A Multivariate Model for Coastal Water Quality Mapping Using Satellite Remote Sensing Images

Area of Science:

• Environmental Science
• Remote Sensing
• Oceanography

Background:

• Coastal water quality monitoring is crucial for environmental management.
• Traditional water sampling methods are labor-intensive and provide limited spatial coverage.
• Satellite remote sensing offers a cost-effective and efficient alternative for large-scale water quality assessment.

Purpose of the Study:

• To demonstrate the feasibility of mapping coastal water quality using satellite remote sensing.
• To develop and validate models for estimating key water quality parameters from satellite imagery.
• To create quantitative coastal water quality maps for a region in northern Taiwan.

Main Methods:

• Conducted in-situ water quality sampling for Secchi disk depth, turbidity, and total suspended solids.
• Acquired concurrent SPOT multispectral satellite images of the study area.
• Developed a spectral reflectance estimation scheme to derive sea surface reflectance from SPOT images.
• Established and compared univariate and multivariate models for water quality estimation using spectral reflectance data.

Main Results:

• Successfully estimated sea surface reflectance from SPOT satellite images.
• The multivariate model, considering wavelength-dependent effects of seawater constituents, outperformed the univariate model.
• Quantitative coastal water quality maps were generated using the validated multivariate model.

Conclusions:

• Satellite remote sensing, particularly using SPOT imagery and a multivariate approach, is a feasible method for coastal water quality mapping.
• The developed models provide accurate estimations of Secchi disk depth, turbidity, and total suspended solids.
• This approach enables efficient and spatially comprehensive monitoring of coastal water quality.