AgMIP GGCMI

The idea behind the Global Gridded Crop Model Intercomparison project (GGCMI), part of the Agricultural Model Intercomparison Project (AgMIP), is to coordinate and improve the use of crop models in order to better understand not only how they work, but also what drives crop performance in different regions and how that might change in the future.

More than 30 research groups around the world are part of GGCMI. Some of these use pure crop models that simulate crop-specific development processes at high levels of detail; others are more general vegetation and/or land system models, of which crops are only a part. This diversity of structures, along with the high number of participating models, lends GGCMI the ability to assess a wide range of model-related uncertainty.

The current phase of GGCMI, phase 3, serves as the organizational nexus for the ISIMIP agricultural sector’s contribution to ISIMIP3. These coordinated standardized model experiments will allow the coordinated evaluation of model performance for the recent past, and exploration of climate change impacts over the 21st century. In addition to being (along with Jonas Jägermeyr) one of the co-coordinators of the ISIMIP agriculture sector, I am also performing the phase 3 runs for LPJ-GUESS.

In the recently-completed second phase of GGCMI, models simulated crop yields in a factorial experiment with axes of climate anomalies (amount of temperature and precipitation deviation from the recent past), CO₂ concentrations, and management strategies (fertilizer and irrigation use). Emulators built using outputs from these runs can be used to approximate yield estimates under changing climate and management in a much more lightweight manner than actually running the models themselves. At the moment, I am working on using these emulators to generate yields for input into the PLUM land-use model (like how LPJ-GUESS feeds PLUM in LandSyMM) to assess the magnitude of model structural uncertainty on land-use projections relative to uncertainty in climate and socioeconomic development scenarios.