# Search Results

## You are looking at 1 - 10 of 11 items for

- Author or Editor: Murray D. Levine x

- Refine by Access: All Content x

^{ }

## Abstract

The Garrett–Munk (GM) spectrum continues to be a useful description of the oceanic internal wave field. However, there are several inconsistencies and ambiguities that make it difficult to use in comparing internal wave fields at different latitudes, stratifications, and water depths. A modified spectral formulation is presented that treats three problems with the Garrett–Munk formulation: the normalization of the energy spectrum as a function of frequency bandwidth, the energy distribution at frequencies below the semidiurnal tide, and the treatment of vertical boundaries and turning points. Addressing these problems leads to a substitution of the GM parameters *E* (nondimensional energy), *b* (vertical length scale), and *N*
_{0} (buoyancy frequency scale) with two new dimensional scales: *E*
_{ref}, the energy per unit mass, and *D*(*ω*), the Wentzel–Kramers–Brillouin (WKB)-scaled thickness of the vertical waveguide. The advantages of the modified spectrum are illustrated by comparing with observations from the equator and the continental shelf.

## Abstract

The Garrett–Munk (GM) spectrum continues to be a useful description of the oceanic internal wave field. However, there are several inconsistencies and ambiguities that make it difficult to use in comparing internal wave fields at different latitudes, stratifications, and water depths. A modified spectral formulation is presented that treats three problems with the Garrett–Munk formulation: the normalization of the energy spectrum as a function of frequency bandwidth, the energy distribution at frequencies below the semidiurnal tide, and the treatment of vertical boundaries and turning points. Addressing these problems leads to a substitution of the GM parameters *E* (nondimensional energy), *b* (vertical length scale), and *N*
_{0} (buoyancy frequency scale) with two new dimensional scales: *E*
_{ref}, the energy per unit mass, and *D*(*ω*), the Wentzel–Kramers–Brillouin (WKB)-scaled thickness of the vertical waveguide. The advantages of the modified spectrum are illustrated by comparing with observations from the equator and the continental shelf.

^{ }

^{ }

## Abstract

We have developed a statistical model describing a random field of internal waves passively oscillating a random, locally horizontally uniform temperature finestructure. Here we define finestructure to be non-internal wavelike temperature fluctuations of *any* vertical scale caused by phenomena such as horizontal intrusions or geostrophic eddies. The model allows one to examine the effects of finestructure upon all relevant measurable statistical quantities of the temperature field as a function of the vertical scale of the finestructure. Also, the effects of the time variation of the finestructure itself are considered.

The model was fit to data obtained during the 3-week Mid-ocean Acoustic Transmission Experiment (MATE) in summer 1977 new Cobb seamount in the northeastern Pacific. Various spectra and coherences estimated from temperature time series and vertical profiles were used to make an assessment of the finestructure as well as establish the consistency of the model.

Temperature variance measured at frequencies above the local Väisäiä frequency was used to estimate the magnitude of the high vertical wavenumber finestructure. This high wavenumber finestructure (0.05–1 m^{−1}) with a (wavenumber)^{−−2.5} dependence can consistently explain all the observed variance in the measured high vertical wavenumber spectra. At lower wavenumbers (0.002–0.020 m^{−1}) a (wavenumber)^{−2} dependence was observed, and the spectral level found to be approximately equally divided between finestructure and internal waves advecting a constant temperature gradient. The contribution of the finestructure effects to the internal wave frequency spectrum was found to be about a factor of 10 less than that of internal waves advecting a constant temperature gradient.

## Abstract

We have developed a statistical model describing a random field of internal waves passively oscillating a random, locally horizontally uniform temperature finestructure. Here we define finestructure to be non-internal wavelike temperature fluctuations of *any* vertical scale caused by phenomena such as horizontal intrusions or geostrophic eddies. The model allows one to examine the effects of finestructure upon all relevant measurable statistical quantities of the temperature field as a function of the vertical scale of the finestructure. Also, the effects of the time variation of the finestructure itself are considered.

The model was fit to data obtained during the 3-week Mid-ocean Acoustic Transmission Experiment (MATE) in summer 1977 new Cobb seamount in the northeastern Pacific. Various spectra and coherences estimated from temperature time series and vertical profiles were used to make an assessment of the finestructure as well as establish the consistency of the model.

Temperature variance measured at frequencies above the local Väisäiä frequency was used to estimate the magnitude of the high vertical wavenumber finestructure. This high wavenumber finestructure (0.05–1 m^{−1}) with a (wavenumber)^{−−2.5} dependence can consistently explain all the observed variance in the measured high vertical wavenumber spectra. At lower wavenumbers (0.002–0.020 m^{−1}) a (wavenumber)^{−2} dependence was observed, and the spectral level found to be approximately equally divided between finestructure and internal waves advecting a constant temperature gradient. The contribution of the finestructure effects to the internal wave frequency spectrum was found to be about a factor of 10 less than that of internal waves advecting a constant temperature gradient.

^{ }

^{ }

## Abstract

Tidal mixing over a slope was explored using moored time series observations on Kaena Ridge extending northwest from Oahu, Hawaii, during the Survey component of the Hawaii Ocean Mixing Experiment (HOME). A mooring was instrumented to sample the velocity and density field of the lower 500 m of the water column to look for indirect evidence of tidally induced mixing and was deployed on a slope in 1453-m water depth for 2 months beginning in November 2000. The semidiurnal barotropic tidal currents at this site have a significant cross-ridge component, favorable for exciting an internal tidal response. A large-amplitude response is expected, given that the slope of the topography (4.5°) is nearly the same as the slope of the internal wave group velocity at semidiurnal frequency. Density overturns were inferred from temperature profiles measured every 2 min. The number and strength of the overturns are greater in the 200 m nearest the bottom, with overturns exceeding 24 m present at any depth nearly 10% of the time. Estimates of turbulent dissipation rate *ε* were made for each overturn by associating the measured Thorpe scale with the Ozmidov scale. The average *ε* between 1300 and 1450 m for the entire experiment is about 10^{−8} m^{2} s^{−3}, corresponding to an average *K _{ρ}
* of 10

^{−3}m

^{2}s

^{−1}. Both

*ε*and

*K*decrease by about an order of magnitude by 1200 m. The occurrence of overturns and the magnitude of

_{ρ}*ε*are both highly correlated with the tide: both with the spring–neap cycle as well as the phase of the semidiurnal tide itself. Dissipation rate varies by at least an order of magnitude over the spring–neap cycle. It appears that tidal frequency vertical shear within 200 m of the boundary leads to significant strain (vertical divergence). Most of the overturns occur during the few hours when the vertical strain is greatest. The buoyancy frequency

*N*calculated from reordering these overturns is a factor of 3 lower than the background

*. This is consistent with the following kinematic description: the internal tide first strains the mean density field, leading to regions of low*N

*N*that subsequently overturn. Less regularly, overturns also occur when the internal tide strain has created relatively high stratification within 200 m of the bottom.

## Abstract

Tidal mixing over a slope was explored using moored time series observations on Kaena Ridge extending northwest from Oahu, Hawaii, during the Survey component of the Hawaii Ocean Mixing Experiment (HOME). A mooring was instrumented to sample the velocity and density field of the lower 500 m of the water column to look for indirect evidence of tidally induced mixing and was deployed on a slope in 1453-m water depth for 2 months beginning in November 2000. The semidiurnal barotropic tidal currents at this site have a significant cross-ridge component, favorable for exciting an internal tidal response. A large-amplitude response is expected, given that the slope of the topography (4.5°) is nearly the same as the slope of the internal wave group velocity at semidiurnal frequency. Density overturns were inferred from temperature profiles measured every 2 min. The number and strength of the overturns are greater in the 200 m nearest the bottom, with overturns exceeding 24 m present at any depth nearly 10% of the time. Estimates of turbulent dissipation rate *ε* were made for each overturn by associating the measured Thorpe scale with the Ozmidov scale. The average *ε* between 1300 and 1450 m for the entire experiment is about 10^{−8} m^{2} s^{−3}, corresponding to an average *K _{ρ}
* of 10

^{−3}m

^{2}s

^{−1}. Both

*ε*and

*K*decrease by about an order of magnitude by 1200 m. The occurrence of overturns and the magnitude of

_{ρ}*ε*are both highly correlated with the tide: both with the spring–neap cycle as well as the phase of the semidiurnal tide itself. Dissipation rate varies by at least an order of magnitude over the spring–neap cycle. It appears that tidal frequency vertical shear within 200 m of the boundary leads to significant strain (vertical divergence). Most of the overturns occur during the few hours when the vertical strain is greatest. The buoyancy frequency

*N*calculated from reordering these overturns is a factor of 3 lower than the background

*. This is consistent with the following kinematic description: the internal tide first strains the mean density field, leading to regions of low*N

*N*that subsequently overturn. Less regularly, overturns also occur when the internal tide strain has created relatively high stratification within 200 m of the bottom.

^{ }

^{ }

## Abstract

Wind-generated inertial currents can radiate from the mixed layer as horizontally and vertically propagating new-inertial internal gravity waves. To study the timescale of the decay of mixed layer energy and the magnitude of the energy transfer to the ocean below, the authors developed a numerical, linear model on a β plane, using baroclinic modes to describe the velocity field. The model is unforced-wave propagation is initiated by specifying the mixed layer currents that would he generated by a moving atmospheric front. The numerical results are interpreted using concepts of modal interference and modal departure that can be evaluated analytically, thereby permitting predictions Of some features of wave field evolution without the need to run the numerical model. The energy exchange with the pycnocline and deep ocean is explored as a function of the propagation speed and direction of the front, the horizontal extent of the storm, and the background stratification.

The timescale of energy transfer from the mixed layer to the pycocline due to modal interference is greatly affected by the β effect, causing much faster energy transfer for currents generated by southward propagating fronts. The timescale is typically not a strong function of mixed layer depth; however. the magnitude of the energy transfer is. Besides modal interference, vertical energy propagation occurs when low modes leave the area- a possibility for storms of finite horizontal extent. The deep stratification and *f* also affect the timescale; climatological examples indicate faster wave evolution at low latitudes.

## Abstract

Wind-generated inertial currents can radiate from the mixed layer as horizontally and vertically propagating new-inertial internal gravity waves. To study the timescale of the decay of mixed layer energy and the magnitude of the energy transfer to the ocean below, the authors developed a numerical, linear model on a β plane, using baroclinic modes to describe the velocity field. The model is unforced-wave propagation is initiated by specifying the mixed layer currents that would he generated by a moving atmospheric front. The numerical results are interpreted using concepts of modal interference and modal departure that can be evaluated analytically, thereby permitting predictions Of some features of wave field evolution without the need to run the numerical model. The energy exchange with the pycnocline and deep ocean is explored as a function of the propagation speed and direction of the front, the horizontal extent of the storm, and the background stratification.

The timescale of energy transfer from the mixed layer to the pycocline due to modal interference is greatly affected by the β effect, causing much faster energy transfer for currents generated by southward propagating fronts. The timescale is typically not a strong function of mixed layer depth; however. the magnitude of the energy transfer is. Besides modal interference, vertical energy propagation occurs when low modes leave the area- a possibility for storms of finite horizontal extent. The deep stratification and *f* also affect the timescale; climatological examples indicate faster wave evolution at low latitudes.

^{ }

^{ }

## Abstract

Observations of near-inertial oscillations collected during the Ocean Storms Experiment in the northeast Pacific Ocean are compared with results from a linear, numerical model on a β plane, developed by Zervakis and Levine. A slab mixed layer model, forced by the observed wind time series, is used to identify three isolated events of local generation in October, January, and March for detailed analysis. Synoptic storm track maps are used to estimate the initial horizontal wavenumber of the mixed layer currents that are used as initial conditions to the modal. A comparison of the modal with the observed currents reveals some differences and similarities. Overall the January and March events are better represented by the model than the October event. The timescale of the initiation of vertical propagation of energy from the mixed layer occurs almost immediately in October rather than after 8 days in January and March-this difference cannot be explained by the model. The observed vertical and temporal structure indicates that the near-Inertial energy propagated as a “beam” of energy through the pycnocline, especially in October. In the model the wave energy appears to accumulate at the top of the pycnocline. Physical processes that might be responsible for the deficiency of the model are discussed.

## Abstract

Observations of near-inertial oscillations collected during the Ocean Storms Experiment in the northeast Pacific Ocean are compared with results from a linear, numerical model on a β plane, developed by Zervakis and Levine. A slab mixed layer model, forced by the observed wind time series, is used to identify three isolated events of local generation in October, January, and March for detailed analysis. Synoptic storm track maps are used to estimate the initial horizontal wavenumber of the mixed layer currents that are used as initial conditions to the modal. A comparison of the modal with the observed currents reveals some differences and similarities. Overall the January and March events are better represented by the model than the October event. The timescale of the initiation of vertical propagation of energy from the mixed layer occurs almost immediately in October rather than after 8 days in January and March-this difference cannot be explained by the model. The observed vertical and temporal structure indicates that the near-Inertial energy propagated as a “beam” of energy through the pycnocline, especially in October. In the model the wave energy appears to accumulate at the top of the pycnocline. Physical processes that might be responsible for the deficiency of the model are discussed.

^{ }

^{ }

^{ }

## Abstract

An error in the calculation of the baroclinic pressure gradient term in the Princeton Ocean Model (POM) was identified while modeling the M_{2} tidal current near its critical latitude in the southern Weddell Sea. The error arises from the present calculation of density, which involves the subtraction of a background density profile from the density field calculated at each internal time step. The small displacement of sigma surface depths relative to the surface, as surface elevation changes, causes a slight error in the calculation of the vertical and horizontal gradients of potential density. The error is largest at the seabed over rapidly changing bathymetry such as the continental slope. The baroclinic pressure gradient error is typically much smaller than the Coriolis term in the momentum equations and, therefore, usually unimportant. Close to the critical latitude, however, near-resonance between the error and Coriolis terms can cause an energetic and spatially complex spurious inertial mode to develop. The error is significant when modeling tides near their critical latitudes, and will contribute to the error in the baroclinic pressure gradient in other simulations. Two methods were suggested for fixing this problem. The preferred method was tested by applying the new form of POM to the southern Weddell Sea. The new results are consistent with both current meter data and predictions of linear internal wave theory.

## Abstract

An error in the calculation of the baroclinic pressure gradient term in the Princeton Ocean Model (POM) was identified while modeling the M_{2} tidal current near its critical latitude in the southern Weddell Sea. The error arises from the present calculation of density, which involves the subtraction of a background density profile from the density field calculated at each internal time step. The small displacement of sigma surface depths relative to the surface, as surface elevation changes, causes a slight error in the calculation of the vertical and horizontal gradients of potential density. The error is largest at the seabed over rapidly changing bathymetry such as the continental slope. The baroclinic pressure gradient error is typically much smaller than the Coriolis term in the momentum equations and, therefore, usually unimportant. Close to the critical latitude, however, near-resonance between the error and Coriolis terms can cause an energetic and spatially complex spurious inertial mode to develop. The error is significant when modeling tides near their critical latitudes, and will contribute to the error in the baroclinic pressure gradient in other simulations. Two methods were suggested for fixing this problem. The preferred method was tested by applying the new form of POM to the southern Weddell Sea. The new results are consistent with both current meter data and predictions of linear internal wave theory.

^{ }

^{ }

^{ }

## Abstract

A thermistor chain was moored below the pack ice from 50–150 m in the Arctic Ocean for five days in 1981. Oscillations in temperature are attributed to the vertical dispalcement of internal waves. The spectral shape of isotherm dispalcement is consistent with the Garrett-Munk model and other internal wave observations, but the spectral level is significantly lower. Other observations from the Arctic Ocean also exhibit lower internal-wave energy when compared with historical data from lower latitudes. The lower energy may be related to the unique generation and dissipation mechanisms present in the ice-covered Arctic Ocean. Significant peaks in vertical coherence occur at 0.81 and 2.6 cph. The peak at 2.6 cph coincides approximately with the high-frequency spectral cutoff near the local buoyancy frequency; this feature has been observed in many other internal wave experiments. The coherent oscillations at 0.81 cph exhibit a node in vertical dispalcement at 75–100 m. This is consistent with either the second, third or fourth vertical mode calculated from the mean buoyancy frequency profile. Evidence is presented which suggests that, contrary to the Garrett-Munk model, the frequency spectrum does not scale with the Coriolis parameter.

## Abstract

A thermistor chain was moored below the pack ice from 50–150 m in the Arctic Ocean for five days in 1981. Oscillations in temperature are attributed to the vertical dispalcement of internal waves. The spectral shape of isotherm dispalcement is consistent with the Garrett-Munk model and other internal wave observations, but the spectral level is significantly lower. Other observations from the Arctic Ocean also exhibit lower internal-wave energy when compared with historical data from lower latitudes. The lower energy may be related to the unique generation and dissipation mechanisms present in the ice-covered Arctic Ocean. Significant peaks in vertical coherence occur at 0.81 and 2.6 cph. The peak at 2.6 cph coincides approximately with the high-frequency spectral cutoff near the local buoyancy frequency; this feature has been observed in many other internal wave experiments. The coherent oscillations at 0.81 cph exhibit a node in vertical dispalcement at 75–100 m. This is consistent with either the second, third or fourth vertical mode calculated from the mean buoyancy frequency profile. Evidence is presented which suggests that, contrary to the Garrett-Munk model, the frequency spectrum does not scale with the Coriolis parameter.

^{ }

^{ }

^{ }

## Abstract

We describe the spectral analysis of temperature and velocity measurements made in the northeast Pacific as part of the Mixed Layer Experiment (MILE) and attempt to relate the observed fluctuations to internal-wave models of the upper ocean. From the inertial frequency to 1 cph there is good agreement between these upper-ocean data and typical deep-ocean observations as described by the WKB-scaled Garrett-Munk model. The largest deviations from the Garrett-Munk model occur in the vertical-displacement field at high frequency, 1–5 cph, where there is a spectral peak or shoulder and high vertical coherence. These high-frequency features in vertical displacement are successfully modeled using a few standing modes and un-correlated noise, though the velocity spectra are poorly modeled—probably because of contamination by mooring motion. There are significant temporal fluctuations of the high-frequency energy that are not correlated with the local winds but are perhaps associated with the advection of an eddy-like feature.

## Abstract

We describe the spectral analysis of temperature and velocity measurements made in the northeast Pacific as part of the Mixed Layer Experiment (MILE) and attempt to relate the observed fluctuations to internal-wave models of the upper ocean. From the inertial frequency to 1 cph there is good agreement between these upper-ocean data and typical deep-ocean observations as described by the WKB-scaled Garrett-Munk model. The largest deviations from the Garrett-Munk model occur in the vertical-displacement field at high frequency, 1–5 cph, where there is a spectral peak or shoulder and high vertical coherence. These high-frequency features in vertical displacement are successfully modeled using a few standing modes and un-correlated noise, though the velocity spectra are poorly modeled—probably because of contamination by mooring motion. There are significant temporal fluctuations of the high-frequency energy that are not correlated with the local winds but are perhaps associated with the advection of an eddy-like feature.

^{ }

^{ }

^{ }

^{ }

^{ }

^{ }

## Abstract

Results are presented from two dye release experiments conducted in the seasonal thermocline of the Sargasso Sea, one in a region of low horizontal strain rate (~10^{−6} s^{−1}), the second in a region of intermediate horizontal strain rate (~10^{−5} s^{−1}). Both experiments lasted ~6 days, covering spatial scales of 1–10 and 1–50 km for the low and intermediate strain rate regimes, respectively. Diapycnal diffusivities estimated from the two experiments were *κ*
_{
z
} = (2–5) × 10^{−6} m^{2} s^{−1}, while isopycnal diffusivities were *κ*
_{
H
} = (0.2–3) m^{2} s^{−1}, with the range in *κ*
_{
H
} being less a reflection of site-to-site variability, and more due to uncertainties in the background strain rate acting on the patch combined with uncertain time dependence. The Site I (low strain) experiment exhibited minimal stretching, elongating to approximately 10 km over 6 days while maintaining a width of ~5 km, and with a notable vertical tilt in the meridional direction. By contrast, the Site II (intermediate strain) experiment exhibited significant stretching, elongating to more than 50 km in length and advecting more than 150 km while still maintaining a width of order 3–5 km. Early surveys from both experiments showed patchy distributions indicative of small-scale stirring at scales of order a few hundred meters. Later surveys show relatively smooth, coherent distributions with only occasional patchiness, suggestive of a diffusive rather than stirring process at the scales of the now larger patches. Together the two experiments provide important clues as to the rates and underlying processes driving diapycnal and isopycnal mixing at these scales.

## Abstract

Results are presented from two dye release experiments conducted in the seasonal thermocline of the Sargasso Sea, one in a region of low horizontal strain rate (~10^{−6} s^{−1}), the second in a region of intermediate horizontal strain rate (~10^{−5} s^{−1}). Both experiments lasted ~6 days, covering spatial scales of 1–10 and 1–50 km for the low and intermediate strain rate regimes, respectively. Diapycnal diffusivities estimated from the two experiments were *κ*
_{
z
} = (2–5) × 10^{−6} m^{2} s^{−1}, while isopycnal diffusivities were *κ*
_{
H
} = (0.2–3) m^{2} s^{−1}, with the range in *κ*
_{
H
} being less a reflection of site-to-site variability, and more due to uncertainties in the background strain rate acting on the patch combined with uncertain time dependence. The Site I (low strain) experiment exhibited minimal stretching, elongating to approximately 10 km over 6 days while maintaining a width of ~5 km, and with a notable vertical tilt in the meridional direction. By contrast, the Site II (intermediate strain) experiment exhibited significant stretching, elongating to more than 50 km in length and advecting more than 150 km while still maintaining a width of order 3–5 km. Early surveys from both experiments showed patchy distributions indicative of small-scale stirring at scales of order a few hundred meters. Later surveys show relatively smooth, coherent distributions with only occasional patchiness, suggestive of a diffusive rather than stirring process at the scales of the now larger patches. Together the two experiments provide important clues as to the rates and underlying processes driving diapycnal and isopycnal mixing at these scales.

^{ }

^{ }

^{ }

^{ }

^{ }

^{ }

## Abstract

A strong, isolated October storm generated 0.35–0.7 m s^{−1} inertia] frequency currents in the 40-m deep mixed layer of a 300 km×300 km region of the northeast Pacific Ocean. The authors describe the evolution of these currents and the background flow in which they evolve for nearly a month following the storm. Instruments included CTD profilers, 36 surface drifters, an array of 7 moorings, and air-deployed velocity profilers. The authors then test whether the theory of linear internal waves propagating in a homogeneous ocean can explain the observed evolution of the inertial frequency currents.

The subinertial frequency flow is weak, with typical currents of 5 cm s^{−1}, and steady over the period of interest. The storm generates inertial frequency currents in and somewhat below the mixed layer with a horizontal scale much larger than the Rossby radius of deformation, reflecting the large-scale and rapid translation speed of the storm. This scale is too large for significant linear propagation of the inertial currents to occur. It steadily decreases owing to the latitudinal variation in *f*, that is, β, until after about 10 days it becomes sufficiently small for wave propagation to occur. Inertial energy then spreads downward from the mixed layer, decreasing the mixed layer inertial energy and increasing the inertial energy below the mixed layer. A strong maximum in inertial energy is formed at 100 m ("the Beam"). By 21 days after the storm. both mixed layer inertial energy and inertial frequency shear maximum just below the mixed layer have been reduced to background levels. The total depth-average inertial energy decreases by about 40% during this period.

Linear internal wave theory can only partially explain the observed evolution of the inertial frequency currents. The decrease in horizontal wavelength is accurately predicted as due to the β effect. The decrease in depth-average inertial energy is explained by southward propagation of the lowest few modes. The superinertial frequency and clockwise rotation of phase with depth are qualitatively consistent with linear theory. However, linear theory underpredicts the initial rate at which inertial energy is lost from the mixed layer by 20%–50% and cannot explain the decrease of mixed layer energy and shear to background levels in 21 days.

## Abstract

A strong, isolated October storm generated 0.35–0.7 m s^{−1} inertia] frequency currents in the 40-m deep mixed layer of a 300 km×300 km region of the northeast Pacific Ocean. The authors describe the evolution of these currents and the background flow in which they evolve for nearly a month following the storm. Instruments included CTD profilers, 36 surface drifters, an array of 7 moorings, and air-deployed velocity profilers. The authors then test whether the theory of linear internal waves propagating in a homogeneous ocean can explain the observed evolution of the inertial frequency currents.

The subinertial frequency flow is weak, with typical currents of 5 cm s^{−1}, and steady over the period of interest. The storm generates inertial frequency currents in and somewhat below the mixed layer with a horizontal scale much larger than the Rossby radius of deformation, reflecting the large-scale and rapid translation speed of the storm. This scale is too large for significant linear propagation of the inertial currents to occur. It steadily decreases owing to the latitudinal variation in *f*, that is, β, until after about 10 days it becomes sufficiently small for wave propagation to occur. Inertial energy then spreads downward from the mixed layer, decreasing the mixed layer inertial energy and increasing the inertial energy below the mixed layer. A strong maximum in inertial energy is formed at 100 m ("the Beam"). By 21 days after the storm. both mixed layer inertial energy and inertial frequency shear maximum just below the mixed layer have been reduced to background levels. The total depth-average inertial energy decreases by about 40% during this period.

Linear internal wave theory can only partially explain the observed evolution of the inertial frequency currents. The decrease in horizontal wavelength is accurately predicted as due to the β effect. The decrease in depth-average inertial energy is explained by southward propagation of the lowest few modes. The superinertial frequency and clockwise rotation of phase with depth are qualitatively consistent with linear theory. However, linear theory underpredicts the initial rate at which inertial energy is lost from the mixed layer by 20%–50% and cannot explain the decrease of mixed layer energy and shear to background levels in 21 days.