Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks.
Operational Arctic sea ice forecasts are of crucial importance to commercial and scientific activities in the Arctic region. Currently, numerical climate models, including General Circulation Models (GCMs) and regional climate models, are widely used to generate the Arctic sea ice predictions at wea...
Main Authors: | , , , |
---|---|
Format: | Lecture |
Language: | English |
Published: |
Zenodo
2020
|
Subjects: | |
Online Access: | https://doi.org/10.5281/zenodo.3899798 |
id |
ftzenodo:oai:zenodo.org:3899798 |
---|---|
record_format |
openpolar |
spelling |
ftzenodo:oai:zenodo.org:3899798 2024-09-15T17:57:59+00:00 Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. Yang Liu Laurens Bogaardt Jisk Attema Wilco Hazeleger 2020-05-06 https://doi.org/10.5281/zenodo.3899798 eng eng Zenodo https://zenodo.org/communities/eu https://zenodo.org/communities/blue-actionh2020 https://doi.org/10.5281/zenodo.3899797 https://doi.org/10.5281/zenodo.3899798 oai:zenodo.org:3899798 info:eu-repo/semantics/openAccess Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode info:eu-repo/semantics/lecture 2020 ftzenodo https://doi.org/10.5281/zenodo.389979810.5281/zenodo.3899797 2024-07-25T16:37:34Z Operational Arctic sea ice forecasts are of crucial importance to commercial and scientific activities in the Arctic region. Currently, numerical climate models, including General Circulation Models (GCMs) and regional climate models, are widely used to generate the Arctic sea ice predictions at weather time-scales. However, these numerical climate models require near real-time input of weather conditions to assure the quality of the predictions and these are hard to obtain and the simulations are computationally expensive. In this study, we propose a deep learning approach to forecasts of sea ice in the Barents sea at weather time scales. To work with such spatial-temporal sequence problems, Convolutional Long Short Term Memory Networks (ConvLSTM) are useful. ConvLSTM are LSTM (Long-Short Term Memory) networks with convolutional cells embedded in the LSTM cells. This approach is unsupervised learning and it can make use of enormous amounts of historical records of weather and climate. With input fields from atmospheric (ERA-Interim) and oceanic (ORAS4) reanalysis data sets, we demonstrate that the ConvLSTM is able to learn the variability of the Arctic sea ice within historical records and effectively predict regional sea ice concentration patterns at weekly to monthly time scales. Based on the known sources of predictability, sensitivity tests with different climate fields were also performed. The influences of different predictors on the quality of predictions are evaluated. This method outperforms predictions with climatology and persistence and is promising to act as a fast and cost-efficient operational sea ice forecast system in the future. EGU 2020 Session ITS4.3/AS5.2 Lecture Barents Sea Sea ice Zenodo |
institution |
Open Polar |
collection |
Zenodo |
op_collection_id |
ftzenodo |
language |
English |
description |
Operational Arctic sea ice forecasts are of crucial importance to commercial and scientific activities in the Arctic region. Currently, numerical climate models, including General Circulation Models (GCMs) and regional climate models, are widely used to generate the Arctic sea ice predictions at weather time-scales. However, these numerical climate models require near real-time input of weather conditions to assure the quality of the predictions and these are hard to obtain and the simulations are computationally expensive. In this study, we propose a deep learning approach to forecasts of sea ice in the Barents sea at weather time scales. To work with such spatial-temporal sequence problems, Convolutional Long Short Term Memory Networks (ConvLSTM) are useful. ConvLSTM are LSTM (Long-Short Term Memory) networks with convolutional cells embedded in the LSTM cells. This approach is unsupervised learning and it can make use of enormous amounts of historical records of weather and climate. With input fields from atmospheric (ERA-Interim) and oceanic (ORAS4) reanalysis data sets, we demonstrate that the ConvLSTM is able to learn the variability of the Arctic sea ice within historical records and effectively predict regional sea ice concentration patterns at weekly to monthly time scales. Based on the known sources of predictability, sensitivity tests with different climate fields were also performed. The influences of different predictors on the quality of predictions are evaluated. This method outperforms predictions with climatology and persistence and is promising to act as a fast and cost-efficient operational sea ice forecast system in the future. EGU 2020 Session ITS4.3/AS5.2 |
format |
Lecture |
author |
Yang Liu Laurens Bogaardt Jisk Attema Wilco Hazeleger |
spellingShingle |
Yang Liu Laurens Bogaardt Jisk Attema Wilco Hazeleger Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. |
author_facet |
Yang Liu Laurens Bogaardt Jisk Attema Wilco Hazeleger |
author_sort |
Yang Liu |
title |
Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. |
title_short |
Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. |
title_full |
Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. |
title_fullStr |
Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. |
title_full_unstemmed |
Extended Range Arctic Sea Ice Forecast with Convolutional Long-Short Term Memory Networks. |
title_sort |
extended range arctic sea ice forecast with convolutional long-short term memory networks. |
publisher |
Zenodo |
publishDate |
2020 |
url |
https://doi.org/10.5281/zenodo.3899798 |
genre |
Barents Sea Sea ice |
genre_facet |
Barents Sea Sea ice |
op_relation |
https://zenodo.org/communities/eu https://zenodo.org/communities/blue-actionh2020 https://doi.org/10.5281/zenodo.3899797 https://doi.org/10.5281/zenodo.3899798 oai:zenodo.org:3899798 |
op_rights |
info:eu-repo/semantics/openAccess Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode |
op_doi |
https://doi.org/10.5281/zenodo.389979810.5281/zenodo.3899797 |
_version_ |
1810434197624455168 |