sensible colours for mess plots

[plantarum]

When plotting the results of the mess analysis, the default behaviour uses a colour ramp that is difficult to interpret. Using the example from the docs, its impossible to see where 0 is, so we can't tell which areas are outside the range of the reference values:

As I understand it, the expected values range from maximum similarity at 100, down to 'minimum' similarity at 0, with negative values indicating increasingly 'dissimilar' values. We could use the col argument to plot.SpatRaster to produce more intuitive colors:

colMat <- data.frame(from = c(50, 25, 5, 0, -5, -25, -50, -100,
                              -Inf),
                     to = c(100, 50, 25, 5, 0, -5, -25, -50,
                              -100),
                     col = c("green", "aquamarine", "lightblue", "blue", 
                             "beige", "yellow", "brown",
                             "orange", "red"))

plot(mss[[10]], col = colMat)

Those are not particularly ideal colours or intervals, but that seems better than the default.

This could be included in the example in the docs so users could set their own colours. Or perhaps it would be worth creating a class for mess output, such that a new plot method could be provided that would generate a nice colour scheme? If either of these sound useful I could put together a pull request.

• tyler

[rhijmans]

Thank you for the suggestions. I have added something similar to the example.

I do not want to create a new class. Instead I now plan to add support in "terra" for setting a "color table" with intervals to a SpatRaster such that the mess function could add a default legend to the output SpatRaster.

[plantarum]

Perfect, thanks! I didn't know about the breaks argument, that makes for a straightforward plot.