Almost global existence for the stochastic Navier-Stokes equations with small $H^{1/2}$ data
Mustafa Sencer Aydın, Igor Kukavica, Fanhui Xu
TL;DR
This work studies global existence for the stochastic Navier-Stokes equations with multiplicative noise on $\mathbb{T}^3$ for initial data in $H^{1/2}$, achieving almost global-in-time solutions with high probability when data and noise are small. A key innovation is an infinite decomposition of the initial data into smoother components $u_0=\sum_{k\ge0} v_0^{(k)}$, enabling a cascade of SNSE-type problems with cutoffs and a fixed-point scheme to control the nonlinear advection via $H^{\frac12+\delta}$ regularity. The authors prove global energy bounds and probabilistic stopping-time controls, then pass to the limit to obtain a probabilistically strong solution up to a stopping time $\tau$, with $\mathbb{P}(\tau=\infty)\ge 1-p_0$ under small data/noise, and a local-in-time existence result when the noise is not necessarily small. The approach leverages the stochastic heat equation as a core auxiliary tool to manage regularity and energy, offering a novel pathway around criticality for SNSE with multiplicative noise. These results advance understanding of stochastic fluid models at critical regularity and provide quantitative probabilistic guarantees for global behavior in low-regularity regimes.
Abstract
We address the global existence of solutions to the stochastic Navier-Stokes equations with multiplicative noise and with initial data in $H^{1/2}(\mathbb{T}^{3})$. We prove that the solution exists globally in time with probability arbitrarily close to~$1$ if the initial data and noise are sufficiently small. If the noise is not assumed to be small, then the solution is global on a sufficiently small deterministic time interval with probability arbitrarily close to~$1$.
