[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-127932-en":3,"doc-seo-127932-105":31,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},127932,687207024478,"Liam","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Interpretable Multiscale Machine Learning-Based Parameterizations of Convection for ICON","Machine learning–based parameterizations are designed to improve Earth System Model (ESM) representation of subgrid processes or to accelerate computation. This work uses a filtering technique in ICON to separate convection from related effects in realistic storm-resolving simulations, then benchmarks multiple offline ML algorithms on convective fluxes. Shapley values show that an unablated U-Net achieves best offline performance while learning reverse non-causal precipitation relations, whereas removing non-causal precipitation connections yields more stable online coupling, improved precipitation predictions, and 180-day simulation stability at the cost of added smoothing bias.","RESEARCH ARTICLE  \n10.1029/2024MS004398  \nKey Points:  \n• We train/benchmark machine learning models on convective fluxes derived from realistic coarse‐grained data of storm‐resolving simulations  \n• Shapley values reveal that the best offline model, a U‐Net, learns non‐ causal links to precipitation and shows poor online performance  \n• A model, without non‐causal precipitation connections, runs more stable coupled to ICON and indicates better precipitation predictions  \nSupporting Information:  \nSupporting Information may be found in the online version of this article.  \nCorrespondence to:  \nH. Heuer,  \n[helge.heuer@dlr.de](helge.heuer@dlr.de)  \nCitation:  \nHeuer, H., Schwabe, M., Gentine, P., Giorgetta, M. A., & Eyring, V. (2024) . Interpretable multiscale machine learning‐ based parameterizations of convection for ICON. Journal of Advances in Modeling Earth Systems, 16, e2024MS004398 .  \n[https://doi.org/10.1029/2024MS004398](https://doi.org/10.1029/2024MS004398)  \nReceived 12 APR 2024  \nAccepted 1 AUG 2024  \n© 2024 The Author(s) . Journal of Advances in Modeling Earth Systems published by Wiley Periodicals LLC on behalf of American Geophysical Union. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.  \nInterpretable Multiscale Machine Learning‐Based Parameterizations of Convection for ICON  \nHelge Heuer1 , Mierk Schwabe1 , Pierre Gentine2 , Marco A. Giorgetta3 , and Veronika Eyring1,4   \n1Deutsches Zentrum für Luft‐ und Raumfahrt e.V. (DLR), Institut für Physik der Atmosphäre, Wessling, Germany, 2Center for Learning the Earth with Artificial Intelligence and Physics (LEAP), Columbia University, New York, NY, USA, 3Max Planck Institute for Meteorology, Hamburg, Germany, 4University of Bremen, Institute of Environmental Physics (IUP), Bremen, Germany  \nAbstract Machine learning (ML)‐based parameterizations have been developed for Earth System Models (ESMs) with the goal to better represent subgrid‐scale processes or to accelerate computations. ML‐based parameterizations within hybrid ESMs have successfully learned subgrid‐scale processes from short high‐ resolution simulations. However, most studies used a particular ML method to parameterize the subgrid tendencies or fluxes originating from the compound effect of various small‐scale processes (e.g., radiation, convection, gravity waves) in mostly idealized settings or from superparameterizations. Here, we use a filtering technique to explicitly separate convection from these processes in simulations with the Icosahedral Non‐ hydrostatic modeling framework (ICON) in a realistic setting and benchmark various ML algorithms against each other offline. We discover that an unablated U‐Net, while showing the best offline performance, learns reverse causal relations between convective precipitation and subgrid fluxes. While we were able to connect the learned relations of the U‐Net to physical processes this was not possible for the non‐deep learning‐based Gradient Boosted Trees. The ML algorithms are then coupled online to the host ICON model. Our best online performing model, an ablated U‐Net excluding precipitating tracer species, indicates higher agreement for simulated precipitation extremes and mean with the high‐resolution simulation compared to the traditional scheme. However, a smoothing bias is introduced both in water vapor path and mean precipitation. Online, theablated U‐Net significantly improves stability compared to the non‐ablated U‐Net and runs stable for the full simulation period of 180 days. Our results hint to the potential to significantly reduce systematic errors with hybrid ESMs.  \nPlain Language Summary Due to their computational costs, it is currently not feasible to run more accurate high‐resolution climate models on a global domain on climate (century) time‐scales. However, high‐ accuracy climate simulat","cbCaicq0IutiGBXa","https://ap.wps.com/l/cbCaicq0IutiGBXa","pdf",2655432,3,1,26,"English","en",105,"# Introduction\n## Machine learning for ESM parameterizations\n## Filtering convection in ICON simulations\n## Offline benchmarking of ML algorithms\n## Online coupling and precipitation evaluation\n## Stability and error implications","[{\"question\":\"How does the study isolate convection from other processes in ICON?\",\"answer\":\"A filtering technique is applied to explicitly separate convection from related subgrid components in realistic storm-resolving ICON simulations.\"},{\"question\":\"What do Shapley values reveal about the unablated U-Net model?\",\"answer\":\"Shapley values indicate the best offline model learns non-causal, reverse links between convective precipitation and subgrid fluxes, leading to poor online performance.\"},{\"question\":\"What is the effect of ablating non-causal precipitation connections when coupling online to ICON?\",\"answer\":\"The ablated U-Net runs more stably for 180 days and shows better agreement for precipitation extremes and mean, while introducing a smoothing bias in water vapor path and mean precipitation.\"}]","Interpretable Multiscale Machine Learning-Based Parameterizations of Convection for ICON | PDF",1785943071,66,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"interpretable-multiscale-machine-learning-based-parameterizations-of-convection-for-icon","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,49,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":48},"https://docshare.wps.com/document/","Document",2,{"item":50,"name":12,"@type":44,"position":20},"https://docshare.wps.com/document/research-report/",{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/interpretable-multiscale-machine-learning-based-parameterizations-of-convection-for-icon/127932/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-25","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"How does the study isolate convection from other processes in ICON?","Question",{"text":76,"@type":77},"A filtering technique is applied to explicitly separate convection from related subgrid components in realistic storm-resolving ICON simulations.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What do Shapley values reveal about the unablated U-Net model?",{"text":81,"@type":77},"Shapley values indicate the best offline model learns non-causal, reverse links between convective precipitation and subgrid fluxes, leading to poor online performance.",{"name":83,"@type":74,"acceptedAnswer":84},"What is the effect of ablating non-causal precipitation connections when coupling online to ICON?",{"text":85,"@type":77},"The ablated U-Net runs more stably for 180 days and shows better agreement for precipitation extremes and mean, while introducing a smoothing bias in water vapor path and mean precipitation.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":48,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]