IF somebody said you could have a tailor-made app for your farm, to predict crop yields, it no doubt would pique your interest.
It would be interesting to see how you would react if, on further inquiry, you discovered it had more accuracy than online production model Yield Prophet and costs only $70.
According to Esperance farmer Mic Fels, who developed the app, called iPaddockYield, most farmers have reacted with a mix of interest and scepticism.
“It sounds too good to be true and generally that’s the basis of judgement by society,” Mic said. “I’ve been using the basics of iPaddockYield for the past 15 years and it has helped generate some solid profits, mainly through informed management decisions that mitigate over-spending.
“We’ve managed to grow our business from 2000ha in the 1990s to now running a 10,500ha cropping program and owning 8000ha of land.
“A big chunk of that has been financed by the extra profits from using iPaddockYield.”
He developed it into an app because he wanted to share the benefits of this knowledge with other farmers after winning the 2012 GRDC Grain Inventors Award.
Last year iPaddock also won the Best Invention and Best New Release Awards at the Newdegate Machinery Field Days.
The reaction from farmers has always been positive, but uptake hasn’t exactly run Mic off his feet.
“There has been wide uptake and there’s a spread of farmers around Australia which will hopefully help promote the concept in different rainfall zones,” Mic said. “And now we’re employing it on our new property at Three Springs.
“I am, and always have been my own best case study”
“I’m a pretty busy farmer and farming pays the bills, so I am relying on word of mouth rather than a big sales campaign to sell the apps. If people like them, they will sell themselves”
According to Mic, iPaddockYield differs from the French-Schultz and Yield Prophet systems by using the farmer’s historical records to calculate their own yield-to-rainfall algorithm.
What you’re trying to do is convert rainfall into grain.
“Basically you feed the app by entering your historical rainfall and crop yield records, which it uses to generate a tailored model of your yield versus rainfall,” Mic said. “The more accurate the information the better and ideally it should be10 years of records.
“The emphasis is on accurate records and in our case at Esperance (Wittenoom Hills), we achieve about 95 per cent forecast accuracy for the July yield forecast.
“But even in less reliable areas we are consistently seeing correlations better than 70pc.
“You can do your own sums comparing it against other predictive methods.”
Joining farmer ranks, this sceptical writer naturally wanted the “proof of the pudding”.
I chose last year for two reasons, both of which were based on the difficulty of forecasting and input management.
At Esperance, for example, the season started as a big challenge and finished okay whereas at Three Springs, crops jumped out of the ground but got hit by a hot August.
“Over the past five years, iPaddock has accurately picked my crop yield by July,” Mic said.
“It is well proven that the most profitable use of nitrogen gives wheat protein around 11pc, which is actually not hard to achieve if you know the actual yield you are fertilising for.
“Our protein graph over the past five years is consistent around our target of 11pc, so just that bit of information tells me we’re maximising profits with our N fertiliser.
“You can’t control rainfall but you can control your N inputs.
“There’s a tendency to under-fertilise in wet years and over-fertilise in drier years and either way you easily burn hundreds of thousands of dollars by not matching the rate to the realistic yield potential.
“The problem is they are ‘hidden losses’ because the loss occurs before we even harvest the crop.
“If someone actually gave us all that money, and then we had to give it back we would be howling from the rooftops.”
Cropping 6000ha in Esperance, last year saw a dry January-April period with only 19mm of rainfall recorded.
Yet iPaddockYield was forecasting a yield of 2.2t/ha against a long term farm average of 3.1t/ha.
“You don’t make a profit growing a 2.2t crop with the inputs for a 3t crop,” Mic said. “Based on the 2.2t/ha figure, I made changes, cutting $20,000 out of my fertiliser budget for seeding.
“My mindset then was that I had to think like an eastern Wheatbelt farmer to make a profit. The cheque book went into lock down.
“I cut a lot of Flexi-N use and backed off on phosphorous rates on the heavy country.
“I was aggressively targeting a profit and I started dry sowing canola on April 10 and didn’t stop the program.
“We got our first decent rain in mid-May while sowing barley and it turned out to be a life-saver with 54mm being recorded, which bumped the iPaddock App prediction to 2.5t/ha.
“Because it was still below average we didn’t change anything but the relatively dry 25mm in June saw me reduce urea spreading to 27kg/ha of N rather than our long term average of 47kg/ha and we just used a single shot of the cheaper fungicides mixed with our broadie spray.”
At that stage the forecast was still sitting on 2.4t/ha.
“Our fortunes changed though with 80mm in July, and the forecast jumped up to 2.9t/ha,” Mic said. “But we were only fertilised for 2.4t/ha so we spread another 20kg/ha of N, which brought us up to our long term average of 47kg of N required to grow a 3t/ha crop.
“At that stage we also increased our grain sales target into the cash and swaps markets, based on the 2.9t/ha target.
“We also committed on buying a third header, because two headers would handle a 2.5t crop but we needed three if we were going to take off 15,000t in our cosy Esperance climate.
“We got 20mm in August which reduced the forecast back to 2.6t/ha, but by now we had done everything needed for 2.6t so we just shut the gate.
“A fair bit of Flexi-N was still going out around the district in August and September.
“The 50mm in September saw the forecast back at 2.8t/ha so we consolidated our sales targets leading into harvest and ended up with an average of 3.1t/ha, which was 300kg/ha more than I would have achieved in the past on the same rainfall.”
Assessing the season, Mic said the key profit drivers were getting the crop in early, adding more N in July to match the increased yield potential, but resisting the urge to put late N on in spring even though the crops looked good.
“Protein is the best indicator of N efficiency, and we averaged 11pc again last year while the district average was over 12 pc,” he said.
“Using DAFWA’s Select Your Nitrogen model, our saving on N fertiliser alone last year was $120,000 compared to if we had used district practice.
“That is real profit which we are using to pay off debt, buy more land and a few toys and family holidays as well.
“We didn’t do anything special, we just had access to some well timed strategic information from the iPaddockYield app.
“At $70 I’m always surprised that people wouldn’t just give it a go, even if they were completely sceptical about it.
“I think farmers make a lot of assumptions about how iPaddockYield works, and all the reasons why it wouldn’t work for them, but once they see the simplicity of it and how well it works in so many different areas, they are usually a bit shocked.”
Mic said he always aimed to beat the App’s prediction because if he did, it showed improvement in management decisions.
“It is the only presently available way to compare what you are doing now, with what you have done in the past,” he said.
Mic has found this priceless when making big changes, such as changing to discs and narrow row spacings for his seeding gear.
“There’s always an element of gut feeling because you never stick to the same procedures every year,” he said. “If the crop had only yielded 2.5t/ha that would have told me we got it wrong with the changes we made last year, but thankfully it all worked out very well.
“Our crop yields were average compared to others in the district but our costs were low, we didn’t blow out protein and the bank is happy with our gross margins.”
Mic only cropped 1000ha at their new Three Springs property last year and wasn’t able to use the app there, because he didn’t have any historical information.
“After being used to knowing my yield potential all through the year, I felt like I was fighting with both hands tied behind my back,” he said. “I don’t know how people who don’t use it sleep at night.”
In March this year he was able to access seven years of data from a neighbour, which he ran through the app and found that it picked last year at Three Springs to a tee.
“I was virtually flying blind at Three Springs with no experience in the district but the data we got proved to be pretty close to the mark,” he said.
So here’s what happened at Three Springs in 2014.
February: (First rains, 14mm) Predicted target 1.2t/ha. March (26mm) 1.7t/ha. April (29mm) 2t/ha. May (58mm) 1.9t/ha. June (30mm) 1.8t/ha. July (49mm) 1.4t/ha, following a dry spell. August (51mm) 1.5t/ha. September (54mm) 1.6t/ha.
“We finished with an average 1.7t/ha so it turns out iPaddockYield was spot on,” Mic said. “Interestingly in May, there was a lot of excitement about the good break and the district was talking 3t/ha, which definitely influenced input decisions.
“Yet it turns out iPaddockYield was only forecasting 1.9t/ha by the end of May.
“Like most in the area we over-fertilised, ending up with 12.2pc protein in our wheat.
“Hopefully we can get it right this year, now that we have the app running at both farms.
“The app shows its ability to throw out a reality check where you don’t get emotional and it stops you throwing money at the crop because you tend to go on the high side in such scenarios as last year.
“I’d rather feed the crop based on more rationale and accurate analysis.
“And as the year went on people blamed the dry spell for not getting 3t/ha when in reality it was never a 3t season.”
The new version of iPaddockYield which is about to be released also includes a neat daily rainfall recorder.
You simply enter the rainfall amount for the day and it calculates monthly total, year-to-date total complete with a graph and deciles by month and year.
This year, iPaddockYield is currently forecasting a bumper 3.6t/ha at Mic’s Esperance properties, and 1.6t/ha at Three Springs.
“There’s a lot more confidence in forward selling your crop with the information you gain from the app,” he said.
p The iPaddockYield app can be downloaded onto iPhone and iPad from the App Store or the website ipaddock.com.au and you can follow Mic on twitter at @ipaddockapps for the latest news and releases.