GC Insights: Rainbow colour maps remain widely used in the geosciences

Westaway, Richard M.

doi:https://doi.org/10.5194/gc-5-83-2022

Articles | Volume 5, issue 1

https://doi.org/10.5194/gc-5-83-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

GC Insights

https://doi.org/10.5194/gc-5-83-2022

© Author(s) 2022. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 5, issue 1

GC Insights

|

11 Mar 2022

GC Insights |

| 11 Mar 2022

GC Insights: Rainbow colour maps remain widely used in the geosciences

Richard M. Westaway

Download

Final revised paper (published on 11 Mar 2022)
Preprint (discussion started on 26 Oct 2021)

Interactive discussion

Status: closed

CC1:
'Comment on gc-2021-39', Rolf Hut, 26 Oct 2021

I applaud dr. Westaway for investigating this important issue specifically for the geosciences. This is an important study to be brought to the attention of the editorial teams and support staff of our journals.
I do, however, not agree with dr. Westaways point that the responsibillity is "with editors and reviewers to identify and correct colour issues as part of the review process" (line 101 of the prepint). Editors and especially reviewers are already over-asked to provide unpaid labour for the amount of papers that need review and should, in my honest opinion, focus on the scientific merit of the work. Current review processes tend to focus (again, imho) too much on the points and comma's of a manuscript and less on the rigor of the science presented. Checks like the correct use of colour maps should be part of the paper validation step by the support staff of the publisher and could even be included in automated checks. This is, I think, partly what APCs are supposed to be for.
Looking forward to dr. Westaway's reaction and thoughts on this.
dr. ir. Rolf Hut

Citation: https://doi.org/10.5194/gc-2021-39-CC1
- AC1: 'Reply on CC1', Richard Westaway, 26 Oct 2021
  
  I thank Dr Hut for his interest in the manuscript and for his comment about the role of editors and reviewers.
  
  Putting aside the 'fairness' of the peer review system, I think where responsibility lies for identifying colour issues is an interesting question. Based on current review criteria, "presentation quality" (including "quality of figures") is one of the three principal evaluation criteria used by the editors of EGU journals for initial manuscript evaluation, implying it is currently (rightly or wrongly) part of these editors' roles to check for colour issues (though I appreciate this will vary from publisher to publisher). Whether this could be done in a more automated way - one that reduces the burden on editors - is another good question and yes, perhaps ideally, all colour issues would be identified and corrected before a paper gets sent out for peer review.
  
  While I agree with your assertion that the scientific merit of the work should be reviewers' primary focus, I think it is widely accepted that reviewers are also ensuring that results are reported correctly and unambiguously. I know, for example, that the AGU specifically advise reviewers to ask: “Are all parts of the text, references, graphics and tables necessary for the new results and main points to be understood?”. It implies, to me at least, that this should include checking that figures are not misleading due to the use of rainbow or red-green colour schemes.
  
  Thus, on balance, I stand behind my (full) original statement that more publisher guidelines about colour issues "would appear to place increased responsibility with editors and reviewers to identify and correct colour issues as part of the review process". I should stress I am not saying that this is necessarily the best or most efficient way to eliminate colour issues, just my interpretation of the current situation.
  
  Citation: https://doi.org/10.5194/gc-2021-39-AC1
CC2: 'rainbow logo', Klemens Hocke, 03 Nov 2021

How critical is the rainbow logo of Geoscience Communication?

Citation: https://doi.org/10.5194/gc-2021-39-CC2
RC1:
'Comment on gc-2021-39', Fabio Crameri, 17 Nov 2021

This is a solid, significant, and insightful study revealing and quantifying the widespread use of data-distorting and to many readers inaccessible colour combinations, like the rainbow colour map or the simple red and green combination. The author makes a good choice to focus on his findings, namely an extensive survey of papers published in geoscience journals rather than diving too deep into explaining details like the human perception of colour, uniformity of colour gradients, or available colour maps, which has been done already multiple times elsewhere. As such it is a concise and well-presented insight to any geoscientist and geoscience communicator and fits as a publication in Geoscience Communication. I only have minor comments to take care of by the author and suggest it for publication afterwards.

Line 77: Regarding point (i) on the increased awareness of the importance of scientific colour maps: Our recent perspective piece in Nature Communication on the topic (Crameri et al., 2020 - already in reference list - written by Geoscientists) is, now with >145k accesses, the most widely accessed paper of the journal published in the year 2020 (see https://www.nature.com/collections/gegdbjhfbj/). I guess that is a valid reference point to support this statement?

Lines 47-79: An important additional reason why there is less colour issues in recent years, might also be because there is more science-proof alternative colour maps available. The Scientific colour maps package (Crameri 2018) has, for example, in the last two years gained many more users, with currently >4.7k downloads, of which more than half (>2.6k) are for the latest version (7.0.0), which was released less than a year ago (on 2.2.2021) – See http://doi.org/10.5281/zenodo.1243862 for details. Moreover, the Scientific colour maps package provides all colour map types and classes, which has not been the case before, meaning that e.g., scientific cyclic or categorical colour maps were either rare, not readily usable with common visualisation software, or simply non-existent, and hence simply not used.

Crameri, F. (2018). Scientific colour maps. Zenodo. http://doi.org/10.5281/zenodo.1243862

Line 84: Maybe this paragraph could be introduced by the relevant finding of Moreland (2016), that the widespread use of rainbow is the main reason scientists propagate it further.

Moreland, K. Why we use bad color maps and what you can do about it. Electron. Imaging 2016, 1–6 (2016).

Lines 100-101: Just to back up the author in a previous discussion comment: The last sentence is a fair statement by the author, especially considering that adding visual errors to data of up to 7% and more (across the data range displayed; see Crameri 2017, GMD) is severely misleading (to both authors and readers). Colour misuse is not about personal preference or opinions, it is quantifiable and fact, in the same way statistical misuse is. So, if a reviewer or editor is supposed to ensure publications free of the latter, then this should clearly also apply to the misuse of colour. Until we researchers receive a solid education in scientific visualisation (i.e., one of the most widely used scientific methodologies), a thorough quality check by reviewers and editors is really the only way forward.

Fabio Crameri

17.11.2021

Citation: https://doi.org/10.5194/gc-2021-39-RC1
- AC2: 'Reply on RC1', Richard Westaway, 10 Dec 2021
  
  I thank Dr Crameri for his supportive comments.
  Line 77: I agree that the fact that Crameri et al. (2020) is one of the most downloaded papers from that year is a useful proxy to indicate that there is increased awareness of the importance of colour maps, and I will add this detail to the manuscript accordingly.
  Lines 47-79: Similarly, I welcome the suggestion that the increase in users of the Scientific colour maps package is a useful indication that more alternative colour maps are available (and, more to the point, that they are being used), and will use this to support point (iii) on line 78.
  Line 84: I thank Dr Crameri for bringing the Moreland (2016) paper to my attention. I believe this is a useful reference to use to suggest that "inertia" might help explain continued use of rainbow colour maps. However I propose that this is more effectively done in the Introduction (lines 25-26) as a potential reason why rainbow colour maps remain frequently used despite their known flaws. I would favour restricting discussion on line 84 and onwards to guidance (or lack thereof) associated with the publishing process.
  
  Citation: https://doi.org/10.5194/gc-2021-39-AC2
RC2:
'Comment on gc-2021-39', Anonymous Referee #2, 08 Dec 2021

While it is concerning to note that over half the papers surveyed contained visualisations that were problematic I feel the paper is not really very interesting or useful to the general reader. It would be made more interesting to readers if the discussion could be expanded to consider the principal factors contributing to the generation of poor visualisations and to suggest some more concrete proposals to improve the situation.

Certainly a lack of guidance from journals and conferences is an issue. Journals and conferences typically mandate LaTeX or Word templates to ensure a standard of typesetting and layout, it would seem visualisation deserves a similar level of guidance. The requirements for manuscript figures as specified by Ocean Science is excellent.

However, while lack of guidance to authors is an issue perhaps the main contributors to the problem are the vendors of visualisation packages who, in many cases, provide poor default colour maps and fail to provide much, if any, guidance in their use.

Another contributing factor is that while authors, reviewers and editors have ready access to numerous guides for english writing and expression there is no concise and accessible equivalent to Strunk and White for visualisation. Also, while courses in scientific writing are widely given at educational institutions equivalent courses in scientific visualisation are perhaps not so common.

Citation: https://doi.org/10.5194/gc-2021-39-RC2
- AC3: 'Reply on RC2', Richard Westaway, 10 Dec 2021
  
  I thank reviewer #2 for taking the time to review the manuscript and for raising some interesting points.
  It was disappointing to read that the reviewer felt that the paper was not useful to the general reader, but without specific feedback on why they felt this was the case, it is hard for me to make any further comment. Discussion was deliberately not expanded to wider topics as (i) I felt these have been covered elswehere and (ii) as a "GC Insights" paper, the manuscript was intended to consiely present an idea but not explore it in detail. Nonetheless, I will endeavour to add further references to the manuscript and/or make it clearer where further details can be obtained (e.g. Crameri et al 2020 for factors; Stoelzle and Stein 2021 for proposals).
  I agree that lack of guidance remains an issue and that, historically at least, software and visualisation packages provided poor default colour options. However I think that situation has changed in recent years, with many packages now offering a much better range of colour map options (e.g. ColourBrewer https://colorbrewer2.org/ and Scientific colour maps http://doi.org/10.5281/zenodo.1243862).
  Finally, I agree that 'training' in scientific visualisation has not been (and is likely still not) as common or seen as important as training in other aspects of scientific communication, and that this is likely another reason for the persistance of rainbow and red-green colour schemes.
  
  Citation: https://doi.org/10.5194/gc-2021-39-AC3
AC4: 'Comment on gc-2021-39', Richard Westaway, 11 Jan 2022

Final Author Comment. Please see coments AC2 and AC3 for my responses to reviewers comments.

Citation: https://doi.org/10.5194/gc-2021-39-AC4

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

ED: Publish subject to minor revisions (further review by editor) (01 Feb 2022) by Leslie Almberg

AR by Richard Westaway on behalf of the Authors (04 Feb 2022) Author's response Author's tracked changes Manuscript

ED: Publish as is (08 Feb 2022) by Leslie Almberg

ED: Publish as is (08 Feb 2022) by Sam Illingworth (Executive editor)

AR by Richard Westaway on behalf of the Authors (09 Feb 2022)

Short summary

Rainbow and red–green colour schemes are widely used in scientific publications despite being known to be problematic, particularly to people with colour blindness. This study investigates to what extent such colour schemes are used in geoscience publications. Based on a desk survey of over 2500 papers from six journals between 2005 and 2020, it is found that over half of the published papers have maps or graphs with colour issues that are therefore potentially misleading to readers.