Spatio-temporal Process Research Articles

A spatiotemporal motion prediction network based on multi-level feature disentanglement

The prediction task is significantly challenged by the intricate scene information and motion variations present in spatiotemporal data. Existing prediction methods struggle to accurately forecast long-term outcomes, particularly for transient motions characterized by notable trends, such as hand lifts, jumps, or vehicle turns. To address these challenges, we introduce a Spatiotemporal Motion Prediction Network based on Multi-level Feature Disentanglement (FDPNet). The model delineates spatiotemporal prediction into two distinct stages: feature disentanglement and motion prediction. We first devise a Multi-level Feature Disentanglement (MFD) model to disentangle the multilayer features of motion within the temporal sequence, encompassing period, trend, and residual components. This disentanglement is based on disentangling spatiotemporal coupling, enabling the network to comprehensively grasp the genuine laws governing motion in the spatiotemporal evolution process. Second, to enhance the prediction accuracy of the network over extended periods, we introduce the Motion Differential Self-Attention LSTM unit (MDSA-LSTM). This unit employs differential operations to extract inter-frame motion trends, elevating the network's proficiency in capturing spatiotemporal correlations over long distances through an enhanced self-attention mechanism. FDPNet attains state-of-the-art performance on the Moving MNIST, UCF101, KITTI, and Caltech pedestrian datasets. These outcomes substantiate the substantial potential of this research within the realm of spatiotemporal prediction.

Design of spatiotemporal variability index for climatic variables

Spatiotemporal signals are currently gaining wide popularity owing to their richness in terms of information content from both spatial and temporal aspects. Climatic studies involve the usage of such signals from different spatial locations to understand the climatic conditions, anomalies and for prediction of similar climatic zones. Measurement of the complex inter-relationships in the spatiotemporal processes is of vital importance in this regard. The predominant frequency components associated with annual, semi-annual, and quarterly seasonal variations exhibit interchangeability across spatial and temporal dimensions, thereby contributing significantly to the variability within the spatiotemporal region under investigation. Employing frequency domain analysis enables the examination of these crucial seasonal components, even amidst noise, provided effective frequency domain-based techniques are employed. This paper is based on the measurement of the variability of spatiotemporal climatic variables considered from different geographic zones using multivariate variational mode decomposition. The main advantage of the proposed index lies in the ability to measure the variability of the signals acquired from either spatially close or far-off locations on Earth. Also, the index harnesses the hidden frequency components in the spatiotemporal signals under test by comparing the aligned periodic components across different zones. The correctness of the proposed index is tested through exhaustive simulation studies and also against the weather measurements like the Earth skin temperature, humidity and precipitation acquired from spatially close and far-off locations over a period of 40 years.

Minimum contrast for the first-order intensity estimation of spatial and spatio-temporal point processes

In this paper, we harness a result in point process theory, specifically the expectation of the weighted K-function, where the weighting is done by the true first-order intensity function. This theoretical result can be employed as an estimation method to derive parameter estimates for a particular model assumed for the data. The underlying motivation is to avoid the difficulties associated with dealing with complex likelihoods in point process models and their maximization. The exploited result makes our method theoretically applicable to any model specification. In this paper, we restrict our study to Poisson models, whose likelihood represents the base for many more complex point process models. In this context, our proposed method can estimate the vector of local parameters that correspond to the points within the analyzed point pattern without introducing any additional complexity compared to the global estimation. We illustrate the method through simulation studies for both purely spatial and spatio-temporal point processes and show complex scenarios based on the Poisson model through the analysis of two real datasets concerning environmental problems.

Hydrological profile observation scheme based on optical fiber sensing for polar sea ice buoy monitoring

The monitoring of hydrological elements in the polar region is the basis for the study of the dynamic environment under the ice. The traditional cross-season subglacial hydrological environment monitoring mainly relies on tether-type vertical profile measurement ice-based buoys, which have the advantages such as high reliability, high measurement accuracy, and real-time communication, while also has disadvantages of high-cost, large volume and weight, high power consumption, and complex layout. Therefore, it is urgent to develop a new type of ice-based profile buoy with low-cost, miniaturization, low power consumption, convenient deployment, and high reliability. In this paper, a novel optical fiber sensing scheme for ice-based buoy monitoring is proposed, which uses arrayed fiber grating to measure seawater temperature and depth profile and uses a dual-conduction mode resonance mechanism to measure seawater salinity. The temperature, depth, and salinity of seawater can be detected by an all-optical fiber technology in real-time. Preliminary experiments show that the temperature accuracy is ±0.1 °C in the range of -5∼35 °C, the salinity accuracy is ±0.03‰ in the range of 30‰∼40‰, and the vertical spatial resolution of depth can be adjusted in the range of 0∼1000 m, which can better meet the requirements of polar hydrological multi-layer profile observation. It can provide an innovative technology and equipment support for studying the spatiotemporal change process of the polar subglacial ocean.

Earthfarsser: Versatile Spatio-Temporal Dynamical Systems Modeling in One Model

Efficiently modeling spatio-temporal (ST) physical processes and observations presents a challenging problem for the deep learning community. Many recent studies have concentrated on meticulously reconciling various advantages, leading to designed models that are neither simple nor practical. To address this issue, this paper presents a systematic study on existing shortcomings faced by off-the-shelf models, including lack of local fidelity, poor prediction performance over long time-steps, low scalability, and inefficiency. To systematically address the aforementioned problems, we propose an EarthFarseer, a concise framework that combines parallel local convolutions and global Fourier-based transformer architectures, enabling dynamically capture the local-global spatial interactions and dependencies. EarthFarseer also incorporates a multi-scale fully convolutional and Fourier architectures to efficiently and effectively capture the temporal evolution. Our proposal demonstrates strong adaptability across various tasks and datasets, with fast convergence and better local fidelity in long time-steps predictions. Extensive experiments and visualizations over eight human society physical and natural physical datasets demonstrates the state-of-the-art performance of EarthFarseer. We release our code at https://github.com/easylearningscores/EarthFarseer.

Revisiting spatiotemporal evolution process and mechanism of a giant reservoir landslide during weather extremes

An updated comprehension of landslide kinematics is a prerequisite for developing effective risk assessment and emergency management in the context of climate variability and anthropogenic disturbances. This study employs a multidisciplinary approach, taking the Outang landslide in China's Three Gorges Reservoir area as a case study. We examined the spatiotemporal surface processes and time- and subzone-dependent hydrometeorological factors from 2007 to 2021, particularly identifying rainfall regimes and reservoir water levels that trigger landslide movements. Given the uneven distribution of GNSS surface markers, we introduced weighted monthly mean displacement (D) and weighted marker frequency (f), enhancing the accuracy of our insights into distinct active subzones of the landslide. Preliminary results revealed July, September and June as months of higher susceptibility, as well as months with low water levels. The early-stage deformation control by reservoir water levels is being increasingly overshadowed by rainfall, due to extreme weather events and localized anti-slip measures. Long-term surface processes have also highlighted the transformation of landslide failure modes towards a compound pattern, from a predominantly retrogressive to a thrust-dominated pattern, and even overall translation coupled with localized rotation. Furthermore, the comparative analysis of normalized displacement/strain demonstrates the power and potential of integrating multi-source monitoring data. By revisiting the evolution processes and potential deformation mechanisms of the Outang landslide, we are getting closer to understanding more about the dynamics of giant landslides in reservoir areas, and illuminate the pathway towards enhanced risk management and emergency response strategies.

Study on precursor features of coal and rock loading failure based on difference network

In order to reveal the spatio-temporal precursor features of coal and rock loading failure, and accurately identify the critical state of coal and rock failure, the development of coal and rock loading failure is regarded as the spatio-temporal evolution dynamical process of the complex network system, the spatio-temporal evolution dynamical characteristics of coal and rock dynamic disasters are analyzed. The precursor features indicators I(t) and Ia(t) of the coal and rock failure based on the difference network are proposed, Brazilian splitting experiments of different types of coal and rock are carried out. Combined with the damage theory, the load model of coal and rock damage characterization based Ia(t) is established. The results show that when the coal-rock system is in the warning critical period, the damage is accelerated, with high potential energy, low rebound and weak robustness, which is reflected in the acoustic emission (AE) and electromagnetic radiation (EMR) data at different spatial monitoring points. When the coal-rock system is in the stable stage, the structure of the correlation network at adjacent moment is similar, and the number of edges in the difference network is small. While when coal or rock is about to fracture, the correlation network at adjacent moment has great structure differences, and there are many edges in the difference network. I(t) and Ia(t) synthesize the signal characteristics of different spatial monitoring points, and reflect the spatio-temporal evolution process and damage state of coal and rock. The damage characterization load based on Ia(t) is in good agreement with the measured load, and the correlation coefficients between the measured load and the calculated load based on Ia(t) of most rock samples are more than 0.80. The research results can be used as the precursor features of coal and rock failure, and have certain significance for the early warning of coal and rock dynamic disasters.

A 3D virtual geographic environment for flood representation towards risk communication

Risk communication seeks to develop a shared understanding of disaster among stakeholders, thereby amplifying public awareness and empowering them to respond more effectively to emergencies. However, existing studies have overemphasized specialized numerical modelling, making the professional output challenging to understand and use by non-research stakeholders. In this context, this article proposes a 3D virtual geographic environment for flood representation towards risk communication, which integrates flood modelling, parallel computation, and 3D representation in a pipeline. Finally, a section of the Rhine River in Bonn, Germany, is selected for experiment analysis. The experimental results show that the proposed approach is capable of flood modelling and 3D representation within a few hours, the parallel speedup ratio reached 6.45. The intuitive flood scene with 3D city models is beneficial for promoting flood risk communication and is particularly helpful for participants without direct experience of floods to understand its spatiotemporal process. It also can be embedded in the Geospatial Infrastructure Management Ecosystem (GeoIME) cloud application for intelligent flood systems.

Bayesian tip-dated phylogeny and biogeography of Cissampelideae (Menispermaceae): Mitigating the effects of homoplastic morphological characters.

The integration of morphological and molecular data is essential to understand the affinities of fossil taxa and spatio-temporal evolutionary processes of organisms. However, homoplastic morphological characters can mislead the placement of fossil taxa and impact downstream analyses. Here, we provide an example of how to mitigate effectively the effect of morphological homoplasy on the placement of fossil taxa and biogeographic inferences of Cissampelideae. We assembled three data types, morphological data only, morphological data with a molecular scaffold and combined morphological and molecular data. By removing high-level homoplastic morphological data or reweighting the morphological characters, we conducted 15 parsimony, 12 undated Bayesian and four dated Bayesian analyses. Our results show that the 14 selected Cissampelideae fossil taxa are placed poorly when based only on morphological data, but the addition of molecular scaffold and combination of morphological and molecular data greatly improve the resolution of fossil nodes. We raise the monotypic Stephania subg. Botryodiscia to generic status and discover that three fossils previously assigned to Stephania should be members of Diploclisia. The Bayesian tip-dated tree recovered by removing homoplastic morphological characters with a Rescaled Consistency Index <0.25 has the highest stratigraphic fit and consequently generates more reasonable biogeographic reconstruction for Cissampelideae. Cissampelideae began to diversify in Asia in the latest Cretaceous and subsequently dispersed to South America around the Cretaceous-Palaeogene boundary. Two dispersal events from Asia to Africa occurred in the Early Eocene and the Late Eocene-Late Oligocene, respectively. These findings provide guidelines and practical methods for mitigating the effects of homoplastic morphological characters on fossil placements and Bayesian tip-dating, as well as insights into the past tropical floristic exchanges among different continents.

Knowledge Graph Representation of Multi-Source Urban Storm Surge Hazard Information Based on Spatio-Temporal Coding and the Hazard Events Ontology Model

Coastal cities are increasingly vulnerable to urban storm surge hazards and the secondary hazards they cause (e.g., coastal flooding). Accurate representation of the spatio-temporal process of hazard event development is essential for effective emergency response. However, current knowledge graph representations face the challenge of integrating multi-source information with various spatial and temporal scales. To address this challenge, we propose a new information model for storm surge hazard events, involving a two-step process. First, a hazard event ontology is designed to model the components and hierarchical relationships of hazard event information. Second, we utilize multi-scale time segment integer coding and geographical coordinate subdividing grid coding to create a spatio-temporal framework, for modeling spatio-temporal features and spatio-temporal relationships. Using the 2018 typhoon Mangkhut storm surge event in Shenzhen as a case study and the hazard event information model as a schema layer, a storm surge event knowledge graph is constructed, demonstrating the integration and formal representation of heterogeneous hazard event information and enabling the fast retrieval of disasters in a given spatial or temporal range.

The Spatiotemporal Clustering of Short‐Duration Rainstorms in Shanghai City Using a Sub‐Hourly Gauge Network

AbstractImpacted by both the large‐scale climate change and the local hydrometeorological atmosphere, the spatiotemporal structure and properties of short‐duration extreme rainfall events in urban areas remain a complex issue. This study explores the spatiotemporal pattern variability of short‐duration rainstorm events using an exceptional sub‐hourly gauge network in Shanghai, China. A catalog of 207 “independent” rainstorms is extracted. The hierarchical clustering method is used to examine the spatial and temporal features including rainfall duration, intensity, and coverage for each district with different degrees of urbanization in the 2014–2018 period. The results point to a pronounced spatial heterogeneity with heavier rainfall magnitude in the north. In the urban core area, intensive short‐duration rainstorms hit more frequently than in other districts. Rainstorms can be divided into three types from the spatiotemporal process: Type delayed‐slow, Type advanced‐fast, and Type advanced‐slow. The delayed‐peak with a higher slope occurs more in the west and the advanced‐peak more along the coastal line. The other type of advanced‐peak storm with a reverse trend of intensity and coverage concentrates in June–August of the year and the 16:00–20:00 of the day. While 65.6% of the delayed‐peak storms only occur in one district, more than half of the advanced‐peak appear without other types at the same time. These findings show the importance of observation‐based analysis in understanding the spatiotemporal pattern of short‐duration storms and provide a reference for storm design of sub‐regions in urban areas.

ADP-Based Event-Triggered Constrained Optimal Control on Spatiotemporal Process: Application to Temperature Field in Roller Kiln.

The precise control of the spatiotemporal process in a roller kiln is crucial in the production of Ni-Co-Mn layered cathode material of lithium-ion batteries. Since the product is extremely sensitive to temperature distribution, temperature field control is of great significance. In this article, an event-triggered optimal control (ETOC) method with input constraints for the temperature field is proposed, which takes up an important position in reducing the communication and computation costs. A nonquadratic cost function is adopted to describe the system performance with input constraints. First, we present the problem description of the temperature field event-triggered control, where this field is described by a partial differential equation (PDE). Then, the event-triggered condition is designed according to the information of system states and control inputs. On this basis, a framework of the event-triggered adaptive dynamic programming (ETADP) method that is based on the model reduction technology is proposed for the PDE system. A critic network is used to approach the optimal performance index by a neural network (NN) together with that an actor network is used to optimize the control strategy. Furthermore, an upper bound of the performance index and a lower bound of interexecution times, as well as the stabilities of the impulsive dynamic system and the closed-loop PDE system, are also proved. Simulation verification demonstrates the effectiveness of the proposed method.

Crystallization processes of quartz in a granitic magma: Implications for the magma chamber processes of Okueyama granite, Kyushu, Japan

Cathodoluminescence (CL) characterization combined with titanium concentration of quartz crystals in granite is a prevalent tool for elucidating the crystal growth of quartz in granitic magmas. The CL pattern yields to the internal structure, and the titanium concentration allows crystallization temperature determination based on TitaniQ geothermometry. Using two approaches, 1) observations of quartz crystals using thin sections and 2) observations based on multiple sections of separated quartz, we identified the overall characteristics of quartz crystals within a magma chamber and the 3D internal structure and growth characteristics of quartz grains. Thin section observations allow for the characterization of quartz containing zircon inclusions, indicating that zircon crystallization was followed by quartz crystallization; therefore, the onset of quartz crystallization occurred later than the onset of zircon crystallization in the cooling magma chamber. Such petrological restrictions permit allow for suitable geothermometer use and appropriate TiO2 activity estimation, and thus, accurate quartz crystallization temperature determination. Observations using multiple sections of separated quartz crystals provide 3D internal structures and true onset parts within the grains. Progressive variation in local activities of TiO2 in the concentration boundary layer at the growing quartz crystal interface in the magma chamber can be evaluated using a combination of the crystallization temperature and variations in titanium concentration within the grain, yielding a sequential crystallization model of oscillatory zoning quartz with local activities of TiO2 in the cooling magma chamber. The difference in CL characteristics and crystallization temperature among the lithofacies in granite contributes to clarifying the spatiotemporal magma chamber processes.

Learning the intrinsic dynamics of spatio-temporal processes through Latent Dynamics Networks

Predicting the evolution of systems with spatio-temporal dynamics in response to external stimuli is essential for scientific progress. Traditional equations-based approaches leverage first principles through the numerical approximation of differential equations, thus demanding extensive computational resources. In contrast, data-driven approaches leverage deep learning algorithms to describe system evolution in low-dimensional spaces. We introduce an architecture, termed Latent Dynamics Network, capable of uncovering low-dimensional intrinsic dynamics in potentially non-Markovian systems. Latent Dynamics Networks automatically discover a low-dimensional manifold while learning the system dynamics, eliminating the need for training an auto-encoder and avoiding operations in the high-dimensional space. They predict the evolution, even in time-extrapolation scenarios, of space-dependent fields without relying on predetermined grids, thus enabling weight-sharing across query-points. Lightweight and easy-to-train, Latent Dynamics Networks demonstrate superior accuracy (normalized error 5 times smaller) in highly-nonlinear problems with significantly fewer trainable parameters (more than 10 times fewer) compared to state-of-the-art methods.

Severe convective storms’ reproduction: empirical analysis from the marked self-exciting point processes point of view

The paper focuses on the evaluation of hailstorms’ and thunderstorms winds’ events in the United States of America, in the period from 1996 to 2022, under the marked spatio-temporal self-exciting point processes point of view. The aim of the present article is the assessment and description of the spatio-temporal spontaneous and reproducing activity of severe hailstorms’ and thunderstorms winds’ processes. The present application shows how the spatio-temporal pattern is well-fitted and clearly explainable, according to the flexible semi-parametric ETAS model fitting.

Developmental loss of NMDA receptors results in supernumerary forebrain neurons through delayed maturation of transit-amplifying neuroblasts

Developmental neurogenesis is a tightly regulated spatiotemporal process with its dysregulation implicated in neurodevelopmental disorders. NMDA receptors are glutamate-gated ion channels that are widely expressed in the early nervous system, yet their contribution to neurogenesis is poorly understood. Notably, a variety of mutations in genes encoding NMDA receptor subunits are associated with neurodevelopmental disorders. To rigorously define the role of NMDA receptors in developmental neurogenesis, we used a mutant zebrafish line (grin1−/−) that lacks all NMDA receptors yet survives to 10 days post-fertilization, offering the opportunity to study post-embryonic neurodevelopment in the absence of NMDA receptors. Focusing on the forebrain, we find that these fish have a progressive supernumerary neuron phenotype confined to the telencephalon at the end of embryonic neurogenesis, but which extends to all forebrain regions during postembryonic neurogenesis. This enhanced neuron population does not arise directly from increased numbers or mitotic activity of radial glia cells, the principal neural stem cells. Rather, it stems from a lack of timely maturation of transit-amplifying neuroblasts into post-mitotic neurons, as indicated by a decrease in expression of the ontogenetically-expressed chloride transporter, KCC2. Pharmacological blockade with MK-801 recapitulates the grin1−/− supernumerary neuron phenotype, indicating a requirement for ionotropic signaling. Thus, NMDA receptors are required for suppression of indirect, transit amplifying cell-driven neurogenesis by promoting maturational termination of mitosis. Loss of suppression results in neuronal overpopulation that can fundamentally change brain circuitry and may be a key factor in pathogenesis of neurodevelopmental disorders caused by NMDA receptor dysfunction.

Technical note: Accurate, reliable, and high-resolution air quality predictions by improving the Copernicus Atmosphere Monitoring Service using a novel statistical post-processing method

Abstract. Starting from the regional air quality forecasts produced by the Copernicus Atmosphere Monitoring Service (CAMS), we propose a novel post-processing approach to improve and downscale results on a finer scale. Our approach is based on the combination of ensemble model output statistics (EMOS) with a spatio-temporal interpolation process performed through the stochastic partial differential equation–integrated nested laplace approximation (SPDE-INLA). Our interpolation approach includes several spatial and spatio-temporal predictors, including meteorological variables. A use case is provided that scales down the CAMS forecasts on the Italian peninsula. The calibration is focused on the concentrations of several air quality pollutants (PM10, PM2.5, NO2, and O3) at a daily resolution from a set of 750 monitoring sites, distributed throughout the Italian country. Our results show the key role that conditioning variables play in improving the forecast capabilities of ensemble predictions, thus allowing for a net improvement in the calibration with respect to ordinary EMOS strategies. From a deterministic point of view, the performance of the predictive model shows a significant improvement in the performance of the raw ensemble forecast, with an almost-zero bias, significantly reduced root mean square errors, and correlations that are almost always higher than 0.9 for each pollutant; moreover, the post-processing approach is able to significantly improve the prediction of exceedances, even for very low thresholds, such as those recently recommended by the World Health Organisation. This is particularly significant if a forecasting approach is used to predict air quality conditions and plan adequate human health protection measures, even for low alert thresholds. From a probabilistic point of view, the quality of the forecast was verified in terms of reliability and credible intervals. After post-processing, the predictive probability density functions were sharp and much better calibrated than the raw ensemble forecast. Finally, we present some additional results based on a set of gridded (4 km × 4 km) maps covering the entire Italian country for the detection of areas where pollution peaks occur (exceedances of the current and/or proposed regulatory thresholds).

Contemporary and historical human migration patterns shape hepatitis B virus diversity.

Infection by hepatitis B virus (HBV) is responsible for approximately 296 million chronic cases of hepatitis B, and roughly 880,000 deaths annually. The global burden of HBV is distributed unevenly, largely owing to the heterogeneous geographic distribution of its subtypes, each of which demonstrates different severity and responsiveness to antiviral therapy. It is therefore crucial to the global public health response to HBV that the spatiotemporal spread of each genotype is well characterized. In this study, we describe a collection of 133 newly sequenced HBV strains from recent African immigrants upon their arrival in Belgium. We incorporate these sequences-all of which we determine to come from genotypes A, D, and E-into a large-scale phylogeographic study with genomes sampled across the globe. We focus on investigating the spatio-temporal processes shaping the evolutionary history of the three genotypes we observe. We incorporate several recently published ancient HBV genomes for genotypes A and D to aid our analysis. We show that different spatio-temporal processes underlie the A, D, and E genotypes with the former two having originated in southeastern Asia, after which they spread across the world. The HBV E genotype is estimated to have originated in Africa, after which it spread to Europe and the Americas. Our results highlight the use of phylogeographic reconstruction as a tool to understand the recent spatiotemporal dynamics of HBV, and highlight the importance of supporting vulnerable populations in accordance with the needs presented by specific HBV genotypes.

Stelfi: An R package for fitting Hawkes and log-Gaussian Cox point process models.

Modelling spatial and temporal patterns in ecology is imperative to understand the complex processes inherent in ecological phenomena. Log-Gaussian Cox processes are a popular choice among ecologists to describe the spatiotemporal distribution of point-referenced data. In addition, point pattern models where events instigate others nearby (i.e., self-exciting behaviour) are becoming increasingly popular to infer the contagious nature of events (e.g., animal sightings). While there are existing R packages that facilitate fitting spatiotemporal point processes and, separately, self-exciting models, none incorporate both. We present an R package, stelfi, that fits spatiotemporal self-exciting and log-Gaussian Cox process models using Template Model Builder through a range of custom-written C++ templates. We illustrate the use of stelfi's functions fitting models to Sasquatch (bigfoot) sightings data within the USA. The structure of these data is typical of many seen in ecology studies. We show, from a temporal Hawkes process to a spatiotemporal self-exciting model, how the models offered by the package enable additional insights into the temporal and spatial progression of point pattern data. We present extensions to these well-known models that include spatiotemporal self-excitation and joint likelihood models, which are better suited to capture the complex mechanisms inherent in many ecological data. The package stelfi offers user-friendly functionality, is open source, and is available from CRAN. It offers the implementation of complex spatiotemporal point process models in R for applications even beyond the field of ecology.

Non-Separable Covariance Kernels for Spatiotemporal Gaussian Processes Based on a Hybrid Spectral Method and the Harmonic Oscillator

Non-Separable Covariance Kernels for Spatiotemporal Gaussian Processes Based on a Hybrid Spectral Method and the Harmonic Oscillator