Esperance grower Tyson Redding discusses yield-prediction tools.
PHOTO: James Tolmie
Advances in crop-yield predictive technology continue to accelerate and, if found to be accurate for local environments, offer opportunities for targeted fertiliser and herbicide management, more confidence in grain marketing and higher profits.
Being able to estimate in-crop what a particular paddock might yield paves the way for efficient input decision-making and crop enterprise planning as the season unfolds.
In the Esperance port zone last year, the South East Premium Wheat Growers Association (SEPWA) worked with four growers across the region to test five in-season yield-predictive systems.
The focus of this GRDC-funded and Esperance Regional Cropping Solutions Network (RCSN) group-supported research was on:
- practicalities of implementation;
- complexities of the tools;
- accuracy of crop responses compared to predictions; and
- whether growers liked using them.
SEPWA project officer Alice Butler says the systems evaluated were Yield Prophet® and ProductionWise (both based on the Agricultural Production Systems sIMulator (APSIM) model), iPaddockYield, Potential Yield Calculator (PYCAL) and nitrogen (N)-rich strips.
She says the project found the more complex tools Yield Prophet® and ProductionWise did not necessarily deliver more accurate yield predictions (mainly due to a lack of regionally specific soil-characterisation data).
These also generally took more time to set up, required more understanding and support from advisers and had greater room for error.
Owners: Kentyn and Jill Wehr, Tyson Redding
Location: Scaddan, Western Australia
Area: 2950 hectares
Enterprise: 100 per cent cropping
Annual average rainfall: 400 to 450 millimetres
Soil types: Sand over clay, shallow sand, duplex
2016 crop program: 1075ha wheat, 1100ha canola, 600ha barley, 175ha field peas
Ideal crop rotation: wheat, barley, canola
“The simpler tools of iPaddockYield, PYCAL and N-rich strips were easier to implement by growers and, in a lot of cases, out-performed the more complex tools in yield prediction accuracy,” Alice said.
“We did find that during the season, yield predictions varied with the different models used at the various sites.”
Alice says the project found ProductionWise was the closest to predicting yield for Baudin barley at the Condingup site.
She says PYCAL and iPaddockYield had the closest yield predictions for Mace wheat at the Scaddan and Mount Ney trial sites and iPaddock Yield had the closest prediction for Hindmarsh barley at the Neridup site.
“But when the growers were asked which tool they liked best and will continue to use, three growers said N-rich strips and one grower said iPaddock Yield and N-rich strips, because of the simplicity of these tools,” she says.
“The level of understanding of the person setting up each tool also directly impacted on how accurately the model replicated the real-world environment. The more simple tools appear to be suited to grower set-up and the more complex ones are probably better used with advisers – or adviser and grower collaboration – to be reliable and robust.”
Tyson Redding was one of the growers involved in the SEPWA project last year.
With his uncle and aunt, Kentyn and Jill Wehr, he is already using a wide range of precision agriculture tools to optimise productivity on their 2950-hectare Scaddan property.
The family incorporates data from soil testing, pH monitoring, biomass imagery, GPS guidance, yield at harvest, electromagnetic (EM), nitrogen and phosphorus maps, and soil moisture probes into variable-rate technology (VRT) and controlled-traffic farming (CTF) systems for targeted and accurate fertiliser and herbicide input application.
Tyson says the next step is to be able to accurately and confidently predict yield during the growing season, allowing more efficient management of inputs, risk, grain marketing and enterprise gross margins. To this end, the family has been using Yield Prophet® in recent years.
“All of these tools and technologies are vital because we are making many decisions in the paddock every day and we need to get these right to optimise profits,” Tyson says.
He estimates there are potential yield gains of 0.5 to 1 tonne per hectare to be made from good integration of in-season yield-predictive and precision agriculture tools.
Research has shown iPaddock Yield provided the closest prediction to actual yield at the Wehrs’ Scaddan property during the SEPWA project last year.
Tyson says the N-rich strips, which he set up using 200 litres/ha of Flexi-N in 100-metre-long and 2m-wide strips in several paddocks on a range of soil types, also proved to be a great visual reference for crop performance during the season.
He says SEPWA used a GreenSeeker® crop sensor to compare normalised difference vegetation indices (NDVI) between the N-rich strips and the ‘standard’ nitrogen rate on the surrounding paddocks. This data can be useful in calculating how much extra nitrogen is needed to reach target yield and grain protein.
“The N-rich strips gave us an instant ‘read’ of how the crop was going and whether it needed more fertiliser,” Tyson says.
“Coupled with the iPaddock Yield system, which is based on historical yield and rainfall data, we got good predictions of how the season was going and when and what fertiliser we needed to add to the crop.”
Tyson says in future, the best in-crop decision-making tools will be those that provide more precise information about crop development, potential yield and the season.
“Decisions are being made on thego and information needs to be quick and useful,” he says.
Table 1 In-season management tools used in the GRDC-funded SEPWA Grower Tools project, the associated required information to set these up and indicative cost.
|In-season management tool
|Information required for set-up
||Rainfall, soil moisture at sowing and soil evaporation value. PYCAL was used in the ProductionWise platform, as it used actual soil moisture at sowing.
||Free – but this project used the one that was part of the ProductionWise package (which costs $550 for up to 5000ha for one year).
||10 years of monthly rainfall and corresponding average farm crop yield.
||$179 – iPad/iPhone App.
||Runs off the APSIM model, which requires soil characterisation, soil-test results, rainfall data from nearest weather station, last year’s yield, and nitrogen, as well as NDVI from satellite imagery to produce its prediction.
||Starting package is $550 for up to 5000ha for one year – web-based platform.
|High rate of nitrogen strip that is compared to paddock standard using a GreenSeeker® to measure NDVI. The NDVI values for paddock and N-rich strip are then plugged into the N Calculator to give various N rates to produce various protein levels.
||Strip laid out for $50/paddock.
GreenSeeker® to read the strip $600.
N Calculator $50/paddock.
|Runs off the APSIM model, which requires soil characterisation, soil-test results, rainfall data, last year’s yield, crop rooting depth and N application.
||$180/paddock plus soil-coring costs – web-based platform.
0408 660 122
Alice Butler, SEPWA project officer
Crop yield predictive technology – Tyson Redding – video
0447 631 115
GRDC Hot Topics Water Use Efficiency and tools for growers
Ground Cover Supplement – 'Decision Support Tools'
Suspect barley samples needed
Time fungicides for powdery mildew control
GRDC Project Code