Machine Learning and Deep Learning for Multi-Source Precipitation Integration in the Yangtze River Basin

 Accurate precipitation estimation is crucial for hydrological modeling and flood forecasting in the Yangtze River Basin (YRB), China. This study explores the use of machine learning (ML) and deep learning (DL) methods to fuse multi-source precipitation data, including satellite, radar, and ground-based observations. We apply models such as Random Forest (RF), Support Vector Machines (SVM), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM) networks to improve precipitation estimation accuracy. Performance is evaluated using metrics like Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Our results demonstrate that deep learning models, particularly CNNs and LSTMs, outperform traditional ML methods in terms of accuracy and spatial consistency. This work provides a robust approach to multi-source data fusion, enhancing precipitation monitoring and hydrological applications in the YRB.