Abstract:Runoff forecasting is crucial for sustainable water resource management. Despite the widespread application of deep learning methods in this field, there is still a need for improvement in the modeling and utilization of multi-scale information. For the first time, we introduce the Neural Basis Expansion Analysis with Exogenous Variable (NBEATSx) model to perform runoff prediction for a full exploration in rich temporal characteristics of runoff sequences. To harness wavelet transform (WT) multi-scale information capabilities, we developed the WT-NBEATSx forecasting model, integrating WT and NBEATSx. This model was further enhanced by incorporating a Long Short-Term Memory (LSTM) model for superior long-term dependency detection and a Random Forest (RF) model as a meta-model. The result is the advanced multi-model fusion forecasting model WT-NBEATSx-LSTM-RF (WNLR). This approach significantly enhances performance in runoff prediction. Utilizing a daily scale runoff and meteorological dataset from the Yangtze River Source region in China from 2006 to 2018, we systematically evaluated the performance of the WNLR model in runoff prediction tasks. Compared with LSTM, Gated Recurrent Units (GRUs), and NBEATSx models, the WNLR model not only significantly outperforms the original NBEATSx model but also surpasses other comparison models, particularly in accurately extracting cyclical change patterns, with NSE scores of 0.986, 0.974, and 0.973 for 5-, 10-, and 15-day forecasts, respectively. Additionally, compared to the standalone LSTM and GRU models, the introduction of wavelet transforms to form WT-LSTM and WT-GRU notably improved prediction performance and robustness, especially in long-term forecasts, where NSE increased by 32% and 1.5%, respectively. This study preliminarily proves the effectiveness of combining the cyclical characteristics of NBEATSx and wavelet transforms and creatively proposes a new deep learning model integrating LSTM and RF, providing new insights for further considering multi-scale features of complex runoff time series, thereby enhancing runoff prediction effectiveness.

Evaluate Transformer model and Self-Attention mechanism in the Yangtze River basin runoff prediction

A novel deep learning rainfall–runoff model based on Transformer combined with base flow separation

A Ladder Water Level Prediction Model for the Yangtze River Based on Transfer Learning and Transformer

Research on Runoff Prediction Based on Time2Vec-TCN-Transformer Driven by Multi-Source Data

Runoff Predictions in New-Gauged Basins Using Two Transformer-Based Models

An attention-based LSTM model for long-term runoff forecasting and factor recognition

Monthly-scale hydro-climatic forecasting and climate change impact evaluation based on a novel DCNN-Transformer network

A Surrogate Model for the Variable Infiltration Capacity Model Using Deep Learning Artificial Neural Network

A New Rainfall-Runoff Model Using Improved LSTM with Attentive Long and Short Lag-Time

Enhancing streamflow simulation in large and human-regulated basins: Long short-term memory with multiscale attributes

Enhanced LSTM model for daily runoff prediction in the upper huai river basin, china

Deep dive into predictive excellence: Transformer's impact on groundwater level prediction

Deep insight into daily runoff forecasting based on a CNN-LSTM model

Research on Runoff Simulations Using Deep-Learning Methods

Application of a Multi-Model Fusion Forecasting Approach in Runoff Prediction: A Case Study of the Yangtze River Source Region

Interpretable probabilistic modeling method for runoff prediction: A case study in Yangtze River basin, China

Short-term runoff prediction using deep learning multi-dimensional ensemble method

Research on runoff process vectorization and integration of deep learning algorithms for flood forecasting

The Application of Machine Learning in Runoff Prediction

Comparative Study of Multi-Combination Models for Medium- and Long-Term Runoff Prediction in Weihe River

A deep learning-based hybrid approach for multi-time-ahead streamflow prediction in an arid region of Northwest China