Well-posedness and large deviations for 2D stochastic constrained Navier-Stokes equations driven by Lévy noise in the Marcus canonical form (Q2232180)

\par The authors consider the following constrained stochastic two-dimensional Navier-Stokes equation (NSE) driven by Lévy noise in the Marcus canonical form: \[ \begin{aligned} &d u(t) + \{ A u(t) + B( u(t)) \} dt \\ &= \vert \nabla u(t) \vert_{L^2}^2 u(t) dt + F u(t) \circ d W(t) + \sum_{j=1}^N g_j ( u(t)) \bullet d L_j^{(d)} (t), \quad t \geq 0, \end{aligned} \tag{1} \] where \(A\) is the Stokes operator, \(B(u) := - \Pi[ (u \cdot \nabla) u ]\), and \[ F u(s) \circ d W(s) = \sum_{j=1}^N \Pi [ ( f_j \cdot \nabla) u ](s) \circ d W_j(s). \] There, the periodic boundary condition and the initial condition \(u(0) = u_0\) such that \(\vert u_0 \vert_{L^2} = 1\) are equipped with the above equation (1). \(f_j\) are constant vectors, \(W\) is a \(\mathbb{R}^N\)-valued Wiener process which is the continuous component of the associated Lévy noise \(L\), and \(W_j\) are \(\mathbb R\)-valued i.i.d. standard Brownian motions; \(\circ\) denotes the Stratonovich form. For each \(j=1,2, \dots, N\), \(g_j\) is a bounded linear operator from \(H^1\) to \(H^1\), where \(H^1\) is a Hilbert space with the scalar product given by \[ \langle u, v \rangle_{H^1} := \langle u, v \rangle_{L^2} + \langle \nabla u, \nabla v \rangle_{L^2} \] for \(u, v \in H^1( {\mathbb T}^2 )\). Moreover, \(L^{(d)}\) \(=\) \(( L_1^{(d)}, \dots, L_N^{(d)} )\) denotes the discontinuous component of the associated \({\mathbb R}^N\)-valued Lévy process \(L\), and define \[ L^{(d)}(t) := \int_0^t \int_{ {\mathbb B} } \ell \cdot \tilde{\eta}( ds, d \ell), \] where \(\ell = ( \ell_1, \dots, \ell_N) \in {\mathbb B}\) (= the closed unit ball in \({\mathbb R}^N\) ). When \(\eta\) is the time-homogeneous Poisson random measure associated to \(L\), then the intensity measure of \(\eta\) is equal to \(Leb \otimes \nu\) (where \(\nu\) is the jump intensity measure), and \(\tilde{\eta}\) \(:=\) \(\eta - Leb \otimes \nu\) denotes the corresponding compensated time-homogeneous Poisson random measure. We assume that \(W\) and \(L^{(d)}\) are mutually independent processes. Let \(H\) \(:=\) \(\{ u \in {\mathbb L}_0^2\): \(\nabla \cdot u = 0 \}\), and let \(\Pi\) be the Leray-Helmholtz projection : \(L^1( {\mathbb T}^2, {\mathbb R}^2) \to H\). Here, \(\bullet\) denotes the Marcus product. \par Here is the important key point about the reason why the Marcus product \(\bullet\) is employed in the above equation. The analysis of noise in this article is motivated by the requirement that the noise must preserve the invariance property under coordinate transformation and this issue is extremely important for the norm-preserving condition. Such invariance property is known in the literature for Stratonovich integrals. However, one needs an analogue of the Stratonovich integral in the case of stochastic integral with respect to compensated Poisson random measure. All about this requires the Marcus product technique here. \par In the first part of the paper, the authors show the existence and uniqueness of smooth pathwise solutions (martingale solution) to (1), and further discuss the existence of a strong solution. Those can be realized based upon the construction relying on Faedo-Galerkin approximations, compactness method and the Skorokhod representation theorem. However, significant difficulty and novelty of the problem arise due to the presence of the discontinuous component of the corresponding Lévy noise in the Marcus form. In the second part of the paper, the authors prove Wentzell-Freidlin-type large deviation principle (LDP) for the small noise asymptotic of solutions to (1) using the weak convergence method. For other related works, see, e.g., [\textit{E. Motyl}, Potential Anal. 38, No. 3, 863--912 (2013; Zbl 1282.35282)] for stochastic Navier-Stokes equations driven by Lévy noise, [\textit{Z. Brzeźniak} et al., J. Differ. Equations 264, No. 4, 2833--2864 (2018; Zbl 1379.35209)] for 2D constrained Navier-Stokes equations, and [\textit{A. Budhiraja} et al., Stochastic Processes Appl. 123, No. 2, 523--560 (2013; Zbl 1259.60065)] for large deviations for stochastic partial differential equations driven by a Poisson random measure.