Profiling Dust Mass Concentration in Northwest China Using a Joint Lidar and Sun-Photometer Setting

72-hour real-time forecasting of ambient PM2.5 by hybrid graph deep neural network with aggregated neighborhood spatiotemporal information

The observation-based air pollution forecasting method has high computational efficiency over traditional numerical models, but a poor ability in long-term (after 6 h) forecasting due to a lack of detailed representation of atmospheric processes associated with the pollution transport. To address such limitation, here we propose a novel real-time air pollution forecasting model that applies a hybrid graph deep neural network (GNN_LSTM) to dynamically capture the spatiotemporal correlations among neighborhood monitoring sites to better represent the physical mechanism of pollutant transport across the space with the graph structure which is established with features (angle, wind speed, and wind direction) of neighborhood sites to quantify their interactions. Such design substantially improves the model performance in 72-hour PM_2.5 forecasting over the whole Beijing-Tianjin-Hebei region (overall R² increases from 0.6 to 0.79), particularly for polluted episodes (PM_2.5 concentration > 55 µg/m³) with pronounced regional transport to be captured by GNN_LSTM model. The inclusion of the AOD feature further enhances the model performance in predicting PM_2.5 over the sites where the AOD can inform additional aloft PM_2.5 pollution features related to regional transport. The importance of neighborhood site (particularly for those in the upwind flow pathway of the target area) features for long-term PM_2.5 forecast is demonstrated by the increased performance in predicting PM_2.5 in the target city (Beijing) with the inclusion of additional 128 neighborhood sites. Moreover, the newly developed GNN_LSTM model also implies the "source"-receptor relationship, as impacts from distanced sites associated with regional transport grow along with the forecasting time (from 0% to 38% in 72 h) following the wind flow. Such results suggest the great potential of GNN_LSTM in long-term air quality forecasting and air pollution prevention.

Forecasting Fine Particulate Matter Concentrations by In-Depth Learning Model According to Random Forest and Bilateral Long- and Short-Term Memory Neural Networks

Accurate prediction of fine particulate matter concentration in the future is important for human health due to the necessity of an early warning system. Generally, deep learning methods, when widely used, perform better in forecasting the concentration of PM2.5. However, the source information is limited, and the dynamic process is uncertain. The method of predicting short-term (3 h) and long-term trends has not been achieved. In order to deal with the issue, the research employed a novel mixed forecasting model by coupling the random forest (RF) variable selection and bidirectional long- and short-term memory (BiLSTM) neural net in order to forecast concentrations of PM2.5/0~12 h. Consequently, the average absolute percentage error of 1, 6, and 12 h shows that the PM2.5 concentration prediction is 3.73, 9.33, and 12.68 μg/m3 for Beijing, 1.33, 3.38, and 4.60 μg/m3 for Guangzhou, 1.37, 4.19, and 6.35 μg/m3 for Xi’an, and 2.20, 7.75, and 10.07 μg/m3 for Shenyang, respectively. Moreover, the results show that the suggested mixed model is an advanced method that can offer high accuracy of PM2.5 concentrations from 1 to 12 h post.

24-Hour prediction of PM2.5 concentrations by combining empirical mode decomposition and bidirectional long short-term memory neural network

Accurate prediction of the future PM_2.5 concentration is crucial to human health and ecological environmental protection. Nowadays, deep learning methods show advantages in the prediction of PM_2.5 concentration, but few of the studies can achieve accurate prediction of short term (within 6 h) concentration and also catch longer term (6-24 h) change trends. To address this issue, this study constructs a novel hybrid prediction model by combining the empirical mode decomposition (EMD) method, sample entropy (SE) index and bidirectional long and short-term memory neural network (BiLSTM) to predict 0-24 h PM_2.5 concentrations. The experimental results show that the hybrid model has good performance on PM_2.5 prediction with R² = 0.987, RMSE = 2.77 μg/m³ at T + 1 moment and R² = 0.904, RMSE = 7.51 μg/m³ at T + 6 moment. The novel model improves the accuracy on short-term (within 6 h) prediction of PM_2.5 concentrations by at least 50% compared with other single deep learning models. Moreover, it well catches the variation trend of PM_2.5 concentrations after 6 h till 24 h.

CALIOP-Based Quantification of Central Asian Dust Transport

Central Asia is one of the most important sources of mineral saline dust worldwide. A comprehensive understanding of Central Asian dust transport is essential for evaluating its impacts on human health, ecological safety, weather and climate. This study first puts forward an observation-based climatology of Central Asian dust transport flux by using the 3-D dust detection of Cloud-Aerosol LiDAR with Orthogonal Polarization (CALIOP). The seasonal difference of transport flux and downstream contribution are evaluated and compared with those of the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). Central Asian dust can be transported not only southward in summer under the effect of the South Asian summer monsoon, but also eastward in other seasons under the control of the westerly jet. Additionally, the transport of Central Asian dust across the Pamir Plateau to the Tibetan Plateau is also non-negligible, especially during spring (with a transport flux rate of 150 kg m−1 day−1). The annual CALIOP-based downstream contribution of Central Asian dust to South Asian (164.01 Tg) is 2.1 times that to East Asia (78.36 Tg). This can be attributed to the blocking effect of the higher terrain between Central and East Asia. Additionally, the downstream contributions to South and East Asia from MERRA-2 are only 0.36 and 0.84 times that of CALIOP, respectively. This difference implies the overestimation of the wet and dry depositions of the model, especially in the low latitude zone. The quantification of the Central Asian dust transport allows a better understanding of the Central Asian dust cycle, and supports the calibration/validation of aerosol-related modules of regional and global climate models.

Polarization Lidar Measurements of Dust Optical Properties at the Junction of the Taklimakan Desert–Tibetan Plateau

Previous studies have shown that dust aerosols may accelerate the melting of snow and glaciers over the Tibetan Plateau. To investigate the vertical structure of dust aerosols, we conducted a ground-based observation by using multi-wavelength polarization lidar which is designed for continuous network measurements. In this study, we used the lidar observation from September to October 2020 at the Ruoqiang site (39.0°N, 88.2°E; 894 m ASL), located at the junction of the Taklimakan Desert–Tibetan Plateau. Our results showed that dust aerosols can be lifted up to 5 km from the ground, which is comparable with the elevation of the Tibetan Plateau in autumn with a mass concentration of 400–900 μg m−3. Moreover, the particle depolarization ratio (PDR) of the lifted dust aerosols at 532 nm and 355 nm are 0.34 ± 0.03 and 0.25 ± 0.04, respectively, indicating the high degree of non-sphericity in shape. In addition, extinction-related Ångström exponents are very small (0.11 ± 0.24), implying the large values in size. Based on ground-based lidar observation, this study proved that coarse non-spherical Taklimakan dust with high concentration can be transported to the Tibetan Plateau, suggesting its possible impacts on the regional climate and ecosystem.