cmglaciology

Colormaps for glaciology, packaged for Matplotlib.

Name	Type	Intended use
speed	Sequential	Ice surface speed
akbathtopo	Topographic	Bathymetry and surface elevation, Alaska
aktopo	Topographic	Surface elevation (land only), Alaska
grisbathtopo	Topographic	Bathymetry and surface elevation, Greenland
gristopo	Topographic	Surface elevation (land only), Greenland

Contributions are welcome. Do you have a favorite colormap for ice thickness? Please let us know.

Install

python -m pip install cmglaciology

For the development version:

python -m pip install git+https://github.com/pism/cmglaciology.git

or, from a clone:

python -m pip install -e .

Usage example

import cmglaciology.cm as cmg
import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(0, 1, 100)[np.newaxis, :]

plt.imshow(x, aspect='auto', cmap=cmg.speed)
plt.axis('off')
plt.show()

For a reversed colormap, append _r to the colormap name:

plt.imshow(x, aspect='auto', cmap=cmg.speed_r)

For a discretized colormap, use the .resampled method on any of the colormaps:

plt.imshow(x, aspect='auto', cmap=cmg.speed.resampled(25))

Alternatively, the registered name string can be used. Names are prefixed with cmg.:

import cmglaciology  # required in order to register the colormaps with Matplotlib
...
plt.imshow(x, aspect='auto', cmap='cmg.speed')
plt.imshow(x, aspect='auto', cmap='cmg.speed_r')

For an image of all the available colormaps:

from cmglaciology import show_cmaps

show_cmaps()

To also see how the colormaps look with a color vision deficiency, pass cvd=True. This needs the colorspacious package, which the cvd extra installs:

python -m pip install "cmglaciology[cvd]"

show_cmaps(cvd=True)

simulate_cvd returns a single colormap as it is seen with a deficiency, to use in a plot of your own:

from cmglaciology import simulate_cvd

plt.imshow(x, aspect='auto', cmap=simulate_cvd(cmg.speed, "deuteranomaly"))

See the figure at the bottom.

Notes on the colormaps

The colormaps were designed in QGIS against data values, so their color stops are not evenly spaced. The stops are (data value → position in the colormap is linear between the first and last value):

• speed: 10, 30, 100, 250, 750 (m/yr)
• akbathtopo: -2000, 0, 1, 1250, 2000, 2500, 3000, 3500 (m)
• aktopo: 0, 1, 1250, 2000, 2500, 3000, 3500 (m)
• grisbathtopo: -500, 0, 1, 1500, 2000 (m)
• gristopo: 0, 1, 1500, 2000 (m)

To reproduce the QGIS rendering, use vmin/vmax equal to the first and last stop. The topographic maps have a sharp break at sea level (0–1 m).

Adding a colormap

1. Export the color ramp from QGIS (Layer Properties → Symbology → Save color map to file) and put the file in qgis/.
2. Run python scripts/qgis2txt.py to regenerate the 256-level RGB tables in cmglaciology/cmaps/.
3. Add the name to one of the groups at the top of cmglaciology/cm.py.
4. Run python cm.py from the cmglaciology/ directory to refresh colormaps.png, python scripts/show_cvd.py to refresh colormaps_cvd.png, and pytest to check everything.

Releasing

Development happens on main; releases are cut from the release branch by the ASFHyP3 actions workflows in .github/workflows/, as in pism-terra:

1. Every pull request to main adds an entry to CHANGELOG.md under a heading with the next version number, e.g. ## [0.2.0] (or carries the bumpless label if it changes nothing a user would notice).
2. To release, open a pull request from main to release and give it one of the labels major, minor, patch or bumpless.
3. Merging it tags the new version (v0.2.0), which creates a GitHub release with the matching changelog section as its notes, brings release back into main, and publishes the package to PyPI.

The version number of an installed package comes from the git tags, through setuptools_scm.

Acknowledgements

The structure of this package — one text file of RGB values per colormap, loaded and registered with Matplotlib on import, with reversed _r versions and a show_cmaps() overview — is taken from cmcrameri by Callum Rollo and contributors, which packages Fabio Crameri's Scientific colour maps. cmglaciology/cm.py and the tests are adapted from cmcrameri's code, which is distributed under the MIT License, Copyright (c) 2020 Callum Rollo. Many thanks to Callum and the cmcrameri contributors for making this so easy to reuse. Any errors in this adaptation are our own.

No colormap data from cmcrameri or the Scientific colour maps is included here. If you need perceptually uniform colormaps, use cmcrameri directly.

License

MIT, see LICENSE.txt.

Color vision deficiency

The colormaps as they are seen with normal vision and with protanopia, deuteranopia and tritanopia, in which the red, green or blue cones do not work:

The figure is made with show_cmaps(cvd=True), by python scripts/show_cvd.py. The simulation follows Machado et al. (2009), as implemented in colorspacious. With protanopia and deuteranopia, the orange-to-magenta upper part of speed turns into a single olive ramp, so high speeds are harder to tell apart.

Example of Kennicott Glacier, Alaska, using aktopo and speed colormaps, rendered with pyvista: