Abstract
The objective of this work was to determine the effects of sample diversity on mid-infrared diffuse reflectance spectroscopic calibrations for biological measures in soils. A total of 136 soil samples representing a landscape within a small watershed were taken from a 20-ha field. The samples were obtained in a grid pattern from the first 10 centimeters. Soil samples were analyzed by mid-infrared diffuse reflectance spectroscopy and by conventional methods for total C and N, biomass C and N, acid and alkaline phosphatase, and arylsulfatase. Dried and ground samples, non-KBr diluted, were scanned from 4000 to 400 cm-1 (4 cm-1 resolution, 64 co-added scans) on a DigiLab FTS-60 Fourier transform spectrometer using a custom made sample transport that allowed a sample area of 50 by 2-mm to be scanned. Results using partial least squares regression showed that diffuse reflectance infrared fourier transform spectroscopy was capable of determining, to some degree, compositional parameters (total C and N) and biological activity as reflected by three enzymes. Based on calibration errors relative to the mean analyte values, the results for the measures of biological activity were as good as those found in previous studies using a less diverse (mineralogically) set of samples, although calibrations for biomass C and N were not feasible. In general, the results using mid-infrared spectra were also more accurate than those found using near-infrared spectra for the same samples.