Publication
Clustering Ocean Structures Using ARMOR 3D Data for Marine Pattern Recognition
| Summary: | Understanding ocean structure and its variations over time is critical for climate studies, marine ecosystem research, and oceanographic monitoring. This thesis investigates the application of clustering techniques to multivariate ocean data, focusing on the ARMOR3D dataset from the Copernicus Marine Environment Monitoring Service (CMEMS). The study aims to identify patterns in ocean structures across time and depth using clustering algorithms, particularly K-means, with validation against established Ecological Marine Units (EMUs). To achieve this, the research employs a depth-stratified clustering approach, where oceanographic variables are analyzed across different depths and seasons. The clustering results from 2010 and 2022 reveal distinct seasonal and interannual changes, with surface layers exhibiting the most significant variations, while deeper ocean layers remain relatively stable. A depth-wise cluster intersection analysis ensures consistency in labeling across depth levels, facilitating a better understanding of vertical oceanographic structures. The clustering results align well with EMU classifications, supporting the effectiveness of multivariate clustering in capturing marine pattern variability. These findings contribute to advancing oceanographic data analysis by demonstrating how clustering techniques can enhance the identification and interpretation of global ocean structures. |
|---|---|
| Main Authors: | Lucas, Miguel Filipe Baptista |
| Subject: | Multivariate Ocean Clustering Temporal and Depth Analysis Marine Pattern Recognition SDG 13 - Climate action SDG 14 - Life below water |
| Year: | 2025 |
| Country: | Portugal |
| Document type: | master thesis |
| Access type: | open access |
| Associated institution: | Universidade Nova de Lisboa |
| Language: | English |
| Origin: | Repositório Institucional da UNL |
| Summary: | Understanding ocean structure and its variations over time is critical for climate studies, marine ecosystem research, and oceanographic monitoring. This thesis investigates the application of clustering techniques to multivariate ocean data, focusing on the ARMOR3D dataset from the Copernicus Marine Environment Monitoring Service (CMEMS). The study aims to identify patterns in ocean structures across time and depth using clustering algorithms, particularly K-means, with validation against established Ecological Marine Units (EMUs). To achieve this, the research employs a depth-stratified clustering approach, where oceanographic variables are analyzed across different depths and seasons. The clustering results from 2010 and 2022 reveal distinct seasonal and interannual changes, with surface layers exhibiting the most significant variations, while deeper ocean layers remain relatively stable. A depth-wise cluster intersection analysis ensures consistency in labeling across depth levels, facilitating a better understanding of vertical oceanographic structures. The clustering results align well with EMU classifications, supporting the effectiveness of multivariate clustering in capturing marine pattern variability. These findings contribute to advancing oceanographic data analysis by demonstrating how clustering techniques can enhance the identification and interpretation of global ocean structures. |
|---|