On-farm input trials show where the money is
GroundCover™ Issue: 94 | 17 Aug 2011
Grain grower Kym Curnow estimates variable rate technology (VRT) has saved his business more than $117,000 over three years by allowing him to target inputs to soil type on his Western Australian farm.
Mr Curnow, who farms 2700 hectares with wife Rosanne at Scaddan, says the phosphorus and nitrogen savings during the past three years have been handy. He does worry he might be mining the phosphorus bank although soil test results show levels are being maintained.
That concern aside, Kym Curnow says the savings have freed up funds to buy a real time kinematic (RTK) controller for variable rate (VR) spreading, rather than using a contractor.
Mr Curnow’s foray into VRT started in 2005 when he bought a new farm and wanted to quickly improve production with site-specific gypsum application. “We’d just finished applying a blanket rate of gypsum over everything and the paddocks still weren’t performing,” he says.
He took his agronomist’s advice and implemented an on-farm VR gypsum test to investigate the scope for lifting production on heavier soil types that tended to become waterlogged after rain.
In 2005, Mr Curnow engaged Precision Agronomics Australia (PAA) to undertake geophysical surveys of his soils and ground-truth the results with tests to confirm the physical and chemical characteristics of the soil.
PAA produced EM38 maps to determine which areas were likely to respond to gypsum (Figure 1 below). Orange areas indicated soils that might not respond to gypsum, while green and blue areas showed soils likely to respond.
In 2006, PAA produced a prescription map that Mr Curnow gave to a contractor to differentially apply gypsum.
As part of the work, Mr Curnow’s agronomist, Quenten Knight of PAA, established a long-term gypsum trial in one paddock to test the optimum rate of gypsum (0, 2.5, 5 or 10 tonnes a hectare) for each soil on the farm.
The trial also acts as a ‘barometer’ to indicate when gypsum reapplication is necessary. This is done by looking at how different crops planted on different zones perform over time with the differing rates of gypsum.
After the first year of gypsum application (2007), wheat yield results showed there was a strong response to gypsum (Figure 2), but the 10t/ha treatment was toxic as high rates of gypsum contain high levels of salt. The salt takes several years to leach away.
In 2009 (three years after the gypsum application), the canola yield increased by 100 kilograms a hectare on the least responsive soil (Figure 3).
On responsive soils productivity improvements were higher. On the nil gypsum treatment, yields were 1.4t/ha compared with 1.8t/ha where gypsum had been applied: a 400kg/ha yield response three years after application.
Quenten Knight says the results show quite clearly that adding even a small amount of gypsum can “increase the size of that bucket” and improve water infiltration, allowing 200 to 400kg/ha more grain to be produced, depending on the soil type and its responsiveness to gypsum.
“When you look at the data, there’s not a lot of difference in response between applying 2.5t/ha and 10t/ha of gypsum,” he says. “But monitoring the trial over time will allow Mr Curnow to decide if the extra expense of applying 5 to 10t/ha of gypsum is more cost-effective than regular applications of 2.5t/ha.”
Mr Knight says the yield in the 2.5t/ha test strip will eventually start to decrease, while the yields in the 5 and 10t/ha test strips should remain high for longer. “This (declining yield) will indicate when it is time to reapply gypsum to areas that have had a lower rate applied,” he says.
Although Mr Curnow’s test strips have shown little difference in yield response to the different rates of gypsum, Mr Knight says others have shown a significant response.
He points to grain grower Lloyd Burrell’s tests at Ravensthorpe as an example and says there are other cases where higher rates are required to give production an immediate “kick”.
In Mr Burrell’s trial, testing crop responses to 0, 2.5, 5 and 10t/ha showed there was no significant difference between the treatments one year after applying gypsum.
However, in year two (2009), field peas in the 10t treatment yielded 35 per cent more than the control.
Three years after applying gypsum, Mr Burrell’s trials showed the nil treatment yielded 1.4t/ha, the 2.5t treatment yielded 1.8t/ha, the 5t treatment yielded 2.1t/ha and the 10t treatment yielded 2.3t/ha.
The water use efficiency improvements had taken the crop from achieving only 55 per cent of its yield potential to 85 per cent of its potential.
To estimate the return on investment, PAA totalled the gross income from each treatment from 2008 to 2010 (barley 2008, field peas 2009 and wheat 2010) and then subtracted the initial gypsum treatment costs.
While the 10t/ha treatment produced the most grain after three years on Mr Burrell’s farm, the highest return on gypsum investment was achieved with the 5t/ha treatment.
According to Quenten Knight, these results show that “you have to view every trial in isolation and that’s why all growers need to implement their own on-farm test strips to assist decision-making”.
When it comes to the economics, Mr Curnow’s 2009 yield data showed that applying gypsum on responsive soils could add an extra $100 a hectare to his returns (Figure 4, page 16). At a cost of about $35/ha for gypsum application, Mr Knight estimates the net return at $65/ha.
Buoyed by these results, Mr Curnow was subsequently keen to trial VR phosphorus.
In 2008, he engaged PAA again to produce an EM38 map and establish production zones.
Mr Knight established a trial in one paddock to determine which soil types would respond to increased phosphorus rates.
After harvest 2008, the trial results showed none of his soils were responsive, which allowed Mr Curnow to adopt a phosphorus-replacement strategy and a more focused soil-testing program.
To estimate the phosphorus rates required for 2009, Mr Knight placed the 2008 yield map over the initial EM38 map of the paddock. Some parts of the paddock yielded 3.6t/ha.
For these areas Mr Knight calculated the phosphorus rate based on a removal rate of phosphorus by crop type. For this particular paddock, the phosphorus rate applied went from a low rate of 10 units to a high rate of 15 units.
In 2009, Mr Curnow used prescription application maps produced by PAA to apply VR phosphorus across 1900ha of his cereal program.
Before moving to VRT, his standard approach was to apply a blanket application of 14kg/ha of phosphorus across his entire farm. With VRT, some areas were given 5kg/ha while others received up to 17kg/ha (10kg/ha of phosphorus on average). This meant that instead of using 154t of fertiliser, Mr Curnow’s phosphate spend (at $5.60/kg) was reduced to 112t using VRT.
Across 1900 hectares, VRT saved $42,000, or $21/ha, due to the high cost of phosphate in 2009.
In 2010, cereals were sown across 2100ha. Had Mr Curnow used a blanket application of 14kg/ha of phosphorus, he would have used 215t of fertiliser. At 10kg/ha of phosphorus (on average) ranging from 5kg/ha up to 16kg/ha, only 158t of fertiliser (at a cost of $4.11/kg of phosphorus) was used. The fertiliser saving for 2010 was $33,000 or $15/ha.
In 2011, Mr Curnow applied 173t of fertiliser (10.5kg/ha of phosphorus) compared with 230t if he was still applying a blanket rate of 14kg/ha of phosphorus – a saving of 57t of fertiliser. Mr Knight says this saved Mr Curnow another $42,180 or $18.30/ha.
Adding the 2009, 2010 and 2011 savings together across Mr Curnow’s cereal program alone produced a $117,180 saving in phosphorus after three years.
In addition to phosphorus, Mr Curnow is using VRT to vary nitrogen application. A base rate of nitrogen is applied at sowing and VRT is then used four to six weeks afterwards to apply liquid urea (Flexi-N), at varying rates, based on soil test results, previous paddock history and the seasonal outlook. Some areas are given no extra nitrogen while other sites receive 50 to 70 litres a hectare.
Mr Curnow estimates the cost of adding a VR liquid nitrogen system to his Hardi tow-behind boomspray at $9000.
- By Nicole Baxter
GRDC Project Code DAW00193
Was this page helpful?