MAgPIE - An Open Source land-use modeling framework

[4.14.1] - 2026-08-25

changed

• .Rprofile add r-universe repo, use envvars if present
• 14_yields pasture yield correction added to dynRegPastrTau_apr26 realization to match managementcalib_aug19
• CI test-code.yaml: use ubuntu-latest and checkout@v7
• Dockerfile now based on ubuntu 26.04, R 4.6, gams 54.1
• inputdata updated input data to rev4.132, repairing a FAO item-name mapping mismatch in the preprocessing (present since rev4.121): factor requirements in f38_fac_req_fao.csv for others, cottn_pro, groundnut, cassav_sp, puls_pro and oilpalm return to pre-rev4.121 levels; added f14_yld_past_switch.csv
• main.gms reslim reduced from 1000000 to 900 seconds, so a solver stuck inside one iteration is interrupted and retried instead of consuming the whole job allocation
• Makefile make reset-renv resets renv
• renv/activate.R updated to version 1.2.4
• scripts saveToResultsArchive saves to inbox folder if available

added

• .devcontainer/devcontainer.json A new configuration for development containers, which allow for reproducible, prepared development environments
• 14_yields/dynRegPastrTau_apr26 New realization, allows changing tau factor spillover to pastures by region and timestep.
• 22_land_conservation new options for IPLC land conservation
• 80_optimization/nlp_ipopt New realization, using IPOPT instead of CONOPT4 (and the fallback CONOPT3) as the NLP solver for the MAgPIE model.
• Dockerfile Re-added a Dockerfile, which can be used to build a local docker image as well as a GH codespace
• scenario_config_susmip.csv A set of sceanrios for the SusMIP excercise in the PRISMA project
• scripts added $RSCRIPT_SLURM_HOOK to run on slurm compute nodes via apptainer
• scripts/npi_ndc/start_npi_ndc.R Added ndcdelay afforestation policy (PRISMA T6.4 "Asymmetric Roll-back") applying ndc targets with milestone target-years delayed by country categorization. Plus, added AFFEXP which derives afforestation targets based on the share of potential forest land and speed of afforestation.
• scripts/start/extra/ipopt.R Start script for solving MAgPIE with IPOPT.

removed

• scripts/projects/fsec.R Removed FSEC_nitrogenPollution (grid-level nitrogen pollution downscaling) from the FSEC run output pipeline.

fixed

• 09_drivers, 14_yields, 15_food Minor stylistic improvements to GAMS code following capitalization consistency rules in gms::codeCheck.
• 21_trade Bugfix and refinement of bilateral trade realization to avoid infeasibiliteis in SSP4 and SSP5.
• 35_natveg/14_yields Fix youngsecdf wood production: derive its growing stock (im_growing_stock_ysf) from the uncalibrated secondary-forest carbon curve — the same curve its carbon density uses — instead of the FRA-2025-calibrated im_growing_stock(...,"secdforest"). Previously young secondary forest on other land yielded calibrated (high) wood volumes while booking uncalibrated (low) carbon, letting the optimiser evade land-CO2 caps/prices by relocating wood harvest onto youngsecdf. Result-changing for scenarios with land-CO2 pricing or AFOLU caps; explains the "other-land wood harvest" anomaly flagged under PR #876's Known limitations.
• 59_som Carry the soil carbon stock (pcm_carbon_stock(...,"soilc",...)) forward each timestep in postsolve (both cellpool_jan23 and static_jan19), so the soil term in vm_emissions_reg (q52_emis_co2_actual) is a per-timestep flux instead of a cumulative-since-initialisation change.
• scripts/output/projects FSDP_collect.R, FSDP_collect2.R and peatland.R selected yield variables under names magpie4 no longer emits, so their yield series were silently empty (FSDP since 2024-04-25, peatland since 2024-02-26). Updated to the Productivity|Yields tree; magpie4 dependency raised to >= 2.80.0 accordingly.