Deep learning speeds up ice flow modelling by several orders of magnitude
This paper introduces the Instructed Glacier Model (IGM) – a model that simulates ice dynamics, mass balance and its coupling to predict the evolution of glaciers, icefields or ice sheets. The novelty of IGM is that it models the ice flow by a Convolutional Neural Network, which is trained from data...
Main Authors: | , , , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
International Glaciological Society
2022
|
Subjects: | |
Online Access: | https://www.zora.uzh.ch/id/eprint/220916/ https://www.zora.uzh.ch/id/eprint/220916/1/ZORA_div_class_title_deep_learning_speeds_up_ice_flow_modelling_by_several_orders_of_magnitude_div.pdf https://doi.org/10.5167/uzh-220916 https://doi.org/10.1017/jog.2021.120 |
id |
ftunivzuerich:oai:www.zora.uzh.ch:220916 |
---|---|
record_format |
openpolar |
spelling |
ftunivzuerich:oai:www.zora.uzh.ch:220916 2024-06-23T07:54:16+00:00 Deep learning speeds up ice flow modelling by several orders of magnitude Jouvet, Guillaume Cordonnier, Guillaume Kim, Byungsoo Lüthi, Martin P Vieli, Andreas Aschwanden, Andy 2022-08-01 application/pdf https://www.zora.uzh.ch/id/eprint/220916/ https://www.zora.uzh.ch/id/eprint/220916/1/ZORA_div_class_title_deep_learning_speeds_up_ice_flow_modelling_by_several_orders_of_magnitude_div.pdf https://doi.org/10.5167/uzh-220916 https://doi.org/10.1017/jog.2021.120 eng eng International Glaciological Society https://www.zora.uzh.ch/id/eprint/220916/1/ZORA_div_class_title_deep_learning_speeds_up_ice_flow_modelling_by_several_orders_of_magnitude_div.pdf doi:10.5167/uzh-220916 doi:10.1017/jog.2021.120 urn:issn:0022-1430 info:eu-repo/semantics/openAccess Creative Commons: Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) http://creativecommons.org/licenses/by-nc-nd/4.0/ Jouvet, Guillaume; Cordonnier, Guillaume; Kim, Byungsoo; Lüthi, Martin P; Vieli, Andreas; Aschwanden, Andy (2022). Deep learning speeds up ice flow modelling by several orders of magnitude. Journal of Glaciology, 68(270):651-664. Institute of Geography 910 Geography & travel Earth-Surface Processes Journal Article PeerReviewed info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion 2022 ftunivzuerich https://doi.org/10.5167/uzh-22091610.1017/jog.2021.120 2024-05-29T01:13:38Z This paper introduces the Instructed Glacier Model (IGM) – a model that simulates ice dynamics, mass balance and its coupling to predict the evolution of glaciers, icefields or ice sheets. The novelty of IGM is that it models the ice flow by a Convolutional Neural Network, which is trained from data generated with hybrid SIA + SSA or Stokes ice flow models. By doing so, the most computationally demanding model component is substituted by a cheap emulator. Once trained with representative data, we demonstrate that IGM permits to model mountain glaciers up to 1000 × faster than Stokes ones on Central Processing Units (CPU) with fidelity levels above 90% in terms of ice flow solutions leading to nearly identical transient thickness evolution. Switching to the GPU often permits additional significant speed-ups, especially when emulating Stokes dynamics or/and modelling at high spatial resolution. IGM is an open-source Python code which deals with two-dimensional (2-D) gridded input and output data. Together with a companion library of trained ice flow emulators, IGM permits user-friendly, highly efficient and mechanically state-of-the-art glacier and icefields simulations. Article in Journal/Newspaper Journal of Glaciology University of Zurich (UZH): ZORA (Zurich Open Repository and Archive |
institution |
Open Polar |
collection |
University of Zurich (UZH): ZORA (Zurich Open Repository and Archive |
op_collection_id |
ftunivzuerich |
language |
English |
topic |
Institute of Geography 910 Geography & travel Earth-Surface Processes |
spellingShingle |
Institute of Geography 910 Geography & travel Earth-Surface Processes Jouvet, Guillaume Cordonnier, Guillaume Kim, Byungsoo Lüthi, Martin P Vieli, Andreas Aschwanden, Andy Deep learning speeds up ice flow modelling by several orders of magnitude |
topic_facet |
Institute of Geography 910 Geography & travel Earth-Surface Processes |
description |
This paper introduces the Instructed Glacier Model (IGM) – a model that simulates ice dynamics, mass balance and its coupling to predict the evolution of glaciers, icefields or ice sheets. The novelty of IGM is that it models the ice flow by a Convolutional Neural Network, which is trained from data generated with hybrid SIA + SSA or Stokes ice flow models. By doing so, the most computationally demanding model component is substituted by a cheap emulator. Once trained with representative data, we demonstrate that IGM permits to model mountain glaciers up to 1000 × faster than Stokes ones on Central Processing Units (CPU) with fidelity levels above 90% in terms of ice flow solutions leading to nearly identical transient thickness evolution. Switching to the GPU often permits additional significant speed-ups, especially when emulating Stokes dynamics or/and modelling at high spatial resolution. IGM is an open-source Python code which deals with two-dimensional (2-D) gridded input and output data. Together with a companion library of trained ice flow emulators, IGM permits user-friendly, highly efficient and mechanically state-of-the-art glacier and icefields simulations. |
format |
Article in Journal/Newspaper |
author |
Jouvet, Guillaume Cordonnier, Guillaume Kim, Byungsoo Lüthi, Martin P Vieli, Andreas Aschwanden, Andy |
author_facet |
Jouvet, Guillaume Cordonnier, Guillaume Kim, Byungsoo Lüthi, Martin P Vieli, Andreas Aschwanden, Andy |
author_sort |
Jouvet, Guillaume |
title |
Deep learning speeds up ice flow modelling by several orders of magnitude |
title_short |
Deep learning speeds up ice flow modelling by several orders of magnitude |
title_full |
Deep learning speeds up ice flow modelling by several orders of magnitude |
title_fullStr |
Deep learning speeds up ice flow modelling by several orders of magnitude |
title_full_unstemmed |
Deep learning speeds up ice flow modelling by several orders of magnitude |
title_sort |
deep learning speeds up ice flow modelling by several orders of magnitude |
publisher |
International Glaciological Society |
publishDate |
2022 |
url |
https://www.zora.uzh.ch/id/eprint/220916/ https://www.zora.uzh.ch/id/eprint/220916/1/ZORA_div_class_title_deep_learning_speeds_up_ice_flow_modelling_by_several_orders_of_magnitude_div.pdf https://doi.org/10.5167/uzh-220916 https://doi.org/10.1017/jog.2021.120 |
genre |
Journal of Glaciology |
genre_facet |
Journal of Glaciology |
op_source |
Jouvet, Guillaume; Cordonnier, Guillaume; Kim, Byungsoo; Lüthi, Martin P; Vieli, Andreas; Aschwanden, Andy (2022). Deep learning speeds up ice flow modelling by several orders of magnitude. Journal of Glaciology, 68(270):651-664. |
op_relation |
https://www.zora.uzh.ch/id/eprint/220916/1/ZORA_div_class_title_deep_learning_speeds_up_ice_flow_modelling_by_several_orders_of_magnitude_div.pdf doi:10.5167/uzh-220916 doi:10.1017/jog.2021.120 urn:issn:0022-1430 |
op_rights |
info:eu-repo/semantics/openAccess Creative Commons: Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) http://creativecommons.org/licenses/by-nc-nd/4.0/ |
op_doi |
https://doi.org/10.5167/uzh-22091610.1017/jog.2021.120 |
_version_ |
1802646377750069248 |