Quantifying organic carbon in particulate and mineral-associated fractions of calcareous soils – A method comparison

Accurate quantification of soil organic carbon (OC) pools is essential to study the dynamics of OC in soils. Therefore, organic matter is often separated into physical fractions with distinct turnover times, e.g. particulate (POM) and mineral-associated organic matter (MAOM), after which their OC content is measured. Calcareous soils are under-represented in such fractionation studies, because the required analytical differentiation between OC and inorganic carbon (IC) is not straightforward and implies more processing time, sample mass and equipment. Here, we review the performance of four methods to quantify OC in POM and MAOM fractions of soils containing 3–16 g IC kg−1 and 10–25 g OC kg−1 (n = 16). We assessed the similarity and consistency of the data obtained with the different methods. Furthermore, we checked how their ability to distinguish OC from IC was influenced by the fractionation and in particular by the dispersion with sodium hexametaphosphate (SHMP). The four methods were: 1) Elemental carbon analysis after removal of IC by acid fumigation (OCfum); 2) Elemental carbon analysis after removal of IC with aqueous acid (OCaq); 3) Elemental carbon as determined by Rock-Eval 6 thermal analysis (OCRE6); and 4) Elemental carbon analysis after removal of OC by loss on ignition and subsequent quantification of OC by difference between total carbon and IC (OCLOI). We found that the OCfum, OCaq, and OCLOI methods produced similar OC contents for bulk soils, whereas OCRE6 results were slightly but significantly lower. Total recovered OCRE6 contents of the summed POM and MAOM fractions were similar to the bulk soil OCRE6 contents. In contrast, the recovered OCfum, OCaq and OCLOI contents were slightly higher than respective bulk soil OC contents (108–112%), especially in soils with high IC/OC ratios. We show that this was not caused by operational artefacts or chemical changes that occurred during OC fractionation, but rather likely indicated error propagation. We conclude that all 4 methods can reliably quantify OC in POM and MAOM fractions in calcareous soils and that the optimal choice depends on the required accuracy of the results and the available sample mass, lab equipment and processing time. Based on our findings and practical considerations, we provide guidance for the selection of the most optimal OC quantification method to study OC dynamics in calcareous soils.