Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods

In the Arctic Ocean, obtaining water levels from satellite altimetry is hampered by the presence of sea ice. Hence, water level retrieval requires accurate detection of fractures in the sea ice (leads). This paper describes a thorough assessment of various surface type classification methods, includ...

Full description

Bibliographic Details
Published in:Marine Geodesy
Main Authors: Bij de Vaate, I. (author), Martin, Ericka (author), Slobbe, D.C. (author), Naeije, M.C. (author), Verlaan, M. (author)
Format: Article in Journal/Newspaper
Language:English
Published: 2022
Subjects:
Online Access:http://resolver.tudelft.nl/uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa
https://doi.org/10.1080/01490419.2022.2089412
_version_ 1821818525893787648
author Bij de Vaate, I. (author)
Martin, Ericka (author)
Slobbe, D.C. (author)
Naeije, M.C. (author)
Verlaan, M. (author)
author_facet Bij de Vaate, I. (author)
Martin, Ericka (author)
Slobbe, D.C. (author)
Naeije, M.C. (author)
Verlaan, M. (author)
author_sort Bij de Vaate, I. (author)
collection Delft University of Technology: Institutional Repository
container_start_page 1
container_title Marine Geodesy
description In the Arctic Ocean, obtaining water levels from satellite altimetry is hampered by the presence of sea ice. Hence, water level retrieval requires accurate detection of fractures in the sea ice (leads). This paper describes a thorough assessment of various surface type classification methods, including a thresholding method, nine supervised-, and two unsupervised machine learning methods, applied to Sentinel-3 Synthetic Aperture Radar Altimeter data. For the first time, the simultaneously sensed images from the Ocean and Land Color Instrument, onboard Sentinel-3, were used for training and validation of the classifiers. This product allows to identify leads that are at least 300 meters wide. Applied to data from winter months, the supervised Adaptive Boosting, Artificial Neural Network, Naïve-Bayes, and Linear Discriminant classifiers showed robust results with overall accuracies of up to 92%. The unsupervised Kmedoids classifier produced excellent results with accuracies up to 92.74% and is an attractive classifier when ground truth data is limited. All classifiers perform poorly on summer data, rendering surface classifications that are solely based on altimetry data from summer months unsuitable. Finally, the Adaptive Boosting, Artificial Neural Network, and Bootstrap Aggregation classifiers obtain the highest accuracies when the altimetry observations include measurements from the open ocean. Physical and Space Geodesy Astrodynamics & Space Missions Mathematical Physics
format Article in Journal/Newspaper
genre Arctic
Arctic Ocean
Sea ice
genre_facet Arctic
Arctic Ocean
Sea ice
geographic Arctic
Arctic Ocean
geographic_facet Arctic
Arctic Ocean
id fttudelft:oai:tudelft.nl:uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa
institution Open Polar
language English
op_collection_id fttudelft
op_container_end_page 34
op_doi https://doi.org/10.1080/01490419.2022.2089412
op_relation http://www.scopus.com/inward/record.url?scp=85133721554&partnerID=8YFLogxK
Marine Geodesy: an international journal of ocean surveys, mapping and sensing--0149-0419--1ecf91d1-a328-4e46-b073-07cdd79948c0
http://resolver.tudelft.nl/uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa
https://doi.org/10.1080/01490419.2022.2089412
op_rights © 2022 I. Bij de Vaate, Ericka Martin, D.C. Slobbe, M.C. Naeije, M. Verlaan
publishDate 2022
record_format openpolar
spelling fttudelft:oai:tudelft.nl:uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa 2025-01-16T20:23:57+00:00 Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods Bij de Vaate, I. (author) Martin, Ericka (author) Slobbe, D.C. (author) Naeije, M.C. (author) Verlaan, M. (author) 2022 http://resolver.tudelft.nl/uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa https://doi.org/10.1080/01490419.2022.2089412 en eng http://www.scopus.com/inward/record.url?scp=85133721554&partnerID=8YFLogxK Marine Geodesy: an international journal of ocean surveys, mapping and sensing--0149-0419--1ecf91d1-a328-4e46-b073-07cdd79948c0 http://resolver.tudelft.nl/uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa https://doi.org/10.1080/01490419.2022.2089412 © 2022 I. Bij de Vaate, Ericka Martin, D.C. Slobbe, M.C. Naeije, M. Verlaan Arctic Ocean classification lead detection machine learning Sentinel-3 synthetic aperture radar journal article 2022 fttudelft https://doi.org/10.1080/01490419.2022.2089412 2024-01-24T23:33:10Z In the Arctic Ocean, obtaining water levels from satellite altimetry is hampered by the presence of sea ice. Hence, water level retrieval requires accurate detection of fractures in the sea ice (leads). This paper describes a thorough assessment of various surface type classification methods, including a thresholding method, nine supervised-, and two unsupervised machine learning methods, applied to Sentinel-3 Synthetic Aperture Radar Altimeter data. For the first time, the simultaneously sensed images from the Ocean and Land Color Instrument, onboard Sentinel-3, were used for training and validation of the classifiers. This product allows to identify leads that are at least 300 meters wide. Applied to data from winter months, the supervised Adaptive Boosting, Artificial Neural Network, Naïve-Bayes, and Linear Discriminant classifiers showed robust results with overall accuracies of up to 92%. The unsupervised Kmedoids classifier produced excellent results with accuracies up to 92.74% and is an attractive classifier when ground truth data is limited. All classifiers perform poorly on summer data, rendering surface classifications that are solely based on altimetry data from summer months unsuitable. Finally, the Adaptive Boosting, Artificial Neural Network, and Bootstrap Aggregation classifiers obtain the highest accuracies when the altimetry observations include measurements from the open ocean. Physical and Space Geodesy Astrodynamics & Space Missions Mathematical Physics Article in Journal/Newspaper Arctic Arctic Ocean Sea ice Delft University of Technology: Institutional Repository Arctic Arctic Ocean Marine Geodesy 1 34
spellingShingle Arctic Ocean
classification
lead detection
machine learning
Sentinel-3
synthetic aperture radar
Bij de Vaate, I. (author)
Martin, Ericka (author)
Slobbe, D.C. (author)
Naeije, M.C. (author)
Verlaan, M. (author)
Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods
title Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods
title_full Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods
title_fullStr Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods
title_full_unstemmed Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods
title_short Lead Detection in the Arctic Ocean from Sentinel-3 Satellite Data: A Comprehensive Assessment of Thresholding and Machine Learning Classification Methods
title_sort lead detection in the arctic ocean from sentinel-3 satellite data: a comprehensive assessment of thresholding and machine learning classification methods
topic Arctic Ocean
classification
lead detection
machine learning
Sentinel-3
synthetic aperture radar
topic_facet Arctic Ocean
classification
lead detection
machine learning
Sentinel-3
synthetic aperture radar
url http://resolver.tudelft.nl/uuid:0f1ff6db-a877-4466-9d1e-0f2ca3fbf0aa
https://doi.org/10.1080/01490419.2022.2089412