Advancements in land-use-specific sediment source apportionment: from concentration-dependent mathematical mixtures to novel lignin derived methoxy isotopes

Accelerated soil erosion poses a significant global threat to soil health. Sediment source fingerprinting aids in the identification and apportionment of the main erosion sources. Retracing sediments to their sources and respective land-use by sediment fingerprinting offers a semi-empirical, field-based approach to determine the proportional contribution of different soil erosion sources and potentially can provide information to policy makers, land managers and researchers. Given the pivotal role of land-use change in historical and contemporary soil erosion systems, current tracers are often limited by their inability to discriminate amongst land-uses. Thus, the development of tracers which are able to discriminate between land-uses represents a crucial advancement in this field. While compound specific stable isotope (CSSI) tracers are immensely beneficial for land-use-specific sediment source apportionment, they still have several unresolved issues. The aim of this thesis is to enhance land-use-specific sediment source apportionment by addressing these issues and offering advancements in the field to enable more accurate representations of sediment dynamics in the environment.
Initially, I developed a simple approach to evaluate the performance of isotopic mixing models using mathematical mixtures. Isotopic tracers depend not only on their isotopic values and mixing proportions but also on the concentration of each tracer in each source. Therefore, I devised a novel concentration-dependent mathematical mixture tool. Utilizing this tool, I employed a 'brute force' method to investigate how the number of fatty acid (FA) tracers influences the model's performance. Contrary to the previous assumption that the Bayesian framework mixing model (MixSIAR) handles all conservative tracers beneficially, I discovered that tracer redundancy, wherein tracers which share similar mixing spaces and co-linearity (one-dimensional mixing line), negatively impact the model's performance.
Our findings using the concentration-dependent mathematical mixture tool also highlighted the need for an additional land-use-specific tracer to expand the one-dimensional mixing line formed by δ13C FA tracers with co-linearity (similar mixing spaces). Consequently, I incorporated the δ13C values of lignin-derived methoxy groups (LMeO) as an additional tracer. This approach was applied to investigate the sediment history of Lake Baldegg over the past 130 years. In particular, I successfully distinguished between the inputs of plant debris (POMterr) and mineral-associated organic matter (MOAM). By assigning POMterr as its own distinct source, I was able to remove the POMterr contribution from the sediment source apportionments and apportion only the MOAM fraction.
Given our focus on Lake Baldegg’s historically deposited sediment, Suess corrections are necessary to account for the changing atmospheric δ13C composition. For more representative Suess corrections, I accounted for multiple tracer turnover times (10, 30, and 100 years) in the Suess correction. This approach facilitated achieving a more representative account of past sediment dynamics within the Baldegg catchment and stands as an important component in using isotopic tracers to estimate historic sediment dynamics. Using both the POMterr removal and Suess correction methods, our estimates for sediment source apportionment aligned well with the policy and land-use change in the catchment. While our results of the historic apportionment are highly credible, the conservativeness of δ13C LMeO during transport and deposition required further investigation.
To assess the conservativeness of δ13C LMeO values during degradation, I utilized the dual isotopes of LMeO (δ2H LMeO and δ13C LMeO) across the degradation continuum from the litter layer to the MAOM fraction. Since δ2H MeO (lignin and pectin MeO groups) values are known to be stable during degradation, I was able to disentangle isotopic fractionation from source mixing and demonstrated the stability of δ13C LMeO values during degradation. Furthermore, and importantly, I found that the dual isotope approach allowed for the discrimination of the litter layer, above-ground woody material, and root lignin. This method then enabled the apportionment of lignin sources in organic and mineral horizons. By applying this method to two contrasting soil types, a podzol and stagnosol, I elucidated different soil type-specific lignin mixing and sources across the degradation continuum. While the stagnosol demonstrated minimal translocation of above-ground lignin sources to the MOAM fraction, the podzol showed the accumulation of lignin from above-ground sources in the MOAM. Considering the high percentage of lignin in the terrestrial biosphere, this novel, simple, solvent-free, and rapid method of demonstrating lignin dynamics may hold great potential in terms of understanding and modelling carbon sequestration.
Considering the high abundance of LMeO in the terrestrial biosphere and the extremely depleted δ13C LMeO values of leaf litter (∼60‰), I assessed how much of the well-known but lesser understood bulk 13C enrichment with organic matter degradation can be explained by the transition of lignin sources from leaf litter to roots. A mass balance approach was used to determine the δ13C values of the non-LMeO fraction. Using the difference in enrichment of the δ13C non-LMeO and δ13C bulk values, up to 14% can be explained by LMeO. Our findings suggest that models using δ13C bulk as a proxy for carbon turnover may overestimate degradation.
By providing a method to test the accuracy of concentration-dependent mixing models and offering an alternative CSSI tracer capable of discriminating between POMterr and MOAM, this thesis makes significant contributions to advancing sediment fingerprinting using CSSI tracers. Additionally, the use of LMeO in determining lignin mixing dynamics in soils represents an important step in understanding carbon sequestration. Simultaneously, this thesis both resolves some of the issues with CSSI sediment source apportionment and opens new questions and tools awaiting exploration by curious and inquisitive researchers in the future.