Application of the Support Vector Regression Method for Turbidity Assessment with MODIS on a Shallow Coral Reef Lagoon (Voh-Koné-Pouembout, New Caledonia)

A novel Support Vector Regression (SVR) model successfully addresses the challenges of turbidity assessment in shallow, oligotrophic coral reef lagoons using MODIS satellite imagery. Unlike traditional algorithms, which struggle with bottom reflection in such environments, this model incorporates bathymetry and bottom color data alongside reflectance, significantly improving accuracy in retrieving in-situ turbidity values. Tested on New Caledonia's Voh-Koné-Pouembout lagoon—a critical part of the world's second-longest continuous coral reef—it demonstrates potential for global application in similar oligotrophic shallow lagoons. This breakthrough enhances monitoring capabilities for biodiversity and ecosystem health in pristine tropical lagoons, particularly those affected by river plumes from ultramafic watersheds.

Wattelez, G., Dupouy, C., Lefèvre, J., Ouillon, S., Fernandez, J.-M., & Juillot, F. (2017). Application of the Support Vector Regression Method for Turbidity Assessment with MODIS on a Shallow Coral Reef Lagoon (Voh-Koné-Pouembout, New Caledonia). Water, 9(10), 737. https://doi.org/10.3390/w9100737