IQ-SASS - Improved Quantitative Source Assessment of organic matter in Soils and Sediments using molecular markers and inverse modeling

Understanding carbon cycling in present and past ecosystems has been of rising interest partially as a consequence of the growing awareness of global (climate) change with an aim to better explain and predict future ecosystem responses. To that end, quantitative source assessment of organic matter in soils and sediments is essential and molecular tools like molecular markers have often been successfully applied in this respect. Compound classes such as free extractable lipids including alkanes, fatty acids, and alkanols as well as bound lipids like lignin, suberin, cutin, and phospholipid fatty acids have been proven useful to trace specific sources of organic matter in soils and sediments. However, the question often arose of what individual compound classes could tell about the overall organic matter as frequently only single compound classes were studied, although the sum of the mentioned compound classes can tell much more about bulk organic matter than single compound classes. So far, the indications from individual compound classes are often interpreted separately, which could lead to contradicting results if several compound classes are considered. Thus, there is a high demand for a more comprehensive understanding of organic matter cycling and a method for more reliable deciphering different sources in soils and sediments by using and evaluating multiple compound classes simultaneously.In this project, we aim at developing a universal tool to quantitatively assess the sources of organic matter through inverse modeling based on the input of molecular composition in soils and sediments and potential sources like vegetation, microbial biomass, and pyrogenic organic matter. The basis for this tool was developed by colleagues at the University of Amsterdam as a MATLAB script titled the VERHIB model (Vegetation Reconstruction with the Help of Inverse Modeling and Biomarkers) that was successfully applied in a study on Ecuadorian peat soils. Although preliminary test applications on other data sets (lacustrine sediments, soils, peat) highlight the great potential of the model, it is, so far, not very user-friendly and needs major revisions and validation to apply it to a broader variety of settings. For this purpose, we will include a couple of data sets from different soil, peat and sediment sequences with well-known vegetation histories, which are partially based on historical land use records for soils and manipulative experiments on soils as well as independently reconstructed vegetation history by means of existing pollen records or compound-specific isotope records for peat and sediments. One of the existing data sets (Lake Baldegg sediments) will be used for the model improvement, as several independent vegetation records are available, enabling model verification during its optimization and transfer to an "R" script, which will be made widely accessible at the end of the project. The other data sets will be used for validation in different environmental settings, including archives from different climate zones to demonstrate the universal applicability of the improved model script. One important data set that will be used for model improvement and validation will be an afforestation sequence at Jaun (Switzerland), which consists of a sequential afforestation of forests with an age of 30 to 120 years on previous pasture and where the subsequent vegetation shift was successfully traced in bulk soil organic matter. This sequence will be investigated for a variety of compound classes in the proposed project and is connected to the complementary SNF-Ambizione project led by Dr. K. Gavazov (WSL) that focuses on fungal and more specifically mycorrhizal biomass. Thus, it provides an ideal test case to trace subsequent vegetation shifts in soil sequences using the improved model. To conduct the planned research, one PhD student will investigate the organic matter composition of the afforestation sequence at Jaun to mechanistically understand how subsequent afforestation changes composition of various compound classes in the soil, with special emphasis on tracing incorporation and degradation using the existing afforestation sequence and establishment of decomposition experiments in close collaboration with the running Ambizione. The second PhD student will focus on the transfer of the VERHIB model from MATLAB to R software, followed by the improvement of the model script and validation in other environmental settings and different archives. In the final 18 months of the project a web-based tool will be developed to enter data sets and to make the model accessible to the wider scientific community.


Project Funding: Swiss National Science Foundation (SNF)