Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods
Supraglacial lakes in Greenland are highly dynamic hydrological features in which glacial meltwater cumulates, allowing for the loss and transport of freshwater from a glacial surface to the ocean or a nearby waterbody. Standard supraglacial lake monitoring techniques, specifically image segmentatio...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2023
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs15174360 |
id |
ftmdpi:oai:mdpi.com:/2072-4292/15/17/4360/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2072-4292/15/17/4360/ 2023-10-09T21:51:45+02:00 Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods Katrina Lutz Zahra Bahrami Matthias Braun agris 2023-09-04 application/pdf https://doi.org/10.3390/rs15174360 eng eng Multidisciplinary Digital Publishing Institute https://dx.doi.org/10.3390/rs15174360 https://creativecommons.org/licenses/by/4.0/ Remote Sensing Volume 15 Issue 17 Pages: 4360 meltwater supraglacial lakes remote sensing Sentinel-2 deep learning U-Net Greenland Zachariæ Isstrøm Nioghalvfjerdsbræ Text 2023 ftmdpi https://doi.org/10.3390/rs15174360 2023-09-10T23:54:35Z Supraglacial lakes in Greenland are highly dynamic hydrological features in which glacial meltwater cumulates, allowing for the loss and transport of freshwater from a glacial surface to the ocean or a nearby waterbody. Standard supraglacial lake monitoring techniques, specifically image segmentation, rely heavily on a series of region-dependent thresholds, limiting the adaptability of the algorithm to different illumination and surface variations, while being susceptible to the inclusion of false positives such as shadows. In this study, a supraglacial lake segmentation algorithm is developed for Sentinel-2 images based on a deep learning architecture (U-Net) to evaluate the suitability of artificial intelligence techniques in this domain. Additionally, a deep learning-based cloud segmentation tool developed specifically for polar regions is implemented in the processing chain to remove cloudy imagery from the analysis. Using this technique, a time series of supraglacial lake development is created for the 2016 to 2022 melt seasons over Nioghalvfjerdsbræ (79°N Glacier) and Zachariæ Isstrøm in Northeast Greenland, an area that covers 26,302 km2 and represents roughly 10% of the Northeast Greenland Ice Stream. The total lake area was found to have a strong interannual variability, with the largest peak lake area of 380 km2 in 2019 and the smallest peak lake area of 67 km2 in 2018. These results were then compared against an algorithm based on a thresholding technique to evaluate the agreement of the methodologies. The deep learning-based time series shows a similar trend to that produced by a previously published thresholding technique, while being smoother and more encompassing of meltwater in higher-melt periods. Additionally, while not completely eliminating them, the deep learning model significantly reduces the inclusion of shadows as false positives. Overall, the use of deep learning on multispectral images for the purpose of supraglacial lake segmentation proves to be advantageous. Text glacier Greenland MDPI Open Access Publishing Greenland Peak Lake ENVELOPE(-106.984,-106.984,56.250,56.250) Remote Sensing 15 17 4360 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
meltwater supraglacial lakes remote sensing Sentinel-2 deep learning U-Net Greenland Zachariæ Isstrøm Nioghalvfjerdsbræ |
spellingShingle |
meltwater supraglacial lakes remote sensing Sentinel-2 deep learning U-Net Greenland Zachariæ Isstrøm Nioghalvfjerdsbræ Katrina Lutz Zahra Bahrami Matthias Braun Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods |
topic_facet |
meltwater supraglacial lakes remote sensing Sentinel-2 deep learning U-Net Greenland Zachariæ Isstrøm Nioghalvfjerdsbræ |
description |
Supraglacial lakes in Greenland are highly dynamic hydrological features in which glacial meltwater cumulates, allowing for the loss and transport of freshwater from a glacial surface to the ocean or a nearby waterbody. Standard supraglacial lake monitoring techniques, specifically image segmentation, rely heavily on a series of region-dependent thresholds, limiting the adaptability of the algorithm to different illumination and surface variations, while being susceptible to the inclusion of false positives such as shadows. In this study, a supraglacial lake segmentation algorithm is developed for Sentinel-2 images based on a deep learning architecture (U-Net) to evaluate the suitability of artificial intelligence techniques in this domain. Additionally, a deep learning-based cloud segmentation tool developed specifically for polar regions is implemented in the processing chain to remove cloudy imagery from the analysis. Using this technique, a time series of supraglacial lake development is created for the 2016 to 2022 melt seasons over Nioghalvfjerdsbræ (79°N Glacier) and Zachariæ Isstrøm in Northeast Greenland, an area that covers 26,302 km2 and represents roughly 10% of the Northeast Greenland Ice Stream. The total lake area was found to have a strong interannual variability, with the largest peak lake area of 380 km2 in 2019 and the smallest peak lake area of 67 km2 in 2018. These results were then compared against an algorithm based on a thresholding technique to evaluate the agreement of the methodologies. The deep learning-based time series shows a similar trend to that produced by a previously published thresholding technique, while being smoother and more encompassing of meltwater in higher-melt periods. Additionally, while not completely eliminating them, the deep learning model significantly reduces the inclusion of shadows as false positives. Overall, the use of deep learning on multispectral images for the purpose of supraglacial lake segmentation proves to be advantageous. |
format |
Text |
author |
Katrina Lutz Zahra Bahrami Matthias Braun |
author_facet |
Katrina Lutz Zahra Bahrami Matthias Braun |
author_sort |
Katrina Lutz |
title |
Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods |
title_short |
Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods |
title_full |
Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods |
title_fullStr |
Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods |
title_full_unstemmed |
Supraglacial Lake Evolution over Northeast Greenland Using Deep Learning Methods |
title_sort |
supraglacial lake evolution over northeast greenland using deep learning methods |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2023 |
url |
https://doi.org/10.3390/rs15174360 |
op_coverage |
agris |
long_lat |
ENVELOPE(-106.984,-106.984,56.250,56.250) |
geographic |
Greenland Peak Lake |
geographic_facet |
Greenland Peak Lake |
genre |
glacier Greenland |
genre_facet |
glacier Greenland |
op_source |
Remote Sensing Volume 15 Issue 17 Pages: 4360 |
op_relation |
https://dx.doi.org/10.3390/rs15174360 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/rs15174360 |
container_title |
Remote Sensing |
container_volume |
15 |
container_issue |
17 |
container_start_page |
4360 |
_version_ |
1779314859512430592 |