Predictions of lower crop yields call for yield gap studies

Wheat, corn, and rice account for about 85% of global cereal production, and these crops contribute a majority of calories eaten around the world. Their production is vital if we are to feed a continually growing population. Predicting future yields of these crops will give researchers and farmers a better idea of what’s ahead. A new study suggests that previous predictions have been overly optimistic and that, in fact, yields in some areas have already reached a plateau.

headshot of Patricio Grassini

In the study recently published in Nature Communications, researchers from the University of Nebraska-Lincoln analyzed the trends of crop yields over the past four decades using comprehensive and powerful statistics. Patricio Grassini, lead author of the study, and his colleagues, Kenneth Cassman and Kent Eskridge, then compared their analysis against predictions published in other studies.

Many previous studies assumed that yields will increase at an exponential rate over time to 2030 or 2050. The new study showed that exponential rates of yield increase are not supported by historical trends or what is known about crop growth. Instead, exponential predictions are likely to greatly overestimate future increases in yields.

“For example, the goal for average U.S. corn yield of 300 or more bushels per acre by 2030 would require an annual yield gain that is four times greater than the rate of increase from 1965 to 2011,” says Grassini, an assistant professor in the Department of Agronomy and Horticulture. “In our study, we found no evidence of exponential yield gain at all for the cases we analyzed – 36 countries and regions.”

The model that Grassini and his colleagues found to match historical trends best was a linear model. And while some other studies used a linear model, they assumed that linear rate of increase would continue forever. But in Grassini’s study, there were interruptions and changes in the linear trend over time.

“We found that a linear model can reproduce well the shape of the trends during the last 40 years,” he explains. “But in many cases, we saw a break in that linear trend. We found that the rate of yield gain slowed down or that yields reached a plateau.”

The new study shows that there is often a second phase in the linear trend, a point at which the slope of the line decreases. The shallower slope means not that crop yields themselves are decreasing but that the rate at which yields are increasing is slowing down.

Even more striking were the cases in which crop yields appeared to plateau during the second phase of the linear trend. Yields had actually remained flat over the last 10 or 20 years in 31% of the cases Grassini and his colleagues studied.

“For some of the most productive areas around the world, yields have not increased for one or two decades,” says Grassini. “We can no longer assume a steady rate of gains in the future. It is more realistic to project yields based on linear models with breakpoints and plateaus.”

What is causing these crop yield plateaus that Grassini and his colleagues observe? There are two potential explanations. In some areas, the plateaus are occurring at very high yield levels. In these cases, such as for rice in China and California or wheat in northwest Europe, it is likely that current yields are close to the yield potential. Yield potential is the yield of a crop when growth is only limited by sun, temperature, and carbon dioxide supply, and also precipitation in rainfed systems. This is the biophysical limit of the crop. In areas where average farm yield has reached about 80% of the yield potential, it is very difficult for farmers to increase yield any further.

farmer standing in her corn field in Tanzania

But some of the plateaus are occurring at very low yields, for example, for corn in sub-Saharan Africa. In these areas, current crop yields are far below the potential yield. Yield plateaus in these cases are not due to a biophysical limit, but to external obstacles. And it is in these areas that Grassini sees the most potential to increase yields.

“In the places where their plateaus have very low yield levels, there is, in theory, more opportunity to increase yields,” he explains. “There is a large exploitable gap between the actual and potential level of productivity. However, these places also have the poorest access to technology, infrastructure, markets, and capital required for agricultural development.”

To better understand the gaps between actual and maximum yield, Grassini and colleagues from the University of Nebraska-Lincoln and Wageningen University are working on a project called the Global Yield Gap Atlas. The project aims to estimate yield potential for cropping systems around the globe and compare potential yields to actual yields in order to estimate yield gaps. Mapping areas where the yield gap is large can help prioritize research and inform agricultural policies.

“It is crucial to know the size of the yield gap because otherwise you don’t know exactly what the room for improvement in productivity is in a given region,” says Grassini. “You don’t know how much the productivity can be increased locally, regionally, or globally. Having an estimate of what would be the maximum productivity on a given piece of land is crucial for having an accurate estimate of food security in the future.”

Grassini hopes that the new study and others like it will bring attention to yield gaps and direct research funding toward measuring and taking advantage of those gaps. Fully understanding yield gaps throughout the world is a rewarding but immense project, he says, one that requires boots on the ground.

“It’s a huge challenge. The yield gap is widespread, but causes for the gap are different in each area. We need a local approach, agronomists working locally to identify yield gaps and yield constraints. Then we can find possible solutions in an economical and sustainable way.”

To learn more about the Yield Gap Atlas, go to