Application of novel artificial bee colony optimized ANN and data preprocessing techniques for monthly streamflow...

Area of Science:

• Hydrology and Water Resource Management
• Computational Intelligence in Environmental Science

Background:

• Accurate streamflow estimation is critical for sustainable water resource management, disaster preparedness, and various water-related applications.
• Traditional hydrological models often face challenges in accurately predicting streamflow, particularly in regions prone to extreme events like droughts and floods.

Purpose of the Study:

• To develop and evaluate novel hybrid models for enhanced streamflow estimation.
• To assess the efficacy of combining Artificial Bee Colony (ABC) optimized Artificial Neural Networks (ANN) with advanced signal decomposition techniques.

Main Methods:

• Developed an Artificial Bee Colony-Artificial Neural Network (ABC-ANN) hybrid model.
• Integrated the ABC-ANN model with Local Mean Decomposition (LMD) and Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) signal decomposition techniques.
• Applied these hybrid models (LMD-ABC-ANN and CEEMDAN-ABC-ANN) for streamflow prediction in the East Black Sea Region, Türkiye.

Main Results:

• The study successfully evaluated the performance of the novel LMD-ABC-ANN and CEEMDAN-ABC-ANN hybrid approaches.
• Demonstrated the potential of these advanced hybrid models in improving streamflow prediction accuracy.

Conclusions:

• The developed hybrid models offer reliable strategies for enhancing streamflow estimation.
• These findings provide valuable resources for water resource planners and policymakers in managing water resources effectively.