UNL agronomists spearhead standards for using crop data

jbrehm2, May 11, 2015 | View original publication

UNL agronomists spearhead standards for using crop data

An international team led by UNL researchers has published the first systematic guidelines for using data to more reliably simulate crop yields.

The guidelines could apply to any crop model used for estimating gaps between potential vs. actual yields, assessing a country’s ability to feed its population, and predicting the impacts of changes in climate or land use.

By accounting for various levels of data availability and quality, the team’s guidelines attempt to reflect agricultural realities while still addressing a long-standing need for practical and consistent standards, UNL agronomist Patricio Grassini said.

“The guidelines we are establishing are flexible, because we understand that, in some areas of the world, it is difficult to find high-quality data,” said Grassini, assistant professor of agronomy and horticulture. “Therefore, we propose different options for data sources depending on the [nature] of data in a given country.”

To illustrate the versatility of its guidelines, the team used them in determining the gap between actual and potential corn yields across regions in Nebraska, Argentina and Kenya – sites that vary in terms of environmental conditions and data availability.

Grassini said his team is aiming to ensure that researchers use data responsibly, particularly given the prevalence and ever-growing importance of the models that rely on it.

“(Virtually) all assessments that have to do with land use, food security, climate change and associated topics are based on computer models,” Grassini said. “However, it is common to find studies … in which researchers did not pay much attention to the input data used as the basis for these assessments, which are then subject to a high degree of error. We are trying to rectify this and really set the standards for reliable crop yield simulations.”

The standards address several classes of variables critical to projecting agricultural production and related outcomes. These range from the contextual factors of weather, soil types, planting rotations and crop varieties to methodological issues surrounding the spatial scales and time frames of available data.

For instance, the protocols specify how many years of weather data are needed to ensure a reasonable estimate of average crop yields and variability – while noting that the answer depends partly on the amount of rainfall or irrigation a given area receives.

“This is also a very effective tool to identify gaps in data, because if you run these protocols for a given country, you can really tell which are the major data bottlenecks,” Grassini said. “So you can look at a country and say that what’s missing is weather data or soil or something else. It really helps to prioritize efforts and funding to collect the data that are missing and improve their quality.”

Grassini said the team’s guidelines rely on observable data whenever possible. The team advocates a “bottom-up” approach that begins with local simulations and builds toward regional and global levels according to weather-based parameters.

“We’ve found that if you go the other way around – that is, if you start by using these huge global databases and try to simulate something in every single grid of the world – you really lose precision,” he said.

Grassini and his colleagues also endorse the transparency of publicly available data. They’ve published their guidelines in two open-access online articles through the journal Field Crops Research.

“The uncomfortable truth is that, in many parts of the world, data have a price,” said Grassini. “That’s a very important constraint for agricultural research and development. Relevant data should remain in the public sector for the sake of the common good.”

This philosophy motivated the creation of the Global Yield Gap and Water Productivity Atlas, which uses consistent protocols to analyze actual yields against estimates of food production capacity in numerous countries around the world.

“Ultimately, reliable crop yield simulations – at the local, regional and global levels – are needed to better inform policymaking and prioritize agronomic research and development,” Grassini said. “That’s the ultimate goal.”

The lessons learned while building the atlas helped inform the data guidelines recently published by Grassini and his colleagues, who include Justin Van Wart, postdoctoral researcher in agronomy and horticulture; Haishun Yang, associate professor of agronomy and horticulture; and Kenneth Cassman, a Robert B. Daugherty Professor of agronomy.

The UNL researchers developed their protocols with colleagues from the Netherlands-based Wageningen University; the International Crops Research Institute for the Semi-Arid Tropics in Nairobi, Kenya; and the Africa Rice Center in Cotonou, Benin.

Funding for the team’s research and atlas came from the Bill and Melinda Gates Foundation, the University of Nebraska’s Robert B. Daugherty Water for Food Institute, Wageningen University and the U.S. Agency for International Development.