New approach to measure soil particulate organic matter in intact samples using x-ray computed microtomography

Alexandra Kravchenko, W. Negassa, A. K. Guber, Sonja Schmidt

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Particulate soil organic matter (POM) is the biologically and chemically active fraction of soil organic matter that plays an important role in soil C processes. Measuring POM in soil samples with intact structure is crucial for a better understanding the role of the soil structure in providing physical protection to soil C. However, so far a lack of tools for POM identification has hindered progress. The objective of this study was to develop a procedure for identification and quantitative characterization of POM within intact soil samples using X-ray computed microtomography (μ-CT) images and to test performance of the proposed procedure. We used 16 4- to 6-mm soil aggregates collected at the 0- to 15-cm depth from a Typic Hapludalf soil and scanned at 6.3-μm resolution. The developed procedure combined image preprocessing steps with discriminant analysis classification. Image preprocessing was used for preliminary POM identification based on the range of gray-scale values (GVs), shape, and size of POM pieces. Final POM identification was achieved with discriminant analysis conducted using statistical and geostatistical characteristics. The POM identified in the intact individual aggregates using the proposed procedure was in a good agreement with POM measured using a conventional laboratory method (R2 = 0.73). Of particular importance for accurate identification of POM in the images was the information on spatial characteristics of POM’s GVs. To illustrate new possibilities that the use of X-ray μ-CT opens in analyzing POM within intact soil samples, we also explored and reported GVs and size–volume distributions of POM particles in the studied aggregates.
Original languageEnglish
Pages (from-to)1177-1185
JournalSoil Science Society of America Journal
Volume78
Issue number4
DOIs
Publication statusPublished - 25 Jul 2014

    Fingerprint

Cite this