Colorful science

One of the joys of the modern age of publishing is that things can be published in color.  That, unfortunately, carries its own hazard.  Many have noted, for instance, that just under a tenth of all people are color blind in one way or another and it seems like one of the professional news magazines carries an admonition about this from time to time (for instance, in GSA Today and in EOS).  Back in the day of the black and white press, this was no issue, but now folks practically use color even when it isn’t really necessary. The result is that some figures are hostile to every twelfth reader or so. Now for papers on a computer, color blind folks can try software that takes the stuff that looks the same to them and make it look different, but that is a burden and imperfect solution. So generally folks are encouraged to use tools to make sure that colors are distinguishable. One such approach is to substitute a diverging color map for a default rainbow color map. (Another problem with the rainbow is there is way too much green in it).

This is all noble and good, and GG is well aware of publishing some figures that are color-blind-hostile (see below), but this shouldn’t be the end of it.  GG has not been too happy with the diverging color map either; this is really accentuated in printed materials because the intensity of the color might vary across a page (also, the apparent intensity varies depending on the background).  You might look at a point and try and estimate a value and fail because you are comparing your impression of one color with another on a scale bar. All too frequently a diverging color map ends up being a three-value scale (red, white, and blue).

So in addition to finding a color scheme that can appeal to all readers, GG would like to encourage a return of sorts to an older style of things, namely the contour plot. This doesn’t require contours per se, but does mean that instead of using continuous colors that the spectrum have steps. Some workers do this regularly (for instance, Thorsten Becker’s papers do this well) and others not so much. Consider two examples we discussed for other reasons long ago:


Mantle topography from Vs model of Shen et al, from Levandowski et al. 2014.


From Becker et al., EPSL 2014

GG is a coauthor on the top figure and, well, good luck deciding if an element in Nevada is -1.2 or -1.4, while the same kind of information is present in the bottom figure, which in addition to having a friendlier color scheme, the use of discrete color levels allows the reader to know what the value is (within the uncertainty of an individual color level).  You can even sneak in some more information if you are careful by adding shading to an image like this (though often shading is used to show something separate, like relief/topography).

All too often we churn out figures quickly using default settings to look at things and don’t revisit our choices when it comes to picking out the figures to publish.  Arguably a good thing to have at a journal might be a visual editor (somebody to complement the copy editor) who might help catch issues like these.


2 responses to “Colorful science”

  1. Paul Braterman says :

    Continuous colour maps are even worse than you suggest. The way we perceive a specific colour depends on those surrounding it, in nonobvious ways, and we can’t stop our brains from doing this even if we are aware of it. And a patch of colour on the map, surrounded by other rapidly changing colours in every direction, is being compared with a patch in the very different context of the scale bar.

    Stepped colour is much easier to read without such distortion, across both steep and shallow gradients.


Trackbacks / Pingbacks

  1. Mapa Culpa | The Grumpy Geophysicist - December 18, 2017

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: