Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks
Within the rapidly changing Arctic region, accurate sea ice forecasts are of crucial importance for navigation activities, such as the planning of shipping routes. Numerical climate models have been widely used to generate Arctic sea ice forecasts at different time scales, but they are highly depend...
Published in: | Frontiers in Marine Science |
---|---|
Main Authors: | , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Frontiers Media S.A.
2022
|
Subjects: | |
Online Access: | https://doi.org/10.3389/fmars.2022.860403 https://doaj.org/article/8a58ffe94b8d4660b40fc185ded43d7a |
id |
ftdoajarticles:oai:doaj.org/article:8a58ffe94b8d4660b40fc185ded43d7a |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:8a58ffe94b8d4660b40fc185ded43d7a 2023-05-15T14:35:35+02:00 Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks Jianfen Wei Renlong Hang Jing-Jia Luo 2022-06-01T00:00:00Z https://doi.org/10.3389/fmars.2022.860403 https://doaj.org/article/8a58ffe94b8d4660b40fc185ded43d7a EN eng Frontiers Media S.A. https://www.frontiersin.org/articles/10.3389/fmars.2022.860403/full https://doaj.org/toc/2296-7745 2296-7745 doi:10.3389/fmars.2022.860403 https://doaj.org/article/8a58ffe94b8d4660b40fc185ded43d7a Frontiers in Marine Science, Vol 9 (2022) Arctic sea ice forecast deep learning attention-based LSTM Science Q General. Including nature conservation geographical distribution QH1-199.5 article 2022 ftdoajarticles https://doi.org/10.3389/fmars.2022.860403 2022-12-31T02:03:21Z Within the rapidly changing Arctic region, accurate sea ice forecasts are of crucial importance for navigation activities, such as the planning of shipping routes. Numerical climate models have been widely used to generate Arctic sea ice forecasts at different time scales, but they are highly dependent on the initial conditions and are computationally expensive. Recently, with the increasing availability of geoscience data and the advances in deep learning algorithms, the use of artificial intelligence (AI)-based sea ice prediction methods has gained significant attention. In this study, we propose a supervised deep learning approach, namely attention-based long short-term memory networks (LSTMs), to forecast pan-Arctic sea ice at monthly time scales. Our method makes use of historical sea ice concentration (SIC) observations during 1979–2020, from passive microwave brightness temperatures. Based on the persistence of SIC anomalies, which is known as one of the dominant sources of sea ice predictability, our approach exploits the temporal relationships of sea ice conditions across different time windows of the training period. We demonstrate that the attention-based LSTM is able to learn the variations of the Arctic sea ice and can skillfully forecast pan-Arctic SIC on monthly time scale. By designing the loss function and utilizing the attention mechanism, our approach generally improves the accuracy of sea ice forecasts compared to traditional LSTM networks. Moreover, it outperforms forecasts with the climatology and persistence based empirical models, as well as two dynamical models from the Copernicus Climate Change Service (C3S) datastore. This approach shows great promise in enhancing forecasts of Arctic sea ice using AI methods. Article in Journal/Newspaper Arctic Climate change Sea ice Directory of Open Access Journals: DOAJ Articles Arctic Frontiers in Marine Science 9 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
Arctic sea ice forecast deep learning attention-based LSTM Science Q General. Including nature conservation geographical distribution QH1-199.5 |
spellingShingle |
Arctic sea ice forecast deep learning attention-based LSTM Science Q General. Including nature conservation geographical distribution QH1-199.5 Jianfen Wei Renlong Hang Jing-Jia Luo Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks |
topic_facet |
Arctic sea ice forecast deep learning attention-based LSTM Science Q General. Including nature conservation geographical distribution QH1-199.5 |
description |
Within the rapidly changing Arctic region, accurate sea ice forecasts are of crucial importance for navigation activities, such as the planning of shipping routes. Numerical climate models have been widely used to generate Arctic sea ice forecasts at different time scales, but they are highly dependent on the initial conditions and are computationally expensive. Recently, with the increasing availability of geoscience data and the advances in deep learning algorithms, the use of artificial intelligence (AI)-based sea ice prediction methods has gained significant attention. In this study, we propose a supervised deep learning approach, namely attention-based long short-term memory networks (LSTMs), to forecast pan-Arctic sea ice at monthly time scales. Our method makes use of historical sea ice concentration (SIC) observations during 1979–2020, from passive microwave brightness temperatures. Based on the persistence of SIC anomalies, which is known as one of the dominant sources of sea ice predictability, our approach exploits the temporal relationships of sea ice conditions across different time windows of the training period. We demonstrate that the attention-based LSTM is able to learn the variations of the Arctic sea ice and can skillfully forecast pan-Arctic SIC on monthly time scale. By designing the loss function and utilizing the attention mechanism, our approach generally improves the accuracy of sea ice forecasts compared to traditional LSTM networks. Moreover, it outperforms forecasts with the climatology and persistence based empirical models, as well as two dynamical models from the Copernicus Climate Change Service (C3S) datastore. This approach shows great promise in enhancing forecasts of Arctic sea ice using AI methods. |
format |
Article in Journal/Newspaper |
author |
Jianfen Wei Renlong Hang Jing-Jia Luo |
author_facet |
Jianfen Wei Renlong Hang Jing-Jia Luo |
author_sort |
Jianfen Wei |
title |
Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks |
title_short |
Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks |
title_full |
Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks |
title_fullStr |
Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks |
title_full_unstemmed |
Prediction of Pan-Arctic Sea Ice Using Attention-Based LSTM Neural Networks |
title_sort |
prediction of pan-arctic sea ice using attention-based lstm neural networks |
publisher |
Frontiers Media S.A. |
publishDate |
2022 |
url |
https://doi.org/10.3389/fmars.2022.860403 https://doaj.org/article/8a58ffe94b8d4660b40fc185ded43d7a |
geographic |
Arctic |
geographic_facet |
Arctic |
genre |
Arctic Climate change Sea ice |
genre_facet |
Arctic Climate change Sea ice |
op_source |
Frontiers in Marine Science, Vol 9 (2022) |
op_relation |
https://www.frontiersin.org/articles/10.3389/fmars.2022.860403/full https://doaj.org/toc/2296-7745 2296-7745 doi:10.3389/fmars.2022.860403 https://doaj.org/article/8a58ffe94b8d4660b40fc185ded43d7a |
op_doi |
https://doi.org/10.3389/fmars.2022.860403 |
container_title |
Frontiers in Marine Science |
container_volume |
9 |
_version_ |
1766308379580432384 |