Uncertainty analysis of coupling multiple hydrologic models and multiple objective functions in Han River, China

Area of Science:

• Hydrology
• Environmental Science
• Water Resource Management

Background:

• Accurate rainfall-runoff simulation is crucial for water resource management.
• Quantifying prediction uncertainty in hydrological models remains a challenge.
• Integrating multiple models and objective functions can potentially improve simulation performance.

Purpose of the Study:

• To evaluate the effectiveness of combining multiple hydrological models with various objective functions.
• To apply Bayesian model averaging (BMA) for improved rainfall-runoff prediction and uncertainty estimation.
• To compare the performance of individual models against the BMA approach.

Main Methods:

• Three hydrological models were employed to simulate rainfall-runoff processes.
• Three objective functions (squared transformed flows, squared root transformed flows, logarithmic transformed flows) were used.
• Bayesian model averaging (BMA) was utilized to combine predictions from nine individual models.

Main Results:

• The coupled approach of multiple models and objective functions generally outperformed individual models.
• BMA provided reliable mean flow predictions and uncertainty intervals.
• Distinct uncertainty components (individual model, total BMA, between-model, within-model) were characterized.

Conclusions:

• Combining multiple hydrological models and objective functions enhances rainfall-runoff simulation accuracy.
• Bayesian model averaging is an effective tool for improving hydrological predictions and uncertainty quantification.
• The study demonstrates a robust framework for hydrological modeling in the Han River basin.