摘要: The large amount of dissolved and particulate material discharged by the Amazon River into the Equatorial Atlantic Ocean cause distinct spectral response of its waters as compared to the nearby ocean waters. This paper shows the application of K-means clustering algorithm for classifying water masses in the region under the Amazon River plume influence according to their spectral behavior. Salinity and temperature data from 67 oceanographic stations were related to Sea-viewing Wide Field-of-view Sensor (SeaWiFS) remote sensing reflectances values and the following bio-optical products:(i) chlorophyll-a concentration, (ii) water attenuation coefficient and (iii) absorption coefficient for dissolved and detrital material. Four different water masses were identified such as:(1) oceanic water, (2) intermediate oceanic water, (3) intermediate river plume water and (4) Amazon River plume water. The spectral behavior of these water masses allowed concluding that the main active optical component of the waters in the region is the colored dissolved organic matter originated mostly from the Amazon River.