Skip to main content
ARS Home » Research » Publications at this Location » Publication #123140
Title: II POTENZIALE DELLA SPECTROSCOPIA NIR PER LA DETERMINAZIONE DI MISURE DI ATTIVITA BIOLOGICA NE SUOLO

Author
item MIMMO, TANJA - U OF BOLOGNA, BOLOGNA, IT
item GALLETTI, GUIDO - U OF BOLOGNA, BOLOGNA, IT
item Reeves Iii, James
item McCarty, Gregory

Submitted to: University of Bologna Italy Thesis
Publication Type: Other
Publication Acceptance Date: 2/21/2001
Publication Date: 3/25/2001
Citation: N/A

Interpretive Summary: Spectroscopy utilizes the interaction of light with materials to determine their composition. The objective of this work was to determine the effects of sample diversity on near-and mid- infrared spectroscopic calibrations (process for relating spectra to compositions) for biological activity in soil. One hundred and thirty six soil samples were taken from a 20 ha field within a small watershed. The samples were obtained in a grid pattern from the top ten centimeters. Samples were analyzed spectroscopically and by conventional methods for total N, total C, and several measures of biological activity. Near-infrared based results were not successful for biological measures, but were somewhat successful for total C and total N. Mid-infrared results were an improvement over the near-infrared results for total N and C and biological activity as reflected by three enzymes. While the mid-infrared results were not as good as found in previous studies using less diverse sample sets, they were much better than achieved using near-infrared spectra for this data set. In general, the results using mid-infrared spectra were more accurate for the determination of soil parameters than those found using near-infrared spectra. Also, results using either spectral region indicate that sample diversity has a large and detrimental effect on the ability to accurately determine biological activity in soils.

Technical Abstract: The objective of this work was to determine the effects of sample diversity on near-and mid-infrared spectroscopic calibrations for biological activity in soil. One hundred and thirty six soil samples were taken from a 20 ha field consisting of a toposequence within a small watershed. The samples were obtained in a grid pattern from the top ten centimeters. Samples were analyzed spectroscopically and by conventional methods for total N, total C, alkaline and acid phosphatase, arylsulfatase, biomass N and biomass C. Samples were scanned "as is", dried and dried and ground in the near-infrared using a scanning monochromator, and dried and ground in the mid-infrared using a Fourier Transform spectrometer. Near-infrared calibrations were not successful for biological measures as reflected by the three enzymes, biomass N or C, but were somewhat successful for total C and total N (R2 of approx .6 and .8, resp.). Mid-infrared results were an improvement over the near-infrared results for total N and C (R2 of approx. 0.85 and .95, resp.) and biological activity as reflected by the three enzymes, but again were unsatisfactory for measures of biomass N and C. With the best R2 for the three enzymes in the range of 0.7 the results were not as good as found in previous studies using less diverse sample sets, but were much better than achieved using near-infrared spectra for this data set. In general, the results using mid-infrared spectra were more accurate for the determination of soil parameters than those found using near-infrared spectra.