Home / aus / New research accurately predicts Australian wheat yield months before harvest

New research accurately predicts Australian wheat yield months before harvest



wheat

Credit: CC0 Public Domain

Topping the list of Australia's most important crops, wheat is grown in more than half of the country's arable lands and is an important export item. With so much driving on wheat, an accurate yield forecast is needed to predict regional and global food security and commodity markets. A new study published in Agricultural and forest meteorology shows how machine learning methods can accurately predict wheat yield for the land two months before the crop matures.


"We have tested various machine learning approaches and integrated large-scale climate and satellite data to arrive at a reliable and accurate prediction of wheat production for the whole of Australia," said Kaiyu Guan, assistant professor at the Department of Natural Resources and Environmental Sciences at the University of Illinois, professor of Blue Waters at the National Center for Supercomputing Applications and principal investigator. "The incredible team of international staff who contributed to this study has significantly improved our ability to predict wheat yield for Australia."

People have tried to predict the harvest yield almost as long as there have been crops. With increasing computing power and access to different data sources, predictions continue to improve. Scientists have developed fairly accurate crop yield estimates in recent years using climate data, satellite data, or both, but Guan says it was unclear whether one dataset was more useful than another.

"In this study we use a comprehensive analysis to identify the predictive power of climate and satellite data. We wanted to know what each contributes," he says. "We discovered that climate data alone is pretty good, but satellite data provides extra information and brings efficiency performance to a higher level."

Using both climate and satellite data sets, the researchers were able to predict wheat yield two months before the end of the growing season with an accuracy of around 75 percent.

"In particular, we discovered that satellite data can gradually capture crop yield variability, which also reflects accumulated climate information. Climate information that cannot be captured with satellite data serves as a unique contribution to wheat yield prediction throughout the growing season," says Yaping Cai, PhD student and lead author of the research.

Co-author David Lobell of Stanford University adds: "We also compared the predictive power of a traditional statistical method with three device learning algorithms, and computer learning algorithms outperformed the traditional method anyway." Lobell started the project during a sabbatical 2015 in Australia.

The researchers say the results can be used to improve predictions about Australia's wheat harvest in the future, with possible ripple effects on the Australian and regional economy. Moreover, they are optimistic that the method itself can be translated into other crops in other parts of the world.

The article "Integration of satellite and climate data to predict wheat yield in Australia using machine learning approaches" is published in Agricultural and forest meteorology.


Climate extremes explain 18% to 43% of the global variations in crop yield


More information:
Cai et al., Integration of satellite and climate data to predict wheat yield in Australia using machine learning approaches, Agricultural and forest meteorology (2019). DOI: 10.1016 / j.agrformet.2019.03.010

Supplied by
University of Illinois at Urbana-Champaign

Quote:
New research accurately predicts Australian wheat yield months before harvest (2019, May 13)
recovered May 13, 2019
from https://phys.org/news/2019-05-accurately-australian-wheat-yield-months.html

This document is copyrighted. Apart from any fair trade for private investigation or investigation, no
portion may be reproduced without written permission. The content is provided for information purposes only.


Source link