Robotic farmers? Judging the potential or pitfalls of digital agriculture for Australian farms

A combination of low-cost sensors, cheap and increasingly powerful computing capacity and expanding mobile phone and internet access is resulting in the generation of rapidly growing volumes of digital data associated with farming operations. While holding the promise of productivity improvements by enabling management at the square metre rather than the paddock scale, there are concerns about data privacy and the potential use of digital information for government compliance purposes. 

The Australian Farm Institute asked two experts in ‘digital agriculture’ to provide their views on the potential implications of these developments for Australian farmers.

     

Brett Whelan
Associate Professor in Precision Agriculture,
Precision Agriculture Laboratory,
Faculty of Agriculture and Environment,
The University of Sydney (USYD)

    
Andrew Weidemann
Chairman,
Grain Producers Australia (GPA),
and farmer, Weidemann Pastoral Co

Q1.    There is a lot of hype about the benefits that might be available to farmers arising from the large volumes of electronic data that can now be generated by sensors and machinery. What is your assessment of the percentage that farmers might be able to improve yields or decrease costs through the use of electronic data and associated technologies?

Brett Whelan, USYD
These are exciting times for farm management and research. The data recorded by sensors placed in-field, on-vehicle, on-animal or on a remote platform will only become more important to the goal of optimising input use and productivity in agricultural systems. The information extracted will be used strategically to benefit multiple aspects of the systems from breeding rights through to product quality management. Given this, I see the use of electronic data being transformative across many industries, with the benefit large and component contributions difficult to partition.

However, from a broadacre cropping perspective, the size of the benefits at the farm level will depend on the farming enterprise and the current management. Efficiencies in machinery operation and logistics, better soil and environmental resource description, improved weather predictions and greater tactical use of resources in-season will allow for more certainty in achieving strategic productivity targets. This may not lead to a reduction of variable costs, but should mean that financial returns from the investment in variable costs (gross margin) will be increased. An improvement in gross margin of between 10% and 20% should be easily achievable for most operations.

Andrew Weidemann, GPA
In my opinion the electronic age has been a revolution for farmers in respect of education and the opportunities from major advancements, particularly in machinery. In 1996, we started yield mapping on our farm with the help of header mounted monitors. The data that we recorded on yield and moisture at that time measured the variability across each paddock. The yield maps raised questions that were at that time hard to quantify but today with the advancements in products, such as APSIM yield profit soil characterisation, and as farmers generally we have been educated by these tools to better understand the correlation between rainfall and soil type.

Technology advancements by machinery manufacturers has, in my opinion, now driven the next generation of advancements through auto steer and this has changed the view of farming today to a more sophisticated level as we continually learn how to better use these tools to improve productivity, and now also better able to record and vary various applications of products across paddocks.

App technology has now driven an easier hand-held recording process as well which can be used – as in our case where we are utilising this technology in providing and supporting the information which sits behind our grain marketing and storage program.

If you look at the overall result of these tools we certainly have learnt how to better apply things such as nitrogen which applied with full knowledge of available soil moisture can vastly improve the overall productivity of the entire cereal or canola program. In most cases we have worked on a rule of an application of $1 aiming to return $4 for the investment made. Or conversely if we don’t apply the nitrogen, based on the overall prediction of the season based on the scientific outlook for rain, then we save money. In my opinion, most farmers employ these types of assessments on a regular basis and, as confidence and improvements in weather information improves, we can expect better productivity and ultimately profitability growth from these specific technologies.

Q2.    To what extent do you think farmers should be wary about the potential for their farm data to be distributed to unauthorised persons, or even accessed by government regulators?

Brett Whelan, USYD
There are a number of levels of farm data that will be valuable, but the value will only be realised when the data is used to extract information for the business/industry/society. At the moment many farmers are generating/collecting data without extracting the value. That is a waste.

I expect that when wireless broadband reaches useful upload speeds across the farming regions, then the options for automatically gathering and storing vehicle operational, farm resource, environmental and field operation/production data will mature. This development will come with a strong commitment by the agriculture community to open data protocols and data privacy controls. Combined, these advances will allow manufacturers, farmers, input suppliers and advisors to share data for a variety of different purposes which may range from directing future site-specific field operations back on the same farm, to aggregating information from a region on yield response for different crops/varieties to local conditions and treatments. Both these examples would extract value from the data and benefit the data provider. I am optimistic that farmers who are prepared to engage in these developments will be rewarded by accessing new information and analyses to improve decision-making at a range of different operational scales within their businesses.

Andrew Weidemann, GPA
Producers should own their data and as time progresses the value of the farmer generated data will become more important to a range of sectors but they should be prepared to pay for it. I am aware there is a precedent in other countries which allows for data to be collated and then shared in a cooperative arrangement which actually pays a return to the producers. Data cooperatives are a thing for the future and will be something that I think may well be developed in the coming years, as we come to terms with the data age. We need to be careful though in how this information is collected and the format of the data will be vitally important. Governments have a way of over-analysing farmers and in my opinion this has rarely seen a return to farmers more broadly, whereas the private sector tends to drive outcomes from research data collected.

Q3.   Thinking ahead five to 10 years, what do you think will be the major developments that will emerge in the ‘digital agriculture’ space? Do you think data and robots will substantially replace farmers, just as has occurred to workers employed in manufacturing industries?

Brett Whelan, USYD
Increasing volumes of digital data along with remote storage and fast data transfer will finally help achieve the real practical goal of precision agriculture, which is to increase the number of (correct) decisions per hectare/animal/machine/season made in the business of farm management.

Initially targeted at the cropping industries, the real-time telemetry of machinery operational status will become commonplace to allow remote expert adjustment and optimisation. The data will be stored for analysis to improve the use of fuel and labour over time.

Software systems will come of age that merge data streams from diverse sources and scales with adaptable crop production and environmental models that will feed information into key management decisions. Such a system will require data generation and capture through a range of platforms for such attributes as yield, crop vigour, disease, pests, soil, environment and economics/markets. The data will need to be automatically stored in/transferred between online ‘data dormitories’ in formats that are openly accessible.

Industry will then provide analysis to drive the required prescriptive agriculture decisions. These will be provided as alternative options based on assessing the probable outcomes of models based on understanding the causal relationships involving the measured attributes. This will be a substantial improvement in trustworthy decision support and more inline with producing an ‘augmented advisor’ than replacement with robots.

I expect there will be some specific automation of vehicle-based operations such as targeted weeding, ground preparation/sowing and harvesting in some crops in the next five years. However, I believe the timeline for any substantial shift from Australian farmers to intelligent robotic machinery managing whole-farm operations is greater than 10 years.

Andrew Weidemann, GPA
I don’t think we will see farmers generally replaced by robots but I can see the day when we will see driverless tractors and spatial imagery along with drone and or satellite pictures of the fields being easier to access, and most likely from the phone.

The thing the industry will have to work through though is: will this technology lead to more profitability? or will it just take more important time to manage?

Back to November 2015 index