{"paper":{"title":"Njord: A Probabilistic Graph Neural Network for Ensemble Ocean Forecasting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A probabilistic graph neural network for ocean forecasting achieves the lowest errors on a global benchmark while providing uncertainty estimates.","cross_cats":["physics.ao-ph"],"primary_cat":"cs.LG","authors_text":"Daniel Holmberg, Erik Larsson, Fredrik Lindsten, Joel Oskarsson, Teemu Roos","submitted_at":"2026-05-14T23:17:21Z","abstract_excerpt":"Ocean dynamics are inherently chaotic, yet existing machine learning ocean models produce only deterministic forecasts. We introduce Njord, a probabilistic data-driven model for ocean forecasting, applicable to both global and regional domains. Njord combines a deep latent variable framework with a graph neural network architecture, enabling sampling each forecast step in a single forward pass. We apply Njord globally at 0.25{\\deg} resolution and regionally to the Baltic Sea at 2 km resolution. To scale to these large ocean grids we introduce K-means cluster meshes that adapt to irregular sea "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"On the global OceanBench benchmark, Njord achieves the lowest errors on average across upper-ocean variables when evaluated against real-world observations, with the largest improvements in surface temperature prediction.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That K-means cluster meshes adapt sufficiently well to irregular sea-surface geometry to allow accurate and efficient scaling of the graph neural network to global 0.25-degree and regional 2 km grids.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Njord is a probabilistic GNN model using latent variables and adaptive K-means meshes that produces ensemble forecasts and outperforms deterministic ML baselines on global OceanBench and Baltic Sea domains.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A probabilistic graph neural network for ocean forecasting achieves the lowest errors on a global benchmark while providing uncertainty estimates.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"7e9c71582a64d9759e136d3b014f81fa2c162998edcab5c9bc83d381dbbee2ef"},"source":{"id":"2605.15470","kind":"arxiv","version":1},"verdict":{"id":"adbe142c-8224-40dd-b1c3-348c93d70984","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T14:39:20.562050Z","strongest_claim":"On the global OceanBench benchmark, Njord achieves the lowest errors on average across upper-ocean variables when evaluated against real-world observations, with the largest improvements in surface temperature prediction.","one_line_summary":"Njord is a probabilistic GNN model using latent variables and adaptive K-means meshes that produces ensemble forecasts and outperforms deterministic ML baselines on global OceanBench and Baltic Sea domains.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That K-means cluster meshes adapt sufficiently well to irregular sea-surface geometry to allow accurate and efficient scaling of the graph neural network to global 0.25-degree and regional 2 km grids.","pith_extraction_headline":"A probabilistic graph neural network for ocean forecasting achieves the lowest errors on a global benchmark while providing uncertainty estimates."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15470/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"cited_work_retraction","ran_at":"2026-05-19T15:22:19.471242Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T15:01:17.580396Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T14:49:58.570631Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T14:21:54.089210Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"citation_quote_validity","ran_at":"2026-05-19T13:49:41.411551Z","status":"skipped","version":"0.1.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.663458Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"c2e8f0e76011fb012772b9c1909130809a6dcd2d42b5d672cf2945d7176a3ca6"},"references":{"count":46,"sample":[{"doi":"","year":2019,"title":"From observation to information and users: The Copernicus Marine Service perspective.Frontiers in Marine Science, 6:234, 2019","work_id":"ea660b61-3a4d-4a1d-bc67-7384b6816102","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Evolution of the Copernicus Marine Service global ocean analysis and forecasting high-resolution system: Potential benefit for a wide range of users","work_id":"0aa62b17-5b73-4f5b-8224-9ae582217bff","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Nemo-Nordic 2.0: Operational marine forecast model for the Baltic Sea.Geoscientific Model Development, 14(9):5731–5749, 2021","work_id":"47d21787-d065-49a5-b697-9c17260aa40e","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"GLONET: Mercator’s end-to-end neural global ocean forecasting system.Journal of Geophysical Research: Machine Learning and Computation, 2(3), 2025","work_id":"6cbe6f05-6825-44e6-9672-09c3192db546","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Accurate Mediter- ranean Sea forecasting via graph-based deep learning.Scientific Reports, 15(45051), 2025","work_id":"90fe3b9c-6554-41cc-8821-440c65cdaca4","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":46,"snapshot_sha256":"92d2d97affc7c9f3ab5b596cd1b35c027a0bccba841c107241627d2595f57752","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"be38c72342fccc9bf87482343291b671d77bf369e4ca09e519d45e0e0b514314"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}