Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data
Abstract Classifying the state of the atmosphere into a finite number of large-scale circulation regimes is a popular way of investigating teleconnections, the predictability of severe weather events, and climate change. Here, we investigate a supervised machine learning approach based on deformable...
Published in: | Scientific Reports |
---|---|
Main Authors: | , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Nature Portfolio
2022
|
Subjects: | |
Online Access: | https://doi.org/10.1038/s41598-022-12167-8 https://doaj.org/article/5755b551cbce4e54b4067db86786c400 |
id |
ftdoajarticles:oai:doaj.org/article:5755b551cbce4e54b4067db86786c400 |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:5755b551cbce4e54b4067db86786c400 2023-05-15T17:33:06+02:00 Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data Andreas Holm Nielsen Alexandros Iosifidis Henrik Karstoft 2022-05-01T00:00:00Z https://doi.org/10.1038/s41598-022-12167-8 https://doaj.org/article/5755b551cbce4e54b4067db86786c400 EN eng Nature Portfolio https://doi.org/10.1038/s41598-022-12167-8 https://doaj.org/toc/2045-2322 doi:10.1038/s41598-022-12167-8 2045-2322 https://doaj.org/article/5755b551cbce4e54b4067db86786c400 Scientific Reports, Vol 12, Iss 1, Pp 1-12 (2022) Medicine R Science Q article 2022 ftdoajarticles https://doi.org/10.1038/s41598-022-12167-8 2022-12-30T21:49:09Z Abstract Classifying the state of the atmosphere into a finite number of large-scale circulation regimes is a popular way of investigating teleconnections, the predictability of severe weather events, and climate change. Here, we investigate a supervised machine learning approach based on deformable convolutional neural networks (deCNNs) and transfer learning to forecast the North Atlantic-European weather regimes during extended boreal winter for 1–15 days into the future. We apply state-of-the-art interpretation techniques from the machine learning literature to attribute particular regions of interest or potential teleconnections relevant for any given weather cluster prediction or regime transition. We demonstrate superior forecasting performance relative to several classical meteorological benchmarks, as well as logistic regression and random forests. Due to its wider field of view, we also observe deCNN achieving considerably better performance than regular convolutional neural networks at lead times beyond 5–6 days. Finally, we find transfer learning to be of paramount importance, similar to previous data-driven atmospheric forecasting studies. Article in Journal/Newspaper North Atlantic Directory of Open Access Journals: DOAJ Articles Scientific Reports 12 1 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
Medicine R Science Q |
spellingShingle |
Medicine R Science Q Andreas Holm Nielsen Alexandros Iosifidis Henrik Karstoft Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
topic_facet |
Medicine R Science Q |
description |
Abstract Classifying the state of the atmosphere into a finite number of large-scale circulation regimes is a popular way of investigating teleconnections, the predictability of severe weather events, and climate change. Here, we investigate a supervised machine learning approach based on deformable convolutional neural networks (deCNNs) and transfer learning to forecast the North Atlantic-European weather regimes during extended boreal winter for 1–15 days into the future. We apply state-of-the-art interpretation techniques from the machine learning literature to attribute particular regions of interest or potential teleconnections relevant for any given weather cluster prediction or regime transition. We demonstrate superior forecasting performance relative to several classical meteorological benchmarks, as well as logistic regression and random forests. Due to its wider field of view, we also observe deCNN achieving considerably better performance than regular convolutional neural networks at lead times beyond 5–6 days. Finally, we find transfer learning to be of paramount importance, similar to previous data-driven atmospheric forecasting studies. |
format |
Article in Journal/Newspaper |
author |
Andreas Holm Nielsen Alexandros Iosifidis Henrik Karstoft |
author_facet |
Andreas Holm Nielsen Alexandros Iosifidis Henrik Karstoft |
author_sort |
Andreas Holm Nielsen |
title |
Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
title_short |
Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
title_full |
Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
title_fullStr |
Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
title_full_unstemmed |
Forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
title_sort |
forecasting large-scale circulation regimes using deformable convolutional neural networks and global spatiotemporal climate data |
publisher |
Nature Portfolio |
publishDate |
2022 |
url |
https://doi.org/10.1038/s41598-022-12167-8 https://doaj.org/article/5755b551cbce4e54b4067db86786c400 |
genre |
North Atlantic |
genre_facet |
North Atlantic |
op_source |
Scientific Reports, Vol 12, Iss 1, Pp 1-12 (2022) |
op_relation |
https://doi.org/10.1038/s41598-022-12167-8 https://doaj.org/toc/2045-2322 doi:10.1038/s41598-022-12167-8 2045-2322 https://doaj.org/article/5755b551cbce4e54b4067db86786c400 |
op_doi |
https://doi.org/10.1038/s41598-022-12167-8 |
container_title |
Scientific Reports |
container_volume |
12 |
container_issue |
1 |
_version_ |
1766131488883998720 |