Skip to Content
MathematicsMathematics
  • Article
  • Open Access

16 November 2026

13 Pages

Vanishing Diffusivity and Viscosity Limit of 3D Boussinesq Equations with a Slip Boundary Condition

and
School of Mathematics and Statistics, Pu’er University, Pu’er 665000, China
*
Author to whom correspondence should be addressed.

Abstract

This paper investigates the regularity of solutions to the three-dimensional Boussinesq equations with slip boundary conditions in a cubic domain. Uniform H 3 bounds for strong solutions are established. Furthermore, the vanishing diffusivity and viscosity limits are discussed, and explicit convergence rates are obtained: the strong solution converges to a weak solution of the inviscid temperature system at a rate of κ in H 1 , and to the unique strong solution of the Euler–Boussinesq system at a rate of ( κ + ν ) 1 2 in H 2 .

1. Introduction

We consider the three-dimensional heat-conductive Boussinesq system, which is governed by the following coupled partial differential equations.
∂ t u − ν Δ u + ( u · ∇ ) u + ∇ p = θ e 3 in Ω × ( 0 , T ] , ∂ t θ − κ Δ θ + u · ∇ θ = 0 in Ω × ( 0 , T ] , ∇ · u = 0 in Ω × ( 0 , T ] , u · n = 0 , ∇ × u × n = 0 , ∂ θ ∂ n = 0 on ∂ Ω × ( 0 , T ] , ( θ , u ) | t = 0 = ( θ 0 , u 0 ) in Ω ,
where Ω ⊂ R 3 is a simply connected bounded open domain and the vector u denotes the velocity of the fluid, while the scalars θ , p represent the deviation from the basic temperature and the hydrostatic pressure, respectively. e 3 is the third standard basis vector of R 3 and n denotes the outward unit normal to ∂ Ω . The positive parameters ν , κ are the kinematic viscosity coefficient and the thermal conductivity coefficient, respectively. From a mathematical perspective, the Boussinesq equations have been extensively studied. Numerous results concerning existence, well-posedness, inviscid limit, and large-time behavior have been proved; see, for instance, refs. [1,2,3,4,5,6,7] and their references therein.
Recently, problems related to the Boussinesq equations have been investigated in [8,9,10,11,12,13,14]. In this work, we focus on the vanishing viscosity and diffusivity limits. Under slip boundary conditions, Berselli and Spirito [15] studied the vanishing viscosity limit for the viscous Boussinesq equations without diffusion and obtained weak convergence to the Euler–Boussinesq equations. Jiang et al. [16] established the vanishing diffusivity limit for the two-dimensional Boussinesq equations in R + 2 with no-slip boundary conditions for the velocity field and inhomogeneous Dirichlet boundary conditions for the temperature. Wang et al. [17] discussed the zero dissipation limit for the initial-boundary value problem of the multi-dimensional Boussinesq equations with viscosity and heat conductivity, proving L 2 convergence of solutions via a standard energy method, as both the viscosity and heat conductivity coefficients tend to zero. Zhang [18] studied the three-dimensional heat-conductive Boussinesq equations with slip boundary conditions for the velocity field and Neumann boundary conditions for the temperature, establishing the vanishing diffusivity limit and obtaining convergence rates for both the velocity field and the temperature.
The authors in [19] improved the convergence rate of solutions for the viscous Navier–Stokes equations with slip-without-friction boundary conditions involving vorticity:
u · n = 0 on Γ × ( 0 , T ] , ω × n = 0 on Γ × ( 0 , T ] ,
toward those of the Euler equations, under initial data satisfying the Euler equations
u 0 · n = 0 on Γ , ω 0 = 0 on Γ ,
in W s , p ( Ω ) ( s ∈ [ 1 , 2 ) , p > 3 ) . The existence of weak solutions for the nonhomogeneous Navier–Stokes system with Navier friction boundary conditions
u · n = 0 on ∂ Ω × ( 0 , T ] , [ D ( u ) n + α u ] t a n = 0 on ∂ Ω × ( 0 , T ] ,
where D ( u ) = 1 2 ( ∂ i u j + ∂ j u i ) , was proved, and the convergence of weak solutions of the nonhomogeneous Navier–Stokes system to strong solutions of the Euler equations was studied in [20].
The vanishing viscosity limit for the three-dimensional incompressible magnetohydrodynamic (MHD) system in a general bounded smooth domain Ω ⊂ R 3 with generalized Navier slip boundary conditions,
v · n = 0 , n × ( ∇ × v ) = [ β v ] τ on ∂ Ω × ( 0 , T ] , H · n = 0 , n × ( ∇ × H ) = [ β H ] τ on ∂ Ω × ( 0 , T ] ,
was investigated, and convergence rates of solutions of the viscous MHD system to the corresponding ideal MHD system were obtained in [21].
For domains with flat boundaries, Xiao and Xin [22] studied the vanishing viscosity limit for the incompressible Navier–Stokes equations with slip boundary conditions, proving strong convergence in L q ( 0 , T ; H 3 ( Ω ) ) ∩ C ( [ 0 , T ] ; H 2 ( Ω ) ) for q ≥ 1 . Moreover, Beirão da Veiga and Crispo [23] established strong convergence in C ( [ 0 , T 0 ] ; W s , p ( Ω ) ) for any s < 3 and p > 3 2 using an L p approach. Subsequently, the results in [23] were extended to W k , p ( Ω ) for arbitrarily large k and p in [24].
The novelty of this work is best highlighted by a direct comparison with [18]. First, regarding regularity, [18] achieved θ ∈ L ∞ ( 0 , T ; H 2 ( Ω ) ) ∩ L 2 ( 0 , T ; H 3 ( Ω ) ) and u ∈ L ∞ ( 0 , T ; H 3 ( Ω ) ) ∩ L 2 ( 0 , T ; H 4 ( Ω ) ) . By developing a boundary-layer analysis tailored to the slip boundary conditions, we strictly improve the regularity of the temperature to perfectly match the velocity field, establishing ( θ , u ) ∈ L ∞ ( 0 , T ; H 3 ( Ω ) ) ∩ L 2 ( 0 , T ; H 4 ( Ω ) ) . Second, whereas the uniform bound in [18] is independent of κ , our estimate is uniform with respect to both κ and ν . Third, for the vanishing parameter limits, instead of treating a single parameter as in [18] (which yields a rate of κ in the mixed lower-regularity norms L 2 ( Ω ) × H 1 ( Ω ) ), we address the joint limit κ , ν → 0 and prove a strong uniform H 2 ( Ω ) convergence rate of O ( κ + ν ) towards the unique strong solution of the Euler–Boussinesq system.

2. Preliminaries

Throughout the remainder of this paper, let Ω ⊂ R 3 denote the cubic domain ( 0 , 1 ) 3 , and define
∂ Ω = x : 0 ≤ x 1 , x 2 ≤ 1 , and x 3 = 0 or x 3 = 1 .
Boundary conditions are imposed only on the two opposite faces, while periodicity is assumed on the remaining faces.
For 1 ≤ q ≤ ∞ , ∥ · ∥ L q denotes the norm in the Lebesgue space L q ( Ω ) , and ∥ · ∥ W k , q denotes the norm in the Sobolev space W k , q ( Ω ) . In the case q = 2 , we write H k ( Ω ) with norm ∥ · ∥ H k ( Ω ) . Moreover, ( u , v ) = 0 , 0 , 0 ∫ Ω u · v d x represents the inner product of u and v. The symbols ∇, ∇ · , ∇ × , and Δ denote the gradient, divergence, curl, and Laplace operators, respectively. For convenience, we introduce
w = ∇ × u , ζ = ∇ × w , χ = ∇ × ζ .
The following trace inequality will also be used (see [19]).
Lemma 1.
Let Ω be a smooth bounded open set and let 1 < p < + ∞ . Then there exists a constant C > 0 such that
∥ u ∥ L p ( ∂ Ω ) ≤ C ∥ u ∥ L p ( Ω ) 1 − 1 p ∥ u ∥ W 1 , p ( Ω ) 1 p .
It is well known that there exists a constant C = C ( s , p , Ω ) such that
∥ u ∥ W s , p ( Ω ) ≤ C ( ∥ ∇ · u ∥ W s − 1 , p ( Ω ) + ∥ ∇ × u ∥ W s − 1 , p ( Ω ) + ∥ u · n ∥ W s − 1 / p , p ( ∂ Ω ) + ∥ u ∥ W s − 1 , p ( Ω ) ) ,
for any s ≥ 1 and u ∈ W s , p ( Ω ) , see [25]. Based on this estimate, one can establish norm-equivalence relations for certain functions derived from u. Since the divergence-free condition ( ∇ · u = 0 ) and the slip boundary condition ( u · n = 0 ) hold, the lower-order term ∥ u ∥ W s − 1 , p in (8) can be absorbed via the standard Poincaré inequality. These are summarized in the following lemma, which is adapted from Beirão [23].
Lemma 2.
Let u, w, ζ, χ be as defined above. Then, for each non-negative integer k, the following norm equivalences hold:
∥ w ∥ H k ≃ ∥ u ∥ H k + 1 ; ∥ ζ ∥ H k ≃ ∥ u ∥ H k + 2 ; ∥ χ ∥ H k ≃ ∥ u ∥ H k + 3 .
Lemma 3.
Let u be a vector field in Ω, and assume that u 3 = ω 1 = ω 2 = 0 on ∂ Ω . Then the vector fields ( u · ∇ ) ω and ( ω · ∇ ) u are normal to ∂ Ω .
Finally, we introduce some boundary-related results that will be used in the subsequent proofs.
Proposition 1.
Let ( θ , u ) be a sufficiently smooth pair of functions. If ( θ , u ) is a solution to system (1), then
n · ∇ Δ θ = 0 on ∂ Ω .
Proof. 
Applying the gradient operator ∇ to the second equation in (1) yields
∂ t ∇ θ − κ ∇ Δ θ + ∇ ( u · ∇ θ ) = 0
The first term in (11) is tangential to the boundary, as the boundary condition ∂ θ ∂ n = 0 holds for all t, differentiating in time preserves this property. We now show that the third term is also tangential. Using the vector identity
∇ ( u · ∇ θ ) = u · ∇ 2 θ + ( ∇ θ · ∇ ) u + u × ( ∇ × ∇ θ ) + ∇ θ × ( ∇ × u ) .
For a flat boundary, the conditions u · n = 0 and ∂ θ ∂ n = 0 imply, in local coordinates,
u 3 = ∂ 3 θ = 0 on ∂ Ω .
Consequently,
( ∇ θ · ∇ ) u · n = ∂ 1 θ ∂ 1 u 3 + ∂ 2 θ ∂ 2 u 3 + ∂ 3 θ ∂ 3 u 3 = 0 , on ∂ Ω , u · ∇ 2 θ · n = u 1 ∂ 1 ∂ 3 θ + u 2 ∂ 2 ∂ 3 θ + u 3 ∂ 3 ∂ 3 θ = 0 , on ∂ Ω .
Moreover,
( ∇ θ × w ) · n = − ( w × n ) · ∇ θ = 0 on ∂ Ω ,
since w × n = ( ∇ × u ) × n = 0 on ∂ Ω by the boundary condition.
Then the vector fields u · ∇ 2 θ , ∇ θ · ∇ u and ∇ θ × w are tangent to boundary. Hence each term on the right-hand side of (12) is tangential to the boundary, where we have used ∇ × ∇ θ = 0 , so ∇ ( u · ∇ θ ) is tangential as well. Returning to (11), since ∂ t ∇ θ and ∇ ( u · ∇ θ ) are both tangential, it follows that ∇ Δ θ is also tangential on ∂ Ω , i.e., n · ∇ Δ θ = 0 there. □
Lemma 4.
Under the assumptions of Proposition 1, the following identity holds.
ν χ × n = − ν Δ w × n = ∇ × ( θ e 3 ) × n on ∂ Ω .
Proof. 
Taking ∇ × operator to the first equation of system (1), we have
∂ t w − ν Δ w + ( u · ∇ ) w + ( w · ∇ ) u = ∇ × ( θ e 3 ) .
The boundary condition ∇ × u × n = 0 implies ∂ t w is normal to ∂ Ω , and on the other hand, we know that ( u · ∇ ) w and ( w · ∇ ) u are normal to ∂ Ω from [15]; hence their cross-products with n vanish. Taking the cross-product of (17) with n and using the above observations, we obtain
− ν Δ w × n = ∇ × ( θ e 3 ) × n on ∂ Ω .
Finally, since Δ w = − χ , we have ν χ × n = − ν Δ w × n . Combining this with the previous identity yields the desired result, completing the proof of Lemma 4. □
Proposition 2.
Under the assumptions of Proposition 1, one has
ν ( n · ∇ ) χ · χ = 0 on ∂ Ω .
Proof. 
From (16), we obtain
ν χ 1 = ∂ 2 θ , ν χ 2 = − ∂ 1 θ , on ∂ Ω .
Since ∇ · χ = 0 , together with (19) it follows that
ν ∂ 3 χ 3 = − ν ( ∂ 1 χ 1 + ∂ 2 χ 2 ) = 0 , on ∂ Ω .
Using (13) and (20), we compute
ν ( ∂ i χ j ) n i χ j = ν ( ∂ 3 χ j ) χ j = ν ( ∂ 3 χ 1 χ 1 + ∂ 3 χ 2 χ 2 + ∂ 3 χ 3 χ 3 ) = ∂ 3 ∂ 2 θ χ 1 − ∂ 3 ∂ 1 θ χ 2 + ν ∂ 3 χ 3 χ 3 = 0 , on ∂ Ω .
Thus (18) holds, completing the proof of Proposition 2. □

3. Uniform Estimates

In this section, we discuss the regularity of strong solutions to system (1). In order to establish strong convergence in the next section, we first obtain uniform bounds for the solutions. More precisely, we have the following theorem.
Theorem 1.
Assume that the initial data ( θ 0 , u 0 ) ∈ H 3 ( Ω ) satisfy the boundary conditions in system (1), and that u 0 is divergence-free in Ω. Then there exists a time T ∗ > 0 , independent of κ and ν, such that for any T 0 ∈ [ 0 , T ∗ ) , system (1) admits a unique local strong solution ( θ , u , p ) on [ 0 , T 0 ] satisfying
( θ , u ) ∈ L ∞ ( 0 , T 0 ; H 3 ( Ω ) ) ∩ L 2 ( 0 , T 0 ; H 4 ( Ω ) ) , p ∈ L ∞ ( 0 , T 0 ; H 3 ( Ω ) ) , ( ∂ t θ , ∂ t u ) ∈ L ∞ ( 0 , T 0 ; H 1 ( Ω ) ) ∩ L 2 ( 0 , T 0 ; H 2 ( Ω ) ) ,
and obeying the following uniform bound.
∥ ( θ , u ) ∥ L ∞ ( 0 , T 0 ; H 3 ( Ω ) ) + κ ∥ θ ∥ L 2 ( 0 , T 0 ; H 4 ( Ω ) ) + ν ∥ u ∥ L 2 ( 0 , T 0 ; H 4 ( Ω ) ) ≤ C ,
where the constant C is independent of ν and κ.
Proof. 
The proof is based on the classical energy method. Since the technical challenge of non-vanishing boundary integrals arising from the slip boundary conditions has been resolved in the previous section, we proceed directly to the higher-order energy estimates. By applying ordinary differential equation comparison theory to the resulting differential inequalities, we establish the local existence and the uniform bound. The detailed technical steps are provided below.
Applying the operator ( ∇ × ) 3 to the first equation of system (1), we get
∂ t χ − ν Δ χ + ∑ i , j , k C i , j , k 1 ( ∂ i j u ) ( ∂ k ω ) + ∑ i , j , k C i , j , k 2 ( ∂ i j ω ) ( ∂ k u ) = ( ∇ × ) 3 ( θ e 3 ) ,
where C i , j , k 1 and C i , j , k 2 are constants.
Taking the inner product of (23) with χ and using the Hölder inequality, Young’s inequality, and the boundary condition (18), we have
1 2 d d t ∥ χ ∥ L 2 2 + ν ∥ ∇ χ ∥ L 2 2 = − ∫ Ω ( ∇ × ) 3 ( u · ∇ ) u · χ d x + ∫ Ω ( ∇ × ) 3 ( θ e 3 ) · χ d x + ν ∫ ∂ Ω ( ∂ i χ j ) n i χ j d S ≤ C ∫ Ω ( | D u | | D 2 w | + | D 2 u | | D w | ) | χ | d x + ∫ Ω | ( ∇ × ) 3 ( θ e 3 ) | | χ | d x ≤ ∥ D u ∥ L ∞ ∥ D 2 w ∥ L 2 ∥ χ ∥ L 2 + ∥ | D 2 u | 2 ∥ L 4 ∥ χ ∥ L 2 + ∥ ∇ Δ θ ∥ L 2 ∥ χ ∥ L 2 .
By the Sobolev embedding inequality, we have
∥ D u ∥ L ∞ ≤ C ∥ D 3 u ∥ L 2 ,
and
∥ | D 2 u | 2 ∥ L 2 = ∥ D 2 u ∥ L 4 2 ≤ C ∥ D 3 u ∥ L 2 2 .
Thus, using the equivalence (9) and Young’s inequality, we obtain
1 2 d d t ∥ χ ∥ L 2 2 + ν ∥ ∇ χ ∥ L 2 2 ≤ C ∥ χ ∥ L 2 3 + ∥ ∇ Δ θ ∥ L 2 2 + ∥ χ ∥ L 2 2 .
Next, we apply ∇ Δ to the second equation of system (1) and take the inner product with ∇ Δ θ . Using Hölder inequality, Young’s inequality, (25), (26), and the boundary condition in Proposition 1, we obtain
1 2 d d t ∥ ∇ Δ θ ∥ L 2 2 + κ ∥ Δ 2 θ ∥ L 2 2 = − ∫ Ω ∇ Δ ( u · ∇ θ ) · ∇ Δ θ d x + κ ∫ ∂ Ω ( Δ 2 θ ) n i Δ ∂ i θ d S ≤ C ∫ Ω ( | D 3 u | | D θ | + | D 2 u | | D 2 θ | + | D u | | D 3 θ | ) | ∇ Δ θ | d x ≤ C ( ∥ D 3 u ∥ L 2 ∥ ∇ θ ∥ L ∞ + ∥ D 2 u ∥ L 4 ∥ D 2 θ ∥ L 4 + ∥ D u ∥ L ∞ ∥ ∇ Δ θ ∥ L 2 ) ∥ ∇ Δ θ ∥ L 2 ≤ C ( ∥ ∇ Δ θ ∥ L 2 3 + ∥ χ ∥ L 2 3 ) .
Combining (24) and (28), we get
1 2 d d t ( ∥ ∇ Δ θ ∥ L 2 2 + ∥ χ ∥ L 2 2 ) + κ ∥ Δ 2 θ ∥ L 2 2 + ν ∥ ∇ χ ∥ L 2 2 ≤ ( ∥ ∇ Δ θ ∥ L 2 2 + ∥ χ ∥ L 2 2 ) + C ( ∥ ∇ Δ θ ∥ L 2 2 + ∥ χ ∥ L 2 2 ) 3 2 ,
where C denotes a constant independent of κ and ν .
Define y ( t ) = ∥ ∇ Δ θ ( t ) ∥ L 2 2 + ∥ χ ( t ) ∥ L 2 2 . Then (29) implies
d d t y ≤ y + C y 3 2 .
Consider the ordinary differential equation d d t y ˜ = y ˜ + C y ˜ 3 / 2 with initial value y ˜ ( 0 ) = y ( 0 ) . By comparison, there exists a time T ∗ > 0 , depending only on y ( 0 ) , such that y ( t ) remains bounded for all t ∈ [ 0 , T ∗ ) . Consequently, for any T 0 ∈ ( 0 , T ∗ ) , we have
∥ ( θ , u ) ∥ L ∞ ( 0 , T 0 ; H 3 ( Ω ) ) ≤ C ,
where the constant C is independent of ν and κ . Returning to (29) and integrating over ( 0 , t ) for t ∈ ( 0 , T 0 ) , together with (30), we obtain
κ ∥ θ ∥ L 2 ( 0 , T 0 ; H 4 ( Ω ) ) + ν ∥ u ∥ L 2 ( 0 , T 0 ; H 4 ( Ω ) ) ≤ C .
Taking the divergence of the velocity equation, the pressure satisfies
− Δ p = div ( u · ∇ ) u − θ e 3 .
Standard elliptic regularity yields p ∈ L ∞ ( 0 , T ; H 3 ( Ω ) ) since the right-hand side belongs to L ∞ ( 0 , T ; H 1 ( Ω ) ) . Since the pressure p is uniquely determined by ( θ , u ) via the pressure-Poisson equation, for the sake of brevity, the uniform estimates and convergence results in the subsequent theorems are formulated primarily in terms of the dynamical variables ( θ , u ) .
Finally, from system (1), together with (30) and (31), we deduce
∥ ( ∂ t θ , ∂ t u ) ∥ L ∞ ( 0 , T 0 ; H 1 ( Ω ) ) ≤ C , ∥ ( ∂ t θ , ∂ t u ) ∥ L 2 ( 0 , T 0 ; H 2 ( Ω ) ) ≤ C .
The uniqueness of the strong solution follows from a standard argument. This completes the proof of Theorem 1. □

4. Vanishing Diffusivity Limit

In this section, we discuss the vanishing diffusivity limit. In view of the uniform bounds obtained in Theorem 1, and inspired by [15], we establish the convergence of the strong solution of system (1) to a weak solution of the viscous Boussinesq equations (i.e., the zero diffusivity case):
∂ t u ν − ν Δ u ν + ( u ν · ∇ ) u ν + ∇ p ν = θ ν e 3 in Ω × ( 0 , T ] , ∂ t θ ν + u ν · ∇ θ ν = 0 in Ω × ( 0 , T ] , ∇ · u ν = 0 in Ω × ( 0 , T ] , u ν · n = 0 , n × ∇ × u ν = 0 on ∂ Ω × ( 0 , T ] , ( θ ν , u ν ) | t = 0 = ( θ 0 , u 0 ) in Ω .
The existence of weak solutions for system (33) can be found in [15]. We now present the convergence rate; specifically, we prove the following theorem.
Theorem 2.
Let ( θ , u ) be the strong solution of system (1) given in Theorem 1, and let ( θ ν , u ν ) be a weak solution of system (33). Then there exists a constant C independent of κ such that for all t ∈ [ 0 , T 0 ] ,
∥ θ − θ ν ∥ H 1 2 + ∥ u − u ν ∥ H 1 2 ≤ C κ 2 .
Proof. 
The main challenge here is the low regularity of the weak solution ( θ ν , u ν ) . To avoid requiring higher-order regularity for the weak solution, we perform the error estimates by employing the strong solution’s high regularity from Theorem 1 as a multiplier. By transferring the higher-order derivatives to the strong solution via integration by parts, the error bounds reduce to a Grönwall-type inequality, which yields the rate.
We denote the differences θ ¯ = θ − θ ν , u ¯ = u − u ν , and p ¯ = p − p ν , which satisfy
∂ t u ¯ − ν Δ u ¯ + ( u ν · ∇ ) u ¯ + ∇ p ¯ = − ( u ¯ · ∇ ) u + θ ¯ e 3 in Ω × ( 0 , T ] , ∂ t θ ¯ + ( u ν · ∇ ) θ ˜ = κ Δ θ − ( u ¯ · ∇ ) θ in Ω × ( 0 , T ] , ∇ · u ¯ = 0 in Ω × ( 0 , T ] , u ¯ · n = 0 , n × ∇ × u ¯ = 0 on ∂ Ω × ( 0 , T ] , ( θ ¯ , u ¯ ) | t = 0 = ( 0 , 0 ) in Ω .
Applying the curl operator ∇ × to the first equation of (34) and taking the inner product with w ¯ , we obtain
1 2 d d t ∥ w ¯ ∥ L 2 2 + ν ∥ ζ ¯ ∥ L 2 2 = − ∇ × ( u ¯ · ∇ ) u , w ¯ − ∇ × ( u ν · ∇ ) u ¯ , w ¯ + ∇ × ( θ ¯ e 3 ) , w ¯ .
For the first term on the right-hand side of (35), integration by parts and Hölder’s inequality yield
− ∇ × ( u ¯ · ∇ ) u , w ¯ = − ∫ Ω ( u ¯ · ∇ ) u · ζ ¯ d x ≤ ∥ u ¯ ∥ L 6 ∥ ∇ u ∥ L 3 ∥ ζ ¯ ∥ L 2 ≤ C ∥ w ¯ ∥ L 2 2 ∥ u ∥ L 2 ∥ χ ∥ L 2 + ν 2 ∥ ζ ¯ ∥ L 2 2 ≤ C ∥ w ¯ ∥ L 2 2 + ν 4 ∥ ζ ¯ ∥ L 2 2 .
Similarly, for the second term, we have
− ∇ × ( u ν · ∇ ) u ¯ , w ¯ = − ∫ Ω ( u ν · ∇ ) u ¯ · ζ ¯ d x ≤ ∥ u ν ∥ L ∞ ∥ ∇ u ¯ ∥ L 2 ∥ ζ ¯ ∥ L 2 ≤ C ∥ u ν ∥ L 2 ∥ χ ν ∥ L 2 ∥ w ¯ ∥ L 2 2 + ν 2 ∥ ζ ¯ ∥ L 2 2 ≤ C ∥ w ¯ ∥ L 2 2 + ν 4 ∥ ζ ¯ ∥ L 2 2 .
Finally, for the last term in (35), integration by parts gives
∇ × ( θ ¯ e 3 ) , w ¯ = ∫ Ω ( θ ¯ e 3 ) · ζ ¯ d x ≤ C ∥ θ ¯ ∥ L 2 2 + ν 4 ∥ ζ ¯ ∥ L 2 2 .
Substituting (36)–(38) into (35), we arrive at
1 2 d d t ∥ w ¯ ∥ L 2 2 + ν 4 ∥ ζ ¯ ∥ L 2 2 ≤ C ( ∥ w ¯ ∥ L 2 2 + ∥ θ ¯ ∥ L 2 2 ) .
By applying the gradient operator to the second equation in (34) and taking the inner product with ∇ θ ¯ , we have
1 2 d d t ∥ ∇ θ ¯ ∥ L 2 2 = κ ( ∇ Δ θ , ∇ θ ¯ ) − ∇ ( u ¯ · ∇ ) θ , ∇ θ ¯ − ∇ ( u · ∇ ) θ ¯ , ∇ θ ¯ .
For the first term on the right-hand side of (40), Hölder’s inequality together with the uniform bound from Theorem 1 yields
κ ( ∇ Δ θ , ∇ θ ¯ ) ≤ κ ∥ ∇ Δ θ ∥ L 2 ∥ ∇ θ ¯ ∥ L 2 ≤ C κ ∥ ∇ θ ¯ ∥ L 2 ≤ C ∥ ∇ θ ¯ ∥ L 2 2 + C κ 2 .
For the second term, integration by parts and Hölder’s inequality give
− ∇ ( u ¯ · ∇ ) θ , ∇ θ ¯ ≤ ∥ ∇ ( u ¯ · ∇ ) θ ∥ L 2 ∥ ∇ θ ¯ ∥ L 2 ≤ C ∥ D u ¯ ∥ L 2 ∥ ∇ θ ∥ L ∞ + ∥ u ¯ ∥ L 6 ∥ D 2 θ ∥ L 3 ∥ ∇ θ ¯ ∥ L 2 ≤ C ∥ w ¯ ∥ L 2 ∥ θ ∥ H 3 ∥ ∇ θ ¯ ∥ L 2 ≤ C ∥ w ¯ ∥ L 2 2 + C ∥ ∇ θ ¯ ∥ L 2 2 .
For the third term, we use the vector identity (12). Noting that u · ∇ 2 θ ¯ is orthogonal to ∇ θ ¯ after integration by parts, we obtain
− ∇ ( u · ∇ ) θ ¯ , ∇ θ ¯ = − u · ∇ 2 θ ¯ + ( ∇ θ ¯ · ∇ ) u + u × ( ∇ × ∇ θ ¯ ) + ∇ θ ¯ × ( ∇ × u ) , ∇ θ ¯ = − ( ∇ θ ¯ · ∇ ) u + ∇ θ ¯ × ( ∇ × u ) , ∇ θ ¯ ≤ ∥ w ∥ L ∞ ∥ ∇ θ ¯ ∥ L 2 2 ≤ C ∥ u ∥ H 3 ∥ ∇ θ ¯ ∥ L 2 2 .
Putting (41)–(43) into (40), we arrive at
1 2 d d t ∥ ∇ θ ¯ ∥ L 2 2 ≤ C ( ∥ w ¯ ∥ L 2 2 + ∥ ∇ θ ¯ ∥ L 2 2 ) + C κ 2 .
Adding (39) and (44) yields
1 2 d d t ( ∥ ∇ θ ¯ ∥ L 2 2 + ∥ w ¯ ∥ L 2 2 ) + ν 4 ∥ ζ ¯ ∥ L 2 2 ≤ C ( ∥ w ¯ ∥ L 2 2 + ∥ ∇ θ ¯ ∥ L 2 2 ) + C κ 2 .
Applying Grönwall’s inequality, we get the convergence rate
∥ ∇ θ ¯ ∥ L 2 2 + ∥ w ¯ ∥ L 2 2 + ν 4 ∫ 0 t ∥ ζ ¯ ( s ) ∥ L 2 2 d s ≤ C ( T 0 ) κ 2 .
Thus, the proof of Theorem 2 is complete. □
Remark 1.
Theorem 2 shows that the unique strong solution ( θ , u ) of the full system converges strongly in H 1 to a weak solution ( θ ν , u ν ) of the system (1) as κ → 0 . Since ( θ , u ) itself belongs to L ∞ ( 0 , T ; H 3 ) , the strong convergence implies that ( θ ν , u ν ) inherits the H 1 regularity. In fact, for smooth initial data ( θ 0 , u 0 ) ∈ H 3 , the limiting system admits a unique strong solution, and the convergence rate obtained above confirms that the weak solution obtained via the vanishing diffusivity limit coincides with this strong solution. Thus, while weak solutions of the system (33) may exist with only L 2 regularity, the particular solution selected by the limit process possesses higher regularity due to the strong convergence from the heat-conductive approximation.

5. Vanishing Diffusivity and Viscosity Limit

In this section, we discuss the vanishing diffusivity and viscosity limit. In view of the uniform bounds obtained in Theorem 1 and inspired by [26], we establish the convergence of the strong solution of system (1) to the corresponding Euler–Boussinesq system:
∂ t u E + ( u E · ∇ ) u E + ∇ p E = θ E e 3 in Ω × ( 0 , T ] , ∂ t θ E + u E · ∇ θ E = 0 in Ω × ( 0 , T ] , ∇ · u E = 0 in Ω × ( 0 , T ] , u E · n = 0 on ∂ Ω × ( 0 , T ] , ( θ E , u E ) | t = 0 = ( θ 0 , u 0 ) in Ω .
Theorem 3.
Let ( θ κ , ν , u κ , ν ) be the strong solution of system (1) given by Theorem 1 on [ 0 , T 0 ] . Then, as ν , κ → 0 , ( θ κ , ν , u κ , ν ) converges to the unique solution ( θ E , u E ) of system (47) in the sense that
( θ κ , ν , u κ , ν ) → ( θ E , u E ) i n L 2 ( 0 , T 0 ; H 3 ( Ω ) ) ∩ C ( [ 0 , T 0 ] ; H 2 ( Ω ) ) ,
and the following uniform bounds hold:
∥ ( θ E , u E ) ∥ L ∞ ( 0 , T 0 ; H 3 ( Ω ) ) ≤ C , ∥ ( ∂ t θ E , ∂ t u E ) ∥ L ∞ ( 0 , T 0 ; H 1 ( Ω ) ) ≤ C .
Proof. 
The proof relies on the standard compactness argument, specifically, the Aubin–Lions lemma. Based on the uniform bounds from Theorem 1, we first extract a weakly or weakly-* convergent subsequence. The uniform bounds on the time derivatives allow us to apply the Aubin–Lions compactness theorem to obtain strong convergence in the desired space. We then pass to the limit in the nonlinear terms. Finally, since the limit Euler–Boussinesq system admits a unique strong solution, a standard subsequence argument guarantees the convergence of the entire sequence.
From Theorem 1, it is easy for us to know that the family ( θ κ , ν , u κ , ν ) κ , ν > 0 is uniformly bounded in L ∞ ( 0 , T 0 ; H 3 ( Ω ) ) ∩ L 2 ( 0 , T 0 ; H 4 ( Ω ) ) , and the time derivatives are uniformly bounded in L ∞ ( 0 , T 0 ; H 1 ( Ω ) ) ∩ L 2 ( 0 , T 0 ; H 2 ( Ω ) ) . By classical compactness arguments, there exists a subsequence ( κ n , ν n ) → 0 and a limit ( θ E , u E ) such that the convergence in (48) holds. Passing to the limit in the weak formulation yields that ( θ E , u E ) is a strong solution of (47), and the bounds in (49) follow from the uniform bounds. Uniqueness of the strong solution for (47) with H 3 initial data then implies convergence of the entire sequence. □
Theorem 4.
Suppose ( θ , u ) is the strong solution of system (1) from Theorem 1, and ( θ E , u E ) is the corresponding strong solution of system (47). Then there exists a constant C independent of ν and κ such that for all t ∈ [ 0 , T 0 ] ,
∥ θ − θ E ∥ H 2 2 + ∥ u − u E ∥ H 2 2 ≤ C ( κ + ν ) .
Proof. 
The proof relies on constructing energy estimates for the error functions θ ˜ , u ˜ , and p ˜ . Due to the high regularity required by the H 2 norm, we apply the curl operator twice to the velocity equation and the Laplacian to the temperature equation. The main technical challenge lies in handling the viscous and heat conduction terms: integration by parts generates boundary integrals that do not vanish, which we control using the trace theorem and Young’s inequality. The resulting differential inequalities for the error energy are then combined and resolved via Grönwall’s inequality, where the time-integrals of the highest-order derivatives are absorbed by the uniform bounds established in Theorem 1, yielding the sharp rate O ( κ + ν ) .
Define the differences θ ˜ = θ − θ E , u ˜ = u − u E , and p ˜ = p − p E . They satisfy
∂ t u ˜ + ( u E · ∇ ) u ˜ + ∇ p ˜ = ν Δ u − ( u ˜ · ∇ ) u + θ ˜ e 3 in Ω × ( 0 , T ] , ∂ t θ ˜ + ( u E · ∇ ) θ ˜ = κ Δ θ − ( u ˜ · ∇ ) θ in Ω × ( 0 , T ] , ∇ · u ˜ = 0 in Ω × ( 0 , T ] , u ˜ · n = 0 on ∂ Ω × ( 0 , T ] , ( θ ˜ , u ˜ ) | t = 0 = ( 0 , 0 ) in Ω .
Let w ˜ = ∇ × u ˜ , ζ ˜ = ∇ × w ˜ . Applying ( ∇ × ) 2 to the first equation in (51) and taking the inner product with ζ ˜ yields
1 2 d d t ∥ ζ ˜ ∥ 2 2 = ν ( Δ ζ , ζ ˜ ) − ( ∇ × ) 2 ( ( u ˜ · ∇ ) u + ( u E · ∇ ) u ˜ ) , ζ ˜ + ( ∇ × ) 2 ( θ ˜ e 3 ) , ζ ˜ .
For the viscous term, by integrating by parts, we use the trace theorem to obtain
ν ( Δ ζ , ζ ˜ ) = − ν ∫ Ω χ · χ ˜ d x − ν ∫ ∂ Ω ( χ × n ) · ζ ˜ d x ≤ ν ∥ χ ∥ L 2 ∥ χ ˜ ∥ L 2 + ν ∥ χ ∥ L 2 ( ∂ Ω ) ∥ ζ ˜ ∥ L 2 ( ∂ Ω ) ≤ C ( ν + ν ∥ χ ∥ L 2 1 2 ∥ ∇ χ ∥ L 2 1 2 ∥ ζ ˜ ∥ L 2 1 2 ∥ ∇ ζ ˜ ∥ L 2 1 2 ) ≤ C ( ν + ν 4 3 ∥ ∇ χ ∥ L 2 2 3 + ∥ ζ ˜ ∥ L 2 2 ) ≤ C ( ν + ν 2 ∥ ∇ χ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) .
where we use Young’s inequality, the trace theorem and the uniform bound on | χ | L 2 from Theorem 1.
For the nonlinear terms, using Sobolev embeddings and the uniform bounds from Theorem 1 (and similarly for ( θ E , u E ) from Theorem 3), we obtain
− ( ∇ × ) 2 ( ( u ˜ · ∇ ) u + ( u E · ∇ ) u ˜ ) , ζ ˜ ≤ C ( ∑ i = 0 2 | D i u ˜ | | D 3 − i u | + ∑ i = 1 2 | D i u E | | D 3 − i u ˜ | , | ζ ˜ | ) ≤ C ( ∥ D 3 u ∥ L 2 + ∥ D 3 u E ∥ L 2 ) ∥ ζ ˜ ∥ L 2 2 ≤ C ∥ ζ ˜ ∥ L 2 2 .
By direct estimation, we obtain
( ∇ × ) 2 ( θ ˜ e 3 ) , ζ ˜ ≤ C ( ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) .
Collecting with (53)–(55) together, from (52), we get
1 2 d d t ∥ ζ ˜ ∥ 2 2 ≤ C ( ν + ν 2 ∥ ∇ χ ∥ L 2 2 + ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) .
Next, by applying the operator Δ to the second equation in (51), and taking the inner product with Δ θ ˜ , we have
1 2 d d t ∥ Δ θ ˜ ∥ 2 2 = κ ( Δ 2 θ , Δ θ ˜ ) − Δ ( ( u ˜ · ∇ ) θ + ( u E · ∇ ) θ ˜ ) , Δ θ ˜ .
For the heat conduction term, integration by parts yields
κ ( Δ 2 θ , Δ θ ˜ ) = − κ ∫ Ω ∇ Δ θ · ∇ Δ θ ˜ d x ≤ κ ∥ ∇ Δ θ ∥ L 2 ∥ ∇ Δ θ ˜ ∥ L 2 ≤ C κ .
For the convective terms, similarly we obtain
− Δ ( ( u ˜ · ∇ ) θ + ( u E · ∇ ) θ ˜ ) , Δ θ ˜ ≤ C ( ∑ i = 0 2 | D i u ˜ | | D 3 − i θ | + ∑ i = 1 2 | D i u E | | D 3 − i θ ˜ | , | Δ θ ˜ | ) ≤ C ( ∥ D 3 θ ∥ L 2 + ∥ D 3 u E ∥ L 2 ) ( ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) ≤ C ( ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) .
From (57)–(59), we have
1 2 d d t ∥ Δ θ ˜ ∥ 2 2 ≤ C ( κ + ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) .
Adding (56) and (60) yields
1 2 d d t ( ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ) ≤ C ( κ + ν + ∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 + ν 2 ∥ ∇ χ ∥ L 2 2 ) .
Applying Grönwall’s inequality and (31), we get the convergence rate
∥ Δ θ ˜ ∥ L 2 2 + ∥ ζ ˜ ∥ L 2 2 ≤ C ( κ + ν ) .
Thus, we complete the proof of Theorem 4. □

6. Conclusions

In this paper, we established uniform H 3 estimates for strong solutions to the three-dimensional Boussinesq system with slip boundary conditions, and proved the joint vanishing limit of viscosity ν and thermal diffusivity κ .
Comparison with Navier–Stokes limits. While the vanishing viscosity limit is classical for the Navier–Stokes equations, the Boussinesq system introduces a more complex structure through the nonlinear coupling between velocity and temperature. This coupling generates additional nonlinear interactions in the higher-order energy estimates, making boundary control significantly more challenging. Unlike classical Navier–Stokes studies dealing with a single parameter, our results address the joint limit of two independent parameters. We proved that a tailored boundary-layer analysis effectively controls the boundary integrals, rigorously justifying the inviscid Euler–Boussinesq system as an approximation, thereby extending the classical vanishing viscosity theory to thermally coupled models.
Numerical and geophysical implications. In geophysical fluid dynamics (e.g., atmospheric dynamics and mantle convection), viscosity and diffusivity are typically extremely small. Our uniform H 3 bound ensures numerical stability as ( ν , κ ) → 0 , while the explicit convergence rate of O ( κ + ν ) in H 2 provides a rigorous theoretical error bound. This theoretically justifies the numerical practice of using idealized inviscid models for large-scale simulations.
Intuitive interpretation. To make the results more accessible to readers outside the pure PDE community, we briefly interpret the core assumptions and estimates. The H 3 initial data and divergence-free condition ( ∇ · u 0 = 0 ) ensure sufficient smoothness and incompressibility of the flow. The slip boundary conditions ( u · n = 0 and ω × n = 0 ) physically mean that the fluid cannot penetrate the boundary but can slide along it without friction (e.g., free-surface flows). Most importantly, the uniform estimate (22) guarantees that the high-order derivatives remain bounded independently of the dissipation parameters ( κ , ν ) , indicating a stable boundary layer even in the vanishing viscosity/diffusivity limit.
Future research. A natural extension is to general bounded domains, which poses significant mathematical challenges due to the geometric complexity and boundary-layer effects. In contrast, extending the results to the whole space R 3 is mathematically simpler, as it circumvents the boundary-layer analysis entirely. Exploring related coupled systems or higher-dimensional settings also presents interesting avenues for future study.

Author Contributions

H.L.: Conceptualization, Methodology, Formal Analysis, Writing—Original Draft. J.W.: Validation, Writing—Review & Editing, Supervision. All authors have read and agreed to the published version of the manuscript.

Funding

This work is supported by the Yunnan Provincial Department of Education Science Research Fund Project (Grant No. 2025J1070), Pu’er University Yunnan Circular Agriculture Industry Research Institute Industry-Academia-Research Collaboration Project (Grant No. 2025PEUXJ04), Special Basic Cooperative Research Programs of Yunnan Provincial Undergraduate Universities’ Association (Grant No. 202501BA070001-121) and Yunnan Provincial Xingdian Talent Program: Young Talent Specific Project (Grant No. XDYC-QNRC-2022-0038).

Institutional Review Board Statement

Not applicable.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Acknowledgments

The authors would like to express his great gratitude to Yuelong Xiao.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Morimoto, H. Nonstationary Boussinesq equations. J. Fac. Sci. Univ. Tokyo Sect. IA Math. 1992, 39, 61–75. [Google Scholar]
  2. Lorca, S.; Boldrini, J. The initial value problem for a generalized Boussinesq model. Nonlinear Anal. 1999, 36, 457–480. [Google Scholar] [CrossRef] [Scilit]
  3. Hou, T.; Li, C. Global well-posedness of the viscous Boussinesq equations. Discret. Cont. Dyn. Syst. A 2005, 12, 1–12. [Google Scholar] [CrossRef] [Scilit]
  4. Danchin, R.; Paicu, M. Existence and uniqueness results for the Boussinesq system with data in Lorenz spaces. Phys. D Nonlinear Phenom. 2008, 237, 1444–1460. [Google Scholar] [CrossRef] [Scilit]
  5. de Almeida, M.; Ferreira, L. On the well posedness and large-time behavior for Boussinesq equations in Morrey spaces. Diff. Integral Eqs. 2011, 24, 719–742. [Google Scholar] [CrossRef] [Scilit]
  6. Villamizar-Roa, E.; Ortega-Torres, E. On a generalized Boussinesq model around a rotating obstacle: Existence of strong solutions. Discret. Contin. Dyn. Syst. 2011, 15, 825–847. [Google Scholar] [CrossRef] [Scilit]
  7. Ferreira, L.; Villamizar-Roa, E. Strong solutions and inviscid limit for Boussinesq system with partial viscosity. Commun. Math. Sci. 2013, 11, 421–439. [Google Scholar] [CrossRef] [Scilit]
  8. Hassainia, Z.; Hmidi, T. On the inviscid Boussinesq system with rough initial data. J. Math. Anal. Appl. 2015, 430, 777–809. [Google Scholar] [CrossRef] [Scilit]
  9. Liu, H.; Gao, H. Global well-posedness and long time decay of the 3D Boussinesq equations. J. Differ. Equ. 2017, 263, 8649–8665. [Google Scholar] [CrossRef] [Scilit]
  10. Wang, S.; Wang, Y.; Liu, J. Regularity criteria to the incompressible axisymmetric Boussinesq equations. Appl. Math. Lett. 2021, 112, 106800. [Google Scholar] [CrossRef] [Scilit]
  11. Ye, Z. Global well-posedness for a model of 2D temperature-dependent Boussinesq equations without diffusivity. J. Differ. Equ. 2021, 271, 107–127. [Google Scholar] [CrossRef] [Scilit]
  12. Huang, J.; Wei, Z.; Yao, Z. A Lagrangian approach for inhomogeneous Boussinesq equations. Math. Methods Appl. Sci. 2020, 43, 3556–3568. [Google Scholar] [CrossRef] [Scilit]
  13. Bleitner, F.; Carlson, E.; Nobili, C. Large-time behavior of the 2D thermally non-diffusive Boussinesq equations with Navier-slip boundary conditions. Z. Angew. Math. Phys. 2025, 76, 58. [Google Scholar] [CrossRef] [Scilit]
  14. Li, J. Asymptotic stability to semi-stationary Boussinesq equations without thermal conduction. J. Math. Phys. 2024, 65, 083102. [Google Scholar] [CrossRef] [Scilit]
  15. Berselli, L.; Spirito, S. An elementary approach to the inviscid limits for the 3D Navier-Stokes equations with slip boundary conditions and applications to the 3D Boussinesq equations. NoDEA Nonlinear Differ. Equ. Appl. 2014, 21, 149–166. [Google Scholar] [CrossRef] [Scilit]
  16. Jiang, S.; Zhang, J.; Zhao, J. Boundary-layer effects for the 2-D Boussinesq equations with vanishing diffusivity limit in the half plane. J. Differ. Equ. 2011, 250, 3907–3936. [Google Scholar] [CrossRef] [Scilit]
  17. Wang, J.; Xie, F. Zero dissipation limit and stability of boundary layers for the heat conductive Boussinesq equations in a bounded domain. Proc. Roy. Soc. Edinb. Sect. A 2015, 145, 611–637. [Google Scholar] [CrossRef] [Scilit]
  18. Zhang, Z. Vanishing diffusivity limit for the 3D heat-conductive Boussinesq equations with a slip boundary condition. J. Math. Anal. Appl. 2018, 460, 47–57. [Google Scholar] [CrossRef] [Scilit]
  19. Berselli, L.C.; Spirito, S. On the vanishing viscosity limit of 3D Navier-Stokes equations under slip boundary conditions in general domains. Comm. Math. Phys. 2012, 316, 171–198. [Google Scholar] [CrossRef] [Scilit]
  20. Ferreira, L.C.F.; Planas, G.; Villamizar-Roa, E.J. On the nonhomogeneous Navier-Stokes system with Navier friction boundary conditions. SIAM J. Math. Anal. 2013, 45, 2576–2595. [Google Scholar] [CrossRef] [Scilit]
  21. Guo, B.; Wang, G. Vanishing viscosity limit for the 3D magnetohydrodynamic system with generalized Navier slip boundary conditions. Math. Methods Appl. Sci. 2016, 39, 4526–4534. [Google Scholar] [CrossRef] [Scilit]
  22. Xiao, Y.; Xin, Z. On the vanishing viscosity limit for the 3D Navier-Stokes equations with a slip boundary condition. Comm. Pure Appl. Math. 2007, 60, 1027–1055. [Google Scholar] [CrossRef] [Scilit]
  23. Beirão da Veiga, H.; Crispo, F. Sharp inviscid limit results under Navier type boundary conditions. An Lp theory. J. Math. Fluid Mech. 2010, 12, 397–411. [Google Scholar] [CrossRef] [Scilit]
  24. Beirão da Veiga, H.; Crispo, F. Concerning the Wk,p-inviscid limit for 3-D flows under a slip boundary condition. J. Math. Fluid Mech. 2011, 13, 117–135. [Google Scholar] [CrossRef] [Scilit]
  25. Bourguignon, J.P.; Brezis, H. Remarks on the Euler equation. J. Funct. Anal. 1974, 15, 341–363. [Google Scholar] [CrossRef] [Scilit]
  26. Xiao, Y.; Xin, Z.; Wu, J. Vanishing viscosity limit for the 3D magnetohydrodynamic system with a slip boundary condition. J. Funct. Anal. 2009, 257, 3375–3394. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.