LPGF: Local Decoupled Pretraining and Global Scale-Wise Fusion Model Using Intrinsic Supervisions for Precipitation Nowcasting

LPGF: Local Decoupled Pretraining and Global Scale-Wise Fusion Model Using Intrinsic Supervisions for Precipitation Nowcasting | 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 LPGF: Local Decoupled Pretraining and Global Scale-Wise Fusion Model Using Intrinsic Supervisions for Precipitation Nowcasting Yang Zhao, Wei Fang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6936062/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract In recent years, deep learning-based methods for precipitation nowcasting have achieved notable successes. However, most of the existing studies focused on the elaborate design of end-to-end models, which inevitably ignore the rich supervisory signals from input data. To address the above challenges, we propose a precipitation nowcasting model named Local Decoupled Pretraining and Global Scale-Wise Fusion (LPGF), which use intrinsic supervisions for pretraining to decouple interaction features and local evolutionary features. To realize the effective use of low-frequency evolutionary features and high-frequency interaction features, we construct Scale-Wise Fusion Block (SW-FB). By analyzing the spatial frequency of the feature maps counting the proportion of maximal direction operators in each direction, we prove the effectiveness of feature decoupling and the interpretability of the model. The results show that LPGF has stronger applicability to different precipitation datasets than other models and has applicability to severe precipitation events. Precipitation nowcasting decoupling interpretability Intrinsic Supervisions Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted 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-6936062","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":482200344,"identity":"3278e079-67e7-45c6-b4d7-b2167a8ea6bd","order_by":0,"name":"Yang Zhao","email":"","orcid":"","institution":"Nanjing University of Information Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Yang","middleName":"","lastName":"Zhao","suffix":""},{"id":482200345,"identity":"80440c60-cf32-49b0-bf30-030bae929758","order_by":1,"name":"Wei Fang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA40lEQVRIiWNgGAWjYBAC9gYwJSHHwMAIZkJIfIDnAESLMVjxARK0MCSCVRKnhf3s4dc8FRbp2yWSmz9/YLCR3XCA+dkDvFp48tKsec5I5O6ckdgmcYAhzXjDATZzA3xa7BlyzIxz2yRyN9xIbAM67HDihgM8bBJ4beF/A9TyTyLd4EZi84cDDP+J0CKRY/w4t0EiAailAeiwA8RoeWPG/OeYhOGGMw/bJM4YJBvPPMxmRsBhOcYfZ9TUyRscT3/8oaLCTrbvePMzvFqAANkZoKBiJqAepOQDYTWjYBSMglEwogEAN5lK39RqMmQAAAAASUVORK5CYII=","orcid":"","institution":"Nanjing University of Information Science and Technology","correspondingAuthor":true,"prefix":"","firstName":"Wei","middleName":"","lastName":"Fang","suffix":""}],"badges":[],"createdAt":"2025-06-20 06:53:38","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6936062/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6936062/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97370032,"identity":"082acae9-fe50-4f77-9d52-b85e2cc2a24d","added_by":"auto","created_at":"2025-12-03 16:26:34","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4175184,"visible":true,"origin":"","legend":"","description":"","filename":"LPGF.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6936062/v1_covered_fbe16d80-d6eb-466a-8129-583916c2fafa.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"LPGF: Local Decoupled Pretraining and Global Scale-Wise Fusion Model Using Intrinsic Supervisions for Precipitation Nowcasting","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Precipitation nowcasting, decoupling, interpretability, Intrinsic Supervisions","lastPublishedDoi":"10.21203/rs.3.rs-6936062/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6936062/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn recent years, deep learning-based methods for precipitation nowcasting have achieved notable successes. However, most of the existing studies focused on the elaborate design of end-to-end models, which inevitably ignore the rich supervisory signals from input data. To address the above challenges, we propose a precipitation nowcasting model named Local Decoupled Pretraining and Global Scale-Wise Fusion (LPGF), which use intrinsic supervisions for pretraining to decouple interaction features and local evolutionary features. To realize the effective use of low-frequency evolutionary features and high-frequency interaction features, we construct Scale-Wise Fusion Block (SW-FB). By analyzing the spatial frequency of the feature maps counting the proportion of maximal direction operators in each direction, we prove the effectiveness of feature decoupling and the interpretability of the model. The results show that LPGF has stronger applicability to different precipitation datasets than other models and has applicability to severe precipitation events.\u003c/p\u003e","manuscriptTitle":"LPGF: Local Decoupled Pretraining and Global Scale-Wise Fusion Model Using Intrinsic Supervisions for Precipitation Nowcasting","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-09 05:32:03","doi":"10.21203/rs.3.rs-6936062/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e5fa6097-2620-4856-99bf-27449ab1c146","owner":[],"postedDate":"July 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-03T02:53:36+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-09 05:32:03","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6936062","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6936062","identity":"rs-6936062","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}