Altimetry
Modeling
Historic
Fluxes
Generation
Wave-wave
Dissipation

Results

The Hawaii Ocean-Mixing Experiment was designed to investigate the loss of energy from the surface tide as it flows over the Hawaiian Ridge and how that energy produces internal waves and turbulent dissipation. The components of the experiment are described here. This page highlights some of the scientific results that these experiments have produced. Where possible, links to further information are provided. Also, please check the Publications page.

Satellite Altimetry

Dissipation of the Barotropic Tide

Generalized inversion has been used to infer the dissipation of the barotropic (i.e., depth-integrated) tide along the Hawaiian Ridge. The divergence of the barotropic energy flux, shown above for the M2 constituent, is a large source of energy for the tidal internal waves and the turbulent mixing which were observed during the data-collection phase of the HOME program. The eight major tidal constituents (M2, S2, N2, K2, K1, O1, P1, and Q1) altogether dissipate nearly 26GW (1 GW = 1 billion watts), of which 19GW is due to the principle lunar semi-diurnal tide, M2.

For more information see http://www.coas.oregonstate.edu/research/po/research/tide/

Coherent Internal Tide Energy Flux

The internal tides have a small, but detectable, signal in the elevation of the ocean surface. Using satellite altimetry, several researchers have been able to quantify the coherent (phase-locked) component of the internal tides and estimate the energy flux of internal waves. The figure above shows several estimates of the Ridge-normal internal wave energy flux at the M2 frequency. The estimates of Ray (red) and Egbert (green) were made from satellite data, while those of Kang (blue) and Merrifield (black) were obtained from numerical models. The peak Ridge-normal energy fluxes due to the coherent internal tides are significantly smaller than what has been computed from the models. Ongoing research is aimed at using altimeter data to quantify the incoherent internal tide and understand the source of phase variability.

For more information see http://www.coas.oregonstate.edu/research/po/research/tide/

Modeling

Historic Data Analysis

Energy Fluxes

Farfield internal wave fluxes

Estimates of baroclinic energy flux at the Farfield site (430 km SW of Kauaii channel, October-November 2001) are obtained based on repeated profiles of velocity and density obtained from the Research Platform FLIP. The Farfield baroclinic energy flux is borne primarily by first-mode semidiurnal waves. The variability of the Farfield depth-integrated (80-730 m) energy flux vectors in the semidiurnal band (b) follows the cycle of its forcing, but the diurnal energy flux (c) varies in accord to the fortnightly cycle of the barotropic semidiurnal tide, rather than with the diurnal forcing. This observation suggests nonlinear interactions. The average cross-Ridge profiles of the energy flux in the (d) semidiurnal and (e) diurnal bands is also shown. The vertical structures in both bands is close to that of mode 1, although greater depth variability is seen in (e).

For more information, see http://opg1.ucsd.edu/~lrainville/HOME/Hawaii.html

Near-ridge internal wave fluxes

An extensive survey along the 3000-m isobath of the Hawaiian Ridge was carried out in 2000 using the Absolute Velocity Profiler (AVP). AVP profile time-series of velocity and density at 14 stations allow estimation of the semidiurnal internal tide energy-flux radiating away from the ridge (red arrows). Model energy-flux hot and cold spots (blue arrows) are consistent with the observations though with some differences in magnitude and direction.

Ridge Generation

Baroclinic generation at the ridge crest

An intensive survey of the internal tide was undertaken on Kaena Ridge. Baroclinic generation occurs over the steep upper flanks and forms beams that propagate upward and downward from x=-10 km as shown above. Superposition of waves from the northern and southern flanks of the ridge (and travelling in opposite directions) produce a standing wave pattern over the ridge top. The standing wave has no net energy flux, but the interference of waves produces regions of high and low APE:HKE that correspond to the ridge flanks (high APE) and ridge top (low APE). Compare for example the APE:HKE at stations 2 (flank) and 5 (crest). Only at the furthest offshore station does the vertical structure of the energetics (its flux and partitioning of APE and HKE) approach that of a standing mode.

For more information see http://kai.coas.oregonstate.edu/home/

Evidence for propogation of rays away from the ridge

Repeated across-ridge sections using SeaSoar and shipboard Doppler sonar allowed phase averaging of the internal tide. The resulting velocity variance is high along paths consistent with internal-wave rays at the M2 frequency (top panel). The rays apparently emanate from both sides of the ridge crest, and then propagate over the top of the ridge (bottom panel). The observations show that the rays do not survive the first surface bounce, possibly caused by enhanced dissipation near the surface.

For more information see http://chowder.ucsd.edu

Wave-wave Interactions

Historic anaylsis of mooring data

Rotary power spectra of currents measured during the RTE87 experiment (Dushaw et al., 1995) at a location 1500km north of the Hawaiian chain. The record was 4 months long, from 5/17/87 to 9/18/87, and was acquired at a depth of 1159 m. Clockwise rotating currents are in blue; counter-clockwise currents are in red. Ordinate units are (cm/s)2/(rad/s). Note the familiar, and strongly clockwise, inertial and internal tide peaks (the latter centered at the frequency of the M2 tide). At a frequency equal to the difference between the M2 and inertial frequencies, there is a strong counter-clockwise motion which we believe is the result of a non-linear interaction (but not a simple advection) between the inertial and semi-diurnal internal tide waves. Data in this region has been reanalyzed for HOME as it is presumably in the 'beam' of baroclinic tidal energy emanating from the chain.

Tidal Dissipation

Integrated dissipation near the ridge

Three instruments were used to directly sample turbulent dissipation near the ridge, two vertical profilers and a towed body, over six cruises. A summary of the diffusivity (panel a) and dissipation (panel b) near Kauai Channel are shown above. Kauai Channel is a promontory between two islands where there is strong internal wave generation and energy. Turbulence was high near the bottom and over the ridge crest, and then decayed away from topography; by 60 km, values had fallen to the open-ocean values. An empirical relation between the tidal energy (E) and the vertical integral of dissipation (D~sqrt(E)) and Mark Merrifield's numerical modeling allowed a rough extrapolation along the ridge giving the total dissipation as 4+/-2 GW.

For more information see http://opg1.ucsd.edu/~jklymak/HomeResearch.html and http://kai.coas.oregonstate.edu/home/

Near-bottom dissipation

Moorings placed on the flank of the ridge give estimates of dissipation rates by looking at density inversions. The data above is taken from a mooring in 1450 m of water. Magnitude of cross-ridge tidal velocity (purple) and estimated epsilon (black) plotted as a function of time. The epsilon series is the vertical average from 1141 to 1445 m, tidal velocity is from the model of Egbert (1997) using 8 tidal constituents. Both time series are smoothed with a 25-h running mean. The log scale is set such that 1 decade in velocity is equal in length to 3 decades in epsilon.

Edge-effects

The loosely tethered profile AMP (Advanced Microstructure Profiler) was used to measure turbulent kinetic energy dissipation over the Kaena Ridge. Additional dissipations were inferred from density overturns recorded by the towed body SWIMS in shallow water near Oahu. Shown is the depth-average dissipation value from each profile. The largest dissipations occur in water shallower than 500 m, the largest values are associated with internal hydraulic jumps. Increased dissipation is also observed near the 400 m high semount (~ -158.65E, 21.73N).

Island-effects: Mamala Bay

Energy flux and dissipation in Mamala Bay. Arrows represent energy- flux vectors from the shipboard survey (red) and from moored and dropped measurements at other times (blue). The size of each circle indicates the total energy. The color of each circle indicates the depth-integrated dissipation rate (estimated from SWIMS II overturning scales), in W m-2. The scales are indicated at center top. The lower color scale indicates the depthmean dissipation rate in W kg-1 if the depth is 400 m. Flux exhibits a coherent but strongly time-dependent pattern resulting from a superposition of waves generated at remote sites to the east and west. Their propagation into the Bay, and the resulting interference pattern, is modulated by slowly-varying stratification along their route. A strong convergence in the western Bay is balanced by increased dissipation there, largely due to shear instability associated with a pronounced ridge.

For more information, see http://opd.apl.washington.edu/scistaff/bios/alford/alford3dplus.html#intern.

Implications

A simple geometrical scaling argument (Armi, 1978; Toole et al., 1997) used to infer basin-average diapycnal diffusivities from observed diffusivities is re-examined. The dependences upon the resolution of the bathymetry used and on the prescribed spatial structure of the near-boundary diffusivities are tested. It is found that (i) employing more accurate, high resolution bathymetry (ETOPO5 versus Smith & Sandwell here) and (ii) applying a diffusivity structure decaying exponentially in both vertical and horizontal directions away from the boundary (Exponential model), rather than on the vertical direction only (Step-function model), yield large increases in the estimated basin-average diapycnal diffusivities with the biggest impact coming from (ii). The figure shows estimated basin-average diffusivity profiles for the North Pacific basin.