Experimental Study of the Influence of Vegetation on Longitudinal Dispersion

Understanding the dispersion processes within marshes is important to assess their filtering abilities and their ecological limits. A laboratory experiment was designed to study the influence of marsh vegetation on longitudinal dispersion. The flow through vegetation was modeled by a recirculating current going through an array of wooden dowels in a 24 m long flume. The longitudinal dispersion coefficient K was estimated by releasing dye upstream of the array and measuring time series of concentration downstream of the array. The study included flow velocities ranging from 3 to 8 cm.s -1 and dowel densities ranging from 0.5% to 5%. Within the accuracy of our estimation, the observed longitudinal dispersion coefficient did not vary significantly for the range of flow velocity and dowel density used in this study. The shape of the concentration curves also demonstrated the existence of a delay phenomenon that affected 20% to 30% of a slug mass forming a 1 m long tail. This tail was attributed to the backflow region behind each dowel which trapped the dye and released it after some delay. A simple numerical model confirmed that delay zones, such as the backflow regions, can substantially increase the global longitudinal dispersion process. Finally, an estimate of each dispersive process likely to occur in the experiments showed that shear flow dispersion due to the bed shear is the other dominant process. Because of enhanced diffusivity this process is diminished within the dowel array. This could explain why K does not dramatically increase through the vegetation. Thesis Supervisor: Heidi M. Nepf Title: Assistant Professor