How to plot data at observation points

Section author: Florian Zill (Helmholtz Centre for Environmental Research GmbH - UFZ)

In this example we plot the data values on observation points over all timesteps. Since requested observation points don’t necessarily coincide with actual nodes of the mesh different interpolation options are available. See ogstools.meshlib.mesh_series.MeshSeries.probe for more details. Here we use a component transport example from the ogs benchmark gallery (https://www.opengeosys.org/docs/benchmarks/hydro-component/elder/).

import matplotlib.pyplot as plt
import numpy as np

import ogstools as ot
from ogstools import examples

mesh_series = examples.load_meshseries_CT_2D_XDMF().scale(time=("s", "a"))
si = ot.variables.saturation

To read your own data as a mesh series you can do:

from ogstools.meshlib import MeshSeries
mesh_series = MeshSeries("filepath/filename_pvd_or_xdmf")

You can also use a variable from the available presets instead of needing to create your own: Variable presets and data transformation

Let’s define 4 observation points and plot them on the mesh.

points = np.asarray([[x, 0, z] for z in [60, 40] for x in [0, 40, 80, 120]])
fig = mesh_series.mesh(0).plot_contourf(si)
fig.axes[0].scatter(points[:, 0], points[:, 2], s=50, fc="none", ec="r", lw=3)
for i, pt in enumerate(points):
    fig.axes[0].annotate(str(i), (pt[0], pt[2] - 5), va="top", fontsize=32)

And now probe the points and the values over time:

labels = [
    f"{i}: {label}"
    for i, label in enumerate(ot.plot.utils.justified_labels(points))
]
fig = mesh_series.plot_probe(points=points[:4], variable=si, labels=labels[:4])

You can also pass create your own matplotlib figure and pass the axes object. Additionally, you can pass any keyword arguments which are known by matplotlibs plot function to further customize the curves. In this case marker and linewidth are not part of the API of plot_probe but get processed correctly anyway. If you want to have more freedom with the data you can just do the probing, adapt to your needs and then do the plotting yourself:

fig, axs = plt.subplots(nrows=2, figsize=[10, 5])
mesh_series.plot_probe(
    points[:4], si, ax=axs[0], colors=["k"], labels=labels[:4], marker="."
)
mesh_series.plot_probe(
    points[4:], si, ax=axs[1], linestyles=["-"], labels=labels[4:], linewidth=1
)
# add the mean of the observation point timeseries
values = si.transform(mesh_series.probe(points, data_name=si.data_name))
mean_values = np.mean(values.reshape((-1, 2, 4)), axis=-1)
ts = mesh_series.timevalues
for index in range(2):
    fig.axes[index].plot(ts, mean_values[:, index], "rk"[index], label="mean")
    fig.axes[index].legend()