Luzon Straight experiments
Internal tides are known to draw around 1 Terrawatt of energy from the Earth-Moon gravitational system. It is widely believed that this energy-flux ultimately plays an important role in ocean mixing, which in turn impacts the Earth’s climate. Numerical simulations and satellite altimetry data suggest that over 70% of the internal tide energy-flux comes from a handful of steep ocean ridges, which act as generation hot spots. Many of these ridges, however, have complicated shapes, making comprehensive field studies and numerical simulations challenging. As such, a clear picture of internal tide generation by three-dimensional topography has yet to be established. Recently, there have been significant breakthroughs in using laboratory experiments to study internal tide generation. Professors Peacock and Dauxois have published experimental results concerning internal tide generation by two-dimensional ocean ridges and continental slopes, respectively. These studies reveal wave fields that are dominated by large vertical wavelengths (low-modes), which are relatively unaffected by small-scale instabilities and overturnings. In a recent collaborative study by Professors Peacock and Dauxois, supported by the MIT-France program, the scattering of low-mode internal tides was investigated. These experiments demonstrated that finite-amplitude topography presents a considerable barrier to the propagation of low-modes, providing a means to direct energy to smaller scales that are more conducive to breaking and mixing. The focus of this proposal is an ambitious new research project that is only possible because of the prior support of the MIT-France program. A new collaboration involving a recently-hired junior faculty member at ENS de Lyon, Prof Joubaud, will take the research in a novel direction by performing the first laboratory experimental studies of internal tide generation for a threedimensional, realistic topographic feature in a density stratification that is representative of the ocean. The research will form the basis of Ph.D. thesis for two new graduate students, one each from MIT and ENS de Lyon. To date, there have been no experimental studies of internal tide generation by three-dimensional topography. The only related laboratory experiments are idealized studies of the internal wave fields generated by an oscillating sphere. A very-limited number of numerical studies involving idealized three-dimensional topographic features, such as a Gaussian seamount, have been performed. These studies suggest that if a ridge is over three times longer than it is wide, the resulting wave field approaches that produced by a two-dimensional ridge of the same width. Large scale numerical simulations have focused on principal generation sites, such as the Hawaiian Island chain and the Luzon ridge system between Taiwan and the Philippines. On such large scales, however, there are concerns about how accurately the simulations can predict the radiated wave field given limited resolution of the topographic data and limited spatiotemporal resolution of the dynamics. There currently exists no experimental data that can provide a benchmark to test these numerical models. The experiment will be performed at the Coriolis facility in Grenoble, France, which is the world largest rotating table; this facility is the only one able to host our project thanks to its great size and support facilities. The wave tank is 13 meters in diameter and 1 meter deep, and can be rotated to simulate the background rotation of the Earth. The density stratification is obtained using mixture of fresh and salt water, which allows for a very robust stratification that can persist over several weeks. The filling process uses computer controlled pumps, and the facility is operated by one chief scientist and two technicians, who will help the research team design and carry out the experiment. Our goal is to try and directly model the generation of internal tides at the Luzon Strait between Taiwan and the Philippines. This is the site of the generation of one of the strongest internal tides in the world, which develops to produce some of the largest amplitude solitary waves ever reported. It is still unknown whether these large amplitude waves observed originate from generation hot spots between the ridges or are basically just evolve from a strong internal tide radiated by the ridge. This site is therefore of particular scientific interest and provides an excellent configuration for testing large-scale numerical models. Prof Peacock has participated in two scientific research cruises in the region and therefore has access to the latest bathymetry and stratification data. The topography will be built from moulded PVC by the technicians at Coriolis, who have previous experience in building such structures. Given the vast horizontal distances covered by the actual ridge system ( 200 km wide, 300 km long) in comparison to the ridge height ( 4 km tall), it is impractical to try and directly reproduce the actual aspect ratio in the lab, even in a facility as large as Coriolis. The aspect ratio of around 50:1 for the actual ridge means that a 0.1 cm tall realization in the lab would need to be 5 meters wide. We will therefore re-scale the horizontal dimensions by a factor of 6 and account for this dynamically by operating at forcing frequencies and density stratifications that maintain the ratio of the angle of energy propagation to the topographic slope (referred to as the criticality). Within the linear regime this ensures that the dynamics are entirely equivalent to the ocean problem, and the dynamics will differ only a little from the full problem, which is only weakly nonlinear (irrespective of this the experiment still provides a data-set for numerical simulations to make comparisons with). Although the Coriolis platform is very large, damping material will be added at the sides of the tank as to prevent undesired reflected waves from the side walls. The topography will therefore be 0.3 meters tall and 2.5 meters wide and 3 meters long, which is perfectly reasonable for the Coriolis platform. It will be mounted upside down, so that it floats on the free surface and extends down into the water from above. This arrangement is dynamically similar to the real ocean problem because fluid densities change by only about 1-2% going from the free-surface to the bottom of the tank, so there is no real difference between the up and down directions. This arrangement has several practical advantages. There is easy access to the topography, which can be removed and repositioned if desired. The pulley system that will also be used to oscillate the topography, thereby simulating the tides, does not need to be submerged. And finally, the region of strong stratification (the pycnocline) exists near the bottom of the tank and can therefore be drained and replaced several without having to empty the entire tank. This is particularly important, since it can take several days to empty and fill the tank, at a cost of thousands of dollars for the salt used to stratify the fluid. Experimental data will be obtained using Particle Image Velocimetry (PIV). A new facility at Coriolis will be used that allows us to scan through the wave field rapidly getting information over a 3D volume. By oscillating a mirror at a high speed, a 2D laser sheet can be moved to measure the particles positions at many different planes, very close and parallel one to each other. The three dimensional motions are then obtained by interpolating the particle positions in the volume and estimating their evolution in space with time. This technique require specific post-processing of the data that will be done using a freeware recently developed at Coriolis platform. Furthermore, the data can be analyzed with classic PIV computations in 2D, which allow us to test and compare our results with proven techniques. Added to this technique of visualization, an array of acoustic probes will be used to study the solitary waves that propagate away from the topography, at the pycnocline. The basic principle of these instruments is to monitor the position of an interface (a jump in density in our case) by measuring the traveling time of ultrasonic sound pulses through this interface. The technique has already been used with success to track solitary waves in a complicated stratification. According to the observed activity between the ridges in the first experiments, added measurements of the local mixing will be done and two complementary techniques can be used. One considers the time evolution of averaged density profiles in the mixing region measured with high precision conductivity and temperature sensors, whereas the other one is based on injecting dye in the fluid and follow its concentration with time. We will build up data sets for the wave field radiated to the East and West of the ridge system, increasing the amplitude of forcing in order to determine when the systems starts to go from being distinctly linear to distinctly nonlinear (a simple diagnostic for this is whether double the side-toside amplitude of forcing doubles the strength of the radiated wave field). If possible, data will also be gathered between the ridge, where intense internal wave activity can be expected. Since three dimensional PIV measurements produce a huge amount of data, it will require careful coordination to gather all this data in both vertical and horizontal planes. During the runs we will seek to identify the existence of any generation hot-spots at the ridge. If these exist, we will perform focused studies in these regions. After this, we will progress onto studying asymmetric tidal forcing, which may play a key role in the generation of solitary waves. We will repeat our experiment for a realistic tidal forcing obtained from a geophysical data set and study the nature of the radiated wave fields. A big question that we seek to answer is whether asymmetric forcing produces clearly asymmetric internal tides. These experiments will provide a significant breakthrough for the Oceanographic community. Not only it will bring insight into phenomena that are very difficult to observed during field studies and of high computational costs for numerical simulations, it will also generate a benchmark data set that can be used as a test bed for numerical simulations for years to come.