A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport

High-resolution spatiotemporal simulations effectively capture the complexities of atmospheric plume dispersion in complex terrain. However, their high computational cost makes them impractical for applications requiring rapid responses or iterative processes, such as optimization, uncertainty quant...

Full description

Saved in:
Bibliographic Details
Main Authors: M. Giselle Fernández-Godino, Wai Tong Chung, Akshay A. Gowardhan, Matthias Ihme, Qingkai Kong, Donald D. Lucas, Stephen C. Myers
Format: Article
Language:English
Published: KeAi Communications Co. Ltd. 2025-06-01
Series:Artificial Intelligence in Geosciences
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666544125000164
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849433436917334016
author M. Giselle Fernández-Godino
Wai Tong Chung
Akshay A. Gowardhan
Matthias Ihme
Qingkai Kong
Donald D. Lucas
Stephen C. Myers
author_facet M. Giselle Fernández-Godino
Wai Tong Chung
Akshay A. Gowardhan
Matthias Ihme
Qingkai Kong
Donald D. Lucas
Stephen C. Myers
author_sort M. Giselle Fernández-Godino
collection DOAJ
description High-resolution spatiotemporal simulations effectively capture the complexities of atmospheric plume dispersion in complex terrain. However, their high computational cost makes them impractical for applications requiring rapid responses or iterative processes, such as optimization, uncertainty quantification, or inverse modeling. To address this challenge, this work introduces the Dual-Stage Temporal Three-dimensional UNet Super-resolution (DST3D-UNet-SR) model, a highly efficient deep learning model for plume dispersion predictions. DST3D-UNet-SR is composed of two sequential modules: the temporal module (TM), which predicts the transient evolution of a plume in complex terrain from low-resolution temporal data, and the spatial refinement module (SRM), which subsequently enhances the spatial resolution of the TM predictions. We train DST3D-UNet-SR using a comprehensive dataset derived from high-resolution large eddy simulations (LES) of plume transport. We propose the DST3D-UNet-SR model to significantly accelerate LES of three-dimensional (3D) plume dispersion by three orders of magnitude. Additionally, the model demonstrates the ability to dynamically adapt to evolving conditions through the incorporation of new observational data, substantially improving prediction accuracy in high-concentration regions near the source.
format Article
id doaj-art-8ca6065bb3f648e8b5f720e6a7e28f43
institution Kabale University
issn 2666-5441
language English
publishDate 2025-06-01
publisher KeAi Communications Co. Ltd.
record_format Article
series Artificial Intelligence in Geosciences
spelling doaj-art-8ca6065bb3f648e8b5f720e6a7e28f432025-08-20T03:27:02ZengKeAi Communications Co. Ltd.Artificial Intelligence in Geosciences2666-54412025-06-016110012010.1016/j.aiig.2025.100120A staged deep learning approach to spatial refinement in 3D temporal atmospheric transportM. Giselle Fernández-Godino0Wai Tong Chung1Akshay A. Gowardhan2Matthias Ihme3Qingkai Kong4Donald D. Lucas5Stephen C. Myers6Lawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550, United States of America; Corresponding author.Stanford University, 440 Escondido Hall, Stanford, CA 94305, United States of AmericaLawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550, United States of AmericaStanford University, 440 Escondido Hall, Stanford, CA 94305, United States of AmericaLawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550, United States of AmericaLawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550, United States of AmericaLawrence Livermore National Laboratory, 7000 East Avenue, Livermore, CA 94550, United States of AmericaHigh-resolution spatiotemporal simulations effectively capture the complexities of atmospheric plume dispersion in complex terrain. However, their high computational cost makes them impractical for applications requiring rapid responses or iterative processes, such as optimization, uncertainty quantification, or inverse modeling. To address this challenge, this work introduces the Dual-Stage Temporal Three-dimensional UNet Super-resolution (DST3D-UNet-SR) model, a highly efficient deep learning model for plume dispersion predictions. DST3D-UNet-SR is composed of two sequential modules: the temporal module (TM), which predicts the transient evolution of a plume in complex terrain from low-resolution temporal data, and the spatial refinement module (SRM), which subsequently enhances the spatial resolution of the TM predictions. We train DST3D-UNet-SR using a comprehensive dataset derived from high-resolution large eddy simulations (LES) of plume transport. We propose the DST3D-UNet-SR model to significantly accelerate LES of three-dimensional (3D) plume dispersion by three orders of magnitude. Additionally, the model demonstrates the ability to dynamically adapt to evolving conditions through the incorporation of new observational data, substantially improving prediction accuracy in high-concentration regions near the source.http://www.sciencedirect.com/science/article/pii/S2666544125000164Atmospheric sciencesGeosciencesPlume transport3D temporal sequencesArtificial intelligenceCNN
spellingShingle M. Giselle Fernández-Godino
Wai Tong Chung
Akshay A. Gowardhan
Matthias Ihme
Qingkai Kong
Donald D. Lucas
Stephen C. Myers
A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport
Artificial Intelligence in Geosciences
Atmospheric sciences
Geosciences
Plume transport
3D temporal sequences
Artificial intelligence
CNN
title A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport
title_full A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport
title_fullStr A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport
title_full_unstemmed A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport
title_short A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport
title_sort staged deep learning approach to spatial refinement in 3d temporal atmospheric transport
topic Atmospheric sciences
Geosciences
Plume transport
3D temporal sequences
Artificial intelligence
CNN
url http://www.sciencedirect.com/science/article/pii/S2666544125000164
work_keys_str_mv AT mgisellefernandezgodino astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT waitongchung astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT akshayagowardhan astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT matthiasihme astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT qingkaikong astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT donalddlucas astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT stephencmyers astageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT mgisellefernandezgodino stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT waitongchung stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT akshayagowardhan stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT matthiasihme stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT qingkaikong stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT donalddlucas stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport
AT stephencmyers stageddeeplearningapproachtospatialrefinementin3dtemporalatmospherictransport