A procedure for isolating soil organic matter fractions suitable for modeling

Fractions of soil organic matter (SOM) were obtained from three soils using alternative physical fractionation procedures, and evaluated against the requirements of model pools. We compared two-stage density fractionation (isolating free and intra-aggregate fractions, before and after dispersion, respectively) with particle-size separation of dispersed soil. For full comparison, the organomineral fraction residual from density fractionation was also size separated. In standardizing the density-based method, we found recovery of intra-aggregate organic matter highly sensitive to separation density as compared with the free. Recovery of the intra-aggregate was also influenced by dispersion energy. The greatest amount was obtained using a combination of the highest density (1.80 g cm(-3)) and dispersion energy (1500 J g(-1)). Analysis by C-13 nuclear magnetic resonance (NMR) showed O-alkyL/alkyl-C ratios 1.38 to 2.30 times greater in intra-aggregate organic matter than in the free. Diffuse reflectance Fourier transform infrared spectroscopy (DRIFT) also indicated a greater proportion of aliphatic hydrocarbon, carboxylic anions, and aromatic C in intraaggregate organic matter. The findings suggest this fraction comprises more decomposed and transformed organic matter relative to the free. Higher signal/noise ratios in NMR spectra of particle-size fractions (compared with their organomineral equivalents) were attributed to C in particulate SONT, not removed by prior density separation. Whilst particle-size fractions confuse particulate SOM with that attached to mineral surfaces, fractions isolated by two-stage density separation are small in number and display distinct chemical properties. We suggest they provide a sound basis for a model of SOM turnover based on measurable pools.