Soil horizon delineation using vis-NIR and pXRF data

Abstract

In this study, we evaluate the use of visible-near infrared (vis-NIR) and portable X-ray fluorescence (pXRF) for delineating soil horizons and how these are affected by soil moisture in a Mollisol (A, Ab, 2E, and 2Bt) profile of south-central Wisconsin. The soil profile was scanned in the field (moist) and soil samples were taken that were scanned in the laboratory (dry) using a vis-NIR spectrometer and a pXRF spectrometer. The CIE L*a*b* color was extracted from the vis-NIR spectra, and the elemental ratios and weathering indices were calculated from the pXRF data. The CIE L*a*b* color, elemental concentrations, elemental ratios, and weathering indices of the vertical transect were interpolated into 1-cm interval up to 100-cm soil depth for moist and dry measurements which were further used in the clustering analysis for identifying soil horizons. It was found that soil moisture decreased L*, a*, and b* color coordinates and the elemental concentrations. Soil moisture had little effect on the vertical distribution of the CIE L*a*b* color and pXRF measurements. Moisture effects can be partly reduced by taking elemental ratios and calculating weathering indices, or corrected by a linear function. The color variables (a* and b*), Ca, and Calcium/Titanium ratio (CTR) distinguished A horizons (Ap, A2, and Ab), whereas the Al, Fe, Ruxton index (R), and Sesquioxide ratio (Saf) distinguished 2E, 2Bt1, and 2Bt2 horizons. The CIE L*a*b* color and pXRF data were excellent variables for delineating soil horizons, and the measurements on moist soil samples performed better than those measured on dry soil samples.