Earth Science Informatics最新文献

Satellite cloud image sequences contain rich spatial and temporal information, and forecasting future cloud image sequences is of great significance for meteorological research. Traditional satellite cloud image prediction methods usually ignore nonlinear variations in cloud masses, leading to large errors in prediction results and low prediction efficiency. The use of existing video prediction methods for satellite cloud image sequence prediction also suffers from problems of blurred prediction images and the accumulation of sequence errors. To address these issues, we propose a Multi-feature Extraction and High-resolution Generative Network (MEHGNet) for the prediction of satellite cloud image sequences, which consists of an encoder, a translator, a decoder, and a generator. To learn the spatial features and spatiotemporal dependencies of cloud images, 2D convolution multi-head attention mechanisms and local residue connections are introduced to the encoder and decoder. The generator preserves detailed features and improves the resolution of the predicted images using the generative ability of generative adversarial networks. In addition, a motion-aware loss function is proposed to learn high-level features of motion variations among cloud image sequences. Experiments on satellite datasets demonstrate that the proposed method is superior compared to other prediction methods.

Accurate well log data is critical for subsurface characterisation and decision-making in the petroleum exploration. We explore and compare the effectiveness of three distinct deep leaning (DL) approaches—Long Short-Term Memory, Bidirectional Long Short-Term Memory, and Convolutional Long Short-Term Memory networks—in predicting missing well log data, a common challenge in the data acquired by Energy and Production (E&P) companies. Our analysis revealed the complex, nonlinear relationships present in geophysical logs through correlation matrix and determining the rank of predictor features through Minimum Redundancy Maximum Relevance (MRMR) analysis. To weigh these models, we used real-field wireline log datasets from the Bhogpara oil field of Upper Assam basin. The performance of each model is evaluated through root mean square error, correlation coefficients, mean absolute error and variance between actual and predicted values. The uncertainty of the models was facilitated by Monte Carlo simulation. Deep learning models accurately predicted neutron porosity logs from gamma-ray, resistivity, density, and photoelectric factor logs. The high correlation coefficients during the training (exceeding 0.90) and test (exceeding 0.97) phases illustrated the predictive precision of the DL models. Conv-LSTM consistently outperforms LSTM and Bi-LSTM, indicating the integration of convolutional layers in feature extraction offers a significant advantage in capturing intricate patterns in log data. The research showcases the effectiveness of deep learning architectures in predicting missing logs, a crucial aspect for E&P companies, as log data is vital for decision-making. The study presents a novel method for preserving data integrity and facilitating informed decision-making.

With the advancement of technology and the maturity of various aerial imaging techniques, data proprietors have awareness of the importance of secure protection for remote sensing images. In order to protect sensitive data of images, we propose a secure encoding scheme for compressing remote sensing images to decrease potential risks of data disclosure associated with such images. First, we designed the Sin chaos paradigm for constructing chaotic systems in various dimensions. As a result through relevant experiments, this chaos paradigm demonstrated effective scalability and stability. In addition, DNA transposition methods have been introduced to extend DNA encoding, expanding the range of DNA encoding from 1 to 4 and achieving dynamic selection of DNA transposition methods. This method reduces potential threats that conflict with fixed DNA encoding methods. In addition, in order to ensure the security of symmetric encryption and the efficiency of asymmetric encryption during key transmission, an elliptical curve “ring” key hiding strategy is adopted. Although the key embedding occupies 1.2% of the space in the ciphertext image, data redundancy realizes the implicit transmission of the key, improving the decryption efficiency of remote sensing images. In response to the above research, we propose a secure compression encoding scheme based on Sin chaotic paradigm and DNA transposition to ensure the security of remote sensing images. After cropping the original remote sensing image to a size of 1/16, the original image can still be decrypted. In addition, when the noise attack reaches 0.3, the ciphertext image can also be restored. Performance analysis and experimental data results show that our proposed secure compression encoding scheme has excellent robustness and security.

This comprehensive review paper examines the integration of Artificial Intelligence (AI) and Machine Learning (ML) tools in earthquake engineering, specifically focusing on damage assessment and retrofitting strategies. The paper begins with an introduction to AI and its significance in structural engineering, highlighting the need for advanced methodologies to address seismic challenges. A detailed review of recent applications of ML, Pattern Recognition (PR), and Deep Learning (DL) in earthquake engineering is provided, showcasing their capabilities in surpassing the limitations of traditional models. The advantages of employing these algorithmic methods in damage assessment, retrofitting designs, risk prediction, and structural optimization are discussed extensively. Furthermore, the paper identifies potential research avenues and emerging trends in AI/ML applications for earthquake resilience, while also addressing the challenges and limitations associated with these technologies. Overall, this review paper offers a comprehensive overview of the current state-of-the-art in AI and ML tools for earthquake damage assessment and retrofitting strategies, paving the way for future advancements in seismic resilience engineering.

The current research employed the least absolute shrinkage and selection operator (Lasso) and Elastic-net algorithms to examine their potential utilization in MVT Pb-Zn prospectivity modeling. In training the model, both Elastic-net and Lasso regularization approaches include a penalty term to the loss function. Since this penalty term limits the feature coefficients, the model is motivated to prioritize the most informative features and penalize the less relevant ones. The Varcheh district in western Iran was the source of the geological, geochemical, tectonic, and alteration dataset. We applied stratified 5-fold cross-validation to train the dataset, ensuring consistent and comprehensive performance evaluation across different data subsets. This method improved data utilization and provided more reliable performance estimates by averaging metrics over multiple folds, thereby enhancing the model’s generalization assessment. The hyperparameters were adjusted using random search, quickly finding near-optimal solutions. Our investigation revealed that Elastic-net exhibited superior prediction accuracy and model robustness compared to Lasso. The combination of L1 and L2 regularization in Elastic-net, offers a more adaptable technique than Lasso, which just utilizes L1 regularization. This feature enables Elastic-net to handle scenarios in which there have been correlated predictors successfully.

The dynamics of Sea Surface Temperature (SST) are crucial for maintaining the balance of marine ecosystems. While existing artificial intelligence methods offer powerful tools for SST prediction, they struggle with data sparsity issues. To enhance SST prediction accuracy under sparse data conditions, this study proposes an innovative prediction model: TL-iTransformer. This model is based on the iTransformer architecture and incorporates transfer learning techniques specifically tailored for SST prediction. We begin by extracting SST features from data-rich sea areas (source sea areas) using a transfer learning strategy, integrating these features into the iTransformer model for pre-training. This process imparts the model with a priori knowledge and basic prediction capabilities, enabling it to adapt to data-sparse sea areas (target sea areas). The model is then fine-tuned using domain adaptive techniques to accurately capture the data characteristics and distribution patterns of the target sea area. We conducted a series of experiments using a real SST dataset from the sea area of British Columbia, Canada. The results demonstrate that TL-iTransformer maintains the Mean Absolute Error (MAE) and Mean Squared Error (MSE) within 0.144 and 0.356, respectively, under data sparsity conditions. Additionally, it outperforms four mainstream time-series prediction baseline models as the prediction time span increases. The proposed model can effectively address the issue of SST prediction in situations with sparse data.

The article “Ionospheric scintillation characteristics over Indian region from latitudinally-aligned geodetic GPS observations” was originally published Online First without open access. After publication in volume 16, issue 3, page 2675–2691, the author decided to opt for Open Choice and to make the article an open access publication.

The main aim of this study is the analysis of the impact that occupants’ characteristics have on thermal comfort assessment, through establishing a novel PMVo model using an approximation method, based on the experimental data. The parameters which are chosen as model’s inputs are the air temperature, mean radiant temperature, relative humidity, basic clothing insulation, air velocity and occupants characteristics – gender, age, height, and body mass, while the output is the PMVo, a novel thermal comfort index. Since existing standards concerning thermal comfort do not consider these occupants’ characteristics, the main novelty of the introduced model is the inclusion of occupants’ characteristics in the thermal comfort assessment. To ensure enhanced precision, the model is established using both linear regression and by training neural network. These two approximation methods are compared to determine which one is more applicable in the context of data approximation. Study shows that regardless of dataset based on which models are established and regardless of testing input values, neural network (R² in the range of 99.87% to 99.96%) is a superior mathematical approximation algorithm compared to the linear regression (R² in the range of 95.3% to 97.5%). Novel neural network based thermal comfort assessment model is used for investigation of occupants’ characteristics impact on thermal comfort assessment. Analysis of the results showed that gender, age, height and body mass may significantly impact thermal comfort indices calculation, which implies the necessity of their inclusion in thermal comfort prediction and evaluation. Thus, the presented PMVo model may be highly beneficial to implement within existing thermal comfort standards, ensuring well-being and satisfaction with conditions of indoor environment for wider range of the occupants.

Graphical abstract

Human lives and property are threatened by Flash floods (FF) worldwide and as a result of the unprecedented conditions of the climate change effects the losses are predicted to increase in the future. As it seems difficult to avoid and prevent them, real-time flash flood detections could be an appropriate solution for damage reduction and better management. Currently, the development of computer vision applications such as deep learning and AI has been advanced. Although AI models have been developed for applications in many fields, their implementations for geosciences are limited based on large amounts of training data and the highly required computational infrastructure. Hence, this work aims to train the latest YOLOv8 model and apply it to real-time flash flood detection for regions of Korea and possibly for other nations. To overcome the shortage of training data, we created small on-site flash flood models and took pictures and footage of them. More than 1500 photos of FF were used for model trains and validations gaining a model mean average precision of above 60% of all training depths (25, 50, 75, and 100 epochs). Despite some model false positives and missed false positive detections using the Korean FF test dataset, the YOLOv8 best model generated bounding boxes (BB) with high confidence values in most FF events. Furthermore, the robustness of the model is highlighted by its ability to smoothly detect the precise positions of the FF areas with high confidence values (best 0.86) when applied for input footage and webcam streams. It is highly encouraged to establish a real-time FF warning system to reduce their negative effects. Although YOLO is effective and fast, like other deep learning models, it requires large input data to ensure higher accuracy and confidence. Future works might explore this aspect, particularly the data acquired in light inefficiency to improve the model detections at night time.

To explore tropical cyclone prediction methods that integrate multimodal meteorological data, this study proposes a novel approach. The proposed model employs an LSTM-based temporal branch to extract temporal sequence features from the CMA dataset and a U-Net-based spatial branch to extract three-dimensional spatial features from the ERA5 dataset. These features are then fused through an encoder-decoder structure to integrate high-dimensional spatiotemporal characteristics. Experimental results demonstrate that the spatiotemporal model significantly improves the prediction accuracy for 24-hour lead times. Subsequently, to further optimize the experimental results, the study introduces an ensemble forecasting framework. This framework enhances prediction accuracy by adjusting the outputs of multiple spatiotemporal model prediction members. The optimization is achieved by solving the objective function that reflects the forecast geographical error, thereby optimizing the weighted coefficients. The experimental results indicate that the ensemble forecasting framework can further optimize prediction outcomes.