Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ...
Although the finite element approach of the Ice-sheet and Sea-level System Model (ISSM) solves ice dynamics problems governed by Stokes equations quickly and accurately, such numerical modeling requires intensive computation on central processing units (CPU). In this study, we develop graph neural n...
Main Authors: | , |
---|---|
Format: | Report |
Language: | unknown |
Published: |
arXiv
2024
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2402.05291 https://arxiv.org/abs/2402.05291 |
id |
ftdatacite:10.48550/arxiv.2402.05291 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2402.05291 2024-03-31T07:53:19+00:00 Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... Rahnemoonfar, Maryam Koo, Younghyun 2024 https://dx.doi.org/10.48550/arxiv.2402.05291 https://arxiv.org/abs/2402.05291 unknown arXiv Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 Machine Learning cs.LG Computational Engineering, Finance, and Science cs.CE FOS Computer and information sciences article Preprint Article CreativeWork 2024 ftdatacite https://doi.org/10.48550/arxiv.2402.05291 2024-03-04T12:40:10Z Although the finite element approach of the Ice-sheet and Sea-level System Model (ISSM) solves ice dynamics problems governed by Stokes equations quickly and accurately, such numerical modeling requires intensive computation on central processing units (CPU). In this study, we develop graph neural networks (GNN) as fast surrogate models to preserve the finite element structure of ISSM. Using the 20-year transient simulations in the Pine Island Glacier (PIG), we train and test three GNNs: graph convolutional network (GCN), graph attention network (GAT), and equivariant graph convolutional network (EGCN). These GNNs reproduce ice thickness and velocity with better accuracy than the classic convolutional neural network (CNN) and multi-layer perception (MLP). In particular, GNNs successfully capture the ice mass loss and acceleration induced by higher basal melting rates in the PIG. When our GNN emulators are implemented on graphic processing units (GPUs), they show up to 50 times faster computational time than ... : 12 pages, 7 figures, 3 tables, Submitted to Nature Communications on Feb 7, 2024 ... Report Ice Sheet Pine Island Glacier DataCite Metadata Store (German National Library of Science and Technology) Pine Island Glacier ENVELOPE(-101.000,-101.000,-75.000,-75.000) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Machine Learning cs.LG Computational Engineering, Finance, and Science cs.CE FOS Computer and information sciences |
spellingShingle |
Machine Learning cs.LG Computational Engineering, Finance, and Science cs.CE FOS Computer and information sciences Rahnemoonfar, Maryam Koo, Younghyun Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... |
topic_facet |
Machine Learning cs.LG Computational Engineering, Finance, and Science cs.CE FOS Computer and information sciences |
description |
Although the finite element approach of the Ice-sheet and Sea-level System Model (ISSM) solves ice dynamics problems governed by Stokes equations quickly and accurately, such numerical modeling requires intensive computation on central processing units (CPU). In this study, we develop graph neural networks (GNN) as fast surrogate models to preserve the finite element structure of ISSM. Using the 20-year transient simulations in the Pine Island Glacier (PIG), we train and test three GNNs: graph convolutional network (GCN), graph attention network (GAT), and equivariant graph convolutional network (EGCN). These GNNs reproduce ice thickness and velocity with better accuracy than the classic convolutional neural network (CNN) and multi-layer perception (MLP). In particular, GNNs successfully capture the ice mass loss and acceleration induced by higher basal melting rates in the PIG. When our GNN emulators are implemented on graphic processing units (GPUs), they show up to 50 times faster computational time than ... : 12 pages, 7 figures, 3 tables, Submitted to Nature Communications on Feb 7, 2024 ... |
format |
Report |
author |
Rahnemoonfar, Maryam Koo, Younghyun |
author_facet |
Rahnemoonfar, Maryam Koo, Younghyun |
author_sort |
Rahnemoonfar, Maryam |
title |
Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... |
title_short |
Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... |
title_full |
Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... |
title_fullStr |
Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... |
title_full_unstemmed |
Graph Neural Networks as Fast and High-fidelity Emulators for Finite-Element Ice Sheet Modeling ... |
title_sort |
graph neural networks as fast and high-fidelity emulators for finite-element ice sheet modeling ... |
publisher |
arXiv |
publishDate |
2024 |
url |
https://dx.doi.org/10.48550/arxiv.2402.05291 https://arxiv.org/abs/2402.05291 |
long_lat |
ENVELOPE(-101.000,-101.000,-75.000,-75.000) |
geographic |
Pine Island Glacier |
geographic_facet |
Pine Island Glacier |
genre |
Ice Sheet Pine Island Glacier |
genre_facet |
Ice Sheet Pine Island Glacier |
op_rights |
Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 |
op_doi |
https://doi.org/10.48550/arxiv.2402.05291 |
_version_ |
1795032926789828608 |