Quantifying the carbon content of aeolian sediments: Which method should we use?

Abstract Total organic carbon (TOC) and inorganic carbon (carbonates) are important soil factors and paleoenvironmental proxies used worldwide. During past decades, several different methods have been proposed for determining these parameters from various sediment archives. Although the automatic elemental analyzer is gradually replacing the traditional quantification methods, it is critical to assess the reliability of all currently used methods to ensure accurate regional comparisons, carbon stock estimations and long-term carbon-cycle dynamics modelling. In this study, three most commonly used methods for determining TOC and carbonate content are compared using sediment samples from a sand-paleosol sequence located in northeastern China. Pearson's correlation between results of the different TOC and carbonate quantification methods produced high correlation coefficients, confirming substantial consistency among them. The automatic carbon analyzer yielded more accurate values with smaller uncertainties, while results of the general chemical method and loss-on-ignition (LOI) produced relatively reliable values with more data variability. Meanwhile, combustion at 500 °C to quantify the TOC content yielded significantly higher values than the automatic carbon analyzer and chemical methods, indicating that LOI method is not as reliable as previously thought. We attribute the overestimation to unsuitable SOM to TOC conversion factors and/or the presence of volatile minerals in the samples. In contrast with other studies, chemical wet oxidation of the sand-paleosol sediments did not significantly overestimate TOC, which commonly occurs in modern soils and lacustrine sediments. Our results suggest that the reliability of different methods for measuring carbon content could be affected by the sediment's origin and environment. Assessment of the results of the different TOC and carbonate quantification methods on additional types of sediment is needed so that researchers can use the most appropriate method for their study and make appropriate paleoenvironmental reconstructions given the sediments and methods used.