How is the free surface influence transported in turbulent open channel flows?

Yoshiyuki Sakai 1, Christian Bauer 2 · 2026

Read How is the free surface influence transported in turbulent open channel flows? by Yoshiyuki Sakai 1, Christian Bauer 2 (2026) 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-2606-24630.txt.

Introduction

Sakai and Bauer trace how a water surface's influence travels downward through turbulent channel flow. Open channels, rivers and canals in the laboratory, differ from closed ducts. The free surface lets turbulence breathe. Using matched direct numerical simulations of open and closed channels at high Reynolds number, in a domain large enough for very large scale motions. The pair dissect the turbulent kinetic energy budget term by term in 2026.

The surface's influence turns out to be communicated primarily through transport terms, rather than local production. Near the surface, pressure transport supplies energy toward the interface while turbulent transport and dissipation fall, and the surplus is exported by viscous diffusion. No single similarity scaling organises all the near surface budget terms: viscous diffusion follows a viscous length scale, dissipation a Kolmogorov scale. Sakai and Bauer's accounting tells modellers which mechanisms a free surface turbulence closure must represent to be believed.

Catalog record

CREATORYoshiyuki Sakai 1, Christian Bauer 2
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE65 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.24630

Browse the full library · About the Tower of Babel

Full text

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

How is the free surface influence transported in turbulent open channel flows? Yoshiyuki Sakai 1, Christian Bauer 2 Abstract keywords To be added during the typesetting process. † † affiliation: 1 Technical University of Munich, TUM School of Engineering and Design, TUM Department of Civil, Geo and Environmental Engineering, Arcisstr. 21, 80333 Munich, Germany 2 German Aerospace Center (DLR), Institute of Aerodynamics and Flow Technology, Bunsenstr. 10, 37073 Göttingen, Germany We investigate how the influence of a free surface is transported in turbulent open channel flow by analysing matched open- and closed-channel direct numerical simulations up to \Rey_{\tau}\approx 900 in a domain large enough to accommodate very-large-scale motions (VLSMs). The turbulent kinetic energy (TKE) budget shows that the surface influence is communicated primarily through transport terms. Near the free surface, pressure transport supplies energy towards the interface, whereas turbulent transport and dissipation are reduced; the resulting energy surplus is exported away from the surface predominantly by viscous diffusion. The near-surface budget terms do not exhibit a single universal similarity scaling: viscous diffusion is organised over the near-surface viscous scale \ell_{\mathrm{V}} , dissipation over the Kolmogorov sublayer scale \ell_{\mathrm{K}} , and pressure-related terms require the mixed velocity scale u_{\mathrm{b}}u_{\tau}^{2}/h . The pressure–strain redistribution further reveals outer–inner coupling: although intense pressure–strain events remain small-scale, their magnitude and directional bias are organised by low-velocity VLSM streaks. The free-surface influence is therefore best understood as a coupled multi-scale process involving local kinematic constraints, Reynolds-number-dependent surface layers, and outer-layer coherent motions. 1 Introduction Turbulent open channel flow (OCF) is a canonical configuration for many free-surface flows, thus of direct relevance in natural science and engineering applications. The OCF configuration features a free-slip rigid-lid on the upper boundary, which is much less kinematically constrained than the no-slip wall on the bottom boundary. Conversely, the closed channel flow configuration (CCF, also known as plane Poiseuille flows) features no-slip walls on both boundaries. This boundary condition difference significantly modifies turbulence dynamics near the upper boundary, and introduces measurable changes over the full channel depth ( 4 ) . The free-slip boundary constrains wall-normal motion only, and it is known to alter the turbulent kinetic energy (TKE) redistribution pathways via the pressure-strain term ( 22 ; 14 ; 9 ; 13 ; 17 ) . If we view this localised modification in the redistribution mechanism as the entry point of the difference between the two channel types, the resulting near-surface effects need to be communicated into and from the channel interior, and it could be possible to trace the energy flow by means of the transport terms in the TKE budget equation. In this view, the energy budgets can serve as essential diagnostics, since the mean and the second-order statistics alone cannot identify mechanisms. Mean streamwise velocity distributions in the two channel types are qualitatively similar, except that a clear wake region is absent in OCF ( 22 ) . Quantitatively, a significantly lower von Kàrmàn constant in law of the wall was reported for OCF ( 19 ) . Whilst the distributions of wall-normal and the spanwise components of Reynolds normal stresses between the two channel types differ only in the near-surface region, the streamwise component exceeds the corresponding CCF level almost throughout the entire channel depth ( 19 ; 4 ) . This accumulation of energy in the streamwise component is associated with emergence of very-large-scale motions (VLSMs) ( 5 ; 7 ; 8 ; 18 ) , which were found to be even larger and more energetic than the CCF counterparts ( 24 ; 4 ) . This unique feature demands notably larger numerical domains for OCF to accommodate the enlarged VLSMs. In CCF, the emergence of VLSMs is responsible for the failure of the wall scaling (i.e. similarity with respect to wall-friction-based velocity) in the streamwise component of Reynolds normal stress ( 10 ) and appearance of a mixed-velocity scaling ( 23 ) . In OCF, more energetic VLSMs cause the streamwise turbulent intensity to be scaled purely with bulk velocity away from the wall and the free-surface ( 2 ) . By revisiting the pioneering work of 6 , our recent contribution ( 4 ) showed that the surface influence appears through a four-layer structure attached to the free surface. The thickness of the top two layers is Reynolds-number dependent, implying that a near-surface spatial resolution comparable to the level used near the no-slip wall is required to resolve them truthfully. Since the inception of the direct numerical simulation of CCF by 12 , the corresponding TKE budgets have been investigated repeatedly ( 15 ) . Despite the long history, however, Reynolds-number dependence of the term-by-term one-point budgets is only sparsely documented ( 11 , e.g.) . Moreover, such investigation is even rarer for the OCF type ( 22 ) , and the budgets near the free surface are particularly under-characterised. This motivates a systematic budget-level comparison of OCF and CCF across Reynolds number, with particular emphasis on the near-surface region. Based on the above, we define the following research questions (RQs): 1. Which budget terms transmit the surface influence into the channel interior? 2. Does a unified similarity scaling exist for the budget terms? 3. Are pressure-driven inter-component transfer events purely local? Or are they macroscopically organised by outer-layer large-scale motions? We address these questions using matched OCF/CCF DNS data ( 3 ) up to friction Reynolds number \Rey_{\tau}=900 in a large domain that accommodates VLSMs. We quantify the full TKE transport budgets, identify term-specific near-surface similarity scalings, and connect pressure–strain redistribution to VLSM organisation via conditional averaging. Those results can provide useful benchmarks and constraints for future modelling efforts. The remainder of this paper is organised as follows. Section 2 describes the DNS database, numerical method, and near-surface layer structure used in the analysis. Section 3 presents the results, beginning with the OCF/CCF comparison of the TKE budgets (§ 3.1 ), followed by the similarity analysis of the near-surface budget terms (§ 3.2 ), and the VLSM organisation of pressure–strain redistribution (§ 3.3 ). Section 4 discusses the resulting physical picture and its implications for turbulence modelling, and § 5 concludes the paper. 2 Methodology 2.1 Numerical methods and DNS database This study analyses our open-source DNS database ( 3 ) , previously investigated in 2 ; 4 . Nevertheless, we summarise the numerical methods and the case specifications for completeness as follows. The incompressible Navier–Stokes equations are solved in a wall-normal velocity–vorticity formulation using a pseudo-spectral method corresponding to the one employed by 12 , in which the flow variables are expanded in Fourier series in the homogeneous streamwise ( x ) and spanwise ( z ) directions over equidistant grid points. In the wall-normal ( y ) direction, the discretisation is based on Chebyshev polynomials on Chebyshev-Gauss-Lobatto points, where the grid resolution is refined towards the no-slip and the free-slip boundaries. The discretised governing equations are integrated in time with a semi-implicit scheme, where the viscous terms are treated by an implicit Euler approach, whereas a third-order low-storage Runge-Kutta scheme integrates the nonlinear terms explicitly. By convention, we define a velocity vector \mathbf{u}=u_{i}=\begin{bmatrix}u&v&w\end{bmatrix}^{\top} in the above Cartesian coordinate system, as well as pressure p . According to the Reynolds decomposition,

Read the complete text