Supplementary Materials for: Machine Learning and Deep Learning in Weather Forecasting: A Systematic Review of Models, Data Sources, Validation Practices, and Operational Challenges

This supplementary material accompanies the systematic review titled "Machine Learning and Deep Learning in Weather Forecasting: A Systematic Review of Models, Data Sources, Validation Practices, and Operational Challenges" (Manuscript ID: 6316808, Advances in Meteorology). Contents: Supplementary Table ST1 – Quality Assessment Scores and Domain Matrix: Ten-criteria quality evaluation for all 99 included studies, classified as High (16–20), Moderate (10–15), or Low (0–9) quality. Supplementary Table ST2 – Taxonomy of AI Tasks in Weather Forecasting: Classification of AI tasks (regression, classification, spatiotemporal prediction, post-processing, downscaling, ensemble) with descriptions, prevalence rates, and example studies. Supplementary Table ST3 – Comparison of Deep Learning Architectures: Comparison of LSTM, CNN, Transformers, GANs, and Foundation Models with strengths, limitations, and representative studies. Supplementary Table ST4 – Validation Strategies Comparison: Comparison of temporal split, random split, cross-validation, and external validation with prevalence rates and risk assessment. Supplementary Table ST5 – Categorization of Acknowledged Limitations: Categorization of limitations (data, spatial, extremes, temporal, operational, computational, interpretability) with counts and example papers. Supplementary Table ST6 – Complete Extracted Dataset: Full data extraction for all 99 included studies (S1–S99) including bibliographic information, AI methodologies, datasets, performance metrics, validation strategies, and acknowledged limitations. Related manuscript: The systematic review manuscript is under revision at Advances in Meteorology (Manuscript ID: 6316808). Repository: The code and analysis scripts supporting this review are available at https://github.com/Fikaduway12/Weather_ai_systematic-review_study