Contents · Previous · Next

H1H^1, the operator domain is H2H^2, and the spatial principal part is uniformly elliptic. The conserved derivative energies are

∥∂tj+1v∥22+∥T∂tjv∥22=constant. \|\partial_t^{j+1}v\|_2^2+ \|\mathcal T\partial_t^jv\|_2^2=\text{constant}.

Time derivatives commute with the equation. Elliptic recovery of spatial derivatives costs the undifferentiated L2L^2 displacement only once, giving ∥S(t)∥Hs+1×Hs≤Cs(1+∣t∣)\|S(t)\|_{H^{s+1}\times H^s}\le C_s(1+|t|). The skew GG need not commute with AA.

The common phase is a zero mode, and far-out normalized phase bumps have ∥AvR∥2→0\|Av_R\|_2\to0. Removing any finite number of symmetry modes leaves this zero threshold. Neither persistence proof below assumes a spectral gap, a compact resolvent or a bounded inverse of AA.

A Quantitative Nonlinear Prefix Interval

Fix integer s≥10s\ge10 and define the relative norm

Bs(t)=∥X(t)∥Hulocs+1+∥Xt(t)∥Hulocs. B_s(t)=\|X(t)\|_{H^{s+1}_{\rm uloc}}+ \|X_t(t)\|_{H^s_{\rm uloc}}.

There are c∗,C∗>0c_*,C_*>0 for the fixed action, carrier and prefix such that

Bs(0)≤ϵ,ϵ(1+T)12≤c∗⟹Bs(t)≤C∗ϵ(1+∣t∣)9(∣t∣≤T).(P1) B_s(0)\le\epsilon,\qquad \epsilon(1+T)^{12}\le c_* \quad\Longrightarrow\quad B_s(t)\le C_*\epsilon(1+|t|)^9\quad (|t|\le T). \tag{P1}

The conclusion includes existence, uniqueness and positive relative support and density on all space. Data are smooth compatible bounded data or the smooth-data closure specified by the local theorem. Spare derivatives justify the calculation; their magnitudes do not replace the stated lower-order smallness. Constants may depend on the finite carrier, prefix, units and positive chart margins.

Here is the nonlinear estimate behind the powers. The exact rotated perturbation action is L2+N3\mathcal L_2+N_3, with

L2=12∣vt∣2+vt⋅B0(x)v−12∣Tv∣2,N3=w0n3(x,X,DX). \mathcal L_2=\frac12|v_t|^2+v_t\cdot B_0(x)v -\frac12|\mathcal Tv|^2,\qquad N_3=w_0 n_3(x,X,DX).

The reference is on shell, so removing the first variation subtracts a divergence. The rotating conserved generator is E−ΩQcommonE-\Omega Q_{\rm common}; the perturbed charge need not equal the reference charge. The remainder is cubic in relative jets, and its physical Hessian and time derivative are O(Bs)O(B_s) on the fixed tube.

Use the estimate weight χδ,y=e−2δ⟨x−y⟩\chi_{\delta,y}=e^{-2\delta\langle x-y\rangle}, 0<δ≤10<\delta\le1, and add δ2∥∂tjv∥L2(χ)2\delta^2\|\partial_t^jv\|_{L^2(\chi)}^2 to the energies for 0≤j≤s0\le j\le s. This changes the estimated norm, not the field equation. If F(v)=0F(v)=0 is the full Euler equation, then

F′(v)∂tjv=Rj,R1=0,∑j=1s∥Rj∥L2(χ)≤CsBsJs. F'(v)\partial_t^jv=R_j,\quad R_1=0,\qquad \sum_{j=1}^s\|R_j\|_{L^2(\chi)}\le C_sB_s\mathcal J_s.

The complete Jacobi operator F′(v)F'(v) retains coefficient variations with respect to both vv and DvDv. Remaining chain-rule factors have total order at most j+1j+1; no uncontrolled derivative of order s+2s+2 is left. Descending elliptic estimates with the fixed reference operator give

Js≤Cs(E0+∥v∥L2(χ))+CsBsJs≤Csδ−1Es,δ,y. \mathcal J_s\le C_s(E_0+\|v\|_{L^2(\chi)})+C_sB_s\mathcal J_s \le C_s\delta^{-1}\mathcal E_{s,\delta,y}.

Only the undifferentiated displacement costs δ−1\delta^{-1}. Under Bs≤c0δ2B_s\le c_0\delta^2 the true Jacobi and zeroth nonlinear energies are equivalent to the regularized positive reference energy. Their exact canonical flux identity yields

D+Es,δ,y≤CE[δ+Bsδ−2]Es,δ,y.(P2) D^+\mathcal E_{s,\delta,y} \le C_E[\delta+B_s\delta^{-2}]\mathcal E_{s,\delta,y}. \tag{P2}

The reference gyro cancels from the energy. The estimate has no hidden fixed exponential growth rate. Weighted local Sobolev multiplication handles intermediate derivatives; it does not assume every factor is bounded pointwise.

Integrability of the actual ten-dimensional tail gives

sup⁡y[∫χδ,yw0 dxw0(y)]1/2≤Cδ−8. \sup_y\left[\frac{\int\chi_{\delta,y}w_0\,dx}{w_0(y)}\right]^{1/2} \le C\delta^{-8}.

For centers within distance O(δ−1)O(\delta^{-1}) use the finite total integral; for farther centers split at ∣x−y∣=∣y∣/2|x-y|=|y|/2 and use the exponential factor. This avoids multiplying a worst density ratio by a separate full collar volume. One further displacement factor gives

Bs(t)≤Cϵδ−9exp⁡(CEδ∣t∣+CEδ−2∫0∣t∣Bs(τ) dτ). B_s(t)\le C\epsilon\delta^{-9} \exp\left(C_E\delta|t|+ C_E\delta^{-2}\int_0^{|t|}B_s(\tau)\,d\tau\right).

Set δ=(1+T)−1\delta=(1+T)^{-1}. The bootstrap Bs≤2Aϵδ−9B_s\le 2A\epsilon\delta^{-9} bounds the nonlinear exponent by 2ACEϵδ−122AC_E\epsilon\delta^{-12}. Fixing that quantity sufficiently small also keeps the energy positive and the chart strict, so continuation proves (P1). The linear relative group obeys the corresponding Cs(1+∣t∣)9C_s(1+|t|)^9 bound.

The sufficient horizon and terminal size are therefore

Contents · Previous · Next