The torrent of information that poured through last week’s Soil, Big Data and the Future of Agriculture conference in Canberra was like the topic of the day: absorbingly interesting, but what does it all mean?
It’s early days, but it seems that for agriculture, “big data” is going to flow through three channels.
One channel will route data collected on farms through professional service providers, who will add value to the data through analysis in return for a profit skimmed from the bits and bytes on their way through.
The other will see farmers collect their own data and, individually or in groups, use it to independently assess their own performance and requirements for profitability.
The third will be industry-owned data, a phenomenon already common in the livestock sectors, where genetic information is collected, analysed and distributed to the benefit of the industry as a whole.
Either way, the value is not in the data — of which there are already terabytes lying idle in farm computers — but what meaning is extracted from it.
For farmers who prefer to outsource technical expertise, big data is underpinning the next stage in the evolution of the “prescription agriculture” model outlined by Pradip Das of Climate Corporation, a subsidiary of Monsanto.
In Dr Das’s words, Climate Corp is in the business of delivering “actionable insights” from data.
The company’s core products are data prediction models that draw on a range of field and climate variables in order to guide the farmer’s delivery of inputs like nitrogen for optimum crop production.
Up to 10,000 modelling simulations are made per field to forecast optimum requirements for crop performance.
“We have been using this data to look at specific physiological processes — how the soil and farm performs in terms of productivity,” Dr Das said.
“All the variables you have on the farm, we try to see how each is related in terms of cause and effect, versus how we look after all the random variables.”
Climate Corp is hooking up with expertise to help make profitable sense of this data. It has acquired Solum, a highly automated soil laboratory, and is working with Novozymes, a Danish industrial biotechnology company, on soil biology.
Harnessing the data
Big data is democratic. It also presents the chance for individual farmers to gain “actionable insights” into their operation.
As young Wimmera farmer Jonathan Dyer told the conference, because of the information he is collecting, each pass he makes across the paddock is also a trial plot.
Tammin, Western Australia, farmer Brad Jones, is already harnessing data with powerful results.
Mr Jones farms 11,000 hectares, and crops 10,000. In 2008, after being battered by a year of highly variable weather, he threw out “tonnes per hectare” as a measure of productivity, and replaced it with “return on investment per unit of input”.
Understanding the new metric meant data, lots of it. Last year, he told the conference, he took 450 soil samples from across the farm, from 0-15 centimetres, and closely analysed them, along with past weather and yield data, to develop a risk profile for different soils.
Crop inputs are applied according to each soil’s risk profile. Soils with a high risk of not delivering a break-even crop get minimal input; soils with a higher likelihood of delivering a break-even result get higher inputs because they are more likely to make good on the higher financial risk.
Mr Jones’ minimal input rates look homeopathic compared to conventional rates. A high-risk soil might only get one unit of phosphorus, for example, trickled into the furrow as a liquid.
That might seem laughable to many, but Mr Jones’ results would not. He told the conference that over five years, on a discounted cashflow analysis, this risk-based approach to cropping has added $2.7 million to the operation’s bottom line.
Alex Ball from Meat and Livestock Australia spoke on the challenges of an industry approach to data aggregation.
Much of the information relevant to livestock performance on the farm is collected several steps past the farmgate, Dr Ball said, and that’s a problem.
New tools are providing valuable information about carcase performance, and can deliver practical data cheaply and quickly, but because of the cultural silos in the cattle and sheep industries, that information is not yet flowing back in ways that deliver value right through the supply chain.
Brands are locking up data in the belief that controlling it enables them to extract maximum value from it.
But in Dr Ball’s view, data gains the most value when all players allow it to flow through the supply chain, so that any party wanting to lift productivity has the full data picture to draw upon.
Big funds for soil data
CSIRO is getting $1.5 million to build an interactive system to digitally link information about farm soils, climate and other data in ways that aid farm decision-making, Agriculture Minister Barnaby Joyce announced at last week's conference.
Funded under the National Landcare Program, the project “takes the science out of the lab, converts it into a useful format and puts it in the hands of farmers,” Mr Joyce said.
“It will increase the digital connections between climate, soil and other information, and aims to give farmers information and improved spatial resolution to help them reduce their input costs and optimise their production.”