Gopi et al. (2026) Machine Learning (ML)-Based Monthly Streamflow Prediction for a River Basin: A Case Study

Identification

• Journal: Lecture notes in civil engineering
• Year: 2026
• Date: 2026-01-01
• Authors: K. Veerendra Gopi, K. Vaishnavi, Akkera Hinduja, K. Navya
• DOI: 10.1007/978-981-95-3775-4_16

Research Groups

• Department of Civil Engineering, VNR VJIET, Hyderabad, India

Short Summary

This study evaluates five machine learning models for monthly streamflow prediction across three gauge stations in the Godavari River basin, finding that the Long Short-Term Memory (LSTM) model consistently outperforms others with high accuracy.

Objective

• To predict monthly streamflow at three gauge stations in the Godavari River basin using various machine learning models, addressing the challenges of rainfall-runoff simulation in semi-arid regions.

Study Configuration

• Spatial Scale: Godavari River basin, India (specifically at three distinct gauge stations).
• Temporal Scale: Monthly streamflow prediction.

Methodology and Data

• Models used: Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), Support Vector Machine (SVM), CatBoost, and Gaussian Process Regression (GPR). Model accuracy was assessed using Coefficient of Correlation (R), Coefficient of Determination (R²), and Nash Sutcliffe Efficiency (NSE).
• Data sources: Data-driven approach relying on observed hydrological and climatic data for model training and prediction. Specific data types (e.g., rainfall, temperature, historical streamflow) are implied but not explicitly detailed beyond "observed data, climate variability, and precise availability of parameters."

Main Results

• The LSTM model demonstrated superior performance in predicting monthly streamflow across all three gauge stations compared to ANN, SVM, CatBoost, and GPR.
• The coefficient of determination (R²) for the LSTM model ranged from 0.75 to 0.92 across the gauge stations, indicating strong predictive accuracy.
• The results suggest that LSTM is a promising tool for future runoff prediction studies in the Godavari River basin, particularly in semi-arid regions where data consistency and accuracy are critical.

Contributions

• This study provides a comparative evaluation of five diverse machine learning models for monthly streamflow prediction in a specific river basin, offering insights into their applicability and performance in a semi-arid context.
• It highlights the superior capability of the LSTM model for hydrological forecasting, contributing to the growing evidence of deep learning's effectiveness in water resource management.
• The research addresses the critical need for accurate runoff forecasting in regions affected by climate variability and data scarcity.

Funding

• Not explicitly mentioned in the provided paper text.

Citation

@article{Gopi2026Machine,
  author = {Gopi, K. Veerendra and Vaishnavi, K. and Hinduja, Akkera and Navya, K.},
  title = {Machine Learning (ML)-Based Monthly Streamflow Prediction for a River Basin: A Case Study},
  journal = {Lecture notes in civil engineering},
  year = {2026},
  doi = {10.1007/978-981-95-3775-4_16},
  url = {https://doi.org/10.1007/978-981-95-3775-4_16}
}