Organic carbon across the terrestrial-to-aquatic continuum: Assessing source and delivery processess using a combined fingerprinting and carbon loss modelling approach.

Quantifying land use sources and understanding the dynamics of organic carbon (OC) in river catchments are essential to reduce both on-site and off-site impacts of soil OC (SOC) erosion. The aim of this research is to improve determination of the dominant terrestrial land-use sources of OC in freshwater sediment at a catchment scale and to assess the likely processes driving spatial and temporal changes in these sources. Four interlinked studies were conducted on two catchments to investigate specific objectives. First, OC fingerprinting and carbon loss modelling (a combination of “net” soil erosion and OC spatial distribution modelling) were carried out using existing OC and n-alkane biomarker data from Carminowe Creek, a mixed land use catchment in Cornwall, UK. This unique combination of two sediment origin techniques crucially identified that riparian woodland disconnected upslope eroded SOC and, concomitantly, provided an input of woodland-derived OC to the streams, giving an increased understanding of sediment and OC transport processes. Secondly, extensive new data from Loch Davan catchment, Aberdeenshire, was used to find the effect of novel combinations of n-alkane concentration ratios, n- alkane compound-specific stable isotopes (CSSI) and short-chain neutral lipid fatty acid (SC-NLFA) biomarkers on land use source discrimination using a Bayesian un-mixing model. In comparison to using only n-alkane ratios, a combination of n-alkane ratios and CSSI improved discrimination between arable and pasture land uses and using a combination of n-alkane ratios and SC-NLFA reduced error when discriminating four land uses (arable, pasture, forest and moorland). Thirdly, in an innovative approach, OC source proportions were identified, in both streambed and suspended sediment (SS), at a headwater sub-catchment and catchment scale. Different drivers of OC dynamics were detectable at the two different scales (sub-catchment and catchment scale), and different dominant land use sources were found in streambed and SS OC leading to improved identification of processes driving spatial and temporal OC dynamics. And finally, soil erosion “hotspots” (i.e. where there is high risk of soil degradation) can be identified by modelling catchment erosion using a variety of different erosion models. The utility of these soil erosion models in identifying hotspots, and guiding Best Management Practices (BMP), depends upon their accuracy and there is a need to assess model usefulness. Thus, a new method was developed and tested using streambed sediment land use -specific yields estimated using OC fingerprinting as a benchmark to determine which erosion model best identified the relative land use OC yields in streambed sediment. The new methods and findings from their application will improve determination of dominant terrestrial land-use sources of OC in freshwater sediment at a catchment scale, and support development of BMP to reduce impacts on land productivity and water quality due to changes in climate and human activity.