No Cover Image

Journal article 865 views 140 downloads

Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks

Ahmed Bouteska Orcid Logo, Marco Lavazza Seranto, Petr Hajek, Mohammad Abedin Orcid Logo

Annals of Operations Research, Volume: 355, Issue: 2, Pages: 1587 - 1605

Swansea University Author: Mohammad Abedin Orcid Logo

  • 65186.VOR.pdf

    PDF | Version of Record

    This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.

    Download (1.96MB)

Check full text

DOI (Published version): 10.1007/s10479-023-05722-7

Abstract

Recent trends in global climate modeling, coupled with the availability of more fine-scale datasets, have opened up opportunities for deep learning-based climate prediction to improve the accuracy of predictions over traditional physics-based models. For this, however, large ensembles of data are ne...

Full description

Published in: Annals of Operations Research
ISSN: 0254-5330 1572-9338
Published: Springer Nature 2025
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa65186
first_indexed 2023-12-02T10:37:35Z
last_indexed 2026-01-23T04:13:09Z
id cronfa65186
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2026-01-21T10:35:27.8354485</datestamp><bib-version>v2</bib-version><id>65186</id><entry>2023-12-02</entry><title>Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks</title><swanseaauthors><author><sid>4ed8c020eae0c9bec4f5d9495d86d415</sid><ORCID>0000-0002-4688-0619</ORCID><firstname>Mohammad</firstname><surname>Abedin</surname><name>Mohammad Abedin</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2023-12-02</date><deptcode>CBAE</deptcode><abstract>Recent trends in global climate modeling, coupled with the availability of more fine-scale datasets, have opened up opportunities for deep learning-based climate prediction to improve the accuracy of predictions over traditional physics-based models. For this, however, large ensembles of data are needed. Generative models have recently proven to be a suitable solution to this problem. For a sound generative model for time-series forecasting, it is essential that temporal dynamics are preserved in that the generated data obey the original data distributions over time. Existing forecasting methods aided by generative models are not adequate for capturing such temporal relationships. Recently, generative models have been proposed that generate realistic time-series data by exploiting the combinations of unsupervised and supervised learning. However, these models suffer from instable learning and mode collapse problems. To overcome these issues, here we propose Wasserstein Time-Series Generative Adversarial Network (WTGAN), a new forecasting model that effectively imitates the dynamics of the original data by generating realistic synthetic time-series data. To validate the proposed forecasting model, we evaluate it by backtesting the challenging decadal climate forecasting problem. We show that the proposed forecasting model outperforms state-of-the- art generative models. Another advantage of the proposed model is that once WTGAN is tuned, generating time-series data is very fast, whereas standard simulators consume considerable computer time. Thus, a large amount of climate data can be generated, which can substantially improve existing data-driven climate forecasting models.</abstract><type>Journal Article</type><journal>Annals of Operations Research</journal><volume>355</volume><journalNumber>2</journalNumber><paginationStart>1587</paginationStart><paginationEnd>1605</paginationEnd><publisher>Springer Nature</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>0254-5330</issnPrint><issnElectronic>1572-9338</issnElectronic><keywords>Forecasting; Climate; Deep learning; Time series; Generative adversarial learning</keywords><publishedDay>1</publishedDay><publishedMonth>12</publishedMonth><publishedYear>2025</publishedYear><publishedDate>2025-12-01</publishedDate><doi>10.1007/s10479-023-05722-7</doi><url/><notes/><college>COLLEGE NANME</college><department>Management School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>CBAE</DepartmentCode><institution>Swansea University</institution><apcterm>SU Library paid the OA fee (TA Institutional Deal)</apcterm><funders>Swansea University</funders><projectreference/><lastEdited>2026-01-21T10:35:27.8354485</lastEdited><Created>2023-12-02T10:36:19.5942091</Created><path><level id="1">Faculty of Humanities and Social Sciences</level><level id="2">School of Management - Accounting and Finance</level></path><authors><author><firstname>Ahmed</firstname><surname>Bouteska</surname><orcid>0000-0002-5710-501x</orcid><order>1</order></author><author><firstname>Marco Lavazza</firstname><surname>Seranto</surname><order>2</order></author><author><firstname>Petr</firstname><surname>Hajek</surname><order>3</order></author><author><firstname>Mohammad</firstname><surname>Abedin</surname><orcid>0000-0002-4688-0619</orcid><order>4</order></author></authors><documents><document><filename>65186__29840__c7b5f52ad79a44eeb9b696c27161091a.pdf</filename><originalFilename>65186.VOR.pdf</originalFilename><uploaded>2024-03-25T15:19:10.7917287</uploaded><type>Output</type><contentLength>2054640</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>http://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling 2026-01-21T10:35:27.8354485 v2 65186 2023-12-02 Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks 4ed8c020eae0c9bec4f5d9495d86d415 0000-0002-4688-0619 Mohammad Abedin Mohammad Abedin true false 2023-12-02 CBAE Recent trends in global climate modeling, coupled with the availability of more fine-scale datasets, have opened up opportunities for deep learning-based climate prediction to improve the accuracy of predictions over traditional physics-based models. For this, however, large ensembles of data are needed. Generative models have recently proven to be a suitable solution to this problem. For a sound generative model for time-series forecasting, it is essential that temporal dynamics are preserved in that the generated data obey the original data distributions over time. Existing forecasting methods aided by generative models are not adequate for capturing such temporal relationships. Recently, generative models have been proposed that generate realistic time-series data by exploiting the combinations of unsupervised and supervised learning. However, these models suffer from instable learning and mode collapse problems. To overcome these issues, here we propose Wasserstein Time-Series Generative Adversarial Network (WTGAN), a new forecasting model that effectively imitates the dynamics of the original data by generating realistic synthetic time-series data. To validate the proposed forecasting model, we evaluate it by backtesting the challenging decadal climate forecasting problem. We show that the proposed forecasting model outperforms state-of-the- art generative models. Another advantage of the proposed model is that once WTGAN is tuned, generating time-series data is very fast, whereas standard simulators consume considerable computer time. Thus, a large amount of climate data can be generated, which can substantially improve existing data-driven climate forecasting models. Journal Article Annals of Operations Research 355 2 1587 1605 Springer Nature 0254-5330 1572-9338 Forecasting; Climate; Deep learning; Time series; Generative adversarial learning 1 12 2025 2025-12-01 10.1007/s10479-023-05722-7 COLLEGE NANME Management School COLLEGE CODE CBAE Swansea University SU Library paid the OA fee (TA Institutional Deal) Swansea University 2026-01-21T10:35:27.8354485 2023-12-02T10:36:19.5942091 Faculty of Humanities and Social Sciences School of Management - Accounting and Finance Ahmed Bouteska 0000-0002-5710-501x 1 Marco Lavazza Seranto 2 Petr Hajek 3 Mohammad Abedin 0000-0002-4688-0619 4 65186__29840__c7b5f52ad79a44eeb9b696c27161091a.pdf 65186.VOR.pdf 2024-03-25T15:19:10.7917287 Output 2054640 application/pdf Version of Record true This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. true eng http://creativecommons.org/licenses/by/4.0/
title Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
spellingShingle Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
Mohammad Abedin
title_short Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
title_full Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
title_fullStr Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
title_full_unstemmed Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
title_sort Data-driven decadal climate forecasting using Wasserstein time-series generative adversarial networks
author_id_str_mv 4ed8c020eae0c9bec4f5d9495d86d415
author_id_fullname_str_mv 4ed8c020eae0c9bec4f5d9495d86d415_***_Mohammad Abedin
author Mohammad Abedin
author2 Ahmed Bouteska
Marco Lavazza Seranto
Petr Hajek
Mohammad Abedin
format Journal article
container_title Annals of Operations Research
container_volume 355
container_issue 2
container_start_page 1587
publishDate 2025
institution Swansea University
issn 0254-5330
1572-9338
doi_str_mv 10.1007/s10479-023-05722-7
publisher Springer Nature
college_str Faculty of Humanities and Social Sciences
hierarchytype
hierarchy_top_id facultyofhumanitiesandsocialsciences
hierarchy_top_title Faculty of Humanities and Social Sciences
hierarchy_parent_id facultyofhumanitiesandsocialsciences
hierarchy_parent_title Faculty of Humanities and Social Sciences
department_str School of Management - Accounting and Finance{{{_:::_}}}Faculty of Humanities and Social Sciences{{{_:::_}}}School of Management - Accounting and Finance
document_store_str 1
active_str 0
description Recent trends in global climate modeling, coupled with the availability of more fine-scale datasets, have opened up opportunities for deep learning-based climate prediction to improve the accuracy of predictions over traditional physics-based models. For this, however, large ensembles of data are needed. Generative models have recently proven to be a suitable solution to this problem. For a sound generative model for time-series forecasting, it is essential that temporal dynamics are preserved in that the generated data obey the original data distributions over time. Existing forecasting methods aided by generative models are not adequate for capturing such temporal relationships. Recently, generative models have been proposed that generate realistic time-series data by exploiting the combinations of unsupervised and supervised learning. However, these models suffer from instable learning and mode collapse problems. To overcome these issues, here we propose Wasserstein Time-Series Generative Adversarial Network (WTGAN), a new forecasting model that effectively imitates the dynamics of the original data by generating realistic synthetic time-series data. To validate the proposed forecasting model, we evaluate it by backtesting the challenging decadal climate forecasting problem. We show that the proposed forecasting model outperforms state-of-the- art generative models. Another advantage of the proposed model is that once WTGAN is tuned, generating time-series data is very fast, whereas standard simulators consume considerable computer time. Thus, a large amount of climate data can be generated, which can substantially improve existing data-driven climate forecasting models.
published_date 2025-12-01T05:17:14Z
_version_ 1856804785498357760
score 11.096047

Similar Items