the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 16 Dec 2024
GCInsights: Consistency in Pyrocartography Starts With Color
Abstract. Fire progression maps provide operational and public information regarding wildland fire spread, size, and proximity to critical assets through time. Cartographic guidance regarding the use of color to denote the sequential nature of fire progression is limited, leading to inconsistency in fire progression maps produced for operational, research, and public applications. Because this inconsistency potentially limits the map’s accessibility and ability to effectively communicate information, I provide colormap recommendations to facilitate consistent, intuitive, and accessible fire progression mapping.
- Preprint
(8689 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on gc-2024-9', Fabio Crameri, 25 Dec 2024
Inconsistencies and misusage of colour in pyrocartography (and beyond) is a major issue and needs to be tackled. This manuscript nicely tackles this problems and provides a clear solution to make understanding, communicating, and tackling fire hazard more accurate, effective, and accessible. I therefore think that it is a very valuable contribution to be published in Geoscience Communication. The article is clearly presented and only has some minor misrepresentation that I think should be fixed before publication.
I have one major issue: The colour map ‚Turbo‘ is presented and used as a perceptually uniform (and I guess colour blind friendly) colour map, which it is not. I think it should be omitted as recommendation here (and elsewhere) to avoid further misconception about its properties. See a more detailed explanation below.
I have one major suggestion: Would simulating and providing colour-vision deficient appearance (as is done for figure A1), and maybe even greyscale conversion, of figure 1 be useful for the purpose of the manuscript? – I think it would and attached some simulations (based on Brettel et al., 1997; happy to share adjusted figures for the author). That way, it would also become more clear that Turbo is not suited for a scientific application (repeated colours along the scale).
Point-by-point comments:
- Line 32: Please clarify what „non-colorsafe colormaps“ means.
- Lines 45–47: To support these statements, we clarify the use of different colour gradient types in:
- Crameri, F., G.E. Shephard, and P.J. Heron (2024). Choosing suitable color palettes for accessible and accurate science figures. Current Protocols, 4, e1126. https://doi.org/10.1002/cpz1.1126
- Line 47: Put the Buckley, 2017 reference into parentheses (i.e., „\citep{xxx} )“ – same problem in other instances, such as line 51
- Lines 48–50: Good point!
- Line 50: „We“ should be „I“ for this single-author paper, shouldn’t it?
- Line 52: Turbo is NOT perceptually uniform. It is confused widely as such though. I guess there is less scrutiny for a colour map developed by a big company than by individual scientists. Turbo is better than a standard rainbow (such as Jet) in terms of perceptual evenness, but it is still not on par with e.g., batlow or YlOrBr, and definitely not perceptually uniform (see e.g., the author’s clarification that „[Turbo] is not a perceptually linear“ and „[Turbo is intended] for day-to-day tasks where perceptual uniformity is not critical“ on https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://research.google/blog/turbo-an-improved-rainbow-colormap-for-visualization/&ved=2ahUKEwjE4snyxbuKAxUjgP0HHU1jOBQQFnoECBoQAQ&usg=AOvVaw3ZCRKST1BPJ1AX72-lwrl4 ). This is visible in Figure 1, where the neighbouring colours in the blueish parts of the color bar clearly differ more strongly than e.g., in its greenish parts.
- Line 54: The Ware et al., 2023 paper is trying to make a case for the use of non-perceptually-uniform and inaccessible colour maps, so it does not fit in here, I think.
- Lines 57–59: Not sure that is a helpful statement. It is unclear what „situation-specific adjustments to colormaps“ means. If it means distorting the uniformity of color gradient of the scale, than this is a bad suggestion, as it would distort the actual data, and for the reader impossible to reproduce (or judge for its validity). In particular in plots showing the spread of fire, it seems key to properly display where the fire spread slowly, and how much more rapidly it spread elsewhere. Distorting the colour scale would suppress this information and misrepresent it.
- Lines 61–63: Along similar lines, I disagree with this statement, as either a colour map is perceptually uniform or it is not. What „used appropriately“ and „better design“ mean here is unclear. Moreover, the paper referenced argues with very limited cases, and provides suggestions that are implementable only by visualisation experts and rather confusing to everyone who is not. And let’s remember, hardly any scientist has received even the basic education in scientific visualisation. The paper’s argument can be mentioned here, but I don’t think it is fair nor clear to the reader to present it equally to other, more broadly applicable studies and more broadly based arguments.
- Figure 1: Given that many people tend to look at figures mostly, these days, the figure caption could be a bit more descriptive to be more helpful: for example, explain that these are recommended colour maps, and what techniques are used in (b,d,f) to increase accessibility.
I enjoyed reading this nice piece and would like to thank the author for their effort!
And finally, for transparency, I am the author of the Scientific colour maps (some of which are shown here).
-
AC1: 'Reply on RC1', Benjamin Hatchett, 20 Feb 2025
Dear Dr. Crameri,
Thank you for your constructive and beneficial but also thought-provokingly critical suggestions for revisions. Your suggestion (and explanation of why it was a major issue) to omit the "Turbo" colormap as well as the suggested idea (shared also by Reviewer 2) to alter Figure 1 and show colourblindness simulations of suggested colourmaps instead of multiple examples of the same colourmaps were particularly appreciated. I have revised the manuscript to follow your suggestions and edits, leading to improvements in content and clarity throughout. Please see the standard response-to-reviewer formatted document in the attached supplement.
Sincerely,
Ben Hatchett
-
RC2: 'Comment on gc-2024-9', Richard Westaway, 08 Jan 2025
Summary
The author explains the utility of fire progression maps (FPM) and argues, given their importance, that greater consideration should be given to the development and adoption of standardised colormaps to portray fire progressions. The paper is well argued, clearly written and introduces an interesting application where the visual communication of spatial information is of great importance to a range of stakeholders. Since I have no prior experience in pyrocartography, the following comments are from a more general colour accessibility and science communication perspective rather than written with any insight in this particular application.General comments
I have two general comments:(1) Figure 1
This figure is intended to illustrate the importance of colormaps for FPMs, and also serves to demonstrate some different color options that are available. However, I find the choice of datasets and colormaps as currently presented is somewhat muddled and suggest that an alternative set of panels might more clearly support the narrative. Currently, Figure 1 presents fire perimeter data from two fire events, demonstrates four different colormaps as well as showing the visual impact of including isochrons. However Figure 1 does not currently include any color blindness simulation or black and white printing challenges (as are shown in Figure A1 for other fire progression mapping examples).
While not possible to show every permutation, I would suggest simplifying what is being shown for example by separating (or only showing one of) the two fire events – plus adding in some examples of colour-vision deficient appearance – such that 12 panels are shown for each fire event (3 panels wide x 4 panels tall), with the panels progressing more logically and allowing easier comparison: 1a Rim Fire Turbo; 1b Rim Fire Turbo with isochrons; 1c Rim Fire Turbo with color blindness simulation; 1d Rim Fire Batlow; 1e Rim Fire Batlow with isochrons; 1f Rim Fire Batlow with color blindness simulation; 1g Rim Fire reversed YlOrBr; 1h Rim Fire reversed YlOrBr with isochrons; 1i Rim Fire reversed YlOrBr with color blindness simulation; 1j Rim Fire reversed Managua; 1k Rim Fire reversed Managua with isochrons; 1l Rim Fire reversed Managua with color blindness simulation. Then I would suggest repeating for the Dixie Fire, i.e.: 2a Dixie Fire Turbo; 2b Rim Fire Turbo with isochrons; 2c Rim Fire Turbo with color blindness simulation; and so on….
Recognising the figure constraints of the GCInsights format, and given that the full 24 panels would be near impossible to display together, I would advocate illustrating perhaps one of the fire events in the main body text (Figure 1 a-l) and moving the other to the Appendix (new Figure A2 a-l). If this is still felt to be too many panels to communicate clearly, my instinct and preference would be that color vision deficient appearance is more important to illustrate (i.e. has a larger effect on visual communication) than the presence/absence of isochrons, which would eliminate the need for four panels per fire event.
The figure captions should also be changed accordingly, and I would suggest expanded to explain each panel more fully.
(2) Preferred colormap(s)
Given the obvious thought that the author has given to the issue of colormaps for FPMs, and the suggestion in the abstract that the paper provides “colormap recommendations”, it might be helpful to readers if the final section of the paper included some specific recommendations for which of the colormaps presented might provide the best basis for standardisation (or alternatively if any of the colormaps presented are clearly less suitable to use in this context). While Figure 1 presents various different colormap and map presentation options, the author does not currently provide any views or evaluation on their relative ability to communicate the desired information. The inclusion of color vision deficient maps in Figure 1 (as suggested above) might help demonstrate how some of the colormaps shown are perhaps less suitable than others (e.g. repeated colors in Turbo).
Specific comments
1. p1, line 4-5: I would suggest the end of the abstract is reworded to “…applications, which potentially limit these map’s accessibility and ability to effectively communicate information. In this paper, I provide colormap recommendations to facilitate consistent, intuitive, and accessible fire progression mapping.”
2. p1, line 6: Change “Wildland fire’s movement..” to “The movement of wildland fires…”
3. p1, line 16: Unnecessary comma after parentheses.
4. p2, line 31: Define or explain “non-colorsafe” (and hence “color-safe” on line 34)
5. p2, line 31 and 34: Inconsistent spelling of colorsafe/color-safe
6. p2, line 34: Define or explain “508 Compliance”
7. p2, line 35: “2022” should not be in parentheses
8. p2, line 36: “2023” should not be in parentheses
9. p2, line 37: “if the maps change” - I suggest you clarify here that you mean that if the colourmaps (or other presentational elements) of the FPMs change. What the maps actually show (i.e. the fire perimeters) would of course be expected to change as new/updated data is included.
10. p2, line 42: “I” not “we”
11. p2, line 46-47: Reference should be in parentheses
12. p2, line 50: “I” not “we”
13. p3, Figure caption: I suggest that the caption is expanded to include more details about what is being shown in Figure 1. However please see my general comments above for further suggestions about Figure 1.
14. p4, line 69: “fires” not “fire”
15. p4, line 71-72: “The four suggested colormaps (Figure 1), one of which is an example in GeoOps (Figure A1d), intend to provide a starting point.” - It is not immediately clear why Figure 1 is stated to show (only) four suggested colormaps rather than six. However please see my general comments above for further suggestions about Figure 1.I look forward to seeing this published in due course!
Citation: https://doi.org/10.5194/gc-2024-9-RC2 -
AC2: 'Reply on RC2', Benjamin Hatchett, 20 Feb 2025
Dear Dr. Westaway,
Thank you for your detailed review and your thoughtful and helpful suggestions for revisions. Your suggestions to re-work Figure 1 to better highlight the colourblindness simulations to better support the narrative of the paper was particularly valuable and appreciated. Your additional suggestions and detail-oriented edits led to improvements throughout. I have revised the manuscript to follow your suggestions and edits, leading to improvements in content and clarity throughout. Please see the standard response-to-reviewer formatted document in the attached supplement.
Sincerely,
Ben Hatchett
-
AC2: 'Reply on RC2', Benjamin Hatchett, 20 Feb 2025
Status: closed
-
RC1: 'Comment on gc-2024-9', Fabio Crameri, 25 Dec 2024
Inconsistencies and misusage of colour in pyrocartography (and beyond) is a major issue and needs to be tackled. This manuscript nicely tackles this problems and provides a clear solution to make understanding, communicating, and tackling fire hazard more accurate, effective, and accessible. I therefore think that it is a very valuable contribution to be published in Geoscience Communication. The article is clearly presented and only has some minor misrepresentation that I think should be fixed before publication.
I have one major issue: The colour map ‚Turbo‘ is presented and used as a perceptually uniform (and I guess colour blind friendly) colour map, which it is not. I think it should be omitted as recommendation here (and elsewhere) to avoid further misconception about its properties. See a more detailed explanation below.
I have one major suggestion: Would simulating and providing colour-vision deficient appearance (as is done for figure A1), and maybe even greyscale conversion, of figure 1 be useful for the purpose of the manuscript? – I think it would and attached some simulations (based on Brettel et al., 1997; happy to share adjusted figures for the author). That way, it would also become more clear that Turbo is not suited for a scientific application (repeated colours along the scale).
Point-by-point comments:
- Line 32: Please clarify what „non-colorsafe colormaps“ means.
- Lines 45–47: To support these statements, we clarify the use of different colour gradient types in:
- Crameri, F., G.E. Shephard, and P.J. Heron (2024). Choosing suitable color palettes for accessible and accurate science figures. Current Protocols, 4, e1126. https://doi.org/10.1002/cpz1.1126
- Line 47: Put the Buckley, 2017 reference into parentheses (i.e., „\citep{xxx} )“ – same problem in other instances, such as line 51
- Lines 48–50: Good point!
- Line 50: „We“ should be „I“ for this single-author paper, shouldn’t it?
- Line 52: Turbo is NOT perceptually uniform. It is confused widely as such though. I guess there is less scrutiny for a colour map developed by a big company than by individual scientists. Turbo is better than a standard rainbow (such as Jet) in terms of perceptual evenness, but it is still not on par with e.g., batlow or YlOrBr, and definitely not perceptually uniform (see e.g., the author’s clarification that „[Turbo] is not a perceptually linear“ and „[Turbo is intended] for day-to-day tasks where perceptual uniformity is not critical“ on https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://research.google/blog/turbo-an-improved-rainbow-colormap-for-visualization/&ved=2ahUKEwjE4snyxbuKAxUjgP0HHU1jOBQQFnoECBoQAQ&usg=AOvVaw3ZCRKST1BPJ1AX72-lwrl4 ). This is visible in Figure 1, where the neighbouring colours in the blueish parts of the color bar clearly differ more strongly than e.g., in its greenish parts.
- Line 54: The Ware et al., 2023 paper is trying to make a case for the use of non-perceptually-uniform and inaccessible colour maps, so it does not fit in here, I think.
- Lines 57–59: Not sure that is a helpful statement. It is unclear what „situation-specific adjustments to colormaps“ means. If it means distorting the uniformity of color gradient of the scale, than this is a bad suggestion, as it would distort the actual data, and for the reader impossible to reproduce (or judge for its validity). In particular in plots showing the spread of fire, it seems key to properly display where the fire spread slowly, and how much more rapidly it spread elsewhere. Distorting the colour scale would suppress this information and misrepresent it.
- Lines 61–63: Along similar lines, I disagree with this statement, as either a colour map is perceptually uniform or it is not. What „used appropriately“ and „better design“ mean here is unclear. Moreover, the paper referenced argues with very limited cases, and provides suggestions that are implementable only by visualisation experts and rather confusing to everyone who is not. And let’s remember, hardly any scientist has received even the basic education in scientific visualisation. The paper’s argument can be mentioned here, but I don’t think it is fair nor clear to the reader to present it equally to other, more broadly applicable studies and more broadly based arguments.
- Figure 1: Given that many people tend to look at figures mostly, these days, the figure caption could be a bit more descriptive to be more helpful: for example, explain that these are recommended colour maps, and what techniques are used in (b,d,f) to increase accessibility.
I enjoyed reading this nice piece and would like to thank the author for their effort!
And finally, for transparency, I am the author of the Scientific colour maps (some of which are shown here).
-
AC1: 'Reply on RC1', Benjamin Hatchett, 20 Feb 2025
Dear Dr. Crameri,
Thank you for your constructive and beneficial but also thought-provokingly critical suggestions for revisions. Your suggestion (and explanation of why it was a major issue) to omit the "Turbo" colormap as well as the suggested idea (shared also by Reviewer 2) to alter Figure 1 and show colourblindness simulations of suggested colourmaps instead of multiple examples of the same colourmaps were particularly appreciated. I have revised the manuscript to follow your suggestions and edits, leading to improvements in content and clarity throughout. Please see the standard response-to-reviewer formatted document in the attached supplement.
Sincerely,
Ben Hatchett
-
RC2: 'Comment on gc-2024-9', Richard Westaway, 08 Jan 2025
Summary
The author explains the utility of fire progression maps (FPM) and argues, given their importance, that greater consideration should be given to the development and adoption of standardised colormaps to portray fire progressions. The paper is well argued, clearly written and introduces an interesting application where the visual communication of spatial information is of great importance to a range of stakeholders. Since I have no prior experience in pyrocartography, the following comments are from a more general colour accessibility and science communication perspective rather than written with any insight in this particular application.General comments
I have two general comments:(1) Figure 1
This figure is intended to illustrate the importance of colormaps for FPMs, and also serves to demonstrate some different color options that are available. However, I find the choice of datasets and colormaps as currently presented is somewhat muddled and suggest that an alternative set of panels might more clearly support the narrative. Currently, Figure 1 presents fire perimeter data from two fire events, demonstrates four different colormaps as well as showing the visual impact of including isochrons. However Figure 1 does not currently include any color blindness simulation or black and white printing challenges (as are shown in Figure A1 for other fire progression mapping examples).
While not possible to show every permutation, I would suggest simplifying what is being shown for example by separating (or only showing one of) the two fire events – plus adding in some examples of colour-vision deficient appearance – such that 12 panels are shown for each fire event (3 panels wide x 4 panels tall), with the panels progressing more logically and allowing easier comparison: 1a Rim Fire Turbo; 1b Rim Fire Turbo with isochrons; 1c Rim Fire Turbo with color blindness simulation; 1d Rim Fire Batlow; 1e Rim Fire Batlow with isochrons; 1f Rim Fire Batlow with color blindness simulation; 1g Rim Fire reversed YlOrBr; 1h Rim Fire reversed YlOrBr with isochrons; 1i Rim Fire reversed YlOrBr with color blindness simulation; 1j Rim Fire reversed Managua; 1k Rim Fire reversed Managua with isochrons; 1l Rim Fire reversed Managua with color blindness simulation. Then I would suggest repeating for the Dixie Fire, i.e.: 2a Dixie Fire Turbo; 2b Rim Fire Turbo with isochrons; 2c Rim Fire Turbo with color blindness simulation; and so on….
Recognising the figure constraints of the GCInsights format, and given that the full 24 panels would be near impossible to display together, I would advocate illustrating perhaps one of the fire events in the main body text (Figure 1 a-l) and moving the other to the Appendix (new Figure A2 a-l). If this is still felt to be too many panels to communicate clearly, my instinct and preference would be that color vision deficient appearance is more important to illustrate (i.e. has a larger effect on visual communication) than the presence/absence of isochrons, which would eliminate the need for four panels per fire event.
The figure captions should also be changed accordingly, and I would suggest expanded to explain each panel more fully.
(2) Preferred colormap(s)
Given the obvious thought that the author has given to the issue of colormaps for FPMs, and the suggestion in the abstract that the paper provides “colormap recommendations”, it might be helpful to readers if the final section of the paper included some specific recommendations for which of the colormaps presented might provide the best basis for standardisation (or alternatively if any of the colormaps presented are clearly less suitable to use in this context). While Figure 1 presents various different colormap and map presentation options, the author does not currently provide any views or evaluation on their relative ability to communicate the desired information. The inclusion of color vision deficient maps in Figure 1 (as suggested above) might help demonstrate how some of the colormaps shown are perhaps less suitable than others (e.g. repeated colors in Turbo).
Specific comments
1. p1, line 4-5: I would suggest the end of the abstract is reworded to “…applications, which potentially limit these map’s accessibility and ability to effectively communicate information. In this paper, I provide colormap recommendations to facilitate consistent, intuitive, and accessible fire progression mapping.”
2. p1, line 6: Change “Wildland fire’s movement..” to “The movement of wildland fires…”
3. p1, line 16: Unnecessary comma after parentheses.
4. p2, line 31: Define or explain “non-colorsafe” (and hence “color-safe” on line 34)
5. p2, line 31 and 34: Inconsistent spelling of colorsafe/color-safe
6. p2, line 34: Define or explain “508 Compliance”
7. p2, line 35: “2022” should not be in parentheses
8. p2, line 36: “2023” should not be in parentheses
9. p2, line 37: “if the maps change” - I suggest you clarify here that you mean that if the colourmaps (or other presentational elements) of the FPMs change. What the maps actually show (i.e. the fire perimeters) would of course be expected to change as new/updated data is included.
10. p2, line 42: “I” not “we”
11. p2, line 46-47: Reference should be in parentheses
12. p2, line 50: “I” not “we”
13. p3, Figure caption: I suggest that the caption is expanded to include more details about what is being shown in Figure 1. However please see my general comments above for further suggestions about Figure 1.
14. p4, line 69: “fires” not “fire”
15. p4, line 71-72: “The four suggested colormaps (Figure 1), one of which is an example in GeoOps (Figure A1d), intend to provide a starting point.” - It is not immediately clear why Figure 1 is stated to show (only) four suggested colormaps rather than six. However please see my general comments above for further suggestions about Figure 1.I look forward to seeing this published in due course!
Citation: https://doi.org/10.5194/gc-2024-9-RC2 -
AC2: 'Reply on RC2', Benjamin Hatchett, 20 Feb 2025
Dear Dr. Westaway,
Thank you for your detailed review and your thoughtful and helpful suggestions for revisions. Your suggestions to re-work Figure 1 to better highlight the colourblindness simulations to better support the narrative of the paper was particularly valuable and appreciated. Your additional suggestions and detail-oriented edits led to improvements throughout. I have revised the manuscript to follow your suggestions and edits, leading to improvements in content and clarity throughout. Please see the standard response-to-reviewer formatted document in the attached supplement.
Sincerely,
Ben Hatchett
-
AC2: 'Reply on RC2', Benjamin Hatchett, 20 Feb 2025
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 123 | 28 | 11 | 162 | 9 | 11 |
- HTML: 123
- PDF: 28
- XML: 11
- Total: 162
- BibTeX: 9
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 79 | 46 |
| Germany | 2 | 9 | 5 |
| United Kingdom | 3 | 9 | 5 |
| France | 4 | 8 | 4 |
| Brazil | 5 | 7 | 4 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 79