Models to estimate hydraulic resistance due to vegetation in emergent wetlands are crucial to wetland design and management. Hydraulic models that consider vegetation rely on an accurate determination of a resistance parameter such as a friction factor or a bulk drag coefficient. At low Reynolds numbers typical of flows in wetlands, hydraulic resistance is orders of magnitude higher than fully turbulent flows and resistance parameters are functions of the flow regime as well as the vegetation density and structure. The exact relationship between hydraulic resistance, flow regime and vegetation properties at low-Reynolds number flows is unclear. The project goal was to improve modeling of emergent wetlands by linking vegetation and flow properties to hydraulic resistance. A 12.2-m x 1.2 m vegetated flume was constructed to evaluate seven models of vegetated hydraulic resistance through woolgrass (Scirpus cyperinus (L.) Kunth), a common native emergent wetland plant. Measurements of vegetation geometry and structure were collected after each set of flume runs. Study results showed at low stem-Reynolds numbers (<100), the drag coefficient is inversely proportional to the Reynolds number and can vary greatly with flow conditions. Empirical models that were developed from data collected in natural wetlands predicted flow velocity most accurately. Using data from this flume study, regression models were developed to predict hydraulic resistance. Results indicated stem Reynolds number, stem diameter, and vegetation area per unit volume were the best predictors of friction factor. Vegetation flexibility and water depth were also important parameters but to a lesser extent. The spatial distribution of hydraulic resistance was estimated in a small floodplain wetland near Stephens City, VA using the regression models developed from the flume data. MODFLOW was used to simulate a 4-hour flood event through the wetland. The vegetated open water surface was modeled as a highly conductive aquifer layer. On average, MODFLOW slightly underpredicted the water surface elevation. However, the model error was within the range of survey error. MODFLOW was not highly sensitive to small changes in the estimated surface hydraulic conductivity caused by small changes in vegetation properties, but large decreases in surface hydraulic conductivity dramatically raised the elevation of the water surface. / Ph. D.

This research is intended to determine the role of riparian vegetation in stream morphology. This experiment examined the effects of riparian vegetation on boundary shear stress (BSS) by completing the following objectives: (1) evaluating the effects of streambank vegetation on near-bank velocity and turbulence; (2) determining a method for measuring BSS; and, (3) examining the effects of streambank vegetation on BSS using an existing model. A second order prototype stream, with individual reaches dominated by the three vegetation types (trees, shrubs, and grass) was modeled using a fixed-bed Froude-scale modeling technique. One model streambank of the prototype stream was constructed for each vegetation type in addition to one bank with only grain roughness. Velocity profiles were measured using an acoustic Doppler velocimeter (ADV) and a miniature propeller (MP). A flush-mounted Dantec MiniCTA system was used to measure shear stress at the streambank wall. The addition of vegetation on a sloping streambank increased the streamwise free stream velocity and decreased the near-bank streamwise velocity. The turbulence caused by the upright shrub treatment increased turbulent kinetic energy and Reynolds stresses near the streambank toe, an area susceptible to fluvial erosion. The presence of dense, semi-rigid vegetation may encourage the formation of a wider channel with a vertical streambank. The small range of CTA shear stress measurements (0.02—2.14 Pa) suggested that one estimate can describe a streambank. The law of the wall technique is not appropriate because the velocity profiles did not follow the necessary logarithmic shape. Vegetative roughness present in channels created secondary flow; turbulence characteristics more appropriately estimated BSS. The BSS model predicted velocity fields in similar distribution to that measured by the ADV and MP. BSS calculated using the ray-isovel method for both velocity measurement devices were different than the measured BSS values, likely due to distortions in the measured velocity field. In general, the predicted BSS distribution increased with water depth and decreased with increasing vegetation density. The predicted BSS at the shrub toe indicated a spike in shear stress consistent with TKE estimates. / Ph. D.