Geographic Projections and Transforms

Geographic data that is on a sphere can be visualized by projecting that data onto a representation of that sphere flattened into 2d space. There exist a number of projection types, which can be found in the the cartopy documentation. With support from cartopy, yt now supports these projection types for geographically loaded data. Underlying data is assumed to have a transform of PlateCarree, which is data on a flattened, rectangular, latitude/longitude grid. This is a a typical format for geographic data.

The distinction between the data transform and projection is worth noting. The data transform is what system your data is defined with and the data projection is what the resulting plot will display. For more information on this difference, refer to the cartopy documentation on these differences. If your data is not of this form, feel free to open an issue or file a pull request on the yt github page for this feature.

It should be noted that these projections are not the same as yt’s ProjectionPlot. For more information on yt’s projection plots, see Types of Projections.

Installing Cartopy

In order to access the geographic projection functionality, you will need to have an installation of cartopy available on your machine. Please refer to Cartopy’s documentation for detailed instructions

Using Basic Transforms

As mentioned above, the default data transform is assumed to be of PlateCarree, which is data on a flattened, rectangular, latitude/longitude grid. To set something other than PlateCarree, the user can access the dictionary in the coordinate handler that defines the coordinate transform to change the default transform type. Because the transform describes the underlying data coordinate system, the loaded dataset will carry this newly set attribute and all future plots will have the user-defined data transform. Also note that the dictionary is ordered by axis type. Because slicing along the altitude may differ from, say, the latitude axis, we may choose to have different transforms for each axis.

ds = yt.load_uniform_grid(data, sizes, 1.0, geometry=("geographic", dims), bbox=bbox)
ds.coordinates.data_transform["altitude"] = "Miller"
p = yt.SlicePlot(ds, "altitude", "AIRDENS")

In this example, the data_transform kwarg has been changed from its default of PlateCarree to Miller. You can check that you have successfully changed the defaults by inspecting the data_transform and data_projection dictionaries in the coordinate handler. For this dataset, that would be accessed by:

print(ds.coordinates.data_transform["altitude"])
print(ds.coordinates.data_projection["altitude"])

Using Basic Projections

All of the transforms available in Cartopy v0.15 and above are accessible with this functionality.

The next few examples will use a GEOS dataset accessible from the yt data downloads page. For details about loading this data, please see cookbook-geographic_projections.

If a geographic dataset is loaded without any defined projection the default option of Mollweide will be displayed.

ds = yt.load_uniform_grid(data, sizes, 1.0, geometry=("geographic", dims), bbox=bbox)
p = yt.SlicePlot(ds, "altitude", "AIRDENS")

If an option other than Mollweide is desired, the plot projection type can be set with the set_mpl_projection function. The next code block illustrates how to set the projection to a Robinson projection from the default PlateCarree.

ds = yt.load_uniform_grid(data, sizes, 1.0, geometry=("geographic", dims), bbox=bbox)
p = yt.SlicePlot(ds, "altitude", "AIRDENS")
p.set_mpl_projection("Robinson")
p.show()

The axes attributes of the plot can be accessed to add in annotations, such as coastlines. The axes are matplotlib GeoAxes so any of the annotations available with matplotlib should be available for customization. Here a Robinson plot is made with coastline annotations.

p.set_mpl_projection("Robinson")
p.render()
p.plots["AIRDENS"].axes.set_global()
p.plots["AIRDENS"].axes.coastlines()
p.show()

p.render() is required here to access the plot axes. When a new projection is called the plot axes are reset and are not available unless set up again.

Additional arguments can be passed to the projection function for further customization. If additional arguments are desired, then rather than passing a string of the projection name, one would pass a 2 or 3-item tuple, the first item of the tuple corresponding to a string of the transform name, and the second and third items corresponding to the args and kwargs of the transform, respectively.

Alternatively, a user can pass a transform object rather than a string or tuple. This allows for users to create and define their own transforms, beyond what is available in cartopy. The type must be a cartopy GeoAxes object or a matplotlib transform object. For creating custom transforms, see the matplotlib example.

The function set_mpl_projection() accepts several input types for varying levels of customization:

set_mpl_projection("ProjectionType")
set_mpl_projection(("ProjectionType", (args)))
set_mpl_projection(("ProjectionType", (args), {kwargs}))
set_mpl_projection(cartopy.crs.PlateCarree())

Further examples of using the geographic transforms with this dataset can be found in cookbook-geographic_projections.