Everyone sees colours slightly differently. With a multitude of colour options available to make complex climate-related maps & line-graphs, which do you choose? Which colour scale is best for you?
Below is an example for mean annual temperature in the UK, which ranges from around 5°C to 11°C. It is shown using four different colour scales (rows). Also shown (columns) are simulations of how those maps might appear to those with strong forms of the two most common forms of colour blindness, deutranopia & protanopia, which together effect around 8% of men and 0.4% of women. You will have a friend or colleague with colour blindness.
Am interested to know which colour scheme works best for you?
For example, I think the rainbow scheme (jet) artificially enhances certain regions and gradients because of the sharp differences in colour hue around yellow, and should not be used. The red-blue scheme, which is great for anomalies, might also overly enhance the boundary between red and blue at 8°C. Parula has less differentiation between colours but doesn’t artificially create boundaries where there are none.
My favourite is cube-helix for this case, because the colours are clear, and it has no obvious sharp boundaries. Other colour schemes are available of course, and which one is best will likely depend on the situation and personal preference.
Also, there might be issues with colour blindness for both jet & parula, as the colours blur together a little more. But, everyone will see colours differently, depending on monitors, the printer used, lighting conditions & and the strength of any colour blindness etc. As a rule-of-thumb, putting red & green in the same scale most easily causes confusion.
As a whole, scientists (including me) need to think more about which colour scale is most appropriate for each different type of plot, whether it be map, line graph or histogram. The choices for maps should be different depending on whether anomalies or absolute values are being shown, or whether the values are single signed or not.