Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions changelog-entries/793.md
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
- Added the `wolf-sheep-soil-creep` tutorial [#793](https://github.com/precice/tutorials/pull/793)
58 changes: 58 additions & 0 deletions wolf-sheep-soil-creep/README.md
Comment thread
uekerman marked this conversation as resolved.
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
---
title: Wolf-sheep-grass model with soil creep
permalink: tutorials-wolf-sheep-soil-creep.html
keywords: MESA, Landlab, wolf-sheep-grass, soil creep, ABM, agent-based modeling
summary: Example of bi-directional ABM-PDE coupling via preCICE.
---

{% note %}
Get the [case files of this tutorial](https://github.com/precice/tutorials/tree/develop/wolf-sheep-soil-creep), as continuously rendered here, or see the [latest released version](https://github.com/precice/tutorials/tree/master/wolf-sheep-soil-creep) (if there is already one). Read how in the [tutorials introduction](https://precice.org/tutorials.html).
{% endnote %}

## Setup

This tutorial is based on Landlab's [Wolf-Sheep-Grass Model with Soil Creep](https://landlab.csdms.io/tutorials/agent_based_modeling/wolf_sheep/wolf_sheep_with_soil_creep.html) example, which couples the [Wolf-Sheep model from Mesa](https://mesa.readthedocs.io/latest/examples/advanced/wolf_sheep.html) with a Landlab soil creep model in a single notebook-based workflow. The [source notebook](https://github.com/landlab/landlab/blob/master/docs/source/tutorials/agent_based_modeling/wolf_sheep/wolf_sheep_with_soil_creep.ipynb) is licensed under [MIT](https://landlab.csdms.io/about/license.html).

The purpose of this tutorial is to demonstrate a simple two-way coupling between an agent-based model (ABM) and a grid-based continuum model via preCICE. The example extends the canonical wolf-sheep-grass ABM with a simple soil-creep feedback on a shared sloping raster grid. When sheep eat grass, the soil beneath the damaged grass becomes more mobile, increasing the local soil transport efficiency and affecting downslope soil transport. The feedback also acts in the other direction: soil thickness affects grass growth by preventing grass from growing where the soil becomes too thin.

In this implementation, the Mesa wolf-sheep-grass model and the Landlab soil creep model are two separate preCICE participants, exchanging grass cover and soil depth during the simulation. The Mesa participant extracts the current grass map from the wolf-sheep-grass model and sends it to the Landlab participant. The Landlab participant uses this grass map to adapt the soil creep coefficient, evolves the soil thickness, and sends the updated soil depth back to the Mesa participant, where it is used to limit grass growth. This example is intentionally simple and should be understood as a demonstration case for coupling ABM and PDE-style models via preCICE, rather than as a physically calibrated landscape evolution model.

## Configuration

preCICE configuration (image generated using the [precice-config-visualizer](https://precice.org/tooling-config-visualization.html)):

![preCICE configuration visualization](images/tutorials-wolf-sheep-soil-creep-precice-config.png)

## Available solvers

Soil-Creep PDE participant:

* Landlab. Numerical modeling of Earth surface dynamics. For more information, have a look at the [Landlab documentation](https://landlab.csdms.io/index.html).

Wolf-Sheep-Grass ABM participant:

* MESA. Agent-based modeling (or ABM) framework. For more information have a look at the [MESA documentation](https://mesa.readthedocs.io/latest/index.html)

## Running the simulation

Open two separate terminals and start the soil creep and wolf sheep participants by calling the respective run scripts `run.sh` located in each of the participants' directories. For example:

```bash
cd soil-creep-landlab
./run.sh
```

and

```bash
cd wolf-sheep-grass-mesa
./run.sh
```

## Post-processing

The soil-creep participant generates Python-based visualizations, which are written to `soil-creep-landlab/output`. The following examples were generated by setting `rng=42` in the `WolfSheepScenario` for the wolf-sheep-grass participant.

![erosion/deposition patterns](images/tutorials-wolf-sheep-soil-creep-erosion_deposition_patterns.png)
![soil thickness](images/tutorials-wolf-sheep-soil-creep-soil_thickness.png)
![grass map](images/tutorials-wolf-sheep-soil-creep-grass_map.png)
1 change: 1 addition & 0 deletions wolf-sheep-soil-creep/clean-tutorial.sh
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
69 changes: 69 additions & 0 deletions wolf-sheep-soil-creep/precice-config.xml
Comment thread
MakisH marked this conversation as resolved.
Original file line number Diff line number Diff line change
@@ -0,0 +1,69 @@
<?xml version="1.0" encoding="UTF-8" ?>
<precice-configuration>
<log>
<sink
filter="%Severity% > debug"
format="---[precice] %ColorizedSeverity% %Message%"
enabled="true" />
</log>

<data:scalar name="Grass" />
<data:scalar name="Soil" />

<mesh name="Wolf-Sheep-Grass-Mesh" dimensions="2">
<use-data name="Grass" />
</mesh>

<mesh name="Soil-Creep-Mesh" dimensions="2">
<use-data name="Grass" />
</mesh>

<mesh name="Soil-Grass-Mesh" dimensions="2">
<use-data name="Soil" />
</mesh>

<mesh name="Soil-Depth-Mesh" dimensions="2">
<use-data name="Soil" />
</mesh>

<participant name="Wolf-Sheep-Grass">
<receive-mesh name="Soil-Depth-Mesh" from="Soil-Creep" />
<provide-mesh name="Wolf-Sheep-Grass-Mesh" />
<provide-mesh name="Soil-Grass-Mesh" />
<write-data name="Grass" mesh="Wolf-Sheep-Grass-Mesh" />
<read-data name="Soil" mesh="Soil-Grass-Mesh" />
<mapping:nearest-neighbor
direction="read"
from="Soil-Depth-Mesh"
to="Soil-Grass-Mesh"
constraint="consistent" />
</participant>

<participant name="Soil-Creep">
<receive-mesh name="Wolf-Sheep-Grass-Mesh" from="Wolf-Sheep-Grass" />
<provide-mesh name="Soil-Creep-Mesh" />
<provide-mesh name="Soil-Depth-Mesh" />
<read-data name="Grass" mesh="Soil-Creep-Mesh" />
<write-data name="Soil" mesh="Soil-Depth-Mesh" />
<mapping:nearest-neighbor
direction="read"
from="Wolf-Sheep-Grass-Mesh"
to="Soil-Creep-Mesh"
constraint="consistent" />
</participant>

<m2n:sockets acceptor="Wolf-Sheep-Grass" connector="Soil-Creep" exchange-directory=".." />

<coupling-scheme:serial-explicit>
<participants first="Soil-Creep" second="Wolf-Sheep-Grass" />
<time-window-size value="2" />
<max-time value="100" />
<exchange
data="Grass"
mesh="Wolf-Sheep-Grass-Mesh"
from="Wolf-Sheep-Grass"
to="Soil-Creep"
initialize="true" />
<exchange data="Soil" mesh="Soil-Depth-Mesh" from="Soil-Creep" to="Wolf-Sheep-Grass" />
</coupling-scheme:serial-explicit>
</precice-configuration>
1 change: 1 addition & 0 deletions wolf-sheep-soil-creep/soil-creep-landlab/.gitignore
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
output/
9 changes: 9 additions & 0 deletions wolf-sheep-soil-creep/soil-creep-landlab/clean.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,9 @@
#!/usr/bin/env sh
set -e -u

rm -rfv ./output/

. ../../tools/cleaning-tools.sh

clean_precice_logs .
clean_case_logs .
4 changes: 4 additions & 0 deletions wolf-sheep-soil-creep/soil-creep-landlab/requirements.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
numpy >1, <2
matplotlib
landlab
pyprecice~=3.0
23 changes: 23 additions & 0 deletions wolf-sheep-soil-creep/soil-creep-landlab/run.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
#!/usr/bin/env bash
set -e -u

. ../../tools/log.sh

exec > >(tee --append "$LOGFILE") 2>&1

if [ ! -v PRECICE_TUTORIALS_NO_VENV ]; then

if [ ! -d ".venv" ]; then
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt && pip freeze > pip-installed-packages.log
else
source .venv/bin/activate
fi

fi

mkdir -p output
python3 soil_creep.py

close_log
116 changes: 116 additions & 0 deletions wolf-sheep-soil-creep/soil-creep-landlab/soil_creep.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,116 @@
import copy
import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
from landlab import RasterModelGrid, imshow_grid
from landlab.components import LinearDiffuser
import precice

initial_soil_depth = 0.3
hstar = 0.2
fast_creep = 0.1
slow_creep = 0.001
Comment on lines +10 to +12
Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you add some comments here on what these parameters mean?


ground_cover_cmap = copy.copy(mpl.colormaps["YlGn"])

# Create a grid the same size as the W-S-G model's grid
width = 20
height = 20

rmg = RasterModelGrid((width, height))

# Create elevation field and have it slope down to the south at 10% gradient
elev = rmg.add_zeros("topographic__elevation", at="node")
elev[:] = 0.1 * rmg.y_of_node

# Have one open boundary on the south side
rmg.set_closed_boundaries_at_grid_edges(True, True, True, False)

# Remember the starting elevation so we can calculate cumulative erosion/deposition
initial_elev = np.zeros(rmg.number_of_nodes)
initial_elev[:] = elev

# Also remember the elevation of the prior time step, so we can difference
prior_elev = np.zeros(rmg.number_of_nodes)

# Create a field for the creep coefficient, and set parameters for two
# rates: slow (full grass cover) and fast (partial or "eaten" grass cover)
creep_coef = rmg.add_zeros("creep_coefficient", at="node")

# Create a soil-thickness field
soil = rmg.add_zeros("soil__depth", at="node")
Comment thread
joargu marked this conversation as resolved.
soil[:] = initial_soil_depth

# Instantiate a LinearDiffuser (soil creep) Landlab component
diffuser = LinearDiffuser(rmg, linear_diffusivity=creep_coef)

# preCICE setup
participant_name = "Soil-Creep"
config_file_name = "../precice-config.xml"
solver_process_index = 0
solver_process_size = 1
participant = precice.Participant(participant_name, config_file_name, solver_process_index, solver_process_size)

positions = [[x, y] for x in range(width) for y in range(height)]
vertex_gm_ids = participant.set_mesh_vertices("Soil-Creep-Mesh", positions)
vertex_soil_ids = participant.set_mesh_vertices("Soil-Depth-Mesh", positions)

participant.initialize()

while participant.is_coupling_ongoing():
solver_dt = 0.2 * rmg.dx * rmg.dx / fast_creep
precice_dt = participant.get_max_time_step_size()
dt = np.minimum(solver_dt, precice_dt)

gm = participant.read_data("Soil-Creep-Mesh", "Grass", vertex_gm_ids, dt)

# Assign the higher creep coefficient to cells where the grass has
# been eaten and not yet recovered; the slower value is assigned to
# "fully grown" grass patches.
creep_coef[gm.flatten() == 1] = fast_creep
creep_coef[gm.flatten() == 2] = slow_creep

# Adjust the creep coefficient to account for soil depth
creep_coef *= 1.0 - np.exp(-soil / hstar)

# Run the soil-creep model
prior_elev[:] = elev
diffuser.run_one_step(dt)
Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I guess this updates elev? Where could I see more about this?
(a comment could probably be good here)


# Update the soil cover
soil += elev - prior_elev

participant.write_data("Soil-Depth-Mesh", "Soil", vertex_soil_ids, soil)

participant.advance(dt)

participant.finalize()

# Calculate and plot the erosion/deposition patterns
plt.figure()
ero_dep = elev - initial_elev
maxchange = np.amax(np.abs(ero_dep))
imshow_grid(
rmg,
ero_dep,
vmin=-maxchange,
vmax=maxchange,
cmap="coolwarm_r",
colorbar_label="Depth of soil accumulation (+) or loss (-), m",
)
plt.savefig("output/erosion_deposition_patterns.png")
plt.close()

# Soil thickness
plt.figure()
imshow_grid(rmg, soil, colorbar_label="Soil thickness, m")
plt.savefig("output/soil_thickness.png")
plt.close()

# Ground cover
plt.figure()
imshow_grid(
rmg, gm, cmap=ground_cover_cmap, colorbar_label="Ground cover (1 = bare, 2 = grass)"
)
plt.savefig("output/grass_map.png")
plt.close()
7 changes: 7 additions & 0 deletions wolf-sheep-soil-creep/wolf-sheep-grass-mesa/clean.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
#!/usr/bin/env sh
set -e -u

. ../../tools/cleaning-tools.sh

clean_precice_logs .
clean_case_logs .
5 changes: 5 additions & 0 deletions wolf-sheep-soil-creep/wolf-sheep-grass-mesa/requirements.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
numpy >1, <2
matplotlib
mesa>=3
pyprecice~=3.0
networkx
22 changes: 22 additions & 0 deletions wolf-sheep-soil-creep/wolf-sheep-grass-mesa/run.sh
Original file line number Diff line number Diff line change
@@ -0,0 +1,22 @@
#!/usr/bin/env bash
set -e -u

. ../../tools/log.sh

exec > >(tee --append "$LOGFILE") 2>&1

if [ ! -v PRECICE_TUTORIALS_NO_VENV ]; then

if [ ! -d ".venv" ]; then
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt && pip freeze > pip-installed-packages.log
else
source .venv/bin/activate
fi

fi

python3 wolf_sheep_grass.py

close_log
Loading