“With more and more climate change occurring, it’s very important for us to understand how we can predict the potential impacts of these climate scenarios before they’ve actually happened.”
The current report, posted on the Social Science Research Network, provided a forecast of the 2012 harvest. The study was completed in December 2012, before the release of official harvest numbers from the National Agricultural Statistics Service (NASS).
This effort is part of an international project called Agriculture Model Intercomparison and Improvement Project (AgMIP), which brings together researchers to share and improve crop and livestock models.
“The purpose of this report was to see if we could predict, given the emerging weather conditions, the crop production for the United States before we get through harvest and assemble the yields.”
Inputs for the model were numerous and included weather variables, soil profiles, planting and maturity dates, and data on irrigated and rainfed harvested areas.
To validate their model, Hatfield and his colleagues performed hindcasts for 1979 to 2011. On average, their estimations for those years deviated from the NASS observations by 8.3 percent. Their estimate for the 2012 corn yield was 7.507 t/ha compared with the observed NASS number of 7.74 t/ha.
The accuracy and consistency of the model are currently being evaluated in a comparison study with 23 other corn models. The models are being tested on a range of production systems in different areas of the world.
Hatfield hopes that as more information about their model is unveiled, improvements can be made to increase its robustness and utility.
One of the ways the team is trying to improve the model is to better understand how different environmental factors interact and affect crop yield.
The researchers are currently studying an approach they call CTW – the interaction between carbon dioxide, temperature, and water. Models have been constructed using observations of these parameters which may not be representative of the conditions under which crops will grow in the future.
“What we’ve been looking at is how that interaction is influencing the plant growth, development, and yield, and how the models currently handle those interactions,” explains Hatfield. “Once we understand that and do some model improvement, we hope that the models will be able to handle future scenarios we haven’t even considered yet.”
As the inputs to the model are better understood and researchers gain more confidence in its performance, the model can be used to look at a variety of climate scenarios.
If a weather event is forecasted to occur, and researchers know the impact of that event ahead of time, the model could help uncover changes to farming systems that might help alleviate some of the effects.
This possibility leads to various questions. Can the model be used as a tool to investigate possible adaptation strategies? Can the forecasts be moved far enough back that producers would have an idea of what’s coming before the season even starts?
“Eventually that’s where we’d like to get to,” says Hatfield. “We’d like to be able to evaluate the scenarios as we go forward, but then roll them back and use the model in a decision-making mode – as a support tool for producers to help them cope with weather changes.” FG
—From Crop Science Society of America news release
