Spatio-temporal Methods Research Articles

Deep learning initialized compressed sensing (Deli-CS) in volumetric spatio-temporal subspace reconstruction.

Spatio-temporal MRI methods offer rapid whole-brain multi-parametric mapping, yet they are often hindered by prolonged reconstruction times or prohibitively burdensome hardware requirements. The aim of this project is to reduce reconstruction time using deep learning. This study focuses on accelerating the reconstruction of volumetric multi-axis spiral projection MRF, aiming for whole-brain T1 and T2 mapping, while ensuring a streamlined approach compatible with clinical requirements. To optimize reconstruction time, the traditional method is first revamped with a memory-efficient GPU implementation. Deep Learning Initialized Compressed Sensing (Deli-CS) is then introduced, which initiates iterative reconstruction with a DL-generated seed point, reducing the number of iterations needed for convergence. The full reconstruction process for volumetric multi-axis spiral projection MRF is completed in just 20min compared to over 2h for the previously published implementation. Comparative analysis demonstrates Deli-CS's efficiency in expediting iterative reconstruction while maintaining high-quality results. By offering a rapid warm start to the iterative reconstruction algorithm, this method substantially reduces processing time while preserving reconstruction quality. Its successful implementation paves the way for advanced spatio-temporal MRI techniques, addressing the challenge of extensive reconstruction times and ensuring efficient, high-quality imaging in a streamlined manner.

Enhanced spatial analysis assessing the association between PFAS-contaminated water and cancer incidence: rationale, study design, and methods

BackgroundCancer is a complex set of diseases, and many have decades-long lag times between possible exposure and diagnosis. Environmental exposures, such as per- and poly-fluoroalkyl substances (PFAS) and area-level risk factors (e.g., socioeconomic variables), vary for people over time and space. Evidence suggests PFAS exposure is associated with several cancers; however, studies to date have various limitations. Few studies have used rigorous spatiotemporal approaches, and, to our knowledge, none have assessed cumulative exposures given residential histories or incorporated chemical mixture modeling. Thus, spatiotemporal analysis using advanced statistical approaches, accounting for spatially structured and unstructured heterogeneity in risk, can be a highly informative strategy for addressing the potential health effects of PFAS exposure.MethodsUsing population-based incident cancer cases and cancer-free controls in a 12-county area of southeastern Pennsylvania, we will apply Bayesian spatiotemporal analysis methods using historically reconstructed PFAS-contaminated water exposure given residential histories, and other potential cancer determinants over time. Bayesian group index models enable assessment of various mixtures of highly correlated PFAS chemical exposures incorporating mobility/residential history, and contextual factors to determine the association of PFAS-related exposures and cancer incidence.DiscussionThe purpose of this paper is to describe the Enhanced PFAS Spatial Analysis study rationale, study design, and methods.

Comparative evaluation of spatiotemporal methods for effective dengue cluster detection with a case study of national surveillance data in Thailand

Dengue fever poses a significant public health burden in tropical regions, including Thailand, where periodic epidemics strain healthcare resources. Effective disease surveillance is essential for timely intervention and resource allocation. Various methods exist for spatiotemporal cluster detection, but their comparative performance remains unclear. This study compared spatiotemporal cluster detection methods using simulated and real dengue surveillance data from Thailand. A simulation study explored diverse disease scenarios, characterized by varying magnitudes and spatial-temporal patterns, while real data analysis utilized monthly national dengue surveillance data from 2018 to 2020. Evaluation metrics included accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. Bayesian models and FlexScan emerged as top performers, demonstrating superior accuracy and sensitivity. Traditional methods such as Getis Ord and Moran’s I showed poorer performance, while other scanning-based approaches like spatial SaTScan exhibited limitations in positive predictive value and tended to identify large clusters due to the inflexibility of its scanning window shape. Bayesian modeling with a space–time interaction term outperformed testing-based cluster detection methods, emphasizing the importance of incorporating spatiotemporal components. Our study highlights the superior performance of Bayesian models and FlexScan in spatiotemporal cluster detection for dengue surveillance. These findings offer valuable guidance for policymakers and public health authorities in refining disease surveillance strategies and resource allocation. Moreover, the insights gained from this research could be valuable for other diseases sharing similar characteristics and settings, broadening the applicability of our findings beyond dengue surveillance.

Framework for Monitoring the Spatiotemporal Distribution and Clustering of the Digital Society Index of Indonesia

Digital disparities remain a significant challenge in Indonesia, particularly across its diverse regions, with uneven access to digital infrastructure, skills, and economic opportunities. This study aims to map these digital disparities at the district level, analyze the spatial distribution and clustering of digital transformation using the Digital Society Index of Indonesia (IMDI), and investigate the key drivers of digital inequality across four core pillars: Infrastructure and Ecosystem, Digital Skills, Empowerment, and Jobs. To measure the IMDI, primary data were collected from the industrial sector and the general population over three years (2022–2024), combined with secondary data on internet usage and service standards. A multistage random sampling approach ensured representativeness, considering demographic variations and industrial segments. The analysis employed spatiotemporal methods to capture temporal trends and spatial clustering. The results revealed a significant IMDI increase from 37.80 in 2022 to 43.18 in 2023, followed by stability at 43.34 in 2024. The hotspots of digital transformation remain concentrated on Java Island, while low spots persist in eastern Indonesia. This study provides critical insights into Indonesia’s digital readiness, identifying priority areas for targeted interventions to bridge the digital divide and foster equitable digital development.

Quantifying Gait Asymmetry in Stroke Patients: A Statistical Parametric Mapping (SPM) Approach.

BACKGROUND The VICON Toolkit enables three-dimensional (3D) motion capture for gait analysis. Statistical parametric mapping (SPM) is a voxel-based neuroimaging approach used to identify region-specific effects. This study aimed to apply SPM to analyze the joint angles of the hip, knee, and ankle during gait in 20 post-stroke patients using the VICON motion capture system. MATERIAL AND METHODS A total of 20 post-stroke patients participated in the study. A 10-camera VICON motion capture system (250 Hz) was used to record 3D kinematic data. Joint angles were assessed in the sagittal, frontal, and transverse planes using SPM. The data were normalized to 100% of the gait cycle, and a paired t test was performed to assess asymmetry between affected (AS) and unaffected sides (UAS), with P<0.05 considered statistically significant. RESULTS Significant asymmetry was observed in the hip and knee joint angles in the sagittal plane. The hip joint showed differences during 0-39% of the stance phase and 40-100% of the swing phase. The knee joint exhibited differences during 98-100% of the stance phase and 0-79% of the swing phase (P<0.05). CONCLUSIONS The study highlights the importance of addressing asymmetry in the hip and knee joints in post-stroke rehabilitation programs. The findings indicate the need for improving hip and knee joint mobility to enhance functional gait outcomes. The use of SPM provides a more comprehensive analysis of gait asymmetry, offering insights that go beyond traditional spatiotemporal methods.

Enhancing fault diagnosis with a hybrid attention mechanism and spatio-temporal feature mining model using small sample data

Fault diagnosis in rotating machinery presents unique challenges due to the limited availability of effective samples, which can lead to a long-tail effect during model training. To overcome these challenges, this paper introduces a novel small-sample fault diagnosis model for rotating machinery named the parallel convolution and bidirectional gated recurrent unit of attention model (PConvBiGA), which focuses on spatial and temporal feature extraction. The model is comprised of two core components: the wide convolutional denoising block (WCDB) and the hybrid attention mechanism spatio-temporal feature mining model (HAM-STM). The WCDB employs a one-dimensional wide convolutional layer coupled with a parallel pooling strategy, which helps preserve essential information from the original vibration data while filtering out weak noise and irrelevant features for diagnosis. To counter the feature loss typically encountered in traditional spatio-temporal feature fusion methods, the HAM-STM is developed. This structure utilizes a hybrid attention mechanism to effectively allocate high-dimensional feature weights across different paths, enhancing the model’s ability to mine features from various channels’ spatio-temporal characteristics. This, in turn, boosts the feature extraction capabilities of the model. Experimental results demonstrate that PConvBiGA successfully extracts more comprehensive features from vibration signals, enabling precise diagnosis with small samples and showing strong generalization and robustness.

An overview of bit-depth enhancement: Algorithm datasets and evaluation

To enhance the visualization of images on high dynamic range (HDR) monitors, it is essential to employ bit-depth enhancement (BDE) methods for converting low bit-depth contents into high bit-depth contents. While many methods have been proposed in recent years, to the best of our knowledge, there is no benchmark to analyze the state-of-the-art methods thoroughly. In this paper, we provide a detailed review of current bit-depth enhancement algorithms, and categorized them into four types: classic pixel-independent methods, traditional spatial context-aware methods, deep learning based spatial BDE methods and fusion based spatio-temporal BDE methods. Meanwhile, we have conducted extensive and fair experimental comparisons to evaluate the effectiveness of each algorithm. Two typical evaluation metrics PSNR and SSIM are employed, and accordingly, we provide a thorough analysis and guidance for future work. This benchmark for bit-depth enhancement aims to benefit related researches in image restoration. The relevant codes and datasets are available at https://github.com/TJUMMG/BDE.

Research on Spatiotemporal Evolution Methods Networks Based on Deep Learning: Arctic Shipping Data as a Case

Research on Spatiotemporal Evolution Methods Networks Based on Deep Learning: Arctic Shipping Data as a Case

Spatio-temporal quasi-experimental methods for rare disease outcomes: the impact of reformulated gasoline on childhood haematologic cancer

Abstract Although some pollutants emitted in vehicle exhaust, such as benzene, are known to cause leukaemia in adults with high exposure levels, less is known about the relationship between traffic-related air pollution (TRAP) and childhood haematologic cancer. In the 1990s, the US EPA enacted the reformulated gasoline program in select areas of the U.S., which drastically reduced ambient TRAP in affected areas. This created an ideal quasi-experiment to study the effects of TRAP on childhood haematologic cancers. However, existing methods for quasi-experimental analyses can perform poorly when outcomes are rare and unstable, as with childhood cancer incidence. We develop Bayesian spatio-temporal matrix completion methods to conduct causal inference in quasi-experimental settings with rare outcomes. Selective information sharing across space and time enables stable estimation, and the Bayesian approach facilitates uncertainty quantification. We evaluate the methods through simulations and apply them to estimate the causal effects of TRAP on childhood leukaemia and lymphoma.

Traffic Flow Prediction Using Novel Spatial-Temporal Multi-Head Attention Graph Convolution Networks

For urban development, accurate traffic flow prediction is crucial. Traffic flow prediction data can be used for optimizing public transportation systems, reducing congestion, planning road networks, and improving overall city infrastructure management. Despite numerous spatiotemporal methods used to forecast traffic flow, information aggregation at each site neither produce optimal results, because of the use non-Euclidean distance data. For more accurate results of traffic flow prediction, we propose a spatial-temporal multi-head attention graph convolution network (STMAGCN). It extracts temporal features using temporal convolutional layers and enriches nodes with spatial attention and multi-head attention mechanisms. By combining them, we can extract space characteristics more effectively. We also introduce a novel graph structure construction method that employs graph-based transformations to convert non-connected data points into a comprehensive graph structure, facilitating the effective application of our STMAGCN. This approach improves the predictive efficiency and accuracy of our model by ensuring seamless integration of disparate data sources. Experimental results indicate that the preferential extraction of spatiotemporal data significantly enhances traffic flow prediction accuracy. Compared with the latest models, our performance on the California Department of Transportation Performance Measurement System (PeMS) PeMS04 and PeMS08 datasets shows a slight improvement, with an average increase of 16.8%. Additionally, the average improvement in time is approximately 445.95%.

Dynamic graph structure and spatio-temporal representations in wind power forecasting

Wind Power Forecasting (WPF) has gained considerable focus as a crucial aspect of the successful integration and operation of wind power. However, due to the stochastic and unstable nature of wind, it poses a real challenge to effectively analyze the correlations among multiple time series data for accurate prediction. In our study, an end-to-end framework called Dynamic Graph structure and Spatio-Temporal representation learning (DSTG) framework is proposed to achieve stable power forecasting by constructing graph data to capture the critical features in the data. Specifically, a graph structure learning (GSL) module is introduced to dynamically construct task-related correlation matrices via backpropagation to mitigate the inherent inconsistency and randomness of wind power data. Additionally, a dual-scale temporal graph learning (DTG) module is further proposed to explore the implicit spatio-temporal features at a fine-grained level using different skip connections from the constructed graph data. Finally, comprehensive experiments are performed on the collected Xuji Group Wind Power (XGWP) dataset, and the results show that DSTG outperforms the state-of-the-art spatio-temporal methods by 10.12% on the average of root mean square error and mean absolute error, demonstrating the effectiveness of DSTG. In conclusion, our model provides a promising approach.

Algorithm for image transmission in the MIMO-OFDM system in the presence of active interference and structural distortions

Problem Statement. With the advent and ever-growing development of wireless communications, the demand for high-speed, reliable and energy-efficient wireless data transmission methods has become an urgent need in our increasingly interconnected world. One of the most promising technologies to meet this need is a combination of multiple-input multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) systems. However, as with any technology, these systems face challenges. One of the significant obstacles to their effective implementation is the presence of active interference and structural distortions in image transmission. This paper is devoted to solving this problem, focusing on improving the adaptive image transmission algorithm in the MIMO-OFDM system under such adverse conditions. Objective. Improving the efficiency of a wireless data transmission system in an environment with multiple reflections and active interference by reducing the probability of a bit error based on the development and research of spatio-temporal methods of signal processing and the development of an adaptive algorithm. Results. The main features of image transmission via MIMO-OFDM communication systems are considered. The main characteristics of such communication systems are given, the structural diagram of the transceiver part, models of interference that occur in the communication system, the structure of pilot signals for channel estimation and the algorithm of operation of such a communication system are considered. The adaptive algorithm for processing spatio-temporal signals in the MIMO-OFDM communication system is considered and improved. The MIMO-OFDM system is simulated, and an in-house model of a simulator of such a communication system is developed, experiments are conducted to study the influence of spatial correlations between antennas on the operation of such a system, the gain from using the adaptive algorithm for processing spatio-temporal signals is shown and substantiated, which demonstrates its applied value. Practical significance. The proposed modification of the adaptive algorithm showed a stable improvement in terms of the quality of transmitted images under conditions of active interference and structural distortions. This makes the practical application of the adaptive algorithm useful, for example, in the situation of transmitting images from a UAV under conditions of a complex interference-signal environment and is capable of significantly improving the quality of transmitted images.

Population Distribution Forecasting Based on the Fusion of Spatiotemporal Basic and External Features: A Case Study of Lujiazui Financial District

Predicting the distribution of people in the time window approaching a disaster is crucial for post-disaster assistance activities and can be useful for evacuation route selection and shelter planning. However, two major limitations have not yet been addressed: (1) Most spatiotemporal prediction models incorporate spatiotemporal features either directly or indirectly, which results in high information redundancy in the parameters of the prediction model and low computational efficiency. (2) These models usually incorporate certain basic and external features, and they can neither change spatiotemporal addressed features according to spatiotemporal features nor change them in real-time according to spatiotemporal features. The spatiotemporal feature embedding methods for these models are inflexible and difficult to interpret. To overcome these problems, a lightweight population density distribution prediction framework that considers both basic and external spatiotemporal features is proposed. In the study, an autoencoder is used to extract spatiotemporal coded information to form a spatiotemporal attention mechanism, and basic and external spatiotemporal feature attention is fused by a fusion framework with learnable weights. The fused spatiotemporal attention is fused with Resnet as the prediction backbone network to predict the people distribution. Comparison and ablation experimental results show that the computational efficiency and interpretability of the prediction framework are improved by maximizing the scalability of the spatiotemporal features of the model by unleashing the scalability of the spatiotemporal features of the model while enhancing the interpretability of the spatiotemporal information as compared to the classical and popular spatiotemporal prediction frameworks. This study has a multiplier effect and provides a reference solution for predicting population distributions in similar regions around the globe.

Rapid predator-prey balance shift follows critical-population-density transmission between cod (Gadus morhua) and capelin (Mallotus villosus)

Sensing limitations have impeded knowledge about how individual predator-prey interactions build to organized multi-species group behaviour across an ecosystem. Population densities of overlapping interacting oceanic fish predator and prey species, however, can be instantaneously distinguished and quantified from roughly the elemental individual to spatial scales spanning thousands of square kilometres by wide-area multispectral underwater-acoustic sensing, as shown here. This enables fundamental mechanisms behind large-scale ordered predator-prey interactions to be investigated. Critical population densities that transition random individual behaviour to ordered group behaviour are found to rapidly propagate to form vast adversarial prey and predator shoals of capelin and surrounding cod in the Barents Sea Arctic ecosystem for these keystone species. This leads to a sudden major shift in predator-prey balance. Only a small change in local behaviour triggers the shift due to an unstable equilibrium. Such unstable equilibria and associated balance shifts at predation hotspots are often overlooked as blind spots in present ocean ecosystem monitoring and assessment due to use of highly undersampled spatio-temporal sampling methods.

A Spatiotemporal Empowerment Framework for China’s National 3D Mapping Program 3dRGLM

Abstract. China has launched a national 3D mapping program to build 3D Realistic geospatial Landscape Model (3dRGLm) over the whole country, covering both cities and countryside. The 3dRGLM is a digital description and representation of the real 3D geospatial spaces, and is defined as a new generation geospatial data product with 3D structures, realistic scenes, and geospatial entities. Such a national or local 3dRGLm will set up a digital 3D realistic geospatial space which can facilitate the vital connection with the real geospatial space, and will serve as a new generation of geospatial information infrastructure for our society. With the acquisition of complex 3D spatial data continues to advance, managing and utilizing the massive volumes of 3D data to provide spatiotemporal empowerment applications has become a significant technological challenge. This paper analyzes the technical characteristics of 3d realistic geospatial landscape model, and then proposes the main content and basic structure of a 3d realistic geospatial landscape model database, as well as the spatiotemporal service empowerment methods based on the database. Finally, a case analysis is provided to serve as a reference for the construction and application of 3d realistic geospatial landscape.

A deep learning approach for deriving wheat phenology from near-surface RGB image series using spatiotemporal fusion

Accurate monitoring of wheat phenological stages is essential for effective crop management and informed agricultural decision-making. Traditional methods often rely on labour-intensive field surveys, which are prone to subjective bias and limited temporal resolution. To address these challenges, this study explores the potential of near-surface cameras combined with an advanced deep-learning approach to derive wheat phenological stages from high-quality, real-time RGB image series. Three deep learning models based on three different spatiotemporal feature fusion methods, namely sequential fusion, synchronous fusion, and parallel fusion, were constructed and evaluated for deriving wheat phenological stages with these near-surface RGB image series. Moreover, the impact of different image resolutions, capture perspectives, and model training strategies on the performance of deep learning models was also investigated. The results indicate that the model using the sequential fusion method is optimal, with an overall accuracy (OA) of 0.935, a mean absolute error (MAE) of 0.069, F1-score (F1) of 0.936, and kappa coefficients (Kappa) of 0.924 in wheat phenological stages. Besides, the enhanced image resolution of 512 × 512 pixels and a suitable image capture perspective, specifically a sensor viewing angle of 40° to 60° vertically, introduce more effective features for phenological stage detection, thereby enhancing the model’s accuracy. Furthermore, concerning the model training, applying a two-step fine-tuning strategy will also enhance the model’s robustness to random variations in perspective. This research introduces an innovative approach for real-time phenological stage detection and provides a solid foundation for precision agriculture. By accurately deriving critical phenological stages, the methodology developed in this study supports the optimization of crop management practices, which may result in improved resource efficiency and sustainability across diverse agricultural settings. The implications of this work extend beyond wheat, offering a scalable solution that can be adapted to monitor other crops, thereby contributing to more efficient and sustainable agricultural systems.

Spatio-temporal occurrence, burden, risk factors and modelling methods for estimating scrub typhus burden from global to subnational resolutions: a systematic review protocol.

Scrub typhus is a neglected life-threatening vector-borne disease mainly caused by the bacterium Orientia tsutsugamushi, which is occasionally transmitted to humans during feeding of larval mites. It has been estimated that more than 1 billion persons are potentially threatened and 1 million clinical cases occur annually across the world; however, it is unclear how this estimate was computed (and what the original source was) and much remains unknown regarding its global burden and risk factors. This systematic review aims to provide a comprehensive overview of the spatial-temporal distribution of scrub typhus, associated burden and risk factors at global, national and subnational resolutions, and to review the burden estimation models used at those different scales. A systematic search for literature on scrub typhus occurrence, risk factors and modelling methods will be conducted. PubMed and five other databases will be searched for published literature, and Google Scholar and nine other databases will be used to search for grey literatures. All titles/abstracts of the searched records will be separately assessed by two reviewers, who will then screen the full-text of potential records to decide eligibility. A pre-formatted spreadsheet will be used by one reviewer to extract data from qualifying research, with a second reviewer checking the results. Data will be tabulated, synthesized descriptively, and summarized narratively for each review question. Where appropriate, meta-analyses will be conducted. The risk of bias will be assessed, and potential publication bias will be detected. This review will provide a comprehensive understanding of the current occurrence, spatial-temporal distribution, and burden of scrub typhus, identify associated risk factors from global to subnational resolutions, consolidate the best practice modeling framework(s) to estimate the burden of scrub typhus at various geographic/temporal resolutions, and decompose the relative contributions of various risk factors at scale. CRD42022315209.

Spatiotemporal patterns and socioeconomic determinants of pulmonary tuberculosis in Dongguan city, China, during 2011–2020: an ecological study

ObjectivePulmonary tuberculosis (PTB) is a critical challenge worldwide, particularly in China. This study aimed to explore the spatiotemporal transmission patterns and socioeconomic factors of PTB in Dongguan city, China.Methods/designAn ecological...

The spatiotemporal associations between esophageal and gastric cancers provide evidence for its joint endoscopic screening in China: a population-based study

BackgroundThe spatiotemporal epidemiological evidence supporting joint endoscopic screening for esophageal cancer (EC) and gastric cancer (GC) remains limited. This study aims to identify combined high-risk regions for EC and GC and determine optimal areas for joint and separate endoscopic screening.MethodsWe analyzed the association of incidence trends between EC and GC in cancer registry areas across China from 2006 to 2016 using spatiotemporal statistical methods. Based on these analyses, we divided different combined risk regions for EC and GC to implement joint endoscopic screening.ResultsFrom 2006 to 2016, national incidence trends for both EC and GC showed a decline, with an average annual percentage change of -3.15 (95% confidence interval [CI]: -5.33 to -0.92) for EC and -3.78 (95% CI: -4.98 to -2.56) for GC. A grey comprehensive correlation analysis revealed a strong temporal association between the incidence trends of EC and GC, with correlations of 79.00% (95% CI: 77.85 to 80.14) in males and 77.62% (95% CI: 76.50 to 78.73) in females. Geographic patterns of EC and GC varied, demonstrating both homogeneity and heterogeneity across different regions. The cancer registry areas were classified into seven distinct combined risk regions, with 33 areas identified as high-risk for both EC and GC, highlighting these regions as priorities for joint endoscopic screening.ConclusionThis study demonstrates a significant spatiotemporal association between EC and GC. The identified combined risk regions provide a valuable basis for optimizing joint endoscopic screening strategies for these cancers.

Predicting Ground Cover with Deep Learning Models—An Application of Spatio-Temporal Prediction Methods to Satellite-Derived Ground Cover Maps in the Great Barrier Reef Catchments

Livestock grazing is a major land use in the Great Barrier Reef Catchment Area (GBRCA). Heightened grazing density coupled with inadequate land management leads to accelerated soil erosion and increased sediment loads being transported downstream. Ultimately, these increased sediment loads impact the water quality of the Great Barrier Reef (GBR) lagoon. Ground cover mapping has been adopted to monitor and assess the land condition in the GBRCA. However, accurate prediction of ground cover remains a vital knowledge gap to inform proactive approaches for improving land conditions. Herein, we explored two deep learning-based spatio-temporal prediction models, including convolutional LSTM (ConvLSTM) and Predictive Recurrent Neural Network (PredRNN), to predict future ground cover. The two models were evaluated on different spatial scales, ranging from a small site (i.e., &lt;5 km2) to the entire GBRCA, with different quantities of training data. Following comparisons against 25% withheld testing data, we found the following: (1) both ConvLSTM and PredRNN accurately predicted the next-season ground cover for not only a single site but also the entire GBRCA. They achieved this with a Mean Absolute Error (MAE) under 5% and a Structural Similarity Index Measure (SSIM) exceeding 0.65; (2) PredRNN superseded ConvLSTM by providing more accurate next-season predictions with better training efficiency; (3) The accuracy of PredRNN varies seasonally and spatially, with lower accuracy observed for low ground cover, which is underestimated. The models assessed in this study can serve as an early-alert tool to produce high-accuracy and high-resolution ground cover prediction one season earlier than observation for the entire GBRCA, which enables local authorities and grazing property owners to take preventive measures to improve land conditions. This study also offers a new perspective on the future utilization of predictive spatio-temporal models, particularly over large spatial scales and across varying environmental sites.