02225naa a2200229 a 450000100080000000500110000800800410001910000170006024501200007726000090019752014730020665300280167965300290170765300090173665300260174570000160177170000170178770000170180470000230182170000140184477301370185811282372019-01-16       2018    bl uuuu     u00u1 u    #d1 aSILVA, E. B.  aBuilding a spectral library for estimating soil texture in the Santa Catarina State, Brazil.h[electronic resource]  c2018  aSoil properties characterization is required to inform and lead decision-makers into action. To achieve this goal, it is necessary to develop effective methods to soil measure and monitoring. Several soil databases have been used to build large spectral library for improving and facilitating the use of VIS-NIR-SWIR spectroscopy approach in a collaborative way. In this study, the performance of four multivariate statistical techniques (e.g., Partial least-squares - PLS, Support vector machine - SVM, Random forest - RF, Gaussian process regression - GPR) were compared to evaluated their ability to predict sand, silt and clay content (%) using a regional soil spectral library. A total of 1534 observations were collected throughout the State of Santa Catarina (SC). These observations were used as dataset for the calibration (n = 1151) and validation (n = 383) procedures. Soil spectra was obtained using the FieldSpec 3 (ASD) sensor in the VIS-NIR-SWIR range. The spectra were subjected to seven preprocessing techniques including raw spectra (RAW), first derivative (FD-SG), second derivative (SD-SG), continuum removed reflectance (CR), detrending (DT), normalization by range (NBR) and multiplicative scatter correction (MSC). The ranges of sand (1.0 to 99.0%), silt (0.0 to 83.0%) and clay (0.0 to 77.0 %) content are wide and their coefficients of variation are large, indicating great variability of samples as a result of different soil-forming factors.  achemometrics techniques  areflectance spectroscopy  aSand  asilt and Clay content1 aGIASSON, E.1 aCATEN, A. T.1 aDOTTO, A. C.1 aDEMATTÊ, J. A. M.1 aVEIGA, M.  tIn: WORLD CONGRESS OF SOIL SCIENCE, 21., 2018, Rio de Janeiro. Abstracts... Viçosa: Sociedade Brasileira de Ciência do Solo, 2018.