# DCRP101 / X72-R84 — Finite-Lag Joint Path Profile, Second/Third Moment Independence, and the Kelvin–Transport–Riesz Copula Lock **Series:** Independent Navier–Stokes Research Series / X72 Bridge **Date:** 2026-08-20 **Status:** proof-development checkpoint / joint increment-profile round **Immediate predecessor:** `NS_DCRP100_X72R83_FiniteLagDuhamel_RotationalSGSKernel_2026-08-20.md` ## Primary internal dependencies - DCRP24 — cylindrical increment Young profile / fiber escape / covariance realization. - DCRP65 — all exact single-factor Round38 null channels closed; remaining obstruction is genuinely correlational. - DCRP66 — exact cofactor/vorticity two-stress source identity and 4:1 silent-correlation manifold. - DCRP95–96 — sign-coherent Kelvin phase slip and second-moment nematic covariance lock. - DCRP99 — fixed X/Kelvin bounded-lag word. - DCRP100 — finite-lag X-Duhamel forcing split and transport–Riesz source selection. - X72 Round38 — exact transport–Riesz triple-increment pairing. ## Fresh primary-source calibration - R. Yu, *Filtered Vortex Stretching and Subgrid Defects for the Three-Dimensional Navier–Stokes Equations*, arXiv:2606.27560 (2026). - Critical derivative-compatible velocity-increment defects admit cylindrical generalized Young-measure profiles. - Cylindrical control alone does not automatically imply full norm/covariance tightness. - G. L. Eyink, *The Cascade of Circulations in Fluid Turbulence*, arXiv:physics/0606159. - Coarse circulation violation is generated by the turbulent subgrid/vortex force. - G. L. Eyink and H. Aluie, *Localness of energy cascade in hydrodynamic turbulence, I. Smooth coarse-graining*, arXiv:0909.2386. - Smooth coarse-graining naturally organizes quadratic subfilter stresses through scale-local velocity increments. No full Navier–Stokes regularity theorem is claimed. --- # 0. Executive result DCRP100 reduced the late compact finite-lag survivor to a finite source alphabet. DCRP101 attacks the highest-leverage selected branch: \[ \boxed{ \mathsf C_{{\rm TR}\Gamma}^{\ell_*}, } \] meaning: 1. one fixed oriented SGS Kelvin reset recurs: \[ \sigma_\Gamma \delta_\Gamma^{\rm SGS}(n) \ge c_\Gamma>0; \] 2. after one fixed bounded lag \(\ell_*\), the selected X72 event is supported by the transport–Riesz source; 3. after finite sign/subwindow pigeonholing, the transport–Riesz source also has one fixed sign and one uniform normalized gap: \[ \sigma_{\rm TR} \mathcal Q_{\rm TR}^{(\ell_*)}(n) \ge c_{\rm TR}>0 \] on a positive-density generation set. The first proposed intuition was: > both channels are increment-generated, so perhaps the Kelvin second-moment Young profile already forces the transport–Riesz sign. DCRP101 proves that this is false. The Kelvin source is a **quadratic / second-moment** observable of velocity increments. The transport–Riesz source is a **mixed third-order correlation** observable. Even worse for direct coercivity: \[ \boxed{ \text{all one-factor marginals and all pairwise marginals can be identical, while the mixed triple moment is positive, zero, or negative}. } \tag{0.1} \] Thus no theorem of the form \[ \boxed{ Q_\Gamma \Longrightarrow \operatorname{sign}\mathcal Q_{\rm TR} } \] can follow from second moments, one-factor marginals, or pairwise correlations alone. The correct compact survivor is a **joint path-space copula lock** carrying simultaneously: \[ \boxed{ \text{Kelvin second-moment lock} } \] and: \[ \boxed{ \text{transport–Riesz mixed third-moment lock}. } \] This is the new normal form. --- # 1. Scope repair — Round38 identity for a general tensor test D100's finite-lag Duhamel term is not, in general, the Round38 defect-energy pairing. Let: \[ \Phi_X(s) = \mathcal U_E(t_1,s)^*\Psi_X \] be the pulled-back X test. The finite-lag transport–Riesz Duhamel channel is: \[ \boxed{ I_{\rm TR} = \int_{t_0}^{t_1} \left\langle \Phi_X(s), [u\cdot\nabla,\mathcal T_0]q \right\rangle ds. } \tag{1.1} \] In general: \[ \Phi_X\neq E_p. \] Therefore D66's pressure-self-cancellation / cofactor-only reduction cannot be inserted automatically. However the **triple-increment representation itself is more general**. Let: \[ \mathcal K_uq = [u\cdot\nabla,\mathcal T_0]q. \] The trace-free Riesz kernel is: \[ K_0(z), \] with: \[ K_0(-z)=K_0(z), \] so: \[ \nabla K_0(-z) = -\nabla K_0(z). \] Define: \[ G_u(x,y) = [ u(x)-u(y) ] \cdot \nabla K_0(x-y). \] Then: \[ G_u(y,x)=G_u(x,y), \] and the principal-value row/column means vanish. For any sufficiently regular trace-free tensor test \(\Phi\), ## Theorem D101.1 — General Test Triple-Increment Identity \[ \boxed{ \begin{aligned} \left\langle \Phi, [u\cdot\nabla,\mathcal T_0]q \right\rangle = -\frac12 \operatorname{p.v.} \iint & [ \delta_{xy}u \cdot \nabla K_0(x-y) ] \\ &: \delta_{xy}\Phi \, \delta_{xy}q \,dxdy. \end{aligned} } \tag{1.2} \] ### Proof Start with: \[ \langle\Phi,\mathcal K_uq\rangle = \iint \Phi(x):G_u(x,y)q(y)\,dxdy. \] Because: \[ \int G_u(x,y)dy=0, \] \[ \int G_u(x,y)dx=0, \] and: \[ G_u(x,y)=G_u(y,x), \] one has: \[ \begin{aligned} & -\frac12 \iint G_u(x,y): [ \Phi(x)-\Phi(y) ] [ q(x)-q(y) ] dxdy \\ &= \iint \Phi(x):G_u(x,y)q(y)\,dxdy. \end{aligned} \] Substitute the kernel form of \(G_u\). \[ \square \] Thus D100's pulled-back adjoint source retains an exact triple-increment representation. --- # 2. Finite-lag adjoint triple product Apply D101.1 with: \[ \Phi=\Phi_X(s). \] Then: ## Theorem D101.2 — Finite-Lag Adjoint Transport–Riesz Identity \[ \boxed{ \begin{aligned} I_{\rm TR} = -\frac12 \int_{t_0}^{t_1} \operatorname{p.v.} \iint & [ \delta u \cdot \nabla K_0 ] \\ &: \delta\Phi_X \, \delta q \,dxdy\,ds. \end{aligned} } \tag{2.1} \] Therefore the generic finite-lag X source depends on the mixed tuple: \[ \boxed{ ( \delta u, \delta\Phi_X, \delta q ). } \tag{2.2} \] It is **not** a third moment of velocity increments alone. --- # 3. Kelvin is a second-moment observable On the smoothed loop-current branch of D96, the SGS Kelvin reset is: \[ \mathfrak F_\Gamma(R_\ell) = \int R_\ell:A_\Gamma. \] After the D24 full-representation / fiber audit, the Reynolds stress is represented by the centered velocity-increment covariance: \[ Q_\Gamma = \int \zeta\otimes\zeta\,d\nu + Q^c. \] Because: \[ \operatorname{tr}A_\Gamma=0, \] only the deviatoric covariance contributes: \[ \boxed{ \mathfrak F_\Gamma = \int A_\Gamma:Q_\Gamma^0. } \tag{3.1} \] Thus the Kelvin channel is a **second-order** moment functional. On the selected oriented branch: \[ \boxed{ \sigma_\Gamma \mathfrak F_\Gamma \ge c_\Gamma>0. } \tag{3.2} \] --- # 4. Generic transport–Riesz is a mixed third moment Introduce the pair variable: \[ z=x-y. \] At one fixed normalized pair cylinder define abstract variables: \[ a = \delta u\in\mathbb R^3, \] \[ B = \delta\Phi_X \in \mathrm{Sym}_0(3), \] \[ c = \delta q\in\mathbb R. \] Then the local transport–Riesz integrand is: \[ \boxed{ \mathcal T(a,B,c;z) = [ a\cdot\nabla K_0(z) ] : B \, c. } \tag{4.1} \] This is trilinear in: \[ (a,B,c). \] Define the mixed third tensor: \[ \boxed{ M^{(3)}_{i,ab} = \int a_i B_{ab} c \,d\mu(a,B,c,z,\ldots). } \tag{4.2} \] Then the TR functional is a fixed singular-kernel projection of: \[ M^{(3)}. \] The Kelvin functional depends instead on: \[ \boxed{ M^{(2)}_{ij} = \int a_i^\Gamma a_j^\Gamma \,d\mu. } \tag{4.3} \] These are different moment orders and, generically, different time slices inside the fixed-lag path cylinder. --- # 5. Bounded lag requires a path-space profile D99 gives a fixed lag: \[ \ell_*. \] If: \[ \ell_*\neq0, \] the Kelvin event and the TR source event need not be the same local state. Therefore the correct compact object is not one instantaneous Young measure. Augment the normalized generation state to the finite path segment: \[ \boxed{ \mathbf Z_n = ( z_n, z_{n+1}, \ldots, z_{n+\ell_*} ). } \tag{5.1} \] After also selecting one fixed source subwindow inside the D100 Duhamel interval, define the finite cylinder of observables: \[ \boxed{ \Xi_n = ( a^\Gamma, a^{\rm TR}, B^{\rm TR}, c^{\rm TR}, z^{\rm pair}, \text{loop/detector labels} ). } \tag{5.2} \] On a tight compact branch, pass to a subsequence: \[ \boxed{ \mu_n^{(\ell_*)} \rightharpoonup \mu_*^{(\ell_*)}. } \tag{5.3} \] This is the **finite-lag joint path profile**. ### Scope warning Yu's 2026 theorem directly supplies cylindrical Young profiles for the critical velocity-increment sector. It does **not** by itself supply full tightness of the larger tuple: \[ ( \delta u, \delta\Phi_X, \delta q ). \] Therefore if the enlarged tuple loses tightness / representation, record: \[ \boxed{ R_{\rm path/fib} \vee R_{\rm crit} \vee R_{\rm state}. } \tag{5.4} \] D101's full joint-profile theorem applies only on the complementary compact representation branch. --- # 6. Main joint path-profile lock On the compact full-representation branch, D95 gives: \[ \boxed{ \sigma_\Gamma \mathcal L_\Gamma ( M_\Gamma^{(2)} ) \ge c_\Gamma. } \tag{6.1} \] D100 source selection, finite subwindow selection, and sign pigeonholing give: \[ \boxed{ \sigma_{\rm TR} \mathcal L_{\rm TR} ( M_{\rm TR}^{(3)} ) \ge c_{\rm TR}. } \tag{6.2} \] Therefore: ## Theorem D101.3 — Kelvin/TR Second–Third Moment Lock Every compact recurrent branch of type: \[ \mathsf C_{{\rm TR}\Gamma}^{\ell_*} \] has a limiting finite-lag path profile satisfying simultaneously: \[ \boxed{ \sigma_\Gamma \mathcal L_\Gamma ( M_\Gamma^{(2)} ) \ge c_\Gamma>0, } \] and: \[ \boxed{ \sigma_{\rm TR} \mathcal L_{\rm TR} ( M_{\rm TR}^{(3)} ) \ge c_{\rm TR}>0. } \] Thus the final compact profile is locked at **two distinct moment orders**. --- # 7. Parity-copula counterexample The second moment does not determine the mixed third moment. In fact, even the **entire one-variable and pairwise marginals** do not determine it. Let: \[ s,t,r \in \{-1,+1\} \] be independent Rademacher signs. Use an abstract scalarized detector geometry in which: \[ a=s, \qquad B=t. \] Consider three couplings for \(c\). ## Positive triple coupling \[ \boxed{ c=st. } \] Then: \[ abc = s\,t\,(st) = 1. \] Therefore: \[ \boxed{ \mathbb E[abc]=1. } \tag{7.1} \] ## Negative triple coupling \[ \boxed{ c=-st. } \] Then: \[ \boxed{ \mathbb E[abc]=-1. } \tag{7.2} \] ## Decorrelated coupling \[ \boxed{ c=r. } \] Then: \[ \boxed{ \mathbb E[abc] = \mathbb E[str] = 0. } \tag{7.3} \] In **all three models**: \[ a,\ B,\ c \] are individually uniform on: \[ \{-1,+1\}. \] Moreover every pair: \[ (a,B), \quad (a,c), \quad (B,c) \] is independent and has exactly the same pairwise law in all three constructions. Finally: \[ \boxed{ \mathbb E[a^2]=1 } \tag{7.4} \] in every case. Thus: ## Theorem D101.4 — Marginal/Pairwise Insufficiency \[ \boxed{ \text{same second moment} + \text{same one-factor marginals} + \text{same pairwise marginals} \not\Rightarrow \text{same mixed triple correlation}. } \tag{7.5} \] The TR sign is genuine **third-order copula information**. ### Scope This is an abstract moment/coupling counterexample. It is not asserted to be an exact Navier–Stokes pair field. Its purpose is precise: > no theorem based only on the Kelvin covariance and lower-order marginals can determine the transport–Riesz correlation sign. --- # 8. D65 is exactly the same warning in PDE form D65 proves that on the recurrent aligned/no-turnover branch, the three exact factorwise Round38 null channels are gone. The late silent branch cannot be explained by: \[ \delta V=0, \] \[ \delta E_p=0, \] or: \[ \delta q=0. \] Any remaining cancellation must use: - pair/support decorrelation; - tensor/angular orthogonality; - multiscale sign cancellation; - principal-value cancellation. D101's parity-copula theorem is the finite-dimensional moment analogue of this PDE conclusion. The remaining problem is relational. --- # 9. Conditional symmetry kernel Let: \[ \mathscr I_{\rm TR}(a,B,c,z) = [ a\cdot\nabla K_0(z) ] : B \,c. \] Suppose the joint path profile is invariant under a measurable involution: \[ T \] such that: \[ \boxed{ \mathscr I_{\rm TR}(T\Xi) = - \mathscr I_{\rm TR}(\Xi). } \tag{9.1} \] Then: \[ \boxed{ \mathcal L_{\rm TR}(\mu)=0. } \tag{9.2} \] Therefore the active fixed-sign TR branch must break every exact **detector-reversing copula symmetry**. This is stronger than breaking a one-variable sign symmetry. The physical pair exchange: \[ x\leftrightarrow y \] does not automatically produce such a cancellation because all relevant pair factors transform together. --- # 10. Fixed-sign TR recurrence gives linear normalized variation After finite sign pigeonholing on the selected \(\mathsf C_{{\rm TR}\Gamma}^{\ell_*}\) branch, let: \[ \mathcal A_{\rm TR} \] be the positive-density generations carrying: \[ \sigma_{\rm TR} \mathcal Q_{\rm TR}(n) \ge c_{\rm TR}. \] Then: \[ \boxed{ \sum_{\substack{n0. \] Use (11.2): ## Theorem D101.5 — Quantitative Two-Stress Correlation Gap \[ \boxed{ \sigma_{\rm TR} \left( 4\mathfrak Q_{CC} - \mathfrak Q_{C\omega} \right) \ge \sqrt{\frac83} \,c_{\rm TR}. } \tag{12.2} \] Thus the active zero-lag Kelvin/TR branch stays a uniform signed distance from the D66 correlation-silent manifold. This is the precise replacement for the vague phrase: > nonzero third moment. --- # 13. Generic finite lag does NOT inherit the 4:1 reduction automatically For generic: \[ \ell_*\neq0, \] the pulled-back test is: \[ \Phi_X(s) = \mathcal U_E(t_1,s)^*\Psi_X. \] There is no theorem: \[ \boxed{ \Phi_X=E_p. } \] Therefore one cannot write: \[ \delta\Phi_X \rightsquigarrow \delta C_S^0 \] without an additional adjoint-lock theorem. Hence: ## STOP-D101-A \[ \boxed{ \text{D66 4:1 two-stress reduction} } \] is a zero-lag / defect-energy subbranch result, not a generic finite-lag identity. The generic finite-lag survivor retains: \[ \boxed{ ( \delta u, \delta\Phi_X, \delta q ) } \] as its mixed third-order path variable. --- # 14. Joint path-profile normal form Define: \[ \boxed{ \mathsf C_{2+3}^{\ell_*} } \] to be a finite-lag path profile satisfying: ### Kelvin second-order lock \[ \boxed{ \sigma_\Gamma \mathcal L_\Gamma ( M_\Gamma^{(2)} ) \ge c_\Gamma. } \tag{14.1} \] ### transport–Riesz third-order lock \[ \boxed{ \sigma_{\rm TR} \mathcal L_{\rm TR} ( M_{\rm TR}^{(3)} ) \ge c_{\rm TR}. } \tag{14.2} \] ### fixed path phase \[ \boxed{ \ell_* = \text{constant}. } \tag{14.3} \] ### compact finite labels - fixed loop orientation; - fixed X detector; - fixed Duhamel source channel; - fixed selected source subwindow. ### no lower-order determination The mixed third moment is not determined by: - Kelvin covariance; - one-variable marginals; - pairwise marginals. This is a genuine copula/phase constraint. --- # 15. Main D101 compiler Combine D100 with the joint-profile scope audit. ## Theorem D101.6 — Kelvin/TR Path-Copula Compiler \[ \boxed{ \mathsf C_{{\rm TR}\Gamma}^{\ell_*} \Longrightarrow \mathsf C_{2+3}^{\ell_*} \vee R_{\rm path/fib} \vee R_{\rm state} \vee R_{\rm crit}. } \tag{15.1} \] In the zero-lag defect-energy subbranch: \[ \boxed{ \mathsf C_{2+3}^{0} \Longrightarrow \text{Kelvin nematic lock} + \text{signed D66 two-stress correlation gap} } \tag{15.2} \] unless state/fiber/critical escape occurs. No simpler covariance-only closure is available. --- # 16. Why the rotational SGS kernel survives this round D100's rotational SGS force kernel can support Kelvin circulation while suppressing direct symmetric-gradient response. D101 does not eliminate it. To coexist with the selected TR branch, the path must later develop: \[ \mathcal L_{\rm TR}(M^{(3)})\neq0. \] The rotational source itself does not supply that sign. The fixed-lag path profile must carry the missing third-order correlation through: - defect memory; - adjoint test evolution; - cofactor/pressure-source evolution; - transport–Riesz correlation. Thus the rotational Kelvin kernel is now embedded inside the larger \(2+3\) path-copula normal form. --- # 17. Why this is not “just another Young measure” The new object has three structural differences from D24/D96. ## A. different time positions The Kelvin and TR moments may occur at different generations. ## B. different moment orders Kelvin is quadratic. TR is mixed trilinear. ## C. different variables The generic TR tuple is: \[ (\delta u,\delta\Phi_X,\delta q), \] not three copies of the velocity increment. Therefore the correct object is a **finite-lag lifted path profile**, not merely a higher moment of the D24 velocity Young measure. --- # 18. Compactness consequence Suppose: - all relevant path variables are uniformly tight; - no concentration/fiber loss; - detector/source labels are fixed; - the branch stays in the bounded-reservoir normalized class. Then a subsequence converges to: \[ \mu_*^{(\ell_*)}. \] The two signed functionals remain separated from zero: \[ \boxed{ \sigma_\Gamma \mathcal L_\Gamma(\mu_*^{(\ell_*)}) \ge c_\Gamma, } \] \[ \boxed{ \sigma_{\rm TR} \mathcal L_{\rm TR}(\mu_*^{(\ell_*)}) \ge c_{\rm TR}. } \] Thus the late compact survivor is no longer an unstructured infinite-dimensional defect cloud. It is one fixed-lag joint probability/profile object constrained by two explicit signed moment functionals. --- # 19. Status ledger ## PROVED this round ### D101-P1 — general tensor-test transport–Riesz triple-increment identity. ### D101-P2 — D100's pulled-back adjoint TR source retains an exact triple-increment representation. ### D101-P3 — Kelvin SGS reset is a second-order increment-covariance functional. ### D101-P4 — generic finite-lag TR source is a mixed third-order functional of: \[ (\delta u,\delta\Phi_X,\delta q). \] ### D101-P5 — generic bounded lag requires a lifted path-space profile, not one instantaneous Young measure. ### D101-P6 — same second moment does not determine the mixed third moment. ### D101-P7 — even identical one-factor and pairwise marginals do not determine the triple sign. ### D101-P8 — active TR breaks every exact detector-reversing joint symmetry. ### D101-P9 — fixed-sign recurrent TR gives linear normalized positive variation on its positive-density event set. ### D101-P10 — D66 two-stress 4:1 reduction is restricted to the zero-lag defect-energy subbranch. ### D101-P11 — active zero-lag TR has a quantitative signed distance from the D66 4:1 silent manifold. ### D101-P12 — the compact TR/Kelvin survivor reduces to one finite-lag second/third-moment copula lock. --- # 20. What is NOT proved D101 does not prove: - the \(2+3\) joint path profile is impossible; - the full enlarged tuple automatically has a Young-measure representation from Yu's velocity-increment theorem alone; - the Kelvin second moment determines the TR mixed third moment; - the D66 4:1 relation holds for a generic pulled-back adjoint X test; - a rotational Kelvin SGS source necessarily produces later TR correlation; - the two normalized total-variation laws have one finite physical capacity; - global Navier–Stokes regularity. The remaining issue is now **adjoint-path copula rigidity**. --- # 21. STOP-D101 \[ \boxed{ \begin{minipage}{0.94\linewidth} The highest-leverage D100 transport–Riesz/Kelvin branch does not collapse to one higher moment of the old velocity-increment Young measure. The generic finite-lag Duhamel source is paired against the pulled-back X test \(\Phi_X=\mathcal U_E^*\Psi_X\), not against \(E_p\) itself. Nevertheless the Round38 symmetrization extends to any tensor test: \[ \langle\Phi,[u\cdot\nabla,\mathcal T_0]q\rangle = -\tfrac12\operatorname{p.v.}\iint [\delta u\cdot\nabla K_0]:\delta\Phi\,\delta q. \] Hence Kelvin reset is a second-order velocity-increment covariance observable, whereas the finite-lag TR source is a mixed third-order observable of \((\delta u,\delta\Phi_X,\delta q)\). These moment orders are genuinely independent: an explicit Rademacher coupling keeps the Kelvin second moment, every single-factor marginal, and every pairwise marginal identical while changing the triple moment from positive to zero to negative. The remaining information is therefore third-order copula/phase information, exactly matching D65's conclusion that all single-factor null channels are gone and only relational cancellation remains. On a compact no-fiber branch the fixed-lag survivor extracts a joint path profile carrying two signed locks: a Kelvin nematic second moment and a TR mixed third moment. Only in the zero-lag defect-energy subbranch does the D66 pressure-self-null identity reduce the TR factor to strain cofactor and yield the quantitative two-stress gap \[ \sigma_{\rm TR}(4\mathfrak Q_{CC}-\mathfrak Q_{C\omega}) \ge \sqrt{8/3}\,c_{\rm TR}, \] i.e. a fixed distance from the old 4:1 silent manifold. The generic survivor is therefore a finite-lag Kelvin–TR second/third-moment path-copula lock, not a covariance-only normal form. \end{minipage} } \] --- # 22. Next autonomous step ## DCRP102 / X72-R85 — Backward Adjoint X-Test / Path-Copula Rigidity **Working title** > **Can the Pulled-Back X72 Adjoint Test Maintain a Fixed-Sign Third-Order Correlation with the Kelvin-Active Increment Field over a Fixed Lag without Entering an Adjoint Symmetry / Gradient / State Escape?** Primary tasks: 1. start from: \[ \mathsf C_{2+3}^{\ell_*}; \] 2. derive the exact backward adjoint equation: \[ -\partial_s\Phi_X = \mathcal A_E(s)^*\Phi_X; \] 3. obtain evolution laws for: \[ \delta\Phi_X; \] 4. test detector-reversing symmetries of the adjoint flow; 5. classify when: \[ [\delta u\cdot\nabla K_0]: \delta\Phi_X \] can keep one sign after multiplication by \(\delta q\); 6. use D65–68 correlation/shape geometry in the zero-lag or cofactor-locked subbranch; 7. route loss of adjoint/profile tightness to: \[ R_{\rm state}\vee R_{\rm crit}; \] 8. seek: \[ \mathsf C_{2+3}^{\ell_*} \Longrightarrow \text{adjoint phase-lock normal form} \vee R_{\rm state} \vee R_{\rm crit}. \] Desired endpoint: \[ \boxed{ \text{late compact survivor} \Longrightarrow \text{one backward-adjoint copula kernel} \vee R_{\rm state} \vee R_{\rm crit}. } \] **End checkpoint:** DCRP101 / X72-R84.