Soil organic matter fractions as early indicators for carbon stock changes under different land-use?

With respect to carbon sequestration in soil, attempts have been made to identify soil organic matter (SOM) fractions that respond more rapidly to changes in land-use than bulk SOM, which could thus serve as early indicators for the overall stock change. We used a combination of physical fractionation (size and density separation) and chemical characterisation (C-to-N ratios, CuO lignin signature, ¹³C NMR spectroscopy) to identify sensitive SOM fractions in an agricultural system with sandy dystric cambisols in Bavaria, Germany, 7 years after a land-use change. Land-use types included long-term arable land and grassland, and conversion from one system to the other. Soil carbon and nitrogen contents in 0–3 cm increased from 14 to 39 mg organic carbon g⁻¹ soil, and from 1.7 to 3.9 mg nitrogen g⁻¹ soil in the following order: permanent arable, conversion grassland to arable, conversion arable to grassland, and permanent grassland. Wet sieving and ultrasonic dispersion with 22 J ml⁻¹ released <5% and 60% to 80%, respectively, of the amount of particles >20 μm relative to complete dispersion. The most sensitive fraction, with respect to land-use, was SOM in the fraction >20 μm not released after sequential wet sieving and ultrasonic dispersion. In contrast, the proportion of free light (wet sieving, density <1.8 g cm⁻³) and occluded light (ultrasonic dispersion with 22 J ml⁻¹, <1.8 g cm⁻³) particulate organic matter (POM) showed no clear response to land-use. The structural composition of POM indicated its vegetation origin with a selective enrichment of lignin and a loss of O-alkyl C relative to its plant precursors. Decomposition of the occluded light POM was only slightly advanced relative to the free light POM. In mineral fractions <20 μm, SOM was significantly more transformed than in the coarse fractions, as shown by NMR spectroscopy; however, it revealed no specific land-use pattern. An exception to this was the proportion of O-alkyl C in the clay fraction, which increased with SOC content. Ratios of alkyl to O-alkyl C in mineral fractions <20 μm differentiated samples gave a better differentiation of samples than the C-to-N ratios. We conclude that neither free nor occluded light POM are appropriate early indicators for changes in land-use at the investigated sites; however, total SOM, its distribution with depth, and SOM allocated in stable aggregates >20 μm were more sensitive.