Abstract

Sergio M. Saverin,Aryton J. Fernando, Alberto Carmen Meirelles, Machado Anitta Rousseff and Heitor O. Bruna

The use of spectral remote sensing is a reality for the determination of soil attributes. This study aimed to evaluate the soil spectral curves of a dataset and infer the possibility of their use for management purposes. For this, soil samples were collected at Rafard/SP in a regular grid of 100 m of edge, totaled 185 points. These samples were submitted to physical, chemical, mineralogical and spectral analyses in individual laboratories. From the results of the analysis was performed to soil classification and organization chart containing the spectral curves for each soil class. Through visual and statistics evaluations were determined, among the 18 land units of area, groups of classes with similar characteristics for both the surface layer and subsurface. These groupings refer, among other attributes, to the original material, constitution textural, mineralogical composition, etc. It is fundamental to the characterization of the area and formulates practicable strategies for better use of it. Thus, it was noted that the analysis of spectral curves can indeed generate information of great practical importance for the management of soils in this region.