A Vertical Conversation: How Explainable Deep Learning reveals stability-dependent information pathways in the Atmospheric Boundary Layer

A Vertical Conversation: How Explainable Deep Learning reveals stability-dependent information pathways in the Atmospheric Boundary Layer | 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 Article A Vertical Conversation: How Explainable Deep Learning reveals stability-dependent information pathways in the Atmospheric Boundary Layer John Keithley Difuntorum, Marwan Katurji, Jiawei Zhang, Peyman Zawar-Reza This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8998358/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The atmospheric boundary layer (ABL) is a turbulent “vertical conversation” that couples motions across height and time in strongly stability-dependent ways, yet most learning systems neither reveal nor exploit this structure. Here we test whether a deep learning (DL) model can recover these couplings directly from data and whether the resulting insight can guide model design. Using turbulence-resolving large-eddy simulations across neutral, weakly convective and convective regimes, we train a spatiotemporal DL model to predict near-surface windspeed from short sequences of horizontal windspeed at multiple heights. We interrogate trained networks with integrated gradients, temporal/spatial occlusion and region-of-interest attribution to map cones of influence that quantify how predictive information is distributed across height and lag and how it concentrates on coherent turbulent structures. The learned pathways reorganize with stability: neutral cases rely on earlier lags and near-surface streak-like shear structures; weakly convective cases broaden upward and redistribute attribution across intermediate lags; and convective cases become more recency-weighted while emphasizing plume and cell-boundary regions. A temporal-sampling sweep shows that skill improves when input spacing aligns with regime-dependent integral time scales. Together, these results show how explainable deep learning exposes information pathways in the ABL and provides actionable design rules for geophysical models. Earth and environmental sciences/Climate sciences/Atmospheric science/Atmospheric dynamics Physical sciences/Astronomy and planetary science/Planetary science/Atmospheric dynamics Atmospheric Boundary Layer Deep Learning Explainable AI Full Text Additional Declarations There is NO Competing Interest. Supplementary Files MovieS1Neutral.mp4 Movie S1 Neutral Regime MovieS2WeakConvective.mp4 Movie S2 Weakly Convective MovieS3Convective.mp4 Movie S3 Convective Cite Share Download PDF Status: Under Review 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-8998358","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":622981108,"identity":"0206c1b5-7643-45d4-8706-a64b19748ce3","order_by":0,"name":"John Keithley Difuntorum","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIiWNgGAWjYBACxnYGgwMVMN4HIGZjJ6SlGajlDAODBJgzA6SFmZA1zAwGDDAtzDwQEQI6mpk3HjjYdqeOX/qM6WabX9vk+ZgZGD98zMHnMLYCoJZnEpJ9OWa3c/tuG7YxMzBLztyGTwuPweGPbYclDM7wALX03GYEamFj5iWgBWgLVItlz217ErUw/LidSIQWoF8OnDssObOHrexmb8Pt5DZmxma8fjFsb9784UDZYX5+HuZtN378uW07v7354IeP+LQ0oNjZBiYbsKmEA3lU7h+8ikfBKBgFo2CEAgDDrlQuVCaZuAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-1272-333X","institution":"University of Canterbury","correspondingAuthor":true,"prefix":"","firstName":"John","middleName":"Keithley","lastName":"Difuntorum","suffix":""},{"id":622981109,"identity":"a9141df3-dc8c-4c82-82e9-350cc4e15cd3","order_by":1,"name":"Marwan Katurji","email":"","orcid":"https://orcid.org/0000-0002-3368-1469","institution":"University of Canterbury","correspondingAuthor":false,"prefix":"","firstName":"Marwan","middleName":"","lastName":"Katurji","suffix":""},{"id":622981110,"identity":"6a0655a4-aaa0-4022-900d-19db20c7cff1","order_by":2,"name":"Jiawei Zhang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jiawei","middleName":"","lastName":"Zhang","suffix":""},{"id":622981111,"identity":"058717df-c3a5-4d83-b774-c5fc387b076a","order_by":3,"name":"Peyman Zawar-Reza","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Peyman","middleName":"","lastName":"Zawar-Reza","suffix":""}],"badges":[],"createdAt":"2026-03-01 00:05:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8998358/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8998358/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107488997,"identity":"4c2b1179-b720-487b-bcb4-eb214730223a","added_by":"auto","created_at":"2026-04-22 02:46:24","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4142147,"visible":true,"origin":"","legend":"Article File","description":"","filename":"VerticalConversation1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8998358/v1_covered_34f06ab4-dfee-455c-96a6-01c87d4e89e2.pdf"},{"id":107474766,"identity":"18096087-4a03-464d-8e65-94116a8dae17","added_by":"auto","created_at":"2026-04-21 22:05:33","extension":"mp4","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":37481415,"visible":true,"origin":"","legend":"Movie S1 Neutral Regime","description":"","filename":"MovieS1Neutral.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8998358/v1/5d93d33a948f74b90f14fe94.mp4"},{"id":107474767,"identity":"6edb0ec2-72ee-472f-9154-dea9e05625b5","added_by":"auto","created_at":"2026-04-21 22:05:33","extension":"mp4","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":37490115,"visible":true,"origin":"","legend":"Movie S2 Weakly Convective","description":"","filename":"MovieS2WeakConvective.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8998358/v1/d7f4f30af261ec1877efb765.mp4"},{"id":107474768,"identity":"cb0c4398-69b2-406a-8776-e610cb9de45a","added_by":"auto","created_at":"2026-04-21 22:05:33","extension":"mp4","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":37487642,"visible":true,"origin":"","legend":"Movie S3 Convective","description":"","filename":"MovieS3Convective.mp4","url":"https://assets-eu.researchsquare.com/files/rs-8998358/v1/2242e073952dc078f091f25f.mp4"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"A Vertical Conversation: How Explainable Deep Learning reveals stability-dependent information pathways in the Atmospheric Boundary Layer","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Atmospheric Boundary Layer, Deep Learning, Explainable AI","lastPublishedDoi":"10.21203/rs.3.rs-8998358/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8998358/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The atmospheric boundary layer (ABL) is a turbulent “vertical conversation” that couples motions across height and time in strongly stability-dependent ways, yet most learning systems neither reveal nor exploit this structure. Here we test whether a deep learning (DL) model can recover these couplings directly from data and whether the resulting insight can guide model design. Using turbulence-resolving large-eddy simulations across neutral, weakly convective and convective regimes, we train a spatiotemporal DL model to predict near-surface windspeed from short sequences of horizontal windspeed at multiple heights. We interrogate trained networks with integrated gradients, temporal/spatial occlusion and region-of-interest attribution to map cones of influence that quantify how predictive information is distributed across height and lag and how it concentrates on coherent turbulent structures. The learned pathways reorganize with stability: neutral cases rely on earlier lags and near-surface streak-like shear structures; weakly convective cases broaden upward and redistribute attribution across intermediate lags; and convective cases become more recency-weighted while emphasizing plume and cell-boundary regions. A temporal-sampling sweep shows that skill improves when input spacing aligns with regime-dependent integral time scales. Together, these results show how explainable deep learning exposes information pathways in the ABL and provides actionable design rules for geophysical models.","manuscriptTitle":"A Vertical Conversation: How Explainable Deep Learning reveals stability-dependent information pathways in the Atmospheric Boundary Layer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-21 22:05:28","doi":"10.21203/rs.3.rs-8998358/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-earth-and-environment","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commsenv","sideBox":"Learn more about [Communications Earth and Environment](https://www.nature.com/commsenv/)","snPcode":"","submissionUrl":"","title":"Communications Earth \u0026 Environment","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"7b61fc2f-ad66-4616-b521-1e505f35ebe0","owner":[],"postedDate":"April 21st, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-08T19:53:10+00:00","index":1,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-08T02:55:03+00:00","index":3,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-06T09:24:06+00:00","index":2,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":66297157,"name":"Earth and environmental sciences/Climate sciences/Atmospheric science/Atmospheric dynamics"},{"id":66297158,"name":"Physical sciences/Astronomy and planetary science/Planetary science/Atmospheric dynamics"}],"tags":[],"updatedAt":"2026-04-21T22:05:28+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-21 22:05:28","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8998358","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8998358","identity":"rs-8998358","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}