Time-varying parameterization of a reservoir water quality model using data assimilation

Area of Science:

• Environmental Science
• Hydrology
• Water Quality Modeling

Background:

• Complex water quality models face challenges like structural misspecification and unrealistic constant parameterization.
• Data assimilation with dual state-parameter updating offers adaptive calibration for evolving environmental conditions.

Purpose of the Study:

• To pioneer the adjustment of temporal parameter evolution in a reservoir water quality model.
• To improve temperature and dissolved oxygen (DO) profile estimations using an ensemble Kalman filter framework.

Main Methods:

• Implemented a dual state-parameter updating framework using the ensemble Kalman filter.
• Optimized parameter, state, and observation vectors for effective parameter updates.
• Evaluated multiple parameter evolution strategies for a CE-QUAL-W2 model of Wadi Dayqah Reservoir.

Main Results:

• Significantly improved water quality estimation in the reservoir.
• Achieved a 16.8% reduction in root mean square error for temperature profile estimation.
• Achieved a 43% reduction in root mean square error for dissolved oxygen (DO) profile estimation.

Conclusions:

• Time-evolving parameter adjustment via dual state-parameter updating enhances reservoir water quality modeling.
• The adaptive framework improves model flexibility and alleviates process misspecifications.
• This approach offers a robust method for calibrating complex water quality models in stratified reservoirs.