ARS | Publication request: SOIL CARBON DETERMINATION USING INFRARED SPECTROSCOPY (DRIFTS AND NIRS)

Submitted to: Brazilian Soil Science Congress
Publication Type: Proceedings
Publication Acceptance Date: June 30, 2005
Publication Date: August 1, 2005
Citation: Madari, B.E., Reeves III, J.B., Coelho, M.R., Coelho, R.M., Machado, P.L., De-Polli, H., Benites, V.M., Souza, L.F. 2005. Soil carbon determination using infrared spectroscopy (drifts and nirs). Brazilian Soil Science Congress, July 17-22, 2005, Recife-PE, Brazil. CD.

Interpretive Summary: Spectroscopy using light beyond the range of human sight has come to be used to determine the composition of a wide variety of agricultural materials. Recently efforts have been made to use the same techniques to soil composition in a rapid and relatively inexpensive manner. Also, there is high interest in rapid methods for the determination of carbon (C) in soils for the purpose of determining whether C is being sequestered or lost from soil stocks. The objective of this work was to test mid- and near infrared diffuse reflectance spectroscopy (DRIFTS and NIRS) for the quantification of total carbon (TC) in tropical soils, from the Brazilian National Soil Collection (Embrapa Solos). High temperature (925 °C) combustion, an expensive and time consuming traditional method, was used as reference method for the calibrations (Equations relating spectra to composition). Calibration results between R2 = 0.811 and 0.948 were obtained for DRIFTS, and R2 = 0.750 and 0.952 for NIRS. In general, DRIFTS gave better calibration results. The accuracy of the calibration decreased with a decrease in the mean TC content and carbon range of the calibration sets using either spectral ranges. The reason for this could be that (1) spectral quality for carbon declines as carbon concentration declines (2) the accuracy of the standard method for the calibration (combustion) is lower for lower carbon concentration.

Technical Abstract: The objective of this work was to test mid- and near infrared diffuse reflectance spectroscopy (DRIFTS and NIRS) for the quantification of total carbon (TC) in tropical soils, from the National Soil Collection of Embrapa Solos, using PLS calibration. High temperature (925 °C) combustion was used as reference method for the calibrations. Calibration results between R2 = 0.811 and 0.948 were obtained for DRIFTS, and R2 = 0.750 and 0.952 for NIRS. In general, DRIFTS gave better calibration results. The accuracy of the calibration was decreasing with the decrease in the mean TC content and carbon range of the calibration sets using both spectral ranges. The reason for this could be that (1) spectral quality for carbon declines as carbon concentration declines (2) the accuracy of the standard method for the calibration (combustion) is lower for lower carbon concentration.