Trained Statistical Models for Pacific Bluefin Tuna Recruitment Prediction (Wang et al., 2026)

Description This dataset provides archived statistical models developed for the manuscript: Title: Trans-basin teleconnections regulate variability in Pacific bluefin tuna recruitment Authors: Hui-Yu Wang, Wan-Ling Tseng, Je-Wei Sheu, Yi-Chi Wang Journal: Ecological Informatics Manuscript status: Under revision. The study investigates how large-scale trans-basin climate teleconnections influence interannual variability in Pacific bluefin tuna recruitment. The initial dataset release contains 40 trained Generalized Additive Models (GAMs) and 40 trained Random Forest (RF) models based on six singular value decomposition (SVD)-derived teleconnection predictors. The models were developed using a 40-fold leave-one-year-out cross-validation (LOOCV) framework covering 1981–2020, with fold-specific predictor construction to prevent information leakage. The archived models support verification and reproducibility of the observational recruitment prediction analyses presented in the associated manuscript. Future versions may include additional trained models based on regional sea surface temperature (RegSST) predictors and the Pacific Decadal Oscillation (PDO) and Atlantic Multidecadal Oscillation (AMO) indices. Publication note: The associated manuscript is currently under revision at Ecological Informatics. The final journal citation and DOI will be added once available.