rLPJGUESS

An R package that wraps LPJ-GUESS, a dynamic global vegetation model, so that it can be driven from R rather than through its own configuration files.

The practical motivation is calibration and sensitivity analysis: those require running the model hundreds or thousands of times, which is only feasible in parallel. rLPJGUESS handles that on a workstation via SOCK clusters and on HPC systems via MPI, which is what makes it usable together with samplers such as BayesianTools.

Not on CRAN — install from the GitHub releases or build from source.