05.14.01 · symplectic / topological-hydrodynamics

Helicity as a Casimir invariant of the ideal fluid

shipped3 tiersLean: none

Anchor (Master): Arnold 1974 *The asymptotic Hopf invariant and its applications* (Proc. Summer School Diff. Eqns., Erevan; repr. Sel. Math. Sov. 5, 1986); Moreau 1961 *Constantes d'un îlot tourbillonnaire en fluide parfait barotrope* (C. R. Acad. Sci. Paris 252); Moffatt 1969 *The degree of knottedness of tangled vortex lines* (J. Fluid Mech. 35, 117-129); Arnold-Khesin *Topological Methods in Hydrodynamics* (Springer Applied Math. Sci. 125, 2nd ed. 2021) Ch. I §1-§4 and Ch. III §1-§2; Marsden-Ratiu *Introduction to Mechanics and Symmetry* §13.7; Khesin-Wendt *The Geometry of Infinite-Dimensional Groups* (Springer 2009) Ch. IV

Intuition Beginner

Imagine the flow of an ideal fluid — water with no friction — filling a closed region of space. At every point the fluid moves in some direction, and the field of all these arrows is the velocity. Wherever the velocity twists around itself, the fluid carries a little local spin, a tendency to rotate, which is captured by a second field called the vorticity. Helicity is one single number, computed from the whole flow at one instant, that measures how much the velocity and the vorticity wind around together — the total amount of corkscrewing in the motion.

A right-handed corkscrew and a left-handed one contribute with opposite signs, so a flow with no net handedness has helicity zero. A flow where the spin tends to point along the direction of motion has large positive helicity.

The surprise is that this number does not change as the fluid evolves. As the flow swirls and stretches and folds, the velocity field changes completely from one moment to the next — yet the helicity stays locked at its starting value. This is a conservation law, like the conservation of energy, but it comes from a different source. Energy is conserved because the equations do not care about the clock. Helicity is conserved for a deeper reason: an ideal fluid does not care which parcel of water you painted which colour. You may relabel the parcels however you like and the physics is the same. That relabelling freedom is a symmetry, and helicity is the bookkeeping quantity it protects.

The right way to picture what helicity counts is linkage. Bundle the vorticity into thin tubes threaded through the fluid. If two tubes are linked like two rings of a chain, that linkage shows up in the helicity. Unlinked tubes give small helicity; knotted, threaded tubes give large helicity. Because the fluid carries its vortex tubes along without ever letting them cut through one another, the linkage can never change, and so neither can the helicity. This picture connects a number from calculus to a fact about knots, and it opens the whole subject of topological fluid mechanics.

Visual Beginner

Picture two closed loops of vortex tube floating in the fluid, each one a doughnut-shaped ring along which the fluid spins. Draw them threaded through one another, like two links of a chain. Each loop carries an amount of swirl, a circulation; call these numbers $a$ and $b$ . The helicity of this two-loop configuration turns out to be twice the product $a \times b$ , multiplied by how many times the loops pass through each other — their linking number.

The picture makes three points at once. First, two rings lying side by side, not threaded through one another, have linking number zero and contribute nothing to the helicity. Second, the more times the rings wind through each other, the larger the helicity. Third, and most important, you cannot unlink the rings without passing one tube through the other — and the fluid never lets that happen. So the linkage is frozen, the helicity is frozen, and a flow that starts with linked vortex tubes can never evolve into one with unlinked tubes. The number on the page is a topological fact wearing the costume of an integral.

Worked example Beginner

Take the two-ring picture and put real numbers on it. Suppose ring one carries circulation $a = 3$ and ring two carries circulation $b = 2$ , and suppose the two rings are linked exactly once, so their linking number is $n = 1$ . The formula for the helicity of a pair of linked vortex tubes is $$ \mathcal{H} = 2,n,a,b. $$

Step 1. Plug in the numbers: $H = 2 \times 1 \times 3 \times 2 = 12$ .

Step 2. Now let the fluid evolve. The rings drift, wobble, and stretch into complicated shapes. The circulation of each ring is itself a conserved quantity of ideal flow (this is Kelvin's circulation law, developed in continuum mechanics 09.07.01), so $a$ stays $3$ and $b$ stays $2$ . The linking number $n$ is a whole number that counts how many times one ring passes through the other; a whole number cannot change a little bit, and the only way it could jump would be for one tube to cut through the other, which ideal flow forbids. So $n$ stays $1$ .

Step 3. Therefore $H = 2 \times 1 \times 3 \times 2 = 12$ at every later time. The helicity is frozen at $12$ .

Step 4. Compare a side-by-side pair, unlinked, with the same circulations $a = 3$ , $b = 2$ but linking number $n = 0$ . Then $H = 2 \times 0 \times 3 \times 2 = 0$ , and it stays $0$ forever. The two configurations have different helicities and can never be deformed into one another by an ideal flow.

What this tells us: a single number, computed once, decides whether one flow can ever turn into another. The helicity $12$ and the helicity $0$ live in different worlds, separated by a barrier the fluid cannot cross.

Check your understanding Beginner

Exercise (easy, true-false).

True or false: helicity is conserved by ideal flow for the same reason that energy is conserved — because the equations do not depend on time.

Hint

Energy conservation comes from time-symmetry. Helicity conservation comes from a different symmetry — the freedom to relabel fluid parcels.

Answer

False. Energy conservation does come from time-independence of the equations. Helicity conservation comes from a separate symmetry: the relabelling freedom of an ideal fluid, which does not care which parcel of fluid is which. Helicity is conserved no matter what the energy of the flow is doing; it is a different kind of conservation law, called a Casimir invariant. Feedback: the deep point is that helicity is conserved for every possible energy rule, not just the one nature picks — it is protected by the geometry of the fluid, not by the particular forces in play.

Exercise (easy, multiple choice).

A flow has all its vortex tubes lying parallel and unlinked. What can you say about its helicity?

A. It is large and positive
B. It is zero
C. It is negative
D. It cannot be determined

Hint

Helicity measures linkage. Unlinked tubes contribute no linkage.

Answer

B. Helicity counts the linkage of vortex tubes, weighted by their circulations. Parallel, unlinked tubes have linking number zero and contribute nothing, so the helicity is zero. Feedback-correct: correct — and because helicity is frozen, such a flow can never spontaneously develop linked vortex tubes under ideal evolution. Feedback-wrong: large positive or negative helicity would require the tubes to be linked or to corkscrew around one another, which the parallel-and-unlinked configuration excludes.

Formal definition Intermediate+

Let $(M, g)$ be a closed oriented Riemannian 3-manifold (no boundary), with Riemannian volume form $μ \in Ω^{3} (M)$ . For a vector field $u \in X (M)$ write $u^{♭} \in Ω^{1} (M)$ for the metric-dual 1-form. The vorticity is the vector field $ω = curl u$ , defined by $ι_{ω} μ = d u^{♭}$ ; equivalently $ω = (⋆ d u^{♭})^{♯}$ where $⋆$ is the Hodge star.

Configuration space. The configuration space of an ideal incompressible fluid is the group $G = SDiff (M)$ of volume-preserving diffeomorphisms of $M$ . Its Lie algebra is $g = X_{vol} (M)$ , the divergence-free vector fields, with bracket $[u, v] = - [u, v]_{Lie}$ (the right-invariant convention of 05.09.05). The Euler equation of an ideal incompressible fluid, $$ \partial_t u + (u \cdot \nabla) u = -\nabla p, \qquad \operatorname{div} u = 0, $$ is the Euler-Arnold equation on $G$ for the $L^{2}$ kinetic-energy metric 05.09.05.

The dual and the vorticity. The regular dual $g^{*}$ is realised as $Ω^{1} (M) / d Ω^{0} (M)$ : a divergence-free $u$ pairs with a coset $[α]$ by $⟨[α], u ⟩ = \int_{M} α (u) μ$ , and the ambiguity $α \mapsto α + df$ pairs to zero against divergence-free fields. The class $[u^{♭}] \in g^{*}$ is the momentum; its exterior derivative $d u^{♭} = ι_{ω} μ$ is the vorticity, a well-defined element of the exact 2-forms $d Ω^{1} (M)$ .

Helicity. The helicity of a divergence-free vector field $u$ on $M$ is $$ \mathcal{H}(u) ;=; \int_M u \cdot \omega ; \mu ;=; \int_M u^\flat \wedge d u^\flat . $$ Both expressions agree: $u \cdot ω μ = u^{♭} \land ι_{ω} μ = u^{♭} \land d u^{♭}$ . On a closed $M$ the integral is independent of the choice of representative $u^{♭}$ within $[u^{♭}]$ , because replacing $u^{♭}$ by $u^{♭} + df$ changes the integrand by $df \land d u^{♭} = d (f d u^{♭})$ , which integrates to zero by Stokes' theorem (the same gauge-independence holds when $u$ is tangent to a boundary with $ω$ null on it). Thus $H$ descends to a well-defined functional on $g^{*}$ .

Lie-Poisson structure. The dual $g^{*}$ carries the Lie-Poisson bracket 05.03.01 $$ {F, G}(\mu_*) = \Big\langle \mu_*, \Big[ \frac{\delta F}{\delta \mu_*}, \frac{\delta G}{\delta \mu_*} \Big] \Big\rangle , $$ whose symplectic leaves are the coadjoint orbits — here the isovorticity sets, the orbits of the vorticity under pullback by $SDiff (M)$ . A functional $C : g^{*} \to R$ is a Casimir if ${C, F} = 0$ for every $F$ , equivalently if $ad_{δ C / δ μ_{*}}^{*} μ_{*} = 0$ , equivalently if $C$ is constant on every coadjoint orbit.

Sign and orientation conventions. Helicity changes sign under an orientation-reversing diffeomorphism of $M$ (it is a pseudoscalar): a right-handed flow and its mirror image have opposite helicity. We fix the orientation of $M$ once and for all through the volume form $μ$ , and all helicities below are computed with that fixed orientation.

Key theorem with proof Intermediate+

Theorem (Moreau 1961; Moffatt 1969; Arnold 1974 — helicity is a Casimir). Let $M$ be a closed oriented Riemannian 3-manifold. The helicity $H (u) = \int_{M} u^{♭} \land d u^{♭}$ is a Casimir of the Lie-Poisson structure on $\mathfrak{g}^ = \mathfrak{X}_{\mathrm{vol}}(M)^ $. C o n se q u e n tl y$ \mathcal{H}$ is conserved along the Euler flow for every choice of Hamiltonian — in particular along the ideal-fluid Euler equation.

Proof. Work with the vorticity 2-form $ϖ := d u^{♭} = ι_{ω} μ$ , the natural coordinate on $g^{*}$ . The coadjoint action of $SDiff (M)$ on $g^{*}$ is pullback of the vorticity: for $ϕ \in SDiff (M)$ , the coadjoint orbit of $ϖ$ consists of all $ϕ^{*} ϖ$ . The infinitesimal coadjoint action by $v \in g$ is the Lie derivative, $ad_{v}^{*} ϖ = L_{v} ϖ$ .

First, $H$ is constant on each coadjoint orbit. Let $ϕ \in SDiff (M)$ and let $u$ have momentum $[u^{♭}]$ and vorticity $ϖ$ . The pulled-back field $ϕ^{*} u$ has momentum $[ϕ^{*} u^{♭}]$ and vorticity $ϕ^{*} ϖ = d (ϕ^{*} u^{♭})$ . Then $$ \mathcal{H}(\phi^* u) = \int_M \phi^* u^\flat \wedge d(\phi^* u^\flat) = \int_M \phi^!\big( u^\flat \wedge d u^\flat \big) = \int_M u^\flat \wedge d u^\flat = \mathcal{H}(u), $$ using that $\phi^ $co mm u t es w i t h$ d $an d w i t h w e d g e, an d t ha t$ \phi $, p r eser v in g$ \mu $an d t h eor i e n t a t i o n, p r eser v es t h e in t e g r a l . S o$ \mathcal{H} $t ak eso n e v a l u eo n t h e w h o l e i so v or t i c i t y or bi tt h r o ug h$ u$.

Infinitesimally, this says $L_{v} H = 0$ for every divergence-free $v$ . Compute the variation directly. Under $u^{♭} \mapsto u^{♭} + ϵ L_{v} u^{♭}$ , $$ \frac{d}{d\epsilon}\Big|_0 \mathcal{H} = \int_M \big( \mathcal{L}_v u^\flat \wedge d u^\flat + u^\flat \wedge d,\mathcal{L}_v u^\flat \big) = \int_M \mathcal{L}_v\big( u^\flat \wedge d u^\flat \big), $$ where the second equality regroups the two terms using $d L_{v} = L_{v} d$ . By Cartan's formula $L_{v} = d ι_{v} + ι_{v} d$ applied to the 3-form $u^{♭} \land d u^{♭}$ (top degree, so $d$ of it is zero), $L_{v} (u^{♭} \land d u^{♭}) = d ι_{v} (u^{♭} \land d u^{♭})$ , an exact form, and its integral over the closed manifold $M$ vanishes by Stokes. Hence the directional derivative of $H$ along every coadjoint direction is zero.

Finally, the variational derivative $δ H / δ ϖ$ pairs to zero against every coadjoint vector, which is the statement $ad_{δ H / δ ϖ}^{*} ϖ = 0$ , the defining condition of a Casimir 05.03.01. A Casimir Poisson-commutes with every functional, so for any Hamiltonian $H$ the bracket ${H, H} = 0$ , and along the Hamiltonian flow $\dot{H} = {H, H} = 0$ . Taking $H$ to be the kinetic energy $\frac{1}{2} \int_{M} ∣ u ∣^{2} μ$ gives conservation of helicity along the ideal-fluid Euler equation. $□$

Bridge. This result builds toward the entire topological theory of the chapter, and the foundational reason is the same one that organises the rigid body in 05.03.01: a Casimir is constant on coadjoint orbits because the orbits are the symplectic leaves, and a functional constant on leaves Poisson-commutes with everything. This is exactly the rigid-body Casimir $∣ L ∣^{2}$ — the squared angular momentum, constant on the spheres that foliate $so (3)^{*}$ — transported to the infinite-dimensional fluid group, where the spheres become isovorticity orbits and the Casimir becomes helicity. The bridge is the observation that conservation here is independent of the Hamiltonian: helicity is conserved by ideal Euler flow not because of what the energy is, but because of the relabelling symmetry that degenerates the Poisson structure, and this generalises every Hamiltonian we could write on $g^{*}$ . The handedness that the integral measures appears again in 05.14.02 as an average asymptotic linking number, and the same invariant migrates to the magnetic setting in 05.14.04; putting these together, helicity is the first and most basic of the topological obstructions that make an ideal fluid remember the knottedness of its own vortex lines.

Exercises Intermediate+

Exercise 4 (medium, short-answer).

Explain why the family of functionals $C_{F} (ω) = \int_{M} F (ω)$ — Casimirs of the two-dimensional Euler equation, with $ω$ the scalar vorticity and $F$ arbitrary — has no direct three-dimensional analogue, and why helicity is the natural 3D replacement.

Hint

In 2D the vorticity is a scalar advected by the flow; in 3D it is a vector that is also stretched. Which pointwise functions of $ω$ survive stretching?

Answer

In two dimensions the vorticity $ω$ is a scalar transported as a density along the flow, $\partial_{t} ω + u \cdot \nabla ω = 0$ , so any $\int_{M} F (ω) μ$ (the enstrophy Casimirs and their generalisations) is conserved: the distribution of vorticity values is merely rearranged. In three dimensions the vorticity is a vector field obeying the Helmholtz equation $\partial_{t} ω = L_{u} ω$ — transported and stretched — so pointwise functions of $∣ ω ∣$ are not conserved; vortex stretching changes the magnitude. The quantity that survives is not a pointwise function of $ω$ but the global pairing $\int_{M} u^{♭} \land d u^{♭}$ of the momentum with its own vorticity, which is invariant under the relabelling group because it is built from the de Rham wedge, insensitive to how the flow stretches individual tubes. Helicity is thus the 3D Casimir, replacing the infinite 2D family with (generically) a single global invariant. Rubric: full credit for the 2D-rearrangement / 3D-stretching contrast and the identification of helicity as the surviving invariant.

Exercise 5 (medium, short-answer).

State the Helmholtz vorticity transport equation and use it to give a direct (non-Casimir) proof that $\frac{d}{d t} H = 0$ along the Euler flow on a closed $M$ .

Hint

The vorticity 2-form $ϖ = d u^{♭}$ satisfies $\partial_{t} ϖ + L_{u} ϖ = 0$ . Differentiate $H = \int_{M} u^{♭} \land ϖ$ in time and use the Euler equation $\partial_{t} u^{♭} = - L_{u} u^{♭} + d (\frac{1}{2} ∣ u ∣^{2} - p)$ .

Answer

Helmholtz transport: $\partial_{t} ϖ = - L_{u} ϖ$ , with $ϖ = d u^{♭}$ . The Euler equation in geometric form reads $\partial_{t} u^{♭} = - L_{u} u^{♭} + d B$ where $B = \frac{1}{2} ∣ u ∣^{2} - p$ is the Bernoulli function. Differentiate $H = \int_{M} u^{♭} \land ϖ$ : $$ \dot{\mathcal{H}} = \int_M \partial_t u^\flat \wedge \varpi + u^\flat \wedge \partial_t \varpi = \int_M (-\mathcal{L}_u u^\flat + dB)\wedge \varpi - u^\flat \wedge \mathcal{L}_u \varpi. $$ The two Lie-derivative terms combine: $L_{u} u^{♭} \land ϖ + u^{♭} \land L_{u} ϖ = L_{u} (u^{♭} \land ϖ)$ , whose integral is zero (Cartan + Stokes on closed $M$ , since $u$ is divergence-free). The remaining term is $\int_{M} d B \land ϖ = \int_{M} d B \land d u^{♭} = \int_{M} d (B d u^{♭}) = 0$ by Stokes. Hence $\dot{H} = 0$ . This recovers the Casimir conclusion by a hands-on computation. Rubric: full credit for the transport equation, the Lie-derivative cancellation, and the Bernoulli-term Stokes argument.

Exercise 6 (medium, short-answer).

On a domain $M \subset R^{3}$ with boundary, helicity is generally not gauge-invariant. State the boundary condition on $u$ and $ω$ under which $\int_{M} u^{♭} \land d u^{♭}$ is well-defined, and identify the obstruction when that condition fails.

Hint

Re-run the gauge computation $\int_{M} df \land d u^{♭} = \int_{M} d (f d u^{♭})$ on a domain with boundary; the Stokes term now picks up $\int_{\partial M} f d u^{♭}$ .

Answer

On a domain with boundary, the gauge change adds $\int_{M} d (f d u^{♭}) = \int_{\partial M} f d u^{♭} = \int_{\partial M} f ι_{ω} μ$ by Stokes. This vanishes for all $f$ if and only if the pullback of $ι_{ω} μ$ to $\partial M$ is zero, i.e. the vorticity has no flux through the boundary: $ω$ is tangent to $\partial M$ (the field-line condition $ω \cdot n = 0$ on $\partial M$ ). Under this null-boundary condition — vortex lines closed inside $M$ or tangent to the wall — helicity is gauge-independent and a genuine invariant. When vortex lines pierce the boundary, helicity depends on the gauge of $u^{♭}$ and one must instead use relative helicity, subtracting a reference potential field with the same boundary flux (Berger-Field 1984). Rubric: full credit for the tangency/null-flux condition and naming the relative-helicity remedy.

Exercise 7 (hard, short-answer).

Show that helicity is independent of the Riemannian metric used to define the curl — that is, it depends on $M$ only through its volume form and orientation, not through $g$ . Explain why this makes helicity a topological rather than a geometric invariant.

Hint

Rewrite $H$ purely in terms of the divergence-free flux 2-form $β = ι_{u} μ$ and its potential, with no Hodge star in sight.

Answer

A divergence-free field $u$ corresponds to the closed 2-form $β = ι_{u} μ$ (closed because $d β = L_{u} μ = (div u) μ = 0$ ). On a closed $M$ with $[β] = 0 \in H^{2} (M)$ , write $β = d α$ . The helicity equals $H (u) = \int_{M} α \land d α$ , the self-linking / Hopf integral of the 1-form $α$ , with no metric appearing: the wedge and exterior derivative are purely smooth-structure operations, and $μ$ enters only to define the correspondence $u \leftrightarrow β$ and the orientation. One checks $α \land d α = u^{♭} \land d u^{♭}$ up to an exact form when $α = u^{♭}$ , but the point is that the value needs only the volume form and orientation. Therefore two metrics with the same volume form assign the same helicity to the same field. This is why helicity is topological: it is invariant under the volume-preserving (not merely isometric) diffeomorphism group, the full relabelling symmetry, and it measures linking — a homotopy-theoretic datum — rather than any length or angle. The metric is scaffolding for the curl; it leaves no trace in the answer. Rubric: full credit for the metric-free reformulation via $β = d α$ and the topological-invariance conclusion.

Exercise 8 (hard, short-answer).

Derive the linked-vortex-ring formula $H = 2 n a b$ used in the Beginner worked example, starting from $H = \int_{M} u \cdot ω μ$ with the vorticity concentrated on two thin tubes $T_{1}, T_{2}$ of circulations $a, b$ and linking number $n$ .

Hint

Model $ω$ as supported on the two tubes, write $u$ as the Biot-Savart field of $ω$ , and reduce $\int u \cdot ω$ to a double sum of Gauss linking integrals between the tubes.

Answer

Concentrate vorticity on two thin closed tubes: $ω μ \approx a δ_{T_{1}} + b δ_{T_{2}}$ in the sense of currents, where $δ_{T_{i}}$ is the 1-current of the core curve $C_{i}$ . The velocity is the Biot-Savart field $u = curl^{- 1} ω$ , so $H = \int_{M} u \cdot ω μ = ⟨ curl^{- 1} ω, ω ⟩$ . Expanding $ω = a C_{1} + b C_{2}$ bilinearly gives four terms: $a^{2} Slk (C_{1}) + b^{2} Slk (C_{2}) + 2 ab Lk (C_{1}, C_{2})$ , where $Slk$ is the self-linking (writhe + twist) of a single tube and $Lk$ is the Gauss linking integral $\frac{1}{4 π} \oint\oint \frac{( x _{1} - x _{2} ) \cdot ( d x _{1} \times d x _{2} )}{∣ x _{1} - x _{2} ∣ ^{3}} = n$ . For untwisted, unknotted tubes the self-linking terms vanish, leaving the cross term $H = 2 ab Lk (C_{1}, C_{2}) = 2 nab$ . The factor $2$ is the symmetry of the bilinear pairing — tube 1 links tube 2 and tube 2 links tube 1, contributing equally. This is Moffatt's 1969 identification of helicity with total linkage, and the appearance of the Gauss integral is the seed of the asymptotic-linking theorem 05.14.02. Rubric: full credit for the current model, the bilinear expansion, the Gauss-integral identification of $n$ , and the symmetry factor.

Advanced results Master

The Casimir property places helicity inside the general theory of degenerate Poisson structures 05.03.01. The Lie-Poisson bracket on $X_{vol} (M)^{*}$ has, as symplectic leaves, the isovorticity coadjoint orbits; a functional constant on every leaf is a Casimir, and the catalogue of Casimirs is precisely the catalogue of orbit invariants. In two dimensions the orbits are classified by the rearrangement class of the scalar vorticity, and the Casimirs form the infinite family $C_{F} = \int_{M} F (ω) μ$ for arbitrary $F$ . In three dimensions the orbit structure is governed by the topology of the vorticity field, and helicity is the leading invariant — generically the only smooth Casimir for an aperiodic vorticity field, supplemented by the circulation Casimirs $\oint_{γ} u^{♭}$ on any closed vortex line $γ$ when such lines exist.

The metric-independence established in Exercise 7 sharpens to a precise statement: helicity is invariant under the action of the full group $SDiff (M)$ , the relabelling symmetry, and depends on $u$ only through the isotopy class of its vorticity field's integral curves together with their circulations. This is the algebraic shadow of the asymptotic-linking interpretation of 05.14.02: helicity is the integral over $M \times M$ of the asymptotic linking number of pairs of orbits of the vorticity flow, Arnold's asymptotic Hopf invariant. The Casimir computation of this unit and the ergodic computation of 05.14.02 are two faces of the same invariant — the present unit derives its conservation, the next derives its topological meaning.

There is one structural caveat that distinguishes the fluid case from finite dimensions. For a finite-dimensional Lie-Poisson space the Casimirs cut out the symplectic leaves exactly; for $SDiff (M)^{*}$ the coadjoint orbits are not closed and the Casimir functionals do not separate them. Helicity is a continuous invariant of the orbit, but two fields with equal helicity need not lie on the same orbit — the signature of the vorticity (its full linking structure) is finer than its single helicity number. This is the infinite-dimensional reason the helicity-as-linking theorem 05.14.02 is an integral average of linking, not a complete invariant, and it is the entry point to the Hopf-invariant unknotting obstruction of 05.14.03, where higher topological data beyond helicity controls which vortex configurations are mutually reachable.

Synthesis. Putting these together, helicity is the foundational reason an ideal fluid remembers topology: this is exactly the rigid-body Casimir $∣ L ∣^{2}$ of 05.03.01 reincarnated on the diffeomorphism group, and it generalises to the magnetic-helicity Casimir $\int A \land d A$ of ideal magnetohydrodynamics 05.14.04, which is dual to it under the velocity-magnetic-field correspondence. The central insight is that conservation here is Hamiltonian-independent — the bridge from the Lie-Poisson degeneracy of 05.03.01 to the frozen-linkage picture of 05.14.02 runs through a single fact, that a functional constant on coadjoint orbits commutes with every observable. This pattern recurs throughout the chapter: the energy-Casimir stability method of 05.14.05 sums energy and helicity-type Casimirs to manufacture a Lyapunov functional, and the Hopf-invariant obstruction of 05.14.03 refines the single number into the full linking signature. Helicity is where the geometry of the coadjoint orbit and the topology of knotted vortex tubes first coincide, and every later result in topological hydrodynamics builds toward exploiting that coincidence.

Full proof set Master

Proposition 1 (coadjoint action is vorticity pullback). For the right-invariant convention on $G = SDiff (M)$ , the coadjoint action of $ϕ \in G$ on $\mathfrak{g}^ $i s$ \operatorname{Ad}^_\phi[u^\flat] = [\phi^ u^\flat] $, an d in f ini t es ima l l y$ \operatorname{ad}^_v[u^\flat] = [\mathcal{L}_v u^\flat] $, e q u i v a l e n tl y$ \operatorname{ad}^_v \varpi = \mathcal{L}_v \varpi $o n t h e v or t i c i t y$ \varpi = du^\flat$.*

Proof. The coadjoint action is dual to the adjoint action $Ad_{ϕ} v = ϕ_{*} v$ on $g$ . For $[α] \in g^{*}$ and $v \in g$ , $⟨ Ad_{ϕ}^{*} [α], v ⟩ = ⟨[α], Ad_{ϕ^{- 1}} v ⟩ = \int_{M} α (ϕ_{*}^{- 1} v) μ$ . Change variables by $ϕ$ , using $ϕ^{*} μ = μ$ (volume-preserving) and $ϕ^{*} (α (ϕ_{*}^{- 1} v)) = (ϕ^{*} α) (v)$ : the integral becomes $\int_{M} (ϕ^{*} α) (v) μ = ⟨[ϕ^{*} α], v ⟩$ . Hence $Ad_{ϕ}^{*} [α] = [ϕ^{*} α]$ . Differentiating along a path $ϕ_{t}$ with $\dot{ϕ}_{0} = v$ gives $ad_{v}^{*} [α] = [L_{v} α]$ . Applying $d$ (which commutes with $L_{v}$ and with $ϕ^{*}$ ) to the representative gives the vorticity form $ad_{v}^{*} ϖ = L_{v} ϖ$ . $□$

Proposition 2 (helicity is constant on coadjoint orbits, with vanishing first variation). With $H ([u^{♭}]) = \int_{M} u^{♭} \land d u^{♭}$ , the directional derivative $\langle \delta\mathcal{H}, \operatorname{ad}^_v\varpi\rangle = 0 $f or a l l$ v \in \mathfrak{g} $, so$ \delta\mathcal{H}/\delta\varpi $annihi l a t ese v er y co a d j o in t d i r ec t i o nan d$ \mathcal{H}$ is a Casimir.*

Proof. The first variation of $H$ in the direction $ad_{v}^{*} ϖ = L_{v} ϖ$ (Proposition 1) corresponds to varying $u^{♭}$ by $L_{v} u^{♭}$ . Compute $$ \delta\mathcal{H} = \int_M \big(\mathcal{L}_v u^\flat \wedge du^\flat + u^\flat \wedge d\mathcal{L}_v u^\flat\big). $$ Because $d$ and $L_{v}$ commute, $d L_{v} u^{♭} = L_{v} d u^{♭}$ , and the integrand is the Leibniz expansion $L_{v} (u^{♭} \land d u^{♭})$ . The 3-form $u^{♭} \land d u^{♭}$ is top-degree, so $d (u^{♭} \land d u^{♭}) = 0$ , and Cartan's magic formula gives $L_{v} (u^{♭} \land d u^{♭}) = (d ι_{v} + ι_{v} d) (u^{♭} \land d u^{♭}) = d ι_{v} (u^{♭} \land d u^{♭})$ . Thus $δ H = \int_{M} d ι_{v} (u^{♭} \land d u^{♭}) = 0$ by Stokes on the closed $M$ . As this holds for every $v \in g$ , the variational gradient $δ H / δ ϖ$ pairs to zero with the entire image of $ad^{*}$ , i.e. with the tangent space to the coadjoint orbit, which is the Casimir condition. $□$

Proposition 3 (conservation under Euler flow). Let $H (ϖ) = \frac{1}{2} \int_{M} ∣ u ∣^{2} μ$ be the kinetic energy and let $ϖ (t)$ solve the Lie-Poisson (Euler) equation $\dot\varpi = -\operatorname{ad}^_{\delta H/\delta\varpi}\varpi $. T h e n$ \frac{d}{dt}\mathcal{H}(\varpi(t)) = 0$.*

Proof. The Lie-Poisson evolution of any functional $F$ is $\dot{F} = {F, H}$ . For the Casimir $H$ , the bracket ${H, H} = ⟨ ϖ, [δ H / δ ϖ, δ H / δ ϖ]⟩ = - ⟨ ad_{δ H / δ ϖ}^{*} ϖ, δ H / δ ϖ ⟩$ . By Proposition 2, $ad_{δ H / δ ϖ}^{*} ϖ = 0$ , so ${H, H} = 0$ identically, for every $H$ — in particular the kinetic energy. Hence $\dot{H} = 0$ along the Euler flow. The conclusion is insensitive to the inertia operator defining $H$ : any right-invariant metric, any Hamiltonian whatsoever on $g^{*}$ , conserves helicity. $□$

Connections Master

05.03.01 supplies the engine: the Lie-Poisson bracket on $g^{*}$ , the identification of symplectic leaves with coadjoint orbits, and the definition of a Casimir as a functional constant on those leaves. Helicity is the instance of that abstract structure for $g = X_{vol} (M)$ ; the rigid-body Casimir $∣ L ∣^{2}$ on $so (3)^{*}$ is the finite-dimensional model this unit transports to the fluid group.
05.09.05 provides the dynamical setting: the ideal-fluid Euler equation as the Euler-Arnold geodesic equation on $SDiff (M)$ with the $L^{2}$ metric, the vorticity transport (Helmholtz) equation as coadjoint flow, and the right-invariant bracket convention used throughout. The conservation proof of this unit is the Casimir layer sitting on top of that geodesic dynamics.
09.07.01 gives the classical-fluids substrate: Kelvin's circulation theorem and the Helmholtz vortex laws, which guarantee that circulations and vortex-line linkages are frozen — the physical facts that the Casimir formalism repackages as orbit invariance. The linked-ring helicity formula of Exercise 8 rests on Kelvin's conservation of each tube's circulation.
05.14.02 is the immediate successor: Arnold's theorem that helicity equals the average asymptotic linking number of the vorticity flow, the asymptotic Hopf invariant. The Casimir conservation derived here and the ergodic linking interpretation derived there are two readings of the same number.
05.14.04 carries the invariant into ideal magnetohydrodynamics, where the magnetic helicity $\int A \land d A$ is the Casimir of the semidirect-product Lie-Poisson structure and Alfvén's frozen-flux theorem replaces Helmholtz transport. Magnetic helicity is the velocity-to-magnetic-field dual of fluid helicity.

Historical & philosophical context Master

The conserved integral $\int_{M} u \cdot ω d^{3} x$ was first identified as a constant of the Euler equation by Jean-Jacques Moreau in a 1961 Comptes Rendus note ^{[Moreau 1961]}, in the context of barotropic perfect fluids; the result went largely unnoticed. It was rediscovered and given its topological meaning by H. Keith Moffatt in 1969 ^{[Moffatt 1969]}, who named the quantity helicity by analogy with the particle-physics term and proved that for vorticity concentrated on thin tubes it equals the total linkage of the tubes weighted by their circulations — the first bridge between a fluid invariant and the Gauss linking integral. Vladimir Arnold then placed helicity inside the geometric theory of the Euler equation as a geodesic flow on the volume-preserving diffeomorphism group, recognising it as a Casimir of the Lie-Poisson structure and, in his 1974 Erevan lectures ^{[Arnold 1974]}, as the asymptotic Hopf invariant — the ergodic average of the asymptotic linking number of pairs of vorticity orbits. The synthesis of these strands is the opening movement of Arnold and Khesin's Topological Methods in Hydrodynamics, which makes the coadjoint-orbit picture the organising frame for the entire subject.

Moffatt's identification of a hydrodynamic invariant with a topological linking number was a decisive event in twentieth-century fluid mechanics: it revived, in rigorous form, the nineteenth-century vision of Helmholtz and Kelvin that vortex lines carry a permanent identity, and it connected directly to the knot-theoretic programme Kelvin had hoped would explain atomic structure. The Casimir reading sharpens why the conservation is robust. Conservation of energy follows from time-translation invariance of the Lagrangian; conservation of helicity follows from a different and larger symmetry, the particle-relabelling symmetry that an ideal fluid possesses because its dynamics depends only on the volume form, not on any labelling of fluid parcels. By Noether's theorem in its Lie-Poisson form, that symmetry degenerates the Poisson structure, and the helicity is the invariant that the degeneracy protects — conserved for every Hamiltonian, not merely the kinetic-energy Hamiltonian nature selects.

Bibliography Master

@article{Moreau1961,
  author  = {Moreau, Jean-Jacques},
  title   = {Constantes d'un \^ilot tourbillonnaire en fluide parfait barotrope},
  journal = {Comptes Rendus de l'Acad\'emie des Sciences},
  volume  = {252},
  pages   = {2810--2812},
  year    = {1961}
}

@article{Moffatt1969,
  author  = {Moffatt, H. K.},
  title   = {The degree of knottedness of tangled vortex lines},
  journal = {Journal of Fluid Mechanics},
  volume  = {35},
  number  = {1},
  pages   = {117--129},
  year    = {1969},
  doi     = {10.1017/S0022112069000991}
}

@article{Arnold1974,
  author  = {Arnold, Vladimir I.},
  title   = {The asymptotic {H}opf invariant and its applications},
  journal = {Selecta Mathematica Sovietica},
  volume  = {5},
  number  = {4},
  pages   = {327--345},
  year    = {1986},
  note    = {Originally Proc. Summer School in Differential Equations, Erevan, 1974}
}

@book{ArnoldKhesin2021,
  author    = {Arnold, Vladimir I. and Khesin, Boris A.},
  title     = {Topological Methods in Hydrodynamics},
  series    = {Applied Mathematical Sciences},
  volume    = {125},
  edition   = {2},
  publisher = {Springer},
  year      = {2021}
}

@book{MarsdenRatiu1999,
  author    = {Marsden, Jerrold E. and Ratiu, Tudor S.},
  title     = {Introduction to Mechanics and Symmetry},
  series    = {Texts in Applied Mathematics},
  volume    = {17},
  edition   = {2},
  publisher = {Springer},
  year      = {1999}
}

@article{Moffatt1992,
  author  = {Moffatt, H. K. and Tsinober, A.},
  title   = {Helicity in laminar and turbulent flow},
  journal = {Annual Review of Fluid Mechanics},
  volume  = {24},
  pages   = {281--312},
  year    = {1992},
  doi     = {10.1146/annurev.fl.24.010192.001433}
}

Prerequisites

05.03.01
05.09.05
03.04.04

Tier anchors

beginner: Moffatt-Tsinober 1992 *Helicity in Laminar and Turbulent Flow* (Annu. Rev. Fluid Mech. 24) §1 informal; Arnold-Khesin *Topological Methods in Hydrodynamics* (Springer Applied Math. Sci. 125, 2nd ed. 2021) Ch. I §1 informal
intermediate: Arnold-Khesin *Topological Methods in Hydrodynamics* Ch. I §1-§4, Ch. III §1; Marsden-Ratiu *Introduction to Mechanics and Symmetry* (Springer TAM 17, 2nd ed. 1999) §13.7; Moffatt 1969 *The degree of knottedness of tangled vortex lines* (J. Fluid Mech. 35)
master: Arnold 1974 *The asymptotic Hopf invariant and its applications* (Proc. Summer School Diff. Eqns., Erevan; repr. Sel. Math. Sov. 5, 1986); Moreau 1961 *Constantes d'un îlot tourbillonnaire en fluide parfait barotrope* (C. R. Acad. Sci. Paris 252); Moffatt 1969 *The degree of knottedness of tangled vortex lines* (J. Fluid Mech. 35, 117-129); Arnold-Khesin *Topological Methods in Hydrodynamics* (Springer Applied Math. Sci. 125, 2nd ed. 2021) Ch. I §1-§4 and Ch. III §1-§2; Marsden-Ratiu *Introduction to Mechanics and Symmetry* §13.7; Khesin-Wendt *The Geometry of Infinite-Dimensional Groups* (Springer 2009) Ch. IV

References

Moffatt 1969 — The degree of knottedness of tangled vortex lines · J. Fluid Mech. 35, 117-129, originator of the helicity invariant and its interpretation as total linkage of vortex lines · source being verified
Moreau 1961 — Constantes d'un îlot tourbillonnaire en fluide parfait barotrope · C. R. Acad. Sci. Paris 252, 2810-2812, first identification of the helicity integral as a conserved quantity of the Euler equation · source being verified
Arnold 1974 — The asymptotic Hopf invariant and its applications · Sel. Math. Sov. 5 (1986), 327-345; helicity as the asymptotic linking number and as a Casimir of the ideal-fluid Lie-Poisson structure · source being verified
Arnold, Khesin — Topological Methods in Hydrodynamics · Springer Applied Math. Sci. 125, 2nd ed. 2021, Ch. I §1-§4 (Euler equation as geodesic, Lie-Poisson, Casimirs) and Ch. III §1-§2 (helicity, asymptotic linking)
Marsden, Ratiu — Introduction to Mechanics and Symmetry · Springer TAM 17, 2nd ed. 1999, §13.7 (the ideal fluid as a Lie-Poisson system; circulation and enstrophy Casimirs) · source being verified

Estimated time

beginner: 18m
intermediate: 45m
master: 85m