the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Aug 2026
Knowing the ocean: an epistemological and communication framework for Physical Oceanography
Abstract. Physical Oceanography is the branch of marine sciences concerned with understanding the oceans as a geophysical fluid system, focusing on the processes that govern the spatial organisation and temporal variability of momentum, heat, and mass across a wide range of scales. Knowledge in Physical Oceanography is produced through a combination of observations, theory, numerical models, and statistical inference; yet its epistemological structure remains largely implicit. This paper develops a framework to clarify how such knowledge is generated, validated, and constrained by the ocean’s vast spatial extent, continuous variability, and multiscale dynamics. We propose a two-dimensional epistemic map that classifies knowledge claims by their aim and representational form, providing a structured vocabulary for interpreting the diversity of oceanographic research. We illustrate the map with an exploratory classification of scientific synopses, and use it to read recent tendencies in the discipline. We further examine the epistemic status of oceanic numerical models and reanalysis systems, and present a case study of wind-wave spectral representation that exposes the underlying assumptions and limits of common practices. Building on these elements, we propose a diagnostic reporting template aimed at making explicit the assumptions, validation criteria, and epistemic payoff of individual studies. By rendering the epistemic structure of a result legible, the framework supports clearer interpretation, comparison, and communication of physical-oceanographic knowledge, both within the discipline and across the wider Earth-system sciences.
- Preprint
(1399 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (extended)
-
RC1: 'Comment on egusphere-2026-4208', Anonymous Referee #1, 13 Sep 2026
reply
The editor and authors should understand that I read this paper not as a philosopher of science, but as an active physical oceanographer. Having said that, the paper proved far more interesting than I expected from the title, and I believe that it should be accepted, subject to the addressing of the issues described below. In many ways the paper is wildly ambitious, and consequently there are many points, large and small to consider. (The journal demand for labelling things as "major" or "minor" changes is not helpful.)Despite its length, the paper skips over any true historical view: The comment by Walter Munk does not appear until very late in the paper, but a very strong argument can be made that PO underwent a true revolution beginning in the 1970s. Line 280 finally mentions the growth of global systems. The field went from one dominated by ship-borne observations, a system that took 100+ years to scantily observe the system, and leading to the concept that it was a laminar, almost unchanging flow, best described by the static fields in the colored atlases that dominated the field for 100 years. With the technological revolution following the 1970s, the demands for understanding climate, and the field having implemented such programs as the World Ocean Circulation Experiment, the picture shifted almost completely to a system that is observed globally (in some elements) every day, was fully turbulent, and far more dependent for understanding upon the fluid dynamical equations buried in the computer codes. That this huge change did not have a zero-order bearing on e.g., the changes in Tables 1,2, is hard to believe.A related comment is that the paper ignores the sociology of the science. The technical revolution destroyed a scientific way of life built around the use of ships, with then dominant scientists being primarily naturalists who knew no mathematics and despised it, working for weeks and months in isolation on ships, and ultimately being replaced by applied mathematicians, highly trained engineers, and computer scientists, et al. and either remote from the ships, or in constant contact with the outside world if at sea. The perception of ocean physics was strongly colored by the historical ship-based culture.Some serious attention needs to be given throughout to the question of who is the likely readership and will they understand what is being said?Section 6 might be moved to an Appendix, as I wonder how many readers of the document will understand the sudden introduction of mathematical elements? The section needs to emphasize that measurements are almost always discrete and that the continuum mathematics is a theoretical crutch. (More comments below.)The authors should be aware that in the PO community, the Oreskes 2022 book is widely regarded as nonsensical. (See e.g., review by Briscoe, in Oceanography (34) P. 78+ )Smaller points:Line 86+ Why is it useful to have an "explicit theory of knowledge"? Any examples?Liine 163. The Evensen reference is to the atmosphere. Better to have an oceanographic one.:ine 169. What is the meaning of "model-mediated"?Circa line 205. The most notable difference between PO and medicine/epidemiology, economics, is that PO has an a priori descriptive set of equations in the modified Navier-Stokes system from Newtonian mechanics. Those other fields have none--relying wholly on purely empirical laws.Line 319. The discussion of observations is inadequate. Listed as "sea-truth". But there is no such thing as "truth", and there is no such thing as a pure observation. Even a mercury thermometer relies on a (simple) model, one connecting the measurement of a length (of mercury) to a temperature. Understanding the power, and limits, of observations is a hugely important ongoing part of the field. "Matching" observations can degrade a model.Line 340+ Will the reader know what "data assimilation" is? Will she understand the difference (if any) from "reanalysis"? Line 370+ A considerable literature exists pointing out that conventional data assimilation does violate the core---violating energy, mass, and other conservation laws. (Meteorological practice is a good example of a system that works well for forecasting (weather), but not useful for basic understanding of principles such as energy conservation.) Recent paper by S. Williamson and others, 2023 in Ocean Sci., has various references.Under Lakatosian reading, the "hydrostatic approximation" is hardly a core element---numerous models are non-hydrostatic, including almost all dealing with continental margins. And fluid mechanics in practice is *always* based upon a series of approximations, all of which are constantly being subjected to question.Line 485. The authors might consider the current frenzy over AMOC "shutdown" as an example of using models beyond their skill level.Line 530. Many spectra have been computed and are useful for non-sinusoidal phenomena. The sentence is confused. The section goes on to argue that the spectrum is "not directly measured", whatever that actually means. But few phenomena are directly measured: the mercury thermometer is a trivial example of the inevitable requirement of interfacing instruments and models. Another example is salinity: typically measured by conductivity. Is that a direct measurement (even ignoring the complicated circuitry of a modern salinometer)?The term "geostrophy" notoriously does not appear in Sverdrup et al. (1942).ReplyCitation: https://doi.org/
10.5194/egusphere-2026-4208-RC1 -
RC2: 'Comment on egusphere-2026-4208', Anonymous Referee #2, 02 Nov 2026
reply
Summary:
The paper aims to develop an epistemological and communication framework for physical oceanography (PO), filling a genuine gap: PO has lacked an explicit, practice-facing treatment of how its knowledge is generated, validated, and limited. The authors propose a two-dimensional epistemic map (aim: descriptive/explanatory/predictive; form: statistical/causal/mechanistic), illustrate it through an LLM-based classification of Journal of Physical Oceanography abstracts (1995 vs. 2025), analyse the epistemic status of ocean models and reanalysis products, present an insightful case study of the wind-wave spectrum as an epistemic construct, and introduce a diagnostic reporting template for individual studies.
Due to my background, I would like to give some fruit for thought from a social science perspective:
Social science is incorporated, but could be more substantial
- The authors have clearly engaged with the literature — Douglas, Edwards, and Oreskes are apt choices — but the paper stops short of carrying their argument through. The introduction makes a persuasive case that societal needs shape what counts as oceanographic knowledge, yet the frameworks that follow — the epistemic map, the corpus analysis, the reporting template — leave that insight a little unused: values, funding structures, and the social organisation of expertise do not reappear as objects of analysis in their own right. As it stands, the social dimension is introduced with conviction and then set aside, which is the gap between citing this research and incorporating it. Here are some ideas: add a third, explicitly social axis (or at minimum a dedicated section) covering (i) the co-production of oceanographic knowledge with institutional and funding structures, drawing on Edwards and Oreskes at depth; (ii) social epistemology of trust, expertise, and consensus as it applies to model-based fields; and (iii) an empirically informed treatment of communication that engages actual audiences rather than idealised readers.
- There is no engagement with the sociology of scientific knowledge or social epistemology as it bears on the authors' own claims: robustness analysis (Schupbach is cited but only as a technical device), the social organisation of consensus (IPCC-style expert judgement is mentioned once), peer review, or the tacit-knowledge literature beyond a single Polanyi citation. Longino (science as social knowledge), Hardwig on trust in experts, or Collins & Evans on expertise would be directly relevant to a "practice-facing" epistemology.
- Citizen science, Indigenous and local knowledge in marine contexts, participatory observation, and the extensive marine social-science literature on ocean governance and the "blue economy" are absent entirely — which could be very beneficial for a paper claiming relevance to "societal applications" (see more below)
- The title promises a communication framework, yet there is no engagement with empirical work on how uncertainty is actually understood by audiences (e.g., the calibrated-language literature beyond a mention of IPCC, risk-perception research, or the substantial body of work on communicating model uncertainty to decision-makers). The reporting template (Sect. 7) is asserted to improve communication; no evidence or user testing with the intended audiences (reviewers, decision-makers, other disciplines) is offered
Section 2:
The "ocean" of physical oceanography is itself a historical artifact — and the paper treats it as given.
The paper does not reflect the field of Marine Social Sciences und Blue Humanities. A thorough reading of the relevant literature would be highly beneficial for the overall argument as there is a whole specific discourse on ocean science in the humanities. The blue humanities would say the paper has built a rigorous epistemology of the ocean-as-known-by-PO, while leaving unexamined the ocean-as-made-by-history (which also has tremendous effects on the epistemology of ocean science). See e.g.:
Bogusz, T. 2025. Valuating marine knowledge: Heterogeneous collaborations at the Concarneau marine station. History and Philosophy of the Life Sciences, 47(3), 37.;
Bogusz, T. et al. 2024. Nature and society in marine research – Perspectives for interdisciplinary cooperation in the Anthropocene. Zenodo. https://doi.org/10.5281/zenodo.14887580; 28;
Eriksson, B. & W.J. Boonstra. 2025. Toward contested seas? A review of geopolitical trends at sea in the 21st century. The Anthropocene Review, 20530196251334759;
Hannigan, J. 2017. Toward a sociology of oceans. Can. Rev. Sociol. 54, 8-27;
Jouffray, J. B et al. 2020. The blue acceleration: the trajectory of human expansion into the ocean. One earth, 2(1), 43-54;
Kołodziej, A., & Kołodziej-Durnaś, A. 2022. Maritime sociology in the making. Maritime Spaces and Society: 3-24;
Pelke N & FC Simonn FC (2023) Ocean-human relations in the Anthropocene: mapping trends in sustainability-related marine social sciences articles. Front. Mar. Sci. 10:1210133. doi: https://doi.org/10.3389/fmars.2023.1210133;
Ríos C.V. & B.B. Chicaguala. 2024. A Maritime Sociology for Sustainability Science. Current Opinion in Environmental Sustainability 70, 101474
Methodology
I understand and appreciate that the authors frame Section 3.2 explicitly as an "epistemic experiment" — a demonstration of the coding procedure rather than an empirical characterisation of the discipline, and I do not read the section as overclaiming. My question is more fundamental: given that framing, what does the section actually contribute? The two-axis map is already illustrated through worked examples in Section 3.1, and the tables are, by the authors' own account, hypothesis-generating only, sensitive to codebook choices, and inseparable from classification noise and drift in writing style. This leaves the section in an awkward position: too methodologically fragile to support its quantitative claims, yet too insubstantial to function as the demonstration it claims to be — since any qualitative example from Section 3.1 already shows "what the map makes visible" at least as convincingly.
Further remarks:
Line 100-105: who is meant by “working physical ocenographers''? Academics? Actors in the field outside of academia? Please reflect general use of terminology and a priori assumptions which the paper actually tries to question and could have the potential to add value to the own understanding of the field
Line 320: be careful with using "truth" as category for scientific discovery. Despite only referring to Kuhn and Popper, please urgently see Bruno Latour and Steve Woolgar's book Laboratory Life: The Construction of Scientific Facts (1979) which examines how scientific facts are socially and materially produced inside laboratories rather than simply discovered as pre-existing truths.
Line 445 – 455: Technology is never neutral: “I have the basic conviction that most of the progress that's been done in oceanography was the result of applying some new technology, not because somebody had some great new ideas. A new technology almost always leads to a new understanding, and I think it's a perfectly good way to go" (Munk, 2008). ---Very interesting take, which can be theorized by social science literature: see Latour, Bruno. "Morality and Technology: The End of the Means." Translated by Couze Venn, Theory, Culture & Society, vol. 19, no. 5-6, Dec. 2002, pp. 247–260.
Closing remarks
I find the idea behind this paper genuinely intriguing. Bringing an explicit, practice-facing epistemology to physical oceanography is a worthwhile and overdue project, and the two-dimensional map, the wave-spectrum case study, and the diagnostic template together show what such a framework could actually do for a working discipline.
My request for revisions comes, above all, from a social-science perspective. The paper itself makes the argument that societal needs, and the social conditions of knowledge production must be analyzed and made explicit. I am therefore asking the authors to carry their own insight through: a dedicated treatment of the co-production of oceanographic knowledge with its institutional context, of the social organisation of expertise and trust in model-based fields, and of communication with real rather than idealised audiences. Equally, the corpus experiment would benefit from modest methodological validation so that its suggestive trends can be interpreted with confidence.
These corrections are substantive but tractable, and none of them undermines the framework — rather, they would complete it. Addressing them would make a promising paper not only philosophically rigorous but also consistent with its own central claim
Citation: https://doi.org/10.5194/egusphere-2026-4208-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 369 | 178 | 86 | 633 | 101 | 92 |
- HTML: 369
- PDF: 178
- XML: 86
- Total: 633
- BibTeX: 101
- EndNote: 92
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
