Predicting Agricultural Impacts of Large-scale Drought

Joshua Elliott, a fellow at the Computation Institute at the University of Chicago, and several colleagues are studying the effects of the drought of 2012 on corn yields in the US. In a recent paper, the team makes the case that the severe nature of the drought has demonstrated a significant need for better analytical tools.

elliottThe researchers are undertaking a model-based assessment of the 2012 US growing season using the parallel System for Integrating Impact Models and Sectors (pSIMS). The system is a high performance computing framework that fuses independent climate and agriculture models at large scales, producing 5-arcminute spatial resolution (about 10 km) simulations. The pSIMS framework is written in Swift, an open source parallel scripting language developed at the Computation Institute and prototyped through the University of Chicago Computing Cooperative (UC3) campus grid, the Open Science Grid (OSG), and   the Extreme Science and Engineering Discovery Environment (XSEDE).
“The goal is to create a probabilistic framework to estimate and forecast climate change impacts on a variety of temporal and spatial scales—ranging from local field scale all the way up to continental and global scales—and in time ranging from seasonal scales to multi-decadal climate change scales,” explains Elliott. “The framework runs several different models at once using the same set of assumptions and scenarios.”

Agricultural impacts of large-scale drought

The median deviation of simulated 2012 county-level yields from linear trend as a percentage of county-specific trend yields from 1979 to 2011.

The team is validating their model against county-level data released by the US Department of Agriculture. By fine-tuning the underlying mechanics, they hope to determine how soon accurate predictions can be made about an upcoming harvest. Extending that lead time will ultimately give farmers and decision makers more opportunities to adjust for adverse weather – and researchers more opportunities to simulate approaches to improving crop resilience.
“The Open Science Grid provides us with easy access to compute resources an order of magnitude larger than we otherwise could get from campus clusters,” Elliott says. Ultimately, the OSG gives Elliott compute cycles without the overhead that sometimes comes with large scale resources—such as difficult custom architectures, required proposals (and subsequent wait times), and complicated access procedures. “Grad students without degrees in computer science—or in our case students from geoscience—are able to get accounts and get started on the OSG in relatively short order,” says Elliott.

Elliott also made a point of praising the OSG support system for researchers. “For example, in some cases, code may need to be optimized for distributed computing. The community is full of great people who can help solve all sorts of problems. Researchers should definitely take advantage of the OSG.”

Elliott notes that they have only scratched the surface of possible applications for high performance computing and big data in simulating crop yields and climate change impacts. He thinks this type of work is essential for helping stakeholders at all scales—from local farmers to agri-businesses to governments and even international aid agencies—respond to the challenges of drought, food insecurity, and climate change over the coming years and decades.

“Agricultural risk management and adaptation planning at seasonal to multi-decadal time scales has become more important than ever in the context of global socio-economic and environmental change,” Elliott says. “Countries around the globe will be facing difficult choices over the coming decades, and we’re hoping these tools and analyses can help. Resources like the Open Science Grid are essential if we want to solve big data challenges like this.”

“A world four degrees warmer than it is now is not a world that we’ve ever seen before,” Elliott says. “Studying years like 2012 in detail can potentially be very useful for helping us understand whether our models can hope to accurately capture the future.”

~ Rob Mitchum (communications manager, Computation Institute) and Greg Moore

For more on Elliott’s research and its insights into the future of agriculture, read Rob Mitchum’s April 17 iSGTW newsletter article: