|Abstract or Summary
- Coastal vegetation dampens waves which can provide benefits to the local area, but existing literature shows large variations in the degree of wave height attenuation depending on plant properties and wave conditions. Better knowledge of how to predict the wave height decay accurately in different types of vegetation may help to evaluate the ecosystem services provided by coastal vegetation. This thesis presents two manuscripts which evaluate and improve existing analytical and numerical models for predicting wave height attenuation in vegetation through comparison with small-scale and prototype-scale experiments with artificial and natural vegetation, respectively. The first manuscript used a quarter scale physical model experiment and a phase resolving numerical model (FUNWAVE) to evaluate random wave attenuation through two types of synthetic vegetation. Measurements were made for a control channel, each vegetation type separately, and the two vegetation types combined. The wave height attenuation was modeled with existing equations and gave an average root mean square error of 1.5%. The linear combination of the attenuation coefficients obtained from the individual plants provided a good prediction of the attenuation coefficient for the cases of the combined, heterogeneous vegetation. FUNWAVE was used to model wave attenuation for the individual and combined tests using a bottom friction factor calibrated for each run. The numerical attenuation followed the same trends as the measured data, with an average root mean square error of 1.7%. Similar to the physical model study, it was found that the linear combination of the calibrated model friction factors for the cases of the individual plants reasonably predicted the wave height attenuation for the cases of the combined vegetation with an average root mean square error of 3.2%. The second manuscript used a large wave flume and a numerical model to investigate wave height attenuation caused by emergent coastal vegetation. The experiments used threesquare bulrush (Schoenoplectus pungens) and tested regular and irregular waves typical of conditions in a coastal estuary. The attenuation curves were fit to existing equations to determine the drag coefficient, C[subscript d]. The drag coefficients were fit as functions of the Reynolds number (Re) and Keulegan Carpenter number (KC) with R² values ranging from 0.31 to 0.67 for the Re number and from 0.38 to 0.77 for the KC number, indicating that either Re or KC can be used to predict C[subscript d]. The measured attenuation was modeled with SWAN-VEG, a phase averaged spectral wave model which includes a module to account for attenuation due to vegetation with a tunable coefficient, C[subscript d]. C[subscript d] was determined two ways. First, C[subscript d] was estimated from the Re and KC values using the empirical equations tuned to the measurements. For this case, the model produced root mean square errors ranging from 6.8% to 8.0%. Second, C[subscript d] was calibrated directly with the observations for each case. For this method, the model produced root mean square errors ranging from 1.4% to 5.9%, indicating that the model accuracy could be improved if the empirical equations to predict C[subscript d] based on Re or KC could be improved. Finally, the observed and predicted spectra were compared. In the frequencies near the peak energy, the modeled spectra matched most cases with good agreement, but the model was less accurate at higher frequencies near the second harmonic.