Rapid Inference of Nitrogen Oxide Emissions Based on a Top-Down Method with a Physically Informed Variational Autoencoder

A Physically Constrained Deep-Learning Fusion Method for Estimating Surface NO2 Concentration from Satellite and Ground Monitors

Accurate estimation of atmospheric chemical concentrations from multiple observations is crucial for assessing the health effects of air pollution. However, existing methods are limited by imbalanced samples from observations. Here, we introduce a novel deep-learning model-measurement fusion method (DeepMMF) constrained by physical laws inferred from a chemical transport model (CTM) to estimate NO2 concentrations over the Continental United States (CONUS). By pretraining with spatiotemporally complete CTM simulations, fine-tuning with satellite and ground measurements, and employing a novel optimization strategy for selecting proper prior emission, DeepMMF delivers improved NO2 estimates, showing greater consistency and daily variation alignment with observations (with NMB reduced from -0.3 to -0.1 compared to original CTM simulations). More importantly, DeepMMF effectively addressed the sample imbalance issue that causes overestimation (by over 100%) of downwind or rural concentrations in other methods. It achieves a higher R2 of 0.98 and a lower RMSE of 1.45 ppb compared to surface NO2 observations, overperforming other approaches, which show R2 values of 0.4-0.7 and RMSEs of 3-6 ppb. The method also offers a synergistic advantage by adjusting corresponding emissions, in agreement with changes (-10% to -20%) reported in the NEI between 2019 and 2020. Our results demonstrate the great potential of DeepMMF in data fusion to better support air pollution exposure estimation and forecasting.

Perspectives on Advancing Multimodal Learning in Environmental Science and Engineering Studies

The environment faces increasing anthropogenic impacts, resulting in a rapid increase in environmental issues that undermine the natural capital essential for human wellbeing. These issues are complex and often influenced by various factors represented by data with different modalities. While machine learning (ML) provides data-driven tools for addressing the environmental issues, the current ML models in environmental science and engineering (ES&E) often neglect the utilization of multimodal data. With the advancement in deep learning, multimodal learning (MML) holds promise for comprehensive descriptions of the environmental issues by harnessing data from diverse modalities. This advancement has the potential to significantly elevate the accuracy and robustness of prediction models in ES&E studies, providing enhanced solutions for various environmental modeling tasks. This perspective summarizes MML methodologies and proposes potential applications of MML models in ES&E studies, including environmental quality assessment, prediction of chemical hazards, and optimization of pollution control techniques. Additionally, we discuss the challenges associated with implementing MML in ES&E and propose future research directions in this domain.

Development of surface observation-based two-step emissions adjustment and its application on CO, NOx, and SO2 emissions in China and South Korea

It is challenging to estimate local emission conditions of a downwind area solely based on concentrations in the downwind area. This is because air pollutants that have a long residence time in the atmosphere can be transported over long distances and influence air quality in downwind areas. In this study, a Two-step Emissions Adjustment (TEA) approach was developed to adjust downwind emissions of target air pollutants with surface observations, considering their long-range transported emission impacts from upwind areas calculated from air quality simulations. Using the TEA approach, CO, NOx, and SO2 emissions were adjusted in China and South Korea between 2016 and 2021 based on existing bottom-up emissions inventories. Simulations with the adjusted emissions showed that the 6-year average normalized mean biases of the monthly mean concentrations of CO, NOx, and SO2 improved to 0.3 %, -2 %, and 2 %, respectively, in China, and to 5 %, 7 %, and 4 %, respectively, in South Korea. When analyzing the emission trends, it was estimated that the annual emissions of CO, NOx, and SO2 in China decreased at a rate of 7.2 %, 4.5 %, and 10.6 % per year, respectively. The decrease rate of emissions for each of these pollutants was similar to that of ambient concentrations. When considering upwind emission impacts in the emissions adjustment, CO emissions increased by 1.3 %/year in South Korea, despite CO concentrations in the country decreasing during the study period. During the study period, NOx and SO2 emissions in South Korea decreased by 3.9 % and 0.5 %/year, respectively. Moreover, the TEA approach can account for drastic short-term emission changes (e.g., social distancing due to COVID-19). Therefore, the TEA approach can be used to adjust emissions and improve reproducibility of concentrations of air pollutants suitable for health studies for areas where upwind emission impacts are significant.

Advancing shipping NOx pollution estimation through a satellite-based approach

Estimating shipping nitrogen oxides (NOx) emissions and their associated ambient NO2 impacts is a complex and time-consuming task. In this study, a satellite-based ship pollution estimation model (SAT-SHIP) is developed to estimate regional shipping NOx emissions and their contribution to ambient NO2 concentrations in China. Unlike the traditional bottom-up approach, SAT-SHIP employs satellite observations with varying wind patterns to improve the top-down emission inversion methods for individual sectors amidst irregular emission plume signals. Through SAT-SHIP, shipping NOx emissions for 17 ports in China are estimated. The results show that SAT-SHIP performed comparably with the bottom-up approach, with an R2 value of 0.8. Additionally, SAT-SHIP reveals that the shipping sector in port areas contributes ∼21 and 11% to NO2 concentrations in the Yangtze River Delta and Pearl River Delta areas of China, respectively, which is consistent with the results from chemical transportation model simulations. This approach has practical implications for policymakers seeking to identify pollution sources and develop effective strategies to mitigate air pollution.

Development of a recurrent spatiotemporal deep-learning method coupled with data fusion for correction of hourly ozone forecasts

Ambient ozone (O3) predictions can be very challenging mainly due to the highly nonlinear photochemistry among its precursors, and meteorological conditions and regional transport can further complicate the O3 formation processes. The emission-based chemical transport models (CTM) are broadly used to predict O3 formation, but they may deviate from observations due to input uncertainties such as emissions and meteorological data, in addition to the treatment of O3 nonlinear chemistry. In this study, an innovative recurrent spatiotemporal deep-learning (RSDL) method with model-monitor coupled convolutional recurrent neural networks (ConvRNN) has been developed to improve O3 predictions of CTM. The RSDL method was first used to build the ConvRNN within a 24-h scale to characterize the spatiotemporal relationships between the monitored O3 data and CTM simulations, and then incorporated the recurrent pattern to achieve 72-h multi-site forecasts based on a pilot study over the Pearl River Delta (PRD) region of China. The results showed that the RSDL method predicted O3 with high accuracy over this case study, with an increase of 27.54% in the correlation coefficient (R) average for all sites as well as an increase in R of 0.14-0.21 for all cities compared to CTM. Moreover, the regional distribution of CTM was further improved by the RSDL predictions with the data fusion technique, which greatly reduced the underpredictions of O3 concentrations, particularly in high O3-level areas (concentrations >160 μg/m3), with a 33.55% reduction in the mean absolute error (MAE).