DISCRIMINANT ANALYSIS OF SOILS UNDER DIFFERENT LAND USES IN THE ATLANTIC RAIN FOREST AREA USING ORGANIC MATTER ATTRIBUTES

In the mountain region of the State of Rio de Janeiro (Atlantic Rain Forest biome), small farmers practice shifting cultivation in the slash and burn system. In this work, soil surfaces samples (0-15cm) of a Yellow Red Latossolo under five different vegetal coverings (Atlantic Rain Forest, annual culture, coffee, banana and pasture), had been taken in the small farming area of Bom Jardim RJ for soil chemical characterization and carbon content analysis in the different humic fractions. The samples taken under forest and pasture could be clearly isolated from the other groups for the discriminante model created. For the banana, coffee and culture groups, a certain overlapping was observed, indicating a similarity of the soil organic matter attributes between these coverings. The model allowed to correctly classify 88% of the analyzed samples. The fulvic acid (AF), organic carbon (C), total nitrogen (N), and the C/N ratio had been selected by the model, with the fulvic acids having the attribute of higher relative weight. This result suggests that the humic acids fractionation allows the detection of alterations in the soil that are not possible to be identified for the simple determination of the total organic carbon. The discriminate pattern of areas (forest-pasture and banana-coffe-culture) denotes that fertilizer could be related with changes in some important indicator attributes such as fulvic acids.