Statistical and machine learning methods for evaluating trends in air quality under changing meteorological conditions

Quantifying the contributions of meteorology, emissions, and transport to ground-level ozone in the Pearl River Delta, China

Ozone pollution presents a growing air quality threat in urban agglomerations in China. It remains challenge to distinguish the roles of emissions of precursors, chemical production and transportations in shaping the ground-level ozone trends, largely due to complicated interactions among these 3 major processes. This study elucidates the formation factors of ozone pollution and categorizes them into local emissions (anthropogenic and biogenic emissions), transport (precursor transport and direct transport from various regions), and meteorology. Particularly, we attribute meteorology, which affects biogenic emissions and chemical formation as well as transportation, to a perturbation term with fluctuating ranges. The Community Multiscale Air Quality (CMAQ) model was utilized to implement this framework, using the Pearl River Delta region as a case study, to simulate a severe ozone pollution episode in autumn 2019 that affected the entire country. Our findings demonstrate that the average impact of meteorological conditions changed consistently with the variation of ozone pollution levels, indicating that meteorological conditions can exert significant control over the degree of ozone pollution. As the maximum daily 8-hour average (MDA8) ozone concentrations increased from 20 % below to 30 % above the National Ambient Air Quality Standard II, contributions from emissions and precursor transport were enhanced. Concurrently, direct transport within Guangdong province rose from 13.8 % to 22.7 %, underscoring the importance of regional joint prevention and control measures under adverse weather conditions. Regarding biogenic emissions and precursor transport that cannot be directly controlled, we found that their contributions were generally greater in urban areas with high nitrogen oxides (NOx) levels, primarily due to the stronger atmospheric oxidation capacity facilitating ozone formation. Our results indicate that not only local anthropogenic emissions can be controlled in urban areas, but also the impacts of local biogenic emissions and precursor transport can be potentially regulated through reducing atmospheric oxidation capacity.

Constructing transferable and interpretable machine learning models for black carbon concentrations

Black carbon (BC) has received increasing attention from researchers due to its adverse health effects. However, in-situ BC measurements are often not included as a regulated variable in air quality monitoring networks. Machine learning (ML) models have been studied extensively to serve as virtual sensors to complement the reference instruments. This study evaluates and compares three white-box (WB) and four black-box (BB) ML models to estimate BC concentrations, with the focus to show their transferability and interpretability. We train the models with the long-term air pollutant and weather measurements in Barcelona urban background site, and test them in other European urban and traffic sites. Despite the difference in geographical locations and measurement sites, BC correlates the strongest with particle number concentration of accumulation mode (PNacc, r = 0.73-0.85) and nitrogen dioxide (NO2, r = 0.68-0.85) and the weakest with meteorological parameters. Due to its similarity of correlation behaviour, the ML models trained in Barcelona performs prominently at the traffic site in Helsinki (R2 = 0.80-0.86; mean absolute error MAE = 3.90-4.73 %) and at the urban background site in Dresden (R2 = 0.79-0.84; MAE = 4.23-4.82 %). WB models appear to explain less variability of BC than BB models, long short-term memory (LSTM) model of which outperforms the rest of the models. In terms of interpretability, we adopt several methods for individual model to quantify and normalize the relative importance of each input feature. The overall static relative importance commonly used for WB models demonstrate varying results from the dynamic values utilized to show local contribution used for BB models. PNacc and NO2 on average have the strongest absolute static contribution; however, they simultaneously impact the estimation positively and negatively at different sites. This comprehensive analysis demonstrates that the possibility of these interpretable air pollutant ML models to be transfered across space and time.

Achievements and challenges in improving air quality in China: Analysis of the long-term trends from 2014 to 2022

Due to the implementation of air pollution control measures in China, air quality has significantly improved, although there are still additional issues to be addressed. This study used the long-term trends of air pollutants to discuss the achievements and challenges in further improving air quality in China. The Kolmogorov-Zurbenko (KZ) filter and multiple-linear regression (MLR) were used to quantify the meteorology-related and emission-related trends of air pollutants from 2014 to 2022 in China. The KZ filter analysis showed that PM2.5 decreased by 7.36 ± 2.92% yr-1, while daily maximum 8-h ozone (MDA8 O3) showed an increasing trend with 3.71 ± 2.89% yr-1 in China. The decrease in PM2.5 and increase in MDA8 O3 were primarily attributed to changes in emission, with the relative contribution of 85.8% and 86.0%, respectively. Meteorology variations, including increased ambient temperature, boundary layer height, and reduced relative humidity, also contributed to the reduction of PM2.5 and the enhancement of MDA8 O3. The emission-related trends of PM2.5 and MDA8 O3 exhibited continuous decrease and increase, respectively, from 2014 to 2022, while the variation rates slowed during 2018-2020 compared to that during 2014-2017, highlighting the challenges in further improving air quality, particularly in simultaneously reducing PM2.5 and O3. This study recommends reducing NH3 emissions from the agriculture sector in rural areas and transport emissions in urban areas to further decrease PM2.5 levels. Addressing O3 pollution requires the reduction of O3 precursor gases based on site-specific atmospheric chemistry considerations.

Predicting Atmospheric Water-Soluble Organic Mass Reversibly Partitioned to Aerosol Liquid Water in the Eastern United States

Water-soluble organic matter (WSOM) formed through aqueous processes contributes substantially to total atmospheric aerosol, however, the impact of water evaporation on particle concentrations is highly uncertain. Herein, we present a novel approach to predict the amount of evaporated organic mass induced by sample drying using multivariate polynomial regression and random forest (RF) machine learning models. The impact of particle drying on fine WSOM was monitored during three consecutive summers in Baltimore, MD (2015, 2016, and 2017). The amount of evaporated organic mass was dependent on relative humidity (RH), WSOM concentrations, isoprene concentrations, and NOx/isoprene ratios. Different models corresponding to each class were fitted (trained and tested) to data from the summers of 2015 and 2016 while model validation was performed using summer 2017 data. Using the coefficient of determination (R2) and the root-mean-square error (RMSE), it was concluded that an RF model with 100 decision trees had the best performance (R2 of 0.81) and the lowest normalized mean error (NME < 1%) leading to low model uncertainties. The relative feature importance for the RF model was calculated to be 0.55, 0.2, 0.15, and 0.1 for WSOM concentrations, RH levels, isoprene concentrations, and NOx/isoprene ratios, respectively. The machine learning model was thus used to predict summertime concentrations of evaporated organics in Yorkville, Georgia, and Centerville, Alabama in 2016 and 2013, respectively. Results presented herein have implications for measurements that rely on sample drying using a machine learning approach for the analysis and interpretation of atmospheric data sets to elucidate their complex behavior.

Unraveling complex causal processes that affect sustainability requires more integration between empirical and modeling approaches

Scientists seek to understand the causal processes that generate sustainability problems and determine effective solutions. Yet, causal inquiry in nature-society systems is hampered by conceptual and methodological challenges that arise from nature-society interdependencies and the complex dynamics they create. Here, we demonstrate how sustainability scientists can address these challenges and make more robust causal claims through better integration between empirical analyses and process- or agent-based modeling. To illustrate how these different epistemological traditions can be integrated, we present four studies of air pollution regulation, natural resource management, and the spread of COVID-19. The studies show how integration can improve empirical estimates of causal effects, inform future research designs and data collection, enhance understanding of the complex dynamics that underlie observed temporal patterns, and elucidate causal mechanisms and the contexts in which they operate. These advances in causal understanding can help sustainability scientists develop better theories of phenomena where social and ecological processes are dynamically intertwined and prior causal knowledge and data are limited. The improved causal understanding also enhances governance by helping scientists and practitioners choose among potential interventions, decide when and how the timing of an intervention matters, and anticipate unexpected outcomes. Methodological integration, however, requires skills and efforts of all involved to learn how members of the respective other tradition think and analyze nature-society systems.

Assessment of meteorological and air quality drivers of elevated ambient ozone in Beijing via machine learning approach

Over the past few years, surface ozone (O3) pollution has dominated China's air pollution as particulate matter has decreased. In Beijing, the annual average concentrations of ground-level O3 from 2015 to 2020 regularly increased from 57.32 to 62.72 μg/m3, showing a change of almost 9.4%, with a 1.6% per year increase. The meteorological factors are the primary influencer of elevated O3 levels; however, their importance and heterogeneity of variables remain rarely understood. In this study, we used 13 meteorological factors and 6 air quality (AQ) parameters to estimate their influencing score using the random forest (RF) algorithm to explain and predict ambient O3. Among the meteorological variables and overall, both land surface temperature and temperature at 2 m from the surface emerged as the most influential factors, while NO2 stood out as the highest influencing factor from the AQ parameters. Indeed, it is crucial and imperative to reduce the temperature caused by climate change in order to effectively control ambient O3 levels in Beijing. Overall, meteorological factors alone exhibited a higher coefficient of determination (R2) value of 0.80, compared with AQ variables of 0.58, for the post-lockdown period. In addition, we calculated the number of days O3 concentration levels exceeded the WHO standard and newly proposed peak-season maximum daily 8-h average (MDA8) O3 guideline for Beijing. The exceedance number of days from the WHO standard of MDA8 ambient O3 was observed to be the highest in June, and each studied year crossed peak season guidelines by almost 2 times margin. This study demonstrates the contributions of meteorological variables and AQ parameters in surging ambient O3 and highlights the importance of future research toward devising an optimum strategy to combat growing O3 pollution in urban areas.

Estimating particulate matter concentrations and meteorological contributions in China during 2000-2020

Estimating the effects of airborne particulate matter (PM) on climate and human health is highly dependent on the accurate prediction of its concentration and size distribution. High-complexity machine learning models have been widely used for PM concentration prediction, but such models are often considered as "black boxes", lacking interpretability. Here, a simple structure lightGBM model is built for ground PM estimation, and the SHAP approach is used to separate the meteorological contributions due to its strong influence on PM concentration. The models show good performance with correlation coefficient (R²) of 0.84-0.88, 0.80-0.85, and 0.71-0.79, for PM_2.5, PM₁₀, and PM_2.5-10 (2.5-10 μm), respectively. The lightGBM model trains 45 times faster than the XGBoost model while showing similar accuracy. More importantly, the models have small performance gaps between training and predicting (delta R²: 0.07-0.12), effectively reducing overfitting risk. The PM datasets (10 km daily) of three size ranges are then generated over China from 2000 to 2020. The SHAP method shows good agreement with the meteorological normalization approach in separating the meteorological contributions (R² > 0.5). In the Beijing-Tianjin-Hebei region (BTH), meteorology has greater influence on PM_2.5-10 (-5.66%-9.99%) than PM_2.5 and PM₁₀. In the Yangtze River Delta (YRD), and the Pearl River Delta (PRD), albedo has a large contribution to PM_2.5 concentration under the influence of solar radiation. Notably, relative humidity (RH) has different seasonal effects on PM of three size ranges. In the BTH region, RH has negative effects on PM_2.5 (-0.52 μg/m³) and positive effects on PM₁₀ (1.01 μg/m³) and PM_2.5-10 (3.39 μg/m³) in spring, but has opposite effects in summer. The results of SHAP approach are consistent with existing conclusions and imply its feasibility in explaining haze formation. The generated PM datasets are useful in health assessment, environmental management, and climate change studies.

Daily Local-Level Estimates of Ambient Wildfire Smoke PM2.5 for the Contiguous US

Smoke from wildfires is a growing health risk across the US. Understanding the spatial and temporal patterns of such exposure and its population health impacts requires separating smoke-driven pollutants from non-smoke pollutants and a long time series to quantify patterns and measure health impacts. We develop a parsimonious and accurate machine learning model of daily wildfire-driven PM_2.5 concentrations using a combination of ground, satellite, and reanalysis data sources that are easy to update. We apply our model across the contiguous US from 2006 to 2020, generating daily estimates of smoke PM_2.5 over a 10 km-by-10 km grid and use these data to characterize levels and trends in smoke PM_2.5. Smoke contributions to daily PM_2.5 concentrations have increased by up to 5 μg/m³ in the Western US over the last decade, reversing decades of policy-driven improvements in overall air quality, with concentrations growing fastest for higher income populations and predominantly Hispanic populations. The number of people in locations with at least 1 day of smoke PM_2.5 above 100 μg/m³ per year has increased 27-fold over the last decade, including nearly 25 million people in 2020 alone. Our data set can bolster efforts to comprehensively understand the drivers and societal impacts of trends and extremes in wildfire smoke.