## 1. Introduction

Internal gravity waves play an important role in the energy cycle of ocean motions. When they break, the wave field provides energy to small-scale turbulence in the interior of the ocean, which drives density mixing and increases the potential energy of the large-scale circulation (Wunsch and Ferrari 2004). This process is thought to be of importance for the global meridional overturning circulation. Prominent sources energizing the oceanic internal wave field are near-inertial waves excited by wind stress fluctuations at the surface (Alford 2001; Rimac et al. 2013) and the scattering of the barotropic tide into baroclinic waves at topography (Nycander 2005; Falahat et al. 2014). Several other generation processes over a broad range of frequencies have been discussed (see, e.g., Olbers 1983) also involving interactions with the geostrophically balanced circulation, as, for example, through the dissipation of mesoscale eddies by spontaneous wave emission or other processes (Ford et al. 2000; Molemaker et al. 2010; Tandon and Garrett 1996; Eden and Greatbatch 2008; Brüggemann and Eden 2015) and the generation of lee waves by large-scale currents or mesoscale eddies flowing over topography (e.g., Nikurashin and Ferrari 2011; Scott et al. 2011; Trossman et al. 2016).

Whatever their generation mechanism, internal gravity waves can interact with the mean flow. When waves are propagating in a vertically sheared mean flow, they exchange energy with the mean flow, provided there is dissipation and forcing of the wave amplitude. The effect is generally called wave–mean flow interaction and, specifically in the atmospheric literature, gravity wave drag. The process is of importance for the dynamics of the upper atmosphere (e.g., Miller et al. 1989; Fritts and Alexander 2003; Muraschko et al. 2015), and the direction of the energy exchange can be from the mean flow to the waves or vice versa. Furthermore, the deformation of the wave paths by the mean flow can lead to critical layers where the wave energy is transferred to small-scale turbulence. A basic overview of the wave driving of the middle atmosphere is given in Becker (2012), and an overview of the gravity wave schemes currently used in global atmospheric models is given in Alexander et al. (2010).

In the ocean, gravity wave drag is less often considered, but it is the focus of the present study. While early theoretical considerations have predicted rather large wave-induced viscosities (Müller 1976), the observational estimates in the ocean (e.g., Frankignoul and Joyce 1979; Ruddick and Joyce 1979; Brown and Owens 1981) suggest a small effect. Recent reviews by Polzin (2008, 2010) confirm this finding. An exception is the wave–mean flow interactions in the deep equatorial jets studied by Muench and Kunze (1999, 2000). In their model, the jets are driven by the divergence of wave-induced momentum flux, created by internal waves that encounter critical layers.

Müller (1976) considers the concept of a wave-induced viscosity, which appears attractive and logical when appealing to the corresponding problem of mean transport processes induced by particle propagation and collisions in a Boltzmann gas (see, e.g., Huang 1987). The interaction of a small-scale random internal gravity wave field with a large-scale mean flow leads to a perturbation of the wave spectrum that is vertically asymmetric and anisotropic. This part is proportional to the mean shear. Müller (1976) assumes that nonlinear wave–wave interactions relax the actual perturbed energy spectrum toward vertical symmetry of upward and downward propagating energy and toward horizontal isotropy. The theory by Müller (1976) predicts very large viscosities, inconsistent with the observational estimates. The reason is that the waves do not interact with the same shear for a long time but are prone to reflection at the top and bottom of the ocean. Waves that, for example, gain energy from the shear flow on the way up will lose the same amount after reflection. This is a fundamental difference from the atmospheric setting of wave drag, where reflection at the top is missing and which seems to be more aligned with Müller’s theory.^{1} Still, despite small viscosities on average in the ocean and thus a small effect on the large-scale circulation, gravity wave drag may become an important source of oceanic internal gravity waves, as suggested by Polzin (2010). Since energy can be transferred by gravity wave drag from the geostrophically balanced flow to waves, this process may also be of importance for the dissipation of balanced flow.

Here, we revisit the topic of internal wave–mean flow interaction in the ocean and develop a simple closure of the process for use in ocean circulation models. In section 2, we discuss single internal gravity waves propagating in a horizontal mean flow with vertical shear in the Wentzel–Kramers–Brillouin (WKB) approximation (see, e.g., Olbers et al. 2012), to point out some principal properties of the wave–mean flow interaction in terms of energy transfers. Section 3 revisits the continuous description of a wave field using the radiative transfer equation including wave–mean flow interaction, which is used in section 4 to derive integrated energy compartments for the interaction with a unidirectional flow. Section 5 details simple closures for wave–wave interactions and dissipation for these compartments, which are used in section 6 for numerical integrations. These exemplify the effects of wave–wave interactions and dissipation on wave–mean flow interaction in some important idealized cases. The last section provides a short summary and discussion. A companion paper (Eden and Olbers 2017, hereinafter Part II) will deal with arbitrary configurations of the mean flow and the wave effects in the momentum equation.

## 2. Single waves

^{2}mean flow

**U**(

*z*,

*t*) with a vertical shear and a stability frequency

*N*(

*z*), with horizontal wavenumber vector

**k**, vertical wavenumber

*m*, and intrinsic (non-Doppler shifted) frequency

*ω*. Assuming that the vertical wavelength 2

*π*/

*m*and the wave period 2

*π*/

*ω*of the wave remain small compared to the vertical and temporal scales of

**U**and

*N*, respectively, the WKB approximation can be applied. The intrinsic frequency then obeys the local intrinsic dispersion relation

*ω*≡ Ω(

**k**,

*m*,

*z*) for internal gravity waves, which may be expressed in the formwhere

*k*= |

**k**| and

*σ*= −sign(

*m*) is positive for upward propagating waves (

*m*< 0) and negative for downward propagating waves (

*m*> 0). The (Doppler shifted) frequency of encounter then satisfies

**x**,

*z*) is given by the lateral and vertical group velocity,where

*ψ*(

*z*,

*t*) (e.g., Olbers et al. 2012). The intrinsic frequency

*ω*and the vertical wavenumber

*m*are changing because of a shear of the mean flow

**U**. Since both

*N*and

**U**are assumed to lack any lateral dependency, the horizontal wavenumber stays constant,

*E*of the wave. Under WKB conditions the wave action

*A*=

*E*/

*ω*is conserved (Bretherton and Garrett 1968; Whitham 1970; Lighthill 1978), which meansThe energy

*E*of the wave is not conserved, but there is exchange with the energy of the mean flow. To concentrate on this interaction of the wave with the mean flow, we assume for the rest of this section that

**k**=

*k*(cos

*ϕ*, sin

*ϕ*). Here

*ϕ*is the direction of the wave vector. The time scale

*τ*

_{C}characterizes the energy exchange of the wave with the mean flow and both signs

*σ*and

*γ*determine its direction. For a shear representative of the interior ocean

*τ*

_{C}is of order of hours for high-frequency waves with

*ω*/

*N*=

*O*(1), while it is much larger for near-inertial waves. Note that

*τ*

_{C}changes along the ray because

*ω*,

*c*, and

*C*change. The direction

*ϕ*of the horizontal wave path remains constant because of the assumption of lateral homogeneity.

From the second relation in Eq. (5), it follows that the intrinsic frequency *ω* is decreasing for *σγ* > 0. Under this condition *ω* may approach the Coriolis frequency *f*, and ultimately, the wave proceeds into a critical layer where breaking must occur because the vertical wavelength reduces dramatically. Conservation of action tells us that the initial action *E*_{0}/*ω*_{0} = *E*/*ω* is conserved so the amount *E*_{0}*f*/*ω*_{0} of the initial energy must be transferred to turbulence by breaking. In the following discussion, however, we ignore critical layer effects for simplicity.

*σ*, that is, on the upward or downward propagation direction of the wave, and on the sign

*γ*, that is, on the direction of the lateral wave propagation with respect to the mean flow shear. Let

*γ*, the energy and frequency

*E*

^{+}and

*ω*

^{+}of an upward propagating wave are decreasing, such that the wave loses energy to the mean flow. A downward propagating wave, on the other hand, receives energy

*E*

^{−}from the mean flow for positive

*γ*and its frequency

*ω*

^{−}also increases. If

*γ*is negative,

*E*

^{+}and

*ω*

^{+}increase and

*E*

^{−}and

*ω*

^{−}decrease, that is, the upward propagating wave receives and the downward propagating wave loses energy to the mean flow. The process is thus completely symmetric (in absence of any dissipative effects).

In the ocean, we can expect wave reflection at the surface and the bottom, such that an upward propagating wave that has received (or lost) energy from (to) the mean flow will become after reflection a downward propagating wave, which will lose (receive) the same amount of energy to (from) the mean flow, returning to the previous level, since the horizontal wave propagation and thus *γ* has not changed during the reflection process. The wave–mean flow interaction is thus reversible without any net effect, as long as no other dissipative or diabatic processes, such as wave breaking due to gravitational or shear instability, become important and convert the energy to a different form before the process reverses (and as long as we do not include horizontal anisotropies, which makes the situation more complicated). This is an example of the well-known nonacceleration theorem (often called noninteraction theorem) by Boyd (1976) and Andrews and McIntyre (1976). Note that in the atmosphere, surface reflection and the related symmetrization is missing and the wave–mean flow interaction may lead to the familiar gravity wave drag with significant importance for the large-scale circulation.

Because of the reflection at the surface and the bottom, a gravity wave field with a large extent of symmetry in upward or downward propagation direction is observed in the ocean (Garrett and Munk 1975; Munk 1981). Consider a couple of waves that are initially symmetric in *σ*, that is, *E*^{+} = *E*^{−}. The total energy *E* = *E*^{+} + *E*^{−} is then initially conserved since *E* = *E*^{+} − *E*^{−}, which vanishes initially. Any asymmetry in the wave field given by Δ*E* is thus indicative of an energy exchange with the mean flow. The changes of the asymmetry Δ*E* are given by *γ*, the asymmetry Δ*E* will become negative and the total energy *E* will increase, while for negative *γ*, the asymmetry Δ*E* will become positive and the total energy *E* will also increase. The energy exchange is thus directed from mean flow to the waves for both signs of *γ*. However, after reflection of the waves the roles of *E*^{+} and *E*^{−} are exchanged, and for both signs of *γ* the asymmetry will decay and the waves again lose energy to the mean flow. Because of the reflection, the process is thus completely reversible even in the presence of asymmetries in the wave field as long as no diabatic or dissipative processes are involved. Such processes are thus essential for a net energy transfer due to wave–mean flow interaction.

## 3. Continuous wave field

Consider now an ensemble of weakly interacting gravity waves instead of a single wave. The wave field is described by its energy spectrum **k**, *m*) and physical space (**x**, *z*) at time *t*. The quantity *d***k***dm* centered at the wavenumber vector (**k**, *m*) in wavenumber space. In the same way, the wave action spectrum

*m*,

*ω*, and

*ϕ*, where

*ω*is the intrinsic frequency and

*ϕ*denotes the wavenumber angle with

**k**=

*k*(cos

*ϕ*, sin

*ϕ*). It is in fact convenient to use the space

**k**,

*m*) in the current discussion. A governing equation for the wave field—the radiation balance or radiative transfer equation—can be formulated for the action spectrum

*S*a source that represents all processes affecting the wave field, except for propagation and refraction. Such processes contain wave–wave interactions and dissipation of waves by gravitational or shear instability but also the forcing.

*ϕ*still figures as coordinate and continues to be of importance for the wave–mean flow interaction, as we have seen in the discussion of single wave packets. A way to treat lateral inhomogeneities in

_{t}Ω = 0 as before, that is, that the stability frequency

*N*stays constant in time.

## 4. Energy compartments for unidirectional flow

*ω*and separately over negative and positive vertical wavenumbers

*m*, but not yet over the wavenumber direction

*ϕ*, since the direction of wave propagation was shown above to be important for the sign of the interaction with the mean flow. The integration in negative and positive

*m*and all

*ω*of Eq. (9) yieldswith the energy compartments

*m*< 0) and downward (

*m*> 0) propagating waves given byHere

*m*

_{l}denotes a low wavenumber cutoff. As in Olbers and Eden (2013), the integral of the second term on the left hand side of Eq. (9) was evaluated assuming a factorized form of the internal gravity wave spectrum

*A*and

*B*given by Cairns and Williams (1976) and Munk (1981) [which is the Garrett–Munk (GM76) spectrum], which are supposed to be identical for

*σ*= ±1. This leads tousing

*k*needs to be replaced in the integral with the dispersion relation Eq. (1). The integral of the first term on the right-hand side of Eq. (9) was evaluated using the same factorized spectrum for

*c*

_{0}and Λ can be readily evaluated using the functions

*A*(

*m*) and

*B*(

*m*) given by Cairns and Williams (1976) and Munk (1981) and become positive functions of

*N*/

*f*. The mean vertical group velocity

*c*

_{0}is given in Olbers and Eden (2013) and Λ is given in the appendix B.

The Eqs. (10) require knowledge of the processes described by *S*. Olbers and Eden (2013) discuss their effect in terms of energy compartments that have also been integrated in *ϕ*. The wave–wave interactions must conserve total energy; therefore, this nonlinear effect vanishes in the corresponding equation for *τ*_{υ} on the order of days. This closure, together with one for the dissipation of gravity waves, was key to construct a simple model for the propagation, interaction, and dissipation of gravity waves in the ocean, and we will use this concept here as well and extend it with the wave–mean flow interaction.

**U**= (

*U*, 0) and

*C*= ∂

_{z}

*U*cos

*ϕ*. Suppose that ∂

_{z}

*U*> 0. We have learned from the discussion of single wave packets and also see from Eq. (10) that

*π*/2 <

*ϕ*<

*π*/2 (

*γ*> 0) and increases in the other half-space (

*γ*< 0). The opposite is true for

_{z}

*U*< 0 the relations for

*ϕ*aswhere

*ϕ*in the same way, which yieldsandWe have defined by

*e*) and westward (

*w*) directions, as given by Eq. (14), for all frequencies and the negative and positive vertical wavenumber domains, as in Eq. (11). It now becomes necessary to form equations for

Note that critical layers are not considered in the present approach. The waves are assumed to freely propagate through the water column to the surface or the bottom, except for the WKB interaction with the mean flow. However, when parts of the wave field approach critical layers, a fraction of the fluxes

## 5. Closures for wave–wave interactions and dissipation

*ωS*in an analogous way as in Olbers and Eden (2013), it is convenient to consider, instead of

*S*=

*S*

_{ww}+

*S*

_{diss}needs to be decomposed into wave–wave interactions

*S*

_{ww}and dissipation

*S*

_{diss}by gravitational or shear instability. Wave–wave interactions conserve total energy, and it is assumed that this is also true when integrated over a single half-space in

*ϕ*. Note that this would hold for a horizontally isotropic spectrum such as the GM76 spectrum. Therefore, the corresponding integrals of

*S*

_{ww}vanish in the equations for

*E*

_{e}and

*E*

_{w}. The dissipation

*S*

_{diss}is assumed to be symmetric in upward or downward propagating waves and will therefore vanish in the equations for Δ

*E*

_{e}and Δ

*E*

_{w}. This yieldsFollowing Olbers and Eden (2013), the effect of the wave–wave interactions are parameterized as linear damping of Δ

*E*

_{e}and Δ

*E*

_{w}, respectively,with a time scale

*τ*

_{υ}= 1/

*α*

_{υ}on the order of days. The parameter

*α*

_{υ}is discussed in more detail in Olbers and Eden (2013). Its order of magnitude can be motivated from the Langevin rate in the scattering integral that results from

*S*

_{ww}using the weak interaction assumption for the relaxation of a spectral distortion in the GM76 spectrum. As in Olbers and Eden (2013),

*α*

_{υ}is taken constant here for simplicity. The choice to exchange energy between

*E*=

*E*

_{e}+

*E*

_{w},with a function

*μ*=

*μ*(

*f*/

*N*). The choice of

*μ*for Internal Wave Dissipation, Energetics and Mixing (IDEMIX) is discussed in detail in Olbers and Eden (2013). The function is found by analytical means by McComas and Müller (1981) and by Henyey et al. (1986) using an Eikonal approach, and is also used to estimate gravity wave dissipation from finescale observations (Gregg 1989; Polzin et al. 1995; Sun et al. 1999). Adding the expression for

*E*

_{e}and

*E*

_{w}yields the corresponding expression for dissipation

*μE*

^{2}. For ∂

_{z}

*U*= 0, that is, ignoring the wave–mean flow interaction, and adding

*E*=

*E*

_{e}+

*E*

_{w}and Δ

*E*= Δ

*E*

_{e}+ Δ

*E*

_{w}, the resulting model is identical to the one in Olbers and Eden (2013).

## 6. Numerical integrations

Figure 1 shows *N*(*z*) and *U*(*z*) profiles that we use to integrate Eqs. (19) numerically. The Coriolis frequency is set to *f* = 7.3 × 10^{−5} s^{−1}. The stability frequency *N* is an exponential function with a decay scale of 800 m, the mean flow *U* has amplitude 0.5 m s^{−1}, and a shear zone is centered at 500 m with a vertical scale of 200 m. The mean vertical group velocity *c*_{0} and the parameter Λ are functions of *N*/*f*; both have maxima at depth where *N* becomes small. The dissipation parameter *μ*, which is also a function of *N*/*f*, on the other hand, decreases with depth.

Consider a situation without dissipation and forcing. We also start without mean flow. It is then appropriate to use the energy compartment *ϕ*, and similarly for the energy compartment *E*^{+} and *E*^{−} for the case without wave–wave interactions, that is, for *α*_{υ} = 1/*τ*_{υ} = 0. Both compartments are initialized as a Gaussian profile centered at 1200 m depth with a decay scale of 200 m with amplitude of 10^{−3} m^{2} s^{−2}. As expected, *E*^{+} propagates initially upward, while *E*^{−} propagates downward. At the surface (bottom) *E*^{+} (*E*^{−}) is reflected into *E*^{−} (*E*^{+}) and propagates downward (upward) again and so forth, and only numerical artifacts will affect the amplitudes and shapes of the signal during later periods of the cycle. The increase in amplitude toward the surface in both *E*^{+} and *E*^{−} is related to the increase in *N* with height.

Switching on the wave–wave interactions by setting *α*_{υ} = 1/3 day^{−1}, the compartments *E*^{+} and *E*^{−} start to interact and energy is scattered between the upward and the downward waves. After 20 days the process leads to an almost symmetric distribution in upward and downward wave energies, and the upward and downward propagation of the compartments is hardly seen anymore. Using the stationary version of Δ*E* = Δ*E*_{e} + Δ*E*_{w} in the equation for total energy *E* = *E*_{e} + *E*_{w} yields in the absence of dissipation the simple diffusion equation *E* and Δ*E* shown in Fig. 3a. While Δ*E* = 0, that is, a completely symmetric state has been reached, *E* features a surface maximum due to the dependency of *c*_{0} on *N*.

Consider now a case with the mean flow as shown in Fig. 1, but still without dissipation and forcing, and also without wave–wave interactions for the moment, that is, with *α*_{υ} = 0. Figure 4a shows *U* = 0. The perturbation in

This behavior does not change for the same case but including now the wave–wave interactions, that is, setting *α*_{υ} = 1/3 day^{−1}. The scattering between *E*_{e} = Δ*E*_{w} = 0. However, while the energy of the eastward or along-flow propagating part of the waves *E*_{e} is concentrated below the shear zone, the energy of the westward or counterflow propagating part of the waves *E*_{w} is concentrated above the shear zone, that is, the wave field features a strong anisotropy in *ϕ* above and below the shear zone. But since there is as much wave energy propagating into as from the shear zone for both compartments *E*_{e} and *E*_{w}, since Δ*E*_{e} = Δ*E*_{w} = 0, there is no net energy gain by the waves or the mean flow. Since the wave–wave interactions conserve total energy and there is thus no dissipation, the net effect of the waves on the mean flow vanishes, as seen above for the consideration of single waves, and in agreement with the nonacceleration theorem by Andrews and McIntyre (1976).

Only for a case with dissipation, that is, using now the closures [Eq. (21)] with the parameter *μ* as shown in Fig. 1, do we expect a net wave–mean flow interaction. To obtain a stationary case in the numerical simulation, we now also implement forcing at the bottom by specifying a flux ^{−6} m^{3} s^{−3} into both Δ*E*_{e} and Δ*E*_{w}, which aims to resemble the generation of gravity waves by the barotropic tides over topography with a realistic amplitude. Figure 5 shows the spinup of the energy compartments *E*_{e}, *E*_{w}, Δ*E*_{e}, and Δ*E*_{w}. Again, upward and downward propagating signals can be seen in the compartments, originating from the bottom and reflected at the surface and the bottom into the respective complementary compartments. As before for the case without dissipation, *E*_{w}. For the *E*_{e} compartment, the opposite happens: while *E*_{e} > 0 since energy is supplied from below.

Since the *E*_{w} compartment accumulates more energy above the shear zone and thus Δ*E*_{w} > Δ*E*_{e}. The net energy transfer from the mean flow to the waves is given from Eq. (19) by _{z}*U* > 0 and ∂_{z}*U* < 0 and bottom and surface forcing (not shown). However, a different profile of *U* can generate locally a net transfer from the mean flow to the waves, that is, a negative wave drag. A detailed analysis on this issue is given in Part II, treating also the three-dimensional case and arbitrary configurations of the mean flow.

## 7. Summary and discussion

Internal gravity waves interacting with a mean shear flow strictly follow the nonacceleration theorem of Boyd (1976) and Andrews and McIntyre (1976) if adiabatic conditions apply. Then, the energy exchange with the mean flow of a pair of waves with the same horizontal wavenumber and opposing vertical wavenumbers (or opposing horizontal wavenumbers and same vertical wavenumbers) vanishes. The same is true for a single wave propagating upward or downward in a shear flow when reflections at the ocean’s surface and bottom (or interior turning points) are considered and averaging over up–down cycles is applied. Here the waves gain (lose) energy propagating upward or downward and lose (gain) the same amount when reversing.

The nonacceleration constraint is only broken when dissipative effects are active, for example, friction, wave breaking, or encounters of critical layers. In addition to the dissipation, the waves are then affected by a net energy exchange with the mean flow, which, depending on the shear configuration, can be of either sign. The waves then exchange energy with the mean flow depending on the vertical asymmetry (difference between upward and downward propagating energies) and horizontal anisotropy of the wave field. On the other hand, these properties are also excited by the interaction with the mean flow, that is, an initially symmetric and isotropic wave spectrum is modulated toward asymmetry and anisotropy by the interaction process.

The present study discusses these effects in detail for single waves as well as a random ensemble of waves in a horizontal shear flow, applying the WKB method for three-dimensionally propagating waves and describing the wave–mean flow interaction by wave action conservation and the appropriate radiation balance equation for the wave spectrum. The latter is integrated in wavenumber space, similar to previous IDEMIX models (Olbers and Eden 2013; Eden and Olbers 2014), to consider the interaction of integrated energy compartments among themselves and with the mean shear. A key assumption for our approach is the assumption that the wave field is close to a representative empirical spectral shape. The present analysis is restricted to a shear flow with uniform direction and the feedback of the interaction on the mean flow is not yet considered. Part II generalizes the results to arbitrary configurations and the full interaction problem. For unidirectional flow, four compartments are the minimum to adequately represent the interactions, taking care of the up–down differences and along-flow and counterflow anisotropies. The results of the approach are four compartment equations that can be integrated by numerical time stepping.

For practical use, the effect of nonlinear wave–wave interaction needs to be parameterized in the compartment equations. We have used here the simple closures that are also used by Olbers and Eden (2013). These closures rely on the parameterization for the wave dissipation as suggested by McComas and Müller (1981) and Henyey et al. (1986), and on linear damping toward symmetry in upward and downward propagating waves. The former is also used in similar form to estimate wave dissipation from finescale observations (Gregg 1989; Polzin et al. 1995; Sun et al. 1999). As in Olbers and Eden (2013), the time scale *τ*_{υ} for the latter is taken constant for simplicity here. Its value is on the order of days motivated by the Langevin rate in the scattering integral of weakly interaction waves with a spectral distortion. On the other hand, it is clear that *τ*_{υ} differs for different wavenumbers such that a constant value is only a zero-order approximation for the effect. A possible way to determine *τ*_{υ} for a spectral distortion in a continuous GM76-like wave field exactly from the scattering integral is demonstrated in appendix A of Eden and Olbers (2014), but it remains unclear how to generalize this for *τ*_{υ} of a continuous spectrum.

One possible way to obtain an improved expression for *τ*_{υ} and also for the dissipation are numerical calculations of the scattering integral for a given spectrum as done in Lvov et al. (2012). Another way is given by direct numerical simulations of the wave field, which would, however, be computational rather demanding. In such simulations, the impact of mean shear flow on the dissipation and symmetrization of the wave field could also be investigated. Furthermore, the changes of the mean vertical group velocity *c*_{0} and the mean growth rate by wave drag Λ for large deviations of the actual spectrum from the factorized (GM76) spectrum could be estimated in such numerical simulations.

The compartment model for unidirectional flow presented here using the simple closures for the effect of wave–wave interaction is initialized with a simple exponential stratification and a shear zone concentrated in the thermocline and otherwise identical parameterizations as in previous IDEMIX versions, in particular for wave–wave interactions and wave breaking. We have demonstrated the following behavior:

- without wave–wave interactions and dissipation, the wave field is in a cyclic state due to reflection at top and bottom, with vanishing global energy exchange with the mean flow in the time mean;
- including wave–wave interactions symmetrizes the energy spectrum, but conserves total energy, yielding a wave field that is symmetric and that still does not globally exchange energy with the mean flow in the time mean; and
- only when wave dissipation (by wave breaking) is considered does a global nonzero net exchange of energy with the mean flow occur in the time mean. The exchange may have either sign.

The present version of the compartment model is studied as a test bed for the general configuration, that is, interaction with a nonuniform shear including the implied acceleration or deceleration of the mean flow. In addition, in the present version critical layers have no direct manifestation, though one could argue that they contribute to overall dissipation by the implemented parameterization of wave breaking. The extra transfer of energy to turbulence in critical layers and the general interaction problem will be considered in a follow-up paper. The extension of the present study to an arbitrary configuration of the mean flow and the wave effects in the momentum equation is considered in Part II.

## Acknowledgments

This paper is a contribution to the Collaborative Research Centre TRR 181 “Energy Transfers in Atmosphere and Ocean” funded by the German Research Foundation.

## APPENDIX A

### Conservation of Energy

*N*and

**U**(

*z*) result in

*ω*

_{enc}= constant and

*N*then leads towhere

*c*is positive by definition while ∂

*c*/∂

*ω*> 0 for most of the internal wave frequency range (for

*E*.

## APPENDIX B

### The Parameter Λ

*n*

_{B}= (2/

*π*)/[1 − 2/

*π*arcsin(

*f*/

*N*)], the parameter Λ is given byThe integral can be solved in closed form, which yieldsNote that Λ is a positive and decreasing function of

*N*/

*f*. It is displayed in Fig. B1.

## REFERENCES

Alexander, M., and et al. , 2010: Recent developments in gravity-wave effects in climate models and the global distribution of gravity-wave momentum flux from observations and models.

,*Quart. J. Roy. Meteor. Soc.***136**, 1103–1124, doi:10.1002/qj.637.Alford, M., 2001: Internal swell generation: The spatial distribution of energy flux from the wind to mixed layer near-inertial motions.

,*J. Phys. Oceanogr.***31**, 2359–2368, doi:10.1175/1520-0485(2001)031<2359:ISGTSD>2.0.CO;2.Andrews, D. G., and M. E. McIntyre, 1976: Planetary waves in horizontal and vertical shear: The generalized Eliassen–Palm relation and the zonal mean acceleration.

,*J. Atmos. Sci.***33**, 2031–2048, doi:10.1175/1520-0469(1976)033<2031:PWIHAV>2.0.CO;2.Becker, E., 2012: Dynamical control of the middle atmosphere.

,*Space Sci. Rev.***168**, 283–314, doi:10.1007/s11214-011-9841-5.Boyd, J. P., 1976: The noninteraction of waves with the zonally averaged flow on a spherical earth and the interrelationships on eddy fluxes of energy, heat and momentum.

,*J. Atmos. Sci.***33**, 2285–2291, doi:10.1175/1520-0469(1976)033<2285:TNOWWT>2.0.CO;2.Bretherton, F. P., and C. J. Garrett, 1968: Wavetrains in inhomogeneous moving media.

,*Proc. Roy. Soc. London***302A**, 529–554, doi:10.1098/rspa.1968.0034.Brown, E. D., and W. B. Owens, 1981: Observations of the horizontal interactions between the internal wave field and the mesoscale flow.

,*J. Phys. Oceanogr.***11**, 1474–1480, doi:10.1175/1520-0485(1981)011<1474:OOTHIB>2.0.CO;2.Brüggemann, N., and C. Eden, 2015: Routes to dissipation under different dynamical conditions.

,*J. Phys. Oceanogr.***45**, 2149–2168, doi:10.1175/JPO-D-14-0205.1.Bühler, O., and M. E. Mcintyre, 2005: Wave capture and wave–vortex duality.

,*J. Fluid Mech.***534**, 67–95, doi:10.1017/S0022112005004374.Cairns, J. L., and G. O. Williams, 1976: Internal wave observations from a midwater float.

,*J. Geophys. Res.***81**, 1943–1950, doi:10.1029/JC081i012p01943.Eden, C., and R. J. Greatbatch, 2008: Diapycnal mixing by mesoscale eddies.

,*Ocean Modell.***23**, 113–120, doi:10.1016/j.ocemod.2008.04.006.Eden, C., and D. Olbers, 2014: An energy compartment model for propagation, nonlinear interaction and dissipation of internal gravity waves.

,*J. Phys. Oceanogr.***44**, 2093–2106, doi:10.1175/JPO-D-13-0224.1.Eden, C., and D. Olbers, 2017: A closure for internal wave–mean flow interaction. Part II: Wave drag.

,*J. Phys. Oceanogr.***47**, 1403–1412, doi:10.1175/JPO-D-16-0056.1.Falahat, S., J. Nycander, F. Roquet, and M. Zarroug, 2014: Global calculation of tidal energy conversion into vertical normal modes.

,*J. Phys. Oceanogr.***44**, 3225–3244, doi:10.1175/JPO-D-14-0002.1.Ford, R., M. McIntyre, and W. Norton, 2000: Balance and the slow quasimanifold: Some explicit results.

,*J. Atmos. Sci.***57**, 1236–1254, doi:10.1175/1520-0469(2000)057<1236:BATSQS>2.0.CO;2.Frankignoul, C., and T. M. Joyce, 1979: On the internal wave variability during the internal wave experiment (IWEX).

,*J. Geophys. Res.***84**, 769–776, doi:10.1029/JC084iC02p00769.Fritts, D. C., and M. J. Alexander, 2003: Gravity wave dynamics and effects in the middle atmosphere.

,*Rev. Geophys.***41**, 1003, doi:10.1029/2001RG000106.Garrett, C., and W. Munk, 1975: Space-time scales of internal waves: A progress report.

,*J. Geophys. Res.***80**, 291–297, doi:10.1029/JC080i003p00291.Gregg, M. C., 1989: Scaling turbulent dissipation in the thermocline.

,*J. Geophys. Res.***94**, 9686–9698, doi:10.1029/JC094iC07p09686.Henyey, F., J. Wright, and S. Flatté, 1986: Energy and action flow through the internal wave field: An eikonal approach.

,*J. Geophys. Res.***91**, 8487–8495, doi:10.1029/JC091iC07p08487.Huang, K., 1987:

*Statistical Mechanics*. 2nd ed. Wiley, 493 pp.Lighthill, J., 1978:

*Waves in Fluids*. Cambridge University Press, 516 pp.Lvov, Y. V., K. L. Polzin, and N. Yokoyama, 2012: Resonant and near-resonant internal wave interactions.

,*J. Phys. Oceanogr.***42**, 669–691, doi:10.1175/2011JPO4129.1.McComas, C. H., 1977: Equilibrium mechanisms within the oceanic internal wave field.

,*J. Phys. Oceanogr.***7**, 836–845, doi:10.1175/1520-0485(1977)007<0836:EMWTOI>2.0.CO;2.McComas, C. H., and P. Müller, 1981: Time scales of resonant interactions among oceanic internal waves.

,*J. Phys. Oceanogr.***11**, 139–147, doi:10.1175/1520-0485(1981)011<0139:TSORIA>2.0.CO;2.Miller, M., T. Palmer, and R. Swinbank, 1989: Parametrization and influence of subgridscale orography in general circulation and numerical weather prediction models.

,*Meteor. Atmos. Phys.***40**, 84–109, doi:10.1007/BF01027469.Molemaker, M. J., J. C. McWilliams, and X. Capet, 2010: Balanced and unbalanced routes to dissipation in an equilibrated Eady flow.

,*J. Fluid Mech.***654**, 35–63, doi:10.1017/S0022112009993272.Muench, J. E., and E. Kunze, 1999: Internal wave interactions with equatorial deep jets. Part I: Momentum-flux divergence.

,*J. Phys. Oceanogr.***29**, 1453–1467, doi:10.1175/1520-0485(1999)029<1453:IWIWED>2.0.CO;2.Muench, J. E., and E. Kunze, 2000: Internal wave interactions with equatorial deep jets. Part II: Acceleration of the jets.

,*J. Phys. Oceanogr.***30**, 2099–2110, doi:10.1175/1520-0485(2000)030<2099:IWIWED>2.0.CO;2.Müller, P., 1976: On the diffusion of momentum and mass by internal gravity waves.

,*J. Fluid Mech.***77**, 789–823, doi:10.1017/S0022112076002899.Müller, P., J. McWilliams, and M. Molemaker, 2005: Routes to dissipation in the ocean: The 2D/3D turbulence conundrum.

*Marine Turbulence: Theories, Observations and Models*, H. Baumert, J. Simpson, and J. Sündermann, Eds., Cambridge University Press, 397–405.Munk, W., 1981: Internal waves and small-scale processes.

*Evolution of Physical Oceanography*, B. A. Warren and C. Wunsch, Eds., MIT Press, 264–291.Muraschko, J., M. Fruman, U. Achatz, S. Hickel, and Y. Toledo, 2015: On the application of Wentzel–Kramer–Brillouin theory for the simulation of the weakly nonlinear dynamics of gravity waves.

,*Quart. J. Roy. Meteor. Soc.***141**, 676–697, doi:10.1002/qj.2381.Nikurashin, M., and R. Ferrari, 2011: Global energy conversion rate from geostrophic flows into internal lee waves in the deep ocean.

,*Geophys. Res. Lett.***38**, L08610, doi:10.1029/2011GL046576.Nycander, J., 2005: Generation of internal waves in the deep ocean by tides.

,*J. Geophys. Res.***110**, C10028, doi:10.1029/2004JC002487.Olbers, D., 1983: Models of the oceanic internal wave field.

,*Rev. Geophys. Space Phys.***21**, 1567–1606, doi:10.1029/RG021i007p01567.Olbers, D., and C. Eden, 2013: A global model for the diapycnal diffusivity induced by internal gravity waves.

,*J. Phys. Oceanogr.***43**, 1759–1779, doi:10.1175/JPO-D-12-0207.1.Olbers, D., J. Willebrand, and C. Eden, 2012:

*Ocean Dynamics*. Springer, 703 pp.Pinkel, R., 2005: Near-inertial wave propagation in the western arctic.

,*J. Phys. Oceanogr.***35**, 645–665, doi:10.1175/JPO2715.1.Polzin, K. L., 2008: Mesoscale eddy-internal wave coupling. Part I: Symmetry, wave capture, and results from the mid-ocean dynamics experiment.

,*J. Phys. Oceanogr.***38**, 2556–2574, doi:10.1175/2008JPO3666.1.Polzin, K. L., 2010: Mesoscale eddy-internal wave coupling. Part II: Energetics and results from POLYMODE.

,*J. Phys. Oceanogr.***40**, 789–801, doi:10.1175/2009JPO4039.1.Polzin, K. L., J. M. Toole, and R. W. Schmitt, 1995: Finescale parameterizations of turbulent dissipation.

,*J. Phys. Oceanogr.***25**, 306–328, doi:10.1175/1520-0485(1995)025<0306:FPOTD>2.0.CO;2.Rimac, A., J.-S. von Storch, C. Eden, and H. Haak, 2013: The influence of high-resolution wind stress field on the power input to near-inertial motions in the ocean.

,*Geophys. Res. Lett.***40**, 4882–4886, doi:10.1002/grl.50929.Ruddick, B. R., and T. M. Joyce, 1979: Observations of interaction between the internal wavefield and low-frequency flows in the North Atlantic.

,*J. Phys. Oceanogr.***9**, 498–517, doi:10.1175/1520-0485(1979)009<0498:OOIBTI>2.0.CO;2.Scott, R., J. Goff, A. Naveira Garabato, and A. Nurser, 2011: Global rate and spectral characteristics of internal gravity wave generation by geostrophic flow over topography.

,*J. Geophys. Res.***116**, C09029, doi:10.1029/2011JC007005.Sun, S., R. Bleck, C. Rooth, J. Dukowicz, E. Chassignet, and P. Killworth, 1999: Inclusion of thermobaricity in isopycnic-coordinate ocean models.

,*J. Phys. Oceanogr.***29**, 2719–2729, doi:10.1175/1520-0485(1999)029<2719:IOTIIC>2.0.CO;2.Tandon, A., and C. Garrett, 1996: On a recent parameterization of mesoscale eddies.

,*J. Phys. Oceanogr.***26**, 406–416, doi:10.1175/1520-0485(1996)026<0406:OARPOM>2.0.CO;2.Trossman, D. S., B. K. Arbic, J. G. Richman, S. T. Garner, S. R. Jayne, and A. J. Wallcraft, 2016: Impact of topographic internal lee wave drag on an eddying global ocean model.

,*Ocean Modell.***97**, 109–128, doi:10.1016/j.ocemod.2015.10.013.Whitham, G., 1970: Two-timing, variational principles and waves.

,*J. Fluid Mech.***44**, 373–395, doi:10.1017/S002211207000188X.Wunsch, C., and R. Ferrari, 2004: Vertical mixing, energy and the general circulation of the oceans.

,*Annu. Rev. Fluid Mech.***36**, 281–314, doi:10.1146/annurev.fluid.36.050802.122121.

^{1}

Note that if the oceanic waves are dissipated at the surface as observed by Pinkel (2005) for sea ice covered regions, a larger effect of wave drag might become possible.

^{2}

All vectors are two-dimensional here.