数值模式预报臭氧污染

Monitoring the quality of air pollutant concentration forecasts based on chemical transport models (CTMs) currently used in the technology of the Hydrometcentre of Russia indicates the feasibility of the post-processing procedure application. For the first time, artificial neural networks (ANNs) were used to correct ground-level ozone model calculations. Retrospective hourly CTM CHIMERE forecasts for 2019‒2023 formed the training dataset. Experiments were carried out to select the optimal ANN settings. Results of the experimental testing of the best ANN on a week-long summer period with an episode of high ozone concentrations and a spring period with an episode of high ozone concentrations due to active tropospheric mixing are presented. The effectiveness of using ANNs to improve model forecasts of ground-level ozone and its daily dynamics is shown. Keywords: artificial neural networks, ground-level ozone, numerical forecast of pollution, chemical transport model, CHIMERE

Urban photochemical ozone (O3) formation regimes (NOx- and VOC-limited regimes) at nine megacities in East Asia were diagnosed based on near-surface O3 columns from 900 to 700 hPa, nitrogen dioxide (NO2), and formaldehyde (HCHO), which were inferred from measurements by ozone-monitoring instruments (OMI) for 2014–2018. The nine megacities included Beijing, Tianjin, Hebei, Shandong, Shanghai, Seoul, Busan, Tokyo, and Osaka. The space-borne HCHO–to–NO2 ratio (FNR) inferred from the OMI was applied to nine megacities and verified by a series of sensitivity tests of Weather Research and Forecasting model with Chemistry (WRF-Chem) simulations by halving the NOx and VOC emissions. The results showed that the satellite-based FNRs ranged from 1.20 to 2.62 and the regimes over the nine megacities were identified as almost NOx-saturated conditions, while the domain-averaged FNR in East Asia was >2. The results of WRF–Chem sensitivity modeling show that O3 increased when the NOx emissions reduced, whereas VOC emission reduction showed a significant decrease in O3, confirming the characteristics of VOC-limited conditions in all of the nine megacities. When both NOx and VOC emissions were reduced, O3 decreased in most cities, but increased in the three lowest-FNRs megacities, such as Shanghai, Seoul, and Tokyo, where weakened O3 titration caused by NOx reduction had a larger enough effect to offset O3 suppression induced by the decrease in VOCs. Our model results, therefore, indicated that the immediate VOC emission reduction is a key controlling factor to decrease megacity O3 in East Asia, and also suggested that both VOC and NOx reductions may not be of broad utility in O3 abatement in megacities and should be considered judiciously in highly NOx-saturated cities in East Asia.

No abstract available

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.

Issues regarding air quality and related health concerns have prompted this study, which develops an accurate and computationally fast, efficient hybrid modeling system that combines numerical modeling and machine learning for forecasting concentrations of surface ozone. Currently available numerical modeling systems for air quality predictions (e.g., CMAQ) can forecast 24 to 48 h in advance. In this study, we develop a modeling system based on a convolutional neural network (CNN) model that is not only fast but covers a temporal period of two weeks with a resolution as small as a single hour for 255 stations. The CNN model uses meteorology from the Weather Research and Forecasting model (processed by the Meteorology-Chemistry Interface Processor), forecasted air quality from the Community Multi-scale Air Quality Model (CMAQ), and previous 24-h concentrations of various measurable air quality parameters as inputs and predicts the following 14-day hourly surface ozone concentrations. The model achieves an average accuracy of 0.91 in terms of the index of agreement for the first day and 0.78 for the fourteenth day, while the average index of agreement for one day ahead prediction from the CMAQ is 0.77. Through this study, we intend to amalgamate the best features of numerical modeling (i.e., fine spatial resolution) and a deep neural network (i.e., computation speed and accuracy) to achieve more accurate spatio-temporal predictions of hourly ozone concentrations. Although the primary purpose of this study is the prediction of hourly ozone concentrations, the system can be extended to various other pollutants.

No abstract available

No abstract available

Abstract. Atmospheric oxidation capacity (AOC) characterizes the ability of the atmosphere to scavenge air pollutants. However, the processes involved in China, where anthropogenic emissions have changed dramatically in the past decade, are not fully understood. A detailed analysis of different parameters that determine the AOC in China is presented on the basis of numerical simulations performed with the regional chemical–meteorological Weather Research and Forecasting model with Chemistry (WRF-Chem). The model shows that the aerosol effects related to extinction and heterogeneous processes produce a decrease in surface ozone of approximately 8–10 ppbv in NOx-limited rural areas and an increase of 5–10 ppbv in VOC-limited urban areas. In this latter case, the ozone increase is noticeable for aerosol concentrations ranging from 20 to 45 µg m−3 in July 2018. The ozone reduction in NOx-sensitive regions is due to the combined effect of nitrogen dioxide and peroxy radical uptake on particles and of the light extinction by aerosols, which affects the photodissociation rates. The ozone increase in VOC-sensitive areas is attributed to the uptake of NO2 by aerosols, which is offset by the reduced ozone formation associated with HO2 uptake and with aerosol extinction. Our study concludes that more than 90 % of the daytime AOC is due to the reaction of the hydroxyl radical with VOCs and carbon monoxide. In urban areas, during summertime, the main contributions to daytime AOC are the reactions of OH with alkene (30 %–50 %), oxidized volatile organic compounds (OVOCs) (33 %–45 %), and carbon monoxide (20 %–45 %). In rural areas, the largest contribution results from the reaction of OH with alkenes (60 %). Nocturnal AOC is dominantly attributed to the reactions with the nitrate radical (50 %–70 %). Our results shed light on the contribution of aerosol-related NOx loss and the high reactivity of alkenes for photochemical pollution. With the reduction in aerosols and anthropogenic ozone precursors, the chemistry of nitrogen and temperature-sensitive VOCs will become increasingly important. More attention needs to be paid to the role of photodegradable OVOCs and nocturnal oxidants in the formation of secondary pollutants.

No abstract available

Valley topography is recognized for its role in constraining pollutant dispersion, which frequently results in elevated pollutant concentrations within valley regions. However, the specific mechanisms by which valley topography influences daytime ozone (O3) production, nighttime O3 depletion, and diurnal variations in O3 concentrations remain inadequately understood. This study employs the online WRF-Chem air quality model to conduct sensitivity analyses, examining the effects of valley topography on summer O3 concentrations in Lanzhou, a valley city in western China. The results indicate that valley topography generally reduces surface temperature and wind speed while also lowering the planetary boundary layer height (PBLH), which ultimately leads to increased concentrations of primary pollutants. Nevertheless, the impact of valley terrain on O3 is time-dependent, with concentrations decreasing during the day and increasing at night. In Lanzhou, valley topography contributes to a 4.4% reduction in daytime O3 concentrations. This reduction is largely due to the blocking of solar radiation by surrounding mountains, which results in a 7% decrease in shortwave radiation (SR) and a 17% decline in PBLH, subsequently limiting O3 chemical production. Although valley topography restricts pollutant dispersion, the overall effect during the day is a net decrease in O3 levels. In contrast, nighttime valley topography leads to a notable 19.8% increase in O3 concentrations. This increase is driven by mountain breezes transporting O3-rich air from the slopes into urban areas, along with a 67.4% reduction in nitrogen oxide (NO) levels. Despite intensified NO titration within the valley, the combined effect of these local wind patterns contributes to elevated O3 concentrations at night. These findings offer valuable insights into the unique factors contributing to urban O3 pollution in areas with complex valley topography.

The governing equations in atmospheric and ocean computational fluid dynamics models are specific forms of Navier-Stokes equations. Due to the complexity of the model itself, it is difficult to find the optimal parallel solution method. By coupling the atmospheric-ocean numerical model and simulating the interaction between the atmospheric-ocean fluids, the accuracy and prediction ability of the prediction model can be improved, but a more complex numerical calculation is needed. Based on the newly developed Model for Prediction Across Scales atmosphere model (MPAS-atmosphere, MPAS-A) and NOAA/EMC WAVEWATCH III (NWW3) wave model, we use the C-Coupler2 coupler to realize the construction of the global atmosphere-wave coupling numerical forecasting model. Due to the urgent need for operational timeliness of forecast, the parallel performance of the operating system has been analyzed. The parallel I/O, compiling options, parallel partitioning, and other optimizations were tested after the system deployment on the Lenovo cluster of the National Marine Environmental Forecasting Center. After optimization, a well-balanced performance is obtained, and computing and I/O resources are reasonably utilized, thus laying a foundation for the real-time coupling forecasting operation system of global across scales high-resolution atmospheric and ocean wave model.

Chemical transport models (CTMs) are widely used for air quality forecasts, but these models require large computational resources and often suffer from a systematic bias that leads to missed poor air pollution events. For example, a CTM-based operational forecasting system for air quality over the Pacific Northwest, called AIRPACT, uses over 100 processors for several hours to provide 48-h forecasts daily, but struggles to capture unhealthy O3 episodes during the summer and early fall, especially over Kennewick, WA. This research developed machine learning (ML) based O3 forecasts for Kennewick, WA to demonstrate an improved forecast capability. We used the 2017–2020 simulated meteorology and O3 observation data from Kennewick as training datasets. The meteorology datasets are from the Weather Research and Forecasting (WRF) meteorological model forecasts produced daily by the University of Washington. Our ozone forecasting system consists of two ML models, ML1 and ML2, to improve predictability: ML1 uses the random forest (RF) classifier and multiple linear regression (MLR) models, and ML2 uses a two-phase RF regression model with best-fit weighting factors. To avoid overfitting, we evaluate the ML forecasting system with the 10-time, 10-fold, and walk-forward cross-validation analysis. Compared to AIRPACT, ML1 improved forecast skill for high-O3 events and captured 5 out of 10 unhealthy O3 events, while AIRPACT and ML2 missed all the unhealthy events. ML2 showed better forecast skill for less elevated-O3 events. Based on this result, we set up our ML modeling framework to use ML1 for high-O3 events and ML2 for less elevated O3 events. Since May 2019, the ML modeling framework has been used to produce daily 72-h O3 forecasts and has provided forecasts via the web for clean air agency and public use: http://ozonematters.com/. Compared to the testing period, the operational forecasting period has not had unhealthy O3 events. Nevertheless, the ML modeling framework demonstrated a reliable forecasting capability at a selected location with much less computational resources. The ML system uses a single processor for minutes compared to the CTM-based forecasting system using more than 100 processors for hours.

Surface ozone is one of six air pollutants designated as harmful by National Ambient Air Quality Standards because it can adversely impact human health and the environment. Thus, ozone forecasting is a critical task that can help people avoid dangerously high ozone concentrations. Conventional numerical approaches, as well as data-driven forecasting approaches, have been studied for ozone forecasting. Data-driven forecasting models, in particular, have gained momentum with the introduction of machine learning advancements. We consider planetary boundary layer (PBL) height as a new input feature for data-driven ozone forecasting models. PBL has been shown to impact ozone concentrations, making it an important factor in ozone forecasts. In this paper, we investigate the effectiveness of utilization of PBL height on the performance of surface ozone forecasts. We present both surface ozone forecasting models, based on multilayer perceptron (MLP) and bidirectional long short-term memory (LSTM) models. These two models forecast hourly ozone concentrations for an upcoming 24-h period using two types of input data, such as measurement data and PBL height. We consider the predicted values of PBL height obtained from the weather research and forecasting (WRF) model, since it is difficult to gather actual PBL measurements. We evaluate two ozone forecasting models in terms of index of agreement (IOA), mean absolute error (MAE), and root mean square error (RMSE). Results showed that the MLP-based and bidirectional LSTM-based models yielded lower MAE and RMSE when considering forecasted PBL height, but there was no significant changes in IOA when compared with models in which no forecasted PBL data were used. This result suggests that utilizing forecasted PBL height can improve the forecasting performance of data-driven prediction models for surface ozone concentrations.

The ability to forecast the concentration of air pollutants in an urban region is crucial for decision-makers wishing to reduce the impact of pollution on public health through active measures (e.g. temporary traffic closures). In this study, we present a machine learning approach applied to the forecast of the day-ahead maximum value of the ozone concentration for several geographical locations in southern Switzerland. Starting from a dataset containing thousands of historical air quality and weather data as well as numerical weather predictions, the most relevant features are selected using a genetic algorithm and then used to train a number of regression models. After assessing that forcing engineered features suggested by experts in the domain into the initial population of the genetic algorithm does not increase the final forecasters' accuracy, we adopted a procedure entirely agnostic for atmospheric physics. We then used Shapley values to explain the learned models in terms of feature importance and feature interactions in relation to ozone predictions. Our analysis suggests that the trained models effectively learned explanatory cross-dependencies among atmospheric variables, which are described in the ozone photochemistry literature.

No abstract available