A Python ecosystem for reproducible discrete spatial simulation and land-change modelling. An independent alternative to INPE's TerraME/LuccME, which remain the reference: every package is checked against their outputs.
pip install dissmodel📖 Documentation · 🗺️ Repository map · 📦 PyPI · 📝 Under review at JOSS · DOI 10.5281/zenodo.23246106
From open data to a verified model, in code anyone can run.
| Package | What it does |
|---|---|
| dissmodel | Core: models, raster and vector backends, a declarative executor, and experiment records with checksums, specs and Git provenance. |
| disscube | Declarative pipelines that turn open geospatial data into model-ready grids. |
| disslucc | Land use and cover change: continuous and discrete allocation. |
| dissmodel-ca · dissmodel-sysdyn · dissmodel-abm | Extensions: cellular automata, system dynamics and agent-based models. |
| haloexec | Out-of-core execution of cellular automata on rasters larger than memory. |
| dissmodel-configs | Registry of external executors (TOML, reviewed by pull request). |
Case study: brmangue-dissmodel (sea-level rise and mangroves) and its QGIS interface.
Experimental: dissmodel-platform, a prototype execution platform (JupyterLab, FastAPI, workers, S3-compatible storage).
The reference outputs of TerraME and LuccME, and the benchmarks that compare each package against them, live in the LambdaGeo organization: terrame-goldens, luccme-goldens, disscube-benchmark and disslucc-benchmark. The open-lucc-workflow shows a complete traceable experiment from open data.
Developed by the LambdaGeo research group at the Federal University of Maranhão (UFMA). Led by Sérgio Souza Costa.
