Determining soil water status and other soil characteristics by spectral proximal sensing

Automated acquisition of soil information by remote sensing is required to improve soil evaluation. Water is one of the most important elements that interfere with spectral data, and its interaction with electromagnetic energy can help diagnose soil characteristics. The objectives of this study were 1) to evaluate spectral data of wet and dry tropical Brazilian soils with different hydration: 2) to determine a method to identify soil mineralogy; 3) to evaluate clay minerals at different moisture stages and their relationship with soil mineralogy; 4) to test remote sensing techniques for the identification and influence of moisture and, 5) to determine a model to estimate soil moisture using spectral data measured in the laboratory by a spectroradiometer (450-2500 nm). Spectral reflectance was capable of distinguishing different moisture levels. As the content of moisture increased, spectral reflectance in the soil and clay mineral samples decreased. The concavity feature of absorption bands 1400 and 1900 nm decreased during dehydration. The 2200 nm band characterized the presence of kaolinite, which was more evident under drier conditions. The spectral curves of hydrated montmorillonite and kaolinite samples presented more absorption features than those of dried samples. The evaluation of the absorption feature centered at 630 nm was sensitive to the amount of kaolinite. Our results suggest that soil mineralogy could be simultaneously evaluated with wet and dry samples. The absorption features observed for clay patterns were helpful to characterize soil mineralogy. Differences in moisture content could also be detected by the soil line technique. Moisture could be differentiated by analyzing reflectance intensity mainly at the 1550-1750 nm wavelengths. It was also possible to estimate soil moisture by a multiple regression model developed (r(2) value of 0.98) with the intensity dimension of the absorption features centered at 1400, 1900 and 2200 nm. Spectral analysis gave specific soil information (moisture and mineralogy) by using unique basic information from spectral reflectance, thus providing a quick, non-destructive and nonpolluting method of analysis. (c) 2005 Elsevier B.V All rights reserved.