The Model of Agricultural Production and its Impact on the Environment (MAgPIE) is a modular open source framework for modeling global land-systems, which is coupled to the grid-based dynamic vegetation model LPJmL, with a spatial resolution of 0.5°x0.5°. It takes regional economic conditions such as demand for agricultural commodities, technological development and production costs as well as spatially explicit data on potential crop yields, land and water constraints (from LPJmL) into account. Based on these, the model derives specific land use patterns, yields and total costs of agricultural production for each grid cell. The objective function of the land use model is to minimize total cost of production for a given amount of regional food and bioenergy demand. Regional food energy demand is defined for an exogenously given population in 10 food energy categories, based on regional diets. Future trends in food demand are derived from a cross-country regression analysis, based on future scenarios on GDP and population growth.
https://www.pik-potsdam.de/research/projects/activities/land-use-modelling/magpie
A framework description paper has been published in Geoscientific Model Development (GMD): https://doi.org/10.5194/gmd-12-1299-2019
The model documentation for version 4.4.0 can be found at https://rse.pik-potsdam.de/doc/magpie/4.4.0/
A most recent version of the documentation can also be extracted from the model source code via the R package goxygen (https://github.com/pik-piam/goxygen). To extract the documentation, install the package and run the main function (goxygen) in the main folder of the model. The resulting documentation can be found in the folder "doc".
Please find a set of tutorials here https://github.com/magpiemodel/tutorials. This guide will give you a brief technical introduction in how to install, run and use the model and how to analyse the model output.
Please pay attentions to the MAgPIE Coding Etiquette when you modify the code. The Coding Etiquette you find at the beginning of the documentation mentioned above. The Coding Etiquette explains also the naming conventions and other structural characteristics.
Copyright 2008-2021 Potsdam Institute for Climate Impact Research (PIK)
This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, version 3 of the License or later. You should have received a copy of the GNU Affero General Public License along with this program. See the LICENSE file in the root directory. If not, see https://www.gnu.org/licenses/agpl.txt
Under Section 7 of AGPL-3.0, you are granted additional permissions described in the MAgPIE License Exception, version 1.0 (see LICENSE file).
Following the principles of good scientific practice it is recommended to make the source code available in the events of model based publications or model-based consulting.
When using a modified version of MAgPIE which is not identical to versions in the official main repository at https://github.com/magpiemodel add a suffix to the name to allow distinguishing versions (format MAgPIE-suffix).
The model is quite resource heavy and works best on machines with high CPU clock and memory. Recommended is a machine with Windows, MacOS or Linux, with at least 16GB of memory and a Core i7 CPU or similar.
MAgPIE requires GAMS (https://www.gams.com/) including licenses for the solvers CONOPT and (optionally) CPLEX for its core calculations. As the model benefits significantly of recent improvements in GAMS and CONOPT4 it is recommended to work with the most recent versions of both. Please make sure that the GAMS installation path is added to the PATH variable of the system.
In addition R (https://www.r-project.org/) is required for pre- and postprocessing and run management (needs to be added to the PATH variable as well).
Some R packages are required to run MAgPIE. All are either distributed via the offical R CRAN or via a separate repository hosted at PIK (PIK-CRAN). Before proceeding PIK-CRAN should be added to the list of available repositories via:
options(repos = c(CRAN = "@CRAN@", pik = "https://rse.pik-potsdam.de/r/packages"))
Under Windows you need to install Rtools (https://cran.r-project.org/bin/windows/Rtools/) and to add it to the PATH variable. After that you can run the following lines of code:
pkgs <- c("gdxrrw",
"ggplot2",
"citation",
"curl",
"gdx",
"gms", # (>= 0.10)
"magclass",
"madrat",
"mip",
"lucode2",
"magpie4", # (>= 1.104)
"magpiesets",
"lusweave",
"luscale",
"goxygen",
"luplot")
install.packages(pkgs)
For post-processing model outputs Latex is required (https://www.latex-project.org/get/). To be seen by the model it also needs to added to the PATH variable of your system.
To use Docker, copy your gamslice.txt
into the MAgPIE main directory, and build the magpie image using the command
sudo docker build -t magpie .
Basic usage: Run the container (note the use of an absolute path) using
sudo docker run -v /an/absolute/path/to/a/folder/:/home/magpie/output -it magpie
Note: this will run MAgPIE with the default settings, if you want to change them choose the
Advanced usage: Run the container interactively using
sudo docker run -v /an/absolute/path/to/a/folder/:/home/magpie/output -it magpie bash
Model run settings are set in config/default.cfg (or another config file of
the same structure). New model scenarios can be created by adding a column to
config/scenario_config.csv
To run the model execute Rscript start.R (or source("start.R") from within
R) in the main folder of the model.
This will give you a list of available run scripts you can choose from. You can
also add your own run scripts by saving them in the folder scripts/start. To run
a single model run with settings as stated in default.cfg you can choose start
script "default". Make sure that the config file has been set correctly before
starting the model.
We are interested in working with you! Just contact us through GitHub (https://github.com/magpiemodel) or by mail ([email protected]) if you have found and/or fixed a bug, developed a new model feature, have ideas for further model development, suggestions for improvements or anything else. We are open to any kind of contribution. Our aim is to develop an open, transparent and meaningful model of the agricultural land use sector to get a better understanding of the underlying processes and possible futures. Join us doing so!
Model dependencies must be publicly available and should be Open Source. Development aim is to rather minimize than expand dependencies on non-free and/or non open source software. That means that besides currently existing dependencies on GAMS, the GDXRRW R package and the corresponding solvers there should be no additional dependencies of this kind and that these existing dependencies should be resolved in the future if possible.
If a new R package is added as dependency this package should fulfill the following requirements:
- The package is published under an Open Source license
- The package is distributed through CRAN or PIK-CRAN (the PIK-based, but publicly available package repository).
- The package source code is available through a public, version controlled repository such as GitHub
For other dependencies comparable measures should apply. When a dependency is added this dependency should be added to the HOW TO INSTALL section in the README file of the model framework (mentioning the depencendy and explaining how it can be installed). If not all requirements can be fulfilled by the new dependency this case should be discussed with the model maintainer ([email protected]) to find a good solution for it.
In order to allow other researchers to reproduce and use work done with MAgPIE one needs to make sure that all components necessary to perform a run can be shared. One of these components is the input data. As proprietary data usually does not allow its free distribution it should generally be avoided.
When adding a new data source, make sure that it can be freely shared with others. If this is not the case please consider using a different source or solution.
Data preparation should ideally be performed with the madrat data processing framework (https://github.com/pik-piam/madrat). This makes sure that the processing is reproducible and links properly to the already existing data processing for MAgPIE.
In case that these recommendations can not be followed we would be happy if you could discuss that issue with the MAgPIE development team ([email protected]).
By default the results of a model run are written to an individual results folder within the "output/" folder of the model. The two most important output files are the fulldata.gdx and the report.mif. The fulldata.gdx is the technical output of the GAMS optimization and contains all quantities that were used during the optimization in unchanged form. The mif-file is a csv file of a specific format and is synthetized from the fulldata.gdx by post-processing scripts. It can be read in any text editor or spreadsheet program and is well suited for a quick look at the results and for further analysis.
Please contact [email protected]
See file CITATION.cff or the How-to-Cite section in the model documentation for information how to cite the model.
See list of authors in CITATION.cff
See log on GitHub (https://github.com/magpiemodel)