Classificação das Águas da Baía de Guanabara utilizando o sensor OLI/Landsat 8.

Monitoring coastal water bodies is a critical to water quality managing and for the understanding the dynamics of these environments. In this context, the use of remote sensing allows the obtaining of synoptic information of water properties helping in this understanding. The Guanabara Bay is a complex estuarine environment that has an intense anthropic interference that contributes to an increase of the contribution of several optically active components in the water, being a water situation of Case-2. The aim of this work is to determine different types of water in Guanabara Bay using the technique of classification and data mining. The classification technique, performed using the Ecognition8.9 program, was applied to an image of the OLI/Landsat8 sensor, related to the dry period. Seven types of water were detected that had a well defined and distinct characteristics of reflectance. After some tests, combining several descriptors, some classifications were generated that presented patterns of distribution of similar types of water, where an east-west lateral gradient in the central region, different waters in the bottom to the northeast and northwest, a water class characteristic of the central channel and another in the contact of the estuary with the ocean. The use of the classification technique proved to be effective, even when applied in a coastal water body during a period of low river flow, allowing the visualization and identification of the different types of water related to the hydrodynamic characteristics and the disposal of the various tributaries in their environment. Keywords: Classification, Case-2, Optically active components in water, Estuaries, Reflectance.