one decade in North American cities

SignificanceOH is the critical chemical setting removal rates of local pollutants in the atmosphere. The importance of OH to tropospheric chemistry stands in stark contrast to the absence of long-term measurements. Here we synthesize a machine learning technique, satellite observations, and simulations from a state-of-the-art chemical model to estimate OH trends between 2005 and 2014 in 49 North American cities. Compared to the summertime OH in 2005, the OH in 2014 exhibits changes that range from -17 to +11% in different cities. The variation of OH over one decade can be explained by the chemical regime shifts over the years. The identification of chemical regime, in turn, sheds light on the effective policy for controlling ozone.