Effect of vegetation on evolving channel patterns in experimental salt marshes

Abstract

Salt marshes are highly valuable habitats and offer many important ecosystem services. They protect the coast from erosion and serve as pollution filters and nursery ground for fish. Furthermore, they are capable of storing a very high amount of carbon in a relatively short time. Their protection and restoration is therefore crucial for the mitigation of global warming. Salt marshes are shaped by a variety of biogeomorphic feedbacks which depend on many different biotic and abiotic aspects. In order to clarify the conditions needed for such biogeomorphic interactions and the effect of different vegetation colonisation patterns on it, four scaled landscape experiments were conducted in the tidal flume "The Metronome". Two experiments were unvegetated control experiments and two included vegetation. In one of these, the seeds were spread by the flowing water and in the other, circular patches were sown and expanded manually at random locations. Vegetation focused the flow more effectively and led to a split-up into several channels further from the inlet, resulting in a longer and straighter main channel and an overall further landward expansion of the system. Despite these clear morphological differences, quantitative measures of the system, like eroded volume or drainage density, were very similar in all experiments, showing that these are mainly shaped by the hydrodynamic boundary conditions and that the experiments are very comparable. The vegetation-induced morphological differences described above were more pronounced in the experiment with hydrochorous seed spreading than in the experiment with patchy seeding. This is a result of the patch locations being pre-determined randomly. This disables the biogeomorphic feedback between low flow velocities and vegetation establishment, causing further reduction of flow velocities. The completeness of this positive feedback loop is crucial for the emergence of a self-organised landscape. Even though the patches were very dense and covered a larger area of the system than the vegetation in the experiment with hydrochorous seed spreading, they were not able to foster the evolution of a self-organised landscape in a similar way. Only locally, the patches sometimes had strong biogeomorphic effects. Their density enables them to show such effects, like the stabilisation of features or the initiation of channels around them, very clearly, but only if they are in the locally favourable position for that. Since this was not the case everywhere, the overall morphology in the experiment with patchy seeding was less self-organised and more similar to the unvegetated control experiments even though the biogeomorphic feedbacks were more pronounced in some locations. The feedback loops of biogeomorphic feedbacks thus need to be complete to enable the emergence of a fully self-organised landscape.