Self-Supervised Pre-Training for Precipitation Post-Processor

Obtaining a sufficient forecast lead time for local precipitation is essential in preventing hazardous weather events. Global warming-induced climate change increases the challenge of accurately predicting severe precipitation events, such as heavy rainfall. In this paper, we propose a deep learning...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: An, Sojung, Lee, Junha, Jang, Jiyeon, Na, Inchae, Park, Wooyeon, You, Sujeong
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator An, Sojung
Lee, Junha
Jang, Jiyeon
Na, Inchae
Park, Wooyeon
You, Sujeong
description Obtaining a sufficient forecast lead time for local precipitation is essential in preventing hazardous weather events. Global warming-induced climate change increases the challenge of accurately predicting severe precipitation events, such as heavy rainfall. In this paper, we propose a deep learning-based precipitation post-processor for numerical weather prediction (NWP) models. The precipitation post-processor consists of (i) employing self-supervised pre-training, where the parameters of the encoder are pre-trained on the reconstruction of the masked variables of the atmospheric physics domain; and (ii) conducting transfer learning on precipitation segmentation tasks (the target domain) from the pre-trained encoder. In addition, we introduced a heuristic labeling approach to effectively train class-imbalanced datasets. Our experiments on precipitation correction for regional NWP show that the proposed method outperforms other approaches.
doi_str_mv 10.48550/arxiv.2310.20187
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2310_20187</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2310_20187</sourcerecordid><originalsourceid>FETCH-LOGICAL-a677-95eb06e7b178e0e00dca4115319c77eb6a84abfd0d47b5facf1ddf019db190f23</originalsourceid><addsrcrecordid>eNotj71qwzAURrV0KGkfoFP9AkrvtS3LHjqU0D8I1BDv5kq6CoLEMpIb2rdvk3b64Awf5whxh7CuW6XggdJXOK3L6heUgK2-Fo87Pni5-5w5nUJmV_SJ5ZAoTGHaFz6mM7BhDgstIU5FH_Mi-xQt5xzTjbjydMh8-78rMbw8D5s3uf14fd88bSU1WstOsYGGtUHdMjCAs1Qjqgo7qzWbhtqajHfgam2UJ-vROQ_YOYMd-LJaifu_24v_OKdwpPQ9njvGS0f1A3KEQ70</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Self-Supervised Pre-Training for Precipitation Post-Processor</title><source>arXiv.org</source><creator>An, Sojung ; Lee, Junha ; Jang, Jiyeon ; Na, Inchae ; Park, Wooyeon ; You, Sujeong</creator><creatorcontrib>An, Sojung ; Lee, Junha ; Jang, Jiyeon ; Na, Inchae ; Park, Wooyeon ; You, Sujeong</creatorcontrib><description>Obtaining a sufficient forecast lead time for local precipitation is essential in preventing hazardous weather events. Global warming-induced climate change increases the challenge of accurately predicting severe precipitation events, such as heavy rainfall. In this paper, we propose a deep learning-based precipitation post-processor for numerical weather prediction (NWP) models. The precipitation post-processor consists of (i) employing self-supervised pre-training, where the parameters of the encoder are pre-trained on the reconstruction of the masked variables of the atmospheric physics domain; and (ii) conducting transfer learning on precipitation segmentation tasks (the target domain) from the pre-trained encoder. In addition, we introduced a heuristic labeling approach to effectively train class-imbalanced datasets. Our experiments on precipitation correction for regional NWP show that the proposed method outperforms other approaches.</description><identifier>DOI: 10.48550/arxiv.2310.20187</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Learning</subject><creationdate>2023-10</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2310.20187$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2310.20187$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>An, Sojung</creatorcontrib><creatorcontrib>Lee, Junha</creatorcontrib><creatorcontrib>Jang, Jiyeon</creatorcontrib><creatorcontrib>Na, Inchae</creatorcontrib><creatorcontrib>Park, Wooyeon</creatorcontrib><creatorcontrib>You, Sujeong</creatorcontrib><title>Self-Supervised Pre-Training for Precipitation Post-Processor</title><description>Obtaining a sufficient forecast lead time for local precipitation is essential in preventing hazardous weather events. Global warming-induced climate change increases the challenge of accurately predicting severe precipitation events, such as heavy rainfall. In this paper, we propose a deep learning-based precipitation post-processor for numerical weather prediction (NWP) models. The precipitation post-processor consists of (i) employing self-supervised pre-training, where the parameters of the encoder are pre-trained on the reconstruction of the masked variables of the atmospheric physics domain; and (ii) conducting transfer learning on precipitation segmentation tasks (the target domain) from the pre-trained encoder. In addition, we introduced a heuristic labeling approach to effectively train class-imbalanced datasets. Our experiments on precipitation correction for regional NWP show that the proposed method outperforms other approaches.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj71qwzAURrV0KGkfoFP9AkrvtS3LHjqU0D8I1BDv5kq6CoLEMpIb2rdvk3b64Awf5whxh7CuW6XggdJXOK3L6heUgK2-Fo87Pni5-5w5nUJmV_SJ5ZAoTGHaFz6mM7BhDgstIU5FH_Mi-xQt5xzTjbjydMh8-78rMbw8D5s3uf14fd88bSU1WstOsYGGtUHdMjCAs1Qjqgo7qzWbhtqajHfgam2UJ-vROQ_YOYMd-LJaifu_24v_OKdwpPQ9njvGS0f1A3KEQ70</recordid><startdate>20231031</startdate><enddate>20231031</enddate><creator>An, Sojung</creator><creator>Lee, Junha</creator><creator>Jang, Jiyeon</creator><creator>Na, Inchae</creator><creator>Park, Wooyeon</creator><creator>You, Sujeong</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20231031</creationdate><title>Self-Supervised Pre-Training for Precipitation Post-Processor</title><author>An, Sojung ; Lee, Junha ; Jang, Jiyeon ; Na, Inchae ; Park, Wooyeon ; You, Sujeong</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a677-95eb06e7b178e0e00dca4115319c77eb6a84abfd0d47b5facf1ddf019db190f23</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>An, Sojung</creatorcontrib><creatorcontrib>Lee, Junha</creatorcontrib><creatorcontrib>Jang, Jiyeon</creatorcontrib><creatorcontrib>Na, Inchae</creatorcontrib><creatorcontrib>Park, Wooyeon</creatorcontrib><creatorcontrib>You, Sujeong</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>An, Sojung</au><au>Lee, Junha</au><au>Jang, Jiyeon</au><au>Na, Inchae</au><au>Park, Wooyeon</au><au>You, Sujeong</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Self-Supervised Pre-Training for Precipitation Post-Processor</atitle><date>2023-10-31</date><risdate>2023</risdate><abstract>Obtaining a sufficient forecast lead time for local precipitation is essential in preventing hazardous weather events. Global warming-induced climate change increases the challenge of accurately predicting severe precipitation events, such as heavy rainfall. In this paper, we propose a deep learning-based precipitation post-processor for numerical weather prediction (NWP) models. The precipitation post-processor consists of (i) employing self-supervised pre-training, where the parameters of the encoder are pre-trained on the reconstruction of the masked variables of the atmospheric physics domain; and (ii) conducting transfer learning on precipitation segmentation tasks (the target domain) from the pre-trained encoder. In addition, we introduced a heuristic labeling approach to effectively train class-imbalanced datasets. Our experiments on precipitation correction for regional NWP show that the proposed method outperforms other approaches.</abstract><doi>10.48550/arxiv.2310.20187</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2310.20187
ispartof
issn
language eng
recordid cdi_arxiv_primary_2310_20187
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Learning
title Self-Supervised Pre-Training for Precipitation Post-Processor
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-17T20%3A32%3A48IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Self-Supervised%20Pre-Training%20for%20Precipitation%20Post-Processor&rft.au=An,%20Sojung&rft.date=2023-10-31&rft_id=info:doi/10.48550/arxiv.2310.20187&rft_dat=%3Carxiv_GOX%3E2310_20187%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true