Trajectory-user linking (TUL), whereby a trajectory is linked to its owner in location-based social networks, is a fundamental and critical task in spatio-temporal data mining. It plays a key role in personalized recommendation, anomaly detection, and semantic trajectory mining. Existing methods for TUL are either rule-based methods, which link trajectories and users based on conventional trajectory similarities, or learning-based methods, which learn a classification model to map trajectories to their owners. However, rule-based methods ignore the semantic information in the trajectory sequence, and learning-based methods require retraining the model each time a new user is added. In this paper, we propose a Siamese network-based model for trajectory-user linking (TULSN), which uses a Siamese network to capture semantic information in the trajectory, and instead of retraining the model, it requires only a few labeled trajectories per user to identify the user category of the trajectory. The experimental results show that the TULSN outperforms existing baselines and state-of-the-art methods on real-world datasets.

The global atmospheric chemical transport model has become a key technology for air quality forecast and management. However, precise and rapid air quality simulations and forecast are frequently limited by the model’s computational performance. The gas-phase chemistry module is the most time-consuming module in air quality models because its traditional solution method is dynamically stiff. To reduce the solving time of the gas phase chemical module, we built an emulator based on a deep residual neural network emulator (NN) for Carbon Bond Mechanism Z (CBM-Z) mechanism implemented in Global Nested Air Quality Prediction Modeling System. A global high resolution cross-life multi-species dataset was built and trained to evaluate multi-species concentration changes at a single time step of CBM-Z. The results showed that the emulator could accelerate to approximately 300–750 times while maintaining an accuracy similar to that of CBM-Z module (the average correlation coefficient squared was 0.97) at the global scale. This deep learning-based emulator could adequately represent the stiff kinetics of CBM-Z, which involves 47 species and 132 reactions. The emulated ozone (O₃), nitrogen oxides (NO_x), and hydroxyl radical (OH) were consistent with those of the original CBM-Z module in different global regions, heights, and time. Our results suggest that data-driven emulations have great potential in the construction of hybrid models with process-based air quality models, particularly at larger scales.

Secondary inorganic aerosol (SIA) is the key driving factor of fine-particle explosive growth (FPEG) events, which are frequently observed in North China Plain. However, the SIA simulations remain highly uncertain over East Asia. To further investigate this issue, SIA modeling over North China Plain with the 15 km resolution Nested Air Quality Prediction Model System (NAQPMS) was performed from October 2017 to March 2018. Surface observations of SIA at 28 sites were obtained to evaluate the model, which confirmed the biases in the SIA modeling. To identify the source of these biases and reduce them, uncertainty analysis was performed by evaluating the heterogeneous chemical reactions in the model and conducting sensitivity tests on the different reactions. The results suggest that the omission of the SO₂ heterogeneous chemical reaction involving anthropogenic aerosols in the model is probably the key reason for the systematic underestimation of sulfate during the winter season. The uptake coefficient of the “renoxification” reaction is a key source of uncertainty in nitrate simulations, and it is likely to be overestimated by the NAQPMS. Consideration of the SO₂ heterogeneous reaction involving anthropogenic aerosols and optimization of the uptake coefficient of the “renoxification” reaction in the model suitably reproduced the temporal and spatial variations in sulfate, nitrate and ammonium over North China Plain. The biases in the simulations of sulfate, nitrate, ammonium, and particulate matter smaller than 2.5 μm (PM_2.5) were reduced by 84.2%, 54.8%, 81.8%, and 80.9%, respectively. The results of this study provide a reference for the reduction in the model bias of SIA and PM_2.5 and improvement of the simulation of heterogeneous chemical processes.

Aerosol optical depth (AOD) measurements for 2017–2020 in urban Hohhot of the Mongolian plateau, a transition zone between the depopulated zone and East Asian urban agglomeration, were analyzed for the first time. Results show that annual AOD₅₀₀ and Ångström exponent α_440-675 were 0.36 ± 0.09 and 1.11 ± 0.16 (2017), 0.41 ± 0.12 and 0.90 ± 0.28 (2018), 0.38 ± 0.09 and 1.13 ± 0.24 (2019), 0.38 ± 0.12 and 1.17 ± 0.22 (2020), respectively, representing a slightly polluted level with a mixed type of coarse dust aerosol and a fine urban/industrial aerosol. Throughout the year, depopulated-zone continental air flows predominated in Hohhot (i.e., NW-quadrant wind), accounting for 82.12% (spring), 74.54% (summer), 63.61% (autumn), and 100% (winter). The clean and strong NW-quadrant air flows induced by the south movement of a Siberian anticyclone resulted in a low 500-nm AOD of 0.30 ± 0.29, 0.20 ± 0.15, 0.24 ± 0.29, and 0.13 ± 0.08 from spring to winter. Meanwhile, the local emissions from Hohhot city, as well as anthropogenic urban/industrial aerosols transported by southern and western air masses, originating from southern urban agglomeration and western industrial cities (Baotou, Wuhai, etc.), contributed to the highest aerosol loading, with significant transformation rates of the secondary aerosols Sulfate-Nitrate-Ammonium (SNA) of 47.45%, 57.39%, 49.88%, and 45.16–47.36% in PM_2.5 for each season. The extinction fraction of fine aerosols under these anthropogenic trajectories can be as high as 80%, and the largest fine aerosol size was around 0.2–0.25 μm. Dust aerosols were suspending in urban Hohhot all year, although at different levels for different seasons, and the extinction fraction of dust aerosol during sandstorms was generally higher than 70%.

Radar reflectivity data snapshot fine-grained atmospheric variations that cannot be represented well by numerical weather prediction models or satellites, which poses a limit for nowcasts based on model–data fusion techniques. Here, we reveal a multiscale representation (MSR) of the atmosphere by reconstructing the radar echoes from the Weather Research and Forecasting (WRF) model simulations and the Himawari-8 satellite products using U-Net deep networks. Our reconstructions generated the echoes well in terms of patterns, locations, and intensities with a root mean square error (RMSE) of 5.38 dBZ. We find stratified features in this MSR, with small-scale patterns such as echo intensities sensitive to the WRF-simulated dynamic and thermodynamic variables and with larger-scale information about shapes and locations mainly captured from satellite images. Such MSRs with physical interpretations may inspire innovative model–data fusion methods that could overcome the conventional limits of nowcasting.