Abstract
This paper is concerned with two dual aspects of the regularity question for the Navier–Stokes equations. First, we prove a local-in-time localized smoothing effect for local energy solutions. More precisely, if the initial data restricted to the unit ball belongs to the scale-critical space \(L^3\), then the solution is locally smooth in space for some short time, which is quantified. This builds upon the work of Jia and Šverák, who considered the subcritical case. Second, we apply these localized smoothing estimates to prove a concentration phenomenon near a possible Type I blow-up. Namely, we show if \((0, T^*)\) is a singular point, then
This result is inspired by and improves concentration results established by Li, Ozawa, and Wang and Maekawa, Miura, and Prange. We also extend our results to other critical spaces, namely \(L^{3,\infty }\) and the Besov space \({\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }\), \(p\in (3,\infty )\).
Similar content being viewed by others
Notes
In this paper, the parabolic Hölder semi-norm is defined in the following way:
$$\begin{aligned}{}[\cdot ]_{C^{0,\nu }_{par}({\mathbb {R}}^3\times [0,T])}:=[\cdot ]_{C^{0,\nu }_{t}([0,T];L^\infty _x({\mathbb {R}}^3))}+[\cdot ]_{L^\infty _t(0,T;C^{0,2\nu }_{x}({\mathbb {R}}^3))}. \end{aligned}$$The Navier–Stokes equations are invariant under the scaling \((u^{(\lambda )}(x,t), p^{(\lambda )}(x,t))= (\lambda u(\lambda x, \lambda ^2 t), \lambda ^2 p(\lambda x, \lambda ^2 t))\), \(u_{0}^{(\lambda )}(x)= \lambda u_{0}(\lambda x)\). We say that a space \(X\subset {\mathcal {S}}^{'}({\mathbb {R}}^3)\) is critical (or scale-invariant) if its norm is invariant under the above rescaling for the initial data. Likewise, we say \(X_{T}\subset {\mathcal {S}}^{'}({\mathbb {R}}^3)\) is critical if its norm is invariant under the rescaling for the velocity field.
This particular advantage was also exploited by Kukavica in [31].
The choice \(s=d/p\), \(q=1\) is also valid.
By definition \((\cdot )_+:=\max (0,\cdot )\).
References
Albritton, D.: Blow-up criteria for the Navier–Stokes equations in non-endpoint critical Besov spaces. Anal. PDE11(6), 1415–1456, 2018
Albritton, D., Barker, T.: Global weak Besov solutions of the Navier-Stokes equations and applications. Arch. Ration. Mech. Anal. 232, 197–263, 2019. https://doi.org/10.1007/s00205-018-1319-0
Albritton, D., Barker, T.: Localised necessary conditions for singularity formation in the Navier–Stokes equations with curved boundary. arXiv preprint arXiv:1811.00507, 2018
Barker, T.: Existence and weak* stability for the Navier–Stokes system with initial values in critical Besov spaces. arXiv preprint arXiv:1703.06841, 2017
Barker, T.: Uniqueness results for weak Leray–Hopf solutions of the Navier–Stokes system with initial values in critical spaces. J. Math. Fluid Mech. 20(1), 133–160, 2018
Barker, T., Seregin, G.: A necessary condition of potential blowup for the Navier–Stokes system in half-space. Math. Ann. 369(3–4), 1327–1352, 2017
Barker, T., Seregin, G., Šverák, V.: On stability of weak Navier–Stokes solutions with large \(L^{3,\infty }\) initial data. Commun. Partial Differ. Equ. 43(4), 628–651, 2018
Buckmaster, T., Vicol, V.: Nonuniqueness of weak solutions to the Navier–Stokes equation. arXiv preprint arXiv:1709.10033, 2017
Caffarelli, L., Kohn, R., Nirenberg, L.: Partial regularity of suitable weak solutions of the Navier–Stokes equations. Commun. Pure Appl. Math. 35(6), 771–831, 1982
Calderón, C.P.: Existence of weak solutions for the Navier–Stokes equations with initial data in \(L^p\). Trans. Am. Math. Soc. 318(1), 179–200, 1990
Dubois, S.: Uniqueness for some Leray–Hopf solutions to the Navier–Stokes equations. J. Differ. Equ. 189(1), 99–147, 2003
Dong, H., Gu, X.: Boundary partial regularity for the high dimensional Navier–Stokes equations. J. Funct. Anal. 267(8), 2606–2637, 2014
Dong, H., Gu, X.: Partial regularity of solutions to the four-dimensional Navier–Stokes equations. Dyn. Partial Differ. Equ. 11(1), 53–69, 2014
Dong, H., Wang, K.: Interior and boundary regularity for the Navier–Stokes equations in the critical Lebesgue spaces. ArXiv e-prints arXiv:1809.06712, Sept. 2018
Galdi, G.P.: An introduction to the mathematical theory of the Navier–Stokes equations. Springer Monographs in Mathematics. Springer, New York, 2nd edn, 2011. Steady-state problems
Gallagher, I., Koch, G.S., Planchon, F.: Blow-up of critical Besov norms at a potential Navier–Stokes singularity. Commun. Math. Phys. 343(1), 39–82, 2016
Giga, Y.: Solutions for semilinear parabolic equations in \(L^p\) and regularity of weak solutions of the Navier–Stokes system. J. Differ. Equ. 62(2), 186–212, 1986
Giga, Y.: A remark on a Liouville problem with boundary for the Stokes and the Navier–Stokes equations. Discrete Contin. Dyn. Syst. Ser. S6(5), 1277–1289, 2013
Guillod, J., Šverák, V.: Numerical investigations of non-uniqueness for the Navier–Stokes initial value problem in borderline spaces. arXiv preprint arXiv:1704.00560, 2017
Hmidi, T., Keraani, S.: Remarks on the blowup for the \(L^2\)-critical nonlinear Schrödinger equations. SIAM J. Math. Anal. 38(4), 1035–1047, 2006
Holmer, J., Roudenko, S.: On blow-up solutions to the 3D cubic nonlinear Schrödinger equation. Appl. Math. Res. Express. AMRX, (1):Art. ID abm004, 31, 2007
Hunt, R.A.: On \(L(p,\, q)\) spaces. Enseign. Math. (2)12, 249–276, 1966
Jia, H., Šverák, V.: Minimal \(L^3\)-initial data for potential Navier–Stokes singularities. SIAM J. Math. Anal. 45(3), 1448–1459, 2013
Jia, H., Šverák, V.: Local-in-space estimates near initial time for weak solutions of the Navier–Stokes equations and forward self-similar solutions. Invent. Math. 196(1), 233–265, 2014
Jia, H., Šverák, V.: Are the incompressible 3D Navier–Stokes equations locally ill-posed in the natural energy space? J. Funct. Anal. 268(12), 3734–3766, 2015
Kang, K., Miura, H., Tsai, T.-P.: Short time regularity of Navier–Stokes flows with locally \(L^3\) initial data and applications. Int. Math. Res. Not. rnz327 (2020). https://doi.org/10.1093/imrn/rnz327
Kato, T.: Strong \(L^{p}\)-solutions of the Navier–Stokes equation in \({ R}^{m}\), with applications to weak solutions. Math. Z. 187(4), 471–480, 1984
Kikuchi, N., Seregin, G.: Weak solutions to the Cauchy problem for the Navier–Stokes equations satisfying the local energy inequality. In: Nonlinear Equations and Spectral Theory, volume 220 of American Mathematical Society translation Seres 2, pp 141–164. American Mathematical Society, Providence, RI, 2007
Koch, H., Tataru, D.: Well-posedness for the Navier–Stokes equations. Adv. Math. 157(1), 22–35, 2001
Kukavica, I.: On partial regularity for the Navier–Stokes equations. Discrete Contin. Dyn. Syst. 21(3), 717–728, 2008
Kukavica, I.: Partial regularity for the Navier–Stokes equations with a force in a Morrey space. J. Math. Anal. Appl. 374(2), 573–584, 2011
Ladyzhenskaya, O.A., Seregin, G.A.: On partial regularity of suitable weak solutions to the three-dimensional Navier–Stokes equations. J. Math. Fluid Mech. 1(4), 356–387, 1999
Lemarié-Rieusset, P.G.: Recent Developments in the Navier–Stokes Problem, vol. 431. Chapman & Hall/CRC Research Notes in Mathematics. Chapman & Hall/CRC, Boca Raton 2002
Leray, J.: Sur le mouvement d’un liquide visqueux emplissant l’espace. Acta Math. 63(1), 193–248, 1934
Li, K., Ozawa, T., Wang, B.: Dynamical behavior for the solutions of the Navier–Stokes equation. Commun. Pure Appl. Anal. 17(4), 1511–1560, 2018
Li, K., Wang, B.: Blowup criterion for Navier–Stokes equation in critical Besov space with spatial dimensions \( d\ge 4\). arXiv preprint arXiv:1803.04076, 2018
Lin, F.: A new proof of the Caffarelli–Kohn–Nirenberg theorem. Commun. Pure Appl. Math. 51(3), 241–257, 1998
Maekawa, Y., Miura, H., Prange, C.: Estimates for the Navier–Stokes equations in the half-space for non localized data. ArXiv e-prints, Nov. 2017
Maekawa, Y., Terasawa, Y.: The Navier–Stokes equations with initial data in uniformly local \(l^p\) spaces. Differ. Integral Equ. 19(4), 369–400, 2006
Merle, F., Tsutsumi, Y.: \(L^2\) concentration of blow-up solutions for the nonlinear Schrödinger equation with critical power nonlinearity. J. Differ. Equ. 84(2), 205–214, 1990
Meyer, Y.: Wavelets, paraproducts, and Navier–Stokes equations. In Current developments in mathematics, 1996 (Cambridge, MA), pp. 105–212. International Press, Boston, MA, 1997
O’Leary, M.: Conditions for the local boundedness of solutions of the Navier–Stokes system in three dimensions. Commun. Partial Differ. Equ. 28(3–4), 617–636, 2003
O’Neil, R.: Convolution operators and \(l(p, q)\) spaces. Duke Math. J. 30(1), 129–142, 1963
Phuc, N.C.: The Navier–Stokes equations in nonendpoint borderline Lorentz spaces. J. Math. Fluid Mech. 17(4), 741–760, 2015
Planchon, F.: Global strong solutions in Sobolev or Lebesgue spaces to the incompressible Navier–Stokes equations in \({\bf R}^3\). Ann. Inst. H. Poincaré Anal. Non Linéaire13(3), 319–336, 1996
Robinson, J.C., Rodrigo, J.L., Sadowski, W.: The three-dimensional Navier–Stokes equations, volume 157 of Cambridge Studies in Advanced Mathematics. Cambridge University Press, Cambridge, Classical theory, 2016
Seregin, G.: A certain necessary condition of potential blow up for Navier–Stokes equations. Commun. Math. Phys. 312(3), 833–845, 2012
Seregin, G., Šverák, V.: Regularity criteria for Navier–Stokes solutions. In: Giga, Y., Novotný, A. (eds.) Handbook of Mathematical Analysis in Mechanics of Viscous Fluids, pp. 829–867. Springer, Cham 2018
Seregin, G.A.: A new version of the Ladyzhenskaya–Prodi–Serrin condition. Algebra i Analiz18(1), 124–143, 2006
Seregin, G.A., Zajaczkowski, W.: A sufficient condition of local regularity for the Navier–Stokes equations. Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI), 336(Kraev. Zadachi Mat. Fiz. i Smezh. Vopr. Teor. Funkts. 37:46–54, 274, 2006
Acknowledgements
Both authors warmly thank the OxPDE centre, where this work started. The second author would also like to thank Yasunori Maekawa and Hideyuki Miura for stimulating discussions about concentration for blowing-up solutions of the Navier–Stokes equations.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Funding
The second author is partially supported by the project BORDS (ANR-16-CE40-0027-01) operated by the French National Research Agency (ANR). The second author also acknowledges financial support from the IDEX of the University of Bordeaux for the BOLIDE project.
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Communicated by V. Šverák
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: Auxiliary Results
The first result is the classical existence of mild solutions for critical initial data \(a\in L^3_\sigma ({\mathbb {R}}^3)\).
Proposition 15
([17]). There exists universal constants \(\gamma ,\, K_0\in (0,\infty )\) such that the following holds true. For for all \(u_{0,a}\in L^3_\sigma ({\mathbb {R}}^3)\) with \(\Vert u_{0,a}\Vert _{L^3}\leqq \gamma \), there exists a unique smooth mild solution \(a\in C([0,\infty );L^3)\cap L^\infty ((0,\infty );L^3)\) such that \(a(\cdot ,0)=u_{0,a}\), \(a\in C((0,\infty );W^{1,3}\cap BUC_\sigma )\) and
Moreover \(a\in L^5({\mathbb {R}}^3\times (0,\infty ))\) and
The second result is an estimate for local energy solutions, so-called Lemarié-Rieusset solutions [33, Chapter 32 and 33]. Before stating this result, we give the definition of such solutions. See also [28] and [24, Definition 3.1].
Definition 16
(Local energy solutions). A pair (u, p) is called a local energy solution to (1) in \({\mathbb {R}}^3\times (0,\infty )\) with the initial data \(u_0\in L^2_{uloc,\sigma }({\mathbb {R}}^3)\), \(\sup _{|{{\bar{x}}}|\geqq R}\Vert u_0\Vert _{L^2(B_1({{\bar{x}}}))}{\mathop {\longrightarrow }\limits ^{R\rightarrow \infty }} 0\), if (u, p) satisfies the following conditions:
- (i)
We have \(u\in L^\infty _{loc}([0,T); L^2_{uloc,\sigma } ({\mathbb {R}}^3))\) and \(p\in L^\frac{3}{2}_{loc} ({\mathbb {R}}^3\times (0,\infty ))\), and
$$\begin{aligned} \sup _{{{\bar{x}}}\in {\mathbb {R}}^3} \int \nolimits _0^{T} \Vert \nabla u \Vert _{L^2 (B_1({{\bar{x}}})\cap {\mathbb {R}}^3)}^2 d t <\infty , \end{aligned}$$for all finite \(T\in (0, \infty )\).
- (ii)
The pair (u, p) is a solution to (1) in the sense of distributions.
- (iii)
The function \(t\mapsto \langle u(\cdot ,t), \varphi \rangle _{L^2({\mathbb {R}}^3)}\) belongs to C([0, T)) for any compactly supported \(\varphi \in L^2({\mathbb {R}}^3)\). Moreover, for any compact set \(K\subset {\mathbb {R}}^3\),
$$\begin{aligned} \lim _{t\rightarrow 0} \Vert u(\cdot ,t) -u_0 \Vert _{L^2(K)} =0. \end{aligned}$$(98) - (iv)
The pair (u, p) satisfies the local energy inequality: for any \(0\leqq \phi \in C^\infty _c({\mathbb {R}}^3\times (0,\infty ))\), for all \(t\in (0,\infty )\),
$$\begin{aligned} \begin{aligned}&\int \nolimits _{B_1(0)}|u(x,t)|^2\phi (x,t) \,\text {d}x+2\int \nolimits _{0}^t\int \nolimits _{{\mathbb {R}}^3}|\nabla u|^2\phi \,\text {d}x\,\text {d}s\\&\quad \leqq \ \int \nolimits _{0}^t\int \nolimits _{{\mathbb {R}}^3}|u|^2(\partial _t\phi +\Delta \phi )\,\text {d}x\,\text {d}s+\int \nolimits _{0}^t\int \nolimits _{{\mathbb {R}}^3}(|u|^2+2p)u\cdot \nabla \phi \,\text {d}x\,\text {d}s. \end{aligned} \end{aligned}$$
Notice that (98) enables one to transfer the mild decay of the initial data to the solution u.
Proposition 17
([24, 33, Lemma 3.1]). There exist two universal constants \(c_1,\ K_1\in (0,\infty )\) such that for all \(u_0\in L^2_{uloc}({\mathbb {R}}^3)\), for all local energy solution u to (1) with initial data \(u_0\), we have
where
Following [23, 28, 33], if u is a local energy weak solution to (1) in the sense of Definition 16 and a is a mild solution to (1), then \(v-a\) solves
in the sense of distributions, and we have the following global representation formula for the pressure: for all \({{\bar{x}}}\in {\mathbb {R}}^3\), for all \((x,t)\in B_\frac{3}{2}({{\bar{x}}})\times (0,T)\),
Here \(N(x)=-\frac{1}{4\pi |x|}\). Notice that (100) provides a proof of estimate (84) in the case \(a=0\). With \({{\bar{x}}}=0\), we have for all \((x,t)\in B_\frac{3}{2}(0)\times (0,S_{lews})\),
with
where we used Calderón-Zygmund estimates for the first bound.
We conclude this appendix by a useful local representation formula for the pressure. The following lemma is well known, see for instance Caffarelli, Kohn and Nirenberg’s paper [9] (specifically p.782 of [9]).
Lemma 18
Suppose that \(p\in L^{1}(B_{2}(0))\) and \(V_{ij}\in L^{q}(B_{2}(0))\) for some \(q>1\). Furthermore, suppose
in \(B_{2}(0)\) in a distributional sense. Then for \(\varphi \in C_{0}^{\infty }(B_{1}(0))\) we have
Appendix B: The Case of \(\mathbf{L }^{{\varvec{3,\infty }}}\) Initial Data
1.1 B.1. Preliminary Material
Given a measurable subset \(\Omega \subseteq {\mathbb {R}}^{d}\), let us define the Lorentz spaces. For a measurable function \(f:\Omega \rightarrow {\mathbb {R}}\) define:
where \(\mu \) denotes the Lebesgue measure. The Lorentz space \(L^{p,q}(\Omega )\), with \(p\in [1,\infty [\), \(q\in [1,\infty ]\), is the set of all measurable functions g on \(\Omega \) such that the quasinorm \(\Vert g\Vert _{L^{p,q}(\Omega )}\) is finite. Here,
It is known there exists a norm, which is equivalent to the quasinorm defined above, for which \(L^{p,q}(\Omega )\) is a Banach space. For \(p\in [1,\infty )\) and \(1\leqq q_{1}< q_{2}\leqq \infty \), we have the following continuous embeddings:
and the inclusion is known to be strict.
Let X be a Banach space with norm \(\Vert \cdot \Vert _{X}\), \( a<b\), \(p\in [1,\infty )\) and \(q\in [1,\infty ]\). Then \(L^{p,q}(a,b;X)\) will denote the space of strongly measurable X-valued functions f(t) on (a, b) such that
In particular, if \(1\leqq q_{1}< q_{2}\leqq \infty \), we have the following continuous embeddings:
and the inclusion is known to be strict.
Let us recall a known proposition known as ‘O’Neil’s convolution inequality’ (Theorem 2.6 of O’Neil’s paper [43]).
Proposition 19
Suppose \(1< p_{1}, p_{2}, r<\infty \) and \(1\leqq q_{1}, q_{2}, s\leqq \infty \) are such that
and
Suppose that
Then
We will use an inequality that we will refer to as ‘Hunt’s inequality’. The statement below and proof can be found in Hunt’s paper [22] (Theorem 4.5, p. 271 of [22]).
Proposition 20
Suppose that \(0<p,q,r\leqq \infty \) and \(0<s_1,s_2\leqq \infty \). Furthermore, suppose that p, q, r, \(s_1\) and \(s_2\) satisfy the following relations:
and
Then the assumption that \(f\in L^{p,s_1}(\Omega )\) and \(g\in L^{q,s_2}(\Omega )\) implies that \(fg \in L^{r,s}(\Omega )\), with the estimate
As a result of the above Propositions, we have the following estimates with \(B_{R}(0)\subset {\mathbb {R}}^3\) (which we will frequently use):
The first estimate is stated and proven in [11], for example. Now, we state known results for the Navier–Stokes equations with initial data in \(L^{3,\infty }({\mathbb {R}}^3).\) We refer the reader to [41] and [45].
Proposition 21
There exists universal constants \(\gamma ,\, K_0'\in (0,\infty )\) such that the following holds true. For all \(u_{0,a}\in L^{3,\infty }_\sigma ({\mathbb {R}}^3)\), \(\Vert u_{0,a}\Vert _{L^{3,\infty }}\leqq \gamma \), there exists a smooth mild solution \(a\in C_{w^{*}}([0,\infty );L^{3,\infty })\cap L^\infty ((0,\infty );L^{3,\infty })\) such that \(a(\cdot ,0)=u_{0,a}\) and
The mild solution is unique in the class of solutions with small enough \(L^\infty (0,\infty ; L^{3,\infty })\) norm.
1.2 B.2. \(L^{3,\infty }\) Initial Data: Section 2
We briefly describe the changes required for Section 2. With the above Proposition in mind concerning mild solutions, in Section 2 we can no longer assume a is in \(L^{5}_{x,t}\). Instead, we assume that
The first adjustment regards the estimate of the pressure (Lemma 4). In particular, Hunt’s inequality can be used to show that the second and last term in (37) can be replaced by
and
Now we proceed to the adjustments needed for the proof of Theorem 4. In Step 2 and Step 3 the only adjustment is to make extensive use of (115)–(116). In Step 4 we take the adjustment of Lemma 4 into account. Moreover, when estimating the pressure we have to use Hunt’s inequality to estimate \(J_{4}\). In particular, this gives
1.3 B.3. \(L^{3,\infty }\) Initial Data: Section 3
As in Section 2 we can no longer assume a is in \(L^{5}_{x,t}\). Instead, we assume that
where \(\varepsilon _{**}>0\) is some small universal constant. The only difference in Section 3 regards (63). In particular, we can use Young’s inequality in space, followed by O’Neil’s convolution inequality in time and finally Hunt’s inequality in time to see that the following holds: for \(1<q<\infty \),
1.4 B.4. \(L^{3,\infty }\) Initial Data: Section 4
First (86) must be adjusted using Hunt’s inequality as follows:
Then in (93), we must instead make use of (115)–(116).
Appendix C: The Case of Besov Initial Data
1.1 C.1. Preliminaries
Let \(d, m \in {\mathbb {N}}\setminus \{0\}\). We begin by recalling the definition of the homogeneous Besov spaces\(\dot{B}^s_{p,q}({\mathbb {R}}^d;{\mathbb {R}}^m)\). There exists a non-negative radial function \(\varphi \in C^\infty ({\mathbb {R}}^d)\) supported on the annulus \(\{ \xi \in {\mathbb {R}}^d : 3/4 \leqq |\xi | \leqq 8/3 \}\) such that
The homogeneous Littlewood-Paley projectors \({{\dot{\Delta }}}_j\) are defined by
for all tempered distributions f on \({\mathbb {R}}^d\) with values in \({\mathbb {R}}^m\). The notation \(\varphi (2^{-j}D) f\) denotes convolution with the inverse Fourier transform of \(\varphi (2^{-j}\cdot )\) with f.
Let \(p,q \in [1,\infty ]\) and \(s \in (-\infty ,d/p)\).Footnote 4 The homogeneous Besov space \(\dot{B}^s_{p,q}({\mathbb {R}}^d;{\mathbb {R}}^m)\) consists of all tempered distributions f on \({\mathbb {R}}^d\) with values in \({\mathbb {R}}^m\) satisfying
and such that \(\sum _{j \in {\mathbb {Z}}} {{\dot{\Delta }}}_j f\) converges to f in the sense of tempered distributions on \({\mathbb {R}}^d\) with values in \({\mathbb {R}}^m\). In this range of indices, \(\dot{B}^{s}_{p,q}({\mathbb {R}}^d;{\mathbb {R}}^m)\) is a Banach space. When \(s \geqq 3/p\) and \(q > 1\), the spaces must be considered modulo polynomials. Note that other reasonable choices of the function \(\varphi \) defining \({{\dot{\Delta }}}_j\) lead to equivalent norms.
We now recall a particularly useful property of Besov spaces, that is, their characterization in terms of the heat kernel. For all \(s \in (-\infty ,0)\), there exists a constant \(c := c(s) > 0\) such that for all tempered distributions f on \({\mathbb {R}}^3\),
Let \(\Omega \subset {\mathbb {R}}^3\) be a domain with sufficiently smooth boundary. We say \(u\in {\dot{B}}^{s}_{p,q}(\Omega )\) if
- (*)
(extension property) there exists \(E(u)\in {\dot{B}}^{s}_{p,q}({\mathbb {R}}^3)\) such that \(E(u)=u\) on \(\Omega \) as distributions.
Then
In what follows, we will mostly use just one feature of the definition of Besov spaces on bounded domains:
The proof of this uses the definition of Besov spaces on bounded domains and the fact that for \(\varphi \) in the Schwartz class
The proof of this is along the lines of Proposition 2.3 of [36].
We will also make use of a decomposition result for Homogeneous Besov spaces. The statement without (130) can be found in [2]; see also [4].
Lemma 22
Let \(p \in (3,\infty )\). There exist \(\gamma _1, \gamma _2 > 0\), and \(C > 0\), each depending only on p, such that for each divergence-free vector field \(g \in {\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }\cap L^{2}({\mathbb {R}}^3)\) and \(N>0\), there exist divergence-free vector fields \({\bar{g}}^{N}\in {\dot{B}}^{-1+\delta _{2}}_{\infty ,\infty }({\mathbb {R}}^3)\cap {\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }({\mathbb {R}}^3)\cap L^{2}({\mathbb {R}}^3)\) and \({\widetilde{g}}^{N}\in L^2({\mathbb {R}}^3)\cap {\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }({\mathbb {R}}^3)\) with the following properties:
Furthermore,
and
Finally, we state known results for the Navier–Stokes equations with data in \({\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }({\mathbb {R}}^3).\) We refer the reader to [45], for example.
Proposition 23
Let \(S_{mild}\in (0,\infty )\) and \(p\in (3,\infty )\). There exists two constants \(\gamma (p)\in (0,\infty )\) and \(K_0''(p)\in (0,\infty )\) such that the following holds true. For all divergence-free \(u_{0,a}\in {\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }({\mathbb {R}}^3)\),
there exists a smooth mild solution \(a\in C_{w^{*}}([0,S_{mild});{\dot{B}}^{-1+\frac{3}{p}}_{p,\infty })\cap L^\infty ((0,S_{mild});{\dot{B}}^{-1+\frac{3}{p}}_{p,\infty })\) such that \(a(\cdot ,0)=u_{0,a}\) and
The mild solution is unique in the class of solutions with sufficiently small
norm.
1.2 C.2. Besov Initial Data: Section 3
In this section we should now assume
where \(\varepsilon _{**}>0\) is some small universal constant. With this adjustment, the arguments in Section 3 are the same as in the case of \(L^{3,\infty }\) initial data.
1.3 C.3. Besov Initial Data: Section 4
1.3.1 C.3.1. The Extension Operator
Throughout this part we assume
For convenience, we assume without loss of generality that \(p\in (6,\infty )\). Let \(\chi \in C^\infty _c({\mathbb {R}}^3)\) be a cut-off function such that
where \(K_3\in (0,\infty )\) is a universal constant. From the preliminaries we have
Obviously,
Then, we introduce \({{\tilde{u}}}_{0,a}\) given by Bogovskii’s lemma, such that
where \(K_4\in (0,\infty )\) is a universal constant. We extend \({{\tilde{u}}}_{0,a}\) by 0 and let
Clearly,
Since \({\tilde{u}}_{0,a}\) has compact support and \(L_{3}\hookrightarrow {\dot{B}}^{-1+\frac{3}{p}}_{p,\infty }\) we have
Using (132)–(133), together with the heat flow characterisation of Besov spaces gives
Using this and Proposition 23, there exists \({\hat{T}}(M,\gamma )\) and a mild solution \(a(\cdot , u_{0,a})\) associated to \(u_{0,a}\) on \({\mathbb {R}}^3 \times (0, {\hat{T}}(M,\gamma ))\). Furthermore,
Moreover, (134)–(136) and Theorem 3.1 of [5] imply that for \(t\in (0, {\hat{T}}(M,\gamma ))\):
1.3.2 C.3.2. Local Decay Estimates Near the Initial Time
Now clearly \(v=u-a\) has zero initial data locally on the ball \(B_{\frac{3}{2}}(0)\). We next wish to show that for \(t\in (0, \min (1,{\hat{T}},S_{lews}))\) we have
for some \(\nu (p)>0\). With (137) in mind, it is sufficient to show that for \(t\in (0, \min ( 1, S_{lews}) )\):
In order to show this, we use splitting arguments inspired by the work of Cálderón [10]. The arguments we present here closely follow those presented in [5, 23] and [2]. According to Lemma 22 and (134)–(135), we can split \(u_{0,a}\) into two divergence-free pieces:
Define \(u^{N}:= u- e^{t\Delta }\overline{u_{0,a}}^{N}\). Then
We remark that p is the pressure associated to the original local energy solution u. From Proposition 17 and (141), we have
Let \(\phi \in C^\infty _c({\mathbb {R}}^3)\) such that \(0\leqq \phi \leqq 1\), \({{\,\mathrm{supp}\,}}\phi \subset B_\frac{3}{2}(0)\), \(\phi =1\) on \(B_1(0)\), \(|\nabla (\phi ^2)|\leqq K_5\) and \(|\Delta (\phi ^2)|\leqq K_5'\) where \(K_5,\ K_5'\in (0,\infty )\) are a universal constants.
For \(t\in (0,\min (1, S_{lews}))\) we have:
Using (139), we have \(I_{0}\leqq C(M, \gamma ) N^{-2\gamma _{2}}.\) Using (142) and the same arguments as in Section 4 gives
Furthermore, using (142) and (140) we obtain
Next, we may use (140)–(141) to see that
This may be used with (142) to show that
Thus for \(t\in (0,\min (1, S_{lews}))\) we have
Noting that \(u-e^{t\Delta } u_{0,a}= u^{N}- \widetilde{u_{0,a}}^{N}\), we thus obtain for \(t\in (0,\min (1, S_{lews}))\) and \(N\in (0,\infty )\) that
Choosing \(N= t^{-\beta }\), where \(\beta >0\) is sufficiently small, then yields the desired estimate (C.3.2).
1.4 C.4. Besov Initial Data: Section 2
In this section we give the adjustments needed to prove Theorem 4 in the case of a drift a, which rather than satisfying the global \(L^5(Q_1(0,0))\) bound (27), just satisfies
and small, for a fixed \(t_0\in [-1,0]\). This extension is needed to deal with the case of locally Besov initial data \(\dot{B}^{-1+\frac{3}{p}}_{p,\infty }\), with \(p=5\), for which the mild solution just satisfies (131). We actually prove the following theorem which allows to handle any \(p\in (3,\infty )\):
Theorem 6
Let \(t_0\in [-1,0]\) and \(\eta \in (0,1)\) be fixed. For all \(\delta \in (0,3)\), there exists \(C_*(\delta )\in (0,\infty )\), for all \(E\in (0,\infty )\), there exists \(\varepsilon _*(\delta ,\eta ,E)\in (0,\infty )\), for all a such that
and all local suitable solutions v to (24) in \(Q_1(0,0)\) such thatFootnote 5
the conditions
and
imply that for all \(({{\bar{x}}},t)\in {{\bar{Q}}}_{1/2}(0,0)\), for all \(r\in (0,\frac{1}{4}]\),
We note that (143) implies in particular that for all \(s\in (-1,0)\), \(v(\cdot ,s)=0\) almost everywhere on \(B_1(0)\). As was emphasized just below Theorem 4, the constant \(C_*\) only depends on \(\delta \), because it arises when going from scale \(r_n\) to \(r\in (r_{n+1},r_n)\).
The proof goes through using the same general scheme as in Section 2. The main difficulty is that the bound (145) does not imply \(a\in L^\frac{2}{1-\frac{3}{p}}(-1,0;L^p(B_1(0)))\). Hence estimates on the term
carried out in Section 2 do not work as such any longer. One possible way out is to use estimate (C.3.2), which allows us to remove the singularity due to (145). Consequently, there are two main modifications to the argument in Section 2. The first modification is on the bounds (\(A_k\)) and (\(B_k\)) which are iterated. The second modification is on Lemma 4 for the pressure.
Let \(({{\bar{x}}},t)\in {{\bar{Q}}}_{1/2}(0,0)\) be fixed for the rest of this section. For all \(n\in {\mathbb {N}}\), we let \(r_n:=2^{-n}\). Our aim is to propagate for \(k\geqq 2\) the following three bounds:
and
where
for \(\eta '=\frac{\eta }{6}\in (0,\frac{1}{6})\) and constants \(\varepsilon _*(\delta ,\eta ,E),\, C_B(\delta ,\eta )\in (0,\infty )\) to be chosen. Notice that the power \(\frac{3}{4}\eta '\) in (\(A_k''\)) is worse than the corresponding power in (\(A_k'\)). This fact appears in Step 4 of the proof of Theorem 6. It is due to the fourth term in the right hand side of (149) below.
We also need the following modification of Lemma 4:
Lemma 24
(Pressure estimate). There exists a constant \(C_2'\in (0,\infty )\) such that for all \(\rho \in (0,\infty )\), and for all a such that
and for all weak solutions \(q\in L^\frac{3}{2}(Q_\rho (0,0))\) to
we have
for all \(s\in (-r^2,0)\), for all \(0<r\leqq \rho /2\).
Sketch of the proof of Theorem 6
In the whole proof, we define\(M_a\)as in (148). Notice that by assumption (145), \(M_a\leqq \varepsilon _*\). Let us sketch the main differences with respect to the proof of Theorem 4 in Section 2. We focus on the case when \(({{\bar{x}}},t)=(0,0)\), but the argument for general \(({{\bar{x}}},t)\in Q_{\frac{1}{2}}(0,0)\) follows along the same lines.
This step is slightly different from the analogous step in Section 2. Indeed, assumption (146) does not imply any rate of decay near the time \(t_0\). Therefore, we have to combine (146), to get the smallness with respect to \(\varepsilon _*\), with (143) or (144), to get the decay rate in time. To do so, one has to give up a bit of the power \(\eta \). Indeed, instead of \((s-t_0)_+^\eta \), the decay rate in (\(A_k'\)) is \((s-t_0)_+^\frac{\eta }{6}\). We have
The estimate for the pressure using (144) is similar.
We do not give the details for this step. Similar calculations are done below in Step 3. Notice that the terms \(I_4\) and \(I_5\) have to be estimated using (143) and (\(A_k'\)) for \(k=2\). The smallness of a given by (145) enables us to absorb some constants by choosing \(\varepsilon _*\) small enough.
Thanks to the local energy inequality (25), we have for all \(s\in (-r_n^2,0)\),
For \(I_1'\), we have
The term \(I_2'\) is immediate following the estimates of Section 2. Let us write some details for \(I_3'\). We have
Some changes are necessary in order to deal with \(I_4'\) and \(I_5'\). For \(I_4'\),
using the fact that \(M_a\leqq \varepsilon _*\) by assumption. Finally,
This concludes Step 3.
We first prove the estimate (\(A_k'\)). We have
which proves (\(A_k'\)) for \(k=n+1\) by choosing \(\varepsilon _*\) sufficiently small. Let us prove the estimate for the pressure using the bound of Lemma 24. We take \(r=r_{n+1}\) and \(\rho =\frac{1}{4}\). We have
for all \(s\in (-r_{n+1}^2,0)\). We concentrate on the estimates for \(J_2'\), \(J_3'\) and \(J_4'\). The estimate of \(J_1'\) is similar to the one just done above. The estimates of \(J_5'\) and \(J_6'\) do not pose any additional difficulty. For \(J_2'\), we have
For \(J_3'\), the estimate is very close to the bound for \(J_3\) in Section 2. We also split the integral into rings. This yields
Finally for \(J_4'\) splitting again into rings leads to
Hence the estimate (\(A_k''\)) follows for \(\varepsilon _*\) sufficiently small. \(\quad \square \)
Rights and permissions
About this article
Cite this article
Barker, T., Prange, C. Localized Smoothing for the Navier–Stokes Equations and Concentration of Critical Norms Near Singularities. Arch Rational Mech Anal 236, 1487–1541 (2020). https://doi.org/10.1007/s00205-020-01495-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00205-020-01495-6