{"id":"W4415547721","doi":"10.1016/j.ese.2025.100632","title":"A hierarchical transformer and graph neural network model for high-accuracy watershed nitrate prediction","year":2025,"lang":"en","type":"article","venue":"Environmental Science and Ecotechnology","topic":"Hydrological Forecasting Using AI","field":"Environmental Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Ministry of Science and Technology of the People's Republic of China; Ministry of Natural Resources; Ministry of Science and Technology","keywords":"Interpretability; Artificial neural network; Watershed; Water quality; Graph; Transformer; Multivariate statistics; Fuzzy logic","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003681926,0.000542002,0.0005543549,0.0004155607,0.0002335635,0.0005471496,0.001090082,0.0007611695,0.001646688],"category_scores_gemma":[0.00123634,0.0003179914,0.0005709784,0.0005397911,0.0005045415,0.0008504247,0.0006776258,0.0009234639,0.0002297143],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001149341,"about_ca_system_score_gemma":0.001104822,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03216531,"about_ca_topic_score_gemma":0.02644252,"domain_scores_codex":[0.9998832,0.00002620953,0.000006705228,0.00003846784,0.00002591846,0.00001964897],"domain_scores_gemma":[0.9997106,0.0001496262,0.00003345188,0.0000139933,0.00007171866,0.00002045712],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000227211,0.00001181707,0.0003157702,0.00001084534,0.000009355157,0.00001992604,0.000008048622,0.9906914,0.0004988021,0.002010413,0.0002914629,0.006109389],"study_design_scores_gemma":[9.267743e-7,0.000001822521,0.00002250243,3.353847e-7,9.534033e-7,7.926667e-7,4.612745e-7,0.9994434,0.00003207267,0.0004735259,0.00002266558,6.351569e-7],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1355626,0.0006171929,0.8555225,0.0007227635,0.00009849969,0.00004962293,0.0006567182,0.001259577,0.005510518],"genre_scores_gemma":[0.9541712,0.0002441704,0.04106716,0.0001119454,0.00002390709,0.0000646153,0.0004437227,0.00005559509,0.003817709],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.03216531,"threshold_uncertainty_score":0.0639562,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00856923500480631,"score_gpt":0.2092474949964312,"score_spread":0.2006782599916249,"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."}}