the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Spectral Neutrality of Climate Reductions: An Operator Perspective
Abstract. Climate theory relies on a hierarchy of reductions that simplify the governing equations of radiative transfer and geophysical fluid dynamics. Examples include global averaging in energy balance models, quasigeostrophic filtering, the β-plane approximation, and idealized Kelvin–Rossby mode decompositions. These approximations are typically justified asymptotically and are highly successful within their intended regimes. However, they also modify the operators, domains, boundary conditions, or nonlinear functionals that define the admissible variability of the system.
This paper develops an operator-based framework for evaluating the spectral neutrality of climate reductions. A reduction is termed spectrally neutral if it preserves the operator class, admissible function space, domain topology, boundary conditions, and leading spectral structure of the original problem. Many widely used climate reductions are not spectrally neutral in a global sense, even when they remain locally or asymptotically accurate. Two examples are examined in detail. First, nonlinear radiative averaging in energy balance models is interpreted as a projection from a field equation onto a scalar closure, where averaging and nonlinear radiation operators do not commute. Second, the relation between spherical shallow-water dynamics and the β-plane approximation is reconsidered from the viewpoint of operator equivalence. The spherical Laplace tidal operator defines a compact global eigenvalue problem with discrete Hough spectra, whereas the β-plane formulation defines a different operator on a different domain with distinct admissible eigenfunctions. Boundary-value constraints in ocean basins further illustrate that low-frequency adjustment and teleconnections are governed by the spectrum of the full basin operator rather than by local plane-wave dispersion relations alone. The central issue is therefore not whether classical reductions are useful, but whether they preserve the spectral structure of the underlying climate dynamics. This perspective provides a unified framework connecting radiative closures, geometric reductions, and basin-scale wave adjustment.
- Preprint
(729 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 21 Aug 2026)
-
RC1: 'Comment on egusphere-2026-3085', Anonymous Referee #1, 17 Aug 2026
reply
-
AC1: 'Reply on RC1', Gerrit Lohmann, 17 Aug 2026
reply
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-3085/egusphere-2026-3085-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Gerrit Lohmann, 17 Aug 2026
reply
-
RC2: 'Comment on egusphere-2026-3085', Anonymous Referee #2, 17 Aug 2026
reply
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-3085/egusphere-2026-3085-RC2-supplement.pdf
-
AC2: 'Reply on RC2', Gerrit Lohmann, 17 Aug 2026
reply
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-3085/egusphere-2026-3085-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Gerrit Lohmann, 17 Aug 2026
reply
-
RC3: 'Comment on egusphere-2026-3085', Anonymous Referee #3, 17 Aug 2026
reply
The manuscript provides a perspective on the problem of model simplifications that leads to modifications of the spectral properties of operators, and proposes a framework for dealing with that problem. Several applications are briefly explored: the operation of averaging of nonlinear terms, the transformation of spherical geometry to beta-plane approximation, and boundary conditions that modify the nature of the solutions that could emerge. The manuscript is an interesting reminder of the impact of the approximations that are always made in modelling the processes of interest in an engaging manner, but the framework and potential applications remain too vague. I would recommend a deep revision. Here a few points that certainly need to be addressed in the manuscript:
Modelling is always made with some approximations. When reading the manuscript, the feeling is that if we solve the problem of boundaries, geometry and operators, everything is solved. This is clearly not the case as many processes and interactions are not properly represented in any modelling framework. This should be stressed. The approach proposed here is a way to solve part of the problems that may arise.
It is also important to refer to the rigorous mathematical analyses that have been made in the past based on asymptotic expansions which are key in the understanding of the power and limitations of the theories used in climate science and in the climate models. These theories have the advantage to put limits and bounds on the validity of the solutions generated by the atmospheric, ocean and Earth system models. Several references are indeed provided in the manuscript, Pedlosky (1987), Vallis (2017), Majda et al (2003). Relating these theories to the modifications of the spectral properties of operators mentioned in the manuscript is to my opinion an important aspect that should be elaborated or at least deeply discussed.
Concerning the theory of averaging, this question was addressed considerably in the nineties, showing that indeed care should be taken on how to handle the averaging process. These could lead to modifications of the dynamical properties (spectral properties of operators) of atmospheric or climate models, and proper account of these modifications needs careful developments. The latter problem is far to be solved even for very simple systems having nonlinearities. I suggest to check the literature and make the link to the current section dealing with the averaging problem.
Citation: https://doi.org/10.5194/egusphere-2026-3085-RC3 -
AC3: 'Reply on RC3', Gerrit Lohmann, 17 Aug 2026
reply
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-3085/egusphere-2026-3085-AC3-supplement.pdf
-
AC3: 'Reply on RC3', Gerrit Lohmann, 17 Aug 2026
reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 29 | 23 | 5 | 57 | 7 | 5 |
- HTML: 29
- PDF: 23
- XML: 5
- Total: 57
- BibTeX: 7
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In the article the author raises an important point, which I wholeheartedly agree with, namely that most if not all simplified models used in geophysical fluid dynamics modify the spectral structure of the equations of motion. Then the notion of spectral neutrality is introduced, where the spectrally neutral simplifications/reductions would preserve the spectral structure of linearized dynamics at least in the "leading eigenspaces".
Unfortunately, the concept of spectral neutrality as it is presented in the article is not thoroughly thought out and developed. The definition of spectral neutrality is given in lines 141-144. On one hand, this definition is too restrictive for practical application. For instance, the second bullet point requires the topology of the domain to be preserved. This automatically makes all plane, channel, or double periodic models non-neutral for atmospheric dynamics, since plane, channel, or torus are not topologically equivalent to the sphere. The situation is much worse for the ocean, where to achieve spectral neutrality one would need in addition to replicate the topology of the Earth' land mass. Thus, the examples of non-neutral models discussed in sections 4 and 5 are trivial based on topological criteria alone. It is hard for me to come up with a non-trivial example of a spectrally neutral reduced model satisfying definition 141-144 besides truncated models in the normal-mode framework on the sphere. The article also does not provide any examples of the spectrally neutral reduced models.
On the other hand, the proposed definition of spectral neutrality is vague. For instance, line 144 mentions "leading eigenspaces, or spectral measure", however no precise definition of this requirement is given. Is it enough to preserve some of the eigenspaces (if so how many?) , or must the property hold for all finite-dimensional subspaces? It is also not clear, whether the neutrality concept applies to linearized dynamics only, or there is an intent to generalize it to nonlinear systems as well. I assume it's the former, however, the article does not state this explicitly.
If we assume that the concept of spectral neutrality applies to linearlized dynamics only, the important question is around which background state the linearization is performed. For instance, eigenvalues/eigenfunctions of Laplace tidal equations, i.e. rotating shallow water equations linearized about the state of no motion differ from those of the rotating shallow water equations linearized about a non-trivial zonal steady state. How is the impact of the background state is then accounted for? Must spectral neutrality hold for at least one background state, single selected background state (e.g. state of no motion), a subclass of background states (e.g. zonal steady states withing some functional class), or all steady states within some functional class?
The mathematical development in section 2 is also vague and confusing. The use of star notation in Eq. (3) indicates that P* is the formal adjoint to P, however, this is not explicitly stated. If P* is indeed an adjoint operator, is it assumed at this point that H and H_r are Hilbert spaces? Further, the Eq. (4) is not consistent with previous definitions of L and P. While PL is well defined, LP is not, as P:H -> H_r and so the image of P is not in the domain of definition of L.
The title of the article does not reflect the content well. The issue of spectral differences between the simplified and underlying "full" models is not unique to climate, it is a common feature of simplified models encountered in geophysical fluid dynamics and beyond. The examples discussed in the article, with the exception of section 3 are also not climate-specific. Thus, I would suggest to remove the mention of climate from the title.
I commend the author for disclosing the use of AI. Unfortunately, the article is suffering from the typical pitfalls of AI use:
1) Incorrect citations. The article
Müller, D. and Maier-Reimer, E.: Tidal theory of the thermal wind, Physical Review E, 61,457, 1468–1485, 2000.
does not exist. This issue and page of Physicall Review E contains a different article
Müller, D., Trapped Rossby waves.
I was unable to track down the article
Müller, D.: Covariant formulation of shallow-water theory, International Journal of Modern Physics B, 11, 223–237, 1997
Could you provide DOI and/or direct link to this article?
I haven't checked all the references, so other similar issues are possible.
2) Repetitive text. For example, compare lines 136-137, 282-283, and 380-381; similarly cf. lines 82-83, 226-227, 265-266, 302-303, 344.
3) Overuse of triples. For example, lines 302-304 "Basin geometry, stratification, and boundary conditions therefore influence the spatial structure and propagation characteristics of low-frequency variability." However, the effects of stratification are not discussed in the prior text at all; the "stratification" was likely added to complete the triple.
In my view, the use of AI here is detrimental to the article. For what it is worth, in my opinion Lohmann(2020) is much better written.
Minor comments:
Lines (18)-(19). "The spherical Laplace tidal operator defines a compact global eigenvalue problem with discrete Hough spectra,..." The article does not specify what a "compact eigenvalue problem" is. I presume that what is meant is that the resolvent operator is compact for all values in the resolvent set. While it is true that Laplace tidal operator has discrete spectra, it does not automatically imply the compactness of the resolvent for a skew-adjoint operator. Thus please either reference the relevant result, or provide an argument for compactness.
Eq. (5): all parameters are not introduced.
Eq. (7): is this a simplifying assumption, or a statement of fact? This should be made clear.
Eq. (10): the left hand side should involve the averaged temperature.
Eq. (11): if the left hand side involves global equilibrium temperature, the right hand side should include the averaged S.
Eq. (13): psi and L_D are not defined .
Eq. (15): psi and L_M (Laplace tidal operator) are not defined.
Line (233) "In the non-rotating limit, the eigenfunctions reduce to spherical harmonics.." This statement is very imprecise. Hough harmonics, which are eigenfunctions of Laplace tidal operator are 3-compoment objects, consisting of zonal wind u, meridional wind v, and geopotential height h. Accordingly, L_M in Eq. (15) operates on 3-component fields. Consequently, L_M can not have spherical harmonics, which are single component objects, as eigenfunctions, regardless of the rotation rate.
Figure 2. The figure caption specifies neither which of the Rossby Hough modes is shown, nor which component (u, v, or h) is shown. Further, it is clear that eigenmodes of shallow water equations on the sphere do not need to vanish at 30N and 60N, which makes Figure 2 unnecessary.
To summarize, the author raises an important topic of the spectral equivalence between the simplified models, which are widely used in geophysical fluid dynamics, and the underlying equations of motion, as well as the consequences of the non-equivalence. However, the article does not go much beyond stating the problem. The proposed notion of spectral neutrality is mathematically vague and there is not a single example of a spectrally neutral reduced model provided. By their very nature simplified models are able to capture some, but not all aspects of underlying dynamics. Thus, perhaps one should not expect full spectral equivalence from a simplified model. Developing a unified practical framework for systematic evaluation of simplified models from spectral standpoint would be a worthy, albeit a difficult endeavor. Regretfully, the article under review does not succeed in doing so, which is why I can not recommend it for publication in the present form.