Improved data assimilation for algal bloom dynamics simulation in the Three Gorges Reservoir using particle filter

Area of Science:

• Environmental Science
• Water Quality Management
• Ecological Modeling

Background:

• Algal blooms are increasing in lakes and reservoirs, threatening water quality.
• Accurate prediction of algal blooms is essential for effective water resource management.
• Existing models require improved predictive capabilities for dynamic environmental conditions.

Purpose of the Study:

• To enhance algal bloom prediction accuracy using data assimilation.
• To integrate observed chlorophyll a data into an existing hydrodynamic model.
• To simultaneously update water quality state variables and model parameters.

Main Methods:

• Utilized a particle filter (PF) for data assimilation.
• Employed the Environmental Fluid Dynamics Code (EFDC) to model algal bloom dynamics.
• Applied the system to Xiangxi Bay (XXB) in the Three Gorges Reservoir (TGR).

Main Results:

• Particle filter assimilation significantly improved the accuracy and reliability of predicted chlorophyll a levels.
• Indirect updates of phosphate (PO4), ammonium (NH4), and nitrate (NO3) showed considerable improvement.
• Increased assimilation frequency effectively suppressed model error accumulation.

Conclusions:

• Particle filter assimilation is an effective method for dynamic parameter correction in algal bloom models.
• High-frequency water quality data assimilation is recommended for enhanced prediction accuracy.
• The developed system offers a robust approach for managing and predicting harmful algal blooms.