Maps are smaller in V1 than in V0.5

[jeanbournhonesqueMF]

What happened?

Greetings,

I'm trying to use earthkit-plots V1.
I found that the size of the maps changed a lot between 0.5 and 1.0 :

Earthkit plots V1

Earthkit plots V0.5

The 0.5 version was way better for me because the map is larger.
I also have the impression the coastlines don't have a black contour anymore ...

Regards,
Jean

What are the steps to reproduce the bug?

I ran the same python script with two different versions of earthkit (using different modules from ecs)

xxxab6-200 ~ (new_epsgram*) $ uv pip show earthkit-plots
Using Python 3.13.13 environment at: /usr/local/apps/python3/3.13.13-01
Name: earthkit-plots
Version: 1.0.1
Location: /usr/local/apps/python3/3.13.13-01/lib/python3.13/site-packages
Requires: adjusttext, cartopy, earthkit-data, earthkit-utils, matplotlib, numpy, pandas, pint, plotly, pyyaml
Required-by: earthkit

and

xxx@ab6-200 ~ (new_epsgram*) $ uv pip show earthkit-plots
Using Python 3.12.11 environment at: /usr/local/apps/python3/3.12.11-01
Name: earthkit-plots
Version: 0.5.0
Location: /usr/local/apps/python3/3.12.11-01/lib/python3.12/site-packages
Requires: adjusttext, cartopy, earthkit-data, earthkit-plots-default-styles, matplotlib, numpy, pandas, pint, plotly, pyyaml
Required-by: earthkit

The following code is the one I used to create the pictures above

import earthkit.data as ekd
import earthkit.plots as ekp
from earthkit.plots.geo import domains
# Change to from earthkit.plots.geography import domains with earthkit 1.0
import cartopy.crs as ccrs


def telecharge_anomalie() -> ekd.FieldList:
    """Returns l'anomalie"""
    return ekd.from_source(
        "mars",
        {
            "class": "od",
            "expver": 1,
            "stream": "eefo",
            "repres": "ll",
            "levtype": "sfc",
            "type": "taem",
            "param": "167.171",  # T2m anomaly
            "area": [80, -40, 25, 70],
            "date": "20260629",
            "grid": [0.5, 0.5],
            "time": "00",
            "step": "0-168",
        },
        log=None,
    ).to_fieldlist()


domaine_europe = domains.Domain.from_bbox(
    bbox=[-18, 26, 35, 63],
    name="Europe",
)
crs = ccrs.Mercator()


def initialisation_carte() -> ekp.Map:
    """Returns une map standards prête à utilisation"""
    carte = ekp.Map(domain=domaine_europe, crs=crs)
    carte.borders()
    carte.gridlines()
    carte.coastlines(color="black")
    return carte


def plot_anomalie() -> None:
    anomalie = telecharge_anomalie()
    style_anomalie = ekp.styles.Style(
        levels=[-20, -10, -6, -3, -1, -0.5, 0.5, 1, 3, 6, 10, 20],
        colors=[
            "#1933B3",
            "#2E6ED9",
            "#5999EB",
            "#99C9F7",
            "#CBEDFE",
            "#FFFFFF",
            "#FFFFFF",
            "#FED3CB",
            "#E57F7F",
            "#D93333",
            "#B21A00",
            "#580B00",
        ],
        extend="both",
    )
    carte_anomalie = initialisation_carte()
    carte_anomalie.contourf(anomalie, style=style_anomalie, alpha=0.8)
    carte_anomalie.title(
        f"Test",
        fontsize=16,
    )
    carte_anomalie.legend(location="right")
    carte_anomalie.save(f"anomalie_test.jpg")


plot_anomalie()

Version

1.0

Platform (OS and architecture)

Linux

Relevant log output

Accompanying data

No response

Organisation

Meteo-France

[JamesVarndell]

Hi @jeanbournhonesqueMF, many thanks for raising this issue.

There are two things happening here - firstly, the default zorder values for things like coastlines and borders were removed at 1.0. The idea was to stick with cartopy defaults - that is, respect the order in which things are plotted for zorder by default. This is what is making your coastlines look fainter, as they are now beneath your 80% opaque data layer.

However, you are not the first person to get caught out by this, and on reflection it seems more reasonable to give coastlines a default zorder above the default data layer zorder. This has been implemented in #238.

As for the issue of default figure sizes, this was reduced as a stylistic choice, but I take on board your comments and will reconsider the default sizing. In the meantime, if you would like a bigger plot, I would recommend passing figsize when you instantiate your Map, like this:
carte = ekp.Map(domain=domaine_europe, crs=crs, figsize=(10, 10))

Many thanks!