Climate Research Domain BERTs: Pretraining, Adaptation, and Evaluation

Climate Research Domain BERTs: Pretraining, Adaptation, and Evaluation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Climate Research Domain BERTs: Pretraining, Adaptation, and Evaluation Andrija Poleksić, Sanda Martinčić-Ipšić This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6644722/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Motivated by the pressing issue of climate change and the growing volume of data, we pretrain three new language models using climate change research papers published in top-tier journals. Adaptation of existing domain-specific models is utilized for CliSciBERT and SciClimateBERT and pretraining from scratch resulted in CliReBERT (Climate Research BERT). The performance assessment is performed on the climate change NLP benchmark ClimaBench. We evaluate SciBERT, ClimateBERT, BERT, RoBERTa and DistilRoBERTa - along with our new models - CliReBERT, CliSciBERT and SciClimateBERT - using five different random seeds on all seven ClimaBench datasets. CliReBERT achieves the highest overall performance with a macro-averaged F1 score of 65.45%, and outperforms all other models on three of the seven tasks. Additionally, CliReBERT demonstrates the most stable fine-tuning behavior, yielding the lowest average standard deviation across seeds (0.0118). The 5-fold stratified cross-validation on the SciDCC dataset showed that CliReBERT achieved the highest overall macro-average F1 score (53.75%), slightly outperforming RoBERTa and DistilRoBERTa, while the domain-adapted models underperformed their base counterparts. The superior performance of CliReBERT is accompanied by the lowest tokenizer fertility, suggesting appropriateness to model domain-specific vocabulary. domain-specific BERT model domain-adaptive pretraining masked language modelling Climate Change ClimaBench text classification Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 28 Jul, 2025 Reviews received at journal 24 Jul, 2025 Reviews received at journal 20 Jul, 2025 Reviewers agreed at journal 18 Jul, 2025 Reviewers agreed at journal 07 Jul, 2025 Reviewers invited by journal 14 May, 2025 Editor assigned by journal 13 May, 2025 Submission checks completed at journal 13 May, 2025 First submitted to journal 12 May, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6644722","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":457904645,"identity":"6117e4f8-f7c7-46c8-b26e-6b8af2e78cb8","order_by":0,"name":"Andrija Poleksić","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABBElEQVRIiWNgGAWjYHACA4YEHhDN3MDYwCDBwA9kHgAJEqGFEaJFsoEYLRAA1gLkHoBwcWrhn9287cMDGTsGc/bGxocz/ljIG9/IPXi4ooAhz7wBuxaJO8eKZyTwJDNY9hxsNtzYJmG47UZewsEzBgzFMgdwWHMjxxjoF2YGgxuJbZIPGyQSzG7kGBxsMGBInIFDhzxESz1IS/vPB38kEoxnENBiANFyGGwL4wY2iQQDCQJaDG+kFQO1HOcB+UVyJtAvM868AWmRwKlF7kbyZsafPdVy5uzNBz/2/KmT52/PMf7Y8McGpxYwYOxh4DFAE5PApwEIfiAidBSMglEwCkYBBgAAq79dDWCpBdIAAAAASUVORK5CYII=","orcid":"","institution":"University of Rijeka","correspondingAuthor":true,"prefix":"","firstName":"Andrija","middleName":"","lastName":"Poleksić","suffix":""},{"id":457904646,"identity":"e77f5538-1080-4bd3-8239-2edb0adb8d79","order_by":1,"name":"Sanda Martinčić-Ipšić","email":"","orcid":"","institution":"University of Rijeka","correspondingAuthor":false,"prefix":"","firstName":"Sanda","middleName":"","lastName":"Martinčić-Ipšić","suffix":""}],"badges":[],"createdAt":"2025-05-12 08:53:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6644722/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6644722/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":83015630,"identity":"f08dd28d-cfda-4e0b-9f3b-67b239e2e27a","added_by":"auto","created_at":"2025-05-19 06:17:21","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2767879,"visible":true,"origin":"","legend":"","description":"","filename":"Journal1ClimateResearchDomainBERTsPretrainingAdaptationandEvaluationv101.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6644722/v1_covered_6ace39e2-d45e-4478-b52c-5a62bdb86e63.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Climate Research Domain BERTs: Pretraining, Adaptation, and Evaluation","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-applied-sciences","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Applied Sciences](https://link.springer.com/journal/42452)","snPcode":"42452","submissionUrl":"https://submission.springernature.com/new-submission/42452/3","title":"Discover Applied Sciences","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"domain-specific BERT model, domain-adaptive pretraining, masked language modelling, Climate Change, ClimaBench, text classification","lastPublishedDoi":"10.21203/rs.3.rs-6644722/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6644722/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMotivated by the pressing issue of climate change and the growing volume of data, we pretrain three new language models using climate change research papers published in top-tier journals. Adaptation of existing domain-specific models is utilized for CliSciBERT and SciClimateBERT and pretraining from scratch resulted in CliReBERT (Climate Research BERT). The performance assessment is performed on the climate change NLP benchmark ClimaBench. We evaluate SciBERT, ClimateBERT, BERT, RoBERTa and DistilRoBERTa - along with our new models - CliReBERT, CliSciBERT and SciClimateBERT - using five different random seeds on all seven ClimaBench datasets.\u003c/p\u003e\n\u003cp\u003eCliReBERT achieves the highest overall performance with a macro-averaged F1 score of 65.45%, and outperforms all other models on three of the seven tasks. Additionally, CliReBERT demonstrates the most stable fine-tuning behavior, yielding the lowest average standard deviation across seeds (0.0118). The 5-fold stratified cross-validation on the SciDCC dataset showed that CliReBERT achieved the highest overall macro-average F1 score (53.75%), slightly outperforming RoBERTa and DistilRoBERTa, while the domain-adapted models underperformed their base counterparts. The superior performance of CliReBERT is accompanied by the lowest tokenizer fertility, suggesting appropriateness to model domain-specific vocabulary.\u003c/p\u003e","manuscriptTitle":"Climate Research Domain BERTs: Pretraining, Adaptation, and Evaluation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-19 06:00:56","doi":"10.21203/rs.3.rs-6644722/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-07-28T10:33:09+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-25T02:18:17+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-21T02:56:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"53046707987484436996162959265271721778","date":"2025-07-18T15:10:00+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"138829159625288544440549906577842855907","date":"2025-07-07T13:43:31+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-05-15T03:32:46+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-05-14T02:51:25+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-05-14T02:48:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Applied Sciences","date":"2025-05-12T08:48:04+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-applied-sciences","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Applied Sciences](https://link.springer.com/journal/42452)","snPcode":"42452","submissionUrl":"https://submission.springernature.com/new-submission/42452/3","title":"Discover Applied Sciences","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"07f7aaaa-43e7-48a8-ac15-070b33675b49","owner":[],"postedDate":"May 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-09-05T11:08:37+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-19 06:00:56","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6644722","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6644722","identity":"rs-6644722","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}