ABSTRACT: Traditional method of soil survey is expensive, slow, and must be carried out by experienced researchers. Thus, advances in soil observation technologies and the need to obtain information quickly by modern techniques have intensified the use of proximal sensing. This study characterized surface reflectance spectra (A horizon) and related them to traditional soil classification, based on morphological, physical, and chemical properties of representative pedogenetic profiles, developed in two toposequences of the Distrito Federal, Brazil. In the toposequences, 15 soil profiles were selected for a complete morphological description and sampling for laboratory analyses. Soil-landscape relationships were established, and profiles were classified to the fourth level of the Brazilian Soil Classification System (SiBCS). Classes of similar soils were grouped based on their surface spectra, resulting in spectral curves of 10 representative soils in the studied area. The Morphological Interpretation of Reflectance Spectrum (MIRS) and second derivative of the Kubelka-Munk (KM) function were applied to the soil spectra. The clustered soils were similar, mainly in terms of color, textural class, and organic matter content. Groups based on soil physical and chemical properties and on surface and subsurface colors were similar to those determined by surface reflectance. The identification of soil-landscape relationships was fundamental to understand the genesis and distribution of the soils, which had similar chemical and physical properties to their parent materials. The analysis of clusters based on soil surface reflectance proved efficient in determining groups of soil classes with similar properties. Surface reflectance data were related to the soil surface and subsurface properties determined by traditional soil sample analyses, since the two approaches formed similar groups. The simultaneous interaction of soil properties was assessed by MIRS analysis, while the second derivative of the KM function adequately quantified the mineralogy of the spectra.

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.

O objetivo deste trabalho foi avaliar o potencial da espectroscopia de reflectância no VIS-NIR-SWIR, para a caracterização granulométrica de amostras de solos de diferentes classes texturais, e obter modelos de predição dos teores de argila, silte e areia no solo. Utilizou-se um conjunto de amostras representativas de Latossolos e Argissolo de cinco locais do Estado do Mato Grosso do Sul. Os espectros do visível e do infravermelho próximo ao infravermelho de ondas curtas (de 350 a 2.500 nm) das amostras foram obtidos e analisados. Empregaram-se a análise de componentes principais (ACP), agrupamento por "fuzzy c-means", regressão logística multinomial (RLM) e regressão por mínimos quadrados parciais. Espectros característicos para as diferentes classes texturais e a segregação de amostras de classes texturais e de locais de coleta com características distintas, por meio da ACP, "fuzzy c-means" e RLM, mostram o potencial semiquantitativo dos dados de reflectância no VIS-NIR-SWIR. Obteve-se quantificação satisfatória quanto à argila (R²=0,92, RPD=3,59), ao silte (R²=0,80, RPD=2,15) e à areia (R²=0,87, RPD=2,62). As técnicas de espectroscopia de reflectância podem auxiliar na determinação da textura e da variabilidade espacial do solo com metodologias semiquantitativas ou quantitativas.

The objective of this work was to evaluate the potential of VIS-NIR-SWIR reflectance spectroscopy for the characterization of soil particle-size distribution of samples from different textural classes, and to obtain models to predict clay, silt, and sand contents in the soil. A representative sample set of Oxisols and Ultisols from five locations in Mato Grosso do Sul state, Brazil, were used. Visible and near-infrared to short-wave infrared (from 350 to 2,500 nm) spectra of the samples were obtained and analyzed. Principal component analysis (PCA), fuzzy c-means cluster analysis, multinomial logistic regression (MLR), and partial least squares regression were used. Characteristic spectra for the different soil texture classes and segregation of samples from texture classes and from sampling sites with distinct characteristics, through PCA, fuzzy c-means, and RLM, show the semiquantitative potential of the VIS-NIR-SWIR reflectance data. Satisfactory quantification was obtained for clay (R²=0.92, RPD=3.59), silt (R²=0.80, RPD=2.15), and sand (R²=0.87, RPD=2.62). The reflectance spectroscopy techniques can help to assess soil texture and soil spacial variability with semiquantitative or quantitative methodologies.