AGROMINDS

AGROMINDS generates annually resolved agro-meteorological indicators for crop-specific growing seasons from daily climate data and crop calendars.

Run a processing job

Install the dependencies, then create a local configuration and adjust its machine-specific paths:

cp config.example.toml config.toml

Select the climate model, crop calendar, and scenario on the command line:

pip install -r requirements.txt
python generate_indicators.py --gcm GFDL-ESM4 --crop mai --ssp historical

config.toml contains paths only and is ignored by Git. The required arguments are:

• --gcm: GFDL-ESM4, IPSL-CM6A-LR, MPI-ESM1-2-HR, MRI-ESM2-0, UKESM1-0-LL, or GSWP3
• --crop: mai, ri1, ri2, soy, or swh
• --ssp: historical, ssp126, ssp245, ssp370, ssp585, or obsclim

Repository layout

• generate_indicators.py: user-facing processing command.
• agrominds/: reusable crop, biophysical, climate-I/O, feature, and pipeline code.
• aux_data/: small static inputs used by the current workflow.
• notebooks/: notebooks and figures used for data exploration and paper visualizations.
• tests/: automated tests and small test fixtures.
• outputs/: local processing outputs; excluded from version control.

The climate forcing data are external inputs and are not included in this repository.

Input data

aux_data/ is a local, gitignored directory for third-party and project-derived inputs. It is intentionally not distributed with the code. See input-data documentation for the required files, authoritative sources, and provenance status.

Reproducing manuscript figures

The notebooks in notebooks/ reproduce the manuscript figures below. Install the dependencies in requirements.txt, configure the output-data location in config.toml, and run each notebook from the notebooks/ directory.

Notebook	Manuscript figure
features_violins.ipynb	Figure 2
features_lines.ipynb	Figure 3
features_maps.ipynb	Figure 4

These notebooks require the generated AGROMINDS Parquet outputs; they do not recreate the full dataset from daily climate forcing.