issue_comments

I think the second example shows this: http://xarray.pydata.org/en/stable/plotting.html#maps

Related: #3169

We should add this to the docs then

I think this can be closed, as the following works:

```python import cartopy.crs as ccrs import xarray as xr

proj = ccrs.PlateCarree()

ds = xr.tutorial.open_dataset("air_temperature") da = ds.air.isel(time=slice(None, 3)) da.plot(transform=proj, col='time', subplot_kws={'projection':proj}) ```

I don't understand what you mean by 'adding you own to the signature'. Adding my own what to what signature?

;-) I meant adding your own keyword argument to the function's signature

I don't understand what you mean by 'adding you own to the signature'. Adding my own what to what signature?

However it looks like no xarray code contains the transform keyword. However I definitely can use it in calls to plot(). I don't understand what is going on. Can someone enlighten me?

Yes, this is what I meant with xarray not being directly "cartopy aware". The reason is that xarray intends to be a general library, and tries to avoid defining too many keyword arguments in order to avoid confusion. However, xarray passes along any kw argument you want to the underlying mpl plot function (https://github.com/pydata/xarray/blob/master/xarray/plot/plot.py#L379). You'll have to catch it first, either from the kwargs dict or by adding your own to the signature (the latter being recommended only if necessary, for the reason I mentioned above)

I am confused.

I planned to identify the plotting function that accepted the transform keyword (used in the code sample above), and add its value in subplot_kws if it wasn't already.

However it looks like no xarray code contains the transform keyword. However I definitely can use it in calls to plot(). I don't understand what is going on. Can someone enlighten me?

I don't really understand what you mean here. Neither facetted nor non-facetted plots have any logic to greate geoaxes automatically.

Oops, I was indeed mixed up.

If we add a projection argument for faceted plots, we should certainly add it for non-faceted plots, too.

This seems totally reasonable to me, and would be nice because it would improve the consistency between facetted and non-facetted plotting.

I don't really understand what you mean here. Neither facetted nor non-facetted plots have any logic to greate geoaxes automatically.

Otherwise, yes, I think that any logic making working with cartopy a little easier would be a good addition. The logic creating the axes is located here

This seems totally reasonable to me, and would be nice because it would improve the consistency between facetted and non-facetted plotting.

Going even further, we might even default to projection=ccrs.PlateCarree() if a cartopy transform is provided (see https://github.com/SciTools/cartopy/issues/645).

@fmaussion @jhamman any opinions?