What question did this study set out to answer?

To develop a robust framework for accurately forecasting runoff by addressing challenges posed by nonlinearity and seasonal changes.

April 16, 2026Open Access

Hybrid framework for robust runoff forecasting via decomposition and machine learning

Key Points

To develop a robust framework for accurately forecasting runoff by addressing challenges posed by nonlinearity and seasonal changes.
Employs a hybrid model combining time-variant filter empirical mode decomposition, least squares support vector machine, and bidirectional long short-term memory for runoff forecasting.
Integrates flood season segmentation to differentiate between seasonal hydrological regimes.
Utilizes the improved black kite algorithm for optimizing ensemble weights based on runoff data from specific stations.
Achieves a reduction in RMSE by 42.5% and 39.8% for the Shebu and Dongjiang stations respectively compared to benchmarks.
Improvements in KGE of up to 27.2% demonstrate enhanced forecasting accuracy.
Effectively captures flood peaks while ensuring stability during non-flood periods, leading to less lag and underestimation.

Abstract

Accurate runoff forecasting is essential for flood control and water resource management, yet strong nonlinearity and seasonal non-stationarity often limit traditional models. This study proposes a hybrid decomposition-integration-optimization framework (TVFEMD-LSSVM-BiLSTM-FSS-IBKA), integrating time-variant filter empirical mode decomposition (TVFEMD), least squares support vector machine (LSSVM), bidirectional long short-term memory (BiLSTM), flood season segmentation (FSS), and an improved black kite algorithm (IBKA). TVFEMD mitigates noise and non-stationarity, LSSVM and BiLSTM capture nonlinear and temporal features, FSS distinguishes seasonal hydrological regimes, and IBKA adaptively optimizes ensemble weights. Using daily runoff data from Shebu and Dongjiang stations, the proposed model reduces RMSE by 42.5% and 39.8%, respectively, compared with the best benchmark, with KGE improvements up to 27.2%. The framework effectively captures flood peaks and maintains stability during non-flood periods, avoiding prediction lag and underestimation. Overall, the proposed model demonstrates strong accuracy, robustness, and applicability for runoff forecasting under complex hydrological conditions.

Mark Helpful

Bookmark

Relay

View Full Paper