Credits: CC moertl @Flickr.com

Agriculture is one of the sector most challenged by climate change and will probably remain as the sector most affected in the future. Consistent global-scale evaluation of crop productivity is essential for assessing the likely impacts of climate change and identifying system vulnerabilities and potential adaptation strategies.
Over the last several years, research groups around the world have developed several process-based Global Gridded Crop Models (GGCMs) to simulate crop productivity and assess climate change impacts at relatively high spatial resolution over global domain. The different and often contrasting methodologies and assumptions invariably lead to a wide range of impact assessment.
Crop models are currently the best tools for assessing its impacts. Yet the changes in crop yields projected by different models in response to the same meteorological forcing can differ substantially. 

In a work recently submitted to Environment Research Letters, entitled “Simulated vs. Empirical Weather Responsiveness of Crop Yield: U. S. Evidence and Implications for the Agricultural Impacts of Climate Change”, CMCC Foundation researchers Malcolm Mistry and Enrica De Cian from ECIP Division, in collaboration with Ian Sue Wing (Boston University, USA) provided a first glimpse into the origins and implications of this divergence through an inter-method comparison, both among the GGCMs, and between GGCMs and historical observations.

The paper focuses in particular on U.S. agriculture while examining yields of rainfed maize, wheat, and soybeans simulated by six GGCMs from the Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP Fast Track), comparing 1972-2004 hindcast yields over the coterminous United States (U.S.) against U.S. Dept. of Agriculture (USDA) time series for greater than 1,000 counties (a county is an administrative unit within a state, with the majority of counties used in the study roughly equivalent to the size of the Italian province of Verona, viz. approximately 3000 sq.km).

They estimated reduced-form econometric models of crop yield responses to temperature and precipitation exposures for both GGCMs and observations.
GGCMs have difficulty in reproducing the observed distribution of percentage yield anomalies, and exhibit aggregate responses that show yields to be more weather-sensitive than in the observational record, not only in response to adverse exposures to extreme high temperature or low precipitation, but over the entire range of heat and moisture conditions.
In general, GGCMs’ yield responses are more sensitive to temperature and precipitation exposures. Translated into future climate change impacts on yields under current management practices, this implies that the GGCMs tend to be more pessimistic while projecting negative impacts on future crop yields.
This disparity is largely attributable to heterogeneity in GGCMs’ responses, as opposed to uncertainty in historical weather forcings, and ultimately drives the divergence in the projected impacts of climate on future crop yields.

For more information, visit the page of the event.