ABSTRACT Digital soil mapping is an alternative for the recognition of soil classes in areas where pedological surveys are not available. The main aim of this study was to obtain a digital soil map using artificial neural networks (ANN) and environmental variables that express soil-landscape relationships. This study was carried out in an area of 11,072 ha located in the Barra Bonita municipality, state of São Paulo, Brazil. A soil survey was obtained from a reference area of approximately 500 ha located in the center of the area studied. With the mapping units identified together with the environmental variables elevation, slope, slope plan, slope profile, convergence index, geology and geomorphic surfaces, a supervised classification by ANN was implemented. The neural network simulator used was the Java NNS with the learning algorithm "back propagation." Reference points were collected for evaluating the performance of the digital map produced. The occurrence of soils in the landscape obtained in the reference area was observed in the following digital classification: medium-textured soils at the highest positions of the landscape, originating from sandstone, and clayey loam soils in the end thirds of the hillsides due to the greater presence of basalt. The variables elevation and slope were the most important factors for discriminating soil class through the ANN. An accuracy level of 82% between the reference points and the digital classification was observed. The methodology proposed allowed for a preliminary soil classification of an area not previously mapped using mapping units obtained in a reference area.

A minimização de impactos ambientais e o aumento da produtividade agrícola dependem, principalmente, do conhecimento de características químicas, físicas e mineralógicas do solo. Os métodos tradicionais utilizados para este fim consomem muito tempo e são de elevado custo financeiro. O objetivo deste trabalho foi o de testar e validar um método de quantificação de atributos do solo pela energia eletromagnética refletida e detectada por espectrorradiometro no visível e infravermelho. Este procedimento surge como apoio ao método convencional de análise de solo. Para sua execução, as etapas de trabalho compreenderam duas fases: (1) criação e calibração de modelos estatísticos de determinação dos atributos do solo obtidos a partir de dados espectrais (obtidos por sensor em laboratório 450-2500 nm) extraídos de amostras de terra em uma área; (2) validação dos modelos estatísticos numa outra área desconhecida e correlações entre os valores estimados e determinados (método convencional). Concluiu-se que as equações dos atributos Fe2O3, Al2O3 e Argila, atingiram R2 > 0,80 podendo ser aplicadas a uma base de dados diferente daquela que foi utilizada na geração das equações, desde que pertença a mesma região de estudo.

Minimizing environmental impacts and increasing crop productivity depend mainly on the knowledge of chemical, physical and mineralogical characteristics of the soil attributes. However, traditional methods are timeconsuming and costly. The objective of this study was to determine and validate a method to quantify soil attributes using UV-Vis-NIR Spectroscopy as an alternative to conventional methods of soil analyses. The work comprised two main phases: (1) creation and calibration of statistical models to determine the soil attributes derived from spectral data extracted from soil samples collected in area 1, (2) validation of statistical models in area 2 and correlations between the estimated and observed values (conventional method) for each soil attribute. The equations of the attributes Fe2O3, Al2O3, and clay reached R2 > 0.80 and may be applied to a different database than the one that was used to generate the equations, provided that they belong to the same study site.