To predict how agriculture will be affected by future climate change, scientists often rely on a single crop model – a computer simulation of how a specific crop’s yield responds to temperature changes. By combining 30 such models into a single study, and comparing each model against data from existing experimental wheat fields around the world, a team of researchers including Stanford professor David Lobell have developed a more powerful and accurate way to predict future wheat yields.

In a new analysis published in Nature Climate Change, the team’s results support previous work suggesting that wheat yields around the world are sensitive to rising temperatures. Using the new method of analysis, the team estimates an average six percent future yield loss for every one degree Celsius rise in global mean temperature.

“Combining 30 models gives us a much greater ability to predict future impacts and understand past impacts,” said Lobell. “This is a clear step forward.”

Lobell is professor of environmental earth system science in the School of Earth Science at Stanford and the deputy director of the Center on Food Security and the Environment. He is a senior fellow at the Stanford Woods Institute and at the Freeman Spogli Institute for International Studies.

The estimated six percent yield loss for every degree increase is equivalent to about a quarter of the current volume of wheat traded globally in 2013. Yields at some sites, notably those in Mexico, Brazil, India and Sudan, show simulated wheat yield losses of more than 20 percent - in Sudan’s case, more than 50 percent - under a scenario in which global mean temperature rises by two degrees Celsius.

With higher temperatures also comes an increase in the variability of wheat yields, both by location and between years. More fluctuation in wheat yields could mean greater global price volatility for the staple crop.

Approximately 70 percent of the wheat produced today is grown either on irrigated plots or in rainy regions. The research team accounted for this factor by focusing its simulations on multiple regional-specific varieties of wheat that are commonly grown under these conditions.

The new paper includes several suggestions for avoiding some of the predicted yield losses. For example, some varieties of wheat are more heat tolerant than others, and farmers in the places hardest hit by rising temperatures could switch varieties to capitalize on this heat resistance. The effects of rising temperatures could also be managed, in part, by adjusting sowing and harvesting dates, or changing the way fertilizers are applied to crops.