A mathematical proof says that two-dimensional incompressible Navier–Stokes equations with dissipation in only one spatial direction do not show anomalous energy loss or energy creation for the weak solutions covered by the theorem. Every solution in that class obeys the model’s anisotropic energy equality, rather than gaining or losing energy outside that balance.
This is a theoretical analysis that quantifies over a mathematical solution class rather than an empirical sample. It asks whether one-directional dissipation is enough to enforce energy equality at the natural energy level, then asks when two weak solutions must be identical.
The theorem gives every covered weak solution a unique representative that is continuous in time in L2, the space of square-integrable functions, and satisfies the full anisotropic energy equality. The conclusion is conditional on the PDE hypotheses specified for that class.
The class under examination
The mathematical population is a weak-solution class in which the velocity u is essentially bounded in time with values in L2 over the two-dimensional plane, while its x1 derivative is square-integrable over time and space. The velocity is also weakly divergence-free, the model’s mathematical condition for incompressibility.
There is no empirical sample behind the conclusion. Instead, the proof ranges over the stated weak-solution class and, for its uniqueness result, over a pair of weak solutions u and U that meet the theorem’s conditions.
Only the x1 direction carries the model’s diffusion, so the proof cannot rely on the same derivative control in x2. It handles the less regular horizontal component u1 as a renormalized solution, allowing an energy balance to be established without assuming that missing vertical derivative.
The vertical component u2 is treated separately. It has a unique representative continuous in time with values in spatial L2 and satisfies its own component energy balance, which contributes to the full identity for the velocity.
The estimates behind the result
To make the one-directional structure work, the analysis uses a directional Riesz estimate, a bound linking spatial quantities. In the notation of the proof, it relates the spatial L2 norm of the Riesz transform of f squared to the product of the spatial L2 norms of f and its x1 derivative.
Another step examines the transport error created by mollifying the equations. For the scalar and vector fields used in the proof, the divergence of the mollified transport commutator converges strongly to zero in spatial L1 as the smoothing scale ε tends to zero. This convergence is used while passing to the energy balance.
The same analysis establishes a unique square-integrable pressure for each weak solution and additional time regularity for the velocity in a negative-one Sobolev space. These are further properties obtained inside the weak-solution framework.
When two solutions must agree
The paper’s second main result is a weak–strong uniqueness statement. It compares two weak solutions, u and U, that start from identical initial data. They are shown to coincide in L2 at every time covered by the theorem if U has the additional vertical derivative regularity specified in the paper: the x2 derivative of its first component, ∂x2 U1, belongs to the stated L2 time-and-space class.
The proof compares the energy of their difference. With the comparison time set to zero and the difference set to zero there, it applies Grönwall’s lemma to show that the difference remains zero at every theorem time.
This is conditional weak–strong uniqueness, not a general uniqueness theorem for every possible weak-solution pair. One solution must have the specified vertical regularity, and the two solutions must have identical initial data.
What the theorem leaves open
The preprint does not settle existence and uniqueness for arbitrary solenoidal L2 initial data; the supplied analysis says those questions are currently not known.
The theorem also does not establish the missing vertical derivative regularity for every weak solution. Its uniqueness application assumes that regularity for one solution, while the authors state that it likely does not hold in general.
Taken together, the results establish an energy identity across the stated weak-solution class and a conditional route to equality of solutions, but they do not establish existence or uniqueness for arbitrary L2 initial data.
The result concerns the equations and solution classes in this mathematical model, not an empirical sample or direct observation of a physical fluid. The supplied analysis provides no empirical or numerical validation.
The document is an arXiv version-one preprint dated 20 August 2026. The authors disclose AI assistance for candidate proof strategies and intermediate arguments, followed by independent reconstruction and verification of every step. No funding statement is reported in the supplied end matter.
Paper data and sources
Original title: Energy rigidity and weak-strong uniqueness for the 2D anisotropic Navier-Stokes equations
Authors: Josef Demmel, Emil Wiedemann
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text