Geostatistical analysis of soil reflectance spectra for field-scale digital soil mapping. A case study.

Knowledge of field-scale soil variability is essential for sustainable soilmanagement. Traditional techniques, based on soil analysis, are costly and timeconsuming. An alternative method would be the use of visible-infrared reflectancespectroscopy coupled with multivariate analysis, specifically principal componentanalysis (PCA) and geostatistics.In this study, after brief reviews regarding reflectance spectroscopy, PCA, andgeostatistics, we presented a methodological approach for digital soil mapping in astudy area of Southern Italy. Reflectance spectra of 240 surface soil samplescollected at geo-referenced sites, were decomposed by PCA. The first threecomponents (PC1, PC2, PC3) explained most (98%) of the total variance of theinitial data set, therefore, they were considered for the assessment of soil spatialvariability by variography and kriging (geostatistics). The resulting PC1, PC2 andPC3 kriging maps were interpreted in the light of the information contents onreflectance spectra and compared with the results of a previous, conventional soilsurvey. The presented strategy seems to be efficient and reliable for mapping soilspatial variability.