A MSLEFT-Transformer precipitation forecasting method fusing multi-scale spatial and temporal features

Key Points

• Accurate precipitation forecasting is achieved with the MSLEFT-Transformer model, enhancing predictive reliability.
• The model integrates multi-scale features, processing both spatial and temporal data effectively to improve forecasts.
• A novel transformer model is employed to analyze large datasets, optimizing the forecasting mechanics of precipitation events.
• Utilizing advanced machine learning techniques like transformers, this approach highlights the potential for better forecasting outcomes.