Skip to main content

Skip to navigation

Departments & Programs


Project Detail for Refining models of maize yield response to warming

The most certain aspect of global warming is that the planet will heat up, and at a pace and magnitude that is unprecedented in the history of human civilization. From the perspective of crops, warming can be very important as it affects a variety of processes that regulate growth, development, and reproduction. But how, exactly, do crop yields change as temperatures rise? How certain can we be in projecting impacts, and what management or technology interventions will help most?

An ability to model the response of crops to warming represents one of the biggest needs for producing credible projections of the impacts of climate change on food production – in many cases an even bigger need than refining errors in climate projections. A particularly hard problem is anticipating how crops will respond as they are exposed more frequently to historically unprecedented temperatures.

One useful way of making progress on this question is to experimentally expose crops to high temperatures. Another is to look at historical production records from individual countries and analyze how harvests rose or fell with temperature variations. Both are useful, but are limited in their potential to get a robust picture of how crops respond to warming. We have therefore begun to pursue an alternate approach of combining site-level measurements of crop yields from experiments conducted around the world.

For example, the International Maize and Wheat Improvement Center (CIMMYT) has had a decade long effort in Africa to develop drought-tolerant maize varieties (see One result of their work is a large dataset, a compilation of trial data from hundreds of trial sites (see figure below), that includes not only variety and yield numbers, but also phenological data such as anthesis, silking, and maturity dates. These data are of course useful for their intended purpose, to identify maize strains that do particularly well under drought, but they also have countless other applications. Although most trial sites don't collect daily weather data, a few are near enough to established weather stations to allow analyses of crop/climate interactions. These interactions have become the main focus of a Stanford-led project.

Using CIMMYT's data set, combined with climate data from NCDC weather stations, we can begin to develop a picture of maize yield responses to various aspects of climate, such as high radiation at flowering, or days above 40°C. We can compare all of these parameters, and countless others, to ultimate yield, and begin to get a sense of which factors affect yield and to what degree. We also plan to use these datasets to more rigorously test existing models of maize growth, as well as help to parameterize these models or add key missing processes. A similar project is also in the beginning stages for spring wheat.