[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84845-en":3,"doc-seo-84845-105":30,"detail-sidebar-cat-0-en-105":91},{"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":13,"seo_description":14,"update_tm":28,"read_time":29},84845,8796095462418,"Noah","https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780",8,"Research & Report","GeoXplain On the Fly Visual Explanations for Weather Foundation Models","Weather and climate foundation models generate high-dimensional forecasts whose internal relationships are hard to interpret using static plots. GeoXplain is an interactive Python visualization toolkit for geospatial attribution maps across climate variables, atmospheric pressure levels, and forecast times. It renders attribution bundles in notebook widgets or web browsers with map and globe modes, linked timelines, pressure-level controls, annotations, and optional physical-field overlays. A model-agnostic design introduces the GeoXplain Aurora Adapter backend for on-demand explanation computation via local GPU, listener, or SLURM support.","GeoXplain: On-the-Fly Visual Explanations for Weather Foundation  \nModels  \nClemens Walter Koprolin, Leonardo Trentini, Benedikt Soja, Mennatallah El-Assady, Christina Humer  \narXiv :2607 .05655v 1 [ cs .HC] 6 Jul 2026  \nETH Zurich, Zurich, Switzerland  \nFig. 1: Example of GeoXplain notebook for explaining a humidity forecast at 850 hPa over Zurich, Switzerland with Integrated Gradients. Attribution data is computed on demand in notebook cell (1), and a humidity forecast overlay is added in cell (2) . The left panel shows the resulting interactive view. Users can switch between attribution fields for temperature (t) and specific humidity (q), using control (3) . Individual pressure-level explanations can be shown or hidden with control (4), and their visual appearance can be adjusted with the panel (5), for example by switching from filled overlays to contour lines. Panel (6) adjusts the imported overlay, including its color map and opacity. Controls (7) switch between topographic and satellite basemaps and enable a 3D globe view. Control (8) exports the current view as an image, while control (9) provides timeline controls for exploring explanations across forecast times.  \nAbstract—Weather and climate foundation models produce high-dimensional forecasts whose learned relationships are difficult to inspect with static plots alone. GeoXplain is an interactive Python-based visualization toolkit for exploring geospatial attribution maps across climate variables, atmospheric pressure levels, and forecast time. The toolkit accepts attribution bundles containing attribution grids together with corresponding metadata and renders them in a notebook widget or browser with map and globe modes, linked timelines, pressure-level controls, target annotations, and optional physical-field overlays. We frame GeoXplain as a model-agnostic earth-system visualization toolkit and present the GeoXplain Aurora Adapter as its first computation backend. The adapter computes explanations for the Aurora foundation model, either in a local GPU process, through a GPU listener, or through a SLURM-backed listener, while preserving the same Python call site for analysts. It currently supports gradient saliency, Integrated Gradients, RISE, ViTCX, multi-frame saliency and Integrated Gradients rollouts, and retrieval of ERA5 overlays. GeoXplain can be installed as a PyPI package with pip install geoxplain. The code is open-source and available at [https://github.com/clemenskoprolin/geoxplain](https://github.com/clemenskoprolin/geoxplain).  \nIndex Terms—Climate visualization, geospatial visualization, explainable AI, weather foundation models, computational notebooks  \n1 INTRODUCTION  \nAI-based weather and earth-system models are rapidly becoming practical scientific instruments [4] . These models are attractive for weatherand climate-related workflows because they can produce global forecasts at comparatively low computational cost while remaining competitive with established numerical systems [5, 6, 17] . At the sametime, their predictions are distributed over space, atmospheric pressure levels, variables, and lead times, which makes it hard to answer basic interpretability questions: Which input fields influence a humid-air forecast over the Alps? Does the answer change across pressure lev-  \n[E-mail: {ckoprolin](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[ |](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[ltrentini](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[ |](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[sojab](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[ |](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[melassad](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[ |](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)[humerc}@ethz.ch](E-mail: {ckoprolin |ltrentini |sojab |melassad |humerc}@ethz.ch)  \nels? Does","cbCaimvolwN2z7GQ","https://ap.wps.com/l/cbCaimvolwN2z7GQ","pdf",4899696,3,1,9,"English","en",105,"# Introduction\n# GeoXplain Toolkit and Workflow\n## Visualization and controls\n## Model-agnostic separation and adapter backend\n# GeoXplain Contributions","[{\"question\":\"What problem does GeoXplain address for weather foundation models?\",\"answer\":\"It tackles the difficulty of interpreting high-dimensional forecast relationships that are not easily inspected with static plots, especially across space, variables, pressure levels, and forecast lead times.\"},{\"question\":\"How does GeoXplain structure the visualization and explanation computation?\",\"answer\":\"GeoXplain separates a visualization layer from model-specific computation by using adapters that generate attribution results on demand, while the viewer renders those results through consistent Python call sites.\"},{\"question\":\"What explanation methods and execution options does the Aurora Adapter support?\",\"answer\":\"It supports gradient saliency, Integrated Gradients, RISE, ViTCX, multi-frame saliency and Integrated Gradients rollouts, and ERA5 overlay retrieval, with computation performed locally on GPU, via a GPU listener, or via a SLURM-backed listener.\"}]",1784198775,23,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"geoxplain-on-the-fly-visual-explanations-for-weather-foundation-models","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":20},"https://docshare.wps.com/document/research-report/",{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/geoxplain-on-the-fly-visual-explanations-for-weather-foundation-models/84845/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-23","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does GeoXplain address for weather foundation models?","Question",{"text":75,"@type":76},"It tackles the difficulty of interpreting high-dimensional forecast relationships that are not easily inspected with static plots, especially across space, variables, pressure levels, and forecast lead times.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does GeoXplain structure the visualization and explanation computation?",{"text":80,"@type":76},"GeoXplain separates a visualization layer from model-specific computation by using adapters that generate attribution results on demand, while the viewer renders those results through consistent Python call sites.",{"name":82,"@type":73,"acceptedAnswer":83},"What explanation methods and execution options does the Aurora Adapter support?",{"text":84,"@type":76},"It supports gradient saliency, Integrated Gradients, RISE, ViTCX, multi-frame saliency and Integrated Gradients rollouts, and ERA5 overlay retrieval, with computation performed locally on GPU, via a GPU listener, or via a SLURM-backed listener.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]