Plan view example of focused 3D breaking wave approach. Two symmetric crests are generated at the paddles (bottom of figure), and as the stem forms, breaking initiates and spreads laterally.
Physical Modeling at the NEES Tsunami Wave Basin at Oregon State University
There are a number of fundamental gaps in tsunami physics understanding and related unvalidated aspects of numerical models, notably 3D breaking and inundation over irregular and rough topography. The reason for this is a lack of experimental data, for which there has not been a facility capable of tackling the appropriate setups. With the NEES tsunami basin at OSU, this is no longer the case, and it is now possible to both investigate the 3D features of nearshore tsunami waves and to validate/calibrate models, which attempt to predict such phenomenon. Each computational module, as well as the integrated hydrodynamic simulator, needs to be validated through comparisons with laboratory data. For this purpose, a suite of targeted experiments is planned. Note that although preliminary design specifications are given below, the developed Numerical Wave Tank (NWT) will be the primary tool used to decide on specific experimental configurations.
In order to simulate the overland flows using the 2HD models, it is crucial to have an accurate parameterization of wave breaking. In current 2HD models, the energy dissipation mechanism is approximated by an eddy viscosity model, which is calibrated by experimental data for normal incident periodic waves, i.e. 2D long crested waves. This empirical model has not been tested carefully for the case of 3D oblique incident waves. It is also not clear if this model works well for the tsunami waveform. To resolve these questions, we plan to perform a series of breaking wave experiments.
Three-Dimensional Focused Breaking Solitary Waves
To understand the 3D wave breaking process, we will start a set of experiments in the constant water depth. The directional wave generator will be programmed to generate a focused solitary wave that breaks at a targeted location. This experiment can be accomplished by specifying the movement of each segment of the wave generator so that the segment generates an oblique incident solitary wave with a specified angle of propagation. Essentially this focused 3D breaking will utilize oblique wave-wave interactions, to create a large breaking, stem-wave in the middle of the tank. With a proper selection of stroke and starting time for each segment of the wave generator, 3D breaking solitary waves with different crest length and breaking severity can be achieved.
In these experiments, we will measure the free surface profiles and velocity in the water columns before, during and after wave breaking to examine breaking criteria and energy loss. The measurements will require wave gauges, acoustic Doppler velocimeters (ADVs), and above- and underwater video cameras. At least two forms of breaking criteria will be considered: (1) geometric criteria based on local wave steepness and (2) a kinematic criterion based on fluid particle and phase speed. Special attention will be paid to the influence of directionality on the wave breaking criteria.
Bathymetric Effects on 3D Tsunami Breaking
To further investigate the effects of bathymetry on the 3D breaking, we propose to construct two beaches. Both beaches will be made with a sand core and a sprayed-on concrete cap. The advantage of using sand as the core material for a beach is the relative ease to modify the beach configuration by removing the cap, re-shaping the core to the desired profile, and installing another concrete cap. The first beach will be a simple planar beach, with a wide smooth mound in the center of the beach. The beach slope will be 1:15 and the mound will be wide enough to create significant refraction. Analysis in the NWT will allow for refinement of the expected breaking location and intensity for a range of input waves. The focused solitary waves discussed in the previous section will be re-examined with the added effects of shoaling and refraction. Normal and oblique tsunami-like incident waves (such as N-waves) will be generated using the directional wave generator. Additionally, "real" tsunami waveforms, taken from numerical simulations of the Indian Ocean tsunami (Wang and Liu, 2006) and Papua New Guinea (Lynett et al, 2003) will be programmed into the wavemaker. This technique represents a simplified, one-way numerical-experimental hybrid simulation. The resulting unique data set will provide the necessary information to validate oblique wave breaking in both 2HD and fully 3D models; a step certainly required before overland flow can be investigated in detail with numerical tools.
The second beach will be more complex, possibly involving an offshore "reef", or a dune system, with breaks, relevant to the local bathymetry effects often observed in post-tsunami surveys. This beach configuration will be chosen to allow for other experiments, such as those looking into over-land flows and harbor oscillations, to utilize the beach. With as much community input as possible, the second beach will eventually be designed to best accommodate payload projects - design the beach to allow the most researchers possible to use it. Thus the specifics of this beach are made intentionally ill defined yet flexible; whatever configuration is chosen, a similar set of wave conditions as for the first beach will be run, and 3D breaking will be investigated.
Effects of Bottom Roughness on 3D Tsunami Runup and Overland Flow
In most existing models of tsunami runup, the effect of coastal terrain roughness (including structures) is modeled using the traditional Chezy-Manning's formula originally developed for uniform open channel flows. Based on Manning's formula the roughness effect was inferred to be small, such that in some modeling studies the effect of bottom roughness is neglected all together. In a recent study at the University of Hawaii, a series of small-scale wave tank experiments were conducted for wave runup on an adjustable slope and different roughness sizes. For runup on mild slopes (e.g., 5 degrees), roughness was observed to have a significant effect, reducing inundation and runup by 50-80% as compared with the inviscid prediction. Results also show that Manning's modeling does not predict the roughness effect accurately for mild slopes. Since coastal terrain usually has very mild slopes (less than 1 degree), it is conjectured that terrain roughness may have a significant effect in reducing tsunami runup and inundation in the field.
While University of Hawaii experiments shed new light onto the issue of beach roughness, the tank dimensions limit the scope of the project, and thus the applicability to real tsunami runup. Here, we will conduct additional experiments utilizing the two beaches discussed in Section 7.2. The roughness type will include both fixed and movable sand and gravel, stripes and wires (simulating bushes and trees), and small blocks (simulating small structures in the path of tsunami inundation). 2HD runup time series will be digitized from overhead cameras, and the inundation limit will be manually traced after each experimental trial. These experimental settings will simulate situations closer to what tsunamis encounter on natural beaches. With this data, accurate bottom dissipation models will be developed for transient tsunami flow, improving significantly upon the traditional approach.