Complexity key when weighing up yield prediction tools

Author: Natalie Lee | Date: 14 Jun 2016

A project to assess crop yield prediction tools has found that their varying levels of complexity are important to consider when weighing up which is the best to use in a particular situation.

“The level of understanding of the person setting up each tool will directly impact on how accurately the model replicates the real environment,” South East Premium Wheat Growers Association (SEPWA) project officer Alice Butler said.

“We found that the simple tools can be easily used by growers, but the more complex ones are consultant tools or require at least consultant and grower collaboration for a reliable and robust set-up.”

A paddock of Mace wheat at Scaddan.

A paddock of Mace wheat at Scaddan which was one of the sites where yield prediction tools were compared by SEPWA in a project funded by the GRDC’s RCSN initiative. Photo by SEPWA

The project, led by SEPWA, was initiated by the Grains Research and Development Corporation's (GRDC) Esperance Regional Cropping Solutions Network (RCSN) group and was conducted with assistance from growers, Farm & General, the Department of Agriculture and Food WA and Precision Agronomics Australia.  

Estimating what a paddock will yield helps growers to manage herbicide and fertiliser inputs during the season, forward sell grain and accurately insure crops. This is important for managing gross margins and ensuring profitability.

Ms Butler said that Yield Prophet® was the best known yield estimation tool, but other options were becoming available to help growers forecast yields and manage inputs.

“To compare the options, SEPWA conducted a study in 2015 whereby growers at Mount Ney, Condingup, Neridup and Scaddan were set up with Yield Prophet®, iPaddockYield, ProductionWise, Potential Yield CALculator (PYCAL) and N-rich Strips,” she said.

“The study found that different models were better at predicting yields for different sites.

“It also highlighted the complexity of yield prediction and that this could cause misunderstanding and lower grower confidence in some of the models.”

Ms Butler said iPaddockYield, N-rich strips and PYCAL were easy to set up and understand while ProductionWise and Yield Prophet® required assistance and a higher level of time investment.

“Yield Prophet® requires soil characterisation, soil test results, rainfall data, the previous year’s yields, crop rooting depth and nitrogen, making it a much more sophisticated model,” she said.

“ProductionWise had the same level of complexity as Yield Prophet®, with both models running off the Agricultural Production Systems sIMulator (APSIM) model, but the additional use of Normalised Difference Vegetation Index (NDVI) information from satellite imagery was valuable.

“Additional features include the ability to create farm maps and run gross margin analysis.”

SEPWA also conducted a State-wide survey of 82 growers to assess their knowledge and use of the tools and how likely they would be to adjust seasonal inputs based on the information generated from them.

“Results confirmed that Yield Prophet® is still the best known and most commonly tried model, with 30 per cent of growers surveyed having used it,” Ms Butler said.

“Overall, 60 per cent of growers indicated they were likely or highly likely to adjust seasonal inputs based on information from a yield or seasonal prediction tool.”

Results from the GRDC RCSN project ‘Grower tools – Yield predictive devices for making in-season management decisions’ are available via this link and this link.

Contact Details

For Interviews

Alice Butler, SEPWA
08 9083 1165, 0404 277 337


Natalie Lee, Cox Inall Communications
08 9864 2034, 0427 189 827

GRDC Project Code SEP00014

Region West