the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Revealing horizontal gravity force in geopotential coordinates via metric tensors
Abstract. Earth gravity force g is represented by geopotential Φ, g =∇Φ , with the three-dimensional gradient operator. True gravity g has horizontal component. Oceanographic and meteorological communities use two approaches to eliminate horizontal gravity force. The first one is to replace Φ by geopotential for uniform Earth mass density (Φuniform) representing no horizontal gravity force, such as spherical, spheroidal, or recently developed Realistic, Ellipsoidal, Analytically Tractable (GREAT) ΦGREAT on the base of |Φ - Φuniform | <<| Φuniform |, however, it is not sufficient to justify because we still need to compare ∇(Φ - Φuniform) to other forces such as Coriolis force and pressure gradient force. The second approach is to use Φ to establish geopotential coordinates. Consider local Cartesian coordinates (ξ, η, ζ) with unit vectors (ξ^, η^, ζ^). The geopotential coordinates (x, y, Z) and corresponding unit vectors (x̂, ŷ, Ẑ) are defined by x = Z = -Φ/g0, g0= 9.81 m/s2. From such a relationship, metric tensors between (ξ, η, ζ) and (x, y, Z), and in turn the horizontal gradient operator are obtained. The pressure gradient is ∇C p = (∂ξp)ξ^ + (∂ηp)η^ + (∂ζp)ζ^ in the local Cartesian coordinates, and ∇G p = (∂xp + ∂xN∂zp)x̂ + (∂yp + ∂yN∂zp)ŷ + (∂Zp)Ẑ in the geopotential coordinates with N the geoidal undulation. The horizontal gravity force exists in horizontal momentum equation with the geopotential coordinates. Importance of the horizontal gravity force is verified using two publicly available datasets. Five concerns with PNAS-e2416636121 are presented in Appendix. The major equation in PNAS-e2416636121, ∇p = (∂ξp)ξ^ + (∂ηp)η^ + (∂ζp)ζ^ = (∂xp)x̂ + (∂yp)ŷ +(|∇Z|∂Zp)Ẑ, is valid only for the gravity with no horizontal component, i.e., for the geopotential coordinates coincident with the local Cartesian coordinates.
- Preprint
(1902 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-2010', Chris Hughes, 23 Apr 2026
-
CC2: 'Reply on RC1', Francis Poulin, 28 Apr 2026
Publisher’s note: the content of this comment was removed on 29 April 2026 since the comment was posted by mistake.
Citation: https://doi.org/10.5194/egusphere-2026-2010-CC2 -
AC4: 'Reply on RC1', Peter Chu, 11 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC4-supplement.pdf
-
RC4: 'Reply on AC4', Chris Hughes, 12 May 2026
-
AC6: 'Reply on RC4', Peter Chu, 13 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC6-supplement.pdf
-
RC6: 'Reply on AC6', Chris Hughes, 20 May 2026
-
AC7: 'Reply on RC6', Peter Chu, 27 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC7-supplement.pdf
-
AC7: 'Reply on RC6', Peter Chu, 27 May 2026
-
RC6: 'Reply on AC6', Chris Hughes, 20 May 2026
-
AC8: 'Reply on RC4', Peter Chu, 27 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC8-supplement.pdf
-
AC6: 'Reply on RC4', Peter Chu, 13 May 2026
-
RC4: 'Reply on AC4', Chris Hughes, 12 May 2026
-
AC10: 'Reply on RC1', Peter Chu, 05 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC10-supplement.pdf
-
CC2: 'Reply on RC1', Francis Poulin, 28 Apr 2026
-
CC1: 'Comment on egusphere-2026-2010', Baylor Fox-Kemper, 28 Apr 2026
One can define a "vertical", or really quasi-vertical, direction in such a way that the geopotential direction is not aligned with it. However, this is an unnecessarily complex and computationally costly choice:
1) The natural way to define the vertical is the direction where a plumb bob hangs (https://en.wikipedia.org/wiki/Plumb_bob). Carpenters have been using this trick to build strong homes for millennia. This combines all forces experienced by a body which is in static location relative to the surface of the earth, i.e., the vector sum of gravity and centrifugal forces.
2) If one insists on having "horizontal" components of gravity, then the hydrostatic balance becomes a partial differential equation rather than an ordinary differential equation, as now geopotential gradient has derivatives in both the "horizontal", latitude-like direction, as well as the "vertical". This engenders a spurious waste of computation.
3) In a coordinate-agnostic framework (see Joel Feske's recently defended PhD thesis and soon-to-be-submitted paper), the geopotential gradient direction can be implemented as vertical even in coordinate systems where no vertical coordinate exists (e.g., tracer coordinates: https://doi.org/10.1017/jfm.2012.638) by a simple projection operator (ki kj) where ki = ∇i (gravitational potential + centrifugal potential).
Citation: https://doi.org/10.5194/egusphere-2026-2010-CC1 -
AC2: 'Reply on CC1', Peter Chu, 11 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC2-supplement.pdf
-
AC13: 'Reply on CC1', Peter Chu, 05 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC13-supplement.pdf
-
AC2: 'Reply on CC1', Peter Chu, 11 May 2026
-
CC3: 'Comment on egusphere-2026-2010', Francis Poulin, 28 Apr 2026
This manuscript has a lot mathematics where the author rewrites the equations of motion using covariant and contravariant metric tensors. I have not gone through the details of the mathematics but I am concerned more about the fundamental assumptions that go into this model. The focus is to look at the horizontal component of gravity and how that affects the dynamics, which the author seems to have studied previously. I don't believe this is appropriate for the oceans or atmosphere. It's true that the Earth's surface is not a geoid, but the oceans and atmosphere will naturally align themselves so that gravity is acting orthogonal to these thin fluids (at large scales). Rather than simply defining two non-dimensional numbers, I think there should be more evidence to support that this horizontal component of gravity contributes to the large scale flows before this is seriously considered for publication.
Citation: https://doi.org/10.5194/egusphere-2026-2010-CC3 -
AC3: 'Reply on CC3', Peter Chu, 11 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC3-supplement.pdf
-
AC14: 'Reply on CC3', Peter Chu, 05 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC14-supplement.pdf
-
AC3: 'Reply on CC3', Peter Chu, 11 May 2026
-
RC2: 'Comment on egusphere-2026-2010', John Thuburn, 28 Apr 2026
-
AC5: 'Reply on RC2', Peter Chu, 11 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC5-supplement.pdf
-
AC11: 'Reply on RC2', Peter Chu, 05 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC11-supplement.pdf
-
AC5: 'Reply on RC2', Peter Chu, 11 May 2026
-
RC3: 'Comment on egusphere-2026-2010', Anonymous Referee #3, 02 May 2026
Publisher’s note: the supplement to this comment was edited on 6 May 2026. The adjustments were minor without effect on the scientific meaning.
-
AC1: 'Reply on RC3', Peter Chu, 07 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC1-supplement.pdf
-
RC5: 'Reply on AC1', John Thuburn, 17 May 2026
The author makes a valid point in the table in several of the Author Comments: the community is perhaps not as careful as we should be in documenting the spherical geopotential approximation in our models. (I'm sure the same is true of atmospheric models.)
Unfortunately, the author continually misses the point that the approximation made is a geometrical one; we start by assuming a geopotential coordinate - with no horizontal component of gravity by definition - and then approximate the geometry of the coordinate surfaces by approximating the metric (see RC4). In approximating the metric we do not change the balance of forces such as hydrostatic balance. In reply to AC1 point 2: yes, the pressure gradient can balance the bumpy geoid geopotential gradient. For a fairly arbitrary geopotential, there exists a hydrostatically balanced pressure field with p = p(Φ) and ρ = ρ(Φ) chosen so that ρ ∇p = ∇Φ (again see RC4). As the author points out, this statement of hydrostatic balance is coordinate independent. As CC3 points out, the atmosphere and ocean tend to satisfy this balance to an excellent approximation, and that remains true in models that make the spherical geopotential approximation.
I don't have time to carefully check all the mathematics in the author comments. However, there remain errors.
1. The metric tensor gij (below (R14) in AC1) is still incorrect.
2. In (R7), if we substitute for the basis vectors ax, ay, aZ in terms of the Cartesian basis vectors and then do the same in (R8) we get two different and contradictory expressions for ∇.
Citation: https://doi.org/10.5194/egusphere-2026-2010-RC5 -
AC9: 'Reply on RC5', Peter Chu, 27 May 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC9-supplement.pdf
-
AC9: 'Reply on RC5', Peter Chu, 27 May 2026
-
RC5: 'Reply on AC1', John Thuburn, 17 May 2026
-
AC12: 'Reply on RC3', Peter Chu, 05 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC12-supplement.pdf
-
AC1: 'Reply on RC3', Peter Chu, 07 May 2026
-
CC4: 'Comment on egusphere-2026-2010', Michael Montgomery, 09 Jun 2026
Publisher’s note: the supplement to this comment was edited on 11 June 2026. The adjustments were minor without effect on the scientific meaning.
-
AC16: 'Reply on CC4', Peter Chu, 10 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC16-supplement.pdf
-
AC16: 'Reply on CC4', Peter Chu, 10 Jun 2026
-
CC5: 'Comment on egusphere-2026-2010', Christopher Pitt Wolfe, 09 Jun 2026
The other reviewers and commenters have done a commendable job pointing out the flaws in the manuscript under consideration---primarily that, since "horizontal'' is defined to be perpendicular to gravity, there can be no horizontal component to gravity. As pointed out by John Thuburn, some of the author's confusion may stem from the imprecise way that coordinate systems are discussed in numerical model documentation. While the modeling community could do better at specifying exactly what is meant by "spherical" and "geopotential" coordinates, the truth is that most members of the field do not find this issue to be confusing.
I'll just add yet another quick derivation of the momentum equations in orthogonal geopotential coordinates demonstrating that the horizontal components contain no contributions from gravity in the hopes that some may find this derivation compelling. Please see the attached PDF.
-
AC15: 'Reply on CC5', Peter Chu, 09 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-2010/egusphere-2026-2010-AC15-supplement.pdf
-
AC15: 'Reply on CC5', Peter Chu, 09 Jun 2026
-
EC1: 'Comment on egusphere-2026-2010', Julian Mak, 10 Jun 2026
Thank you for submitting your manuscript to Ocean Science. In this case I do not encourage submission of a revised version to Ocean Science, for the following reasons:
- All referees have highlighted issues with the underlying premise of the paper, and neither I nor the referees (from what I can read) find the author's rebuttal to be convincing.
- There are technical issues with the fundamentals that the author notes but does not propose a resolution to (e.g., the appropriate co-ordinate transformations not being inverses of each other, which seems pretty basic and crucial to me).
- The author proposes to reduce the maths in a revision. While some comments have highlighted the use of tensor analysis to be non-essential, to me that reduction in technical content is not acceptable: a key part of the present paper is in its mathematical content to back up the author's point of view in the first point, and given a referee has highlighted issues, not addressing it head on indicates flaws in the technical argument.
- The way the author is proposing to revise the paper in my opinion makes it a different paper, and so should be treated accordingly, i.e. withdrawing the current paper and doing a new submission.
However I also do not recommend such a new submission to EGU Ocean Science. As the referees have pointed out, the issue that the author wants to argue would also impact atmospheric models as it is more generally a geophysical fluid dynamics problem; if it is so grossly neglected in the ocean as the author claims, the same effect should hold in the atmosphere (or some argument should be given as to why that is less important in the atmosphere). If the author wants to continue to argue this, they should consider journals that focus on modelling (e.g. Journal of Atmospheric Sciences, Monthly Weather Review, Journal for Advances in Modeling Earth Systems) and engage the atmospheric scientists employing related models with analogous assumptions.
Citation: https://doi.org/10.5194/egusphere-2026-2010-EC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 353 | 100 | 59 | 512 | 16 | 16 |
- HTML: 353
- PDF: 100
- XML: 59
- Total: 512
- BibTeX: 16
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
See attached pdf