A computer model which predicts crop outcomes is going national after
successful grower-group trials in Victoria, reports Helen Olsen
[Photo by Brad Collis: Improved yield security is coming from a computer model that helps manage seasonal variations.]
A computer model that simulates crop growth according to seasonal conditions and helps to predict likely outcomes to assist growers make risk-management decisions is going national after small-scale trials in Victoria during 2004.
The online risk management and decision support program, Yield Prophet, was developed by the Birchip Cropping Group (BCG), and is a simplified version of the Agricultural Production Systems Simulator (APSIM) program, developed by the Agricultural Production Systems Research Unit (APSRU).
APSRU is a joint research unit comprising the Queensland Departments of Primary Industries and Fisheries and Natural Resources and Mines, CSIRO Sustainable Ecosystems and CSIRO Land and Water, and the University of Queensland.
Yield Prophet is an online computer model which simulates crop growth based on presowing conditions, the management inputs provided and subsequent weather conditions. It can be used at any stage of the season to calculate a range of predicted outcomes based on site-specific historical weather data from the Bureau of Meteorology on rainfall, temperature and solar radiation, collected over the past 100 years.
The BCG won the tender to commercialise the APSIM technology and has done so under the name Yield Prophet. It will supply the service to interested growers for a fee.
The model calculates the likely yield outcome, when supplied with relevant information including:
Yield Prophet last year was only validated for wheat, but in 2005 will also be able to simulate barley and sorghum crops and the water and nitrogen balance of fallows.
To assess the reliability of Yield Prophet, BCG tested the model using 20 years of historical yield data from two farms, and ran the model using climate data from the same 20 years. It was able to predict yields with about a 70 per cent success rate.
BCG consultant Dr Harm van Rees says that the model is designed to understand cropping in Australia, which is a very complex issue, and the accuracy of the data put in by growers is crucial to how well it works.
"Growers need to know their soils pretty well, including the available nitrate, soil water and any subsoil limitations," says Dr van Rees. "In areas such as northern NSW and the Wimmera/Mallee region of Victoria, there is excellent information available about the local soils.
"However, in areas such as central NSW and around Lake Bolac in Victoria there is not much information, or it is not clear where the information is held."
Dr van Rees says that BCG is happy to make Yield Prophet available in areas where soil properties are well known, but is more cautious about regions where soil information is not known or not easily obtainable.
Yield Prophet is a project under the Managing Climate Variability Program, in which the GRDC is a partner. It is trying to find if information for these regions exists, and if so who has the information.
Dr van Rees says not all growers who have used Yield Prophet are sure it was of benefit. One reason why some growers decide not to continue is that it takes a good understanding of the implications of management decisions on the farm, and using probabilities of likely outcomes can be confusing.
The coordinator of the Yield Prophet program, James Hunt, reinforces how sensitive Yield Prophet is to the measurement and description of soil water properties. It needs accurate Crop Lower Limit (CLL) and Drained Upper Limit (DUL) values for paddock soils.
"These measurements are time-consuming, and the values will differ between each soil type that a grower has on his or her property, and as such they are a limit to the model," says Mr Hunt. "At the same time this reflects reality, in that no model can work for Australian conditions without taking into account the importance of soil water in our cropping systems.
"And the CLL and DUL values only need to be measured once for each soil type."
Mr Hunt says that during 2004, 47 growers in northern NSW, the Riverina region, the SA Wimmera/Mallee and the high rainfall areas of WA trialled Yield Prophet.
While Yield Prophet had a deliberately low profile during the 2004 testing phase, Mr Hunt is now keen for it to be available to as many growers as possible. "Yield Prophet is about as user-friendly as internet banking, and therefore I expect at least 90 per cent of growers will be able to use it without any problems."
For the 2004 season, Yield Prophet provided three reports to subscribers: an agronomic report, a climate risk report and a nitrogen comparison report.
The agronomic report provides a yield probability outcome for the season, along with data on simulated crop development and soil water and nitrogen status up to the time the report is generated.
The climate report provides one yield outcome probability based on the past 100 years of climate data, and another based on years in that period in which the Southern Oscillation Index (SOI) was behaving as it is at the time of the report.
The nitrogen comparison report is a tool for growers to determine the likely benefits and risks associated with nitrogen fertiliser inputs. It can generate yield and protein outcome probabilities for three nitrogen application regimes, using climate data from the past 100 years and years in which the SOI was similar to the time of the report.
The additional reports that are planned to be part of Yield Prophet from 2005 are: a financial return on nitrogen cost report, a variety by sowing date report and a special service for irrigation farmers on irrigation scheduling and nitrogen requirements.
GRDC Research Code LWR25
For more information: Dr Harm van Rees, 03 5439 3089; James Hunt, 03 9350 4742; www.bcg.org.au/cb_pages/yield_prophet.php