betaFeedback

WorksPhysics and Astronomy

Available potential vorticity and the wave-vortex decomposition for arbitrary stratification

Abstract

We consider a rotating non-hydrostatic flow with arbitrary stratification and argue that 1) the appropriate form of potential vorticity (PV) for this system is in terms of isopycnal deviation and 2) the decomposition into energetically orthogonal solutions is fundamentally a PV-inversion. The new closed-form expression for available potential vorticity (APV) is expressed in terms of isopycnal deviation, following the ideas in Wagner & Young (2015). This form of APV linearizes to quasigeostrophic PV (QGPV) after discarding the nonlinear stretching term and a height nonlinearity, the latter of which is not present in constant stratification. This formulation leads to positive definite definitions of potential enstrophy and total energy expressed in terms of isopycnal deviation, from which the quadratic versions emerge at lowest order. It is exactly these quantities diagonalized by the linear eigenmodes. Internal-gravity waves, geostrophic motions, inertial oscillations, and a mean density anomaly form the energetically and enstrophically orthogonal constituents of flow. The complete state of the fluid can be represented in terms of these physically realizeable modes and determined from the derived projection operators using the horizontal velocity and density anomaly. The projection of the fluid state onto the non-hydrostatic wave modes, reveals that one must first account for the PV portion of the flow before recovering the wave solutions. We apply the physical insights of the decomposition to a mesoscale eddy showing how strict adherence to adiabatic rearrangement places strong constraints on the vertical structure of such eddies, including a skew towards stronger cyclonic eddies in the upper water-column. Finally, the expression for APV is shown to reproduce the height nonlinearity of shallow-water PV, a well know feature that breaks the cyclone-anticyclone symmetry in QGPV.

Introduction

Geostrophic motions and internal gravity waves (IGWs) form two of the primary building blocks of geophysical fluid dynamics and methods for disentangling these types of motions are often referred to as wave-vortex decompositions.1 Broadly speaking there are two classes of wave-vortex decompositions: full-knowledge decompositions and sparse-data decompositions. The full-knowledge decompositions assume complete knowledge of the fluid state and are then challenged with separating the fluid state into energetically orthogonal constituent parts. The challenges associated with such decompositions are theoretical in nature, but result in separations with no ambiguities beyond methodological assumptions. Sparse-data decompositions, on the other hand, are faced with the same task of separating the fluid state into constituent parts, but with the additional constraints imposed by incomplete knowledge of the fluid state due to sparse sampling, e.g., velocity time series from a mooring. The challenges here are far more practical in nature, requiring statistical assumptions to proceed, and thus resulting in uncertainties related to the sparse sampling.

Full-knowledge decompositions are a prerequisite to sparse-data decompositions. After all, if we cannot agree on what constitutes a wave or vortex with complete knowledge of the fluid state, how can we agree on what we are measuring from sparse observations? Thus at the heart of every wave-vortex decomposition is the idea of observability. Each constituent in the decomposition must be a measurable, meaningfully distinct fluid state. Constituents of the wave-vortex decomposition certainly must satisfy all boundary conditions and solve the equations-of-motion, but what makes a solution meaningfully distinct from other solutions? The answer in geophysical fluid dynamics is the same as in quantum mechanics: energy orthogonality.

What is energy orthogonality?

Energy (and enstrophy) orthogonality is an idea that underpins many of the fundamental concepts in geophysical fluids and without it, the inverse energy cascade and forward enstrophy cascade of two-dimensional turbulence would not be well defined concepts, nor would it be possible to construct an internal wave energy spectrum. The importance of energy and enstrophy orthogonality is rarely discussed explicitly, likely because in the simplest examples it follows directly from Parseval’s theorem for Fourier modes. To understand the importance of this concept we spend some time drawing the analogy with Parseval’s theorem for Fourier modes and discussing several examples where orthogonality has been important in geophysical fluids.

The state of a geophysical fluid (for our purposes) is given by (u,v,w,η,p)(u, v, w, \eta, p) for fluid velocity (u,v,w)(u, v, w) with isopycnal deviation η\eta and pressure pp—and any basis capable of representing all physically realizable states of the fluid is said to be complete. Observations of these dynamical variables in space and time are rather trivially recorded on a complete basis, e.g., nn observations are recorded in a vector of length nn. Representing numerical or analytical solutions with, e.g., a cosine basis, requires a bit more care, as the basis must be compatible with the dynamically imposed boundary conditions. That said, there are numerous suitable complete bases capable of representing the physically realizable states of the system. The reverse is also true: complete bases such as cosine and Chebyshev series are suitable for a variety of different physical systems. In contrast to a complete basis, an energetically orthogonal basis is deeply tied to the dynamics.

To lowest order the total volume-integrated available energy of the rotating, non-hydrostatic Boussinesq is

E[(u,v,w,η)]=12∫(u2+v2+w2+N2η2) dV(1)\mathcal{E}[(u,v,w,\eta)] = \frac{1}{2} \int(u^2+v^2+w^2+N^2\eta^2)\,dV \tag*{(1)}

where N2(z)N^2(z) is the squared buoyancy frequency. If the fluid can be linearly decomposed into two parts such that (u,v,w,η)=(u1+u2,v1+v2,w1+w2,η1+η2)(u,v,w,\eta)=(u_1+u_2,v_1+v_2,w_1+w_2,\eta_1+\eta_2) then

E[(u,v,w,η)]=E[(u1,v1,w1,η1)]+E[(u2,v2,w2,η2)]+ϵ12(2)\mathcal{E}[(u,v,w,\eta)] = \mathcal{E}[(u_1,v_1,w_1,\eta_1)] + \mathcal{E}[(u_2,v_2,w_2,\eta_2)] + \epsilon_{12} \tag*{(2)}

where ϵ12\epsilon_{12} is a quadratic cross-term. For a sparse-data decomposition it may be that the best one can hope for is that the cross-term ϵ12\epsilon_{12} is small, but for a full-knowledge decomposition it may be possible to find solutions where the cross-term vanishes, ϵ12=δ12\epsilon_{12}=\delta_{12} and decomposition is energetically orthogonal.

When a basis is both complete and energetically orthogonal then each basis member has a unique energy signature—a feature that allows one to say that energy is moving from somewhere to somewhere else. Moreover, the lack of cross-terms also allows us to write an energy spectrum because the sum of the squares of the basis functions is proportional to the total energy. This is exactly analogous to a Fourier series which partitions the total variance of a periodic function in terms of orthogonal Fourier modes.

The potential enstrophy of fluid is another quadratic quantity that characterizes the fluid,

Z[(u,v,w,η)]=12∫(vx−uy−fη)2 dV(3)Z[(u,v,w,\eta)] = \frac{1}{2}\int(v_x-u_y-f\eta)^2\,dV \tag*{(3)}

and a wave-vortex decomposition will require enstrophy orthogonality following rules analogous to (2).

For two-dimensional doubly-periodic and three-dimensional triply-periodic flows all components of the energetically and enstrophically orthogonal modes are proportional to Fourier modes, and thus energy orthogonality is simply a rescaled version of Fourier mode orthogonality. Due to this almost trivial relationship between the two, it would be easy to miss the significance of energy orthogonality.

Why are energy and enstrophy orthogonality important?

One of the most important results in rotating stratified fluids is the existence of the inertial cascades [1]. The central idea is that energy moves from smaller scales to larger scales while simultaneously enstrophy moves from larger scales to smaller scales. To make such claims and simultaneously observe this dual cascade requires a basis of eigenmodes that are both energetically and enstrophically orthogonal.

The formulation of the oceanic internal wave spectrum also hinges on energy orthogonality. The canonical form of the internal wave spectrum, the Garrett-Munk (GM) spectrum, is constructed in a basis of energetically orthogonal hydrostatic internal wave eigenmodes [2], such that the depth integrated total energy is expressed as a sum of squares of individual eigenmodes. Importantly, energy orthogonality is not just a choice, but a requirement for constructing an energy spectrum. Approximating the hydrostatic eigenmodes with the WKB approximation leads to a formulation of the GM in terms of a stretched coordinate where the eigenmodes look like rescaled Fourier modes and energy orthogonality resembles Fourier mode orthogonality again.

While the formulation of the internal gravity wave spectrum by Garrett and Munk [1972] is sufficient for describing the energy content of interior flow, eigenmodes with an explicit free-surface are necessary when surface (or barotropic) waves are present. Kelly [2016] constructed an energetically orthogonal basis for hydrostatic waves with an explicit free-surface using vertical modes first described by Olbers [1986]. This solved a decades long ‘spurious energy conversion’ problem which, when viewed through the lens of energy orthogonality, was caused by attempts to use a non-orthogonal basis.

To assess the energy and enstrophy fluxes in three dimensions with variable stratification and buoyancy anomalies at the boundaries, requires an energetically enstrophically orthogonal basis. Finding such a basis is not trivial. The basis proposed by Smith and Vanneste [2013] satisfies these conditions for quasigeostrophic flows.

With an energetically orthogonal basis at hand, it then becomes possible to identify the physically important mechanisms that shape the observed spectra and precisely quantify the role of each mode in the transfer of energy. These include methods that limit the available interactions [6, 7], directly diagnose energy transfers [8], or even directly attribute the mechanisms responsible for stirring [9].

Methodologies for separating dynamical effects that formally lack energy orthogonality may very well be connected to an energy-orthogonal decomposition. Lelong et al. [2020] formulate a hybrid spatial-temporal decomposition to isolate the energetics of a mesoscale eddy, inertial oscillations, and a developing internal gravity wave field. The energetic transfers were consistent with the hypothesis that the eddy acted at a catalyst in facilitating the transfer of energy from the inertial oscillations to the internal gravity wave field. With a complete energetically orthogonal basis, Early et al. [2021] measured this and the other fluxes exactly, confirming the hypothesis directly. Other methodologies for isolating dynamical phenomena based on filtering or scaling are almost certainly related to full-knowledge wave-vortex decompositions.

A march towards realism

The prototypical wave-vortex decompositions are either triply-periodic with constant stratification or shallow-water models—both of which are highly idealized. Although the single-layer shallow-water model would appear to be the simpler one, its energy and enstrophy both have non-quadratic terms due to the height nonlinearity [12], a feature which is also present in the variable stratification model presented in this manuscript. Overall, the community continues to pursue a number of different directions that move beyond these prototypical models and incorporate increasingly physically realistic effects.

A complete, energetically orthogonal solution set unambiguously disentangles the fluid state into constituent wave and geostrophic motions, but the lack of ambiguity is only as good as the physical interpretation of the linear eigenmodes.

One line of inquiry is to extend the linear eigenmode decomposition to incorporate higher order effects. It has long been recognized that true balanced motions require nonlinear corrections proportional to the Rossby number and thus a true separation requires additional steps beyond normal mode decomposition [13].

At larger scales order Rossby number effects remain small. Recent work has extended the normal mode decomposition to the shallow-water equations on the sphere [14]. The geostrophic mode is now interpreted as a Rossby mode, and the methodology has been applied both numerically and diagnostically.

Another glaring omission of the prototypical wave-vortex decompositions is the lack of variable stratification. Recent work has extended the wave-vortex decompositions to include non-hydrostatics with variable stratification [11]. Just like shallow-water, energy and enstrophy now contain higher order effects not present in constant stratification.

Numerous other physical effects remain to be incorporated into wave-vortex decompositions. O’Neill et al. [ [15] develop a methodology that includes Ekman effects at the ocean surface boundary. Although the methodology lacks formal proof of orthogonality, it successfully diagnoses fluid flow at the boundary. This and other effects will need to be incorporated into wave-vortex decompositions to make more direct contact with sparse-data decompositions.

The work here

The work presented here continues the march towards increasing realism in wave-vortex decompositions by resolving a number of outstanding issues surrounding the variable stratification decomposition, including the correct definition of available potential vorticity and orthogonality of non-hydrostatic eigenmodes. We focus particularly on observability, ensuring that all states of motion are physically realizeable. In practice, this means that each eigen-state satisfies the linear equations of motion, all boundary conditions, and even all quadratic conservation laws. A notable exception to this is the mean-density anomaly (mda) mode, discussed at length below.

The wave-vortex decompositions hinge on energy and enstrophy orthogonality of the linear solutions and thus expressions for the fully nonlinear energy and potential enstrophy are critical for articulating the order of approximation in the linear solutions. As noted in [11] however, expressions for potential vorticity either contain a large background contribution and thus do not linearize to the correct quantity (i.e., Ertel PV), or do not have closed form solutions [16]. Here we present what we claim is the ‘best’ definition for available potential vorticity (APV) in variable stratification on an ff-plane, which matches the intention of Wagner and Young [2015], but in closed form. This expression first appeared in [17], although with erroneous caveats limiting its applicability as we noted in [18]. The key to our expression for APV is the use of isopycnal deviation as the dependent variable, just as in [19] for the construction of available potential energy (APE). The nonlinear conservation laws for APV and APE include nonlinearities that are more akin to their shallow-water versions, rather than three-dimensional constant stratification. The conservation laws including their volume integrated forms, are described in § 3.

Although the nonlinear conservation laws are most naturally expressed in terms of isopycnal deviation, the linearized conservation laws and equations-of-motion are best formulated in terms of excess density, as described in § 4. The complete energetically and enstrophically orthogonal solution set is presented in § 5. The solution set includes two types of waves modes devoid of potential vorticity, inertial oscillations and internal gravity wave modes, and two types of potential vorticity carrying modes, the mean-density anomaly (mda) mode and the geostrophic modes. The mean-density anomaly modes describe the difference between the mean density density state and the flattened isopycnal states of motion. These modes carry both potential enstrophy and potential energy and are necessary for a complete description of the fluid state.

With a complete energetically orthogonal basis at hand, § 6 expresses the nonlinear equations of motion in terms of this wave-vortex basis. Reduced interaction models follow immediately, including stratified quasigeostrophic potential vorticity equation by restriction to the potential vorticity carrying modes, and wave-only models by restriction to the wave modes.

Finally § 7 shows the implications of these results on our understanding of mesoscale eddies. In particular, strict adherence to the boundary conditions implies that nonlinear mesoscale eddies will favor cyclones, especially with density anomalies near the surface, consistent with past results. The new form of APV includes a height nonlinearity not present in constant stratification that resembles the height nonlinearity in shallow-water. This nonlinear correction is of O(1)O(1) importance, as has long been predicted for geostrophic motions at the mesoscales. § 8 provides discussion and conclusions.

Background

Under the Boussinesq approximation, we consider the inviscid equations of motion

dudt−fv=−1ρ0∂xptot(4)\frac{du}{dt}-fv=-\frac{1}{\rho_0}\partial_x p_{\mathrm{tot}} \tag*{(4)}
dvdt+fu=−1ρ0∂yptot(5)\frac{dv}{dt}+fu=-\frac{1}{\rho_0}\partial_y p_{\mathrm{tot}} \tag*{(5)}
dwdt=−1ρ0∂zptot−1ρ0gρtot(6)\frac{dw}{dt}=-\frac{1}{\rho_0}\partial_z p_{\mathrm{tot}}-\frac{1}{\rho_0}g\rho_{\mathrm{tot}} \tag*{(6)}
dρtotdt=0(7)\frac{d\rho_{\mathrm{tot}}}{dt}=0 \tag*{(7)}
∂xu+∂yv+∂zw=0,(8)\partial_xu+\partial_yv+\partial_zw=0, \tag*{(8)}

where u=(u,v,w)u=(u,v,w) is the fluid velocity in Cartesian coordinates with position vector x=(x,y,z)\boldsymbol{x}=(x,y,z), ptotp_{\mathrm{tot}} is the total pressure, ρtot\rho_{\mathrm{tot}} is the total density, and ff is the constant Coriolis parameter. The boundaries are assumed to be periodic in (x,y)(x,y); and rigid, free-slip (w=0w=0), and flat at z=0z=0 and z=−Dz=-D. Buoyancy anomalies are not permitted at the vertical boundaries in the work considered here. Note that there is no mixing allowed since equations (4) do not include viscosity or diffusivity.

The no-motion and flattened isopycnal states

The most trivial, but important, solution to the equations of motion (4) is the no-motion state which occurs when we let u(x,t)=(u,v,w)=(0,0,0)\boldsymbol{u}(x,t)=(u,v,w)=(0,0,0). The density of the fluid ρtot(x,t)\rho_{\mathrm{tot}}(x,t) must be adiabatically re-arranged to eliminate horizontal pressure and density gradients by moving parcels to their no-motions heights as sketched in Figure 1. The no-motion solution is denoted as (u,v,w,p,ρ)=(0,0,0,−ρ0gz+p~nm(z),ρ0+ρ~nm(z))(u,v,w,p,\rho)=(0,0,0,-\rho_0gz+\widetilde{p}_{\mathrm{nm}}(z),\rho_0+\widetilde{\rho}_{\mathrm{nm}}(z)) where p~nm\widetilde{p}_{\mathrm{nm}} and ρ~nm\widetilde{\rho}_{\mathrm{nm}} are the variable part of the no-motion solution defined as

Schematic of adiabatic re-arrangement and isopycnal displacements

Figure 1. Schematic of adiabatic re-arrangement and isopycnal displacements. η(x,t)\eta(x,t) is positive if it is above its no-motion height.

p~nm(z)=−g∫0zρ~nm(ξ) dξ,(9)\widetilde{p}_{\mathrm{nm}}(z)=-g\int_0^z\widetilde{\rho}_{\mathrm{nm}}(\xi)\,d\xi, \tag*{(9)}

so that ∂zp~nm(z)≡−gρ~nm(z)\partial_z\widetilde{p}_{\mathrm{nm}}(z)\equiv-g\widetilde{\rho}_{\mathrm{nm}}(z) and the equations of motion are satisfied. We note that the Boussinesq approximation assumes that the positive constant ρ0≫∣ρ~nm(z)∣\rho_0\gg\lvert\widetilde{\rho}_{\mathrm{nm}}(z)\rvert and ∣−ρ0gz∣≫∣p~nm(z)∣\lvert-\rho_0gz\rvert\gg\lvert\widetilde{p}_{\mathrm{nm}}(z)\rvert.

Similar to the no-motion state, the flattened-isopycnal state also has w=0w=0 and ρ=ρ0+ρ~nm(z)\rho=\rho_0+\widetilde{\rho}_{\mathrm{nm}}(z). However, this state of the fluid allows for u≠0u\ne0, v≠0v\ne0 and p≠0p\ne0, provided that ∂zp=0\partial_zp=0. As a consequence of flattening isopycnals, these states of motion have no available potential energy or vortex stretching, but do have kinetic energy and vorticity. Thus, unlike the no-motion state, a flattened isopycnal state is an achievable state-of-motion for any initial condition because energy and enstrophy can be conserved.

Importantly, the no-motion density is not the same as the mean density. These two quantities differ by a ‘mean density anomaly’ (mda), a function of zz which may be quite small, or quite large (imagine an inverted fluid). This will be further discussed in section 5.5.

Isopycnal deviation vs excess density

There are two common approaches to expressing re-arrangements of density relative to the no-motion solution: isopycnal deviation (η\eta) or excess density (ρe\rho_e) defined by

ρ~(x,t)≡ρ~nm(z)+ρe(x,t)(10)\tilde{\rho}(x,t) \equiv\tilde{\rho}_{\mathrm{nm}}(z) + \rho_e(x,t) \tag*{(10)}
ρ~(x,t)≡ρ~nm(z−η(x,t)).(11)\tilde{\rho}(x,t) \equiv\tilde{\rho}_{\mathrm{nm}}(z-\eta(x,t)). \tag*{(11)}

The excess density may also be referred to as density anomaly or buoyancy anomaly where buoyancy is defined as b≡−gρe/ρ0b \equiv-g\rho_e/\rho_0. Both formulations in (6) disallow mixing in order to maintain material conservation, but do allow isopycnal overturning. Note that the argument of ρ~nm\tilde{\rho}_{\mathrm{nm}} in (6b) could be used to define an isopycnal coordinate (lines of constant z−ηz-\eta), but this would disallow isopycnal overturns. In practice, computing η\eta from ρe\rho_e requires inverting ρ~nm\tilde{\rho}_{\mathrm{nm}} in equation (6b) with either an analytically specified no-motion density profile, or performing an adiabatic re-arrangement of an existing density field using, e.g., equation (2.3) in [20]. For analytical work, it is useful to approximate the relationship between ρe\rho_e and η\eta. Using ρ~nm(z)\tilde{\rho}_{\mathrm{nm}}(z) and its inverse from the definitions in (6) results in series expansions

References

  1. [1]Geoffrey K. Vallis. Atmospheric and Oceanic Fluid Dynamics: Fundamentals and Large-Scale Circulation. Cambridge University Press, 1 edition, 2006. doi:10.1017/9781107588417.
  2. [2]Christopher Garrett and Walter Munk. Space-Time scales of internal waves. Geophysical Fluid Dynamics, 3(1):225 – 264, 1972. doi:10.1080/03091927208236082.DOI
  3. [5]K. Shafer Smith and Jacques Vanneste. A Surface-Water Basis for Quasigeostrophic Flow. Journal of Physical Oceanography, 43(3):548 – 562, 2013. doi:10.1175/jpo-d-12-0107.1.DOI
  4. [6]Gerardo Hernandez-Duenas, Leslie M Smith, and Samuel N Stechmann. Investigation of Boussinesq dynamics using intermediate models based on wave–vortical interactions. Journal of Fluid Mechanics, 747:247–287, 2014. ISSN 0022-1120. doi:10.1017/jfm.2014.138.
  5. [7]Carsten Eden, Friederike Pollmann, and Dirk Olbers. Numerical evaluation of energy transfers in internal gravity wave spectra of the ocean. Journal of Physical Oceanography, 49(3):737–749, 2019. ISSN 0022-3670. doi:10.1175/jpo-d-18-0075.1.DOI
  6. [8]Carsten Eden, Friederike Pollmann, and Dirk Olbers. Towards a global spectral energy budget for internal gravity waves in the ocean. Journal of Physical Oceanography, 50(4):935–944, 2020. ISSN 0022-3670. doi:10.1175/jpo-d-19-0022.1.DOI
  7. [9]Gerardo Hernández-Dueñas, M Pascale Lelong, and Leslie M Smith. Impact of Wave-Vortical Interactions on Oceanic Submesoscale Lateral Dispersion. Journal of Physical Oceanography, 2021. ISSN 0022-3670. doi:10.1175/jpo-d-20-0299.1.
  8. [11]Jeffrey J. Early, M.P. Lelong, and M.A. Sundermeyer. A generalized wave-vortex decomposition for rotating Boussinesq flows with arbitrary stratification. Journal of Fluid Mechanics, 912:A32, 2021. ISSN 0022-1120. doi:10.1017/jfm.2020.995.DOI
  9. [12]Mark Remmel and Leslie Smith. New intermediate models for rotating shallow water and an investigation of the preference for anticyclones. Journal of Fluid Mechanics, 635:321–359, 2009. ISSN 0022-1120. doi:10.1017/s0022112009007897.DOI
  10. [13]Manita Chouksey, Carsten Eden, Gökçe Tuba Masur, and Marcel Oliver. A comparison of methods to balance geophysical flows. Journal of Fluid Mechanics, 971:A2, 2023. ISSN 0022-1120. doi:10.1017/jfm.2023.602.DOI
  11. [14]Sergiy Vasylkevych and Nedjeljka Žagar. A high-accuracy global prognostic model for the simulation of Rossby and gravity wave dynamics. Quarterly Journal of the Royal Meteorological Society, 147(736):1989–2007, 2021. ISSN 0035-9009. doi:10.1002/qj.4006.
  12. [15]Larry W O’Neill, Dudley B Chelton, Ernesto Rodríguez, Roger Samelson, and Alexander Wineteer. Feasibility of estimating sea surface height anomalies from surface ocean currents and winds. Journal of Atmospheric and Oceanic Technology, 2024. ISSN 0739-0572. doi:10.1175/jet-d-23-0096.1.
  13. [16]G. L. Wagner and W. R. Young. Available potential vorticity and wave-averaged quasi-geostrophic flow. Journal of Fluid Mechanics, 785:401–424, 2015. ISSN 0022-1120. doi:10.1017/jfm.2015.626.DOI
  14. [17]Peter Müller. Ertel’s potential vorticity theorem in physical oceanography. Reviews of Geophysics, 33(1):67–97, 1995. ISSN 8755-1209. doi:10.1029/94rg03215.DOI
  15. [18]Jeffrey J Early, Gerardo Hernández-Dueñas, Leslie M Smith, and M Pascale Lelong. Exact expressions for available potential energy and available potential vorticity. arXiv, 2022. doi:10.48550/arxiv.2212.07405.DOI
  16. [19]Dennis Holliday and Michael E. McIntyre. On potential energy density in an incompressible, stratified fluid. Journal of Fluid Mechanics, 107(-1):221–225, 1981. ISSN 1469-7645. doi:10.1017/s0022112081001742.DOI
  17. [20]Kraig B. Winters and Roy Barkan. Available potential energy density for Boussinesq fluid flow. Journal of Fluid Mechanics, 714:476–488, 2013. ISSN 0022-1120. doi:10.1017/jfm.2012.493.DOI
  18. [21]Tapio Schneider, Isaac M. Held, and Stephen T. Garner. Boundary Effects in Potential Vorticity Dynamics. Journal of the Atmospheric Sciences, 60(8):1024–1040, 2003. ISSN 0022-4928. doi:10.1175/1520-0469(2003)60<1024:beipvd>2.0.co;2.doi.org/10.1175/1520-0469(2003)60<1024:beipvd>2.0.co;2
  19. [22]Guillaume Roullet and Patrice Klein. Available potential energy diagnosis in a direct numerical simulation of rotating stratified turbulence. Journal of Fluid Mechanics, 624:45–55, 2008. ISSN 0022-1120. doi:10.1017/s0022112008004473.DOI
  20. [23]Walter Munk. Internal Waves and Small-Scale Processes. Evolution of Physical Oceanography, pages 264 – 291. 1981.
  21. [24]Guillaume Roullet and Patrice Klein. Cyclone-Anticyclone Asymmetry in Geophysical Turbulence. Physical Review Letters, 104(21):218501, 2010. ISSN 0031-9007. doi:10.1103/physrevlett.104.218501.DOI
  22. [25]Dudley B. Chelton, Michael G. Schlax, and Roger M. Samelson. Global observations of nonlinear mesoscale eddies. Progress In Oceanography, 91(2):167 – 216, 2011. doi:10.1016/j.pocean.2011.01.002. URL http://www.sciencedirect.com/science/article/pii/S0079661111000036.DOI
  23. [26]David L. T. Anderson and Peter D. Killworth. Nonlinear propagation of long Rossby waves. Deep Sea Research Part I: Oceanographic Research Papers, 26(9):1033 – 1049, 1979. doi:10.1016/0198-0149(79)90046-3. URL http://www.sciencedirect.com/science/article/pii/0198014979000463.DOI
  24. [27]J.G. Charney and G.R. Flierl. Oceanic analogues of large-scale atmospheric motions. Evolution of Physical Oceanography, pages 504 – 548. 1981.
  25. [28]V.I. Petviashvili. Red spot of Jupiter and the drift soliton in a plasma. Journal of Experimental and Theoretical Physics Letters, 32:619 – 622, 1980.
  26. [29]B. J. Hoskins, M. E. McIntyre, and A. W. Robertson. On the use and significance of isentropic potential vorticity maps. Quarterly Journal of the Royal Meteorological Society, 111(470):877–946, 1985. ISSN 0035-9009. doi:10.1002/qj.49711147002.DOI

Paper details

vlnc:W_mrzk7jfu
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents