Submitted to: Diffuse Reflectance International Conference Proceedings
Publication Type: Proceedings
Publication Acceptance Date: July 10, 2000
Publication Date: N/A
The objective was to determine the effect of sample diversity on mid- and near-infrared calibrations for total-C, organic-C and inorganic-C in soils. Samples consisted of 237 soils collected in the western US (in an area stretching from the Canadian border to Texas and west of the Missouri river to the Rocky mountains). A sample of each soil was also acidified to remove the inorganic- C or carbonates giving a total of 474 samples. Each sample was analyzed by combustion for total-C, and inorganic-C computed as total-C(non-acidified) - total-C(acidified). Each samples was also scanned in the near-infrared from 400 to 2498 nm on a NIRSystems model 6500 scanning monochromator equipped with a rotating sample cup, and from 4000 to 400 wavenumbers (2500 to 25,000 nm) at 4 wavenumber resolution on a DigiLab FTS60 Fourier transform mid-infrared spectrometer equipped with a custom made sample transport. With total-, organic- and inorganic-C ranging from .1 to 10.4, .02 to 9.8, and 0 to 6.5%, respectively, a wide range of sample compositions were present. While calibrations could be developed using near-infrared spectra, they were never as accurate as those based on mid-infrared spectra. Finally, the presence of inorganic-C was found to interfere with the accurate determination of organic-C when using mid-infrared spectra, although accurate calibrations were still possible. In conclusion, work has shown that mid-infrared calibrations for various forms of C in soil appear to be less affected by sample diversity than are near-infrared calibrations.