A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season
This study proposes a purely data-driven model for the weekly prediction of daily sea ice concentration (SIC) of the pan-Arctic (90 N, 45 N, 180 E, 180 W) during the melting season. The model, SICNet, adopts an encoder-decoder framework with fully convolutional networks (FCNs) and can predict the SI...
Published in: | IEEE Transactions on Geoscience and Remote Sensing |
---|---|
Main Authors: | , , |
Format: | Report |
Language: | English |
Published: |
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
2022
|
Subjects: | |
Online Access: | http://ir.qdio.ac.cn/handle/337002/179153 https://doi.org/10.1109/TGRS.2022.3177600 |
id |
ftchinacasciocas:oai:ir.qdio.ac.cn:337002/179153 |
---|---|
record_format |
openpolar |
spelling |
ftchinacasciocas:oai:ir.qdio.ac.cn:337002/179153 2023-05-15T15:02:04+02:00 A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season Ren, Yibin Li, Xiaofeng Zhang, Wenhao 2022 http://ir.qdio.ac.cn/handle/337002/179153 https://doi.org/10.1109/TGRS.2022.3177600 英语 eng IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING http://ir.qdio.ac.cn/handle/337002/179153 doi:10.1109/TGRS.2022.3177600 Deep fully convolutional networks (FCNs) recursively predicting satellite-derived sea ice concentration (SIC) SIC prediction temporal-spatial attention Geochemistry & Geophysics Engineering Remote Sensing Imaging Science & Photographic Technology Electrical & Electronic CONVOLUTIONAL LONG VARIABILITY FORECASTS 期刊论文 2022 ftchinacasciocas https://doi.org/10.1109/TGRS.2022.3177600 2022-07-29T12:11:51Z This study proposes a purely data-driven model for the weekly prediction of daily sea ice concentration (SIC) of the pan-Arctic (90 N, 45 N, 180 E, 180 W) during the melting season. The model, SICNet, adopts an encoder-decoder framework with fully convolutional networks (FCNs) and can predict the SIC (covering 320 x 224 grids, each with a resolution of 25 km) one-week lead with high accuracy. We design a temporal-spatial attention module (TSAM) to help SICNet capture spatiotemporal dependencies from SIC sequences. The satellite-derived SIC data of 33 years (1988-2020) from the National Snow and Ice Data Center (NSIDC) are employed to train and test the model, 1988-2015 for training, and 2016-2020 for testing. SICNet achieves the mean absolute error (MAE) of 2.67%, the mean absolute percentage error (MAPE) of 8.67%, and the Nash-Sutcliffe efficiency (NSE) of 0.9784 in weekly predicting of SIC during the melting season. SICNet achieves better performance than existing deep-learning-based models. The TSAM reduced the MAE from 2.73% to 2.67%. We evaluate the model's performance by recursively predicting, from seven- to 28-day leads. We employ the binary accuracy (BACC) metric to measure the accuracy of the predicted sea ice extent (SIE) and compare SICNet with the anomaly persistence (Persist). SICNet shows better performance than Persist with an average BACC on the 28th day of 2016-2019 over 90% (90.17%). For the 28-day lead predictions of three extreme minimum SIE in September 2007, 2012, and 2020, SICNet outperforms Persist with an average improvement of 1.84% in BACC and 0.16 milkm(2) in the SIE error. Report Arctic National Snow and Ice Data Center Sea ice Institute of Oceanology, Chinese Academy of Sciences: IOCAS-IR Arctic Nash ENVELOPE(-62.350,-62.350,-74.233,-74.233) Sutcliffe ENVELOPE(-81.383,-81.383,50.683,50.683) IEEE Transactions on Geoscience and Remote Sensing 60 1 19 |
institution |
Open Polar |
collection |
Institute of Oceanology, Chinese Academy of Sciences: IOCAS-IR |
op_collection_id |
ftchinacasciocas |
language |
English |
topic |
Deep fully convolutional networks (FCNs) recursively predicting satellite-derived sea ice concentration (SIC) SIC prediction temporal-spatial attention Geochemistry & Geophysics Engineering Remote Sensing Imaging Science & Photographic Technology Electrical & Electronic CONVOLUTIONAL LONG VARIABILITY FORECASTS |
spellingShingle |
Deep fully convolutional networks (FCNs) recursively predicting satellite-derived sea ice concentration (SIC) SIC prediction temporal-spatial attention Geochemistry & Geophysics Engineering Remote Sensing Imaging Science & Photographic Technology Electrical & Electronic CONVOLUTIONAL LONG VARIABILITY FORECASTS Ren, Yibin Li, Xiaofeng Zhang, Wenhao A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season |
topic_facet |
Deep fully convolutional networks (FCNs) recursively predicting satellite-derived sea ice concentration (SIC) SIC prediction temporal-spatial attention Geochemistry & Geophysics Engineering Remote Sensing Imaging Science & Photographic Technology Electrical & Electronic CONVOLUTIONAL LONG VARIABILITY FORECASTS |
description |
This study proposes a purely data-driven model for the weekly prediction of daily sea ice concentration (SIC) of the pan-Arctic (90 N, 45 N, 180 E, 180 W) during the melting season. The model, SICNet, adopts an encoder-decoder framework with fully convolutional networks (FCNs) and can predict the SIC (covering 320 x 224 grids, each with a resolution of 25 km) one-week lead with high accuracy. We design a temporal-spatial attention module (TSAM) to help SICNet capture spatiotemporal dependencies from SIC sequences. The satellite-derived SIC data of 33 years (1988-2020) from the National Snow and Ice Data Center (NSIDC) are employed to train and test the model, 1988-2015 for training, and 2016-2020 for testing. SICNet achieves the mean absolute error (MAE) of 2.67%, the mean absolute percentage error (MAPE) of 8.67%, and the Nash-Sutcliffe efficiency (NSE) of 0.9784 in weekly predicting of SIC during the melting season. SICNet achieves better performance than existing deep-learning-based models. The TSAM reduced the MAE from 2.73% to 2.67%. We evaluate the model's performance by recursively predicting, from seven- to 28-day leads. We employ the binary accuracy (BACC) metric to measure the accuracy of the predicted sea ice extent (SIE) and compare SICNet with the anomaly persistence (Persist). SICNet shows better performance than Persist with an average BACC on the 28th day of 2016-2019 over 90% (90.17%). For the 28-day lead predictions of three extreme minimum SIE in September 2007, 2012, and 2020, SICNet outperforms Persist with an average improvement of 1.84% in BACC and 0.16 milkm(2) in the SIE error. |
format |
Report |
author |
Ren, Yibin Li, Xiaofeng Zhang, Wenhao |
author_facet |
Ren, Yibin Li, Xiaofeng Zhang, Wenhao |
author_sort |
Ren, Yibin |
title |
A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season |
title_short |
A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season |
title_full |
A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season |
title_fullStr |
A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season |
title_full_unstemmed |
A Data-Driven Deep Learning Model for Weekly Sea Ice Concentration Prediction of the Pan-Arctic During the Melting Season |
title_sort |
data-driven deep learning model for weekly sea ice concentration prediction of the pan-arctic during the melting season |
publisher |
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
publishDate |
2022 |
url |
http://ir.qdio.ac.cn/handle/337002/179153 https://doi.org/10.1109/TGRS.2022.3177600 |
long_lat |
ENVELOPE(-62.350,-62.350,-74.233,-74.233) ENVELOPE(-81.383,-81.383,50.683,50.683) |
geographic |
Arctic Nash Sutcliffe |
geographic_facet |
Arctic Nash Sutcliffe |
genre |
Arctic National Snow and Ice Data Center Sea ice |
genre_facet |
Arctic National Snow and Ice Data Center Sea ice |
op_relation |
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING http://ir.qdio.ac.cn/handle/337002/179153 doi:10.1109/TGRS.2022.3177600 |
op_doi |
https://doi.org/10.1109/TGRS.2022.3177600 |
container_title |
IEEE Transactions on Geoscience and Remote Sensing |
container_volume |
60 |
container_start_page |
1 |
op_container_end_page |
19 |
_version_ |
1766334056700575744 |