Shihang Wang

To address the bottlenecks of weak cross-domain generalization, insufficient representation of complex spatial behaviors, and the scarcity of real anomaly labels in the detection of multi-semantic pedestrian trajectory anomalies (e.g., running, loitering, and wrong-way) in public security, this paper proposes an unsupervised detection framework combining kinematic constraints and a Bidirectional Long Short-Term Memory Autoencoder (Bi-LSTM-AE). While deep graph neural networks and attention-based models capture multi-agent interactions, they are frequently prone to spatial weight drift and severe computational overhead when transferred across disparate surveillance layouts. In contrast, our framework focuses on the intrinsic kinematic manifold of individual trajectories, striking an effective trade-off between cross-domain generalizability and linear computational complexity O(L). To this end, we construct a rigorous cross-domain evaluation paradigm, utilizing normal trajectories from a massive public dataset (SDD) to drive the unsupervised training of the model, aiming to fit a generalized human walking manifold unconstrained by specific physical scenes. To tackle the lack of multi-dimensional evaluation benchmarks, we design a multi-semantic anomaly generation strategy based on underlying kinematic constraints, programmatically synthesizing highly realistic challenging test samples on the ground-truth manifold of the target domain (UCY). Cross-domain experimental results demonstrate that the proposed framework can overcome the Gaussian smoothing trap of generative models and the cross-domain generalization degradation of complex graph networks. In highly challenging unseen scenes, the overall detection F1 score is maintained above 0.90; notably, in the detection of “spatial loop anomalies (loitering)” that relies on a global spatio-temporal receptive field, a precise recall rate of 93.76% is achieved. With linear computational complexity, our method provides a novel solution for highly robust intelligent video surveillance early-warning in the open world.

Land degradation is known to alter soil properties and quality; however, its depth-dependent effects across contrasting land-use types and the key factors limiting soil recovery remain poorly quantified in tropical ecosystems. This study established a forest degradation gradient on Hainan Island, China, encompassing mature forest, secondary forest, rubber plantation, and areca plantation. Soil physical (e.g., bulk density, porosity, water content, field capacity) and chemical (e.g., organic matter, nitrogen, phosphorus, and potassium fractions) properties were measured at three depths (0–20 cm, 20–40 cm, and 40–60 cm). A soil quality index (SQI) was constructed using principal component analysis, and obstacle degree modeling was applied to identify major limiting factors. The results showed that degradation of mature forests significantly reduced topsoil (0–20 cm) quality regardless of subsequent land-use type. In contrast, changes in medium (20–40 cm) and deep (40–60 cm) soil quality were land-use dependent. Conversion to secondary forests and areca plantations resulted in negligible effects, whereas transformation into rubber plantations significantly enhanced soil quality at medium and deep depths. Obstacle degree analysis identified available phosphorus, rather than nitrogen, as the primary limiting factor for soil quality in the region, accounting for 39.7% of all limitations across land-use types. This study demonstrates that the effects of tropical forest degradation on soil quality exhibit dual dependence on both soil depth and land-use type in tropical settings. Furthermore, it highlights the essential role of available phosphorus management in guiding soil restoration and sustainable land-use strategies in these vulnerable ecosystems.

Rapid urbanization inevitably exerts pressure on the surrounding ecological environment, and balancing the relationship between the ecological environment and urbanization is crucial for sustainable urban development. Taking the Yangtze River Delta urban agglomeration (YRDUA) as a case study, this paper utilizes MODIS data and nighttime light data to construct the MODIS Remote Sensing Ecological Index (MRSEI) and Comprehensive Nighttime Light Index (CNLI) distributions to depict ecological environment quality and urbanization levels. Based on this, the Coupled Coordination Degree (CCD) model is employed to calculate the coupling coordination level between the two, and the Geodetector is used to analyze the underlying causes affecting the CCD. The results indicate the following: (1) the overall ecological environment of the YRDUA tends to be stable, but there are significant differences between regions. Areas with deteriorating ecological conditions are concentrated in cities with higher rates of urbanization changes. (2) All cities are developing towards coordination, but there are imbalances in development among different regions. (3) The key factors affecting the CCD are derived from socioeconomic elements rather than natural elements, with the interaction between GDP and DEM having the strongest explanatory power for the CCD. (4) The CNLI is positively correlated with the CCD, the MRSEI is negatively correlated with the CCD, and the level of urbanization is the decisive factor for CCD changes. The research findings can provide theoretical guidance for promoting sustainable urban development.

Reactive nitrogen (Nr) has been confirmed as an indispensable nutrient for the city ecosystem, but high-intensity human activities have led to nitrogen pollution in cities, especially in coastal cities, jeopardizing ecosystem services and human health. Despite this, the characteristics and influencing factors of Nr remain unclear in coastal cities, particularly in the context of rapid urbanization. This study used the material flow analysis method to estimate Nr emissions in Xiamen from 1995 to 2018 and evaluated the characteristics of excessive Nr emissions. The STIRPAT model was used to identify and explore factors contributing to observed Nr levels in coastal cities. As indicated by the results, (1) the quantity of Nr generated by human activities increased 3.5 times from 1995 to 2018. Specifically, the total Nr entering the water environment showed a general increase with fluctuations, exhibiting an average annual growth rate of 3.1%, increasing from 17.2 Gg to 35.1 Gg. (2) Nr loads in the nearby sea increased notably from 8.1 Gg in 1995 to 25.4 Gg in 2018, suggesting a threefold augmentation compared with surface waters and groundwater. (3) NO_x was the gaseous Nr with the greatest effect on the atmosphere in Xiamen, which was primarily due to fossil fuel consumption. (4) Population and per capita GDP were major factors contributing to Nr load in the water environment, while Nr emission to the atmosphere was influenced by population and energy consumption. These findings provide valuable insights for tailored approaches to sustainable nitrogen management in coastal cities.

This study aimed to evaluate and compare the performances of the random forest (RF) and support vector regression (SVR) models combined with different feature selection methods, including recursive feature elimination (RFE), simulated annealing feature selection (SAFS), and selection by filtering (SBF) in predicting soil pH in Anhui Province, East China. We also used the ALL original features to build the RF and SVR models as a comparison. A total of 140 samples were selected, following the principles of randomness, uniformity, and representativeness, to consider the combination of landscape elements, such as topography, parent material, and land use. Auxiliary data, including climatic, topographic, and vegetation indexes, were used for predicting soil pH. The results showed that compared with the use the ALL original modeling features (ALL-RF, ALL-SVR), the combination of the three feature selection algorithms with RF and SVR can eliminate some redundant features and effectively improve the prediction accuracy of the soil pH model. For the RF model, the RMSE and the MAE of the calibration of the RFE-RF model were 0.73 and 0.57 and had the highest R² in four different RF models. The testing set of the RFE-RF model had an R² of 0.61, which was better than that of the ALL-RF (R² = 0.45) model and lower than those of the SAFS-RF (R² = 0.71) and SBF-RF (R² = 0.69) models. For the SVR model, the RFE-RF model was more robust and had better generalization ability. The accuracy of digital soil mapping can be improved through feature selection.

Analyzing land use changes (LUC) in both past and future scenarios is critical to optimize local ecology and formulate policies for sustainable development. We analyzed LUC characteristics in Huaibei City, China from 1985 to 2020, and used the CLUE-S and PLUS models to simulate LU in 2020. Then, we compared the accuracy of the simulation phase and chose the PLUS model to project LU under four scenarios in 2025. The results showed the following: (1) Farmland and grassland areas decreased from 1985 to 2020, while forestland, water, and construction land increased. (2) The LU types in the region are explained well by the driving factors, with all receiver operation characteristic (ROC) values greater than 0.8. (3) The kappa indices for CLUE-S and PLUS analog modeling were 0.727 and 0.759, respectively, with figure of merit (FOM) values of 0.109 and 0.201. (4) Under the farmland and ecological protection scenario (FEP), farmland and forestland areas are protected, increasing by 1727.91 hm² and 86.22 hm², respectively, while construction land decreases by 2001.96 hm². These results confirm that PLUS is significantly better than the CLUE-S model in modeling forestland and water, and slightly better than the CLUE-S model in modeling the rest of the LU type. Urban sustainability is strong in the scenario FEP.

Soil organic matter (SOM) is a key index of soil fertility. Visible and near-infrared (VNIR, 350–2500 nm) reflectance spectroscopy is an effective method for modeling SOM content. Characteristic wavelength screening and spectral transformation may improve the performance of SOM prediction. This study aimed to explore the optimal combination of characteristic wavelength selection and spectral transformation for hyperspectral modeling of SOM. A total of 219 topsoil (0–20 cm) samples were collected from two soil types in the East China. VNIR reflectance spectra were measured in the laboratory. Firstly, after spectral transformation (inverse-log reflectance (LR), continuum removal (CR) and first-order derivative reflectance (FDR)) of VNIR spectra, characteristic wavelengths were selected by competitive adaptive reweighted sampling (CARS) and uninformative variables elimination (UVE) algorithms. Secondly, the SOM prediction models were constructed based on the partial least squares regression (PLSR), random forest (RF) and support vector regression (SVR) methods using the full spectra and selected wavelengths, respectively. Finally, optimal SOM prediction models were selected for two soil types. The results were as follows: (1) The CARS algorithm screened 40–125 characteristic wavelengths from the full spectra. The UVE algorithm screened 105–884 characteristic wavelengths. (2) For two soil types and full spectra, CARS and UVE improved the SOM modeling precision based on the PLSR and SVR methods. The coefficient of determination (R²) value in the validation of the CARS-PLSR (PLSR model combined with CARS) and CARS-SVR (SVR model combined CARS) models ranged from 0.69 to 0.95, and the relative percent deviation (RPD) value ranged from 1.74 to 4.31. Lin’s concordance correlation coefficient (LCCC) values ranged from 0.83 to 0.97. The UVE-PLSR and UVE-SVR models showed moderate precision. (3) The PLSR and SVR modeling accuracies of Paddy soil were better than those for Shajiang black soil. RF models performed worse for both soil types, with the R² values of validation ranging from 0.22 to 0.68 and RPD values ranging from 1.01 to 1.60. (4) For Paddy soil, the optimal SOM prediction models (highest R² and RPD, lowest root mean square error (RMSE)) were CR-CARS-PLSR (R² and RMSE: 0.97 and 1.21 g/kg in calibration sets, 0.95 and 1.72 g/kg in validation sets, RPD: 4.31) and CR-CARS-SVR (R² and RMSE: 0.98 and 1.04 g/kg in calibration sets, 0.91 and 2.24 g/kg in validation sets, RPD: 3.37). For Shajiang black soil, the optimal SOM prediction models were LR-CARS-PLSR (R² and RMSE: 0.95 and 0.93 g/kg in calibration sets, 0.86 and 1.44 g/kg in validation sets, RPD: 2.62) and FDR-CARS-SVR (R² and RMSE: 0.99 and 0.45 g/kg in calibration sets, 0.83 and 1.58 g/kg in validation sets, RPD: 2.38). The results suggested that the CARS algorithm combined CR and FDR can significantly improve the modeling accuracy of SOM content.