HOME Experiment

It is now well established from various in situ and remote observations that energetic semidiurnal internal tides originate from the Hawaiian Ridge (Dushaw et al., 1995; Ray and Mitchum, 1996). The Hawaii Ocean Mixing Experiment (HOME) was designed to investigate the details of the tidal mixing and dissipation along the Hawaiian chain (Rudnick et al., 2003). As part of the HOME modeling effort, the Princeton Ocean Model (POM) has been used to simulate barotropic to baroclinic tidal energy conversion along Kaena Ridge in the Kauai Channel (figure 1.3). The simulations show internal tide energy originating on each side of the ridge near the nearly symmetric crest (Merrifield and Holloway, 2002). The energy propagates along internal wave characteristics; one beam propagates up and away from the ridge, a second propagates up and over the ridge, and a third propagates downward along the ridge flanks where there is either a near- or super-critical slope below the generation site (Figure 2.1). Observations from different instruments and platforms made during the HOME nearfield phase confirm this beam structure (Nash et al., 2005; Martin et al., 2005; Rainville and Pinkel, 2005).

Mechanical energy associated with this down-going beam as it propagates along the ridge flanks is available for turbulent mixing at the slope. These down-going beams along the north and south flanks of the ridge are of central importance to the present study. The numerical model (Merrifield and Holloway, 2002) indicates an $M_2$ baroclinic energy flux of $\sim10 GW$ radiating away from the entire Hawaiian Ridge. In their model, they use a bottom drag friction parametrization to account for the loss of energy of the combined barotropic and baroclinic tide at the bottom boundary. This parametrization accounts only for the turbulent dissipation created by shear stress at the boundary. They estimated that only $0.6 GW$ of $M_2$ tidal energy is lost by frictional dissipation at the boundary, small compared to the $\sim10 GW$ of internal tidal energy radiated away.

Klymak et al. (2005), using direct dissipation measurements from 4 different instruments, estimate $3 \pm 1.5 GW$ of turbulent dissipation along the entire Hawaiian ridge. This estimate is based on rather sparse data collection around the Hawaiian Ridge system, and the authors acknowledge the fact that geographical variations of mixing intensity may bias their estimates one way or another.

To complement the sparse temporal resolution of HOME microstructure measurements near the bottom, deep moorings were deployed near the ridge top and over the flanks of Kaena Ridge, between the islands of Oahu and Kauai. These moorings were designed to sample temperature and current over the first hundreds of meters above the bottom. Indirect estimates of dissipation can be obtained through detection and analysis of turbulent overturns (Levine and Boyd, 2005).

Levine and Boyd (2005) deployed such a mooring near the ridge top ($1450m$ ) on the northern flank (Big Boy mooring, figure 1.3) and found significant overturns ( $\sim150 m$ ) within $300 m$ from the bottom that are linked to semidiurnal tidal phase and amplitude. Levine and Boyd (2005) describe a scenario where the internal tide first strains the mean density field, leading to regions of low N that subsequently overturn. The phase of the tide when overturns occur varies with depth. They also found less frequent overturns occurring when the stratification is high, similar to the type of mixing events discussed in chapter 6 of this dissertation. Levine and Boyd (2005) found average dissipation levels of $\sim10^{-8}Wkg^{-1}$ in the $150 m$ thick layer for the entire experiment.

Rainville and Pinkel (2005) and Carter and Gregg (2005) observed high levels of near-diurnal internal wave energy at the nearby summit of Kaena Ridge. This diurnal energy flux was observed to vary more with the astronomical semidiurnal forcing than with the diurnal forcing. Parametric Subharmonic Instability (PSI) was proposed as a mechanism to transfer energy from the semidiurnal $M_2$ frequency to the near-diurnal $M_2/2$ frequency, but the exact mechanism describing this transfer still remains unclear (Rainville and Pinkel, 2005).