Why Shared-Phase Temperature and Vorticity Waves Do Not Self-Advect

Good to continue. In the previous lesson, we computed the velocity induced by a vorticity wave and found the key geometry:

$$\vartheta =\Theta \sin (\lambda \zeta \cdot x),\varpi =\Omega \cos (\lambda \zeta \cdot x),$$ $$v=\frac{\Omega }{\lambda |\zeta {|}^2}J\zeta \sin (\lambda \zeta \cdot x).$$

The temperature gradient points in the $\zeta$ direction, while the induced velocity points in the perpendicular $J\zeta$ direction. This lesson verifies the payoff: the new wave does not nonlinearly transport either its own temperature or its own vorticity.

This exact cancellation is what makes the oscillatory layer tractable. It does not make the layer dynamically irrelevant: it still interacts with the older affine temperature gradient, and the Boussinesq buoyancy term still converts temperature variation into vorticity. The point is narrower and more valuable: the high-frequency wave does not generate an uncontrolled self-interaction.

---

What “shared phase” means

Set

$$s(x,t)=\lambda \zeta (t)\cdot x,$$

where $\lambda$ is a fixed large frequency and $\zeta (t)\ne 0$ is the evolving phase-gradient vector. The temperature and vorticity waves are

$$\vartheta (x,t)=\Theta (t)\sin s,\varpi (x,t)=\Omega (t)\cos s.$$

At first sight, sine and cosine may seem to have different phases. They are indeed shifted by one quarter of a period as scalar functions. But “shared phase” here means something more structural: both depend on the same carrier variable $s=\lambda \zeta \cdot x$, hence have the same wavefronts and the same wavevector direction $\zeta$.

The wavefronts are the level sets

$$\lambda \zeta \cdot x=\text{constant}.$$

In two dimensions these are parallel lines. Their normal direction is $\zeta$; their tangent direction is $J\zeta$.

The Biot–Savart calculation from the previous lesson gave

$$v(x,t)=\frac{\Omega (t)}{\lambda |\zeta (t){|}^2}J\zeta (t)\sin s.$$

Thus the wave-induced velocity is tangent to the wavefronts:

$$v\parallel J\zeta .$$

Meanwhile, differentiating either wave produces a vector parallel to the normal direction $\zeta$.

The paper’s initial wave calculation states this compactly.

---

Direct verification of the cancellation

Let us calculate the two relevant gradients. Since

$$\nabla s=\lambda \zeta ,$$

we have

$$\nabla \vartheta =\Theta \cos s\nabla s=\lambda \Theta \zeta \cos s.$$

Similarly,

$$\nabla \varpi =-\Omega \sin s\nabla s=-\lambda \Omega \zeta \sin s.$$

Both gradients are parallel to $\zeta$:

$$\nabla \vartheta \parallel \zeta ,\nabla \varpi \parallel \zeta .$$

The velocity is parallel to $J\zeta$, and $J$ is a right-angle rotation. Therefore

$$(J\zeta )\cdot \zeta =0.$$

Now compute the self-advection of temperature:

$$\begin{array}{rl}v\cdot \nabla \vartheta & =\left(\frac{\Omega }{\lambda |\zeta {|}^2}J\zeta \sin s\right)\cdot \left(\lambda \Theta \zeta \cos s\right)\\ & =\frac{\Omega \Theta }{|\zeta {|}^2}(J\zeta \cdot \zeta )\sin s\cos s\\ & =0.\end{array}$$

Likewise, the wave does not transport its own vorticity:

$$\begin{array}{rl}v\cdot \nabla \varpi & =\left(\frac{\Omega }{\lambda |\zeta {|}^2}J\zeta \sin s\right)\cdot \left(-\lambda \Omega \zeta \sin s\right)\\ & =-\frac{{\Omega }^2}{|\zeta {|}^2}(J\zeta \cdot \zeta ){\sin }^{2}s\\ & =0.\end{array}$$

The cancellation is pointwise. It does not rely on spatial averaging, periodic integration, a small-amplitude approximation, or a cancellation between different waves.

It also does not depend on the value of $\lambda$. Frequency changes the sizes of the gradients and velocity, as seen last lesson, but it cannot change a right angle into a non-right angle.

---

The geometric interpretation: velocity slides along contours

For any scalar field $f$, the quantity

$$v\cdot \nabla f$$

is the directional derivative of $f$ in the velocity direction. If $v$ is tangent to a level set of $f$, moving in that direction does not change $f$, so this directional derivative vanishes.

Here the temperature and vorticity both have level sets determined by the common quantity $s$. Their gradients point normal to these level sets. The wave-induced velocity, meanwhile, runs along them.

So the cancellation can be stated geometrically:

The vorticity wave creates a velocity that moves fluid particles along the wavefronts of both the temperature and vorticity waves, rather than across them.

This is a stronger mental model than “the algebra happens to vanish.” It tells you immediately what would break the cancellation. If the induced velocity had a component parallel to $\zeta$, it would carry particles across the wavefronts and change the wave value. Likewise, if the two fields used genuinely different wavevector directions, the velocity associated with one need not be tangent to the contours of the other.

---

The sine and cosine are convenient, not essential

The cancellation is not a trigonometric accident. The paper later replaces sine and cosine with a more general periodic profile. Let $F$ be a smooth periodic function and consider schematically

$$\vartheta =\Theta F(s),\varpi =\Omega F^{\prime }(s),v=\frac{\Omega }{\lambda |\zeta {|}^2}J\zeta F(s).$$

The particular streamfunction is chosen so that the vorticity is the curl of this velocity; that relationship explains the appearance of $F^{\prime }$ in $\varpi$.

Differentiating the scalar fields still yields

$$\nabla \vartheta =\lambda \Theta F^{\prime }(s)\zeta ,$$ $$\nabla \varpi =\lambda \Omega F^{\prime \prime }(s)\zeta .$$

Hence

$$v\cdot \nabla \vartheta =\frac{\Omega \Theta }{|\zeta {|}^2}F(s)F^{\prime }(s)(J\zeta \cdot \zeta )=0,$$

and

$$v\cdot \nabla \varpi =\frac{{\Omega }^2}{|\zeta {|}^2}F(s)F^{\prime \prime }(s)(J\zeta \cdot \zeta )=0.$$

The profiles can change; the mechanism persists because the gradients remain normal to the same phase surfaces, while velocity remains tangent to them.

---

What is cancelled, and what remains

It would be a mistake to conclude that the added wave experiences no dynamics. It eliminates its self-advection, but several essential terms survive when it is placed on the older affine background

$$u_{\mathrm{old}}=D(t)x,{\theta }_{\mathrm{old}}=G(t)\cdot x.$$

The total fields are

$$u=u_{\mathrm{old}}+v,\theta ={\theta }_{\mathrm{old}}+\vartheta .$$

Three distinctions are important.

1. The affine flow transports the phase

The term

$$(Dx)\cdot \nabla \vartheta$$

does not generally vanish. An affine flow can stretch and rotate the wavevector, changing the orientation and physical wavelength of the oscillation. This is an intended part of the construction, not an error. In the next lesson, we will derive the ODE for $\zeta (t)$ that makes the phase remain transported by this background flow.

2. The new velocity acts on the old temperature gradient

Because the old temperature is affine,

$$\nabla {\theta }_{\mathrm{old}}=G.$$

Therefore,

$$v\cdot \nabla {\theta }_{\mathrm{old}}=v\cdot G,$$

which is generally nonzero:

$$v\cdot G=\frac{\Omega }{\lambda |\zeta {|}^2}(J\zeta \cdot G)\sin s.$$

This is one of the key couplings. The new vorticity creates a transverse velocity, and that velocity acts on the pre-existing temperature gradient.

3. Temperature still sources vorticity

The inviscid Boussinesq vorticity equation contains the forcing term

$${\partial }_1\theta .$$

For the added temperature wave, this contributes

$${\partial }_1\vartheta =\lambda {\zeta }_1\Theta \cos s.$$

Thus temperature variation can still produce vorticity. The self-advection cancellation removes an unwanted nonlinear term, but the temperature-to-vorticity coupling remains and is precisely what later permits amplification.

Finally, because $u_{\mathrm{old}}=Dx$ is affine, its vorticity is spatially constant. Consequently,

$$v\cdot \nabla {\omega }_{\mathrm{old}}=0.$$

So, locally, the only significant interactions left are the deliberately designed ones: affine transport of phase, the action of $v$ on $G$, and buoyancy forcing of the vorticity wave.

---

Why this matters for the layered blowup construction

Nonlinear PDE constructions often become difficult because a new high-frequency component interacts with itself and creates additional frequencies or residuals that cannot be controlled. This wave avoids that immediate difficulty.

Its high-frequency content does not feed into self-transport because its induced velocity has exactly the wrong direction to cross its own phase contours. In particular:

• increasing $\lambda$ makes $\nabla \vartheta$ large;
• Biot–Savart keeps the velocity one derivative smaller than vorticity;
• orthogonality removes $v\cdot \nabla \vartheta$ and $v\cdot \nabla \varpi$ exactly;
• the remaining couplings reduce to scalar amplitude effects once the phase is transported correctly.

This is why the construction can treat a wave as a controllable dynamical layer rather than an arbitrary perturbation of a nonlinear fluid equation.

The cancellation is exact only in the idealized uncut affine region. When the authors later localize waves in space, cutoff gradients and Biot–Savart corrections introduce residual terms. Those are real technical issues, but they are not failures of the core phase geometry.

---

Takeaways

For a wave with shared carrier phase

$$s=\lambda \zeta \cdot x,$$

the induced velocity is directed along $J\zeta$, while both temperature and vorticity gradients are directed along $\zeta$:

$$v\parallel J\zeta ,\nabla \vartheta \parallel \zeta ,\nabla \varpi \parallel \zeta .$$

Since

$$J\zeta \cdot \zeta =0,$$

the self-advection terms vanish pointwise:

$$v\cdot \nabla \vartheta =0,v\cdot \nabla \varpi =0.$$

This is a geometric contour-following cancellation, not a special identity of sine and cosine. It removes the wave’s uncontrolled nonlinear self-interaction while retaining the intended interactions with the affine background and the Boussinesq buoyancy source.

Next, we will derive how the affine background transports the wavevector $\zeta (t)$. That phase-transport law is what lets the shared-phase cancellation remain valid as the layer evolves.