How to Create Time Slices

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

In this example we show how to create a filled contourplot of transient data over a sampling line. For this purpose we use a component transport example from the ogs benchmark gallery (https://www.opengeosys.org/docs/benchmarks/hydro-component/elder/).

To see this benchmark results over all timesteps have a look at How to create Animations.

Let’s load the data which we want to investigate.

import matplotlib.pyplot as plt
import numpy as np
import pyvista as pv

import ogstools as ot
from ogstools import examples

mesh_series = examples.load_meshseries_CT_2D_XDMF().scale(time=("s", "a"))
y = mesh_series[0].center[1]  # flat y coordinate of this 2D mesh is not 0
si = ot.variables.saturation

Now we setup two sampling lines.

pts_vert = np.linspace([25, y, -75], [25, y, 75], num=100)
pts_diag = np.linspace([25, y, 75], [100, y, 0], num=100)
fig = mesh_series.mesh(-1).plot_contourf(si, vmin=0)
fig.axes[0].plot(pts_vert[:, 0], pts_vert[:, 2], "-k", linewidth=3)
fig.axes[0].plot(pts_diag[:, 0], pts_diag[:, 2], "-.k", linewidth=3)

[<matplotlib.lines.Line2D object at 0x7b2706103a30>]

Here, we first show a regular line sample plot for the vertical sampling line for each timestep.

fig, ax = plt.subplots(figsize=[15, 8])
for mesh, timevalue in zip(mesh_series, mesh_series.timevalues, strict=True):
    sample = pv.PolyData(pts_vert).sample(mesh)
    color = str(0.8 * timevalue / mesh_series.timevalues[-1])
    label = f"{timevalue:.1f} a"
    fig = ot.plot.line(
        sample, si, "z", ax=ax, label=label, color=color, fontsize=20
    )

As the above kind of plot is getting cluttered for lots of timesteps we provide a function to create a filled contour plot over the transient data. The function plot_time_slice() automatically detects if the line lies on a cardinal direction and labels the y-axes with the changing spatial dimension.

fig = mesh_series.plot_time_slice(si, pts_vert)

By default the plot is smoothened with interpolation. When deactivated, we see the edges of the raw sampled data. When using the interpolation, be sure to adjust the number of sampling points if the MeshSeries contains a lot of small timesteps.

fig = mesh_series.plot_time_slice(si, pts_vert, interpolate=False)

If the line doesn’t point in a cardinal direction the distance along the line is used for the y-axis by default. You can however, specify if you want to use spatial dimension via the argument “y_axis”. This may be useful when plotting data of an edge / boundary of the mesh.

fig = mesh_series.plot_time_slice(si, pts_diag)