The Development and Application of Machine Learning in Atmospheric Environment Studies

Identifying the key factors of mercury exposure in residents of southwestern Iran using machine learning algorithms

It is necessary to predict hair mercury (Hg) levels and specify the related effective factors to develop preventive strategies to reduce Hg exposure in different regions. This study is the first effort to investigate the effectiveness of eight machine learning (ML) models (including multiple linear regression, decision tree regression, least absolute shrinkage and selection operator, multivariate adaptive regression splines, random forest, extreme gradient boosting, K-nearest neighbor, and Gaussian process) for predicting hair Hg levels and identifying the most important factors affecting them in residents of southwestern Iran. All ML models were trained with 70% of the dataset and their performance was evaluated using the determination coefficient (R2), root mean square error (RMSE), and mean absolute error (MAE) based on the remaining dataset. Finally, the Permutation Feature Importance (PFI) method was used to determine the relative importance (RI) of influencing factors. Mean hair Hg (3.31 µg g⁻1) was higher than the United States Environmental Protection Agency (US EPA) and World Health Organization (WHO) limits. It was indicated a high exposure risk for some people in this region. The extreme gradient boosting (XGB) model outperformed other algorithms in modeling hair Hg levels, with R2 = 0.61, RMSE = 2.2, and MAE = 1.25. According to the PFI analysis, weight (RI: 43.4%) and geographic place (RI: 41.8%) were found as the most important demographic factors influencing Hg variation in the study population. Additionally, occupation (RI: 46.1%) and the frequency of fish and canned fish consumption (RI: 22%) were identified as the most significant exposure factors controlling hair Hg variability in southwestern Iran. These findings can be useful for formulating appropriate strategies to reduce the health risk of Hg exposure and improve human health.

Uncovering key sources of regional ozone simulation biases using machine learning and SHAP analysis

Atmospheric chemical transport models (CTMs) are widely used in air quality management, but still have large biases in simulations. Accurately and efficiently identifying key sources of simulation biases is crucial for model improvement. However, traditional approaches, such as sensitivity and uncertainty analyses, are computationally intensive and inefficient, as they require numerous model runs. In this study, we explored the use of machine learning, specifically XGBoost combined with SHAP analysis, as an efficient diagnostic tool for analyzing simulation biases, focusing on ozone modeling in Guangdong Province as a case study. We used the bias of model inputs as features and excluded a dataset that was more susceptible to observational uncertainties to better target bias sources. Results reveal that biases in concentrations of NO2, NO and PM2.5, temperature and biogenic emissions are important sources that lead to O3 simulation biases. Notably, NOx emissions were identified as the primary cause, particularly in VOC-limited regimes during autumn and winter. Additionally, underestimated NOx emissions caused the model to misrepresent the NO2-O3 relationship, leading to an underestimation of the spatial extent of VOC-limited regimes in the PRD. This study demonstrates that enhancing NOx emission estimates reduces O3 simulation biases in the PRD by 34% and enhances the representation of the NO2-O3 relationship.

Change analysis of surface water clarity in the Persian Gulf and the Oman Sea by remote sensing data and an interpretable deep learning model

The health of an ecosystem and the quality of water can be determined by the clarity of the water. The Persian Gulf and Oman Sea have a unique ecosystem, and monitoring their water clarity is necessary for sustainable development. Here, various criteria such as hue angle, chlorophyll-a, Forel-Ule index, organic carbon (OC), precipitation, sea surface salinity (SSS), Secchi disk depth (SDD), and sea surface temperature (SST) were analyzed from 2002 to 2018 using MODIS-Aqua Imagery, statistical tests, and deep learning (DL) models to monitor the water clarity of the Persian Gulf and the Oman Sea. The study found differences in criteria across different regions, with coastal areas showing higher Forel-Ule index and chlorophyll-a values. Positive trends in the Persian Gulf and the Oman Sea were attributed to the Forel-Ule index and OC, while negative trends were seen in SSS and SST in the Persian Gulf. The convolutional neural network (CNN) model was found to perform better than long short-term memory (LSTM) in predicting water clarity. Interpretation techniques were used to determine the importance of criteria in monitoring water clarity, with the Forel-Ule index, hue angle, and OC showing the greatest interaction. Sensitivity analysis revealed that chlorophyll-a and SSS had the most significant impact on water clarity prediction. Overall, this study using DL models and MODIS-Aqua Imagery can help improve water quality and protect the environment.

Predicting plateau atmospheric ozone concentrations by a machine learning approach: A case study of a typical city on the southwestern plateau of China

Atmospheric ozone (O3) has been placed on the priority control pollutant list in China's 14th Five-Year Plan. Due to their unique meteorological conditions, plateau regions contain high concentrations of atmospheric O3. However, traditional experimental methods for determining O3 concentrations using automatic monitoring stations cannot predict O3 trends. In this study, two machine learning models (a nonlinear auto-regressive model with external inputs (NARX) and a temporal convolution network (TCN)) were developed to predict O3 concentrations in a plateau area in the Kunming region by considering the effects of meteorological parameters, air quality parameters, and volatile organic compounds (VOCs). The plateau O3 prediction accuracy of the machine learning models was found to be much higher than those of numerical models that served as a comparison. The O3 values predicted by the machine learning models closely matched the actual monitoring data. The temporal distribution of plateau O3 displayed a high all-day peak from February to May. A correlation analysis between O3 concentrations and feature parameters demonstrated that humidity is the feature with the highest absolute correlation (-0.72), and was negatively correlated with O3 concentrations during all test periods. VOCs and temperatures were also found to have high positive correlation coefficients with O3 during periods of significant O3 pollution. After negating the effects of meteorological parameters, the predicted O3 concentrations decreased significantly, whereas they increased in the absence of NOx. Although individual VOCs were found to greatly affect the O3 concentration, the total VOC (TVOC) concentration had a relatively small effect. The proposed machine learning model was demonstrated to predict plateau O3 concentrations and distinguish how different features affect O3 variations.

A review on advancements in atmospheric microplastics research: The pivotal role of machine learning

Microplastics (MPs), recognized as emerging pollutants, pose significant potential impacts on the environment and human health. The investigation into atmospheric MPs is nascent due to the absence of effective characterization methods, leaving their concentration, distribution, sources, and impacts on human health largely undefined with evidence still emerging. This review compiles the latest literature on the sources, distribution, environmental behaviors, and toxicological effects of atmospheric MPs. It delves into the methodologies for source identification, distribution patterns, and the contemporary approaches to assess the toxicological effects of atmospheric MPs. Significantly, this review emphasizes the role of Machine Learning (ML) and Artificial Intelligence (AI) technologies as novel and promising tools in enhancing the precision and depth of research into atmospheric MPs, including but not limited to the spatiotemporal dynamics, source apportionment, and potential health impacts of atmospheric MPs. The integration of these advanced technologies facilitates a more nuanced understanding of MPs' behavior and effects, marking a pivotal advancement in the field. This review aims to deliver an in-depth view of atmospheric MPs, enhancing knowledge and awareness of their environmental and human health impacts. It calls upon scholars to focus on the research of atmospheric MPs based on new technologies of ML and AI, improving the database as well as offering fresh perspectives on this critical issue.

Determination of major drive of ozone formation and improvement of O3 prediction in typical North China Plain based on interpretable random forest model

O3 pollution in China has become prominent in recent years, and it has become one of the most challenging issues in air pollution control. We used data on atmospheric pollutants and meteorology from 2019 to 2021 to build an interpretable random forest (RF) model, applying this model to predict O3 concentration in 2022 in five cities in the Southwest North China Plain. The model was also used to identify and explain the influence of various factors on O3 formation. The correlation coefficient R2 between the predicted O3 concentration and observed O3 concentration was 0.82, the MAE was 15.15 μg/m3, and the RMSE was 20.29 μg/m3, indicating that the model can effectively predict O3 concentration in the studying area. The results of correlation analysis, feature importance, and the driving factor analysis from SHapley Additive exPlanations (SHAP) model indicated that temperature (T), NO2, and relative humidity (RH) are the top three features affecting O3 prediction, while the weights of wind speed and wind direction were relatively low. Thus, O3 in the southwestern North China Plain may mainly come from the formation of local photochemical activities. The dominant factors behind O3 also varied in different seasons. In spring and autumn, O3 pollution is more likely to occur under high NO2 concentration and high-temperature conditions, while in summer, it is more likely to occur under high-temperature and precipitation-free weather. In winter, NO2 is the dominant factor in O3 formation. Finally, the interpretable RF model is used to predict future O3 concentration based on features provided by Community Multiscale Air Quality (CMAQ) and Weather Research & Forecast (WRF) model, and the simulation performance of CMAQ on O3 concentration is enhanced to a certain extent, improving the prediction of future O3 pollution situations and guiding pollution control.

The application of machine learning to air pollution research: A bibliometric analysis

Machine learning (ML) is an advanced computer algorithm that simulates the human learning process to solve problems. With an explosion of monitoring data and the increasing demand for fast and accurate prediction, ML models have been rapidly developed and applied in air pollution research. In order to explore the status of ML applications in air pollution research, a bibliometric analysis was made based on 2962 articles published from 1990 to 2021. The number of publications increased sharply after 2017, comprising approximately 75% of the total. Institutions in China and United States contributed half of all publications with most research being conducted by individual groups rather than global collaborations. Cluster analysis revealed four main research topics for the application of ML: chemical characterization of pollutants, short-term forecasting, detection improvement and optimizing emission control. The rapid development of ML algorithms has increased the capability to explore the chemical characteristics of multiple pollutants, analyze chemical reactions and their driving factors, and simulate scenarios. Combined with multi-field data, ML models are a powerful tool for analyzing atmospheric chemical processes and evaluating the management of air quality and deserve greater attention in future.

Ozone response modeling to NOx and VOC emissions: Examining machine learning models

Current machine learning (ML) applications in atmospheric science focus on forecasting and bias correction for numerical modeling estimations, but few studies examined the nonlinear response of their predictions to precursor emissions. This study uses ground-level maximum daily 8-hour ozone average (MDA8 O₃) as an example to examine O₃ responses to local anthropogenic NOx and VOC emissions in Taiwan by Response Surface Modeling (RSM). Three different datasets for RSM were examined, including the Community Multiscale Air Quality (CMAQ) model data, ML-measurement-model fusion (ML-MMF) data, and ML data, which respectively represent direct numerical model predictions, numerical predictions adjusted by observations and other auxiliary data, and ML predictions based on observations and other auxiliary data. The results show that both ML-MMF (r = 0.93-0.94) and ML predictions (r = 0.89-0.94) present significantly improved performance in the benchmark case compared with CMAQ predictions (r = 0.41-0.80). While ML-MMF isopleths exhibit O₃ nonlinearity close to actual responses due to their numerical base and observation-based correction, ML isopleths present biased predictions concerning their different controlled ranges of O₃ and distorted O₃ responses to NOx and VOC emission ratios compared with ML-MMF isopleths, which implies that using data without support from CMAQ modeling to predict the air quality could mislead the controlled targets and future trends. Meanwhile, the observation-corrected ML-MMF isopleths also emphasize the impact of transboundary pollution from mainland China on the regional O₃ sensitivity to local NOx and VOC emissions, which transboundary NOx would make all air quality regions in April more sensitive to local VOC emissions and limit the potential effort by reducing local emissions. Future ML applications in atmospheric science like forecasting or bias correction should provide interpretability and explainability, except for meeting statistical performance and providing variable importance. Assessment with interpretable physical and chemical mechanisms and constructing a statistically robust ML model should be equally important.

A review of datasets and methods for deriving spatiotemporal distributions of atmospheric CO2

As the most abundant greenhouse gas, atmospheric carbon dioxide (CO₂) is considered one of the main attributors to climate change. Atmospheric CO₂ concentrations can be measured by ground-based monitoring networks, mobile monitoring campaigns, and carbon-observing satellites. However, the worldwide ground-based monitoring networks are composed of sparsely distributed sites and are inadequate to represent the spatiotemporal distributions of CO₂. Satellite-based remote sensing features repeated, long-term, and large-scale measurements, so it plays a crucial role in monitoring the global distributions of atmospheric CO₂. However, due to the presence of heavy clouds (or aerosols) and the limitation of satellite orbiting tracks, there exist large amounts of missing data in satellite retrievals. Various methods, including chemical transport models (CTMs), geostatistical methods, and regression-based models, have been employed to derive full-coverage spatiotemporal distributions of CO₂ based on the limited CO₂ measurements. This review summarizes the strengths and limitations of these methods. However, CTMs simulation results can have high uncertainty due to imperfect knowledge of the real world, and the interpolation accuracy of all geostatistical methods is limited by the large amount of data gaps in current satellite retrieved CO₂ products. To overcome these limitations, regression-based methods (especially machine learning models) have the ability to predict CO₂ with superior predictive performance, so this review also summarizes the framework of the machine learning approach. Leveraging the ongoing advancements of satellite instrumentation, the satellite-based CO₂ products have been improving dramatically in recent decades, and this review will describe and critically assess the advantages and disadvantages of the currently used systems in detail. For future improvements, we recommend the fusion of data from multiple satellite retrievals and CTMs by using machine learning algorithms in order to obtain even longer-term, larger-scale, finer-resolution, and higher-accuracy CO₂ datasets.

Machine learning may accelerate the recognition and control of microplastic pollution: Future prospects

Microplastics (MPs, sizes <5 mm) have been found to be widely distributed in various environments, such as marine, freshwater, terrestrial and atmospheric systems. Machine learning provides a potential solution for evaluating the ecological risks of MPs based on big data. Compared with traditional models, data-driven machine learning can accelerate the realization of the control of hazardous MPs and reduce the impact of MPs at both local and global scales. However, there are some urgent issues that should be resolved. For example, lack of MP databases and incomparable literatures causing the current MP data cannot fully support big data research. Therefore, it is imperative to formulate a set of standard and universal MP collection and testing protocols. For machine learning, predictions of large-scale MP distribution and the corresponding environmental risks remain lacking. To accelerate studies of MPs in the future, the methods and theories achieved for other particle pollutants, such as nanomaterials and aerosols, can be referenced. Beyond predication alone, the improvement of causality and interpretability of machine learning deserves attention in the studies of MP risks. Overall, this perspective paper provides insights for the development of machine learning methods in research on the environmental risks of MPs.

A Review on Rainfall Measurement Based on Commercial Microwave Links in Wireless Cellular Networks

As one of the most critical elements in the hydrological cycle, real-time and accurate rainfall measurement is of great significance to flood and drought disaster risk assessment and early warning. Using commercial microwave links (CMLs) to conduct rainfall measure is a promising solution due to the advantages of high spatial resolution, low implementation cost, near-surface measurement, and so on. However, because of the temporal and spatial dynamics of rainfall and the atmospheric influence, it is necessary to go through complicated signal processing steps from signal attenuation analysis of a CML to rainfall map. This article first introduces the basic principle and the revolution of CML-based rainfall measurement. Then, the article illustrates different steps of signal process in CML-based rainfall measurement, reviewing the state of the art solutions in each step. In addition, uncertainties and errors involved in each step of signal process as well as their impacts on the accuracy of rainfall measurement are analyzed. Moreover, the article also discusses how machine learning technologies facilitate CML-based rainfall measurement. Additionally, the applications of CML in monitoring phenomena other than rain and the hydrological simulation are summarized. Finally, the challenges and future directions are discussed.