Search Results (92)

Spectrum prediction is essential for implementing dynamic spectrum management and mitigating spectrum congestion. However, spectrum data in real electromagnetic environments exhibit high non-stationarity, multi-scale features, and complex non-Euclidean spatio-temporal coupling characteristics, which limit the prediction accuracy of existing models. To address these issues, this paper proposes an attention-based spectrum prediction method for spatio-temporal feature synergy (SSA-A-BiGCRNN). First, Singular Spectrum Analysis (SSA) is introduced to decompose and reconstruct the non-stationary spectrum signals, filtering out high-frequency burst noise and extracting core evolutionary trends. Second, a spatial topology graph among multiple frequency bands is constructed based on the Spearman rank correlation coefficient. A Bidirectional Graph Convolutional Recurrent Neural Network is then designed to simultaneously capture the spatial dependencies between frequency bands and the bidirectional evolutionary patterns in the time dimension. Finally, an attention mechanism is incorporated during the feature fusion stage to evaluate and focus on critical spatio-temporal information, further enhancing global prediction accuracy. Experimental results based on a real electromagnetic monitoring dataset demonstrate that the proposed model achieves an accuracy of 96.82%, a coefficient of determination ($R^2$) of 0.9966, a Root Mean Square Error (RMSE) of 0.5597, and a Mean Absolute Error (MAE) of 0.4031, significantly outperforming existing models. Full article

Over the past three decades, geographically weighted regression (GWR) and multiscale geographically weighted regression (MGWR) have become essential tools for spatial analysis in urban, environmental, and land-use research. This scoping review systematically maps and synthesizes the global literature on GWR and MGWR published between 1996 and 2026, aiming to identify the research hotspots, evolutionary paths, and cutting-edge trends. Bibliometrics and CiteSpace visualization tools are used to conduct a multi-dimensional visual analysis of thousands of selected articles, including countries, institutions, core authors, highly cited keywords, and key documents. The results show that the current research focuses on spatial heterogeneity, multiscale analysis, GWR model optimization, non-stationarity characterization, and simulation of urban land-use change. Potential future directions include the construction of spatiotemporal integrated models, the integration of high-performance computing, and the expansion of interdisciplinary applications. The results of this study can help scholars fully understand the current research status and future directions, and provide a scientific spatial analysis framework for practitioners in urban planning, land resource management, and environmental assessment. Furthermore, the conclusions can provide theoretical support and a decision-making basis for the government to formulate intelligent and refined urban development policies. Full article

Traditional linear models in urban informatics struggle to capture the complex, non-linear spatial non-stationarity inherent in metropolitan housing markets. To overcome these constraints, this study introduces a data-driven computational framework integrating a Spatio-Temporal Graph Convolutional Network (STGCN) with gradient-based Explainable Artificial Intelligence (XAI) and Geographically Weighted Regression (GWR). This framework is empirically tested using 217,598 apartment transactions in Busan, the Republic of Korea, augmented with high-resolution micro-demographic grids and Digital Elevation Model (DEM) topographical data. Utilizing unsupervised K-Means clustering, the region is spatially stratified into a dense Urban Core and a dispersed Suburban Periphery. The STGCN demonstrates overwhelming predictive superiority ($R^2=0.802$) over the traditional Spatial Error Model ($R^2=0.437$). Crucially, gradient-based XAI and localized GWR coefficients successfully unspool the deep learning “black box,” visualizing hyper-localized economic realities that global linear models obscure. The analysis expose stark regional market segmentation driven by environmental topography, mathematically quantifying non-linear dynamics such as coastal high-floor premiums, severe mountainous altitude penalties, and latent urban reconstruction premiums. Ultimately, this research bridges the gap between predictive computational power and spatial economic interpretability, offering a robust informatics framework for equitable urban planning. Full article

Rapid urbanization and complex topography complicate Urban Heat Island (UHI) spatio-temporal dynamics. Traditional models and coarse-resolution imagery often fail to capture fine-scale, spatially non-stationary seasonal driving mechanisms. This study investigates the multi-dimensional drivers of surface thermal dynamics in Kunming, a typical low-latitude plateau city, using seasonal median LST composite (2018–2025). Integrating eXtreme Gradient Boosting (XGBoost) with eXplainable Artificial Intelligence (XAI) models decoupled the nonlinear impacts of these drivers. Results reveal a seasonal thermal dichotomy: Summer exhibits the most intense UHI effect with extreme peak temperatures, while Spring presents an anomaly where natural and vegetated Local Climate Zones (LCZs) show pronounced warming. SHapley Additive exPlanations (SHAP) analysis identified a seasonal rotation: anthropogenic and structural factors dominate Summer and Autumn warming, whereas natural and topographic regulators govern Spring and Winter. GeoShapley deconstruction demonstrated strong spatial non-stationarity. Building-density warming is amplified in poorly ventilated urban cores, and fragmented vegetation’s cooling is offset by anthropogenic heat during peak summer. This study provides new insights into the seasonal drivers of urban thermal environments in plateau cities. Full article

Accurate regional wind speed forecasting is critical yet challenging due to inherent spatiotemporal correlations and data non-stationarity. This paper proposes a hybrid framework combining Principal Component Analysis (PCA), Variational Mode Decomposition (VMD), and Long Short-Term Memory (LSTM) networks. First, PCA extracts dominant spatial features from a regional wind field (9 × 9 grid), retaining 99.5% of the information to reduce redundancy. Next, an adaptive VMD strategy, optimized by the Sparrow Search Algorithm (SSA), decomposes these components to mitigate temporal non-stationarity. High-correlation sub-signals are then fed into the LSTM predictor. Experimental results demonstrate that the framework achieves an average coefficient of determination ($R^2$) of approximately 0.41 in the first forecasting step. Crucially, it significantly mitigates error accumulation in multi-step forecasting, maintaining a stable $R^2$ of 0.39 in the third step. Conversely, complex spatiotemporal models like ConvLSTM achieve high initial accuracy but suffer severe degradation ($R^2$ dropping from 0.70 to 0.24) alongside significantly higher computational overhead. The proposed strategy effectively prevents overfitting to high-frequency noise, ensuring a computationally efficient and robust solution for multi-step regional wind forecasting. Full article

Gas concentration series in coal mining faces are jointly affected by multiple coupled factors, including geological conditions, mining disturbances, ventilation organization, and gas drainage intensity, and therefore exhibit pronounced nonstationarity, strong fluctuations, spatiotemporal correlations across multiple monitoring points, and occasional abrupt spikes. To address these challenges, this study proposes a gas concentration prediction and early-warning method that integrates CEEMDAN–SST with GraphPINN-TimesFM (Graph Physics-Informed Neural Network–Time Series Foundation Model). First, based on multi-source monitoring data such as wind speed, gas concentrations at multiple monitoring points, and equipment operating status, anomaly removal, operating-condition segmentation, and change-point detection are performed to construct stable operating-state labels. Feature selection is then conducted by combining optimal time-lag correlation, Shapley value contribution, and dynamic time warping. Second, WGAN-GP is employed to augment samples from minority operating conditions, while CEEMDAN–SST is used to decompose and reconstruct the target series so as to reduce the interference of nonstationary noise and enhance sequence predictability. On this basis, TimesFM is adopted as the backbone for long-sequence forecasting to capture long-term dependency features in gas concentration evolution. Furthermore, GraphPINN is introduced to embed the topological associations among monitoring points, airflow transmission delays, and convection–diffusion mechanisms into the training process, thereby enabling collaborative modeling that integrates data-driven learning with physical constraints. Finally, the predictive performance, early-warning capability, and interpretability of the proposed model are systematically evaluated through regression forecasting, warning discrimination, and Shapley-based interpretability analysis. The results demonstrate that the proposed method can effectively improve the accuracy, robustness, and physical consistency of gas concentration prediction under complex operating conditions, thereby providing a new technical pathway for gas over-limit early warning and safety regulation in coal mining faces. Full article

This study evaluates the performance of a semi-dynamic negative binomial model with cubic spline smoothing to characterize the spatiotemporal dynamics of COVID-19 mortality in Peru, a setting marked by significant data inconsistency and reporting delays. Using nationwide weekly mortality data, we compared a Poisson regression against a semi-dynamic NB model with a population offset and cubic splines (df = 6). The models were evaluated using Akaike Information Criterion and log-likelihood to handle overdispersion and temporal non-stationarity. The NB model demonstrated a superior fit, reducing the AIC from 136,596.4 to 75,668.25 and improving log-likelihood by over 30,000 points. Demographic analysis revealed an 81.6% higher risk of death in males (IRR = 1.816; 95% CI: 1.753–1.881) and an exponential gradient with age, peaking at an IRR of 4.717 (95% CI: 4.499–4.945) for individuals ≥80 years. Departmental fixed effects identified significant spatial heterogeneity, with higher diffusion in coastal regions. The semi-dynamic NB model with splines provides a robust, parsimonious, and scalable framework for epidemiological surveillance in resource-limited settings. By effectively correcting for overdispersion and stabilizing weekly reporting fluctuations, this approach offers a reliable tool for public health decision making in environments with fragmented data quality. Full article

In response to the problem that sensors cannot be directly installed at key local positions on the surface of ship hull structures during the transient strong shock process of underwater explosions due to spatial constraints or large plastic deformations, this paper investigates the chaotic-like nonlinear transient behavior of structural dynamic response systems under strong shock and proposes a key position structural response reconstruction method based on dynamic inversion. Since the structural response under a transient strong shock exhibits significant non-stationarity and nonlinearity, signals from neighboring measurement points cannot directly characterize the dynamic behavior at key positions. Therefore, the shock response signals are discretized in both time and space dimensions. The phase space reconstruction method is employed to characterize the motion trajectory of acceleration responses in a two-dimensional phase space, establish mapping functions for system motion evolution, and use their control parameters to characterize the system’s nonlinear dynamic behavior. Furthermore, based on the spatiotemporal dynamic equations, a spatiotemporal coupled mapping model for spatial state points is established to achieve the theoretical inversion of acceleration responses at key positions. This method provides theoretical support for analyzing the dynamic characteristics of structures at key positions under strong shock environments, characterizing the shock environment, and assessing and designing equipment for shock safety. However, the current validation is based on high-fidelity numerical simulations rather than physical prototype tests; therefore, the predictive capability of this method in actual physical environments requires further validation through subsequent physical model tests. Full article

With the rapid development of e-commerce and smart manufacturing, automated warehouse systems have become critical infrastructure for modern logistics. In China’s vast market, the dynamic scheduling of Rail-Guided Vehicles (RGVs) faces significant challenges due to complex task uncertainties, hierarchical supply chain structures, and real-time collision avoidance requirements. Traditional rule-based methods and static optimization models often fail to adapt to such dynamic environments. To address these issues, this paper proposes a novel hybrid deep reinforcement learning framework integrating a Dynamic Graph Neural Network (DGNN) and a Transformer model. The DGNN captures the spatiotemporal dependencies of the warehouse network topology, while the Transformer mechanism enhances long-range feature extraction for task prioritization. Furthermore, we design a centralized Deep Q-network (DQN) framework with parameterized action spaces to coordinate multiple RGVs collaboratively. While the system manages multiple physical vehicles, the learning architecture employs a single-agent global scheduler to avoid the non-stationarity issues inherent in multi-agent reinforcement learning. Experimental results based on real-world data from a large-scale electronics manufacturing warehouse demonstrate that our method reduces average task completion time by 18.5% and improves system throughput by 22.3% compared to state-of-the-art baselines. The proposed approach demonstrates potential for intelligent warehouse management in dynamic industrial scenarios. Full article

Land use changes within urban agglomerations exhibit significant spatiotemporal heterogeneity and regional diversity. In urban agglomeration land simulation, traditional models often struggle to systematically capture these variations. We introduce the GCTP, a novel framework that integrates guided Geographical zoning, Convolutional Neural Networks (CNN)-Transformer, and the Patch-generating Land Use Simulation (PLUS) model. Initially, guided K-means clustering was employed for geographic zoning to characterize regional spatial non-stationarity. Then, a CNN-Transformer network leveraged self-attention mechanisms to capture multi-scale spatial correlations, obtaining pixel-level development probabilities. Finally, these probabilities were fused with PLUS- Land Expansion Analysis Strategy (LEAS) outputs to drive PLUS- Cellular Automata with multi-type Random Seeds (CARS) for patch-level simulation. The results demonstrate the following: (1) The embedding of guided zoning enabled the model to achieve an Overall Accuracy (OA) of 0.941, effectively mitigating global simulation bias. (2) The optimal simulation performance occurred at a fusion weight of 0.81, yielding a Kappa of 0.8917 and an Figure of Merit (FoM) of 0.3830, significantly exceeding a single model. (3) The 2030 simulation indicates that the GCTP model effectively reduces isolated pixels at urban fringes. The GCTP generates neighborhood patterns with high spatial compactness and geographic consistency. This study highlights the significant advantages of integrating long-range spatial perception with geographical heterogeneity constraints in the land expansion simulation of urban agglomerations. The findings support more precise territorial spatial planning practices. Full article

As key headwater regions of the upper Yangtze River, the Yalong and Dadu River basins are expected to experience highly uncertain hydrological responses under climate warming. However, the nonlinear and spatially heterogeneous evolution of streamflow across multiple time-frequency scales remains insufficiently understood. In this study, a SWAT model driven by CMIP6 climate projections under four shared socioeconomic pathways (SSP1-2.6 to SSP5-8.5) was coupled with multivariate wavelet coherence, spatial wavelet transform, and change-point detection methods to investigate the spatiotemporal evolution of streamflow and extreme risks during 2017–2100. Results indicate that precipitation is the primary driver of streamflow variability, with streamflow responding rapidly, while air temperature mainly regulates seasonal intensity via snowmelt. Streamflow seasonal intensity exhibits a northwest-southeast gradient, with low variability upstream and high sensitivity downstream, reflecting precipitation-concentrated, forested canyons where rapid lateral flow and dry-season evapotranspiration amplify flow contrasts. Moreover, hydrological nonstationarity and extreme risks are projected to intensify, with structural regime shifts emerging in the 2040s–2050s and extreme high-flow magnitudes doubling under SSP5-8.5, accompanied by more frequent drought-flood alternations. These findings highlight an upstream buffering-downstream sensitivity pattern, emphasizing the need for spatially differentiated water resources management under nonstationary climate conditions. Full article

Nighttime light (NTL), a crucial indicator of human activity intensity, has not been systematically analyzed for its interactive mechanisms with air pollution and climate change. This study first investigates the spatiotemporal evolution of China’s total nighttime light (TNTL) and average nighttime light (ANTL), alongside key indicators of meteorological parameters and air pollution, at the grid scale from 2000 to 2023. We then employ prefecture-level city data and a geographically and temporally weighted regression (GTWR) model to quantify the spatiotemporally heterogeneous associations of temperature (TMP), precipitation (PRE), fine particulate matter (PM_2.5), ozone (O₃), land use (LUL), topography, and socioeconomic factors with NTL. The results indicate that (1) China’s NTL exhibits a significant overall upward trend, with areas of increase or significant increase comprising 92.04% of the total study area. TNTL growth demonstrates regional heterogeneity, expanding by a factor of 4.91 in East China and 2.65 in Northeast China; (2) meteorological and air pollution indicators display spatiotemporal non-stationarity, with the synergistic effect between O₃ and PRE being the strongest; (3) among NTL drivers, LUL contributes most significantly (0.44), followed by TMP (0.14) > PM_2.5 (−0.33 × 10⁻¹) > O₃ (0.17 × 10⁻¹) > PRE (−0.33 × 10⁻⁶); (4) TMP and PRE may primarily influence NTL by altering ecological conditions and nighttime activity patterns. TMP shows a strong positive correlation with NTL in the junction zone of South, East, and Central China, whereas PRE predominantly exerts a negative influence; (5) air pollution exhibits distinct spatiotemporal effects: high PM_2.5 and O₃ generally correspond to lower NTL, though positive correlations persist in some areas due to industrial structures, highlighting the need for integrated policies that balance air quality management with sustainable urban planning; (6) the 2013 “Air Pollution Prevention and Control Action Plan” significantly strengthened the negative correlation between PM_2.5 and NTL in North China. However, O₃ concentrations increased by 28.9% after 2017, underscoring the challenge of coordinating VOC and NOx controls for long-term atmospheric sustainability. Full article

Sea surface temperature (SST) serves as a fundamental driver of ocean–atmosphere interactions and global climate variability, exhibiting strong nonstationarity, multiscale dynamics, and cross-variable coupling. However, current deep learning models often fail to capture these complex characteristics, limiting their ability to support accurate and physically consistent long-term SST prediction. To address these issues, we propose PAMSTnet, a unified deep learning framework for physics-informed adaptive multiscale spatiotemporal prediction. PAMSTnet leverages three-dimensional empirical wavelet transform (3DEWT) to learn interpretable multiscale spatiotemporal dynamics from raw observations, and applies multivariate spatiotemporal empirical orthogonal function (MSTEOF) to identify dominant cross-variable coupled modes. These physically meaningful representations are integrated into a deep ConvLSTM predictive network (DCPN) to support coordinated multiscale dynamical learning. Furthermore, PAMSTnet introduces physics-informed consistency learning (PICL) to enforce thermodynamic and dynamic constraints, enhancing physical consistency and interpretability. Extensive experiments demonstrate that PAMSTnet achieves superior performance against state-of-the-art baselines in long-term global SST prediction, reducing RMSE by 8.1% and improving ACC by 2.8% compared with the best-performing baseline, particularly under extreme climate events. Interpretation insights further highlight PAMSTnet’s adaptive representation of variable contributions and regional physical drivers. These findings position PAMSTnet as a promising paradigm for developing intelligent ocean prediction systems with enhanced physical consistency and interpretability. Full article

As the largest urban agglomeration and a critical grain production base in northern China, the Beijing–Tianjin–Hebei (BTH) region faces a sharp conflict between rapid urbanization and cropland conservation. Urban expansion inevitably leads to the loss of high-quality agricultural land, posing dual threats to food security and the terrestrial carbon cycle. To accurately assess the ecological costs of this process, this study integrates the CASA model with a time-weighted cumulative model to quantify the spatiotemporal impacts of urban expansion on cropland NPP in the BTH region from 2001 to 2020. Furthermore, a Geographically Weighted Regression (GWR) model was employed to examine the spatially varying effects of key driving factors on cropland NPP loss. The results indicate that urban land in the BTH region expanded by 45.2% over the past two decades, with 91.04% originating from cropland. Despite an overall upward trend in regional cropland NPP driven by climate change and agricultural intensification, the time-weighted cumulative cropland NPP loss attributable to urban encroachment over 2001–2020 reached 29.24 Tg C, which is equivalent to 0.751× the annual total cropland NPP in 2020 (used as a reference benchmark). Crucially, this expansion exhibits distinct ecological selectivity toward high-quality cropland, meaning that urban development has disproportionately encroached upon highly productive land with productivity levels exceeding the regional average. This selective occupation has led to a structural decline in the region’s potential agricultural production capacity. Additionally, GWR results reveal significant spatial non-stationarity in the relationships between cropland NPP loss and its drivers, revealing differentiated response patterns between plains and mountainous areas in terms of socio-economic drivers and physical constraints. These findings expose the hidden threats of urban expansion to food security, providing a crucial scientific basis for formulating differentiated land management policies and coordinating regional urbanization with cropland protection. Full article

Ground deformation is a major geohazard in many urban areas, requiring reliable monitoring and forecasting for hazard mitigation. Although Multi-Temporal InSAR enables high-resolution deformation monitoring, most prediction approaches rely on single-point modeling and fail to exploit spatial dependencies within deformation fields. This study proposes a spatiotemporally synchronous prediction framework for large-scale InSAR deformation fields, integrating sequence preprocessing, spatiotemporal modeling, and deformation pattern analysis. First-order differencing reduces sequence non-stationarity, while a patch-based encoder-decoder structure preserves spatial topology during dimensionality reduction. The core prediction model, built on PredRNNv2, captures the long-term spatiotemporal evolution of InSAR deformation sequences. In addition, independent component analysis (ICA) combined with K-means clustering identifies dominant deformation patterns and their geological associations. The framework is evaluated using synthetic datasets simulating multiple deformation mechanisms and Sentinel-1 InSAR time-series data over the Beijing Plain from 2015 to 2025. Results show that the model accurately captures deformation evolution and identifies transitions associated with groundwater regulation. These findings demonstrate the potential of deep spatiotemporal learning for large-scale InSAR deformation prediction and geohazard mechanism interpretation. Full article