[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125211-en":3,"doc-seo-125211-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},125211,1374391975076,"Riley","https://ap-avatar.wpscdn.com/avatar/14000253ca4ec9f6853?x-image-process=image/resize,m_fixed,w_180,h_180&k=1783305029341752051",8,"Research & Report","Machine Learning Applications in Physical Oceanography","The dissertation addresses limitations of traditional ocean research methods in resolution, interpretability, and real-time applicability by introducing a machine-learning framework for physical oceanography. It integrates Gaussian Process Regression and Vision Transformers to analyze Lagrangian and remotely sensed data. GPR is evaluated for probabilistic interpolation, reconstructing Eulerian velocity fields from sparse drifter observations with quantified uncertainty. ViT is then applied to sea-surface classification from SAR imagery and combined with GPR to detect and interpret dynamical Regions of Interest, enabling interpretable velocity, strain, vorticity, and divergence analysis across complex regimes.","Please do not remove this page  \nMachine Learning Applications in Physical Oceanography  \nXia, Junfei [https://scholarship.miami.edu/esploro/outputs/doctoral/Machine-Learning-Applications-in-Physical-Oceanography/991032634690402976/fi](https://scholarship.miami.edu/esploro/outputs/doctoral/Machine-Learning-Applications-in-Physical-Oceanography/991032634690402976/fi)lesAndLinks?index=0  \nXia, J. (2025) . Machine Learning Applications in Physical Oceanography [University of Miami] . [https://scholarship.miami.edu/esploro/outputs/doctoral/Machine-Learning-Applications-in-Physical-O](https://scholarship.miami.edu/esploro/outputs/doctoral/Machine-Learning-Applications-in-Physical-O)[ceanography/991032634690402976](ceanography/991032634690402976)  \nOpen  \nDownloaded On 2026/04/25 17:42:05-0400  \nPlease do not remove this page  \nUNIVERSITY OF MIAMI  \nMACHINE LEARNING APPLICATIONS IN PHYSICAL OCEANOGRAPHY  \nBy  \nJunfei Xia  \nA DISSERTATION  \nSubmitted to the Faculty  \nof the University of Miami in partial fulfillment of the requirements for the degree of Doctor of Philosophy  \nCoral Gables, Florida  \nMay 2025  \n©2025 Junfei Xia All Rights Reserved  \nUNIVERSITY OF MIAMI  \nA dissertation submitted in partial fulfillment of the requirements for the degree of  \nDoctor of Philosophy  \nMACHINE LEARNING APPLICATIONS IN PHYSICAL OCEANOGRAPHY  \nJunfei Xia  \nApproved:  \nTamay M. ¨Ozgkmen, Ph.D. Professor of Meteorology and Physical Oceanography  \nRoland Romeiser, Dr.rer.nat.  \nProfessor of Meteorology and Physical Oceanography  \nPierre F.J Lermusiaux, Ph.D. Professor of Mechanical Engineering and Ocean Science  \nMassachusetts Institute of Technology  \nMohamed Iskandarani, Ph.D.  \nProfessor of Meteorology and Physical Oceanography  \nNicole Leeper Piquero, Ph.D.  \nInterim Dean of the Graduate School  \nXIA, JUNFEI (Ph.D., Meteorology and Physical Oceanography)  \nMachine Learning Applications in Physical Oceanography  (May 2025)  \nAbstract of a dissertation at the University of Miami.  \nDissertation supervised by Professor Tamay M. ¨Ozgkmen.  \nNo. of pages in text. (157)  \nThe ocean is a dynamically rich environment governed by nonlinear interactions spanning multiple spatial and temporal scales. Traditional methods for studying ocean flows, such as numerical simulations and Eulerian-based measurements, face limitations in resolution, interpretability, and real-time applicability. This dissertation introduces a machine learning (ML)-driven framework to address these limitations by integrating Gaussian Process Regression (GPR) and Vision Transformers (ViT) for enhanced analysis of Lagrangian and remotely sensed oceanographic data.  \nWe begin by evaluating GPR as a probabilistic interpolation tool that reconstructs Eulerian velocity fields from sparse Lagrangian drifter data. Performance is assessed using two model-based datasets: a time-periodic double-gyre simulation and a convergence region derived from the Navy Coastal Ocean Model (NCOM) . Results show that 3D GPR (in 2D space + time) yields smooth velocity estimates with quantifiable uncertainty, particularly excelling in regions with dense sampling or structured dynamics.  \nIn Chapter 3, we extend the application of ViT to sea surface classification from synthetic aperture radar (SAR) imagery. Using the TenGeoP-SARwv and AI4Arctic Sea Ice Challenge datasets, we train and evaluate a custom ViT model to categorize sea surface  \nphenomena. Results show that ViT outperforms traditional convolutional neural networks (CNNs) in both texture-based and texture-structure-based classes, particularly for highresolution HH and HV polarized satellite data. The model demonstrates strong generalization, revealing ViT’s promise in remote sensing applications with minimal need for handcrafted features.  \nChapter 4 fuses both ViT and GPR approaches to identify and interpret Regions of Interest (ROIs) in real-world drifter datasets. A multilabel ViT model is trained on synthetic data representing canonical dynamic","cbCaicp1fKfYNJut","https://ap.wps.com/l/cbCaicp1fKfYNJut","pdf",54428645,1,177,"English","en",105,"# Abstract\n# Background and Motivation\n# Gaussian Process Regression for Velocity Reconstruction\n## Evaluation datasets and performance findings\n# Vision Transformers for Sea Surface Classification\n## SAR datasets and model comparison\n# Fusing ViT and GPR for ROI Detection and Interpretation\n## Attention-based dynamical interval highlighting\n## GPR-based interpolation and visualization\n# Contributions and Impact","[{\"question\":\"What problem does the dissertation aim to solve in physical oceanography?\",\"answer\":\"It tackles limits of traditional approaches in resolution, interpretability, and real-time usability by using machine learning to extract and analyze ocean dynamics from observational data.\"},{\"question\":\"How is Gaussian Process Regression used in the proposed framework?\",\"answer\":\"GPR performs probabilistic interpolation to reconstruct Eulerian velocity fields from sparse Lagrangian drifter data, providing smooth estimates with quantified uncertainty.\"},{\"question\":\"How do Vision Transformers contribute to the dissertation’s ocean analysis tasks?\",\"answer\":\"ViT models classify sea-surface phenomena from SAR imagery and support dynamical regime identification with strong generalization and reduced reliance on handcrafted features.\"}]","Machine Learning Applications in Physical Oceanography | PDF",1785897473,446,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"machine-learning-applications-in-physical-oceanography","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-applications-in-physical-oceanography/125211/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the dissertation aim to solve in physical oceanography?","Question",{"text":75,"@type":76},"It tackles limits of traditional approaches in resolution, interpretability, and real-time usability by using machine learning to extract and analyze ocean dynamics from observational data.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is Gaussian Process Regression used in the proposed framework?",{"text":80,"@type":76},"GPR performs probabilistic interpolation to reconstruct Eulerian velocity fields from sparse Lagrangian drifter data, providing smooth estimates with quantified uncertainty.",{"name":82,"@type":73,"acceptedAnswer":83},"How do Vision Transformers contribute to the dissertation’s ocean analysis tasks?",{"text":84,"@type":76},"ViT models classify sea-surface phenomena from SAR imagery and support dynamical regime identification with strong generalization and reduced reliance on handcrafted features.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]