Startseite Analyticity and Existence of the Keller–Segel–Navier–Stokes Equations in Critical Besov Spaces
Artikel Open Access

Analyticity and Existence of the Keller–Segel–Navier–Stokes Equations in Critical Besov Spaces

  • Minghua Yang ORCID logo , Zunwei Fu ORCID logo EMAIL logo und Suying Liu ORCID logo
Veröffentlicht/Copyright: 20. Januar 2018

Abstract

This paper deals with the Cauchy problem to the Keller–Segel model coupled with the incompressible 3-D Navier–Stokes equations. Based on so-called Gevrey regularity estimates, which are motivated by the works of Foias and Temam [20], we prove that the solutions are analytic for a small interval of time with values in a Gevrey class of functions. As a consequence of Gevrey estimates, we particularly imply higher-order derivatives of solutions in Besov and Lebesgue spaces. Moreover, we prove that the existence of a positive constant C~ such that the initial data (u0,n0,c0):=(u0h,u03,n0,c0) satisfy

C ~ ( ( n 0 , c 0 ) B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 ) + u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) α u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) 1 - α ) 1

for certain conditions on p,q and α implies the global existence of solutions with large initial vertical velocity component.

1 Introduction and Main Results

Consider the following model proposed in [40]:

(1.1) { u t + κ u u - μ Δ u + = - n ϕ , u = 0 in  ( 0 , T ) × Ω , n t + u n - ν Δ n = - ( χ ( c ) n c ) in  ( 0 , T ) × Ω , c t + u c - δ Δ c = - f ( c ) n in  ( 0 , T ) × Ω ,

where the functions c(t,x):(0,T)×Ω+, n(t,x):(0,T)×Ω+, u(t,x):(0,T)×Ω3 or 2, and (t,x):(0,T)×Ω denote the oxygen concentration, cell concentration, fluid velocity, and scalar pressure, respectively. Here, Ω3 or 2 is a spatial domain where the cells or bacteria and fluid move and interact. The constants ν,δ and μ are the corresponding diffusion coefficients for the cells, substrate and fluid, respectively. The nonnegative function f(c) denotes the oxygen consumption rate, and the nonnegative function χ(c) denotes chemotactic sensitivity. The parameter κ measures the strength of nonlinear fluid convection. The time-independent function ϕ=ϕ(x) denotes the potential function produced by different physical mechanisms, e.g., the gravitational force or centrifugal force.

When κ=0 and either c0 or ϕ is sufficiently small, Duan, Lorz and Markowich [19] constructed global existence of weak solutions to the Cauchy problem in 2. Liu and Lorz [30] removed the smallness assumption and obtained global existence of weak solutions for large data when κ=1 in 2. In bounded convex domains Ω2, Winkler [42] proved that the initial-boundary value problem of (1.1) possesses a unique global classical solution. Duan, Lorz and Markowich [19] proved the global-in-time existence of smooth solutions of (1.1) when the initial data are close to the constant equilibriumstates in H3(3). Global weak solutions of system (1.1) were constructed for κ=0 in [42], and for κ=1 in [44].

Lorz [32] studied local-in-time weak solutions in a bounded domain in d, d=2,3, with no-flux boundary condition, and in 2 with inhomogeneous Dirichlet conditions for oxygen. Chae, Kang and Lee [8] established the local regular existence and some blow-up criterions of solutions of (1.1) when the initial data (u0,n0,c0) is in Hm(d)×Hm-1(d)×Hm(d) with m3 and d=2,3. Moreover, Lorz [33] obtained global existence of a system of the elliptic-parabolic Keller–Segel equations coupled with Stokes equations (κ=0) with small initial data in d (d=2,3). In [9], Chae, Kang and Lee established the local existence of regular solutions for both cases that equations of oxygen concentration is of parabolic or hyperbolic type, and also proved global existence under the some smallness conditions of initial data. Zhang [50] obtained the existence and uniqueness of smooth solutions in inhomogeneous Besov spaces for (1.1) in d (d=2,3). Choe and Lkhagvasuren [14] and Minsuk, Bataa and Choe [15] investigated the local and global existence of (1.1) in critical homogeneous Besov spaces. In 2, Zhang and Zheng [51] showed that global weak solutions of (1.1) exist for a large class of initial data. If u=0 of system (1.1), Tao [39] showed that there exists a unique, global and bounded solution if χ is sufficiently small depending upon c0L. For stability and asymptotic behaviors on (1.1), we can refer to [9, 10, 19, 24, 43].

In particular, if the chemotaxis effects are ignored (n=c=0) for Ω=3, then (1.1) becomes the problems related to the classical Navier–Stokes equations

(1.2) { u t + u u - μ Δ u + = 0 in  ( 0 , T ) × 3 , u = 0 in  ( 0 , T ) × 3 , u | t = 0 = u 0 in  3 ,

which have been widely studied during the past eighty or more years. Leray [29] proved the global existence of weak solutions of (1.2), but the uniqueness and regularity of weak solutions remains a big open problem. It has been proved that the Cauchy problem (1.2) is globally well-posed for the small initial data in series of function spaces including particularly the following critical spaces:

H ˙ 1 2 ( 3 ) L 3 ( 3 ) B p , - 1 + 3 / p ( 3 ) BMO - 1 ( 3 ) ( 3 < p < ) ;

see Fujita and Kato [21], Kato [25], Kozono and Yamazaki [27] and Koch and Tataru [26]. Xiao [45, 46] proved this property in the space Qα-1(N).

Recently, the authors of [11, 12, 22, 23, 34, 35, 36, 49] proved the global well-posedness of (1.2) with initial data in some critical spaces and having a large vertical component provided that the horizontal component is small enough. For example, for any 0<ε<6p-1, 1<p<6 and some positive c~, given u0B˙p,1-1+3/p(3) with u0=0 and

u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) α u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) 1 - α c ~ μ ,

whereas

(1.3) α = { 1 p , 1 < p < 5 , ε , 5 p < 6 .

Zhai and Zhang [49] proved that there exists a global solution of (1.2) satisfying

u 𝔏 ( 0 , T ; B ˙ p , 1 - 1 + 3 / p ( 3 ) ) 𝔏 1 ( 0 , T ; B ˙ p , 1 1 + 3 / p ( 3 ) ) .

Biswas [4] introduced 𝕍θ,p and homogeneous Besov-type spaces 𝔹p,-δ, and then established Gevrey regularity of a class of dissipative equations with initial data in 𝔹p,-δ and 𝕍θ,p. Biswas and Swanson [6] studied the Gevrey regularity of Navier–Stokes equations with space-periodic boundary conditions in FB˙p,p-1+d-d/p(d). Bae [1] studied the Gevrey regularity of Lei–Lin [28] solutions of Navier–Stokes equations in FB˙1,1-1. Bae, Biswas and Tadmor [2] obtained the analyticity of the solutions of Navier–Stokes equations for the sufficiently small initial data in critical Besov spaces B˙p,q-1+d/p(d).

In this article, we shall consider χ(c)=κ=μ=ν=δ=1 and f(c)=c. In other words, the following problem for Keller–Segel-Navier–Stokes equations is considered:

(1.4) { u t + u u - Δ u + = - n ϕ , u = 0 in  ( 0 , T ) × 3 , n t + u n - Δ n = - ( n c ) in  ( 0 , T ) × 3 , c t + u c - Δ c = - c n in  ( 0 , T ) × 3 , ( u , n , c ) | t = 0 = ( u 0 , n 0 , c 0 ) in  3 .

We recall that system (1.4) enjoys nice scaling properties. If (u,n,c) solves (1.4), so does

( u λ , n λ , c λ ) := ( λ u ( λ 2 t , λ x ) , λ 2 n ( λ 2 t , λ x ) , c ( λ 2 t , λ x ) )

with initial data

( u λ ( 0 , x ) , n λ ( 0 , x ) , c λ ( 0 , x ) ) := ( λ u 0 ( λ x ) , λ 2 n 0 ( λ x ) , c 0 ( λ x ) ) .

We say that (𝔸,𝔹,) is a critical space for system (1.4) if the norm of (u0,n0,c0) in 𝔸×𝔹× is invariant for all λ>0. We observe that

B ˙ p , 1 - 1 + 3 / p ( 3 ) × B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 )

and B˙p,13/p(3) (for the definition of these spaces, see below) are critical spaces for (1.4) for the initial data (u0,n0,c0) and the gravitational potential ϕ, respectively.

For any operator T:B˙p,rs(3)B˙p,rs(3), we set

u T B ˙ p , r s ( 3 ) := T u B ˙ p , r s ( 3 ) .

Let Λ be the Fourier multiplier whose symbol is given by |ξ|1=i=13|ξi|, and let λ{0,1}. We emphasize that here ΛΛ1 is quantified by the 1 norm rather than the usual 2 norm associated with Λ2:=(-Δ)1/2. Let eλtΛ be a Fourier multiplier operator whose symbol is given by eλt|ξ|1. A function f is said to be Gevrey regular if

e t Λ f B ˙ p , r s ( 3 ) <

for some s and 1p,r. We mention that the finiteness of the corresponding Gevrey norm implies that the functions are analytic.

Let E0:=B˙p,1-1+3/p(3)×B˙q,1-2+3/q(3)×B˙q,13/q(3). We introduce a vector space ΘT:=𝕏T×𝕐T×T and with the usual product norm (u,n,c)ΘT:=u𝕏T+n𝕐T+cT,

𝕏 T λ := { u : u 𝔏 1 ( 0 , T ; e λ t Λ B ˙ p , 1 1 + 3 / p ( 3 ) ) 𝔏 ( 0 , T ; e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) ) } ,
𝕐 T λ := { n : n 𝔏 1 ( 0 , T ; e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) 𝔏 ( 0 , T ; e λ t Λ B ˙ q , 1 - 2 + 3 / q ( 3 ) ) } ,
T λ := { c : c 𝔏 1 ( 0 , T ; e λ t Λ B ˙ q , 1 2 + 3 / q ( 3 ) ) 𝔏 ( 0 , T ; e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) } ,

and

u 𝕏 T λ := u 𝔏 ( 0 , T ; e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + u 𝔏 1 ( 0 , T ; e λ t Λ B ˙ p , 1 1 + 3 / p ( 3 ) ) ,
n 𝕐 T λ := n 𝔏 ( 0 , T ; e λ t Λ B ˙ q , 1 - 2 + 3 / q ( 3 ) ) + n 𝔏 1 ( 0 , T ; e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) ,
c T λ := c 𝔏 ( 0 , T ; e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) + c 𝔏 1 ( 0 , T ; e λ t Λ B ˙ q , 1 2 + 3 / q ( 3 ) ) .

Meanwhile, let

Θ T 𝒞 = 𝒞 ( [ 0 , T ] ; B ˙ p , 1 - 1 + 3 / p ( 3 ) ) 𝒞 ( [ 0 , T ] ; B ˙ q , 1 - 2 + 3 / q ( 3 ) ) 𝒞 ( [ 0 , T ] ; B ˙ q , 1 3 / q ( 3 ) ) .

For simplicity, when T=, λ=1, we denote 𝕏1:=𝕏, 𝕐1:=𝕐, 1:=, Θ1:=Θ and Θ𝒞:=Θ𝒞.

Inspired by the seminal work of Foias and Temam in [20], and also motivated by the papers [5, 1, 2, 4, 6] concerning the Gevrey regularity of solutions, we obtain the Gevrey class regularity for system (1.4) in the context of critical Besov spaces.

Theorem 1.1.

Let 1<p,q<, 1p+1q>13, 1<q<6, 1min{p,q}-1max{p,q}13, u0=0 and (u0,n0,c0)E0. Furthermore, there exist ϵ>0 and

e λ ϵ Λ ϕ B ˙ p , 1 3 / p ( 3 )

such that system (1.4) admits a unique local solution such that (u,n,c)ΘTλΘTC. Here eλϵΛ is a Fourier multiplier whose symbol is given by eλϵ|ξ|1.

The solution (u,n,c) in Theorem 1.1 in turn enables us to establish the following estimates on the high-order derivatives in Besov and Lebesgue spaces.

Corollary 1.2.

For 0<t<T, there exist positive constants C1,C2,C3 and C~ such that the following statements hold:

  1. If m > 0 , then

    ( D m u , D m n , D m c ) B ˙ p , 1 - 1 + 3 / p ( 3 ) × B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 ) C ~ m m m t - m 2 .

  2. If k 1 > - 1 + 3 p and 1 < p 2 , then

    D k 1 u ( t ) L p ( 3 ) C 1 k 1 + 1 - 3 / p ( k 1 + 1 - 3 p ) k 1 + 1 - 3 / p t - k 1 + 1 - 3 / p 2 .

  3. If k 2 > - 2 + 3 q and 1 < q 3 2 , then

    D k 2 n ( t ) L q ( 3 ) C 2 k 2 + 2 - 3 / q ( k 2 + 2 - 3 q ) k 2 + 2 - 3 / q t - k 2 + 2 - 3 / q 2 .

  4. If k 3 > 3 q and 1 < q 2 , then

    D k 3 c ( t ) L q ( 3 ) C 3 k 3 - 3 / q ( k 3 - 3 q ) k 3 - 3 / q t - k 3 - 3 / q 2 .

It is worth mentioning that for any t(0,T) in Theorem 1.1 we obtain the solution

( u ( t ) , n ( t ) , c ( t ) ) C ( 3 ) × C ( 3 ) × C ( 3 ) .

Motivated by the paper [49] concerning the global well-posedness of 3D-incompressible Navier–Stokes equations with the third component of the initial velocity field being large, applying the local existence mentioned in Theorem 1.1 (in this case λ=0) as well as the standard continuity argument, we prove that system (1.4) has a unique global solution. Our main result reads as follows.

Theorem 1.3.

Let 1<pq<6,0<ε<6p-1 and 1p-1q13. Let (u0,n0,c0)E0 with u0=0, and let u0=(u01,u02,u03)=(u0h,u03). If there exists a positive constant C~ such that the initial data (u0,n0,c0) and potential function

ϕ B ˙ p , 1 3 / p ( 3 )

satisfy

(1.5) C ~ ( ( n 0 , c 0 ) B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 ) + u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) α u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) 1 - α ) 1 ,

where

α = { 1 p , 1 < p < 5 , ε , 5 p < 6 ,

then system (1.4) admits a unique global-in-time solution such that (u,n,c)ΘΘC.

We mention that our results do not have any smallness conditions imposed on the third component u03 of the initial velocity field. We also can remove the smallness condition on ϕ by using a special type of iteration scheme (see [14]). The papers [14, 49] play an essential role in the proof of Theorem 1.3.

Notation.

Throughout the paper, c~ and C~ stand for harmless constants. Let A and B be two operators; we denote [A;B]=AB-BA. For a Banach space X and an interval I of , we denote by 𝒞(I;X) the set of continuous functions on I with value in X. Furthermore, (dj)j will be a generic element of 1() so that dj0 and jdj=1.

2 Preliminaries

Let 𝒮(3) be the Schwartz class of rapidly decreasing functions and let 𝒮(3) be the space of tempered distributions. Here and -1 denote Fourier and inverse Fourier transforms of L1(3) functions, respectively, defined by

f = f ^ ( ξ ) = ( 2 π ) - 3 2 3 e - i x ξ f ( x ) 𝑑 x and - 1 f = f ˇ ( x ) = ( 2 π ) - 3 2 3 e i x ξ f ^ ( ξ ) 𝑑 ξ .

More generally, the Fourier transform of any f is in 𝒮(3), given by (f,g)=(f,g) for any g𝒮(3). Let 𝒞 be the annulus {ξ3:34|ξ|83} and let 𝒟(Ω) be a space of smooth compactly supported functions on the domain Ω. There exist radial functions χ and φ, valued in the interval [0,1], belonging respectively to 𝒟(B(0,43)) and 𝒟(𝒞), such that for all ξ3,

χ ( ξ ) + j 0 φ ( 2 - j ξ ) = 1 ,

for all ξ3{0},

j φ ( 2 - j ξ ) = 1 ,

and

| j - j | 2 implies supp φ ( 2 - j ) supp φ ( 2 - j ) = ,
j 1 implies supp χ ( ) supp φ ( 2 - j ) = .

Define the set 𝒞~=B(0,23)+𝒞. Then we have

| j - j | 5 implies 2 j 𝒞 ~ 2 j 𝒞 = .

Furthermore, for all ξ3 we have

1 2 χ 2 ( ξ ) + j 0 φ 2 ( 2 - j ξ ) 1 ,

and for all ξ3{0} we have

1 2 j φ 2 ( 2 - j ξ ) 1 .

From now on, we write h=-1φ and h~=-1χ. The homogeneous dyadic blocks Δj and Sj are defined by

Δ j u := φ ( 2 - j D ) u := 2 3 j 3 h ( 2 j y ) u ( x - y ) 𝑑 y if  j ,
S j u := χ ( 2 - j D ) u := 2 3 j 3 h ~ ( 2 j y ) u ( x - y ) 𝑑 y if  j .

Here D:=(D1,D2,D3) and Dj:=i-1xj(i2=-1). The set {Δj,Sj}j is called the Littlewood–Paley decomposition. Formally, Δj=Sj-Sj-1 is a frequency projection to the annulus {ξ3:2j-1<|ξ|2j}, and

S j = j j - 1 Δ j

is a frequency projection to the ball {ξ3:|ξ|2j}. We denote by 𝒮h(3) the space of tempered distributions f such that limj-Sjf=0 in 𝒮(3). Recall that for any s and (p,r)[1,]×[1,], the homogeneous Besov spaces B˙p,rs(3) are defined by

B ˙ p , r s ( 3 ) := { f 𝒮 h ( 3 ) : f B ˙ p , r s ( 3 ) < } ,

where

f B ˙ p , r s ( 3 ) := { [ k { 2 k s Δ k f L p ( 3 ) } r ] 1 r if  1 p ,  1 r < , s , sup k [ 2 k s Δ k f L p ( 3 ) ] if  1 p , r = , s .

It is well known that if either s<3/p or s=3/p and r=1, then B˙p,rs(3) is a Banach space.

Let us now recall the definition of the Chemin–Lerner space 𝔏ρ(0,T;B˙p,rs(3)) with 0<T, s and 1p,r,ρ (with the usual convention if r= or ρ=). The Chemin–Lerner space is defined by

𝔏 ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) := { f 𝒮 ( ( 0 , T ) , 𝒮 h ( 3 ) ) : f 𝔏 ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) < }

and

f 𝔏 ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) := f 𝔏 T ρ ( B ˙ p , r s ( 3 ) ) := { k [ 2 k s Δ k f L T ρ ( L p ( 3 ) ) ] r } 1 r ,

where

f L T ρ ( L p ) := [ 0 T f L p ( 3 ) ρ 𝑑 τ ] 1 ρ ,

and the usual modifications are needed when r= or ρ=. We equip the space

L ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) := { f 𝒮 ( ( 0 , T ) , 𝒮 h ( 3 ) ) : f L ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) < }

with the norm

f L ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) := f L T ρ ( B ˙ p , r s ( 3 ) ) := = { 0 T [ k { 2 k s Δ k f L p ( 3 ) } r ] ρ r d t } 1 ρ .

From Minkowski’s inequality, it is easy to deduce that

f 𝔏 ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) f L ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) when  ρ r ,
f L ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) f 𝔏 ρ ( 0 , T ; B ˙ p , r s ( 3 ) ) when  r ρ .

If s1 and s2 are real numbers so that s1<s2 and θ(0,1), then we have, for any (p,r,ρ,ρ1,ρ2)[1,]5 and 1ρ=θρ1+1-θρ2,

(2.1) { f B ˙ p , r θ s 1 + ( 1 - θ ) s 2 ( 3 ) C f B ˙ p , r s 1 ( 3 ) θ f B ˙ p , r s 2 ( 3 ) 1 - θ , f L ρ ( 0 , T ; B ˙ p , r θ s 1 + ( 1 - θ ) s 2 ( 3 ) ) C f L ρ 1 ( 0 , T ; B ˙ p , r s 1 ( 3 ) ) θ f L ρ 2 ( 0 , T ; B ˙ p , r s 2 ( 3 ) ) 1 - θ , f 𝔏 ρ ( 0 , T ; B ˙ p , r θ s 1 + ( 1 - θ ) s 2 ( 3 ) ) C f 𝔏 ρ 1 ( 0 , T ; B ˙ p , r s 1 ( 3 ) ) θ f 𝔏 ρ 2 ( 0 , T ; B ˙ p , r s 2 ( 3 ) ) 1 - θ ,

where C is a positive constant independent of f.

Because B˙p,13/p(3) is embedded in L(3), we deduce that whenever 1p, the product of two functions in B˙p,13/p(3) is also in B˙p,13/p(3) and such that for some constant C>0,

(2.2) u v B ˙ p , 1 3 / p ( 3 ) C u B ˙ p , 1 3 / p ( 3 ) v B ˙ p , 1 3 / p ( 3 ) .

The homogeneous paraproduct of v and u is defined by

T u v := j S j - 1 u Δ j v .

The homogeneous remainder of v and u is defined by

R ( u , v ) := | k - j | 1 Δ k u Δ j v := k Δ k u Δ ~ k v and Δ ~ k := Δ k - 1 + Δ k + Δ k + 1 .

We have the following Bony decomposition:

(2.3) u v := T u v + R ( u , v ) + T v u .

Lemma 2.1 ([3, 41]).

Let 0<r<R, let C:={ξR3:r|ξ|R} be an annulus and let

:= ( 0 , R ) = { ξ 3 : 0 | ξ | R }

be a ball. A positive constant C~ exists such that for any nonnegative integer k, any couple (p,q) in [1,]2 with qp1, and any function u of Lp(R3), we have

supp u ^ λ 𝑖𝑚𝑝𝑙𝑖𝑒𝑠 D k u L q ( 3 ) := sup | α | k α u L q ( 3 ) C ~ k + 1 λ k + 3 ( 1 p - 1 q ) u L p ( 3 ) ,
supp u ^ λ 𝒞 𝑖𝑚𝑝𝑙𝑖𝑒𝑠 C ~ - k - 1 λ k u L p ( 3 ) D k u L p ( 3 ) C ~ k + 1 λ k u L p ( 3 ) .

Lemma 2.2.

Assume that θ{0,1}. Let

B t θ ( 𝔣 , 𝔤 ) := e θ t Λ ( e - θ t Λ 𝔣 e - θ t Λ 𝔤 ) = 3 3 e i x ξ e θ t ( | ξ | 1 - | ξ - η | 1 - | η | 1 ) 𝔣 ^ ( ξ - η ) 𝔤 ^ ( η ) 𝑑 η 𝑑 ξ .

Then we have

(2.4) B t θ ( 𝔣 , 𝔤 ) L 𝔭 ( 3 ) 𝔣 L p 1 ( 3 ) 𝔤 L p 2 ( 3 ) , 1 p 1 + 1 p 2 = 1 𝔭 ,  1 < 𝔭 < .

Proof.

When θ=0, the desired result (2.4) is obvious. For the proof of θ=1, we refer to [2, pp. 975, 976]. ∎

Lemma 2.3.

Let sR and 1ρ,p,r. Assume that u0B˙p,rs(R3),

f 𝔏 ρ ( 0 , T ; e λ t Λ B ˙ p , r s - 2 + 2 / ρ ( 3 ) ) ,

and u solves

{ u t - μ Δ u = f in  3 × ( 0 , T ) , u | t = 0 = u 0 in  3 .

Denote ρ2=(1+1ρ1-1ρ)-1. Then there exist two positive constants c~ and C~ such that for all ρ1[ρ,+] we have

u 𝔏 ρ 1 ( 0 , T ; e λ t Λ B ˙ p , 1 s + 2 / ρ 1 ( 3 ) )
C ~ ( q 2 q s Δ q u 0 L p ( 1 - e - c ~ μ T 2 2 q ρ 1 c ~ μ ρ 1 2 2 q ) 1 ρ 1 + q 2 q ( s - 2 + 2 ρ ) e λ t Λ Δ q f L T ρ ( L p ) ( 1 - e - c ~ μ T 2 2 q ρ 2 c ~ μ ρ 2 2 2 q ) 1 ρ 2 ) .

In particular, there holds

μ 1 ρ 1 u 𝔏 ρ 1 ( 0 , T ; e λ t Λ B ˙ p , 1 s + 2 / ρ 1 ( 3 ) ) C ~ ( u 0 B ˙ p , 1 s ( 3 ) + μ 1 ρ - 1 f 𝔏 ρ ( 0 , T ; e λ t Λ B ˙ p , 1 s - 2 + 2 / ρ ( 3 ) ) ) .

Proof.

We can refer the reader to [18, Section 2.2] for the proof of Lemma 2.3.∎

Lemma 2.4.

Let 1<p,q<,1p+1q>13. If uXT and nYT, then there exists a positive constant C such that

u n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 - 2 + 3 / q )
𝒞 u 𝔏 T ( e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) ) n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) + 𝒞 u 𝔏 T 1 ( e λ t Λ B ˙ p , 1 3 / p + 1 ( 3 ) ) n 𝔏 T ( e λ t Λ B ˙ q , 1 - 2 + 3 / q ( 3 ) ) .

Proof.

We refer the reader to [48, 38] for the proof of Lemma 2.4. For the convenience of the reader, we prove Lemma 2.4 here. Let (N,U):=(eλtΛn,eλtΛu). Thanks to Bony’s paraproduct decomposition (2.3), we have

e λ t Λ Δ j ( u n ) = e λ t Λ Δ j ( T e - λ t Λ U e - λ t Λ N + R ( e - λ t Λ U , e - λ t Λ N ) + T e - λ t Λ N e - λ t Λ U ) .

Using (2.1) and Lemmas 2.1 and 2.2, we have

Δ j e λ t Λ ( T e - λ t Λ U e - λ t Λ N ) L T 1 ( L q ) | j - j | 4 2 j k j - 2 2 3 k / p Δ k U L T ( L p ) Δ j N L T 1 ( L q )
| j - j | 4 2 j ( 3 - 3 / q ) d j k j - 2 2 k 2 k ( - 1 + 3 / p ) N 𝔏 T 1 ( B ˙ q , 1 3 / q ) Δ k U L T ( L p )
2 j ( 2 - 3 / q ) d j U 𝔏 T ( B ˙ p , 1 - 1 + 3 / p ) N 𝔏 T 1 ( B ˙ q , 1 3 / q ) .

Thanks to 1<p<q<, by (2.1) and Lemmas 2.1 and 2.2, we arrive at

Δ j e λ t Λ ( T e - λ t Λ N e - λ t Λ U ) L T 1 ( L q )
2 3 j ( 1 / p - 1 / q ) | j - j | 4 Δ j U L T 1 ( L p ) S j - 1 N L T ( L )
2 3 j ( 1 / p - 1 / q ) | j - j | 4 2 j ( - 3 / p - 1 ) d j U 𝔏 T 1 ( B p , 1 3 / p + 1 ) k j - 2 2 3 k 2 k ( - 2 + 3 / q ) Δ k N L T ( L q )
2 j ( 2 - 3 / q ) d j N 𝔏 T ( B ˙ q , 1 - 2 + 3 / q ) U 𝔏 T 1 ( B ˙ p , 1 1 + 3 / p ) .

If 1qp<, then there exists 1<λ such that 1q=1p+1λ. Using (2.1) and Lemmas 2.1 and 2.2, we get that

Δ j e λ t Λ ( T e - λ t Λ N e - λ t Λ U ) L T 1 ( L q ) 2 3 j ( 1 / p - 1 / q )
| j - j | 4 k j - 2 2 k + 3 k ( 1 / q - 1 / λ ) Δ k N L T ( L q ) Δ j U L T 1 ( L p )
| j - j | 4 2 j ( - 3 / p - 1 ) d j U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ) k j - 2 2 k ( 3 - 3 / λ ) 2 k ( - 2 + 3 / q ) Δ k n L T ( L q )
2 j ( 2 - 3 / q ) d j U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ( 3 ) ) N 𝔏 T ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) .

In the case 1p+1q>1, we can find 1<q such that 1q+1q=1. By (2.1) and Lemmas 2.1 and 2.2, we infer that for some fixed constant N0,

Δ j e λ t Λ R ( e - λ t Λ U , e - λ t Λ N ) L T 1 ( L q ) 2 3 j ( 1 - 1 / q ) j j - N 0 2 j 2 j ( 3 / p - 3 / q ) Δ ~ j N L T ( L q ) Δ j U L T 1 ( L p )
2 j ( 1 - 3 / p ) d j N 𝔏 T ( B ˙ q , 1 - 2 + 3 / q ) U 𝔏 T 1 ( B ˙ p , 1 1 + 3 / p ) .

If 1p+1q1, together with (2.1), Lemmas 2.1 and 2.2 and 1q+1p>13 ensure that

Δ j e λ t Λ R ( e - λ t Λ U , e - λ t Λ N ) L T 1 ( L q ) 2 j ( 1 + 3 / p ) j j - N 0 Δ j U L T 1 ( L p ) Δ ~ j N L T ( L q )
2 j ( 1 + 3 / p ) j j - N 0 2 j ( 1 - 3 / p - 3 / q ) d j U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ) 2 j ( - 2 + 3 / q ) Δ ~ j N L T ( L q )
2 j ( 2 - 3 / q ) d j U 𝔏 T 1 ( B ˙ p , 1 1 + 3 / p ) N 𝔏 T ( B ˙ q , 1 - 2 + 3 / q ) .

Multiplying both sides of the above inequality with 2-2+3j/q and then taking the 1()-norm, we obtain the desired result. ∎

Lemma 2.5.

If nYT and cZT, then there exists a positive constant C such that

( n c ) 𝔏 T 1 ( e λ t Λ B ˙ q , 1 - 2 + 3 / q ( 3 ) )
𝒞 n 𝔏 T ( e λ t Λ B ˙ q , 1 - 2 + 3 / q ( 3 ) ) c 𝔏 T 1 ( e λ t Λ B ˙ q , 1 2 + 3 / q ( 3 ) ) + 𝒞 n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 T ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) .

Proof.

Applying [47, Lemma 2.6] and choosing p=q, f=n and ϕ=c, we can obtain the desired result immediately. The proof of Lemma 2.5 is thus completed. ∎

Lemma 2.6.

Let 1<qp<, 1q-1p13 (or 1<p<q<). If uXT and cZT, then there exists a positive constant C such that

u c 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) )
𝒞 u 𝔏 T ( e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) ) c 𝔏 T 1 ( e λ t Λ B ˙ q , 1 2 + 3 / q ( 3 ) ) + 𝒞 u 𝔏 T 1 ( e λ t Λ B ˙ p , 1 1 + 3 / p ( 3 ) ) c 𝔏 T ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) .

Proof.

This can be obtained by exploiting the same argument used in [31]. Let (C,U):=(eλtΛc,eλtΛu). The Bony paraproduct decomposition (2.3) ensures that

e λ t Λ Δ j ( u c ) = e λ t Λ Δ j ( T e - λ t Λ U e - λ t Λ C + R ( e - λ t Λ U , e - λ t Λ N ) + T e - λ t Λ N e - λ t Λ U ) .

If 1qp< and 1q-1p13, then there exists 1<λ such that 1q=1p+1λ; note that 1+3p-3q0. These, in combination with Lemmas 2.1 and 2.2 and (2.1), show that

Δ j e λ t Λ ( T e - λ t Λ C e - λ t Λ U ) L T 1 ( L q ) | j - j | 4 Δ j U L t 1 ( L p ) k j - 2 2 ( 1 + 3 p - 3 / q ) k Δ k C L T ( L q ) 2 3 k / q
| j - j | 4 d j 2 j ( - 3 / q ) U 𝔏 T 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) C 𝔏 T ( B ˙ q , 1 3 / q ( 3 ) )
d j 2 j ( - 3 / q ) U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ( 3 ) ) C 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) .

If 1p<q<, then by Lemmas 2.1 and 2.2 and (2.1) we obtain

Δ j e λ t Λ ( T e - λ t Λ C e - λ t Λ U ) L T 1 ( L q ) 2 3 j ( 1 / p - 1 / q ) | j - j | 4 Δ j U L T 1 ( L p ) S j - 1 C L T ( L )
2 3 j ( 1 / p - 1 / q ) | j - j | 4 2 j ( 3 / p - 1 ) d j U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ( 3 ) ) k j - 2 2 k 2 3 k / q Δ k C L T ( L q )
2 j ( - 3 / q ) d j C 𝔏 T ( B ˙ q , 1 3 / q ( 3 ) ) U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ( 3 ) ) .

To estimate the remaining term R(e-λtΛU,e-λtΛC), we consider the following two cases: In the case 1p+1q>1, we can find 1<q such that 1q+1q=1 by Lemmas 2.1 and 2.2 and (2.1). For some fixed constant N0, we calculate that

Δ j e λ t Λ R ( e - λ t Λ U , e - λ t Λ C ) L T 1 ( L q ) 2 3 j ( 1 - 1 / q ) j j - N 0 2 j Δ ~ j C L T ( L q ) 2 3 k ( 1 / p - 1 / q ) Δ j U L T 1 ( L p )
2 3 j ( 1 - 1 / q ) j j - N 0 d j 2 - 3 j U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ( 3 ) ) C L T ( B ˙ q , 1 3 / q ( 3 ) ) .

Since 1p+1q1, with the help of Lemmas 2.1 and 2.2 and (2.1) we obtain

Δ j e λ t Λ R ( e - λ t Λ U , e - λ t Λ C ) L T 1 ( L q )
2 ( 3 / p + 1 ) j j j - N 0 2 j ( - 1 - 3 / p - 3 / q ) 2 j ( 3 / p + 1 ) Δ j U L T 1 ( L p ) 2 3 j / q Δ ~ j C L T ( L q )
2 - 3 j / q d j U 𝔏 T 1 ( B ˙ p , 1 3 / p + 1 ( 3 ) ) C 𝔏 T ( B ˙ q , 1 3 / q ( 3 ) ) .

By combining all above estimates, multiplying both sides of the above inequality with 23j/q and then taking the 1()-norm, the proof of Lemma 2.6 is finished. ∎

Lemma 2.7.

Let 1<p1p2 and s13p1, s23p2 with s1+s2>3max(0,1p1+1p2-1). If

e λ t Λ a B ˙ p 1 , 1 s 1 ( 3 ) 𝑎𝑛𝑑 e λ t Λ b B ˙ p 2 , 1 s 2 ( 3 ) ,

then we have

a b e λ t Λ B ˙ p 2 , 1 s 1 + s 2 - 3 / p 1 ( 3 ) a e λ t Λ B ˙ p 1 , 1 s 1 ( 3 ) a e λ t Λ B ˙ p 2 , 1 s 2 ( 3 ) .

Proof.

Applying Lemma 2.2 and simple calculations as in proving [36, Lemma 2.2] enable us to conclude the required result, and we omit the details. ∎

Lemma 2.8.

Let 1<p¯q¯ and s3q¯ with s+3q¯>3max(0,1q¯+1p¯-1). If a¯B˙q¯,13/q¯(R3) and b¯B˙p¯,1s(R3), then a¯b¯B˙p¯,1s(R3), and there holds

a ¯ b ¯ e λ t Λ B ˙ p ¯ , 1 s ( 3 ) a ¯ e λ t Λ B ˙ q ¯ , 1 3 / q ¯ ( 3 ) b ¯ e λ t Λ B ˙ p ¯ , 1 s ( 3 ) .

Proof.

By Lemma 2.2, the proof of Lemma 2.8 is a slight modification of [49, Lemma 2.5]. ∎

Lemma 2.9.

Let p>1. If uXT and u=0, then we obtain

u u 𝔏 T 1 ( e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) ) u 𝕏 T u 𝕏 T .

Proof.

Arguing by interpolation (2.1), the algebra property (2.2) and Lemma 2.2, the proof of Lemma 2.9 consists of an argument similar to the ones in the proofs of Lemmas 2.42.6. ∎

Lemma 2.10.

Let 1<q<6. Assume that nYT and cZT. Applying the algebra property (2.2) of Besov spaces yields

c n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) )
n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 T ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) + n 𝔏 T ( e λ t Λ B ˙ q , 1 - 2 + 3 / q ( 3 ) ) c 𝔏 T 1 ( e λ t Λ B ˙ q , 1 2 + 3 / q ( 3 ) ) .

Proof.

The proof resembles the ones of Lemmas 2.42.6. ∎

Lemma 2.11.

Let 1<q<p<, 1q+1p>13 (or 1<pq<, 1q+1p>13, 1p-1q13). If nYT and

ϕ L T ( e λ t Λ B ˙ p , 1 3 / p ( 3 ) ) ,

then we have

n ϕ 𝔏 T 1 ( e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) ) n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) ϕ L T ( e λ t Λ B ˙ p , 1 - 1 + 3 / p ( 3 ) )
(2.5) n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) ϕ L T ( e λ t Λ B ˙ p , 1 3 / p ( 3 ) ) .

Proof.

By Lemmas 2.7 and 2.8, one can conclude (2.5) immediately. ∎

3 Proof of Theorem 1.1

To prove the local Gevrey regularity of Theorem 1.1, note that, applying Lemma 2.11 for any given ϵ>0, we can take T small enough such that T-ϵ<0, and then we have

n ϕ 𝔏 T 1 ( e λ t Λ B ˙ p , 1 - 1 + 3 / p ) C n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) e λ t Λ ϕ L T ( B ˙ p , 1 3 / p ( 3 ) )
= C n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) e λ ( t - ϵ ) Λ e λ ϵ Λ ϕ L T ( B ˙ p , 1 3 / p ( 3 ) )
C n 𝔏 T 1 ( e λ t Λ B ˙ q , 1 3 / q ( 3 ) ) e λ ϵ Λ ϕ B ˙ p , 1 3 / p ( 3 ) .

Borrowing ideas from [14], we can easily prove Theorem 1.1 by a standard appropriate iteration scheme. We omit the details here.

4 Proof of Corollary 1.2: Decay of Besov Norms and Lebesgue Norms

Theorem 1.1 tells us that the solution is locally in the Gevrey regular, that is, the energy bound (u,n,c)Θ< for λ=1. Specifically, we can show that the solution (u,n,c) satisfies

(4.1) sup t > 0 ( e t Λ u , e t Λ n , e t Λ c ) B ˙ p , 1 - 1 + 3 / p ( 3 ) × B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 ) < + .

As we have showed that the estimate (4.1) holds for the solution (u,n,c) of system (1.4) in Gevrey classes, by (1.3), (4.1) and the same argument as for [5, Corollary 3.3], for m>0 there exists a positive constant C~ such that

( D m u , D m n , D m c ) B ˙ p , 1 - 1 + 3 / p ( 3 ) × B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 )
= ( D m e - t Λ e t Λ u , D m e - t Λ e t Λ n , D m e - t Λ e t Λ c ) B ˙ p , 1 - 1 + 3 / p ( 3 ) × B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 )
C ~ m m m t - m 2 ( e t Λ u , e t Λ v , e t Λ w ) L t B ˙ p , 1 - 1 + 3 / p ( 3 ) × L t B ˙ q , 1 - 2 + 3 / q ( 3 ) × L t B ˙ q , 1 3 / q ( 3 )
C ~ m m m t - m 2 ( e t Λ u , e t Λ v , e t Λ w ) 𝔏 t B p , 1 - 1 + 3 / p × 𝔏 t B ˙ q , 1 - 2 + 3 / q ( 3 ) × 𝔏 t B ˙ q , 1 3 / q ( 3 )
(4.2) C ~ m m m t - m 2 .

By using the relation between homogeneous Besov spaces and homogeneous Triebel–Lizorkin spaces F˙p0,p0s(3) (for the definition of Triebel–Lizorkin spaces, see [7, 13, 16]), p02 for 1p02, and F˙p0,2s(3)=W˙s,p0(3):=(-Δ)-s/2Lp0, one thus has

(4.3) B ˙ p 0 , 1 s ( 3 ) B ˙ p 0 , p 0 s ( 3 ) F ˙ p 0 , p 0 s ( 3 ) F ˙ p 0 , 2 s ( 3 ) = W ˙ s , p 0 ( 3 ) .

By applying (4.3) and (4.1) along the same lines as (4.2), for k1>-1+3p and 1<p2 there exists a positive constant C1 so that

D k 1 u L p ( 3 ) = D k 1 + 1 - 3 / p e - t Λ D - 1 + 3 / p e t Λ u L p ( 3 )
C 1 k 1 + 1 - 3 / p ( k 1 + 1 - 3 p ) k 1 + 1 - 3 / p t - ( k 1 + 1 - 3 / p ) / 2 D - 1 + 3 / p e t Λ u L p ( 3 )
= C 1 k 1 + 1 - 3 / p ( k 1 + 1 - 3 p ) k 1 + 1 - 3 / p t - ( k 1 + 1 - 3 / p ) / 2 e t Λ u F ˙ p , 2 - 1 + 3 / p ( 3 )
C t - ( k 1 + 1 - 3 / p ) / 2 e t Λ u B ˙ p , 1 - 1 + 3 / p ( 3 )
(4.4) C 1 k 1 + 1 - 3 / p ( k 1 + 1 - 3 p ) k 1 + 1 - 3 / p t - ( k 1 + 1 - 3 / p ) / 2 .

By an argument similar to the one in (4.4), there exist positive constants C2 and C3 such that

D k 2 n L q ( 3 ) C 2 k 2 + 2 - 3 / q ( k 2 + 2 - 3 q ) k 2 + 2 - 3 / q t - ( k 2 + 2 - 3 / q ) / 2

for k2>-2+3q and 1<q32. Also, when k3>3q and 1<q2, we have

D k 3 c L q ( 3 ) C 3 k 3 - 3 / q ( k 3 - 3 q ) k 3 - 3 / q t - ( k 3 - 3 / q ) / 2 .

This completes the proof of Corollary 1.2.

5 Proof of Theorem 1.3: Global Existence with Large Vertical Velocity Component

The goal of this section is to present the proof of the global existence of system (1.4) with large vertical velocity component. Let u:=(u1,u2,u3)=(uh,u3) and divhuh=1u1+2u2.

Lemma 5.1 ([49]).

Let x:=(x1,x2,x3),1<pm,r and

u = ( u h , u 3 ) 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) )

with divu=0. There hold

Δ j u 3 L t 2 ( L h m ( L v r ) ) d j 2 - j ( 2 m + 1 r ) u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - 1 p + 1 r u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 p - 1 r

and

Δ j u 3 L t 1 ( L h m ( L v r ) ) d j 2 - j ( 2 m + 1 r + 1 ) u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 p + 1 r u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 p - 1 r ,

where

f L t ρ ( L h m ( L v r ) ) := [ 0 t ( | 2 | f | r 𝑑 x 1 𝑑 x 2 | m r 𝑑 x 3 ) ρ m 𝑑 τ ] .

Lemma 5.2 ([49]).

Let p>1; there holds

Δ j ( u 3 u h ) L t 1 ( L p ( 3 ) )
d j 2 - 3 j / p ( u h 𝔏 t ( B ˙ p , 1 1 + 3 / p ( 3 ) ) u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ) 1 / p u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p + u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + 1 / p u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - 1 / p ) .

Lemma 5.3 ([49]).

Let 1<p<6; there holds

Δ j ( u 3 div h u h ) L t 1 ( L p ( 3 ) )
d j 2 j ( 1 - 3 / p ) ( u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + 1 / p u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - 1 / p + u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + α u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - α ) ,

where

α = { 1 p , 1 < p < 5 , ε , 5 p < 6 ,

for 0<ε<6p-1.

The estimate of the pressure .

Taking div to the first equations of system (1.4) yields that

- Δ = div h div h ( u h u h ) + 2 3 div h ( u 3 u h ) + 3 2 ( u 3 ) 2 + div ( n ϕ ) .

By virtue of u=0, it is clear that

(5.1) = ( - Δ ) - 1 [ div h div h ( u h u h ) + 2 3 div h ( u 3 u h ) - 2 3 ( u 3 div h u h ) + div ( n ϕ ) ] .

Proposition 5.4.

Let 1<pq<6,1p-1q13. If (u,n,c)ΘT with u=0, then (5.1) has a unique solution PLT1(B˙p,1-1+3/p(R3)) such that for t[0,T] there holds

(5.2) 𝔏 t 1 ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) C 𝕎 ( u h , u 3 ) + C ~ n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 ) ,

where

α = { 1 p , 1 < p < 5 , ε , 5 p < 6 ,

for 0<ε<6p-1.

Proof.

Applying the operator Δj to (5.1), taking the Lt1(Lp(3)) norm, and using Lemmas 2.9, 2.11, 5.2 and 5.3 yields that

Δ j ( ) L t 1 ( L p ( 3 ) )
2 j Δ j ( u h u h ) L t 1 ( L p ( 3 ) ) + 2 j Δ j ( u 3 u h ) L t 1 ( L p ( 3 ) ) + Δ j ( u 3 div h u h ) L t 1 ( L p ( 3 ) ) + Δ j ( n ϕ ) L t 1 ( L p ( 3 ) )
C 2 j ( 1 - 3 / p ) d j 𝕎 ( u h , u 3 ) + C ~ d j 2 j ( 1 - 3 / p ) n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 ) ,

where

𝕎 ( u h , u 3 ) := u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p
+ u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + 1 / p + u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + α u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - α ,

from which the desired estimate (5.2) follows readily. Finally, according to the definition of Chemin–Lerner norm, we complete the proof. ∎

The estimate of uh and u3.

The first equation of (1.4) implies that

(5.3) t u h - Δ u h = - u u h - h - n h ϕ

and

(5.4) t u 3 - Δ u 3 = - u u 3 - 3 - n 3 ϕ .

Proposition 5.5.

Let 1<pq<6 and 1p-1q13. If (u,n,c)ΘT with u=0, then there hold

u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) )
(5.5) u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + C ~ 𝕎 ( u h , u 3 ) + C ~ n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )

and

u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) )
(5.6) u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) + C ~ u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) + C ~ 𝕎 ( u h , u 3 ) + C ~ n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 ) .

Proof.

We first prove (5.5). By Applying the operator Δj to (5.3) and taking the L2 inner product with |Δjuh|p-2Δjuh (in the case when p(1,2), we need to make some modifications as in [17]), an integration by parts yields

(5.7) 1 p d d t Δ j u h L p ( 3 ) p - 3 Δ Δ j u h | Δ j u h | p - 2 Δ j u h 𝑑 x = - 3 Δ j ( u u h + h + n h ϕ ) | Δ j u h | p - 2 Δ j u h 𝑑 x .

Thanks to [17, 37], there exists a positive constant c¯p fulfilling

(5.8) - 3 Δ Δ j u h | Δ j u h | p - 2 Δ j u h 𝑑 x c ¯ p 2 2 j Δ j u h L p ( 3 ) p .

Similarly proceeding as in [17], from (5.7) and (5.8) we thus infer the inequality

d d t Δ j u h L p ( 3 ) + c ¯ p 2 2 j Δ j u h L p ( 3 ) Δ j ( u u h ) L p ( 3 ) + Δ j h L p ( 3 ) + Δ j ( n h ϕ ) L p ( 3 ) ,

which after time, by integrating over [0,t] and applying Lemmas 2.9, 2.11, 5.2 and Proposition 5.4, yields that

Δ j u h L t ( L p ( 3 ) ) + c ¯ 2 2 j Δ j u h L t 1 ( L p ( 3 ) )
Δ j u 0 h L p ( 3 ) + 2 j Δ j ( u h u h ) L t 1 ( L p ( 3 ) )
    + 2 j Δ j ( u 3 u h ) L p ( 3 ) + Δ j h L t 1 ( L p ( 3 ) ) + Δ j ( n h ϕ ) L t 1 ( L p ( 3 ) )
Δ j u 0 h L p ( 3 ) + C ~ 2 j 2 - 3 j / p d j ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) )
    + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p + u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + 1 / p
    + 𝕎 ( u h , u 3 ) ) + C ~ d j 2 j ( 1 - 3 / p ) n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )
(5.9) Δ j u 0 h L p ( 3 ) + C ~ 2 j ( 1 - 3 / p ) d j 𝕎 ( u h , u 3 ) + C ~ d j 2 j ( 1 - 3 / p ) n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 ) ,

from which we obtain the desired estimate (5.5).

Finally, as in the proof of (5.9), let us apply Δj to (5.4). We need to make some modifications as in [17]. By virtue of Lemmas 2.11, 5.2 and Proposition 5.4, we have

Δ j u 3 L t ( L p ( 3 ) ) + c ¯ p 2 2 j Δ j u 3 L t 1 ( L p ( 3 ) )
Δ j u 0 3 L p ( 3 ) + Δ j ( u u 3 ) L t 1 ( L p ( 3 ) ) + Δ j 3 L t 1 ( L p ( 3 ) ) + Δ j ( n 3 ϕ ) L t 1 ( L p ( 3 ) )
Δ j u 0 3 L p ( 3 ) + 2 j Δ j ( u h u 3 ) L t 1 ( L p ( 3 ) ) + Δ j ( u 3 div h u h ) L t 1 ( L p ( 3 ) )
    + Δ j 3 L t 1 ( L p ( 3 ) ) + Δ j ( n 3 ϕ ) L t 1 ( L p ( 3 ) )
Δ j u 0 3 L p + d j 2 j ( 1 - 3 / p ) u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) + C ~ 2 j ( 1 - 3 / p ) d j 𝕎 ( u h , u 3 )
(5.10)     + C ~ d j 2 j ( 1 - 3 / p ) n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 ) .

Multiplying (5.10) by 2j(-1+3/p) and summing up over j yield (5.6). Therefore, we complete the proof of Proposition 5.5. ∎

The estimate of n.

Let us consider the second equation of (1.4); we have

(5.11) t n - Δ n = - u n - ( n c ) .

Proposition 5.6.

Let 1<pq<6 and 1p-1q13. If (u,n,c)ΘT with u=0, then there holds

n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) + c ¯ n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) )
n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) + ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p ) n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) )
(5.12)     + C ( n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) c 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ( 3 ) ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) ) .

Proof.

Applying the operator Δj to (5.11), similar to (5.7) we obtain

(5.13) 1 q d d t Δ j n L q ( 3 ) q - 3 Δ Δ j n | Δ j n | q - 2 Δ j n 𝑑 x = - 3 Δ j ( u n + ( n c ) ) | Δ j n | q - 2 Δ j n 𝑑 x .

Thanks to [17, Proposition 2.1], there exists a positive constant c¯ so that

(5.14) - 3 Δ Δ j n | Δ j n | q - 2 Δ j n 𝑑 x c ¯ 2 2 j Δ j n L q ( 3 ) q .

It is clear that we have the following homogeneous Bony’s paraproduct decomposition:

(5.15) u n = T u n + R ( u , n ) + T n u .

Besides, because u=0, by using the standard commutator’s argument, one obtains

3 Δ j ( T u n ) | Δ j n | q - 2 Δ j n 𝑑 x
(5.16) = | j - j | 5 3 [ Δ j ; S j - 1 u ] Δ j n | Δ j n | q - 2 Δ j n 𝑑 x + | j - j | 5 3 ( S j - 1 u - S j - 1 u ) Δ j Δ j n | Δ j n | q - 2 Δ j n 𝑑 x .

Therefore, substituting (5.14)–(5.16) into (5.13), and using an argument for the Lq(3) energy estimate as the one in [17, Lemma A.2], we reach

Δ j n L t ( L q ( 3 ) ) + c ¯ 2 2 j 0 t Δ j n L q ( 3 ) 𝑑 τ
Δ j n 0 L q ( 3 ) + C ~ | j - j | 5 [ Δ j ; S j - 1 u ] Δ j n L t 1 ( L q ( 3 ) ) + C | j - j | 5 ( S j - 1 u - S j - 1 u ) Δ j Δ j n L t 1 ( L q ( 3 ) )
(5.17)     + C ~ Δ j ( R ( u , n ) ) L t 1 ( L q ( 3 ) ) + C ~ Δ j ( n c ) L t 1 ( L q ( 3 ) ) .

Applying the classical estimate on commutators [17] and Lemma 2.1 yields

| j - j | 5 [ Δ j ; S j - 1 u ] Δ j n L t 1 ( L q ( 3 ) )
| j - j | 5 k j - 2 2 k ( 1 + 3 / p ) Δ k u h L t 1 ( L p ) Δ j n L t ( L q ( 3 ) ) + | j - j | 5 k j - 2 2 k Δ k u 3 ( τ ) L t 1 ( L ) Δ j n ( τ ) L t ( L q ( 3 ) )
| j - j | 5 d j 2 j ( 2 - 3 / q ) u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) )
    + | j - j | 5 u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p Δ j n L t ( L q ( 3 ) ) d j
(5.18) d j 2 j ( 2 - 3 / q ) ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p ) n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) .

Now, using the same type of computations as in (5.18), we get

| j - j | 5 ( S j - 1 u - S j - 1 u ) Δ j Δ j n L t 1 ( L q ( 3 ) )
| j - j | 5 ( S j - 1 u h - S j - 1 u h ) L t 1 ( L ( 3 ) ) Δ j n L t ( L q ( 3 ) )
    + | j - j | 5 S j - 1 u 3 ( τ ) - S j - 1 u 3 ( τ ) L t 1 ( L ( 3 ) ) Δ j n ( τ ) L t ( L q ( 3 ) )
(5.19) d j 2 j ( 2 - 3 / q ) ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ) 1 / p ) n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) .

Meanwhile, we obtain

Δ j R ( u , n ) L t 1 ( L q ( 3 ) )
j j - N 0 Δ j u h L t 1 ( L q ( 3 ) ) S j + 2 h n L t ( L ( 3 ) ) + j j - N 0 Δ j u 3 L t 1 ( L h q ( ) ( L v ( 2 ) ) ) S j + 2 3 n L t ( L h ( L v q ( ) ( 2 ) ) )
j j - N 0 j ′′ j + 1 ( 2 j ′′ ( 1 + 3 / q ) Δ j ′′ n L t ( L q ( 3 ) ) 2 j ( 3 / p - 3 / q ) Δ j u h L t 1 ( L p ( 3 ) )
    + Δ j u 3 L t 1 ( L h q ( ) ( L v ( 2 ) ) ) 2 j ′′ ( 1 + 3 / q ) Δ j ′′ n L t ( L q ( 3 ) ) )
(5.20) d j 2 j ( 2 - 3 / q ) ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p ) n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) .

So plugging the above inequalities (5.18)–(5.20) into (5.17) and using Lemma 2.5, we conclude

Δ j n L t ( L q ( 3 ) ) + c ¯ 2 2 j 0 t Δ j n L q ( 3 ) 𝑑 τ
Δ j n 0 L q ( 3 ) + d j 2 j ( 2 - 3 / q ) ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p ) n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) )
    + d j 2 j ( 2 - 3 / q ) ( n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) c 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ( 3 ) ) + n 𝔏 T 1 ( B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) ) .

Then multiplying both sides by 2j(-2+3/q) and taking the 1()-norm, we obtain Proposition 5.6. ∎

The estimate of c.

Thanks to the second equation of (1.4), we have

(5.21) t c - Δ c = - u c - c n .

Proposition 5.7.

Let 1<pq<6 and 1p-1q13. If (u,n,c)ΘT with u=0, then there holds

c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) + c ¯ c 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ( 3 ) )
c 0 B ˙ q , 1 3 / q ( 3 ) + C ( ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p )
(5.22)     × c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) ) .

Proof.

Applying the operator Δj to (5.21) and taking the L2 inner product with |Δjc|q-2Δjc (in the case when q(1,2), we need to make some modifications as in [17]), an integration by parts yields

Δ j c L t ( L q ( 3 ) ) + c ¯ 2 2 j 0 t Δ j c L q ( 3 ) 𝑑 τ
Δ j c 0 L q ( 3 ) + C ~ | j - j | 5 [ Δ j ; S j - 1 u ] Δ j c L t 1 ( L q ( 3 ) ) + C ~ | j - j | 5 ( S j - 1 u - S j - 1 u ) Δ j Δ j c L t 1 ( L q ( 3 ) )
    + C ~ Δ j R ( u , c ) L t 1 ( L q ( 3 ) ) + C ~ Δ j ( c n ) L t 1 ( L q ( 3 ) ) .

Repeating the process as in Proposition 5.6, from Lemma 2.5 we conclude

Δ j c L t ( L q ( 3 ) ) + c ¯ 2 2 j 0 t Δ j c L q ( 3 ) 𝑑 τ
Δ j c 0 L q ( 3 ) + d j 2 j ( 2 - 3 / q ) ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) )
    + d j 2 j ( 2 - 3 / q ) n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) .

Multiplying both sides of the above inequality with 23j/q and then taking the 1()-norm, we obtain the desired estimates (5.22). This proves Proposition 5.7. ∎

Proof of Theorem 1.3.

If we take C>0 large enough in (1.5), then there exists a positive time T such that system (1.4) has a unique local solution

( u , n , c ) Θ T Θ T 𝒞 .

Let T* be a maximal time of existence introduced in Theorem 1.1. Hence, to prove Theorem 1.3, we only need to prove that T*= with (u,n,c)ΘΘ𝒞 provided that (1.5) holds. Let

𝔄 0 := 4 u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + η , 𝔅 0 := 4 u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) + 1 , := 4 ( n 0 , c 0 ) B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 ) ,

where η is a positive constant which will be determined later on. We define

𝔗 := { t [ 0 , T * ) : u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 𝔄 0 , u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 𝔅 0 ,
(5.23) ( n , c ) 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) × 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) + c ¯ ( n , c ) 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) × 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ( 3 ) ) } .

Applying (5.12), we have

n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )
ϕ B ˙ p , 1 3 / p ( 3 ) n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) + ϕ B ˙ p , 1 3 / p ( 3 ) ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p )
   × n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) + ϕ B ˙ p , 1 3 / p ( 3 ) ( n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) c 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ( 3 ) ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) )
(5.24) n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) + ( 𝔄 0 + 𝔅 0 1 - 1 / p 𝔄 0 1 / p + 2 ) ϕ B ˙ p , 1 3 / p ( 3 ) .

Applying Propositions 5.6 and 5.7, we obtain

( n , c ) 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) × 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) + c ¯ ( n , c ) 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) × 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ( 3 ) )
( n 0 , c 0 ) B ˙ q , 1 - 2 + 3 / q × B ˙ q , 1 3 / q ( 3 ) + ( u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p )
    × ( n , c ) 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) + n 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) c 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ) c 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) )
( n 0 , c 0 ) B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 - 2 + 3 / q ( 3 ) + C ( 𝔄 0 + 𝔄 0 1 / p 𝔅 0 1 - 1 / p ) ( n , c ) 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) × 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) )
(5.25)     + C 0 ( n , c ) 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) × 𝔏 t ( B ˙ q , 1 2 + 3 / q ( 3 ) ) .

Noting the definition of the Besov spaces norms and using the Cauchy–Schwarz inequality

a b ε 2 a 2 + 1 2 ε b 2

for ε=2c¯ yield

𝕎 ( u h , u 3 ) := u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) )
+ u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 - 1 / p u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 1 / p
+ u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + 1 / p u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 - 1 / p + u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) 1 + α u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ) 1 - α
𝔄 0 u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + 𝔄 0 1 / p 𝔅 0 1 - 1 / p u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) )
+ ( c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) 1 + 1 / p
× ( c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ) + u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) 1 - 1 / p
+ ( c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) 1 + α
× ( c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) 1 - α
(5.26) ( 𝔄 0 + 𝔄 0 α 𝔅 0 1 - α + 𝔄 0 1 / p 𝔅 0 1 - 1 / p ) ( c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) .

Thanks to (5.24), (5.26) and Proposition 5.5, we have

u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) )
u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + 𝕎 ( u h , u 3 ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )
u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + ( 𝔄 0 + 𝔄 0 1 / p 𝔅 0 1 - 1 / p )
    × ( c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )
u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + C ( 𝔄 0 + 𝔄 0 1 / p 𝔅 0 1 - 1 / p ) ( c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) )
(5.27)     + C n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) + C ( 𝔄 0 + 𝔅 0 1 - 1 / p 𝔄 0 1 / p + 2 ) ϕ B ˙ p , 1 3 / p ( 3 ) .

We deduce from (5.24), (5.26) and Proposition 5.5 that

u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) )
u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) + u h 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) u 3 𝔏 t 2 ( B ˙ p , 1 3 / p ( 3 ) ) + 𝕎 ( u h , u 3 ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )
u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) + 𝔄 0 ( c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) + n 𝔏 t 1 ( B ˙ q , 1 3 / q ( 3 ) ) ϕ B ˙ p , 1 3 / p ( 3 )
    + ( 𝔄 0 + 𝔄 0 1 / p 𝔅 0 1 - 1 / p ) ( c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) + u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) ) + 𝔄 0 1 + 1 / p 𝔅 0 1 - 1 / p + 𝔄 0 1 + α 𝔅 0 1 - α
u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) + C 𝔄 0 ( c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ) + u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) )
    + C n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) + C ( 𝔄 0 + 𝔅 0 1 - 1 / p 𝔄 0 1 / p + 2 ) ϕ B ˙ p , 1 3 / p ( 3 )
(5.28)     + C 𝔄 0 2 + C 𝔄 0 1 + α 𝔅 0 1 - α + C 𝔄 0 1 + 1 / p 𝔅 0 1 - 1 / p .

Let

η := 4 C n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) + 4 C 2 ϕ B ˙ p , 1 3 / p ( 3 ) ( 4 u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + 4 C n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) )
+ 𝔅 0 1 - 1 / p ( 4 u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + 4 C n 0 B ˙ q , 1 - 2 + 3 / q ( 3 ) ) 1 / p .

Thanks to (5.23), taking C~>0 large enough in (1.5), we have

(5.29) { C 𝔄 0 < 1 2 , C 0 < c ¯ 2 , C ( 𝔄 0 + 𝔄 0 1 / p 𝔅 0 1 - 1 / p + 𝔄 0 1 + α 𝔅 0 1 - α ) 1 2 , η 4 + C 𝔄 0 2 + C 𝔄 0 1 + α 𝔅 0 1 - α + C 𝔄 0 1 + 1 / p 𝔅 0 1 - 1 / p < 1 4 .

Applying (5.25) and (5.27)–(5.29), we obtain

(5.30) { u h 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u h 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ( 3 ) ) 2 u 0 h B ˙ p , 1 - 1 + 3 / p ( 3 ) + 1 2 η , u 3 𝔏 t ( B ˙ p , 1 - 1 + 3 / p ( 3 ) ) + c ¯ u 3 𝔏 t 1 ( B ˙ p , 1 1 + 3 / p ) 2 u 0 3 B ˙ p , 1 - 1 + 3 / p ( 3 ) + 1 2 , ( n , c ) 𝔏 t ( B ˙ q , 1 - 2 + 3 / q ( 3 ) ) × 𝔏 t ( B ˙ q , 1 3 / q ( 3 ) ) + c ¯ ( n , c ) 𝔏 t 1 ( B ˙ q , 1 3 / q ) × 𝔏 t 1 ( B ˙ q , 1 2 + 3 / q ) 2 ( n 0 , c 0 ) B ˙ q , 1 - 2 + 3 / q ( 3 ) × B ˙ q , 1 3 / q ( 3 ) .

Thus (5.30) contradicts the definition (5.23). In view of the standard continuity argument, we conclude that 𝔗:=T*. Therefore, we complete the proof of Theorem 1.3. ∎


Communicated by Guozhen Lu


Award Identifier / Grant number: 11671185

Award Identifier / Grant number: 11771195

Award Identifier / Grant number: ZR2017MA041

Funding statement: This work was partially supported by NSF of China (grant nos. 11671185, 11771195) and NSF Shandong Province (grant no. ZR2017MA041).

References

[1] H. Bae, Existence and analyticity of Lei–Lin solution to the Navier–Stokes equations, Proc. Amer. Math. Soc. 143 (2015), no. 7, 2887–2892. 10.1090/S0002-9939-2015-12266-6Suche in Google Scholar

[2] H. Bae, A. Biswas and E. Tadmor, Analyticity and decay estimates of the Navier–Stokes equations in critical Besov spaces, Arch. Ration. Mech. Anal. 205 (2012), no. 3, 963–991. 10.1007/s00205-012-0532-5Suche in Google Scholar

[3] H. Bahouri, J.-Y. Chemin and R. Danchin, Fourier Analysis and Nonlinear Partial Differential Equations, Grundlehren Math. Wiss. 343, Springer, Heidelberg, 2011. 10.1007/978-3-642-16830-7Suche in Google Scholar

[4] A. Biswas, Gevrey regularity for a class of dissipative equations with applications to decay, J. Differential Equations 253 (2012), no. 10, 2739–2764. 10.1016/j.jde.2012.08.003Suche in Google Scholar

[5] A. Biswas, V. R. Martinez and P. Silva, On Gevrey regularity of the supercritical SQG equation in critical Besov spaces, J. Funct. Anal. 269 (2015), no. 10, 3083–3119. 10.1016/j.jfa.2015.08.010Suche in Google Scholar

[6] A. Biswas and D. Swanson, Gevrey regularity of solutions to the 3-D Navier–Stokes equations with weighted lp initial data, Indiana Univ. Math. J. 56 (2007), no. 3, 1157–1188. 10.1512/iumj.2007.56.2891Suche in Google Scholar

[7] M. Cannone, Harmonic analysis tools for solving the incompressible Navier–Stokes equations, Handbook of Mathematical Fluid Dynamics. Vol. III, North-Holland, Amsterdam (2004), 161–244. 10.1016/S1874-5792(05)80006-0Suche in Google Scholar

[8] M. Chae, K. Kang and J. Lee, Existence of smooth solutions to coupled chemotaxis-fluid equations, Discrete Contin. Dyn. Syst. 33 (2013), no. 6, 2271–2297. 10.3934/dcds.2013.33.2271Suche in Google Scholar

[9] M. Chae, K. Kang and J. Lee, Global existence and temporal decay in Keller–Segel models coupled to fluid equations, Comm. Partial Differential Equations 39 (2014), no. 7, 1205–1235. 10.1080/03605302.2013.852224Suche in Google Scholar

[10] M. Chae, K. Kang and J. Lee, Asymptotic behaviors of solutions for an aerotaxis model coupled to fluid equations, J. Korean Math. Soc. 53 (2016), no. 1, 127–146. 10.4134/JKMS.2016.53.1.127Suche in Google Scholar

[11] J.-Y. Chemin, Théorèmes d’unicité pour le système de Navier–Stokes tridimensionnel, J. Anal. Math. 77 (1999), 27–50. 10.1007/BF02791256Suche in Google Scholar

[12] J.-Y. Chemin, M. Paicu and P. Zhang, Global large solutions to 3-D inhomogeneous Navier–Stokes system with one slow variable, J. Differential Equations 256 (2014), no. 1, 223–252. 10.1016/j.jde.2013.09.004Suche in Google Scholar

[13] L. Chen, G. Lu and X. Luo, Boundedness of multi-parameter Fourier multiplier operators on Triebel–Lizorkin and Besov–Lipschitz spaces, Nonlinear Anal. 134 (2016), 55–69. 10.1016/j.na.2015.12.016Suche in Google Scholar

[14] H. J. Choe and B. Lkhagvasuren, Global existence result for chemotaxis Navier–Stokes equations in the critical Besov spaces, J. Math. Anal. Appl. 446 (2017), no. 2, 1415–1426. 10.1016/j.jmaa.2016.09.050Suche in Google Scholar

[15] H. J. Choe, B. Lkhagvasuren and M. Yang, Wellposedness of the Keller–Segel Navier–Stokes equations in the critical Besov spaces, Commun. Pure Appl. Anal. 14 (2015), no. 6, 2453–2464. 10.3934/cpaa.2015.14.2453Suche in Google Scholar

[16] W. Dai and G. Lu, Lp estimates for multi-linear and multi-parameter pseudo-differential operators, Bull. Soc. Math. France 143 (2015), no. 3, 567–597. 10.24033/bsmf.2698Suche in Google Scholar

[17] R. Danchin, Local theory in critical spaces for compressible viscous and heat-conductive gases, Comm. Partial Differential Equations 26 (2001), no. 7–8, 1183–1233. 10.1081/PDE-100106132Suche in Google Scholar

[18] R. Danchin, Fourier analysis methods for PDEs, Lecture Notes (2005). Suche in Google Scholar

[19] R. Duan, A. Lorz and P. Markowich, Global solutions to the coupled chemotaxis-fluid equations, Comm. Partial Differential Equations 35 (2010), no. 9, 1635–1673. 10.1080/03605302.2010.497199Suche in Google Scholar

[20] C. Foias and R. Temam, Gevrey class regularity for the solutions of the Navier–Stokes equations, J. Funct. Anal. 87 (1989), no. 2, 359–369. 10.1016/0022-1236(89)90015-3Suche in Google Scholar

[21] H. Fujita and T. Kato, On the Navier–Stokes initial value problem. I, Arch. Ration. Mech. Anal. 16 (1964), 269–315. 10.1007/BF00276188Suche in Google Scholar

[22] G. Gui and P. Zhang, Stability to the global large solutions of 3-D Navier–Stokes equations, Adv. Math. 225 (2010), no. 3, 1248–1284. 10.1007/978-3-642-36028-2_2Suche in Google Scholar

[23] J. Huang, M. Paicu and P. Zhang, Global well-posedness of incompressible inhomogeneous fluid systems with bounded density or non-Lipschitz velocity, Arch. Ration. Mech. Anal. 209 (2013), no. 2, 631–682. 10.1007/s00205-013-0624-xSuche in Google Scholar

[24] J. Jiang, H. Wu and S. Zheng, Global existence and asymptotic behavior of solutions to a chemotaxis-fluid system on general bounded domains, Asymptot. Anal. 92 (2015), no. 3–4, 249–258. 10.3233/ASY-141276Suche in Google Scholar

[25] T. Kato, Strong Lp-solutions of the Navier–Stokes equation in 𝐑m, with applications to weak solutions, Math. Z. 187 (1984), no. 4, 471–480. 10.1007/BF01174182Suche in Google Scholar

[26] H. Koch and D. Tataru, Well-posedness for the Navier–Stokes equations, Adv. Math. 157 (2001), no. 1, 22–35. 10.1006/aima.2000.1937Suche in Google Scholar

[27] H. Kozono and M. Yamazaki, Semilinear heat equations and the Navier–Stokes equation with distributions in new function spaces as initial data, Comm. Partial Differential Equations 19 (1994), 959–1014. 10.1080/03605309408821042Suche in Google Scholar

[28] Z. Lei and F. Lin, Global mild solutions of Navier–Stokes equations, Comm. Pure Appl. Math. 64 (2011), no. 9, 1297–1304. 10.1002/cpa.20361Suche in Google Scholar

[29] J. Leray, Sur le mouvement d’un liquide visqueux emplissant l’espace, Acta Math. 63 (1934), no. 1, 193–248. 10.1007/BF02547354Suche in Google Scholar

[30] J.-G. Liu and A. Lorz, A coupled chemotaxis-fluid model: Global existence, Ann. Inst. H. Poincaré Anal. Non Linéaire 28 (2011), no. 5, 643–652. 10.1016/j.anihpc.2011.04.005Suche in Google Scholar

[31] Q. Liu, Gevrey analyticity of solutions to the 3D nematic liquid crystal flows in critical Besov space, Nonlinear Anal. Real World Appl. 31 (2016), 431–451. 10.1016/j.nonrwa.2016.02.008Suche in Google Scholar

[32] A. Lorz, Coupled chemotaxis fluid model, Math. Models Methods Appl. Sci. 20 (2010), no. 6, 987–1004. 10.1142/S0218202510004507Suche in Google Scholar

[33] A. Lorz, A coupled Keller–Segel–Stokes model: Global existence for small initial data and blow-up delay, Commun. Math. Sci. 10 (2012), no. 2, 555–574. 10.4310/CMS.2012.v10.n2.a7Suche in Google Scholar

[34] M. Paicu, Équation anisotrope de Navier–Stokes dans des espaces critiques, Rev. Mat. Iberoam. 21 (2005), no. 1, 179–235. 10.4171/RMI/420Suche in Google Scholar

[35] M. Paicu and P. Zhang, Global solutions to the 3-D incompressible anisotropic Navier–Stokes system in the critical spaces, Comm. Math. Phys. 307 (2011), no. 3, 713–759. 10.1007/s00220-011-1350-6Suche in Google Scholar

[36] M. Paicu and P. Zhang, Global solutions to the 3-D incompressible inhomogeneous Navier–Stokes system, J. Funct. Anal. 262 (2012), no. 8, 3556–3584. 10.1016/j.jfa.2012.01.022Suche in Google Scholar

[37] F. Planchon, Sur un inégalité de type Poincaré, C. R. Acad. Sci. Paris Sér. I Math. 330 (2000), no. 1, 21–23. 10.1016/S0764-4442(00)88138-0Suche in Google Scholar

[38] E. M. Stein, Singular Integrals and Differentiability Properties of Functions, Princeton Math. Ser. 30, Princeton University Press, Princeton, 1970. 10.1515/9781400883882Suche in Google Scholar

[39] Y. Tao, Boundedness in a chemotaxis model with oxygen consumption by bacteria, J. Math. Anal. Appl. 381 (2011), no. 2, 521–529. 10.1016/j.jmaa.2011.02.041Suche in Google Scholar

[40] I. Tuval, L. Cisneros, C. Dombrowski, W. Wolgemuth, O. Kessler and E. Goldstein, Bacterial swimming and oxygen transport near contact lines, Proc. Natl. Acad. Sci. USA 102 (2005), no. 7, 2277–2282. 10.1073/pnas.0406724102Suche in Google Scholar PubMed PubMed Central

[41] B. Wang, Z. Huo, C. Hao and Z. Guo, Harmonic Analysis Method for Nonlinear Evolution Equations. I, World Scientific Publishing, Hackensack, 2011. 10.1142/8209Suche in Google Scholar

[42] M. Winkler, Global large-data solutions in a chemotaxis-Navier–Stokes system modeling cellular swimming in fluid drops, Comm. Partial Differential Equations 37 (2012), no. 2, 319–351. 10.1080/03605302.2011.591865Suche in Google Scholar

[43] M. Winkler, Stabilization in a two-dimensional chemotaxis-Navier–Stokes system, Arch. Ration. Mech. Anal. 211 (2014), no. 2, 455–487. 10.1007/s00205-013-0678-9Suche in Google Scholar

[44] M. Winkler, Global weak solutions in a three-dimensional chemotaxis–Navier–Stokes system, Ann. Inst. H. Poincaré Anal. Non Linéaire 33 (2016), no. 5, 1329–1352. 10.1016/j.anihpc.2015.05.002Suche in Google Scholar

[45] J. Xiao, Homothetic variant of fractional Sobolev space with application to Navier–Stokes system, Dyn. Partial Differ. Equ. 4 (2007), no. 3, 227–245. 10.4310/DPDE.2007.v4.n3.a2Suche in Google Scholar

[46] J. Xiao, Homothetic variant of fractional Sobolev space with application to Navier–Stokes system revisited, Dyn. Partial Differ. Equ. 11 (2014), no. 2, 167–181. 10.4310/DPDE.2014.v11.n2.a3Suche in Google Scholar

[47] M. Yang, Z. Fu and J. Sun, Existence and Gevrey regularity for a two-species chemotaxis system in homogeneous Besov spaces, Sci. China Math. 60 (2017), no. 10, 1837–1856. 10.1007/s11425-016-0490-ySuche in Google Scholar

[48] M. Yang and J. Sun, Gevrey regularity and existence of Navier–Stokes–Nernst–Planck–Poisson system in critical Besov spaces, Commun. Pure Appl. Anal. 16 (2017), no. 5, 1617–1639. 10.3934/cpaa.2017078Suche in Google Scholar

[49] C. Zhai and T. Zhang, Global well-posedness to the 3-D incompressible inhomogeneous Navier–Stokes equations with a class of large velocity, J. Math. Phys. 56 (2015), no. 9, Article ID 091512. 10.1063/1.4931467Suche in Google Scholar

[50] Q. Zhang, Local well-posedness for the chemotaxis-Navier–Stokes equations in Besov spaces, Nonlinear Anal. Real World Appl. 17 (2014), 89–100. 10.1016/j.nonrwa.2013.10.008Suche in Google Scholar

[51] Q. Zhang and X. Zheng, Global well-posedness for the two-dimensional incompressible chemotaxis-Navier–Stokes equations, SIAM J. Math. Anal. 46 (2014), no. 4, 3078–3105. 10.1137/130936920Suche in Google Scholar

Received: 2017-03-24
Revised: 2018-01-04
Accepted: 2018-01-07
Published Online: 2018-01-20
Published in Print: 2018-08-01

© 2018 Walter de Gruyter GmbH, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.

Heruntergeladen am 17.9.2025 von https://www.degruyterbrill.com/document/doi/10.1515/ans-2017-6046/html
Button zum nach oben scrollen