Research and Economic Development, Office of


Date of this Version



University of Nebraska–Lincoln Office of Research and Economic Development (2010). Proceedings of the 2010 Water for Food Conference. Lincoln.


Copyright 2010, The Board of Regents of the University of Nebraska. All rights reserved.


A Global Assessment of Corn Water Use As Affected by Climate, Genetics and Scarcity

Marty Matlock described a high-resolution water assessment model he and colleagues are developing to determine how much water corn uses globally and to evaluate the balance between rainwater stored as soil moisture (green water) and water from surface water or groundwater sources (blue water). With a framework for assessing these characteristics, the model can analyze various scenarios, such as climate change and water demand by region.

“Our quest is to develop a modeling framework that has utility for decision-makers,”Matlock said.

To achieve high resolution, Matlock and his colleagues divided the globe into geospatial resolution cells of 5 minutes by 5 minutes, or about 10 kilometers by 10 kilometers. After inputting data for each cell, the researchers ran the model to determine yield. Comparing the results between the model’s predicted yield and observed data, the model was calibrated using high resolution input and yield data available for the U.S. heartland (Corn Belt). From potential yield data, researchers can determine water demand.

Matlock and his colleagues chose the CERESMaize simulation model embedded in the Decision Support System for Agrotechnology Transfer (DSSAT) because it uses daily rainfall inputs. It is therefore sensitive to critical threshold water scarcity, a more important element for kernel development than annual rainfall. Using the CERES model required collecting and entering daily data sources into each cell for each characteristic. Temperature and radiation data were acquired from the Climate Research Unit; precipitation data were acquired first from the Tropical Rainfall Measuring Mission and later from the National Climatic Data Center; and soil characteristics came from the ISRIC-World Inventory of Soil Emission Potentials soil dataset.

After running the model, Matlock’s team assessed its predicted values against global crop yield data obtained from Foley et al. published in Science magazine in 2005. The model did well in dryland regions, but predictions did not match observed yields in wetter regions. To calibrate the model using the highest geospatial resolution yield data, they focused on the U.S. heartland region, inputting high-resolution soil, temperature and rainfall data.

“We’re modeling one stalk of corn and extrapolating that to the world,”Matlock said. “If I really wanted this model to be right, I’d quit right now. All models are wrong; some models are useful. The question is, is there utility with this model? And I would argue that, yes, there’s strong utility because of its process-based development.”

To establish the model’s parameters,Matlock and his colleagues developed a set of parameters based on what other researchers use to model at the field or plot level. They first performed calibration runs on a 40-county region, then on a larger region spanning several hundred counties. For single cultivars, the model is sensitive to the four parameters that define the way a single corn stalk responds to precipitation and temperature. In the case of a single cultivar, the predicted versus observed graphs were not effective. However, modeling using nine cultivars and selecting the cultivar that best fit yield resulted in good calibration between predicted and observed yield. Mapping the results showed these four variables are associated with other important variables as well.

The next step will be evaluating the model’s ability to adequately predict water use. The model then can be used to analyze land use impacts on blue water resources; to determine a stress-related water footprint using regional stress factors; and to develop a series of water stress indices, including the impact on base flow under various scenarios, such as climate change, population change and industrial demand.

A lack of regional high-spatial and high-temporal data remains a problem, Matlock said. In addition, he continued, “We lack integrated models for the outcomes of concern: the ‘so what?’ part. We have to build that from scratch because life cycle assessment, risk-based models just don’t cut it for these sorts of social and economic impacts.”

Plant Research Innovations in the University: When Will They Apply to the Real World?

Despite tremendous innovations in plant research today, the challenges of integrating research into the real world leave many of those innovations stuck in the laboratory, Sally Mackenzie said. She described the approach taken by the Center for Plant Science Innovation at the University of Nebraska–Lincoln (UNL) to move research to the field.

Some of the research occurring in universities includes innovations to improve seed nutrient content, modify plant architecture for water use efficiency and alter properties to enhance shelf life.

Many innovations stem from the ability to sequence the genomes of major crop species, which is helping researchers understand the genes and mechanisms that one day may improve plant tolerance to drought and other beneficial characteristics.

These innovations and capabilities are already happening in the laboratory, Mackenzie said, adding that “the innovation is not what limits our ability to actually come up with some interesting solutions.”