Geospatiotemporal Data Research Articles

3D Convolutional Neural Network with Dimension Reduction and Metric Learning for Crop Yield Prediction Based on Remote Sensing Data

Crop yield prediction is essential for tasks like determining the optimal profile of crops to be planted, allocating government resources, effectively planning and preparing for aid distribution, making decisions about imports, and so on. Crop yield prediction using remote sensing data during the growing season is helpful to farm planning and management, which has received increasing attention. Information mining from multichannel geo-spatiotemporal data brings many benefits to crop yield prediction. However, most of the existing methods have not fully utilized the dimension reduction technology and the spatiotemporal feature of the data. In this paper, a new approach is proposed to predict the yield from multispatial images by using the dimension reduction method and a 3D convolutional neural network. In addition, regions with similar crop yields should have similar features learned by the network. Thus, metric learning and multitask learning are used to learn more discriminative features. We evaluated the proposed method on county-level soybean yield prediction in the United States, and the experimental results show the effectiveness of the proposed method. The proposed method provides new ideas for crop yield estimation and effectively improves the accuracy of crop yield estimation.

DeepGPD: A Deep Learning Approach for Modeling Geospatio-Temporal Extreme Events

Geospatio-temporal data are pervasive across numerous application domains.These rich datasets can be harnessed to predict extreme events such as disease outbreaks, flooding, crime spikes, etc. However, since the extreme events are rare, predicting them is a hard problem. Statistical methods based on extreme value theory provide a systematic way for modeling the distribution of extreme values. In particular, the generalized Pareto distribution (GPD) is useful for modeling the distribution of excess values above a certain threshold. However, applying such methods to large-scale geospatio-temporal data is a challenge due to the difficulty in capturing the complex spatial relationships between extreme events at multiple locations. This paper presents a deep learning framework for long-term prediction of the distribution of extreme values at different locations. We highlight its computational challenges and present a novel framework that combines convolutional neural networks with deep set and GPD. We demonstrate the effectiveness of our approach on a real-world dataset for modeling extreme climate events.

Physics-Informed Tensor-Train ConvLSTM for Volumetric Velocity Forecasting of the Loop Current

According to the National Academies, a week long forecast of velocity, vertical structure, and duration of the Loop Current (LC) and its eddies at a given location is a critical step toward understanding their effects on the gulf ecosystems as well as toward anticipating and mitigating the outcomes of anthropogenic and natural disasters in the Gulf of Mexico (GoM). However, creating such a forecast has remained a challenging problem since LC behavior is dominated by dynamic processes across multiple time and spatial scales not resolved at once by conventional numerical models. In this paper, building on the foundation of spatiotemporal predictive learning in video prediction, we develop a physics informed deep learning based prediction model called—Physics-informed Tensor-train ConvLSTM (PITT-ConvLSTM)—for forecasting 3D geo-spatiotemporal sequences. Specifically, we propose (1) a novel 4D higher-order recurrent neural network with empirical orthogonal function analysis to capture the hidden uncorrelated patterns of each hierarchy, (2) a convolutional tensor-train decomposition to capture higher-order space-time correlations, and (3) a mechanism that incorporates prior physics from domain experts by informing the learning in latent space. The advantage of our proposed approach is clear: constrained by the law of physics, the prediction model simultaneously learns good representations for frame dependencies (both short-term and long-term high-level dependency) and inter-hierarchical relations within each time frame. Experiments on geo-spatiotemporal data collected from the GoM demonstrate that the PITT-ConvLSTM model can successfully forecast the volumetric velocity of the LC and its eddies for a period greater than 1 week.

Analysis of geo-spatiotemporal data using machine learning algorithms and reliability enhancement for urbanization decision support

ABSTRACT We present systematic analyses of the temporal dynamics of the growth of Kumasi, the fastest growing city in Ghana using 20-year Landsat time-series data from 2000 to 2020 (with 1986 Landsat image as a baseline). Two classification algorithms – random forest (RF) and support vector machines (SVM) – were used to produce binary (built-up / non-built up) maps for all years within the temporal span. We further implemented an anomaly detection and temporal consistency algorithm followed by a changing logic to correct the classification anomalies due to image contamination from the cloud and other sources. The mean overall accuracies obtained for RF and SVM were 94.9% (kappa = 0.90) and 95.5% (kappa = 0.91), respectively. Our results reveal that the mean built-up area percentages of the metropolis are approximately 74, 65, 47, and 23 for the years 2020, 2010, 2000, and 1986, respectively, representing a mean annual change of 3.5% over the 34 years. With the present lack of labeled data in Ghana for in-depth analyses of the evolution of land use, we believe that this study serves as an initial attempt to a better understanding of the effects of increasing anthropogenic activities due to urbanization, on human and environment health.

Cluster Analysis-Based Approaches for Geospatiotemporal Data Mining of Massive Data Sets for Identification of Forest Threats

We investigate methods for geospatiotemporal data mining of multi-year land surface phenology data (250 m2 Normalized Difference Vegetation Index (NDVI) values derived from the Moderate Resolution Imaging Spectrometer (MODIS) in this study) for the conterminous United States (CONUS) as part of an early warning system for detecting threats to forest ecosystems. The approaches explored here are based on k-means cluster analysis of this massive data set, which provides a basis for defining the bounds of the expected or “normal” phenological patterns that indicate healthy vegetation at a given geographic location. We briefly describe the computational approaches we have used to make cluster analysis of such massive data sets feasible, describe approaches we have explored for distinguishing between normal and abnormal phenology, and present some examples in which we have applied these approaches to identify various forest disturbances in the CONUS.

A Visual Analytics Approach to Understanding Spatiotemporal Hotspots

As data sources become larger and more complex, the ability to effectively explore and analyze patterns among varying sources becomes a critical bottleneck in analytic reasoning. Incoming data contain multiple variables, high signal-to-noise ratio, and a degree of uncertainty, all of which hinder exploration, hypothesis generation/exploration, and decision making. To facilitate the exploration of such data, advanced tool sets are needed that allow the user to interact with their data in a visual environment that provides direct analytic capability for finding data aberrations or hotspots. In this paper, we present a suite of tools designed to facilitate the exploration of spatiotemporal data sets. Our system allows users to search for hotspots in both space and time, combining linked views and interactive filtering to provide users with contextual information about their data and allow the user to develop and explore their hypotheses. Statistical data models and alert detection algorithms are provided to help draw user attention to critical areas. Demographic filtering can then be further applied as hypotheses generated become fine tuned. This paper demonstrates the use of such tools on multiple geospatiotemporal data sets.

Cognitive Fit Between Conceptual Schemas and Internal Problem Representations: The Case of Geospatio–Temporal Conceptual Schema Comprehension

Geospatio-temporal conceptual models provide a mechanism to explicitly represent geospatial and temporal aspects of applications. Such models, which focus on both what and when/where, need to be more expressive than conventional conceptual models (e.g., the ER model), which primarily focus on what is important for a given application. In this study, we view conceptual schema comprehension of geospatio-temporal data semantics in terms of matching the external problem representation (that is, the conceptual schema) to the problem-solving task (that is, syntactic and semantic comprehension tasks), an argument based on the theory of cognitive fit. Our theory suggests that an external problem representation that matches the problem solver's internal task representation will enhance performance, for example, in comprehending such schemas. To assess performance on geospatio-temporal schema comprehension tasks, we conducted a laboratory experiment using two semantically identical conceptual schemas, one of which mapped closely to the internal task representation while the other did not. As expected, we found that the geospatio-temporal conceptual schema that corresponded to the internal representation of the task enhanced the accuracy of schema comprehension; comprehension time was equivalent for both. Cognitive fit between the internal representation of the task and conceptual schemas with geospatio-temporal annotations was, therefore, manifested in accuracy of schema comprehension and not in time for problem solution. Our findings suggest that the annotated schemas facilitate understanding of data semantics represented on the schema.

Augmenting a conceptual model with geospatiotemporal annotations

While many real-world applications need to organize data based on space (e.g., geology, geomarketing, environmental modeling) and/or time (e.g., accounting, inventory management, personnel management), existing conventional conceptual models do not provide a straightforward mechanism to explicitly capture the associated spatial and temporal semantics. As a result, it is left to database designers to discover, design, and implement - on an ad hoc basis - the temporal and spatial concepts that they need. We propose an annotation-based approach that allows a database designer to focus first on nontemporal and nongeospatial aspects (i.e., what) of the application and, subsequently, augment the conceptual schema with geospatiotemporal annotations (i.e., when and where). Via annotations, we enable a supplementary level of abstraction that succinctly encapsulates the geospatiotemporal data semantics and naturally extends the semantics of a conventional conceptual model. An overarching assumption in conceptual modeling has always been that expressiveness and formality need to be balanced with simplicity. We posit that our formally defined annotation-based approach is net only expressive, but also straightforward to understand and implement.