A Model for Soil Organic Carbon Determination Using Portable X-Ray Fluorescence Spectroscopy.

A model for soil organic carbon determination using portable x-ray fluorescence spectroscopy Sara Nuss, David Weindorf, Yuanda Zhu, Beatrix Haggard Lofton, Amanda McWhirt, Josh Lofton, Kayla Shook, and Courtney Nelson Abstract The use of portable x-ray fluorescence spectroscopy (PXRF) has widened scientists’ capabilities for in-situ, multi-elemental analysis of soils. Experimental analysis is currently underway that aims to expand the use of PXRF for determination of organic carbon contents of various soils. PXRF applications have traditionally been limited to detection of elements having an atomic numbers greater than 14. This range excludes several elements (N, C, etc.) of importance for agricultural soils analysis and field classification. By determining the concentrations of Fe in relation to Zr in a soil sample, a model has been developed to predict the soil organic carbon (SOC) content within that sample. The Fe:Zr ratio has shown slight variability for prediction model accuracy for soils on the basis of differing parent material composition. This study has examined model performance using Louisiana soils originating from loess and alluvial parent materials, and assesses model accuracy both spatially and with depth for SOC distribution. Thirty soil cores have been collected from a total of six sites, representing alluvial and loess origination. Cores were subdivided and subject to PXRF scanning to obtain Fe:Zr concentration ratios. Loss on ignition (LOI) and C-N analyzer testing were used to determine SOC contents. Regression analysis results indicate r2 values of 0.85 and 0.87, indicating a strong relationship between Fe:Zr and SOC levels. Additional data concerning model prediction capabilities for SOC distribution spatially and with depth will be determined, in the hopes of providing a rapid, in-situ method for SOC level assessment that can be applied on the basis of parent material composition.