Local well-posedness of strong solutions to the non-isentropic compressible primitive equations with vertical diffusion

1. Problem Statement

The paper addresses the mathematical analysis of the non-isentropic compressible primitive equations (PEs), which are fundamental to modern meteorology and large-scale atmospheric modeling. These equations are derived from the compressible Navier-Stokes equations by replacing the vertical momentum balance with the hydrostatic balance equation.

Key Challenges:

• Loss of Derivative: Unlike the full Navier-Stokes equations, the PEs lack an evolutionary equation for the vertical velocity component ($w$). Instead, $w$ must be recovered from the hydrostatic balance and the continuity equation. This process results in a loss of one spatial derivative, making the system mathematically more singular and the nonlinearities stronger.
• Non-Isentropic Complexity: While previous works (by the same authors and others) have addressed the isentropic case (constant entropy) or incompressible PEs, the non-isentropic case introduces a temperature equation with heat conduction. This adds significant complexity due to the coupling between density, pressure, and temperature, and the lack of dissipation in the pressure equation.
• Specific Setting: The authors consider the system on a horizontally periodic channel $\mathcal{O} = \mathbb{T}^2 \times (0,1)$ with only vertical diffusion for the temperature ($\kappa \partial_{zz} \theta$) and no gravity term. The boundary conditions are no-slip for velocity and Neumann (insulating) for temperature at the top and bottom boundaries.

2. Methodology

The authors employ a rigorous analytical approach combining parabolic regularization, a priori energy estimates, and fixed-point arguments.

A. Reformulation of the System

To handle the missing vertical momentum equation, the authors derive an explicit representation of the vertical velocity $w$ in terms of horizontal velocity $v$, density $\rho$ (or specific volume $\sigma = 1/\rho$), and pressure $p$.

• They decompose variables into vertical averages ($\bar{f}$) and fluctuations ($\tilde{f}$).
• By integrating the continuity and thermodynamic equations, they express $w$ as:
  $$w = \nu \partial_z \sigma - \int_0^z \phi(v, p) \, dz'$$
  where $\phi$ involves horizontal divergences and nonlinear terms.
• This allows them to reformulate the original system (1.1) into an equivalent system (1.14)–(1.17) involving only $v$, $\sigma$, and $p$, where $w$ is a functional of these variables.

B. Parabolic Regularization

Since the original system is not parabolic in the horizontal directions for density and pressure, the authors introduce a regularized system (2.7)–(2.10) by adding artificial horizontal diffusion terms $\epsilon \Delta_h \sigma$ and $\epsilon \Delta_h p$.

• This regularized system is semi-linear and parabolic, allowing for the application of standard local existence theory for smooth initial data.
• The goal is to establish $\epsilon$-independent a priori estimates to pass to the limit $\epsilon \to 0$.

C. Energy Estimates and Regularity Analysis

The core of the proof lies in deriving high-order energy estimates that do not depend on the regularization parameter $\epsilon$.

• Regularity Requirements: The authors identify that to propagate $H^k$ regularity, the initial data must satisfy specific conditions:
  • Horizontal velocity $v_0 \in H^3(\mathcal{O})$.
  • Specific volume $\sigma_0 \in H^2(\mathcal{O})$.
  • Pressure $p_0 \in H^3(\mathbb{T}^2)$.
  • Reasoning: The pressure equation lacks dissipation, requiring higher regularity ($H^3$) to control the nonlinear terms involving $\nabla_h \Delta_h p$. Similarly, the density equation requires $H^2$ to handle the vertical diffusion term $\partial_{zz} \sigma$.
• Key Estimates:
  • $L^2$ and $H^1$ Estimates: Establishing bounds for the solution and its first derivatives.
  • Higher-Order Estimates: Using commutator estimates (Kato-Ponce type) and Sobolev inequalities to bound terms like $\nabla_h \Delta_h p$ and $\Delta^2 v$.
  • Lower Bounds: Proving that density and pressure remain strictly positive ($\sigma \geq \sigma_0 e^{-Ct}$, $p \geq p_0 e^{-Ct}$) using the comparison principle, which is crucial for the well-posedness of the equations.
  • Time Derivative Estimates: Estimating $\partial_t v, \partial_t \sigma, \partial_t p$ to ensure compactness for the limit process.

D. Existence and Uniqueness

• Existence: Using the uniform a priori estimates, the authors apply the Banach-Alaoglu theorem and the Aubin-Lions compactness lemma to extract a convergent subsequence from the regularized solutions as $\epsilon \to 0$. The limit is shown to satisfy the original system in the strong sense.
• Uniqueness and Continuous Dependence: They perform energy estimates on the difference between two solutions. A critical difficulty is the absence of dissipation in the pressure equation. To overcome this, they perform an $H^1$ energy estimate on the pressure difference and an $H^1$ estimate on the velocity difference, carefully balancing the terms using the lower bounds of density and pressure.

3. Key Contributions

1. First Local Well-Posedness for Non-Isentropic Compressible PEs: This is the first rigorous proof of local existence and uniqueness of strong solutions for the non-isentropic compressible primitive equations with vertical diffusion. Previous results were limited to isentropic or incompressible cases.
2. Handling the "Derivative Loss": The paper successfully manages the loss of one spatial derivative inherent in the hydrostatic approximation by reformulating the system and utilizing the specific structure of the vertical velocity recovery.
3. Optimal Regularity Conditions: The authors establish the precise regularity requirements for initial data ($v_0 \in H^3, \sigma_0 \in H^2, p_0 \in H^3$) necessary to close the energy estimates, highlighting the interplay between the lack of horizontal viscosity in the density/pressure equations and the need for higher initial smoothness.
4. Technical Innovation in Estimates: The derivation of $\epsilon$-independent estimates for a system with mixed dissipation (vertical for density/temperature, none for pressure) and strong nonlinearities represents a significant technical advancement in the analysis of geophysical fluid dynamics.

4. Main Results

Theorem 1.1: Given initial data $(v_0, \sigma_0, p_0)$ with the specified regularity ($H^3, H^2, H^3$) and strictly positive lower bounds, there exists a time $T_0 > 0$ such that the system admits a unique strong solution $(v, \sigma, p)$ on $[0, T_0]$.

• The solution satisfies:
  • $v \in C([0, T_0]; H^2) \cap L^\infty(0, T_0; H^3) \cap L^2(0, T_0; H^4)$.
  • $\sigma \in C([0, T_0]; H^1) \cap L^\infty(0, T_0; H^2)$.
  • $p \in C([0, T_0]; H^2) \cap L^\infty(0, T_0; H^3)$.
• The solution depends continuously on the initial data.
• The density and pressure remain strictly positive throughout the evolution.

5. Significance

• Mathematical Rigor: The paper provides a solid theoretical foundation for the compressible primitive equations, which are widely used in weather prediction but mathematically challenging due to the hydrostatic approximation.
• Atmospheric Modeling: By including non-isentropic effects (heat conduction), the results are more applicable to realistic atmospheric scenarios where temperature variations drive dynamics, unlike the simplified isentropic models.
• Future Directions: The techniques developed here (handling derivative loss, mixed dissipation, and high-order energy estimates) pave the way for studying global existence, long-time behavior, and the low Mach number limit for non-isentropic flows, which are noted as open problems for future research.

In summary, this paper resolves a fundamental open problem in the mathematical theory of geophysical fluids by proving that the non-isentropic compressible primitive equations are locally well-posed, despite the inherent singularities introduced by the hydrostatic balance.