r"""
Time-lapse monitoring
=====================

Repeat EM surveys track how the subsurface *changes* — a rising water
table, a spreading plume, managed recharge. :class:`~pycsamt.interp.TimeLapseEM`
holds a baseline model plus one or more monitor surveys on the same grid and
differences them, and the plot classes turn those differences into
monitoring figures. This example simulates a wetting front with three
surveys of the synthetic section.
"""

# %%
# Build the time series
# ---------------------
# We reuse the shared model with an increasing ``wet`` factor, which lowers
# the aquifer resistivity as pore water increases — a stand-in for three
# repeat surveys over a wetting aquifer.

from _interp_data import demo_model

from pycsamt.interp import (
    TimeLapseEM,
    assert_compatible_grids,
)

# Use the difference section (1st figure) as the card thumbnail.
# sphinx_gallery_thumbnail_number = 1

surveys = [
    demo_model(wet=1.0, seed=1),
    demo_model(wet=1.6, seed=2),
    demo_model(wet=2.4, seed=3),
]
assert_compatible_grids(surveys)  # same grid is required to difference
tl = TimeLapseEM(surveys, labels=["baseline", "month 1", "month 2"])
print("surveys:", tl.labels)

# %%
# The difference section
# ----------------------
# :class:`~pycsamt.interp.plot.PlotTimeLapseSection` images
# :math:`\Delta\log_{10}\rho` between a chosen survey and the baseline. A
# coherent negative anomaly (resistivity dropping) traces where water has
# been added since the baseline.

from pycsamt.interp.plot import (
    PlotMultiTimeLapseGrid,
    PlotTimeLapseSection,
)

PlotTimeLapseSection(tl, quantity="rho").plot()

# %%
# All surveys at a glance
# -----------------------
# :class:`~pycsamt.interp.plot.PlotMultiTimeLapseGrid` lays the surveys out
# in a row under one shared colour scale, so the progression of the change
# reads left to right — the standard monitoring montage.

PlotMultiTimeLapseGrid(tl).plot()

# %%
# **Reading it.** The difference section localises the change to the aquifer
# interval, and the grid shows it strengthening survey to survey — exactly
# the wetting front built into the models. On real monitoring data, always
# check that changes sit *inside* known hydrogeological units (here the
# aquifer) and not at the surface or edges, where they usually mean
# acquisition or static-shift drift rather than real subsurface change.