Dynamic boundary conditions with noise for an energy balance model coupled to geophysical flows

Gianmarco Del Sarto, Matthias Hieber, Tarek Zöchling · 2025

Read Dynamic boundary conditions with noise for an energy balance model coupled to geophysical flows by Gianmarco Del Sarto, Matthias Hieber, Tarek Zöchling (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-16449.txt.

Introduction

Gianmarco Del Sarto, Matthias Hieber, and Tarek Zöchling analyze a climate model with noisy boundaries. Their 2025 paper couples an energy balance model to geophysical flow equations. The link runs through a dynamic boundary condition, with and without random noise. Such models mimic how surface energy and ocean motion interact.

The team proves the full system is well-posed in a strong global sense. The proof covers the plain deterministic case for large starting data. It also covers the noisy case in a standard function space. Well-posedness means solutions exist, are unique, and behave stably. That groundwork lets later studies trust simulations built on the model.

Catalog record

CREATORGianmarco Del Sarto, Matthias Hieber, Tarek Zöchling
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE94 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-16449

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Dynamic boundary conditions with noise for an energy balance model coupled to geophysical flows Thanks: All three authors acknowledge the support by the DFG project FOR 5528 and thank Franco Flandoli for his insightful discussions on the topic. Gianmarco Del Sarto, Matthias Hieber, Tarek Zöchling Abstract . This article investigates a Sellers-type energy balance model coupled to the primitive equations by a dynamic boundary condition with and without noise on the boundary. It is shown that this system is globally strongly well-posed both in the deterministic setting for arbitrary large data in W 2 ​ ( 1 − 1 / p ) , p \mathrm{W}^{2(1-\nicefrac{{1}}{{p}}),p} for p ∈ [ 2 , ∞ ) p\in[2,\infty) and in the stochastic setting for arbitrary large data in H 1 \mathrm{H}^{1} . \labelindent =10pt 2020 Mathematics Subject Classification 35Q86, 35Q35, 60H15, 76D03, 35K55 Fachbereich Mathematik, Technische Universität Darmstadt 1. Introduction Consider a two-dimensional energy balance model coupled with the basic equations of geophysical flows, namely, the primitive equations of ocean dynamics. This system integrates an advection-diffusion model for the ocean’s interior temperature (subject to a dynamic boundary condition) with the three-dimensional primitive equations. We study two settings: the deterministic case and the stochastic case, in which the stochastic force is modelled by a cylindrical Wiener process (W_{t})_{t} on the boundary. Both settings are of great interest due to the presence of dynamic boundary conditions and the fact that we deal with unique, global, strong solutions. More specifically, we investigate global, strong well-posedness results for this system, described in the deterministic setting by the set of equations (1.1) \left\{\begin{aligned} \varepsilon\partial_{t}v^{\varepsilon}+u^{\varepsilon}\cdot\nabla v^{\varepsilon}-\Delta v^{\varepsilon}+\nabla_{\mathrm{H}}p^{\varepsilon}&=0,&\quad\text{ in }&\mathcal{O}\times(0,\tau),\\ \partial_{z}p^{\varepsilon}&=-T,&\quad\text{ in }&\mathcal{O}\times(0,\tau)\\ \mathrm{div}\,u^{\varepsilon}&=0,&\quad\text{ in }&\mathcal{O}\times(0,\tau),\\ \partial_{t}T+u^{\varepsilon}\cdot\nabla T-\Delta T&=0,\quad&\text{ in }&\mathcal{O}\times(0,\tau),\\ T|_{\Gamma_{u}}&=\rho,\quad&\text{ in }&G\times(0,\tau),\\ \partial_{t}\rho+v^{\varepsilon}|_{\Gamma_{u}}\cdot\nabla_{\mathrm{H}}\rho-\Delta_{\mathrm{H}}\rho+(\partial_{z}T)|_{\Gamma_{u}}&=R(x,\rho),\quad&\text{ in }&G\times(0,\tau),\\ v^{\varepsilon}(0)&=v_{0},\quad T(0)=T_{0}.\end{aligned}\right. Here, \mathcal{O}=G\times(0,1) denotes the spatial domain, where G=(0,1)^{2} represents the horizontal domain, and where \Gamma_{u}=G\times\{1\} as well as \Gamma_{b}=G\times\{0\} denote the upper (surface) or lower (bottom) boundary of \mathcal{O} , respectively. Furthermore, u^{\varepsilon}=(v^{\varepsilon},w^{\varepsilon}):\mathcal{O}\rightarrow\mathbb{R}^{3} denotes the velocity of the ocean and p^{\varepsilon}:\mathcal{O}\rightarrow\mathbb{R} denotes ocean’s pressure on a climate, slow time-scale (several years), where \varepsilon\sim\nicefrac{{\tau_{\text{fast}}}}{{\tau_{\text{slow}}}} characterises the separation between the fast (weather) and slow (climate) time-scales; see Section 3 for more details. Moreover, T:\mathcal{O}\to\mathbb{R} denotes the temperature of the ocean and \rho=T|_{\Gamma_{u}}:G\to\mathbb{R} the temperature evaluated at the surface. The operators \Delta_{\mathrm{H}} and \nabla_{\mathrm{H}} denote the horizontal Laplacian and the horizontal gradient, that is \Delta_{\mathrm{H}}:=\partial^{2}_{x}+\partial^{2}_{y} and \nabla_{\mathrm{H}}:=(\partial_{x},\partial_{y})^{T} . Moreover, R(x,\rho) is a reaction term modelling the net radiative balance at the surface described in detail below. In the stochastic setting, the force is given through a cylindrical Wiener process (W_{t})_{t} , acting via a given process H_{\rho} on the temperature on the boundary. In this case, the system reads (1.2) \left\{\begin{aligned} \varepsilon\,\mathrm{d}v^{\varepsilon}+u^{\varepsilon}\cdot\nabla v^{\varepsilon}\,\mathrm{d}t-\Delta v^{\varepsilon}\,\mathrm{d}t+\nabla_{\mathrm{H}}p^{\varepsilon}\,\mathrm{d}t&=0,&\quad\text{ in }&\mathcal{O}\times(0,\tau),\\ \partial_{z}p^{\varepsilon}&=-T,&\quad\text{ in }&\mathcal{O}\times(0,\tau)\\ \mathrm{div}\,u^{\varepsilon}&=0,&\quad\text{ in }&G\times(0,\tau),\\ \,\mathrm{d}T+u^{\varepsilon}\cdot\nabla T\,\mathrm{d}t-\Delta T\,\mathrm{d}t&=0,\quad&\text{ in }&\mathcal{O}\times(0,\tau),\\ T|_{\Gamma_{u}}&=\rho,\quad&\text{ in }&G\times(0,\tau),\\ \,\mathrm{d}\rho+\overline{v}^{\varepsilon}\cdot\nabla_{\mathrm{H}}\rho\,\mathrm{d}t-\Delta_{\mathrm{H}}\rho\,\mathrm{d}t+(\partial_{z}T)|_{\Gamma_{u}}\,\mathrm{d}t&=R(x,\rho)\,\mathrm{d}t+H_{\rho}\,\mathrm{d}W,\quad&\text{ in }&G\times(0,\tau),\\ v^{\varepsilon}(0)&=v_{0},\quad T(0)=T_{0},\end{aligned}\right. Note that here the transport term on the boundary is replaced by the transport along the average of the velocity given by \overline{v}^{\varepsilon}=\int_{0}^{1}v^{\varepsilon}(\cdot,\xi)\,\mathrm{d}\xi . Both systems ( 1.1 ) as well as ( 1.2 ) are supplemented by the boundary conditions (1.3) \displaystyle(\partial_{z}v^{\varepsilon})|_{\Gamma_{u}\cup\Gamma_{b}}=w^{\varepsilon}|_{\Gamma_{u}\cup\Gamma_{b}}=0,\hskip 10.00002pt(\partial_{z}T)|_{\Gamma_{b}}=0\ \text{ and }\ u^{\varepsilon},p^{\varepsilon},T,\rho\hskip 10.00002pt\text{ are periodic on }\ \partial G\times[0,1]. Note that the domain G=(0,1)^{2} , when endowed with periodic boundary conditions, can be naturally identified with the torus \mathbb{T}^{2} . Energy balance model serves often as a simplified climate model. For more information on deterministic energy balance models and their applications, we refer e.g., to [ 27 , 5 , 10 , 17 ] . The first pioneering work on stochastic climate models goes back to K. Hasselmann in [ 31 ] , where noise arises as a result of the interaction of the many components constituting the climate system, in particular as the effect of the faster component on the slower ones; see [ 34 ] for a formulation of Hasselmann’s programme in modern mathematical language. For further information on stochastic energy balance models, we refer to [ 26 , 22 , 18 ] . To account for contributions of both slow and fast processes in the noise term, we may consider W=(W^{1},W^{2}), where W^{1} and W^{2} are independent cylindrical Wiener processes, and set H_{\rho}=(H_{\rho}^{1},\sqrt{\varepsilon}H_{\rho}^{2}). Here, H_{\rho}^{1} captures unresolved variability arising from slow processes, while H_{\rho}^{2} , scaled by \sqrt{\varepsilon} , represents the cumulative effect of fast (weather-scale) processes on the slow climate dynamics, according with standard multiscale stochastic modeling approaches; see [ 17 ] . For simplicity of notation and exposition, in the remainder of this work we avoid explicitly considering the two-component formulation of H_{\rho} and treat a single effective noise term. A central aspect in our system is the analysis of the dynamic boundary condition for the surface temperature, given by T|_{\Gamma_{u}}=\rho . Dynamic boundary conditions have, of course, been previously studied, both in the weak and strong sense. We refer here e. g., to the work of Denk, Prüss and Zacher [ 21 ] , who considered dynamic boundary conditions for diffusion, surface diffusion, the linearised Stefan problem, and linearised free boundary value problems for the Navier-Stokes equations. Furthermore, Prüss, Racke, and Zheng [ 45 ] obtained a strong well-posedness result for the Cahn-Hillard equation with dynamic boundary condition. Let us also mention that Díaz and Tello [ 24 ] proved the existence of a weak solution to certain energy balance models, where the fluid is passive , in the sense that the velocity u of the fluid is prescribed and not part of the problem. There is a large literature on stochastic fluid systems or stochastic heat systems with noise or with noise on the boundary. We refer here for example to the work of Cerrai and Freidlin [ 12 ] or Debussche, Fuhrmann and Tessitore [ 16 ] , Brzezniak, Goldys, Peszat and Russo [ 8 ] , Binz, Hieber, Hussein and Saal [ 6 ] . Regarding the primitive equations with additive or multiplicative noise, we refer to [ 30 ] and [ 15 , 29 ] . For further results focusing on linear boundary conditions and solutions with \mathrm{L}^{2}(\mathrm{H}^{1}) regularity, we refer to [ 7 , 48 ] . Note that in contrast to the above cited studies, we consider dynamic boundary conditions in the strong sense subject to noise . The aim of this article is threefold: firstly, we couple the energy balance model with an active fluid, meaning that the ocean’s velocity is not a given function but governed by the coupling with the incompressible primitive equations. More precisely, the coupling is given by a dynamic boundary condition for the temperature on the boundary subject to a transport term on the upper boundary \Gamma_{u} of the form v|_{\Gamma_{u}}\cdot\nabla_{\mathrm{H}}\rho . Here, v denotes the horizontal component of the (unknown) velocity u=(v,w) . Secondly, we aim to prove the existence of a unique, global, strong solution to the system ( 1.1 ) for arbitrary large initial data belonging to \mathrm{W}_{\mathrm{per}}^{2(1-\nicefrac{{1}}{{p}}),p}\times\mathrm{W}_{\mathrm{per}}^{2(1-\nicefrac{{1}}{{p}}),p} for p\in[2,\infty)\setminus\{3\} , satisfying the horizontal solenoidal divergence-free condition. In the special case of p=2 , we thus obtain global strong well-posedness of solutions for arbitrary large data in \mathrm{H}^{1}\times\mathrm{H}^{1} . This is rather surprising, since we would be unable to do so when the geophysical fluid would be replaced, for example, by the Navier-Stokes

Read the complete text