betaFeedback

WorksPhysics and Astronomy

Geometric Mechanics of the Vertical Slice Model

Abstract

The goals of the present work are to: (i) investigate the dynamics of oceanic frontogenesis by taking advantage of the geometric mechanics underlying the class of Vertical Slice Models (VSMs) of ocean dynamics; and (ii) illustrate the versatility and utility of deterministic and stochastic variational approaches by deriving several variants of wave-current interaction models which describe the effects of internal waves propagating within a vertical planar slice embedded in a 3D region of constant horizontal gradient of buoyancy in the direction transverse to the vertical plane.

Introduction

Oceanic fronts form by extracting the gravitational potential energy available from the ‘tilting’ of buoyancy isoclines, represented as horizontal gradients of buoyancy. In turn, the formation of oceanic fronts facilitates the transfer of kinetic energy to small-scale mixing and transport [23, 42]. In particular, a region of constant horizontal gradient of temperature (implying a corresponding gradient of buoyancy) in a three-dimensional, vertically-stratified fluid governed by the Euler-Boussinesq equations can induce the formation of fronts emerging in vertical planar flows transverse to the direction of constant horizontal gradient of temperature [1, 9, 10, 47]. The present work derives Vertical Slice Models (VSMs) to investigate the formation and subsequent evolution of the fronts emerging in a region of constant horizontal gradient of temperature. These VSMs may also be useful for bench-marking numerical schemes, since in 2D they can be run quickly on a single workstation.

Driven by their constant transverse horizontal gradient of temperature, the VSMs in [1, 9, 10, 47] possess a yy-independent solution structure for the flows within the vertical plane which includes the dynamics of the yy-component of the fluid velocity transverse to the vertical plane and the Coriolis force. The yy-independent solution structure for VSM flows is still a solution of the full 3D Euler-Boussinesq fluid equations, as long as the horizontal gradient of the temperature transverse to the vertical slice remains constant. This solution property arises because the pressure gradient within the vertical plane only accesses the yy-independent part of the temperature as it varies with time and space in the vertical plane.

The VSM family can be derived in the Euler–Poincaré (EP) framework of symmetry-reduced Lagrangians in Hamilton’s variational principle [37]. The EP framework involves a constrained Hamilton’s principle expressed in the Eulerian fluid description. Their derivation in the EP framework establishes the following properties of each member of the VSM family: the Kelvin–Noether circulation theorem, conservation of potential vorticity on fluid parcels, a Lie–Poisson Hamiltonian formulation possessing conserved Casimirs arising from particle-relabelling symmetry, a conserved domain-integrated energy and an associated variational principle satisfied by the equilibrium solutions.

Aims of the present work. New theoretical methods of analysing the dynamics of complex fluids advecting a variety of different co-evolving order parameters have been developing in the Euler–Poincaré (EP) framework during the past twenty years. See, e.g., [26, 5, 16, 18] for discussions of these new theoretical methods. The EP framework also offers a means of extending the VSMs to include the new methods for complex fluids, as discussed below.

The new theoretical methods designed for modelling complex fluid flows include the statistical and probabilistic methods needed for data science, as well as the geometric and analytical methods underlying the theory of nonlinear partial differential equations. For example, recent mathematical discoveries have revealed the Poisson structures of the Eulerian description of ideal (non-dissipative) complex fluids such as superfluids, spin glasses, liquid crystals, ferrofluids, etc. derived in [26, 5, 16, 18]. Complex fluids transport order parameters that co-evolve in the frame of the fluid motion, and the dynamics of these order parameters reacts back to affect the fluid motions that transport them.

Motivated by these recent mathematical discoveries, we consider applying the variational methods of the geometric mechanics framework [38] to guide the investigation of the Poisson structures underlying the ocean science of wave and current interactions. We choose this variational approach because of its versatility in deriving the Poisson structures we seek, even in the presence of any random transport that admits the product rule and the chain rule, [27, 17, 12]. The inclusion of randomness in nonlinear ocean dynamics has been shown to be useful for uncertainty quantification and data assimilation methods to reduce uncertainty in computational simulations of ocean models, in particular by using the method of Stochastic Advection by Lie Transport (SALT), which preserves the semidirect-product Poisson structures of classical fluids, [27, 7, 12]. The Euler-Poincaré and Hamilton-Pontryagin versions of Hamilton’s principle have been extended from stochastic paths to geometric rough paths in [12].

The present work illustrates the versatility and utility of the variational framework of geometric mechanics by deriving several variants of wave-current interaction models that describe internal waves propagating in the vertical slice model (VSM) with transverse flow that was originally derived in [9]. Frontogenesis in this model was demonstrated via numerical simulations in [47] and front formation in its solution behaviour has been analysed recently in [1]. From the viewpoint of modelling in ocean dynamics, the problem statement for VSMs is an augmentation of the standard incompressible Euler–Boussinesq equations with a constant gradient of temperature in the transverse direction.

Here, we will first review the derivation of VSMs based on the Euler-Poincaré approach as in [9] and then re-derive it from a variational approach based on a composition of smooth maps to describe the order-parameter dynamics taking place in the frame of the fluid motion. Subsequently, we will use the latter method to augment the VSM to enable inclusion of non-Boussinesq effects and also to add the effects of internal gravity waves propagating in the vertical slice.

The variational approach we take in this paper introduces a certain composition of maps (CoM) into the well-known approach originally due to Clebsch [6]. Remarkably, the CoM approach produces an ‘untangled’ (block-diagonal) Poisson structure for total momentum and order-parameters, as well as an ‘entangled’ Poisson structure for the fluid momentum alone. The latter Poisson structure exhibits a semidirect-product action of the fluid velocity vector field on the order-parameter phase space, as well as an additional symplectic bracket among the order parameters.

Mathematically, the ‘untangled’ version of its Poisson structure separates into the sum of a symplectic two-cocycle bracket in the phase-space of order parameters added to the standard semidirect-product Lie-Poisson bracket for classical fluid dynamics, as derived in the Lagrangian framework of Hamilton’s variational principle [37]. By comparing the two versions of the Poisson structures for these fluid equations with advected co-evolving order parameters, one finds that they exhibit a mathematical equivalence that can be written as T∗Q/G≃T∗(Q/G)⊕g∗T^*Q/G \simeq T^*(Q/G) \oplus\mathfrak{g}^*, where ⊕\oplus denotes the Whitney sum. Perhaps not surprisingly, this duality has persisted and been remarked about throughout the historical investigations of fluid dynamics involving variational principles with a composition of maps.

The history of Clebsch’s approach to the formulation of Hamilton’s principle for classical fluid dynamics is reviewed in [43], where it is supplemented by an additional advection constraint due Lin [41] who introduced it in deriving Landau’s two-fluid classical-quantum model of He2He_2 superfluids. This Clebsch approach with the Lin constraint was supplemented further and applied to He3He_3 superfluids with an additional spin order-parameter in [33], where of course additional two-cocycles were found. Later, it was applied to Yang–Mills charged fluids in [20] and to magneto-hydrodynamics (MHD) and other fluid plasma models, as well as nonlinear elasticity in [34]. The same augmented approach was applied in deriving fluid equations for special-relativistic plasmas coupled to electromagnetic fields in [25] and also for general-relativistic fluids in [24]. Remarkably, the symplectic two-cocycle arising in the Poisson structure for general-relativistic adiabatic fluid dynamics turned out to comprise the Minkowsky metric for space-time gμνg_{\mu\nu} and its canonically conjugate momentum density πμν\pi_{\mu\nu} in the Arnowitt–Deser–Misner theory of general relativity [24]. After this variational approach had been applied to complex fluids in [26], the mathematical essence of the Poisson structures in this approach was proven to arise via Lie group reduction by symmetry with respect to affine actions yielding precisely the T∗Q/G≃T∗(Q/G)⊕g∗T^*Q/G \simeq T^*(Q/G) \oplus\mathfrak{g}^* equivalence as its associated bundle reduction in [4, 16]. This identification and its main mathematical applications for complex fluids are summarised in detail in [16]. A related approach called “metamorphosis” was investigated for shape analysis in image registration with applications to computational anatomy in [39, 3].

While the mathematical setting of affine Poisson structures for this topic is quite rich, [26, 5, 16, 18], the Clebsch-Lin composition of maps variational approach derives the Poisson structure and identifies its corresponding affine Lie group action quite straight-forwardly, as we shall demonstrate later by comparing the original Euler-Poincaré derivation of the EB VSM to the Clebsch-Lin approach.

Hereafter, the Clebsch-Lin approach with affine advection will be referred to simply as composition of maps, as in [30, 32]. The CoM approach can be applied for uncertainty quantification of ocean models by using the Stochastic Advection by Lie Transport (SALT) approach applied in [27, 7, 12] for the composition of several maps. The CoM approach is also compatible with traditional approaches in ocean modelling such as Generalised Lagrangian Mean (GLM). In fact, the CoM approach has been used to provide a stochastic closure for GLM in [31] which will be illustrated further in the present work.

To reprise, the purpose of the present work is to make a concrete application of the stochastic CoM approach for an enhancement of the VSMs of [9] to include the wave-current interaction (WCI) effects of internal gravity waves (IGW) propagating in a vertical slice of Euler-Boussinesq (EB) flow including a transverse velocity. This work is meant to be presented as explicitly as possible so it can provide a useful foundation for further applications of geometric mechanics in ocean modelling.

Motivation of the paper: In the present work, we consider the interaction of internal waves with Euler-Boussinesq flows in a vertical slice of fluid undergoing three dimensional volume-preserving flow whose motion transverse to a vertical plane advects Lagrangian particle labels that depend linearly on the transverse Eulerian coordinate. The constant transverse slope, ss, of the advected Lagrangian particle labels appears as a constant parameter in the slice dynamics along with a new canonically conjugate transverse momentum density, πT\pi_T, dual in the slice to the potential temperature which is advected in three dimensions as a Lagrangian particle label. This concept for a VSM with transverse flow was introduced in [9] and its solution behaviour has been simulated computationally in [47] and analysed recently in [1].

Main goals of the paper: The ocean modelling goal of the present paper is to include IGW motion in the VSMs with transverse flow introduced in [9]. The mathematical goal accompanying the goal for data assimilation is also to determine the Poisson/Hamiltonian structure of the resulting model, and thereby formulate a new model of stochastic parameterisation of advective transport of potential use for quantifying uncertainty in ocean model simulations.

Plan of the paper:

  • Section § 2 recalls the EB VSM given in [9] and considers two Euler-Poincaré derivations of the VSMs.

    • Section § 2.1 provides two equivalent derivations of the EB VSM model treated in [9]. Namely, they are the derivation based on affine Euler-Poincaré reduction and the Clebsch-Lin derivation in the CoM context.

    • Section § 2.3 treats the ‘Entangled’ and ‘Untangled’ versions of the Hamiltonian formulation for the vertical slice model.

    • Section § 2.4 demonstrates some of the solution behaviours exhibited by the VSM derived in the previous sections by showing the result of computational simulations.

  • Section § 3 treats a further asymptotic expansion estimates of IGW interaction dynamics following the wave mean flow interaction (WMFI) closure due to [22] in which the Brunt-Väisälä buoyancy frequency is determined via the Hessian of the fluid pressure.

    • Section § 3.1 treats the Clebsch-Lin derivation of the VSM in the Euler-Boussinesq approximation with the inclusion of a standard WKB model for the IGW Hamiltonian and assumes the existence of a constant Brunt-Väisälä (BV) buoyancy frequency.

  • Section 3.2 reviews the geometry of the generalised Lagrangian mean approach to wave mean flow interaction.

  • Section 3.3 introduces a wave mean flow decomposition of dynamics in the vertical slice of an Euler-Boussinesq fluid.

  • Section 4 then formulates the stochastic parameterisation model for VSMs based on Stochastic Advection by Lie Transport (SALT).

  • Section 5 Concludes with a discussion of the goals achieved in the paper and a survey of potential future work.

  • Appendix A contains an asymptotic expansion which reveals the form of the action for the dynamical system studied in Section 3.1.

  • Appendix B contains the derivation of a new dispersion relation for inertial gravity waves in a vertical slice domain.

At each level of approximation in the sequence of sections in this paper, we identify the opportunities where stochasticity, e.g., SALT, may be introduced that will preserve the geometric Poisson structure of the model. The dynamical effects of the introduction of SALT into the deterministic VSMs derived here and their implications for data science and data assimilation are beyond the scope of the present paper. However, we expect that the framework for applications of the present approach for data calibration, uncertainty quantification and data assimilation considered previously in [7, 8] will certainly carry over for the stochastic IGW effects considered here.

Vertical slice models (VSMs)

The vertical slice model, as proposed by Eady in the 1940s [14, 15, 45], was born from a desire to understand instabilities in atmospheric sciences. It has since proven to be a valuable model for testing numerical methods for geophysical flows, as well as remaining useful for understanding thermal front formation and propagation. Dynamics in a vertical slice model occurs within a vertical (xx-zz) plane, denoted by MM in Figure 1. The slice model features a velocity field tangent to the slice that advects fluid material variables within the slice, as well as a velocity perpendicular to the slice representing transverse flow crossing through the slice at each point. In this section, the vertical slice model will be derived from a variational principle in the Eulerian representation and investigated further by passing to its Hamiltonian formulation. The key idea is to regard the tangential flow in the slice and the perpendicular flow transverse to the slice as the composition of two flow maps.

Image not extracted

Figure 1. The vertical slice domain

Lagrangian formulation of VSMs

Vertical slice models can be derived using variational principles that apply the Lie group structure of their transverse flow explicitly [9]. In general, the flow map of 3D fluids can be modelled by the diffeomorphism group Diff⁡(R3)\operatorname{Diff}(\mathbb{R}^{3}). Let gt∈Diff⁡(R3)g_t \in\operatorname{Diff}(\mathbb{R}^{3}), its action on the coordinates can be written as

gt(X,Y,Z)=(xt(X,Y,Z),yt(X,Y,Z),zt(X,Y,Z)),g_t(X,Y,Z) = (x_t(X,Y,Z), y_t(X,Y,Z), z_t(X,Y,Z)),

where (X,Y,Z)(X,Y,Z) are the initial Lagrangian particle labels at time, t0t_0, and (xt,yt,zt)(x_t,y_t,z_t) denote the Eulerian location of the particle labels (X,Y,Z)(X,Y,Z) at a later time, tt. In the case of VSMs in 3D, their flow map takes the particular form

ϕt(X,Y,Z)=(xt(X,Z),yt(X,Z)+Y,zt(X,Z)),which implies∂ϕt∂Y=(010),(1)\phi_t(X,Y,Z) = (x_t(X,Z), y_t(X,Z) + Y, z_t(X,Z)), \qquad\text{which implies} \qquad\frac{\partial\phi_t}{\partial Y} = \begin{pmatrix} 0\\ 1\\ 0 \end{pmatrix}, \tag*{(1)}

such that the flow in vertical (x(x-z)z) plane is independent of YY and the flow in the yy plane is linear in YY. One may directly verify that diffeomorphisms of the form (1) constitute to a closed subgroup of Diff⁡(R3)\operatorname{Diff}(\mathbb{R}^{3}) and it is isomorphic to the semidirect-product group S:=Diff⁡(M)⋉F(M)S := \operatorname{Diff}(M) \ltimes\mathcal{F}(M). Here, we have M⊆R2M \subseteq\mathbb{R}^{2} as the domain in the vertically sliced plane with coordinates (x,z)(x,z), F(M)\mathcal{F}(M) as the space of smooth functions on MM, and the symbol ⋉\ltimes as the semidirect-product [9, 37]. The Lie algebra corresponding to this group is s:=X(M)⋉F(M)\mathfrak{s} := \mathfrak{X}(M) \ltimes\mathcal{F}(M), where X(M)\mathfrak{X}(M) is the space of vector fields on MM. Our Lie algebra consists of elements of the form (uS,uT)(u_S,u_T), which can be interpreted as components of the velocity within and transverse to the slice, respectively.

In VSMs, there are two types of advected quantities, belonging to spaces denoted by V1∗V_1^* and V2∗V_2^*, which are acted on by the configuration group SS and are assumed to be subspaces of tensor fields T(M)\mathcal{T}(M). Let (ϕ,v)∈S(\phi,v) \in S and a1∈V1∗a_1 \in V_1^* be arbitrary, we define the contragradient representation of SS on V1∗V_1^* to be the pullback by the component of SS belonging to Diff⁡(M)\operatorname{Diff}(M)

ρ(ϕ,v)−1∗a1:=ρϕ−1∗a1:=ϕ∗a1.(2)\rho^*_{(\phi,v)^{-1}}a_1 := \rho^*_{\phi^{-1}}a_1 := \phi^*a_1. \tag*{(2)}

Let (uS,uT)∈s(u_S,u_T) \in\mathfrak{s}, the infinitesimal action induced by the above representation is given by

ρuS∗a1:=ddt∣t=0ρexp⁡((−tuS,−tuT))∗a1=ddt∣t=0ρexp⁡(−tuS)∗a1=LuSa,(3)\rho^*_{u_S}a_1 := \left.\frac{d}{dt}\right|_{t=0}\rho^*_{\exp((-tu_S,-tu_T))}a_1 = \left.\frac{d}{dt}\right|_{t=0}\rho^*_{\exp(-tu_S)}a_1 = \mathcal{L}_{u_S}a, \tag*{(3)}

where L\mathcal{L} is the Lie derivative. Since the induced infinitesimal action is independent of uTu_T, we interpret quantities in V1∗V_1^* as purely advected by the diffeomorphism group Diff⁡(M)\operatorname{Diff}(M) in the slice domain MM. The dual operation to the Lie derivative is the diamond operation, ⋄\diamond, defined through two different pairings as

−⟨b1⋄a1,uS⟩X(M)∗×X(M)=⟨LuSa1,b1⟩V1∗×V1,(4)-\langle b_1 \diamond a_1,u_S\rangle_{\mathfrak{X}(M)^*\times\mathfrak{X}(M)} = \langle\mathcal{L}_{u_S}a_1,b_1\rangle_{V_1^*\times V_1}, \tag*{(4)}

for all b1∈V1b_1 \in V_1. We note that the form of the diamond operation, like the Lie derivative, depends on the concrete choices of the vector space V1V_1 according to its infinitesimal transformation under the Lie derivative. In VSMs, the fluid quantity that is acted on by SS via the contragredient representation is the volume form, D d2x∈Den⁡(M)⊂V1∗D\,d^2x \in\operatorname{Den}(M) \subset V_1^*, which is assumed to be advected. That is, the dynamics of D d2xD\,d^2x is characterised by the pushforward relation

Dt d2xt=ρ(ϕt,vt)∗(D0 d2x0)=ϕ∗(D0 d2x0),(5)D_t\,d^2x_t = \rho^*_{(\phi_t,v_t)}(D_0\,d^2x_0) = \phi_{*}(D_0\,d^2x_0), \tag*{(5)}

where D0 d2x0D_0\,d^2x_0 is the initial reference volume form. Defining the vector field uS⋅∇:=ϕ˙ϕ−1∈X(M)\mathbf{u}_S \cdot\nabla:= \dot{\phi}\phi^{-1} \in\mathfrak{X}(M) where ∇:=(∂x,∂y)\nabla:= (\partial_x,\partial_y) is the gradient operator in the slice domain, we have that the pushforward relation is the solution to the advection equation

∂t(D d2x)+LuS(D d2x)=0⇒∂tD+∇⋅(uSD)=0.(6)\partial_t(D\,d^2x) + \mathcal{L}_{\mathbf{u}_S}(D\,d^2x) = 0 \quad\Rightarrow\quad\partial_tD + \nabla\cdot(\mathbf{u}_S D) = 0. \tag*{(6)}

In relation to the advection of the 3D volume form D d3xD\,d^3x by the full three dimensional flow, one can obtain the advection of in slice volume form D d2xD\,d^2x in the domain MM under the assumptions that both DD and the transverse component of the 3D velocity field uTu_T are independent of the Lagrangian label YY.

References

  1. [1]Alonso-Orán, D. and Bethencourt de León, A., 2019. Stability, well-posedness and blow-up criterion for the Incompressible Slice Model. Physica D: Nonlinear Phenomena, 392, pp.99-118. https://doi.org/10.1016/j.physd.2018.12.005
  2. [2]Andrews, D.G. & McIntyre, M.E., 1978. An exact theory of nonlinear waves on a Lagrangian-mean flow. Journal of fluid Mechanics, 89(4), pp.609-646.DOI
  3. [3]Bruveris, M., Gay-Balmaz, F., Holm, D.D., Ratiu, T.S., 2011. The momentum map representation of images. Journal of Nonlinear Science, 21 (1), (2011), 115-150. https://doi.org/10.1007/s00332-010-9079-5
  4. [4]Cendra, H., Marsden, J.E., Ratiu, T., 2001. Lagrangian Reduction by Stages. Mem. Amer. Math. Soc. V. 152. https://doi.org/10.1090/memo/0722.
  5. [5]Cendra, H., Marsden, J., and Ratiu, T.S., 2004. Cocycles, Compatibility, and Poisson Brackets for Complex Fluids. In: Capriz, G., Mariano, P.M. (eds) Advances in Multield Theories for Continua with Substructure. Modeling and Simulation in Science, Engineering and Technology. Birkhäuser, Boston, MA. https://doi.org/10.1007/978-0-8176-8185-6_3
  6. [6]Clebsch A., 1859 Uber die Integration der hydrodynamischen Gleichungen 1. reine angew. Math. 56 1-10. https://eudml.org/doc/147740
  7. [7]Cotter, C., Crisan, D., Holm, D. D., Pan, W., and Shevchenko, I., 2018. Numerically modelling stochastic Lie transport in fluid dynamics. SIAM Multiscale Model. Simul., 17(1), 192–232. https://doi.org/10.1137/18M1167929
  8. [8]Cotter, C., Crisan, D., Holm, D. D., Pan, W., and Shevchenko, I., 2020. A Particle Filter for Stochastic Advection by Lie Transport: A Case Study for the Damped and Forced Incompressible Two-Dimensional Euler Equation SIAM/ASA Journal on Uncertainty Quantification 8:4, 1446-1492 https://doi.org/10.1137/19M1277606
  9. [9]Cotter, C.J. and Holm, D.D., 2013. A variational formulation of vertical slice models. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 469(2155), p.20120678. http://dx.doi.org/10.1098/rspa.2012.0678
  10. [10]Cotter, C.J. and Holm, D.D., 2014. Variational formulations of sound-proof models. Quarterly Journal of the Royal Meteorological Society, 140(683), pp.1966-1973. https://doi.org/10.1002/qj.2260
  11. [12]Crisan D., Holm D. D., Leahy J.M., and Nilssen T., 2022. Variational principles for fluid dynamics on rough paths. Adv. Math., 404:Paper No. 108409, 75. https://doi.org/10.1016/j.aim.2022.108409
  12. [13]Cotter, C.J. and Shipton, J., 2012. Mixed finite elements for numerical weather prediction. Journal of Computational Physics, 231 (21), 7076–7091.DOI
  13. [14]Eady, E. T., 1948. The theory of development in dynamical meteorology. PhD thesis, Imperial College London.DOI
  14. [15]Eady, E. T., 1949. Long waves and cyclone waves, Tellus, 1. 33–52.DOI
  15. [16]Gay-Balmaz F., Ratiu T.S., 2009. The geometric structure of complex fluids. Adv. Appl. Math., 42, 176-275. https://doi.org/10.1016/j.aam.2008.06.002
  16. [17]Gay-Balmaz, F. and Holm, D.D., 2018. Stochastic geometric models with non-stationary spatial correlations in Lagrangian fluid flows. Journal of nonlinear science, 28, pp.873-904. https://doi.org/10.1007/s00332-017-9431-0
  17. [18]Gay-Balmaz, F., Ratiu, T.S. and Tronci, C., 2013. Equivalent theories of liquid crystal dynamics. Archive for Rational Mechanics and Analysis, 210(3), pp.773-811. https://doi.org/10.1007/s00205-013-0673-1
  18. [19]Gent, P. R., and J. C. Mcwilliams, 1990: Isopycnal Mixing in Ocean Circulation Models. J. Phys. Oceanogr., 20, 150–155, https://doi.org/10.1175/1520-0485(1990)020<0150:IMUCM>2.0.CO;2.
  19. [20]Gibbons J., Holm, D.D. and Kupershmidt B.A., 1983. The Hamiltonian structure of classical chromohydrodynamics, Phys. D 6 179–194DOI
  20. [21]Gilbert, A., & Vanneste, J., 2018. Geometric generalised Lagrangian-mean theories. Journal of Fluid Mechanics, 839, 95-134. https://doi.org/10.1017/jfm.2017.913
  21. [22]Gjaja, I. and Holm, D.D., 1996. Self-consistent Hamiltonian dynamics of wave mean-flow interaction for a rotating stratified incompressible fluid. Physica D: Nonlinear Phenomena, 98(2-4), pp.343-378. https://doi.org/10.1016/0167-2789(96)00104-2
  22. [23]Gula, J., Taylor, J., Shcherbina, A. and Mahadevan, A., 2022. Submesoscale processes and mixing. In Ocean Mixing (pp. 181-214). Elsevier.DOI
  23. [24]Holm, D.D., 1985. Hamiltonian Formalism for General Relativistic Adiabatic Fluids, Physica D 23, 1–36. https://doi.org/10.1016/0167-2789(85)90131-9
  24. [25]Holm, D.D., 1987. Hamiltonian Dynamics and Stability Analysis of Neutral Electromagnetic Fluids with Induction, Physica D 25, 261–287. https://doi.org/10.1016/0167-2789(87)90104-7
  25. [26]Holm, D.D., 2002. Euler-Poincaré Dynamics of Perfect Complex Fluids. In: Newton, P., Holmes, P., Weinstein, A. (eds) Geometry, Mechanics, and Dynamics. Springer, New York, NY. https://doi.org/10.1007/0-387-21791-6_4
  26. [27]Holm, D.D., 2015. Variational principles for stochastic fluid dynamics. Proceedings of the Royal Society A. 471, 20140963. https://doi.org/10.1098/rspa.2014.0963
  27. [28]Holm, D.D., 2019. Stochastic Closures for Wave-Current Interaction Dynamics. J Nonlinear Sci 29: (6) 2987-3031. https://doi.org/10.1007/s00332-019-09565-0
  28. [29]Holm, D.D., 2002. Lagrangian averages, averaged Lagrangians, and the mean effects of fluctuations in fluid dynamics. Chaos. 12 (2), 518–530. https://doi.org/10.1063/1.1460941
  29. [30]Holm, D.D., Hu, R., and Street, O.D., 2023. Coupling of Waves to Sea Surface Currents Via Horizontal Density Gradients. In: Chapron, B. et al. (eds) Stochastic Transport in Upper Ocean Dynamics. STUOD 2021. Mathematics of Planet Earth, vol 10. Springer, Cham. https://doi.org/10.1007/978-3-031-18988-3_8
  30. [31]Holm, D.D., Hu, R. and Street, O.D., 2023. On the interactions between mean flows and inertial gravity waves. arXiv preprint arXiv:2302.04838.arxiv.org/abs/2302.04838
  31. [32]Holm, D.D., Hu, R., & Street, O.D., 2023. Lagrangian reduction and wave mean flow interaction. Physica D: Nonlinear Phenomena, 133847. https://doi.org/10.1016/j.physd.2023.133847
  32. [33]Holm, D.D. and Kupershmidt B.A., 1982. Poisson structures of superfluids. Phys. Lett. A, 91 (9) 425–430 https://doi.org/10.1016/0375-9601(82)90740-X
  33. [34]Holm, D.D. and Kupershmidt B.A., 1983. Poisson Brackets and Clebsch Representations for Magnetohydrodynamics, Multifuild Plasmas, and Elasticity. Physica D 6 347–363. https://doi.org/10.1016/0167-2789(83)90017-9
  34. [37]Holm, D.D., Marsden, J.E. and Ratiu, T.S., 1998. The Euler–Poincaré equations and semidirect products with applications to continuum theories, Adv. in Math., 137 1-81, https://doi.org/10.1006/aima.1998.1721
  35. [38]Holm, D.D., Schmah, T. and Stoica, C., 2009. Geometric mechanics and symmetry: from finite to infinite dimensions (Vol. 12). Oxford University Press.DOI
  36. [39]Holm, D.D., Trouvé, A. and Younes, L., 2009. The Euler-Poincaré theory of metamorphosis. Quarterly of Applied Mathematics, 67 661-685. https://doi.org/10.1090/S0033-569X-09-01134-2
  37. [41]Lin, C.C., 1963. Liquid Helium. In Proceedings of the Enrico Fermi International School of Physics, Course XXI, G. Careri, editor, Academic Press, New York, pp. 93–146.
  38. [42]McWilliams, J.C., 2021. Oceanic frontogenesis. Annual Review of Marine Science, 13, pp.227-253.DOI
  39. [43]Serrin, J., 1959. Mathematical principles of classical fluid mechanics. Fluid Dynamics I/Strömungsmechanik I, pp.125-263.DOI
  40. [44]Street, O. D. and Crisan, D., 2021. Semi-martingale driven variational principles. Proceedings of the Royal Society A. 477, 20200957. https://doi.org/10.1098/rspa.2020.0957
  41. [45]Vallis, G.K., 2017. Atmospheric and oceanic fluid dynamics. Cambridge University Press.
  42. [46]Visram, A.R., Cotter, C.J. and Cullen, M.J.P., 2014. A framework for evaluating model error using asymptotic convergence in the Eady model. Q.J.R. Meteorol. Soc., 140: 1629-1639. https://doi.org/10.1002/qj.2244
  43. [47]Yamazaki, H., Shipton, J., Cullen, M.J., Mitchell, L. and Cotter, C.J., 2017. Vertical slice modelling of nonlinear Eady waves using a compatible finite element method. Journal of Computational Physics, 343, pp.130-149.DOI

Paper details

vlnc:W_jh5hjjvn
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents