{"id":"W4414928464","doi":"10.1016/j.apor.2025.104793","title":"Deep graph neural networks for spatiotemporal forecasting of sub-seasonal sea ice: A case study in Hudson Bay","year":2025,"lang":"en","type":"article","venue":"Applied Ocean Research","topic":"Arctic and Antarctic ice dynamics","field":"Earth and Planetary Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada; McGill University; University of Waterloo","funders":"National Research Council Canada; Alliance de recherche numérique du Canada","keywords":"Interpretability; Artificial neural network; Graph; Convolutional neural network; Sea ice; Deep learning; Bay; Mean squared error","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002180944,0.0001514967,0.0002637853,0.0003408506,0.0003666065,0.00005496871,0.0002896482,0.00009521834,0.00006011525],"category_scores_gemma":[0.0001178771,0.0001357015,0.00006796624,0.001049505,0.0002022363,0.0000934889,0.00006424332,0.0005205512,0.00000215849],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001660119,"about_ca_system_score_gemma":0.0001331702,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003759986,"about_ca_topic_score_gemma":0.01472583,"domain_scores_codex":[0.9979396,0.0001658702,0.0003771378,0.00039314,0.0004486316,0.0006756436],"domain_scores_gemma":[0.9980779,0.001370952,0.00007201425,0.0002225115,0.0001424474,0.0001142333],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003491007,0.0001096304,0.9243029,0.00009053672,0.00002985763,0.0002388911,0.0006431577,0.02421145,9.622834e-7,0.0003342605,0.0001119838,0.04957732],"study_design_scores_gemma":[0.0009427934,0.000253315,0.1381886,0.00002272782,0.00001424428,0.0000360145,0.008975044,0.8499191,0.000005856954,0.001507325,0.00001492401,0.0001200181],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9939138,0.00008539135,0.003399103,0.0001151263,0.00009370768,0.00124756,0.00002559605,0.00001640335,0.00110335],"genre_scores_gemma":[0.9988749,0.00001067986,0.0008247046,0.00005008646,0.0000772312,0.00001044702,0.0001029838,0.000007217352,0.00004174107],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8257077,"threshold_uncertainty_score":0.8217359,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04705419322509276,"score_gpt":0.3092746692788767,"score_spread":0.262220476053784,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}