Satellite-based Optical Water Type Classification of Inland Waters Bodies of India

Abstract

Inland water bodies show diverse water quality and complexity, which can measured by remotely sensed reflectance in various spectral bands. These reflectance properties can be used to classify water into different optical water types (OWT). The OWT’s exhibit distinct spectral characteristics based on absorption and scattering properties of optically active constituents and their concentrations. This study aims to categorize the optical complexity of Indian water bodies in different OWT. More than 12900 water reflectance spectra were extracted over different times of the year from Indian water bodies using Sentinel-2 MultiSpectral Instrument. The reflectance spectra were normalized and clustered into thirteen distinct OWT using the cosine distance k-means algorithm. The thirteen OWT represents two types of oligotrophic clear blue waters, three green color chlorophyll dominated mesotrophic water, five sedimentrich brown eutrophic water, and three vegetation rich hypereutrophic waters. The temporal dynamics of OWT over India reveal distinct eutrophication changes in water types in different seasons. These changes are congruent with the other limnological, topographical and climatological conditions. Clear oligotrophic waters have maximum extent in month of November. Almost 30 percent of water bodies are always eutrophic. More than 60% of Indian waters are in hyper-eutrophic state during pre-monsoon season. This scheme of water body classification is reliable in understanding the long term changes in eutrophication associated ecotone changes.