peer reviewed | Intensive anchoring of leisure boats in seagrass meadows leads to mechanical damages. This anthropogenic impact creates bare mat patches that are not easily recolonized by the plant. Several tools are used to study human impacts on the structure of seagrass meadows but they are not able to assess the indirect and long term implication of mechanical destruction. We chose to investigate the possible changes in the substrate chemistry given contrasted boat impacts. Our observations show that hydrogen sulfide concentrations remain high at 15 and 20 m depth (42.6 µM and 18.8 µM) several months after the highest period of anchoring during the summer. Moreover, our multidisciplinary study reveals that anchoring impacts of large boats at 15 and 20 m depth can potentially change the seascape structure. By taking into account both structural and chemical assessments, different managing strategies must be applied for coastal areas under anthropogenic pressures. | STARE-CAPMED

International audience | Ammonia (NH 3) emitted into the atmosphere from agricultural sources may affect nearby sensitive ecosystems due to high dry deposition fluxes on vegetation and soil surfaces, contributing to critical load exceedances. Ammonia fluxes near sources are simulated by either short-range atmospheric models or regional models using large grid cell sizes. However, studies are missing on the comparison of the results simulated by these two types of models. This paper presents the effect of model formalism, input factors, especially grid cell size and wind speed and the choice of deposition threshold on the spatial patterns of NH 3 dry deposition fluxes and deposition threshold exceedances. We used the Eulerian chemistry-transport model CHIMERE and the Gaussian plume model OPS-ST on two study domains characterised by contrasting land use. We showed that the average annual NH 3 dry deposition fluxes over each whole domain are similar for both models. By contrast, NH 3 dry deposition fluxes near sources are higher when simulated with OPS-ST that provides analytical solutions that can be sampled with small grid cell sizes (i.e., from 25 to 1600 m in this study), than with CHIMERE, which uses large grid cell sizes (i.e., 800 and 1600 m). As a result, the spatial patterns of deposition threshold exceedance were very different between both models. These patterns depend mainly on grid cell size, the input factors and the choice of the deposition threshold value. We show that the model formalism has a relatively small effect on the results and that the differences result mainly from the spatial resolutions to which they can be applied. Simulation results must therefore be interpreted carefully, taking into account the simulation conditions.