|Reeves Iii, James|
Submitted to: Encyclopedia of Soil Science
Publication Type: Book / Chapter
Publication Acceptance Date: December 21, 2004
Publication Date: December 22, 2005
Citation: Reeves III, J.B., McCarty, G.W. 2005. Spectroscopic methods for soil carbon assessment. Encyclopedia of Soil Science. Interpretive Summary: Interest in development of rapid methods for measurement of carbon content in agricultural ecosystems has been stimulated by the need to sequester carbon released to the atmosphere from fossil fuel combustion. One approach to reducing content of atmospheric carbon is to increase carbon storage in agricultural soils, but methods are needed to economically monitor the amount of carbon stored in soil. Spectroscopic techniques based on reflection of infrared light from soil have been shown to be a useful means for rapid detection of carbon stored in soil. Results using a variety of soil samples have demonstrated that two different regions of infrared (near infrared and mid infrared) show great potential for the determination of soil composition, especially soil C. While mid-IR calibrations for soil C appear to be more accurate and robust, the use of this region for quantitative analysis is not nearly as developed as the near infrared region. With reasonable input of resources, spectroscopy has the potential to generate the extensive databases on soil properties needed for generating maps for soil properties that can then be used for management of agricultural lands. The overall results of this work have demonstrated the utility of spectroscopic techniques for rapid detection of soil properties.
Technical Abstract: Diffuse reflectance spectroscopy based on near-infrared radiation (NIRS) has become an important method for analyzing agricultural products where large numbers of measurements are needed to detect variations in product composition that impact market value (i.e., protein in grains and food/feed composition) or to monitor and measure spatial and temporal variation in environmental parameters (i.e., soil carbon and other soil properties). More recently, diffuse reflectance spectroscopy using Fourier transform mid-infrared (DRIFTS) has also been shown to be capable of rapid quantitative analysis of agricultural products and environmental samples. Through the process termed chemometrics, information contained in the entire spectrum is related to the property of interest by use of multivariate statistical techniques such as stepwise regression, principle component analysis or, more recently, partial squares regression. Mathematical models produced by these techniques form the calibration models used to predict properties of unknown samples. With the development of robust calibrations, NIRS and DRIFTS have the ability to analyze samples rapidly and simultaneously for multiple properties with virtually no consumables and minimal sample preparation.