New insights on lake sediment DNA from the catchment: importance of taphonomic and analytical issues on the record quality.

Over the last decade, an increasing number of studies have used lake sediment DNA to trace past landscape changes, agricultural activities or human presence. However, the processes responsible for lake sediment formation and sediment properties might affect DNA records via taphonomic and analytical processes. It is crucial to understand these processes to ensure reliable interpretations for "palaeo" studies. Here, we combined plant and mammal DNA metabarcoding analyses with sedimentological and geochemical analyses from three lake-catchment systems that are characterised by different erosion dynamics. The new insights derived from this approach elucidate and assess issues relating to DNA sources and transfer processes. The sources of eroded materials strongly affect the "catchment-DNA" concentration in the sediments. For instance, erosion of upper organic and organo-mineral soil horizons provides a higher amount of plant DNA in lake sediments than deep horizons, bare soils or glacial flours. Moreover, high erosion rates, along with a well-developed hydrographic network, are proposed as factors positively affecting the representation of the catchment flora. The development of open and agricultural landscapes, which favour the erosion, could thus bias the reconstructed landscape trajectory but help the record of these human activities. Regarding domestic animals, pastoral practices and animal behaviour might affect their DNA record because they control the type of source of DNA ("point" vs. "diffuse").

Coupling indicators and lumped-parameter modeling to assess suspended matter and soluble phosphorus losses.

For ecological and economic issues, evaluating the environmental fate of dissolved and suspended matter in catchments and river ecosystems still remains a challenge for the preservation and management of natural resources. Models are useful tools and may help to cope with this challenge, and especially to define the relationships between the state of natural systems and land and river management/uses. As it is difficult - even impossible - to carry out experiments on natural systems such as catchments, models are also useful to test hypotheses about the underlying processes acting on dissolved and suspended losses. We propose an innovative approach to achieve these objectives. By coupling environmental indicators and lumped modeling, this study aims to develop a conceptual and general framework to evaluate and test the functions that drive particulate and dissolved matter flows at the catchment and landscape scales, while respecting the constraint of parsimony for the number of model parameters. Calculated suspended matter (SM) and soluble reactive phosphorus (SRP) losses agreed well with field data. 210Pbex (excess Pb) activities in core sediments were also compared to those of 210Pbex calculated from the filling of the reservoir. Our models are parsimonious and this does not impair their accuracy in reproducing recorded outflows or evaluating the sedimentation processes associated to particulate outflows. Considering the adequacy of our models, we validate the hypothesis that river bank erosion and water table behavior are the driving processes that govern losses of particulate and solute forms of P, in the studied extensive agriculture conditions.