Pith. sign in

REVIEW 3 major objections 6 minor 23 references

Global well-posedness for the 2-D inhomogeneous incompressible Navier-Stokes system with large initial data in critical spaces

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The 2-D inhomogeneous Navier-Stokes equations have global solutions for large initial data in critical Besov spaces that are 'almost the energy space'.

desk verdict First large-data critical-space global well-posedness for 2D inhomogeneous Navier-Stokes; proof is mostly solid but the short-time bootstrap in Proposition 3.2 is sketched, not fully closed. read the letter →

arxiv 1908.02216 v1 pith:WLRRXFWX submitted 2019-08-06 math.AP

Hammadi Abidi , Guilong Gui This is my paper · ORCID
classification math.AP MSC 35Q3076D05
keywords InhomogeneousNavier-Stokesequationsglobalwell-posednesscriticalBesovspaceslargeinitialdataLittlewood-Paleytheorydensity-dependentincompressiblefluidsuniquenessOsgoodlemma
topics Navier-Stokes Regularity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that the 2-D incompressible inhomogeneous Navier-Stokes system — the standard model for two miscible incompressible fluids with different densities — admits a global-in-time solution for initial data that are large, as long as the data lie in critical Besov spaces whose scaling matches the energy space. Previous global well-posedness results in two dimensions required extra regularity on the density or the velocity; here that extra regularity is dropped, except that uniqueness still asks for a bit more density regularity. The central object is the reformulated unknown $a = 1/\rho - 1$ transported by the flow, and the proof splits the density-dependent viscous coefficient into a smooth low-frequency part plus a small high-frequency remainder, then closes a short-time bootstrap on the velocity's $L^1_t(\dot B^2_{2,1})$ norm. A smart reader would care because the result moves well-posedness theory for this system from small-data or extra-regularity regimes to the scaling-optimal critical spaces, the same threshold where the classical energy-space weak solutions live.

What carries the argument

The mechanism that carries the argument is the decomposition of the density-dependent diffusion coefficient by a dyadic partial sum, a frequency-by-frequency splitting inherited from Littlewood-Paley theory. In the linearized momentum equation $\partial_t u - (1+a)(\Delta u - \nabla\Pi) = f$, one rewrites $1+a$ as $1 + \dot S_m a + (a - \dot S_m a)$, so that the high-frequency remainder $a - \dot S_m a$ is small in the critical norm (the tail of $a_0$'s Besov norm vanishes as $m \to \infty$), while the low-frequency part $\dot S_m a$ is smooth enough for basic energy estimates. This split feeds a short-time bootstrap (Proposition 3.2) that proves $\|u\|_{L^1_t(\dot B^2_{2,1})} + \|\nabla u\|_{L^2_t(L^2)} \lesssim \sum_{j \in \mathbb{Z}} (1 - e^{-ct2^{2j}}) \|\dot\Delta_j u_0\|_{L^2} + \sqrt{t}$ on a small interval $[0,T_1]$, using Littlewood-Paley commutator estimates, the basic energy inequality, and a pressure bound from Proposition 3.1; this bootstrap is the step that makes the estimate independent of the size of the data. Uniqueness is then closed by an Osgood-type lemma applied to the difference of two solutions, where a logarithmic factor yields the divergent integral $\int_0^1 \frac{dx}{x(1-\ln x)}$ that forces the difference to vanish on a short interval, then inductively for all times.

What would settle it

A reader could try to reproduce the closing of (3.21) with explicit constants: for a fixed initial pair $(a_0, u_0)$ with large critical norms, determine whether there exist $\eta > 0$, $k \in \mathbb{Z}$, $t > 0$ such that the coefficient of $\|u\|_{L^1_t(\dot B^2_{2,1})}$ on the right-hand side of (3.21) is at most $1/2$ while the remaining terms are finite; if for some admissible data the absorption condition fails for every choice, then (3.7) is false. Since the paper does not prove the required continuity of $\|\nabla u\|_{L^1_t(L^\infty)} \to 0$ as $t \to 0$ for the approximate solutions, that calculation is the natural place to look.

Watch

Extended reading notes

Core claim

The central claim, in the authors' own terms, is Theorem 1.1: for any $\varepsilon \in (0,1)$, if $u_0$ is divergence-free in $\dot B^0_{2,1}(\mathbb{R}^2)$ and $1/\rho_0 - 1$ belongs to $\dot B^\varepsilon_{2/\varepsilon,1}(\mathbb{R}^2)$ with $m \leq \rho_0 \leq M$ for some positive constants $m$ and $M$, then the system has a global solution $(\rho, u, \nabla\Pi)$ with $1/\rho - 1 \in C(\mathbb{R}_+; \dot B^\varepsilon_{2/\varepsilon,1})$, $u \in C(\mathbb{R}_+; \dot B^0_{2,1}) \cap L^1_{\rm loc}(\mathbb{R}_+; \dot B^2_{2,1})$, and $\partial_t u, \nabla\Pi \in L^1_{\rm loc}(\mathbb{R}_+; \dot B^0_{2,1})$. If in addition $1/\rho_0 - 1 \in B^1_{2,1}$, the solution is unique. The force of the theorem is the 'without smallness' clause: the critical norms of the two initial quantities may be arbitrarily large, and in terms of the scaling transformation (1.4) the space for the density deviation has the same scaling as $L^\infty$ while the space for the velocity has the same scaling as $L^2$, which is exactly the energy space of the classical weak-solution theory.

Load-bearing premise

The load-bearing premise is the unproven continuity step that closes the short-time bootstrap in Proposition 3.2 — namely that for the mollified approximate solutions the norms $\|\nabla u\|_{L^1_t(L^\infty)}$ and $\|\nabla u\|^2_{L^2_t(L^2)}$ tend to $0$ as $t \to 0$ uniformly enough to make the right-hand side of (3.21) absorbable into the left-hand side.

Editorial extensions

If this is right

  • The 2-D inhomogeneous Navier-Stokes system is globally solvable for arbitrarily large initial velocity in $\dot B^0_{2,1}$ and density deviation in $\dot B^\epsilon_{2/\epsilon,1}$, with no smallness condition on the data.
  • Uniqueness holds in the slightly more regular class with $1/\rho_0 - 1 \in B^1_{2,1}$, so the large-data solutions are determined by the initial data, not merely existing.
  • The regularity reached is scaling-optimal: under the scaling transformation (1.4) the spaces for the density deviation and the velocity match the energy spaces $L^\infty$ and $L^2$, closing the gap between energy-space weak solutions and critical-space strong solutions in two dimensions.
  • Earlier 2-D well-posedness theorems required extra regularity on the density or the velocity; this result operates at the critical regularity itself, the same threshold where the classical weak-solution theory lives.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The low/high-frequency splitting used here should adapt to the variable-viscosity system $\mu(\rho)$, since the same coefficient appears in the diffusion term; checking that the commutator and pressure estimates survive the extra factor would test the method's reach.
  • A natural open question is whether uniqueness can be relaxed to exactly the critical density space $\dot B^\epsilon_{2/\epsilon,1}$; the Osgood argument here needs $B^1_{2,1}$ regularity for the density difference, so the logarithmic loss appears to be the obstacle rather than the existence proof.
  • One could test the bootstrap quantitatively by computing the constants in (3.21) for simple families of large data; the theorem does not quantify how the lifespan $T_1$ shrinks as the critical norms grow, so a numerical exploration would show whether the absorption step is robust in practice.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper proves, without any smallness assumption, the global existence and uniqueness of the 2D inhomogeneous incompressible Navier-Stokes system (1.2) for initial data in critical Besov spaces: u0 in \dot B^0_{2,1}(R^2), 1/rho0 - 1 in \dot B^epsilon_{2/epsilon,1}(R^2), with rho0 bounded away from 0 and infinity, plus uniqueness when 1/rho0 - 1 in B^1_{2,1}. The proof combines a short-time L1_t(\dot B^2_{2,1}) estimate for the velocity (Proposition 3.2), estimates for the pressure (Propositions 3.1 and 3.3), a mollification and compactness argument (Section 4, Step 1), and an Osgood-lemma based uniqueness argument (Section 4, Step 2). The paper is written in a standard Littlewood-Paley/Besov framework and extends earlier small-data critical-space results by removing the smallness condition on the velocity data.

Significance. If the proof is correct, this is a substantial improvement over prior global well-posedness results for the 2D inhomogeneous Navier-Stokes system, which required additional regularity of the density or velocity beyond the critical scaling. The claimed class is 'almost the energy space' in the scaling sense, so the result would be close to optimal with respect to the data regularity. The manuscript contains detailed dyadic estimates, uses a natural low-frequency truncation S_k a of the density fluctuation, and does not fit constants to the conclusion; the central theorem is not assumed in the argument. The main technical tool, Proposition 3.2, is however not fully proved as written, and the uniformity needed for the mollified sequence is not established. For these reasons the paper cannot be accepted in its present form, although the strategy appears plausible and the gaps seem repairable.

major comments (3)
  1. [Section 3, end of proof of Proposition 3.2, around (3.21)] The closing step of Proposition 3.2 is not justified. In (3.21) the terms to be absorbed into the left-hand side X(t) = ||u||_{L1_t(\dot B^2_{2,1})} + ||\nabla u||_{L2_t(L^2)} are multiplied by factors such as exp(C||\nabla u||_{L1_t(L^\infty)}). Since \dot B^2_{2,1} embeds into \dot B^1_{\infty,1}, this exponential is controlled only in terms of ||u||_{L1_t(\dot B^2_{2,1})}, which is the unknown quantity being estimated. Taking t small does not by itself make the coefficient small unless one already knows a uniform small bound for X(t) on the whole interval. The proof needs a bootstrap or continuity argument showing, at least for the approximate solutions, that X_n(t) tends to 0 as t tends to 0 uniformly in n, and that the exponential factors remain close to 1 on a common interval [0,T1]. As written, the time T1 in (3.7) may depend on the particular solution, and the later assertion in Section 4 that T*_n >= T1 for a common T1 does not follow. This is load-bearing for the existence part of Theorem 1.1.
  2. [Section 4, Step 1, around (4.1)-(4.2)] The passage from the mollified initial data to a fixed existence interval is not established. The authors infer (4.2) from Proposition 3.2 and then conclude that the lifespans T*_n of the mollified solutions satisfy T*_n >= T1 for some T1 independent of n. This conclusion requires the uniform-in-n bootstrap described above; otherwise T1 = T1(n) could shrink to 0, and the compactness argument on [0,T1] cannot be performed. A standard fix would be to prove a quantitative version of Proposition 3.2 in which T1 and the constants depend only on the norms of the initial data, or to use a different mechanism (for example, a diagonal compactness argument combined with uniform global bounds from [14,20]) to obtain a solution on a fixed time interval.
  3. [Section 4, Step 2, equations (4.3) and (4.9)] The uniqueness proof uses Proposition 2.1 at the endpoint alpha = 1, but Proposition 2.1 is stated only for alpha in [0,1), and the proof of the underlying commutator estimate (Lemma 2.1) restricts to alpha < 1. Specifically, equation (4.3) applies (2.7) to \dot B^1_{2,1}, and (4.9) requires a tail estimate for a2 - S_j a2 in B^1_{2,1} on a small time interval. If the intended estimate for alpha = 1 is true, it needs to be stated and proved or cited from a source that covers this endpoint; alternatively, one can use the fact that u lies in L1_t(\dot B^2_{2,1}) subset L1_t(\dot B^1_{\infty,1}), so the flow is Lipschitz and standard transport-regularity arguments give a in C([0,T]; B^1_{2,1}) and the needed tail smallness. As written, the reference to Proposition 2.1 does not cover the case used.
minor comments (6)
  1. [Section 3, after (3.17)] In the sentence 'Therefore, thanks to (3.2) and (3.18), we obtain from (3.19) that ...' the reference to (3.19) appears to be a typo; the inequality being substituted into is (3.17), since (3.19) is the displayed estimate that follows.
  2. [Section 4, Step 1] The text refers to 'Lemma 3.2' but the statement in Section 3 is Proposition 3.2; the numbering should be made consistent.
  3. [Section 4, Step 2, around (4.7)] After deriving (4.7), the proof says 'Then applying Proposition 3.3 to (4.7) leads to ...' but the displayed estimate with the factor (1 + 2^j ||a2||^2_{...}) and the norms in B^{-2}_{2,\infty} is precisely Proposition 2.3, not Proposition 3.3. Please correct the citation.
  4. [Section 2, proof of Proposition 2.2, (2.12)-(2.13)] There are malformed double norms in several displays, e.g. 'sup_{q>= -1} 2^{-q} || [v, \Delta_q P] \cdot \nabla u ||_{L^1_T(L^2)}' has an extra bar; these should be cleaned up for readability.
  5. [Section 3 and Section 4, commutator estimates] The proof repeatedly invokes 'Lemma 1 of [22]' for commutator estimates involving the Leray projector P. Since this is a key technical tool, the lemma should be stated explicitly in the paper or its exact form should be quoted, rather than referenced in passing.
  6. [Section 4, Step 2, after (4.13)] The inequality 'ln(e + alpha x^{-1}) <= ln(e + alpha)(1 - ln x)' is stated for alpha >= 0 and x in (0,1]; the argument also needs the resulting factor (1 - ln x) to be positive, so the smallness of ||\delta u||_{...} should be stated explicitly when applying Osgood's lemma.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the claimed L1_t(dot B^2_{2,1}) estimate is proven from the PDE, not assumed; self-citations are motivational/technical, not load-bearing.

full rationale

Theorem 1.1 rests on the small-time estimate of Proposition 3.2. The proof of that proposition starts from the equations and derives (3.11)-(3.21) by Littlewood-Paley commutator and product estimates. The right-hand side of (3.21) is made absorbable by choosing eta small, k large, and t small; this uses smallness of the heat-tail terms involving the sum over q >= k of 2^{q epsilon} ||Dot Delta_q a_0||_{L^{2/epsilon}} and of exp(C||grad u||_{L^1_t(L^infty)}) - 1, not a prior bound of the form (3.7). Nothing in the chain redefines the target quantity as an input, and no fitted constant is later renamed a prediction. The paper's self-citations [2] and [4] are used for motivation and for techniques or prior related results; the estimate being proved is not imported from them. The external dependencies ([12] for local well-posedness, [14,20] for global H^1 theory, [22] for a commutator estimate) are independent published results. The only substantive concern is at the end of the proof of Proposition 3.2: a uniformity argument showing that the chosen time T1 can be made independent of the mollification index n is not explicitly supplied, and without it Section 4's compactness argument would not be justified. This is a possible gap in the bootstrap/compactness argument, not a circular reduction: no equation is assumed in the form it proves, and the conclusion is not equivalent by construction to an input. Under the hard rules, that does not raise the circularity score.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The proof introduces no new entities and no fitted constants. The auxiliary choices k, m, j, N, eta, gamma, T are parameters selected large or small during the estimates; they do not appear in the theorem and are not fitted to data. The main unproved inputs are the external theorems [12,14,20], used as black boxes.

assumptions (7)
  • standard math Littlewood-Paley theory: Besov space embeddings, Bernstein inequalities, Bony decomposition, Chemin-Lerner spaces as recalled in Appendix A.
    Used throughout Sections 2-4 for all estimates, product laws, commutator estimates, and transport equation bounds.
  • standard math Osgood's lemma (Lemma 2.2) for the logarithmic Gronwall inequality in the uniqueness proof.
    Provides the uniqueness conclusion for the velocity difference from the integral inequality (4.14).
  • domain assumption Local well-posedness for smooth approximate initial data from Danchin [12].
    Step 1 of Theorem 1.1 regularizes the initial data and invokes [12] to obtain unique local solutions (rho_n, u_n, grad Pi_n); the uniform estimates then give a common lifespan.
  • domain assumption Global well-posedness for H^1 velocity data from Danchin-Mucha [14] and Paicu-Zhang-Zhang [20].
    After obtaining u(t0) in H^1 for some t0 > 0, the proof cites [14,20] to conclude global existence; the entire global-in-time statement depends on these external results.
  • domain assumption Bounded density assumption m <= rho0 <= M, equivalently 1 + a bounded above and below.
    Used repeatedly for energy estimates, coercivity of the elliptic pressure problem (Propositions 2.3 and 3.3), and in Proposition 2.2 via the lower bound 1 + a >= c1.
  • domain assumption The viscosity coefficient is a positive constant set to 1, reducing the general system (1.1) to (1.2).
    The theorem concerns the constant-viscosity model (1.2); the variable-viscosity case is not treated.
  • standard math For u0 in \dot B^0_{2,1}, the heat-semigroup quantity sum_j (1 - e^{-c t 2^{2j}}) ||dot Delta_j u0||_{L^2} tends to 0 as t tends to 0, and for a0 in \dot B^epsilon_{2/epsilon,1}, the tail sum_{q>=k} 2^{q epsilon} ||dot Delta_q a0||_{L^{2/epsilon}} tends to 0 as k tends to infinity.
    These smallness properties allow 'large data' to be treated; they follow from convergence of the Besov norms and are essential for closing Proposition 3.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Global well-posedness for the 2-D inhomogeneous incompressible Navier-Stokes system with large initial data in critical spaces." pith.science (2026). https://pith.science/paper/WLRRXFWX

@misc{pith2026190802216,
  author       = {Pith},
  title        = {Pith review of: Global well-posedness for the 2-D inhomogeneous incompressible Navier-Stokes system with large initial data in critical spaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WLRRXFWX}},
  note         = {Machine review of arXiv:1908.02216}
}
read the original abstract

Without any smallness assumption, we prove the global unique solvability of the 2-D incompressible inhomogeneous Navier-Stokes equations with initial data in the critical Besov space, which is almost the energy space in the sense that they have the same scaling in terms of this 2-D system.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

  1. [1]

    Abidi, ´Equation de Navier-Stokes avec densit´ e et viscosit´ e variables dans l’espace critique, Rev

    H. Abidi, ´Equation de Navier-Stokes avec densit´ e et viscosit´ e variables dans l’espace critique, Rev. Mat. Iberoam., 23 (2) (2007), 537–586

    work page 2007

  2. [2]

    Abidi, G

    H. Abidi, G. Gui, and P. Zhang, On the wellposedness of 3 −D inhomogeneous Navier-Stokes equations in the critical spaces, Arch. Rational Mech. Anal. , 204 (2012), 189–230

    work page 2012

  3. [3]

    Abidi and M

    H. Abidi and M. Paicu, Existence globale pour un fluide inh omog´ ene,Ann. Inst. Fourier (Grenoble), 57 (2007), 883–917

    work page 2007

  4. [4]

    Abidi and P

    H. Abidi and P. Zhang, On the global well-posedness of 2-D inhomogeneous incompressible Navier-Stokes system with variable viscous coefficient, J. Differential equtions , 259 (2015), 3755–3802

    work page 2015

  5. [5]

    A. V. Kazhikov, Solvability of the initial-boundary val ue problem for the equations of the motion of an inhomogeneous viscous incompressible fluid, (R ussian) Dokl. Akad. Nauk SSSR , 216 (1974), 1008-1010

    work page 1974

  6. [6]

    Bahouri, J

    H. Bahouri, J. Y. Chemin and R. Danchin, Fourier Analysis and Nonlinear Partial Differ- ential Equations , Grundlehren der Mathematischen Wissenschaften, Springe r, 2010

    work page 2010

  7. [7]

    Bergh, J

    J. Bergh, J. L¨ ofstr¨ om,Interpolation spaces. An introduction , Springer-Verlag, 1976

    work page 1976

  8. [8]

    J. M. Bony, Calcul symbolique et propagation des singula rit´ es pour les ´ equations aux d´ eriv´ ees partielles non lin´ eaires,Ann. Sci. ´Ecole Norm. Sup. , 14 (1981), 209–246

    work page 1981

Show all 23 references
  1. [9]

    Chemin, Th´ eor´ emes d’unicit´ e pour le syst´ eme de Navier-Stokes tridimensionnel, J

    J.-Y. Chemin, Th´ eor´ emes d’unicit´ e pour le syst´ eme de Navier-Stokes tridimensionnel, J. Anal. Math. , 77, (1999), 27–50

    1999

  2. [10]

    Chemin and N

    J.-Y. Chemin and N. Lerner, Flot de champs de vecteurs no n lipschitziens et ´ equations de Navier-Stokes, J. Differential Equations , 121, (1995), 314–328

    1995

  3. [11]

    Danchin, Local theory in critical spaces for compres sible viscous and heat-conductive gases, Comm

    R. Danchin, Local theory in critical spaces for compres sible viscous and heat-conductive gases, Comm. Partial Differential Equations , 26 (2001), 1183-1233

    2001

  4. [12]

    Danchin, Local and global well-posedness resultats for flows of inhomogenenous viscous fluids, Adv

    R. Danchin, Local and global well-posedness resultats for flows of inhomogenenous viscous fluids, Adv. differential equations , 9 (2004), 353–386

    2004

  5. [13]

    Danchin, The inviscid limit for density-dependent i ncompressible fluids, Anneles de la Facult´ e de Sciences des Toulouse S´ er., 15 (2006), 637–688

    R. Danchin, The inviscid limit for density-dependent i ncompressible fluids, Anneles de la Facult´ e de Sciences des Toulouse S´ er., 15 (2006), 637–688

    2006

  6. [14]

    Danchin and P

    R. Danchin and P. B. Mucha, The incompressible Navier-S tokes equations in vacuum, Comm. Pure. Appl. Math. , 72 (7) (2019), 1351–1385

    2019

  7. [15]

    Danchin and P

    R. Danchin and P. B. Mucha, A Lagrangian approach for the incompressible Navier-Stokes equations with variable density, Comm. Pure. Appl. Math. , 65 (2012), 1458-1480

    2012

  8. [16]

    R. J. DiPerna and P. L. Lions, Equations diff´ erentielle s ordinaires et ´ equations de trans- port avec des coefficients irr´ eguliers. In S´ eminaire EDP 1988-1989, Ecole Polytechnique, Palaiseau, 1989

    1988

  9. [17]

    T. M. Fleet, Differential analysis, Cambridge University Press , 1980. 24

    1980

  10. [18]

    O. A. Ladyˇ zenskaja and V. A. Solonnikov, The unique sol vability of an initial-boundary value problem for viscous incompressible inhomogeneous flu ids. (Russian) Boundary value problems of mathematical physics, and related questions of the theory of functions, 8, Zap. Nauˇ cn...

    1975

  11. [19]

    P. L. Lions, Mathematical Topics in Fluid Mechanics. Vol. 1. Incompressi ble Models, Oxford Lecture Series in Mathematics and its Applications, 3. Oxfo rd Science Publications. The Clarendon Press, Oxford University Press, New York, 1996

    1996

  12. [20]

    Paicu, P

    M. Paicu, P. Zhang and Z. Zhang, Global unique solvabili ty of inhomogeneous Navier-Stokes equations with bounded density, Communications in Partial Differential Equations , 38 (7) (2013), 1208–1234

    2013

  13. [21]

    Peetre, New thoughts on Besov spaces, Duke University Mathematical Series 1, Durham N

    J. Peetre, New thoughts on Besov spaces, Duke University Mathematical Series 1, Durham N. C. 1976

    1976

  14. [22]

    Planchon, An extension of the Beale-Kato-Majda crit erion for the Euler equations, Comm

    F. Planchon, An extension of the Beale-Kato-Majda crit erion for the Euler equations, Comm. Math. Phys. , 232 (2003), 319-326

    2003

  15. [23]

    Triebel, Theory of Function Spaces, Monograph in mat hematics, Vol

    H. Triebel, Theory of Function Spaces, Monograph in mat hematics, Vol. 78 (1983), Birkhauser Verlag, Basel. 25

    1983

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.