Semi-analytical eddy-viscosity and backscattering closures for 2D geophysical turbulence
Pith reviewed 2026-05-22 20:00 UTC · model grok-4.3
The pith
Semi-analytical derivation fixes parameters for eddy-viscosity and backscattering closures in 2D geophysical turbulence.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The semi-analytical derivation provides the parameters of the Leith and Smagorinsky eddy-viscosity closures and the Jansen-Held backscattering closure up to a constant that can be estimated from the turbulent kinetic energy spectrum, enabling LES to correctly reproduce key DNS statistics including those of extreme events and interscale energy and enstrophy transfers while outperforming the dynamic Leith and Smagorinsky closures and the latter with standard parameter.
What carries the argument
Semi-analytical estimation of closure parameters using the turbulent kinetic energy spectrum from DNS snapshots to fix the remaining constant.
If this is right
- LES with the derived parameters reproduces the key statistics of DNS including extreme events and interscale transfers.
- The parameters agree with those obtained from online learning in several setups.
- These closures outperform the dynamic Leith and Smagorinsky and standard parameter baselines.
- The approach works for different setups of 2D geophysical turbulence.
Where Pith is reading between the lines
- If the spectrum provides a robust constant, the method could be applied to new flow regimes without re-derivation or re-calibration.
- This could connect to fully analytical derivations if the renormalization group estimates are used exclusively.
- Testing in 3D turbulence or with different forcing might reveal the limits of the spectrum-based estimation.
Load-bearing premise
The turbulent kinetic energy spectrum from a few DNS snapshots supplies a reliable estimate of the remaining constant without introducing resolution-dependent bias or requiring re-calibration when the flow regime changes.
What would settle it
Running an LES with the semi-analytically derived parameters on a 2D geophysical turbulence case whose spectrum was not used in the derivation and finding that it fails to match the corresponding DNS statistics would falsify the claim.
read the original abstract
Physics-based closures such as eddy-viscosity and backscattering models are widely used for large-eddy simulation (LES) of geophysical turbulence for applications including weather and climate prediction. However, these closures have parameters that are often chosen empirically. Here, for the first time, we semi-analytically derive the parameters of the Leith and Smagorinsky eddy-viscosity closures and the Jansen-Held backscattering closure for 2D geophysical turbulence. The semi-analytical derivation provides these parameters up to a constant that can be estimated from the turbulent kinetic energy spectrum of a few snapshots of direct numerical simulation (DNS) or other high-fidelity (eddy resolving) simulations, or even obtained from earlier analytical work based on renormalization group. The semi-analytically estimated closure parameters agree with those obtained from online (a-posteriori) learning in several setups of 2D geophysical turbulence in our earlier work. LES with closures that use these parameters can correctly reproduce the key statistics of DNS, including those of the extreme events and interscale energy and enstrophy transfers, and outperform the baselines (dynamic Leith and Smagorinsky and the latter with standard parameter).
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript claims to provide the first semi-analytical derivation of the parameters for the Leith and Smagorinsky eddy-viscosity closures and the Jansen-Held backscattering closure in 2D geophysical turbulence. These parameters are obtained up to a single overall multiplicative constant that is estimated from the turbulent kinetic energy spectrum of a few DNS snapshots (or from prior renormalization-group results). LES runs using the resulting closures are reported to reproduce key DNS statistics—including spectra, extreme-event distributions, and interscale energy/enstrophy transfers—while outperforming dynamic Leith/Smagorinsky baselines and agreeing with earlier online-learning results.
Significance. If the spectrum-derived constant proves insensitive to resolution and regime, the work supplies a practical route to physics-based closure coefficients that avoids both purely empirical tuning and the cost of online learning. The reported agreement with prior a-posteriori results and the reproduction of extreme-event statistics would constitute a useful bridge between analytical turbulence theory and operational geophysical LES.
major comments (2)
- [§3.2] §3.2 and the paragraph following Eq. (8): the remaining constant is extracted directly from the TKE spectrum of a small number of DNS snapshots; the manuscript provides neither error bars on this estimate nor a systematic test of its variation with grid resolution or Reynolds number. Because this constant multiplies every closure coefficient, any resolution or regime dependence would render the procedure a per-simulation calibration rather than a resolution-independent semi-analytical derivation, directly affecting the central claim of general applicability.
- [§4.3] §4.3, comparison with online-learning results: the reported numerical agreement is presented without a quantitative measure (e.g., relative difference or confidence interval) of how closely the spectrum-derived constant matches the learned value across the different forcing regimes; this weakens the assertion that the semi-analytical route reproduces the learned parameters without post-hoc adjustment.
minor comments (2)
- [Figure 7] The caption of Figure 7 does not specify the number of DNS snapshots used to compute the spectrum from which the constant is estimated.
- [§2] Notation for the eddy-viscosity coefficient is introduced inconsistently between the Leith and Smagorinsky derivations; a single symbol table would improve clarity.
Simulated Author's Rebuttal
We thank the referee for the constructive comments that help clarify the scope and limitations of our semi-analytical approach. We address each major comment point by point below.
read point-by-point responses
-
Referee: [§3.2] §3.2 and the paragraph following Eq. (8): the remaining constant is extracted directly from the TKE spectrum of a small number of DNS snapshots; the manuscript provides neither error bars on this estimate nor a systematic test of its variation with grid resolution or Reynolds number. Because this constant multiplies every closure coefficient, any resolution or regime dependence would render the procedure a per-simulation calibration rather than a resolution-independent semi-analytical derivation, directly affecting the central claim of general applicability.
Authors: We agree that the absence of error bars and a systematic sensitivity study limits the strength of the general-applicability claim. The constant is obtained from the inertial-range scaling of the TKE spectrum and is therefore expected to be insensitive to resolution once the inertial range is resolved, but this expectation is not quantified in the current manuscript. In the revised version we will (i) report error bars obtained by averaging over multiple independent DNS snapshots and (ii) add a short supplementary analysis (or table) showing the variation of the extracted constant across the resolutions and Reynolds numbers already available in our data set. If the variation remains small, this will support the semi-analytical character; if not, we will explicitly qualify the range of applicability. revision: yes
-
Referee: [§4.3] §4.3, comparison with online-learning results: the reported numerical agreement is presented without a quantitative measure (e.g., relative difference or confidence interval) of how closely the spectrum-derived constant matches the learned value across the different forcing regimes; this weakens the assertion that the semi-analytical route reproduces the learned parameters without post-hoc adjustment.
Authors: We accept that a purely qualitative statement of agreement is insufficient. The manuscript already states that the semi-analytically derived constants agree with the online-learned values in several setups, but no numerical differences or uncertainty estimates are provided. In the revision we will insert a table (or inline values) giving the relative differences and, where multiple snapshots are used, the associated standard deviations or confidence intervals for each forcing regime. This will allow readers to judge the closeness of the match directly. revision: yes
Circularity Check
Semi-analytical derivation supplies independent functional form; constant from spectrum or prior RG work is a separate estimation step.
full rationale
The paper derives the structure of Leith, Smagorinsky, and Jansen-Held parameters semi-analytically and states that only a multiplicative constant remains, which can be taken from the TKE spectrum of a few DNS snapshots, other high-fidelity data, or earlier renormalization-group analytical results. This separation keeps the core derivation self-contained and independent of the specific numerical values used later in LES. Agreement with the authors' prior online-learning results is noted but is not required to establish the semi-analytical expressions themselves. No equation or step reduces the claimed derivation to a fit or self-citation by construction.
Axiom & Free-Parameter Ledger
free parameters (1)
- overall multiplicative constant
axioms (2)
- domain assumption The energy spectrum of a few high-fidelity snapshots supplies a representative estimate of the remaining constant without resolution or regime dependence.
- standard math Standard assumptions of 2D geophysical turbulence (incompressibility, scale separation, and the validity of the chosen closure functional forms) hold.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.