Temporal Fusion Transformer Model for Reliable Prediction and Sub-Seasonal to Seasonal Forecasting of Reservoir Storage

This archive contains the datasets, model implementation, and prediction results associated with the manuscript titled "Temporal Fusion Transformer Model for Reliable Prediction and Sub-Seasonal to Seasonal Forecasting of Reservoir Storage." The record includes the climate dataset (daily precipitation, maximum temperature, and minimum temperature), the reservoir storage dataset (storage and water level), and the complete source code implementing the Temporal Fusion Transformer (TFT) architecture, including training and forecasting procedures. Model prediction outputs and evaluation metrics (R², RMSE, and MAE) are also provided. The materials included in this archive enable full reproducibility of the results presented in the manuscript, with all preprocessing, scaling, and sequence generation steps implemented within the training script.