Computers & Geosciences最新文献

{"title":"OPTIM: A Python-based optimization framework for geophysical problems","authors":"Tao Lei ,&nbsp;Wei Zhang ,&nbsp;Yongming Lu ,&nbsp;Li Yang","doi":"10.1016/j.cageo.2025.105930","DOIUrl":"10.1016/j.cageo.2025.105930","url":null,"abstract":"<div><div>Geophysical inverse problems, such as full waveform inversion, involve significant computational demands and algorithmic complexity. Geophysicists aim to resolve numerous unknown parameters within a limited number of inversion iterations, necessitating both efficient and accurate geophysical modules (e.g., forward modeling and sensitivity kernel calculations) and robust optimization frameworks to drive the inversion process. To facilitate the rapid construction of comprehensive inversion workflows, we present <em>OPTIM</em>, a Python-based open-source local optimization software package. <em>OPTIM</em> structures each optimization step as an independent program, exchanging information between adjacent steps through files and parameters. Its implementation closely follows mathematical formulations, allowing users to easily identify and modify specific modules as needed. Constructing an inversion workflow with the proposed software is analogous to assembling modular components, minimizing concerns about program interfaces and lifecycle management. Through a series of examples, we demonstrate how the proposed software enables efficient inverse workflow construction and large-scale geophysical inversion on multi-node high-performance clusters. <em>OPTIM</em> empowers researchers to rapidly and robustly develop novel geophysical inversion processes without compromising on performance and scalability. This capability significantly streamlines the complexity in solving geophysical inverse problems and accelerates the development cycle.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"201 ","pages":"Article 105930"},"PeriodicalIF":4.2,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143824416","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A method for preserving three spatial features in the upscaling of categorical raster data","authors":"Xiangyuan He,&nbsp;Chen Zhou,&nbsp;Mingzhu Gao,&nbsp;Saisai Sun,&nbsp;Chiying Lyu,&nbsp;Xiaoyi Han","doi":"10.1016/j.cageo.2025.105933","DOIUrl":"10.1016/j.cageo.2025.105933","url":null,"abstract":"<div><div>Raster resampling can be used to modify the resolution of raster data to satisfy specific application requirements for geographical information systems (GIS). However, with an increase in raster cell size, a process known as upscaling, various spatial features are inevitably lost, resulting in reduced data accuracy. Categorical raster data refer to a raster dataset where each specific raster value corresponds to a category, such as land use types or vegetation cover, rather than continuous numerical values. To improve the accuracy of upscaled data, this study proposes a method for preserving the shape, topological, and area features in categorical raster upscaling. First, we refined the shape index calculation to accurately assess the shape of the raster zones and corrected the shape errors using neighborhood operations. Second, we resolved the topological errors by reassigning the cells between the raster zones. Finally, we calculated the number of cells that needed adjustment in each zone and reassigned the cells on the zone boundaries accordingly, to reduce the overall area error. The results demonstrated a 30.5235 % improvement in the accuracy, compared with the accuracy of the nearest neighbor method for upscaling from 5 to 10 m. The effectiveness of the proposed method decreased with increasing target cell size, with the method being ineffective at 35 m. Furthermore, the method demonstrates wide applicability across different datasets. By efficiently and simultaneously maintaining these spatial features during upscaling, our method can offer users more accurate resampled datasets as input for GIS applications, thereby enhancing the precision of the outputs.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"201 ","pages":"Article 105933"},"PeriodicalIF":4.2,"publicationDate":"2025-04-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143817292","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A spatiotemporal mixed-enhanced generative adversarial network for radar-based precipitation nowcasting","authors":"Long He ,&nbsp;Kun Zheng ,&nbsp;Huihua Ruan ,&nbsp;Shuo Yang ,&nbsp;Jinbiao Zhang ,&nbsp;Cong Luo ,&nbsp;Siyu Tang ,&nbsp;Yunlei Yi ,&nbsp;Yugang Tian ,&nbsp;Jianmei Cheng","doi":"10.1016/j.cageo.2025.105919","DOIUrl":"10.1016/j.cageo.2025.105919","url":null,"abstract":"<div><div>Skillful precipitation nowcasting with high resolution and detailed information holds promise for providing reliable alerts about severe weather events to society. Radar echo extrapolation is an essential method for precipitation nowcasting, but traditional methods struggle to capture rapidly changing regions. Deep learning (DL)-based methods exhibit superior performance. However, existing DL-based methods face challenges such as low accuracy, particularly in producing clear forecasts over longer lead times and accurately forecasting moderate to heavy rainfall events. To address these challenges, we developed a novel radar-based precipitation nowcasting model, STMixGAN, which can be described as a nonlinear proximity forecasting model. This model effectively aggregates global-to-local information and imposes constraints to represent the complex evolution of rainfall efficiently. Consequently, STMixGAN produces realistic and spatiotemporally consistent predictions. Using radar observations from South China, STMixGAN successfully forecasted radar maps for the next 1 h using 24 min of input data. Two traditional methods (Persistence and Optical flow) and five DL-based methods (ConvLSTM, Rainformer, IAM4VP, REMNet, and GAN-argcPredNet) were employed as benchmarks to validate STMixGAN’s forecasting capabilities. The experimental results demonstrate STMixGAN’s superior performance and provide valuable insights for enhancing heavy rainfall forecasting.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"200 ","pages":"Article 105919"},"PeriodicalIF":4.2,"publicationDate":"2025-03-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143715501","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Processing pipeline for fully-automated computation of 3D glacier surface flow time series","authors":"Ayush Gupta ,&nbsp;Balaji Devaraju ,&nbsp;Ashutosh Tiwari","doi":"10.1016/j.cageo.2025.105918","DOIUrl":"10.1016/j.cageo.2025.105918","url":null,"abstract":"<div><div>To comprehend glacier dynamics for a region, a time-series study of glacier change is essential. Most of the existing research focus on understanding the long-term changes in glaciers to understand how glaciers adapt to a warming climate. However, generation of a large time series data often requires a substantial amount of computational time and resources, and no automated pipelines exist for calculating glacier surface flow. We address these problems by building an automated pipeline for efficient processing of satellite SAR images, generating extensive time series data for tracking glacier changes. This pipeline employs Sentinel-1 (S-1) interferometric wide swath SAR data to produce seasonal time series of glacier surface displacements over prolonged durations. The pipeline utilizes the ISCE framework for SAR data processing, and introduces a robust offset tracking module designed to perform offset tracking across an image time-series through Normalized cross correlation (NCC) stacking. It performs offset tracking on both ascending and descending S-1 images to compute displacements in both azimuth and range directions. These displacements are later utilized to compute northward, eastward, and vertical surface displacements between consecutive time-steps through weighted least squares, with optimal weights designed to make the model more robust. We demonstrate the proposed pipeline over three valley-type glaciers (Bara Shigri, Geepang Gath, and Samudra Tapu) located in the Chandra basin, Himachal Pradesh, India, generating 3D surface displacement time series from 2017 to 2022. The software, automating the entire process of glacier surface displacement time-series computation, is open access, and can be employed to monitor varieties of glaciers using current and upcoming SAR sensors with frequent revisits.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"200 ","pages":"Article 105918"},"PeriodicalIF":4.2,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143697810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A deep learning method for 3D geological modeling using ET4DD with offset-attention mechanism","authors":"Anjing Ren ,&nbsp;Liang Wu ,&nbsp;Jianglong Xu ,&nbsp;Yanjie Xing ,&nbsp;Qinjun Qiu ,&nbsp;Zhong Xie","doi":"10.1016/j.cageo.2025.105929","DOIUrl":"10.1016/j.cageo.2025.105929","url":null,"abstract":"<div><div>Deep learning-based methods for 3D geological modeling can automatically identify significant geological features, which is crucial for intelligent 3D geological modeling. We propose a 3D geological modeling method based on ET4DD (Enhanced Transformer for Drilling Data). This deep learning model accurately predicts the lithology categories of 3D points. The study area for our experiment is located in the Tianfu New Area of Chengdu, Sichuan Province, China. We conduct data pre-processing operations, including resampling and standardization, on the data collected from 719 boreholes in the study area. The dataset was split into training and test sets at a 4:1 ratio. To validate the effectiveness of the model, we train a standard stacking model that integrates FNN, RF, GBDT, and XGBoost using the same dataset. The comparison shows that ET4DD achieves the highest precision, recall, and F1 score among all models, with respective scores of 97.33 %, 97.33 %, and 97.29 %. We use MapGIS software to visualize the lithology grid cells predicted by ET4DD, and select three subregions from the geological model for detailed comparison with the stacking method, complemented by visualizations of uncertainty. The results demonstrate that our method effectively captures geological variability and reduces the informational complexity of the geological model. In addition, the geological model generated by our method reveals the geological regularities, including the topology of geological bodies, the geometrical form of strata, and the spatial distribution characteristics of lithological units, leading to more accurate and reasonable simulations of actual subsurface geological conditions.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"200 ","pages":"Article 105929"},"PeriodicalIF":4.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143697811","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unsupervised seismic reconstruction via deep learning with one-dimensional signal representation","authors":"Gui Chen ,&nbsp;Yang Liu ,&nbsp;Mi Zhang ,&nbsp;Yuhang Sun ,&nbsp;Haoran Zhang","doi":"10.1016/j.cageo.2025.105916","DOIUrl":"10.1016/j.cageo.2025.105916","url":null,"abstract":"<div><div>The geometry of seismic surveys often requires modification due to constraints in the acquisition environment, leading to incomplete spatial data coverage and a reduction in the quality and resolution of subsurface imaging. Seismic data reconstruction is critical to restore spatial continuity in such cases. Thus, we propose a novel unsupervised deep learning framework with a trace-by-trace recovery strategy to reconstruct incomplete 3D seismic data, which does not require ground-truth data as the training label. In the proposed framework, we design a one-dimensional signal representation neural network (SRNet) based on the Kolmogorov-Arnold network, which depends on the inherent spatial correlation of seismic signals to achieve single-trace signal representation when using the corresponding spatial coordinate values as its input. The parameters of the SRNet are optimized in a self-learning manner using observed traces and their corresponding spatial coordinates, allowing the well-optimized SRNet to predict seismic signals at unobserved locations. Experiments on the synthetic data and field data examples indicate the effectiveness of the proposed method, and its superiority over the benchmark methods.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"200 ","pages":"Article 105916"},"PeriodicalIF":4.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143686777","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Advancing Earth science applications through a semi-automatic monoplotting framework for efficient 3D geo-referencing of monocular oblique visual Data","authors":"Behzad Golparvar,&nbsp;Ruo-Qian Wang","doi":"10.1016/j.cageo.2025.105915","DOIUrl":"10.1016/j.cageo.2025.105915","url":null,"abstract":"<div><div>The widespread availability of high-quality images from smartphones, drones, and digital cameras presents an unprecedented opportunity for global geospatial data collection. However, these images are often captured at oblique angles, making geo-referencing challenging and limiting their usability. Monoplotting, a technique that requires only a single image and a Digital Elevation Model (DEM), addresses these challenges by establishing pixel-level correspondence between imagery and real-world coordinates. However, traditional monoplotting methods are labor-intensive, requiring manual identification of control points in both the image and DEM, as well as manual tuning of camera parameters, which restricts scalability for large-scale databases and near-real-time applications. This paper proposes a novel semi-automatic monoplotting framework that minimizes human intervention. The framework integrates key point detection, geo-referenced 3D point retrieval, regularized gradient-based optimization, pose estimation, back-projection, and pixel mapping to enable efficient geo-referencing. To the best of the authors’ knowledge, this is the first study to incorporate key point detection into monoplotting, and apply regularized gradient-based optimization for camera position and parameter determination, even with unequal numbers of key points from the image and DEM. Numerical experiments with a historical image and a corresponding real-world DEM demonstrate the framework’s effectiveness. The robustness of the method is further evaluated on distorted images, where the distortion strength coefficient is treated as an unknown and estimated through projection optimization. The results confirm the framework’s ability to establish accurate correspondence between the image pixel domain and real-world 3D coordinates. Additionally, integrating machine learning models, such as semantic segmentation, highlights the framework’s advantages in Earth science applications, including snow and glacier characterization.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"200 ","pages":"Article 105915"},"PeriodicalIF":4.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143686778","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A geological knowledge-constrained entity and relation extraction method for text: A case study of granitic pegmatite-type lithium deposits","authors":"Jintao Tao ,&nbsp;Nannan Zhang ,&nbsp;Jinyu Chang ,&nbsp;Li Chen ,&nbsp;Hao Zhang ,&nbsp;Shibin Liao ,&nbsp;Siyuan Li ,&nbsp;Jianpeng Jing","doi":"10.1016/j.cageo.2025.105920","DOIUrl":"10.1016/j.cageo.2025.105920","url":null,"abstract":"<div><div>Geological text data contain rich and valuable information about geological environments and mineral deposits. The automated extraction of geological information from these unstructured texts is crucial for constructing geological knowledge graphs and facilitating knowledge discovery. Numerous studies have introduced methods for geological entity and relation extraction from different perspectives. Although many of these studies effectively utilize geological ontologies or schemas for data labeling, fewer have explicitly examined how these frameworks can constrain and improve the information extraction process. In this study, we propose a Geological Knowledge-constrained Entity and Relation Extraction (GKERE) method that incorporates a geological schema to enhance the extraction process. The GKERE method uses the Robustly Optimize Bidirectional Encoder Representation from Transformers Pre-training Approach to generate character embeddings from geological sentences. It begins with a span-based named entity recognition model to identify entities, then generates entity pairs and predicts their relationships using the geological schema. The schema helps filter out redundant entity pairs and provides information about the types of head/tail entities and their possible relationships, guiding the relation extraction step. To validate the method, we conducted a case study on granitic pegmatite-type lithium deposits. A geological schema was designed, comprising 22 entity types, 16 relationships, and 184 knowledge rules. An entity-relation extraction dataset was then constructed using 68 geological journal articles and four mineral exploration reports. The proposed GKERE method achieves an impressive F1-score of 75.82 % on this dataset, outperforming several baseline models. Results show that the GKERE method significantly enhances geological entity and relation extraction. The introduction of the geological schema not only accelerates computation but also improves model accuracy, making this approach effective for extracting geological information from large-scale textual data and promoting geological knowledge discovery.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"200 ","pages":"Article 105920"},"PeriodicalIF":4.2,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143686779","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The multi-GPU Wetland DEM Ponding Model","authors":"Tonghe Liu ,&nbsp;Sean J. Trim ,&nbsp;Seok-Bum Ko ,&nbsp;Raymond J. Spiteri","doi":"10.1016/j.cageo.2025.105912","DOIUrl":"10.1016/j.cageo.2025.105912","url":null,"abstract":"<div><div>The Wetland DEM (Digital Elevation Model) Ponding Model (<span>WDPM</span>) is software that simulates how runoff water is distributed across the Canadian Prairies. Previous versions of the <span>WDPM</span> are able to run in parallel with a single CPU or GPU. Now that multi-device parallel computing has become an established method to increase computational throughput and efficiency, this study extends <span>WDPM</span> to a multi-GPU parallel algorithm with efficient data transmission methods via overlapping communication with computation. The new implementation is evaluated from several perspectives. First, the output summary and system are compared with the previous implementation to verify correctness and demonstrate convergence. Second, the multi-GPU code is profiled, showing that the algorithm carries out efficient data synchronization through optimized techniques. Finally, the new implementation was tested experimentally and showed improved performance and good scaling. Specifically, a speedup of 2.39 was achieved when using four GPUs compared to using one GPU.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"199 ","pages":"Article 105912"},"PeriodicalIF":4.2,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143619687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantitative bed-type classification for a global comparison of deep-water sedimentary systems","authors":"Soma Budai ,&nbsp;Luca Colombera ,&nbsp;Adam McArthur ,&nbsp;Marco Patacci","doi":"10.1016/j.cageo.2025.105917","DOIUrl":"10.1016/j.cageo.2025.105917","url":null,"abstract":"<div><div>Characterisation of deep-water successions is often undertaken at the scale of sedimentary beds. However, different studies often apply alternative bed-type classification schemes, rendering the quantitative comparison of bed properties of different deep-water systems difficult. In this study a quantitative approach to the development of a universal deep-water bed-type classification scheme is proposed based on the synthesis of a large sedimentological dataset, containing &gt;32,000 deep-water facies and &gt;10,000 beds accumulated in 27 turbidite-dominated systems. The classification scheme is applicable to discriminate and categorise lithological (sand, gravel) layers and is based on: (i) the proportion of, gravel, sand, sandy-mud and muddy-sand in the bed, (ii) the presence and nature of vertical sharp grain-size changes, and (iii) the presence and thickness ratio of laminated sedimentary facies. Comparing the bedding properties of channel-fills, terminal deposits (e.g. lobes or sheets) and levees showed that the three architectural-element types are characterised by differences in bed frequency and thickness, overlying mudstone proportions, vertical bed thickness trends, mud thickness and sand-gravel fraction values. Building on these recognised statistical differences an algorithm was developed that is capable of generating, in a stochastic manner, geologically realistic synthetic sedimentary logs depicting deep-water terminal-deposit, channel-fill and levee elements. The one-dimensional facies modelling is governed by a series of input parameters, including total number of beds, sand-gravel thickness, and sand-gravel fraction. The approach can be tailored to produce synthetic logs for specified types of depositional systems (e.g., categorised according to dominant grain size of deposits, age of deposition and global climate (icehouse vs. greenhouse conditions)). A large number of synthetic sedimentary logs can be generated, which can be utilised as training datasets in machine learning algorithms developed to aid subsurface interpretations of clastic sedimentary successions.</div></div>","PeriodicalId":55221,"journal":{"name":"Computers & Geosciences","volume":"199 ","pages":"Article 105917"},"PeriodicalIF":4.2,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143619686","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}