Predicting Cyanobacterial Blooms Using Hyperspectral Images in a Regulated River

Area of Science:

• Environmental Science
• Water Quality Management
• Remote Sensing

Background:

• Process-based modeling of harmful cyanobacteria relies on accurate initial conditions, which are often derived from interpolated point-based data.
• Current methods for establishing initial conditions may not fully capture the spatial distribution of cyanobacteria, leading to prediction uncertainties.

Purpose of the Study:

• To develop and evaluate an optimal method for applying hyperspectral images to establish initial conditions for the Environmental Fluid Dynamics Code-National Institute of Environment Research (EFDC-NIER) model.
• To assess the impact of grid resolution on water quality modeling and initial conditions derived from cumulative distribution functions.
• To compare cyanobacteria (Microcystis) predictions using initial conditions from hyperspectral images versus traditional point-based data.

Main Methods:

• A four-step automated process in MATLAB was developed to utilize hyperspectral images for setting EFDC-NIER model initial conditions.
• The EFDC-NIER model was implemented with three grid resolutions in the Nakdong River Basin, focusing on a region dominated by Microcystis during summer.
• Evaluated the influence of grid resolution on water quality simulations and initial conditions derived from cumulative distribution functions.

Main Results:

• Hyperspectral images enable the application of detailed, grid-based initial conditions reflecting spatial cyanobacterial distribution.
• The use of hyperspectral data reduced uncertainties in water quality (cyanobacteria) modeling compared to point-based data.
• Grid resolution influenced both water quality modeling outcomes and the accuracy of initial conditions.

Conclusions:

• Hyperspectral imaging offers a superior approach for establishing initial conditions in cyanobacteria models, enhancing prediction accuracy.
• This method significantly reduces uncertainties in short-term harmful algal bloom forecasting.
• The study highlights the importance of spatially detailed initial conditions for effective water quality management.