Big Data Report Web

research report
April 2016
The Implications of Digital

Agriculture and Big Data for
Australian Agriculture
2016 Australian Farm Institute

ISBN 978-1-921808-38-8 (Print and Web)
Australias Independent
Farm Policy Research Institute
The Australian Farm Institute
The Australian Farm Institute is an agricultural policy research organisation that has been established to develop and
promote public policies that maximise the opportunity for Australian farmers to operate their businesses in a profitable
and sustainable manner.
To do this, the Institute carries out or contracts leading academics and consultants to conduct research into farm policy issues
that the Institutes Research Advisory Committee has identified as being of high strategic importance for Australian farmers.
The Institute has a commitment to ensuring research findings are the conclusion of high quality, rigorous and objective
analysis. The Australian Farm Institute promotes the outcomes of the research to policy-makers and the wider community.
The Australian Farm Institute Limited is incorporated as a company limited by guarantee and commenced operations on
23 March 2004. The Institute is governed by a Board of Directors who determine the strategic direction for the Institute.
The Institute utilises funding voluntarily contributed by individuals and corporations to perform its activities. Initial seed
funding has been contributed by the NSW Farmers Association.
Vision
Farm policies that maximise the opportunity for Australian farmers to operate their businesses in a profitable and
sustainable manner.
Objective
To enhance the economic and social wellbeing of farmers and the agricultural sector in Australia by conducting highly
credible public policy research, and promoting the outcomes to policy-makers and the wider community.
Membership/Subscriptions
The Australian Farm Institute has a wide range of Membership and Publication Subscription alternatives that meet the
requirements of individuals, corporate organisations, libraries and small business:
Corporate Membership
Individual Membership
Student Membership
Journal Subscription
Institutional Subscription
For further information please contact the Institute on 61 2 9690 1388 or via email info@farminstitute.org.au or
alternatively visit www.farminstitute.org.au/memberships
Publications
The Institute publishes a variety of publications, including a quarterly Farm Policy Journal, research reports, papers,
newsletters and media articles.
To obtain hard copies of Institute publications, contact the Institute on (02) 9690 1388,
email info@farminstitute.org.au or visit www.farminstitute.org.au/publications
The Implications of Digital Agriculture
and Big Data for Australian Agriculture
April 2016
A report prepared by the Australian Farm Institute

with funding from Dairy Australia Limited, Grain Growers Limited
and the Cotton Research and Development Corporation
2016 Australian Farm Institute

Australian Farm Institute
April 2016
This publication is protected by copyright laws. Apart from any use permitted under the Copyright Act
1968, no part may be reproduced by any process without the written permission of the publisher:
Australian Farm Institute Limited
Suite 73, 61 Marlborough Street

Surry Hills NSW 2010
AUSTRALIA
ABN 29 107 483 661
T: 61 2 9690 1388
F: 61 2 9699 7270
E: info@farminstitute.org.au
W: www.farminstitute.org.au
All rights reserved
The views and opinions expressed in this publication are those of the authors and do not necessarily reflect
those of the Board of the Australian Farm Institute or the Institutes members or corporate sponsors.
Disclaimer
The material in this Report is provided for information only. At the time of publication, information
provided is considered to be true and correct. Changes in circumstances after publication may impact on
the accuracy of this information. To the maximum extent permitted by law, the Australian Farm Institute
disclaims all liability for any loss, damage, expense and/costs incurred by any person arising from the use
of information contained in this Report.
Publication Data
Keogh, M & Henry, M (2016), The Implications of Digital Agriculture and Big Data for Australian
Agriculture, Research Report, Australian Farm Institute, Sydney, Australia.
Design and Production: Australian Farm Institute
Printing: J.A. Wales Printers, Alexandria

iii
Foreword
Agriculture has experienced two major revolutions over the past century. The first was the mechanical
revolution that occurred in the years between the first and second world wars, during which time
horsepower was replaced by mechanical power, with dramatic improvements in productivity. The second
was the scientific revolution (often referred to as the green revolution) which occurred over the period from
the late 1960s to the late 1990s, and involved the application of well-developed science to the sector, again
resulting in significant productivity increases.
It is probably reasonable to argue that agriculture is now undergoing its third major revolution, the digital
agriculture revolution. This revolution has been made possible as a result of the dramatic reduction that
has occurred in the cost of digital and computer technology, and the adaption of this technology in farm
implements and farm monitoring applications.
The potential impact and implications of the digital agricultural revolution are still quite unclear, although
rapidly developing. Digital technology and data applications are emerging to support farm management
decisions, maintain and report on biosecurity issues, support quality assurance and credence systems, map
and analyse land use and crop performance, monitor and manage water, and to track markets and transact
sales and purchases.
The flood of digital information that can now be generated as a part of normal farm operations is leading
to questions about how the information should or can be stored, managed and utilised in ways that enhance
farm productivity and profitability. The early indications are that the digital agriculture revolution will
create very important opportunities for productivity gains, but will require a combination of scientific
knowledge, computing applications and human resource development in order for those gains to be
realised.
The research reported here involved a detailed analysis of global developments in digital agriculture, and
consideration of what will be needed in Australia to ensure that the undoubted benefits that are available
will be quickly realised.
Mick Keogh
Executive Director
Australian Farm Institute
April 2016
The Implications of Digital Agriculture and Big Data for Australian Agriculture | April 2016
iv
Executive summary
The development of farming machinery and decisions enables farmers to economically change
digital technology that is able to generate objective from paddock and herd average management, to
information about the status of soil, water, crops, square metre and individual animal management,
pasture and animals is quickly changing the way in with reported subsequent increases in farm
which farm businesses can be managed in Australia. productivity. While the extent of productivity
The emergence of digital agriculture, and the gains vary across different agricultural production
potential this creates for the application of big data systems robust analyses report gains of the order of
analytics in agriculture, signals the initial stage of 10% to 15% in cropping systems, with about half
a fundamental change away from the skill-based the gains coming from input efficiencies, and the
farm management systems that have prevailed until other half from increases in output. The deployment
present times towards a more industrialised model of digital agriculture systems in livestock industries
of agriculture where decisions are based to a greater is generally less advanced, and therefore estimates
degree on objective data. of possible productivity benefits while significant
are yet to be properly validated.
The earliest stages of this change occurred in the
row and broadacre cropping sectors in the early Although bringing the promise of important
1990s with the development of global positioning productivity gains, digital agriculture also brings
system (GPS) guidance systems, which were then with it questions about the ownership of, and
augmented with autosteer technology and grain use to which, digital information obtained from
harvester yield monitors. Subsequent developments a farm can be put. The service providers that are
included seeder and fertiliser applicators with the marketing digital agriculture systems and data
capacity to vary application rates within a field. storage platforms for farmers are all commercial
More recently, software applications and cloud data service providers and, as is the case more generally
storage facilities have enabled the resulting data to in relation to digital information, there is a lot of
be captured, stored and manipulated, and then used uncertainty about the rules which govern how this
in decision-support tools to guide farm management new digital environment should operate.
decisions.
Generally speaking, digital information generated
Digital agriculture applications have also emerged by machinery and technology used on a farm is
in the livestock and horticulture sectors, including, owned by the farmer, although the Conditions
for example, electronic livestock identification of Use agreements that are routinely signed by
systems, genomics, automated milking systems, computer software users when they first register
automated livestock weighing platforms, telemetric or use a particular application typically curtail the
irrigation and water management systems, remote users data ownership rights, and create exceptions
sensing technologies, and instruments for the which enable the software provider to use the
automated collection of weather and climatic data in different ways, and often to make that data
information. available to third parties.
The use of digital agriculture systems to implement Developing an appropriate regulatory environment
more intensive and data-driven farm management which protects a farmers ownership rights over
April 2016 | The Implications of Digital Agriculture and Big Data for Australian Agriculture
v
farm data is a complex task. Farm machinery occurred in the US. This is because developments
manufacturers typically reserve the ownership rights in the US have been based on the public availability
to machinery performance data, and accept some of detailed soil maps; public access to high-density
degree of control by farmers over the ownership and weather data; the presence of a comprehensive
use to which digital farm production data can be put. mobile telecommunications network throughout key
cropping regions; and the presence of large-scale
Concerns about the misuse of digital agriculture commercial agribusiness service providers which
data by service providers has led to the development have been prepared to invest in the development of
of Codes of Practice or the strengthening of privacy these systems.
regulations in the United States (US) and New
Zealand, with a focus on limiting the uses of data to Australian agriculture will benefit from the
those agreed to by farmers (who are considered the technology spill-in arising from developments in
owners of the data). Complexities arise in situations the US, with imported farm machinery now routinely
where farm data is transferred to third parties (such equipped with the digital control and monitoring
as agronomists or livestock advisors) and also in the systems that have been developed in the US.
case of remote sensing data obtained via satellite or
drone, over which farmers have no control or rights. Whether the computer software systems and
Despite these uncertainties, workable arrangements data storage applications that become available
appear to be emerging that are not overly restrictive to Australian farmers will be modified to suit
for service providers, and which give sufficient Australian cropping systems remains to be seen,
confidence to farmers. and there has only been limited development of off
the shelf systems for the livestock industries.
Digital agriculture and related big data applications
are more advanced in the cropping sector of the There are a range of initiatives that can be adopted
US than is the case in Australia, and are generally by the agriculture sector in Australia to facilitate
more developed in the cropping sectors than in the more rapid development of digital agriculture
the livestock sectors. After an initial phase during systems, and these are detailed in the following
which service providers attempted to develop recommendations arising from this research.
closed proprietary systems to encourage greater
product loyalty, what has emerged in the US is a Recommendation 1:
commitment to open access data arrangements, Australian agricultural industries, Australian
whereby data obtained from different types of agricultural research agencies and relevant IT,
machinery is able to be used on multiple different telecommunications and software organisations
software platforms, and readily transferred between should collaborate in the establishment of the
these. As a consequence, competition has emerged Australian Digital Agriculture Forum, with the
in the provision of data storage and management broad objective of advancing the development
platforms, and a competitive software market has and adoption of digital agricultural applications
also developed. Different systems cans be used and systems in Australia.
by farmers, irrespective of the particular brand
of machinery that they are using, or the storage Recommendation 2:
platform on which their data is held. In addition,
these open access data arrangements permit farmers It is recommended that Australian agricultural
to transmit their data from one service provider to industries, agricultural technology providers
another, with very little loss of functionality. and digital agriculture platforms and software
system providers should adopt as a key principle
Digital agriculture holds the promise of significant that the farmers who own the land or livestock
productivity benefits for Australian farmers, from which digital agricultural production
although the systems and platforms may not data is obtained retain ownership rights over
develop in Australia to the same extent that has that data. This includes the ability to determine
vi
the uses to which that data can be put, and Australian governments should increase
the persons or organisations which can obtain available funding to augment access to mobile
access to that data. Where contractors and telephone and data networks in rural and
sharefarmers are employed, it is recommended regional Australia, and actively investigate the
that a standard contract be developed that potential for public/private investment models as
defines data access protocols for each party. a means of further enhancing data coverage.
Recommendation 3: Recommendation 7:
It is recommended that Australian agricultural Australian governments and rural research and
industries, agricultural technology providers development corporations should collaboratively
and digital agriculture platforms and software develop a strategy to make the detailed data
system providers should commit to open access and relevant metadata associated with publicly
data protocols, modelled on the standards funded research available in accordance with
adopted by the Open Agriculture Data Alliance an open access data protocol, and work to
established in the US. standardise the availability of other relevant
information about research trials.
Recommendation 4:
Recommendation 8:
It is recommended that Australian agricultural
industries, agricultural technology providers and Australian publicly funded agricultural research
digital agriculture platforms and software system organisations have a fundamental role in the
providers should support the appointment of a generation of knowledge to underpin digital
Farm Data Ombudsman to oversee data privacy agriculture applications, models and algorithms,
standards, to establish data use categories, and but should not be involved in the development
to audit compliance by providers with industry of commercial software programs or digital
standards for data privacy. agriculture platforms that will be used by farm
service organisations or farmers.
Recommendation 5:
Recommendation 9:
ustralian governments should increase available
A
funding for soil mapping and weather recording Private-sector digital agriculture applications
stations, and actively investigate the potential for and platforms have the potential to dramatically
public/private investment models and private- change the way in which farmers access
sector collaboration as a mean of improving the production and other information relevant to
soil and climate datasets that are an essential farm management decisions. These systems
foundation of digital agricultural systems. should become the principal information supply
chain for farmers in the future, and public-sector
Recommendation 6: agricultural research agencies will need to develop
ack of access to mobile telephone and data
L new strategies that recognise these systems as the
coverage can be a major impediment to the principal extension pathways of the future.
adoption of digital agricultural systems.
vii
Table of contents
Foreword iii
Executive summary iv
Table of contents vii
List of boxes, figures and tables viii
Glossary of terms ix
Acronyms x
1. Introduction 1
2. Big data in the economy 6
3. The development of digital agriculture 8
3.1 Cropping 10
3.2 Extensive livestock 15
3.3 Dairy 20
3.4 Horticulture 21
3.5 Productivity implications 24
3.6 Compliance and supply chain implications 25
4. The market for farm data storage and analysis 27
5. Agricultural data and privacy 34
5.1 Legal framework relevant to agricultural data 34
5.2 International digital agriculture standards 37
5.3 Case studies of different data flow models 40
6. Technological limitations of digital agriculture 48
6.1 Data errors 48
6.2 Digital agriculture and internet access 51
7. Discussion and recommendations 53
8. References 64
viii
List of boxes, figures and tables
Figure 1: Percentage of regional cropping areas on which variable rate

applications were used in 2011. 9
Figure 2: Harvester fuel use map. 11
Figure 3: Climate Corporation Nitrogen Advisor. 13
Box 1: Farmlink TrueHarvest. 14
Figure 4: TrueHarvest Gap Map. 14
Figure 5: Diagrammatic representation of how various data from technologies
may contribute to identified on-farm needs. 15
Table 1: Estimated benefits from walk-over-weighing and auto-drafting systems. 16
Figure 6: Remote livestock management system. 17
Figure 7: Wireless NLIS animal identification and tracking system to assist with
herd management and breeding using Taggle ear tags. 17
Figure 8: GPS positions of cattle as recorded by UNEtracker collars
over 14consecutive days. 18
Figure 9: Sense-T Pasture Predictor tool. 19
Table 2: Comparsion of AMS performance under Australian and European
management systems. 21
Figure 10: Autonomous robot conducting surveillance in an almond farm at Mildura. 22
Figure 11: Photogrammetry imagery and a composite image obtained using a UAV. 23
Figure 12: Variable rate map constructed from Satamap data for residual
herbicide application. 23
Figure 13: The farm data ecosystem. 28
Figure 14: A map of the interoperability network that has developed
amongst digital agriculture systems and platforms in the US. 29
Figure 15: Australian and New Zealand croplands mapped using SST Software. 32
Figure 16: An example of the 1:25000 soil maps and data accessible
at the individual field level from the USDA. 32
Figure 17: Trademark available to be used by New Zealand companies
compliant with the New Zealand Code of Practice. 39
Figure 18: 3G mobile telephone coverage provided by Telstra. 51
Table 3: Example of standard use options that may be made available for farmers. 57
ix
Glossary of terms
Term Meaning
Agriculture technology providers (ATPs) is the collective term given to businesses providing both
Agricultural technology
technology and associated data management services to farmers. They include seed or machinery
provider (ATP)
suppliers offering yield-monitoring decision services or decision-support tools.
Big data refers to the analysis of datasets with sizes beyond the ability of commonly used software tools
Big data to capture, manage and process within a tolerable elapsed time (Snijders et al. 2012). It can vary in
scope and detail, with some data very granular while other is more general.
Cloud storage is a model of data storage where the digital data is stored in logical pools, the physical
storage spans multiple servers (and often locations), and the physical environment is typically owned
Cloud computing
and managed by a hosting company. These cloud storage providers are responsible for keeping the data
services
available and accessible, and the physical environment protected and running. People and organisations
buy or lease storage capacity from the providers to store user, organisation, or application data.
Since the end of the 1980s technology development has allowed farmers to collect increasing
volumes of objective data at an individualised field or animal level. Digital agriculture refers to farm
Digital agriculture
management systems where decisions are taken using an increasing amount of digital information, in
order to increase productivity and sustainability.
Information and
ICT is a broad term used to refer to technologies that involve the use of computers, computer networks,
communication
telephone networks and internet networks to manage data and information.
technologies (ICT)
Metadata is the term used to describe the nature and structure of data. It describes the variables which
are included in a dataset, their respective formats, and other aspects of the architecture of a data file.
Metadata
Metadata is essential to the ability of different software systems and applications to be designed to
operate using data from a range of different sources.
National Livestock The Australian red meat industry has implemented the National Livestock Identification System (NLIS),
Identification System a scheme to ensure the quality and safety of beef, pork, lamb, sheepmeat and goatmeat. Each animal and
(NLIS) each property is identified with a unique numbers, allowing greater traceability.
Precision agriculture started as management principles on cropping farms. The use of electronic
sensors and GPS guidance systems on machinery allows farmers to adapt input decisions (fertilisation,
Precision agriculture
irrigation) according to field conditions. It started in 1983 in the US (Zwaenepoel & Bars 1997) with the
first technology enabled fertiliser rates to be varied within a field based on soil test and other data.
Variable rate VRA refers to a technology that is used to enable the application of inputs (fertiliser, water) within a
application (VRA) block or field at variable rates calculated using localised data, in order to adapt the amount supplied to
or Variable rate the specific need of the sub-field zone. Relevant rates can be calculated using information registered on
technology (VRT) a map directly using sensors mounted on the machine used to apply the product.
Real-time kinematic (RTK) satellite navigation is a technique used to enhance the precision of position
Positioning system
data derived from satellite-based positioning systems such as GPS, GLONASS, Galileo, BeiDou, and
using satellite data:
GAGAN. It uses measurements of the phase of the signals carrier wave, rather than the information
Real-time kinematic
content of the signal, and relies on a single reference station or an interpolated virtual station to provide
(RTK)
real-time corrections, providing down to centimetre-level accuracy.
x
Acronyms
ABS Australian Bureau of Statistics NLIS National Livestock Identification System
ACBI Australian Centre for Broadband Innovation NSO National Statistics Organisation
ACIPA Australian Centre for Intellectual Property OADA Open Agriculture Data Alliance
inAgriculture
OTC over the counter
AMS automatic milking systems
PA precision agriculture
APIs application programming interfaces
PC personal computer
APPs Australian Privacy Principles
PGR pasture growth rates
ATP Agricultural technology provider
PLMTs precision livestock management technologies
BMP Best Management Practices
PUC passive uplink connection
DNA deoxyribonucleic acid
RFID radio frequency identification
FOI Freedom of Information
Sedex Supplier Ethical Data Exchange
FOO feed on offer
Sheep CRC The Cooperative Research Centre for Sheep
GNSS Global Navigation Satellite System Industry Innovation
GPS global positioning systems TDM total dry matter
GRDC Grains Research and Development UAV unmanned aerial vehicle

Corporation
UNE The University of New England
IP intellectual property
US United States
Mbps megabits per second
USDA United States Department of Agriculture
MPI Ministry for Primary Industries
VRA variable rate application
NBN National Broadband Network
NDVI Normalised Difference Vegetation Index
1
1. Introduction
The development of farming machinery and While much of the development that has occurred
digital technology that is able to generate objective involves applications utilised for crop or plant
information about the status of soil, water, crops, production, digital technology is also increasingly
pasture and animals is quickly changing the way used in the livestock industries.
in which farm businesses can to be managed in
Australia. The emergence of digital agriculture has Applications include electronic livestock
the potential to enable farm operators to manage identification systems, and the use of these in
farms with a much higher degree of precision conjunction with satellite monitoring and robotic
than has been feasible in the past, and when meat processing systems to more efficiently
combined with the power of modern computers and manage broadacre livestock and to provide detailed
specialised software, has enabled the development feedback on individual animals. Data about the
of sophisticated decision-support tools that have performance of animals from specific herds or
the potential to assist farm managers to make better flocks, and the genetic potential of specific breeds
decisions and to manage larger areas of land or or bloodstock lines is also now being collected, and
numbers of livestock. can potentially be distributed widely.
The initial developments in digital agriculture In the more intensive livestock industries
involved global positioning system (GPS) enabled including dairy, pigs and poultry, electronic animal
machinery that were used to implement systems identification in combination with digital sensors is
such as controlled-traffic farming. Subsequent being used to detect individual animals health and
developments included the use of GPS and reproductive status, to monitor productivity, and to
enhanced harvester monitoring technology to individually tailor feed and medications for specific
produce digital yield maps. More recently, the animals.
development of variable-rate planting and spraying
equipment has enabled some crop farmers to The most recent international developments in
increase yields and reduce crop inputs by using digital agriculture involve the utilisation of data
variable application rates across a paddock. derived from a large number of individual farms
in centrally-managed expert systems which
The latest developments in this area include are used to prescribe very specific crop planting
harvesting equipment that is constantly connected and management programs on a field-by-field
to the internet and that can relay crop yield and basis. Historical weather, soil and previous years
machinery performance information in real-time to production data are used in combination with
an off-farm data storage site or digital application, information about the performance of particular
and variable rate planting technology that can crop varieties to formulate the optimum crop
modify planting recipes and fertiliser application planting strategy at a sub-field level, and this
rates on a sub-paddock basis. Related developments information is then utilised in conjunction with
include the utilisation of unmanned aerial digitally-enabled machinery to vary crop planting,
vehicles and satellites to monitor crop or pasture fertilising and spraying across a single field to plant
performance remotely. These have also been used and manage a crop. The crop is then monitored
to monitor flowering and fruiting in orchards and throughout its growth and the expert system can be
vineyards. used to make decisions about fertiliser or pesticide
2 CHAPTER 1: Introduction
applications. Finally, data obtained from harvest Ultimately, it is hoped that the information
equipment is fed back into the system to close gathered in this research will assist the agriculture
the loop and enable further enhancements to sector in Australia to develop a collaborative
performance in subsequent years. framework that clarifies all the associated legal
and other implications of the development of
While still at a development stage, the digital agriculture, and works towards achieving
implementation of robotic technology in the dairy common agreement about issues that may impede
and meat processing sectors is also creating the the widespread adoption of these technologies in
potential to develop similar closed-loop systems, Australia in the future.
whereby genomics, on-farm production data,
and milk and carcass data can all be integrated Objectives
into a single expert system and used to identify
opportunities to enhance productivity or to focus 1. To develop a clear understanding of the potential
production on specialised market opportunities. for digital agriculture to enhance productivity
growth in the Australian agriculture sector.
Digital agriculture provides the potential
for individual farmers to achieve substantial 2. To research and document the full range of
productivity improvements. However, it also legal and other issues associated with the
raises a number of issues for farmers in relation generation, collection and dissemination of
to the ownership of data, the ownership and digital information derived from Australian farm
management of systems and platforms hosting businesses.
that data, the rights that farmers hold over data
3. To develop, in collaboration with all interested
obtained from their farm, the extent to which data
parties, a clear understanding of the legal and
held by machinery companies, farm input suppliers
statutory implications of the development of
and processors can be sold or transferred to third
digital agriculture under Australian law, and
parties, the uses to which data from individual
to identify any deficiencies or potential areas
farms can be put, and even the legal status of that
of conflict that require resolution in order to
data in the event of litigation or a demand by a
remove potential impediments to wide adoption
government authority to access that data.
of the technologies.
The issue of farm data security has become the
4. To consider and analyse related issues
subject of discussions between United States (US)
such as skills capacity, training needs and
farmers and their farm input suppliers. Issues
telecommunications infrastructure that may act
that have come under consideration as potentially
as an impediment to the potential adoption and
contentious uses of farm data include people with
development of digital agriculture in Australia.
real-time access to harvester yield data from a large
number of harvesters using that information to trade
grain derivatives, the potential for farm data to be
Methodology
sold to input suppliers or banks for use in marketing The research undertaken in this project consisted
campaigns, or the potential for the data to be essentially of desktop and industry research.
accessed by anti-farming advocacy groups or a
government regulator in order to mount a campaign Research associated with the first objective
or prosecution against an individual farmer. involved a search of available published literature
on current and future potential applications of
The aim of the project outlined here is to gain a digitised information in farm management in both
clear understanding of the potential for digital the cropping and livestock industries. Most of the
agriculture to enhance productivity growth in published information available related to crop and
Australian agriculture, and to detail some of the horticulture production, although some detailed
legal and other implications of this development. developments in the livestock industries.
CHAPTER 1: Introduction 3
Interviews were conducted with industry Throughout the research, the focus was on finding
participants and researchers both in Australia and ways to develop some common understanding
overseas in order to gain a clear understanding of across agriculture about these issues in order that
both the current and future potential applications of potential impediments to the widespread adoption
digital agriculture. of digital farming can be removed, and Australian
farmers are given the best opportunity to attain
During the research and interviews, the focus the productivity gains the technologies have the
was on gaining an understanding of the nature potential to deliver.
of the digital information collected, the way that
information is manipulated and stored, the uses Scope
to which that information is put, and the potential
for the information to be utilised by third parties Digital agriculture is relatively well developed in
or for purposes other than assisting farmers with the US cropping sector, and especially within the
production decision-making. US corn industry, where it is estimated that up to
40% of growers producing up to 70% of all US
A related part of the international research involved corn commonly utilise digital information systems
obtaining a comprehensive understanding of the to monitor and manage their cropping activities, and
legal and statutory framework associated with the use that information to vary planting, fertilising and
ownership and use of digital agriculture data in pesticide applications within single fields in order to
specific overseas jurisdictions. The focus of this optimise productivity and/or profitability.
work was on the US, given the relatively advanced
state of development of digital agriculture in that Digital agriculture is relatively less well developed
nation. in Australia. Industry surveys have indicated that
approximately 20% of grain growers have used
The research associated with the second objective variable rate fertiliser applications on their crops,
involved discussions with relevant government and but the use of integrated digital information systems
legal experts in Australia at both the national and and software platforms is much less common in
state level about existing laws or legal precedents Australia than is the case for the US.
that may have implications in relation to the
ownership or use of information arising from Based on the above, it was determined that the
digital agriculture. These laws were predominantly scope of the research associated with this project
those associated with ownership and transmission should encompass a detailed investigation of
of digital information, but also involved issues the development of digital agriculture in the US,
suchas: as a means of gaining insights into the likely
progression of developments in Australia.
the nature of contractual arrangements between
the purchasers and suppliers of technology and It was also considered that, while developments
machinery used in farming in most livestock sectors lag those in the cropping
sectors, the potential for digital information
the rights of individuals in relation to systems to develop rapidly in livestock production,
information about their farm collected via especially the more intensive livestock sectors,
remote sensing, and also information relating to is quite large, and therefore warranted close
input use examination.
the extent to which digital information obtained
from a farm can be accessed by government The scope of the research was limited to digital
authorities, third parties or through legal information associated with farm production
processes for commercial purposes or for activities, and did not include matters such as
compliance monitoring or prosecution. the availability of financial or health-related
information about individuals, as these issues are
dealt with or being dealt with by other relevant
4 CHAPTER 1: Introduction
authorities, and are not specific to those involved in Given the specificity of this term to the cropping
farm businesses. sector, it was decided to use the term digital
agriculture in this report to avoid confusion, and to
The research involved an analysis of a range of clarify what is being considered as the application
different technologies, software products and of digital technologies to a much wider spectrum of
associated digital platforms, some of which are agriculture than just the cropping sector.
described in the report. The research did not set
out to provide a detailed list of all available digital Agricultural big data is a reference to a related,
technologies and systems that are being utilised in but different development. The term big data is
agriculture. Other publications have done this, and typically used to refer to computerised analytical
in any event the rate of change that is evident would systems that interrogate extremely large databases
make such a list redundant before it was published. of information in order to identify particular trends
The inclusion of information about a product or and correlations that can subsequently be used in
technology in this report should not be interpreted expert systems or probabilistic decision-support
as an endorsement of that product or technology. tools in order to help users make management
decisions.
Definitions
The most obvious example of the use of big
The terms big data and digital agriculture are data are the customer loyalty card systems that
commonly used in discussing future developments are operated by both major food retail chains in
in agriculture, and it is useful to ensure that these Australia. The data generated by their store sales are
two are clearly defined, as they do not refer to the incorporated into very large databases consisting
same thing, but they are related. of information about the demographic details and
shopping patterns of consumers. These can be
Digital agriculture is the term used to refer to
interrogated to gain a much better understanding
agriculture which involves the use of digital
of the purchasing habits of millions of Australian
sensors and information to support managerial
consumers, and the information arising from that
decision-making. The rapidly declining cost of
analysis can be used to tailor retail offers to the very
digital sensors has meant that an increasing range
specific needs of quite distinct groups of consumers.
of machinery and equipment used on-farm is
now equipped with digital sensors that record In the case of agriculture, big data is a reference
and transmit detailed objective information about to the collection and analysis of extremely large
that machines operations. Machinery that is now datasets derived from digital information systems
typically equipped with digital sensors includes on multiple farms or in multiple different locations.
tractors, harvesters, sprayers, seeders, haymaking As an example, a dataset containing the yield
equipment, livestock weighing scales, robotic results of all the hundreds of wheat variety trials
milking machines, weather stations, water pumps conducted across Australia over the last decade,
and irrigation systems. In all of these examples, which also included detailed information about the
the machinery or equipment is able to generate location of each trial, soil types and test results of
large volumes of objective data, and often has the each of the trial plots, and the temperatures and
capacity to store or transmit these data wirelessly rainfall experienced during the growing period
or to an internet-based storage facility. These, often for each trial, could be analysed and the results
in combination with GPS technology, enable much utilised to create a predictive tool to assist farmers
more objective and spatially precise information to in deciding which wheat variety is best suited to
be generated and used in farm decision-making. their specific location and growing conditions. Such
an application would be an example of a decision-
The term precision agriculture is commonly used
support tool developed through the use of big data
to refer to cropping practices that involve the use
analytics, and made possible through the use of
of GPS guidance systems, variable rate seeders,
information derived from digital agriculture.
fertiliser spreaders, spray rigs and harvesters.
CHAPTER 1: Introduction 5
One of the more challenging aspects of agricultural single farm, by itself, has essentially no value other
big data arises from the fact that digital information than perhaps to the owner. This is because it is able
detailing the performance of a specific crop in a to be instantly duplicated and transmitted, but also
specific paddock in a single year is of very limited because in the absence of contextual information
value to either a farmer or the industry, yet the same about a wide range of production factors or in the
information obtained from multiple crops on many absence of very large amounts of similar data from
farms over a number of years may be very valuable other farms, there is really few actionable decision
in the development of computer systems that are that can be made based on that limited amount
very useful in assisting farmer decision-making. ofdata.
As a consequence, it can often seem to farmers that Just as data about the purchase decisions of a
the digital information that can now be collected single 24 year old female supermarket shopper on a
relating to their farm, such as a harvester yield map, specific Saturday morning in November are of very
has little more than curiosity value. Experience in limited value to a major supermarket chain, so the
the US corn industry, however, is showing that after data obtained during the harvest of a single paddock
perhaps a decade of collecting so-called useless of wheat in Western Victoria during a particular
information, the volume of digital data that is year is also of very limited value. Similar data from
available is enabling the development of robust many thousands of shoppers or wheat paddocks
new decision-support tools and enabling changed may, however, have significant value.
management systems specifically the use of
variable rate application (VRA) systems that are The need for large volumes of data, and the lack
delivering significant productivity improvements. of value of limited amounts of data, creates some
particular challenges in the development of big data
A related concept that farmers often find difficult applications in the agriculture sector.
to understand is that electronic farm data from a
6
2. Big data in the economy

X.0 XXXXX
It is not easy to grasp the scope of the changes that discernment at a commensurate scale, speed
big data has caused or has the potential to cause in a and intensity) can fundamentally change entire
national economy. On the one hand there are more economies, industries and lives (Davies 2015).
and more businesses developing systems to manage,
store, analyse and distribute electronic data, which One significant factor in the growth of the
is captured by the growth of the ICT sector. On availability of data is the increasing tendency of a
the other hand, businesses in all of the traditional wide range of consumer and manufactured products
sectors of the economy are changing the way they to incorporate internet connectivity as a standard
manage their workforce and assets in response to the part of the product. The resulting internet of things
insights that are now available about their relevant is dramatically increasing the number of things that
markets, as a consequence of the development of are connected to the internet, and the volume of
big data and digital information systems. data they are generating. Some sense of this growth
can be gleaned from the statistic that globally, 23
The digital economy contributed $79 billion million cars were connected to the internet in 2013,
(or 5.1%) of Australian GDP in 201314, and this number is projected to grow to 152 million
(Australian Computer Society 2015) based on by 2020 (IHS Inc. 2013).
the latest methodology used to estimate the rising
digital intensity within traditional businesses. This is creating the opportunity to collect digitised
The Australian digital economy is growing in information at a scale that was unimaginable even
significance and is 50% larger in real terms than it a few years ago, and to utilise this information in
was in 2011. If the digital economy was an industry an enormous number of different ways. Everything
it would be larger than Australias agriculture, from daily car traffic flows through tunnels to
transport or retail industries (Deloitte Access demand for airline tickets, to human disease
Economics 2015). management and control strategies are now being
managed via the use of big data applications.
According to a recent Big Data Roundtable held in
Australia earlier this year, three major factors have The growth of the digital economy and big data is
piqued global interest in big data. These are: a global phenomenon, and it is creating challenges
for policy-makers because of the potential it
1. an exponential increase in the amount of has to disrupt traditional industries and business
data that is being collected with forecasts of arrangements. Markets are changing rapidly in
continued rapid growth response, with national boundaries becoming less
relevant, and disruptive technologies allowing
2. an exponential decrease in the cost of computing new entrants to challenge previously dominant
(to process data) and networking/sensors (to organisations or business systems. Those that are
capture data) failing to respond, whether nationally or at the
firm level, are falling by the wayside, and national
3. increased understanding that big data governments are recognising that strategies and
particularly when coupled with big judgment policies are needed to facilitate the responses that
(that is, the capacity for analysis and are required in national economies.
CHAPTER 2: Big data in the economy 7
To unlock the potential of big data, OECD countries cameras and electronic tags in vehicles or carried as
need to develop coherent policies and practices for the part of corporate security systems.
collection, transport, storage, provision and use of data.
These policies cover issues such as privacy protection,
open data access, skills and employment, infrastructure, The issues and possibilities that emerge from
and measurement, among others. (OECD 2013) these developments in the wider economy are
triggering changes as governments, corporations
There are numerous issues that emerge from the and individuals attempt to come to grips with
development of big data in national economies that the implications of these changes. In some ways,
policy-makers are now beginning to grapple with. the agriculture sector has been less directly
The notion of what constitutes personal privacy has exposed to these changes than some other
been altered in ways that were not imaginable even sectors of the economy due to relatively poor
a decade ago, with digital information now able to telecommunications connectedness, and the varied
be used to trace, amongst other things, a persons and biological nature of agricultural production
daily movements, purchasing habits, workplace systems.
performance, financial assets and personal
relationships. The increasing connectedness of farm machinery,
the developments of smartphone-based
The development of cloud data storage systems farming software applications, and the growing
potentially necessitates completely new notions of interconnectedness of the wider business
what constitutes a personal possession, and creates environment in which farm businesses operate
the potential for government surveillance at a level means that the digital transformation of agriculture
most would find highly intrusive. Finding the right is now proceeding at an ever-accelerating pace, and
balance between privacy and public safety has businesses in the sector will increasingly need to
become much more challenging, given the potential respond to these changes, just as are businesses in
opened up by smartphones, closed circuit television the rest of the economy.
8
X.0 XXXXX
3. The development of digital agriculture
The development of agricultural equipment with found that around 20% of Australian grain growers
the capacity to collect and store relevant digital used variable rate application of fertiliser in 2011,
information represented the initial emergence of up from 5% in 2005 (Robertson et al. 2011).
digital agriculture in Australia, and as noted earlier
this is a necessary precursor to the development of Previous research indicated that there was a
agricultural big data applications. The emergence of significant gap between the commercially available
internet enabled devices and cloud storage platforms technology and the level of adoption by farmers.
that can readily be integrated with farm equipment Of course, seemingly slow rates of adoption of new
has subsequently created an environment from which technology are not unique to this technology, nor to
big data applications can more rapidly emerge. the agriculture sector of Australia more generally. For
example, in the US in 2011, despite wide availability,
The big data development cycle relies firstly on the harvester yield monitors were used on only 40% of
availability of digitally-enabled equipment and the grain crop acres, guidance systems on less than 35%
accumulation of digital information of sufficient of planted acres (winter wheat) and, variable rate
volume to provide a robust database which can applicators operated on less than 14% of planted
be utilised to develop predictive and probabilistic acres (Schimmelpfennig & Ebel 2011). A more
decision-support tools. A virtuous cycle is then able recent survey in the US state of Illinois revealed that
to develop, whereby the availability of improved more than 75% of corn farmers surveyed routinely
decision-support tools encourages more farmers to utilised variable rate fertiliser applications, and 40%
uptake digital agriculture, thereby increasing the utilised variable rate planting (Hale Group 2014).
volume of data available which can then be utilised
to continue the improvement of the decision- In Australia, the Grains Research and Development
support tools. Corporation (GRDC) Farm Practices Survey
Report provided results from a survey of cropping
Digital agriculture, the initial step in this cycle, first technology use by over 2500 grain farmers
emerged in the cropping sector in Australia in the in 2011 (Edwards et al. 2012). The survey
1990s, as farm machinery began to be equipped indicated that variable rate applications were
with various digital information systems, including used on 8.1% of cropland area in that year, and
the ability to generate maps of crop yields across that there was significant regional variation. In
a paddock or to store information about water 2015, GrainGrowers conducted an Agriculture
flows and soil moisture levels under irrigation Technology Survey into the adoption of cropping
systems. Despite the growing accessibility of this technologies. The result of that survey was that
information to farmers, many farmers initially 17% of responders claimed to use variable rate
remained uncertain about how to exploit its applicators, although the area over which that
opportunities (Cook & Bramley 2000). technology was utilised was not reported (Grain
Growers Limited 2015).
A review of the literature regarding the adoption
of digital agriculture shows the rate of uptake by The conclusion from these two survey is that the
farmers is increasing after an initial period of low adoption of variable rate applications is much lower
adoption. Surveys conducted by Robertson et al. in Australia than is currently reported for the US.
CHAPTER 3: The development of digital agriculture 9
35
a
30
25
% of crop area
20
15
b
bc bc
bc
10
bc bcd
bcd bcd bcd
bcd
5 cd d
d
0
NSW Central
NSW NE / QLD SE
NSW NW / QLD SW
NSW / VIC Slopes
QLD Central
SA MidNorth / Lower EP
SA / VIC Btown Wimmera
SA / VIC Mallee
TAS
VIC H Rainfall
WA Central
WA East
WA Mallee / Sand
WA North
Figure 1: Percentage of regional cropping areas on which variable rate applications were used in 2011.
Source: Edwards et al. (2012).
It should be noted that the adoption of variable the technology process and the potential benefits
rate application (VRA) cropping technology does farmers can achieve when equipment and data
not necessarily rely on digital agriculture. The systems are used effectively. This is true for the
authors of the survey reported above noted that grain sector and most notably for the livestock
some adopters of VRA were utilising soils test and sector. This point was frequently mentioned in
electromagnetic survey results to manually vary literature discussing these issues.
sowing and fertiliser rates in response to variations
in soil fertility within paddocks. There are two opportunities to close this gap. Early
learning and education of farmers is one such
The adoption rates for digital agriculture in the approach (Eastwood 2008), the other involves
cropping sector are presumably affected by the the development of improved user interfaces and
fact that accessing the available technology can product integration from manufacturers. There
involve major capital investment in the form of were frequent comments in the literature about the
a new harvester, spray rig, seeder (or planter) or limited use of technology by Australian farmers due
tractor. This type of decision is of a markedly to difficulties in integrating software components,
different level of significance to a decision about problems with data interpretation and using
the adoption of a specific new crop variety, for technology to apply agronomic solutions (Jochinke
example, where the financial outlay may only be et al. 2007).
marginally different to that associated with the
current variety. Generally, it seems farmers implicitly recognise the
need for a data strategy associated with the adoption
There is an additional barrier to adoption in that of digital agriculture or precision agriculture (PA).
to adopt and use these technologies, not only In interviews many have expressed or observed
do farmers have to make financial investments, a reticence to invest in new digitally-enabled
but they may require time to learn new skills. technology unless they can see financial benefit or
There is also a gap between the user phase of feel capable of deriving one from the resulting data.
10 CHAPTER 3: The development of digital agriculture
This attitude broadly aligns with the recognised Other studies have examined the costs of investing
need for a data strategy which incorporates data in digital agriculture technology compared to
technology into a wider solution rather than a the costs of conventional farming technology
reliance on the technology to drive operational on a whole farm basis. Most found that while
improvement of its own accord (Jochinke et al. digital agriculture tools have the potential to save
2007). money for farmers by increasing efficiencies in
broadacre cropping systems, the initial time and
Estimating the financial benefits associated with monetary outlay has always limited its adoption
the adoption of digital agriculture (specifically compared to conventional technology (Jochinke
VRA in the cropping sector) is a difficult exercise et al. 2007). However, Robertson et al. noted
and agricultural economists have not yet come to that farmers are taking a stepwise approach to
a common methodology. Case study results from digital agriculture adoption whereby they make
an analysis of on-property benefits from precision sequential investments in components such as GPS,
livestock management technologies (PLMTs) yield mapping, variable rate applicators etc. This
provided useful information and cost-benefit ratios ameliorates some of the initial outlay requirements.
for specific technologies. The economic benefits Growing market size and better understanding of
were assessed using a subjective methodology to farmer requirements are expected to further reduce
estimate that production and cost saving benefits. implementation costs for farmers (Robertson et al.
The distribution profile for each technology 2011).
provided data that is to some extent used to estimate
the risk of an economic return for each given Literature on the productivity gains that are
technology (Swain et al. 2013). estimated to be available as a result of the
deployment of big data applications in agriculture
In a report written by the Hale Group (2014) on is limited due to the relative immaturity of the
behalf of Iowa Agstate Farm Group, estimates of technology. What research exists tends to focus on
potential gains were provided for corn cropping. the cropping sector and specific technologies that
The study compared the differences in returns and harness more accurate application of inputs or the
costs between innovators using best available use of site-specific information to develop models
digital agriculture and data for full variable rate or tools for producers.
cropping systems and farmers not currently utilising
VRA, but using precision agriculture technologies. The following section examines some of the
specific digital technologies being utilised in
Estimates listed in the report were as follows: different sectors of agriculture.
Yield gains achieved were between

5to10bushels of corn per acre, which at 3.1 Cropping
US$3.50/bushel means a potential gain of Historically, fertiliser, seed and pesticide input
US$18 to US$35 per acre gross. recommendations for crop production were
Improved nitrogen efficiency, reducing costs by prescribed as a best-fit average for an entire
US$25 to US$30 per acre. paddock. The advent of variable rate application
technology, more intensive soil testing and
Gross savings of US$43 to US$65 per acre. electromagnetic soil surveys has allowed for
Deducting from this the US$3 to US$10 per acre greater precision in cropping operations by varying
normally charged by service providers for data application rates of fertilisers and pesticides within
storage, manipulation and management resulted a paddock, and by varying crop planting rates.
in net gains of between US$33 and US$62 per The result has been an improvement in crop farm
acre (Hale Group 2014). productivity, and in the efficiency of water and
fertiliser use.
Figure 2: Harvester fuel use map.

Source: AgInnovators (2016).
The relatively intensive use of inputs and machinery management issues. Figure 2 shows an example of
in the cropping sector has meant the sector has a novel analysis based on harvester fuel use rates
been particularly receptive to potential productivity that has identified five separate factors affecting
gains associated with the deployment of digital the efficiency of the harvester operation. These
agriculture. The size of, and level of investment were (1)contour banks, (2) slope, (3) weight
associated with the US corn and soybean industries of the spreader, (4) soil type or steepness, and
has also been an important factor in the level of (5)compaction caused by an irrigator.
development of digital agriculture technologies in
that sector, with large- and small-scale agricultural The increased sophistication of sensors and
innovation and technology companies making internet enabled devices allow streams of data to
very significant investments to develop new be captured from cropping operations, including
technologies. harvesting, spraying and seeding, as well as
data obtained from remote sensing via satellites,
The cropping sector has realised substantial or obtained via the use of unmanned aerial
operational gains from GPS applications that have vehicles (UAVs). Input suppliers and advisors to
been progressively adopted since the 1990s. The the cropping sector in the US have been in the
benefits obtained from the use of GPS and auto- vanguard of the development of data applications
steer systems in broadacre cropping have included that enable farmers and their advisors to store the
the minimisation of soil compaction, the ability to resulting data, retrieve the stored data, integrate
cultivate for reduced disease impact and herbicide various different sources of data relevant to a
dependence, the ability to minimise input overlap specific paddock, and ultimately to utilise suitable
and waste, and opportunities to improve soil water algorithms to make probabilistic projections about
management. The use of GPS technologies also the implications on specific crop management
makes feasible the analysis of a many additional decisions in particular associated with varying
factors that can impact on the efficiency of input inputs across a paddock.
use, and ultimately crop profitability.
More recent software applications have included
An example is the use of maps showing harvester the capacity to retrieve climate and soil data for a
fuel use rates within a paddock to identify particular specific paddock from public information sources
and to integrate this with private production data; Many of the software systems offered in the US
and a soil-test gridding system that enables users to provide data storage and retrieval capabilities, but
map the location where samples have been taken for have only limited capability to integrate available
soil tests, and attach the data arising from soil test data in a way that supports decision-making
results to that same location on the digital map. by users. This data integration function is still
performed manually by crop advisors and farmers,
In the US the private sector has been leading especially when it comes to drawing up variable
the commercialisation of digital agriculture via rate application zones within a paddock (field).
the development of software and hardware tools
that assist with the integration of the rapidly A limited number of services and software systems
growing flow of digital information that can now have developed to the extent that the variety of
be generated. At a recent conference convened in different pieces of digital information are integrated
St. Louis, Missouri, for example, more than 100 in ways that assist farmer or advisor decision-
different technology and software providers were making. One of the earliest of these was the service
present, with a large range of different technologies offered by Monsanto called Field Scripts. Crop
and data integration products. These included: farmers opting to utilise Field Scripts advised their
seed retailer, who assisted the farmers with field
software systems for farmers and their advisors mapping, and the retrieval of historical soil test and
software systems for crop input resellers harvest data relevant to that field. That information
was provided to Monsanto, who combined that
UAV control and image analysis software information with soil, weather and other data
digitised irrigation and water management (including a large repository of variety trial results)
systems using proprietary algorithms to produce a variable
rate script or cropping recipe (on an ipad) that
paddock mapping software the farmer used in conjunction with a Precision
cloud-based data storage and retrieval services Planter controlled variable rate seeder to plant the
corn crop. The crop was then monitored remotely
software systems that are able to access public on a number of occasions during the growing
climate and soil information for a specific season, and the farmers provided with advice on
paddock matters such as fertilisers and pest control. Finally,
major machinery and seed suppliers offering information from the resulting yield map generated
data storage and analysis platforms during harvest was accessed to close the loop
and refine the system for use in subsequent years.
suppliers of digital control systems for a wide
According to Monsanto, the crop farmers utilising
range of different types of farm machinery
the service achieved yield gains of up to 20%.
suppliers of data integration systems and
technologies that enable the transfer of The Field Scripts system was offered for three
information between different machines and successive years, and Monsanto reported strong
operating systems. uptake by farmers, and a high level of product
loyalty by those who tried the system. The system
A number of these products are available in Australia was discontinued at the end of the 2015 cropping
either as part of a package of systems associated year due to servicing costs and risks, although in its
with large-scale farm machinery, or as stand-alone place Monsanto (through its subsidiary the Climate
products able to be used by farmers and their Corporation) has released a range of software
advisors to manage cropping programs. The different products including Fieldview Prime, Fieldview
cropping systems that exist and the differences in Plus and Fieldview Pro (the latter two being
the availability of public climate and soil data have subscription-based systems) that support many of
probably limited the availability and rate of adoption the functions offered via Field Scripts although in a
of many of these systems in Australia. less prescriptive and more user-friendly way.
and employing data scientists to interrogate them

in order to obtain actionable insights. These
organisations are also developing computer models
of specific components of crop production (eg soil
water uptake, nutrient uptake, etc). These can then
be used to develop algorithms and models which
may not necessarily be perfect at first, but which
can be progressively updated as identified gaps in
knowledge and information are filled over time.
An opinion common to US industry representatives

was that the cropping sector is in the early stages
Figure 3: Climate Corporation Nitrogen of a data growth curve, whereby in a few years
Advisor. there will be enough environmental (soil, water,
Source: The Climate Corporation (2015). temperature, rainfall), production and crop data to
The current generation of software provided by the start to make use of quite complex decision-support
Climate Corporation include decision-support tools tools. At present, a lack of robust historical yield
that offer advice about the weather, crop variety and input data inhibits the widespread development
performance and fertiliser depletion rates, rather of analytics products. Crop yield data, for example,
than telling farmers what, how and where to plant is of limited value without records of seasonal
specific crops. One of the most used developments rainfall and fertiliser applications.
by Monsanto has been the Nitrogen Advisor within
the current Climate/Fieldview software family. This Most industry authorities consulted agree that at
user interface application provides farmers with least five years of yield, fertiliser and weather data
projections of nitrogen fertiliser depletion in crops. are needed to make analytics products useful and
robust. Many US farmers have purchased new,
A user can input the intended nitrogen fertiliser digitally-enabled cropping machinery during the
application rate and the software projects forward recent period of relatively high profitability in the
the rate of depletion of that fertiliser in the soil, US corn industry. It will take several more seasons
based on known soil type and the likely climate with this machinery in operation before multi-year
(based on detailed historical climate data) for datasets become widely available in sufficient
that particular field. This provides users with the volume in order to allow for the rapid growth in
opportunity to finetune fertiliser applications, a accuracy and proliferation of predictive analytical
function that is potentially quite valuable given the software applications.
relative importance of nitrogen fertiliser as a crop
input in the US corn industry. The scope of digital products that have been
developed for use by US crop farmers is growing
The application was used to map and make quickly on the back of the expanding pool of
decisions on more than 75 million acres of crops available digital information, and this growth is also
in 2015 (up from 50 million acres in 2014). This being aided by the availability of comprehensive and
represents approximately 45% of all corn and granular public data on climate and soils in the US.
soybean plantings in the US during those seasons.
The more sophisticated product, Fieldview Pro, An example is Farmlink, a company that had its
which is only available as a fee-based subscription genesis in harvester leasing in the US. The data
service, was reportedly utilised on 5 million acres analytics business of the company developed from
in 2015. the realisation that the harvesters they were leasing
across multiple US states were generating reams of
This is just one example of a number of agricultural yield data each year, but potential calibration errors
input suppliers which are collating big datasets meant this data was of questionable value. Farmlink
developed standardised harvester calibration a number of different software applications that are
systems, and also identified technology that enabled primarily targeted at retail crop advisors, enabling
harvester yield data from individual harvesters to them to map clients cropping programs, identify
be transmitted to a central storage facility. The yield and order required inputs, and provide advice to
data is then overlaid with weather, soil, topography clients about available crop protection and fertiliser
and other data which allows Farmlink to benchmark products. This system is used by several major
farm performance between similar growing Australian crop input suppliers, and has been used
environments at the sub-field level in fact for plots to map large areas of crops over recent years.
of around 150 square feet. Software products such as Agworld and Farmware
provide some similar capabilities, although the
Using this system, Farmlink believes they now have latter also functions as a farm notebook to record
reliable harvest data from many thousands of fields, livestock activities.
which can be used to provide crop advisors with
benchmark information to compare the performance Productionwise is a digital crop and farm
of their farmer clients. This information is also being management system that has been developed in
utilised firstly to calibrate satellite imagery, and then Australia by GrainGrowers. It includes mapping
to use this to make harvest and production estimates, and information recording capabilities, weather
which it is believed will closely rival the accuracy information, chemical and fertiliser record-keeping
and timeliness of official government forecasts. functions, grain stock and crop gross margin
information, and utilises satellite and online climate
A number of different software support systems and soil data to provide vegetation and predicted
are available and currently being utilised in the yield information for each paddock. The system
cropping sector in Australia. SST Software provides also incorporates a number of decision-support
tools that are essentially based on the APSIM
crop model, originally developed by the CSIRO
Box 1: Farmlink TrueHarvest with funding provided by the GRDC. The system
is also able to be utilised either by crop advisors
The following figure is an example of
or farmers. This software system does not as yet
Farmlinks Gap Maps which enables crop
include tools to integrate digital data generated from
advisors to compare the performance of their
a range of different sources including harvesters,
clients crops to those of other crop farmers
but is undergoing further development.
growing under identical soil and climatic
conditions. The ultimate goal of much big data analytics in the
cropping sector is to empirically derive optimal
crop management decisions based on the analysis of
objective farm data. The rate of progress in achieving
this is likely to be iterative, as improved computer
applications will encourage greater adoption of
digital technologies by farmers, which will in turn
increase the volume of data available and hence the
robustness of the computer applications.
To a degree the rate of progress in the initial

stages will depend on the development of ancillary
products (such as the Nitrogen Advisor) that deliver
value to farmers and encourage them to collect and
Figure 4: TrueHarvest Gap Map. accumulate digital information associated with their
cropping activities.
Source: Farmlink (2015).
3.2 Extensive livestock 49% for animal production monitoring

While to some extent the availability of digital 413% for animal health monitoring.
technology associated with the management of
livestock enterprises is as advanced as is the case These preliminary findings need to be tested
in the cropping sector, the routine use of these and validated in a range of different agricultural
technologies, and the development of software industries so that more accurate estimates of
platforms to manage the information they produce productivity benefits can be obtained (Griffith et
is less advanced. Individual electronic animal al. 2013). Other reports from different agricultural
identification systems have been operating in the industries have identified similar positive
beef industry in Australia since 2004, for example, productivity benefits from the adoption of digital
but there has not been anywhere near the same level agriculture.
of software or systems development as has occurred
in the cropping sector since that time. The figure below illustrates the relationship
between a number of farm technology systems and
Digital agricultural technologies now enable on-farm needs relevant to the southern livestock
farmers to record numerous attributes of their industries.
livestock, such as reproductive state, weight gain, Soil mapping Soil fertility
feed conversion ratios and killing-out-percentages Pasture biomass Feed allocation
(carcass weight as percentage of live weight) that Pasture quality Pasture yield
can be used to improve farm management decisions. Animal behaviour Feed prediction
A recent report by Meat & Livestock Australia Animal location

- coarse scale
Weed management
(MLA) assessed the areas of greatest potential Animal location - fine scale
Stock auditing
benefit for beef and sheep livestock enterprises Proximity logging

Stock security
arising from the use of digital technologies (Henry Intra-animal monitoring

Animal association
et al. 2012). The research identified the four main (eg rumen) Animal health
areas with the highest productivity benefits: Animal liveweight

and condition
Animal production
soil fertility monitoring for improved pasture Figure 5: Diagrammatic representation of how
production various data from technologies may
feed allocation systems (allocating appropriate contribute to identified on-farm needs.
quality and quantity of feed to different classes Source: Griffith et al. (2013).
of stock in a timely manner)

Some of the specific digital technologies that
animal production monitoring (monitoring are currently available for the management of
animal weight and body condition to improve broadacre livestock are discussed in more detail
reproductive performance and animal growth below.
rates)
animal disease monitoring (early detection of Auto-drafters and walk-over-weighing
subclinical diseases to improve performance and platforms
welfare). Digital information is now able to be collected
on livestock, pasture and the environment using
The MLA report estimated that the potential technologies such as walk-over weighing scales,
productivity benefits arising from improvements in satellite imagery, GPS collars and weather stations.
farm management, based on selected case studies While all these technologies are available, few
were: are in use in commercial livestock production
enterprises in Australia, although there is growing
1326% for soil fertility improvements interest in the possibility of adopting some of these
911% for better feed allocation by larger-scale commercial operators.
Auto-drafting equipment based on animal weight is trialled technologies. It was also estimated that this
now commercially available for the beef and sheep technology provides an overall average improvement
industries, however data can generally only be of 2% in gross margins (Swain et al.2013).
saved on hardware such as a wand via a Bluetooth
connection. An integrated telemetry system to The walk-over-weighing systems can also provide
upload information in real-time to a user interface a platform for a range of other sensors including
via the internet is currently being tested at CSIROs frame height sensors and cameras that deliver
Digital Homestead in rural Queensland. The digital still images of the livestock, however the
project aims to demonstrate and evaluate livestock accuracy and practicality of this function is not yet
technologies to enable better decision-making. determined. Some trial systems have been refined
In the pastoral regions of northern Australia, this so they can record the weekly weight of individual
technology holds the promise of a substantial animals as they walk over the weighing platform.
reduction in mustering costs for those businesses These data can then be cross-referenced with the
which rely on controlled waters. weight change of the herd as a whole (Brown et al.
2012). This allows producers to identify animals
The CSIRO trials also involve sourcing other showing signs of aberrant behaviours and can be
external farm information such as meat processor selectively contained for physical examination.
pricing schedules, local sale results and weather For producers, this information enables them to
forecasts. Information is integrated and displayed monitor individual and herd live-weight and weight
as a dashboard where data can be accessed by gains on a regular basis. For animals destined for
simply clicking on a particular paddock or herd sale this technology facilitates decision-making by
(Delaney 2015). Although actual productivity gains ensuring market specifications are met with respect
from this project cannot be directly related to farms to live-weight.
outside of the project, the results have provided
information on certain technologies that proved Systems which include auto-drafting capability
more economically viable than others. currently cost approximately $30,000. The major
cost saving benefits from this technology are related
The average savings from using a walk-over- to savings in mustering costs, either helicopter or
weighing system were estimated to be around man hours. Table 1 shows improvements that were
10% of current mustering costs, the highest of all made on large-scale beef properties in areas such
Table 1: Estimated benefits from walk-over-weighing and auto-drafting systems.
Property
1 2 3 4 5
Branding percentage 65% 68%

(annual % increase on current rate) (2.9) (1.13)
Carrying capacity for cell (hd) 1000

(% increase) (4.63)
Average daily live-weight gain (kg) 0.3 kg/day 0.53 kg/day 0.3 kg/day 0.4 kg/day
(% improvement) (3.5) (3.88) (1.5) (4.13)
Margin per kg ($) $0.30/kg $0.40/kg

(% increase in margin) (3.75) (3.75)
Helicopter mustering costs ($) $30,000 $128,000 $96,000

(% reduction in flying costs) (7.5) (14.5) (11.25)
Labour costs ($) $ 5600 $ 22,500

(% saving) (17.5) (7.0)
Source: Swain et al. (2013).
Figure 6: Remote livestock management

system.
Source: Ninti One (2013).
as branding percentages, carrying capacity, weight

gains, profit margins, and savings in mustering and
labour costs (Swain et al. 2013). Figure 7: ireless NLIS animal identification
W
and tracking system to assist with
Available commercial technology is claimed to herd management and breeding using
be capable of identifying, weighing and drafting Taggle ear tags.
individual animals when they step through Source: Griffith et al. (2013).
the station with an accuracy of 96% per pass
(see Figure 6) which is estimated to save cattle real-time on an application accessed on a mobile
producers around $68 a head in annual operational device or personal computer (PC). This information
costs (Bem 2013). The animal data captured can provides almost real-time information about the
be readily transferred to downstream management geo-location of the livestock, potentially preventing
applications and monitoring software capable of theft or loss through straying. The system uses
delivering further operational efficiency gains 15minute sample intervals, with the livestock tags
(Tru-Test Limited 2016). lasting up to an estimated three years based on
in-built battery capacity. This also overcomes the
The equipment is solar-powered and uses telemetry battery power challenge of satellite-based location
to transmit data between a remote location and the technologies. A Taggle system currently costs
office of a manager. It takes advantage of water approximately $5000 per tower and $20 per tag.
as the primary attractant for livestock by using an
obligatory walk-over-weighing facility at water This approach has been adopted at Australias first
points. SMART Farm, an initiative led by the Australian
Centre for Broadband Innovation (ACBI) in
Livestock tracking systems collaboration with CSIRO, and the University
of New Englands (UNE) Precision Agriculture
Technology is now becoming available to assist Research Group.
in locating herds on large properties. The Taggle
system employs Global Navigation Satellite System Kirby Smart Farm was one of the first mainland
(GNSS) positioning to provide an estimated location farms connected to the National Broadband Network
of the animal through an ear tag (see Figure 7). The (NBN) fixed wireless service and broadband which
receiving antennas send information to a server has significantly helped transform the farms
that processes the data which is then presented in operations. A low cost wireless cattle tracking system
indicators which can display weight and drafting

statistics and have the ability to calculate daily
Paddock 1
Paddock 5 weight gains, carcass weight and store tag numbers
and condition scores.
Paddock 3 Exclusion area
Paddock 2
The Cooperative Research Centre for Sheep
Industry Innovation (Sheep CRC) has been
collecting high-quality and large-scale datasets for
the industrys Information Nucleus program to help
0 200 400 800 metres Paddock 4 breeders manage expectations about lamb growth
rates, particularly for twins and triplets.
Recorded location of cattle
Paddocks accessable to cattle
The program involved the collection of a
Exclusion area (used for unrelated grazing trial) comprehensive set of measurements from 20,000
animals and the process of data collection, storage
Figure 8: G
PS positions of cattle as recorded and analysis would not have been possible without
by UNEtracker collars over semi-automated data collection and the improving
14consecutive days. ability to transmit, store and utilise digital data. In
Source: Lamb et al. (2008). order to maintain the accuracy of the predictive
equations used in the system, it is also necessary
and video-monitoring technologies help monitor to keep measuring subsequent generations of
individual animal behaviour and location. A smaller animals to ensure that changes in the patterns of
number of other sensors such as weather stations DNA (deoxyribonucleic acid) continue to predict
and light sensors are also deployed across the farm. breeding values accurately. The latest development
The operation utilises radio frequency identification of these data platforms has been the creation of the
(RFID) devices that can be carried externally on the Ram Buyer smartphone app which was released
animal such as an ear tag or internally in an RFID July 2015. Through the use of predictive algorithms
bolus. GPS collar trackers have also been utilised for and by collating data from the Information Nucleus
livestock management at Kirby Farm. Program, there is an opportunity to enhance the
clarity and transparency of seedstock markets, and
Figure 8 shows a 14 day trial of the tracking collars, establish true value-based pricing for the sheep
where positions of collars recorded and distances industry (Rowe & Banks 2015).
travelled per day were calculated.
The same data is valuable when it is also used to
The use of RFID in combination with automated make better decisions on genetic selection and
data capture systems aims at reducing the cost sharing information through the supply chain.
of useful data collection. Currently, the data New measurements being developed for carcass
is often used only at the time of collection for grading within abattoirs will primarily be used for
a management decision that is immediately carcass sorting and payment grids but will also have
implemented. However the true potential value considerable value in contributing to information
of data is achieved when it is combined with available for genetic selection and feedback to
other information for within flock or herd producers in order to improve management and
selection decisions such as culling or feeding production systems.
and reducing the risk of compromised wellbeing.
Basic management systems that are commercially The development of cloud and internet computing
available and commonly used today with RFID systems facilitates the development of centralised
include portable handheld/stick readers, permanent databases that can be used to ensure that data is
readers (fixed into the race), and weigh scale available for multiple purposes. Efficient data
collection and its effective use have the potential

to improve labour productivity considerably.
When combined with the benefits of faster genetic
gain and better livestock management, more
efficient data use also has the potential to improve
productivity through better informed and timelier
management decisions (Rowe & Banks 2015).
Remote and proximal sensors

Currently there are few remote pasture monitoring
technologies that can be utilised by farmers other
than the purchase of satellite imagery. Pastures
from Space is a CSIRO initiative that offers
services that include pasture growth rates (PGR),
feed on offer (FOO), total dry matter (TDM) and
Greenness Imagery, based on satellite data and Figure 9: Sense-T Pasture Predictor tool.
calculated as the Normalised Difference Vegetation Source: University of Tasmania (2015).
Index (NDVI), an index of vegetation greenness

soil moisture and temperature at various depths as
and density. While providing a useful index of
well as above-ground temperature, relative humidity,
vegetative state, there are limits to the use of
solar radiation and leaf wetness. The system saves
this imagery for intensive pasture management,
real-time data into a cloud storage facility, to be
especially in the case where mixed species are
aggregated with spatial and historical data.
involved, and in the absence of robust and detailed
calibration against ground observations and data. One of the key achievements of the Dairy and
Beef Project component within Sense-T, was the
There are also some proximal monitors that observe
development of the Sense-T Pasture Predictor,
vegetation greenness, which can serve as a good
an online tool that helps farmers to forecast their
indicator of pasture and crop health. A number
future pasture growth and make better decisions in
of these sensors have been investigated for dairy
managing their herds, production and costs.
pasture systems including ultrasonic and optical
plant height sensors (Awty 2009). However, the The Pasture Predictor uses data from a range of
sensors developed to date are based predominantly sources, including current weather conditions and
on correlations of pasture height to biomass and forecasts, rainfall events, past climate records and
suffer from the inability to delineate green and real-time soil moisture to provide accurate growth
senescent material and therefore have limitations in forecasts for 30 days and longer-term trends for up
relation to their usefulness for farmers (Trotter et al. to 90 days. Figure 9 above displays forecast pasture
2010). growth over the coming 90 days, expressed in terms
of kilograms of dry matter per hectare per day
Sense-T is a Tasmanian initiative that aims to
(kgDM/ha/d). The red shading indicates percentiles
digitise much of the relevant information that
in the 25% to 75% probability range. In the future,
could potentially be utilised by land managers in
Sense-T aims to develop the forecasts for individual
Tasmania. It is a partnership between the University
properties on a subscription, cost recovery basis
of Tasmania, the CSIRO and the Tasmanian
(University of Tasmania 2015). While pasture
Government, and is funded by the Australian
prediction tools are not new, this tool is expected
Government. It involves the collection of data from
to include improved accuracy and up-to-date data
a range of different public and private sources, with
andanalytics.
a particular focus on wireless sensor systems that are
easy to install and operate, collecting data such as
Virtual fences information services, for use on-farm, within

service centres, and across the industry.
Constructing and maintaining fences is a major
cost for livestock farmers, especially in the pastoral The MISTRO software system assists dairy farmers
regions. The CSIRO is currently testing a virtual to:
fence in which no physical posts and wires
are required. The livestock are confined within Collect information in the paddock or dairy shed
boundaries drawn entirely by a GPS system. The about cows and pastures.
fence exists only as computer code. The system
consists of a wireless sensor network and the use Maintain accurate records on-farm, herd, and
of cattle neck-collars that emit a sound when the cow management.
animals approach the virtual boundary. The Minimise data entry by interacting electronically
prototype successfully demonstrated that within one with herd recording centres.
hour the cattle learn to associate the sound signal
from their collars with the virtual boundary (Jouven Improve herd and feed management through
et al. 2012). better record keeping.
Obtain relevant information from their records
Before the system can become commercially that improves decision-making.
available, the durability and robustness of the cattle
neck-collars needs to be improved. The potential Process their financial transactions and submit
gains from this technology include reduced labour tax returns.
and costs, better use of pastures, protection of Develop cash flow budgets and financial plans.
environmentally sensitive areas and the ability to
collect information such as production, and health
Obtain information from the internet.
and welfare status of individual livestock. Interact with other dairy farmers on the internet.
MISTRO also provides software solutions for
3.3 Dairy herd improvement centres that provide artificial
Digital agricultural applications in the dairy insemination, merchandising, and herd recording
industry had their genesis in individual cow services. Such a system lends itself to future big
identification systems, as is the case in other data applications as a consequence of the range of
livestock industries. The functionality of in-built data collected, including genetic, production and
sensors within the individual ID systems has since financial information.
developed, as new technologies and more durable
battery systems have been developed. Available Since 2009, Australian dairy farmers have
sensors now provide information about animal expressed growing interest in robotic or automatic
activity levels, weight changes, blood composition, milking systems (AMS) (Dairy Australia Limited
milk characteristics and rumen parameters that 2014a). AMS provide greater flexibility of milking
may assist in management decisions supporting times and milking frequency than conventional
the health, welfare and reproductive management milking systems, eliminating the need to milk cows
of individual dairy cows. Biometric data are at regular set times. They also require less labour
increasingly integrated within financial and supply per cow or per litre of milk produced. This allows
chain data to form holistic management products the operator to shift focus to other areas of on-farm
(see NLP technologies case study). management such as feeding animals, animal health
treatments, insemination and calf rearing.
An example of an integrated dairy data management
system is the MISTRO system developed by AMS can also milk and monitor each quarter of a
Gippsland Herd Improvement which aims to cows udder individually, enabling the operator to
provide the dairy industry with leading edge assess production and some milk characteristics
data collection tools, computer software, and at an individual quarter level (compared to the
whole udder in conventional milking systems). This Other recent systems that have been developed
provides better capacity to manage disease issues alongside these activity meters are ear tags that
such as mastitis before animal health is affected. measure in-calf heat detection and rumination
Milk quality parameters such as conductivity and parameters.
milk colour can also be measured regularly, as can
feed and supplement intake. Similar to other livestock industries, the use of
walk-over-weighing scales to weigh individual
Achieving optimal management outcomes for cows has become more common in the dairy
AMS is more challenging under Australian grazing industry, with data automatically recorded on the
conditions than in Europe where cows are housed dairys computer system. Often these systems
indoors most of the time or in a feedlot situation are used to help provide feeding strategies for
where many of the production parameters can be individual cows or groups of cows (Dairy Australia
controlled (Future Dairy 2009). Limited 2014b).
This is highlighted in the data displayed in Table 2

which provides a comparison of milking machine
3.4 Horticulture
utilisation and performance in AMS between As a sector of agriculture with relatively intensive
Australia and Europe. Nevertheless, the systems management systems, high levels of inputs and high
still provide significant opportunities to improve value of production per hectare, the horticulture
dairy herd management as a consequence of the industry is ideally suited to the application of digital
objective digital information available from these technologies to enhance production information
systems. Table 2 shows potential achievable AMS and to aid decision-making. The use of digital
milking machine utilisation levels of well-managed information and technology in horticulture extends
systems during periods of high utilisation (Future from monitoring and management of inputs such
Dairy 2009). as water and fertilisers (for example fertigation
systems controlled by remote telemetry) to
Other management advantages arising from the monitoring of plant, insect and soil conditions,
use of integrated digital information systems (such analysing flowering and fruit setting, through to
as activity meters) in dairy production include robotic harvesting and automated grading, packing
much more efficient and accurate monitoring of the and chilling systems.
reproductive status of individual cows, a critical
element of herd management. Case studies have Intensive horticultural production systems, in
estimated that by using activity meters as a back-up particular those used for large-scale glasshouse
to visual cow heat detection saves at least $7000 per production of flowers and vegetables, have been
year for a 450 cow herd (Dairy Australia Limited developed as highly automated facilities utilising
2015).
Table 2: Comparsion of AMS performance under Australian and European management systems.
Australian pasture Australian pasture

Typical European
based system, based system,
system
Camden: peak of season Camden: annual average
Average milking machine

90% 80% 67%
utilisation
Number of milkings/
170 150 118
machine/day
Milk (litres) harvested/

2300 2000 1384
machine/day
and livestock surveillance, and uses have even

extended to livestock mustering. For pasture and
crop assessment applications, these typically
operate by taking large numbers of individual
images which are downloaded either directly from
a sim card at the completion of the flight or via
wireless communications systems during the flight.
These images are then processed to correct for
various distortions such as topographic relief, lens
distortion and camera tilt, and then digitally stitched
together to create a single, composite image.
The use of both fixed wing and hover drones is

Figure 10: Autonomous robot conducting facilitated by integrated software systems (such
surveillance in an almond farm at as Drone Deploy, see Figure 11) that enable the
Mildura. areas which are to be scanned to be first mapped
Source: The University of Sydney: Faculty of Engineering & Information digitally, and the software then automatically
Technologies (2015).
retrieves relevant meteorological data and creates
digital information to control all parts of the an optimal flight path, taking into account wind
production system. speed and direction. After the drone is launched,
the software then controls the entire flight, while
Digital information systems and robotic simultaneously stitching together images obtained
technologies are being developed for the more from each pass over the field, which are downloaded
extensive horticultural sectors, such as almonds wirelessly during the flight. Within minutes of the
and olives. A key driver of these developments drone landing at the completion of a flight, the
is the relatively high level of labour utilisation in complete digital image of the field is available on the
horticulture, combined with high Australian labour computer as a colour-coded map which can be used
costs. This has the potential to render these sectors to identify plant stress, greenness, vegetation density
uneconomic, unless labour costs can be contained and a number of other crop or pasture characteristics.
or reduced. Robotics creates the potential for
this to occur, although technology is still at the Improved cameras, including those with
developmental stage (Figure 10). hyperspectral capacity, are rapidly increasing the
amount of information that can be obtained using
An important difference between the digital these systems. The additional information is being
information systems that have been developed for used in systems that are being developed for a
the horticulture (and intensive livestock sectors) number of different applications, including crop
and the systems that are evolving for broadacre disease diagnostics, the analysis of vegetation
farming is that all the digital information generated coverage, and soil moisture status.
in the more intensive sectors tends to be generated
and used at the farm level, rather than obtained The GRDC is currently undertaking a project that is
from public sources or distributed via industry-wide evaluating and testing the use of UAVs in broadacre
information systems that give rise to privacy and agriculture businesses in northern NSW. One of
data ownership issues. the major challenges is ensuring that systems are
available that have the processing capacity and
Unmanned aerial vehicles wireless transmission reach required to capture the
Unmanned aerial vehicles (UAVs) or drones data generated from crop paddocks of the size that
are being used for different applications in are common under Australian conditions, which are
cropping and extensive livestock enterprises. many times larger than those that are common in
These include crop and pasture assessment, the US corn belt.
Figure 11: P
hotogrammetry imagery and a
composite image obtained using a
UAV.
Source: Drone Deploy (2015).
An alternative is the use of high speed internet

connections that can be used to transmit data from
the drone to a cloud storage and processing facility,
although the poor quality of high speed internet
access in rural Australia makes this option largely
unfeasible. Even the forthcoming roll out of the
NBN satellite network is not expected to assist
in this regard, as it will only be suitable for fixed
internet access.
Currently, high resolution satellite imagery is

considered cheaper and easier to use than UAVs to
obtain imagery for assessing biomass and fodder
quantity and quality. Nevertheless, in the future cost
effective uses for UAVs are likely to be found for
remote pastoral stations as the technology develops Figure 12: Variable rate map constructed from
and the costs reduce. Satamap data for residual herbicide
application.
Satellite imagery Source: Boughton (2016).
Satellite imagery is expected to remain an

important, if not the dominant form of remote For some applications, the need for ground truthing
imaging for most farm enterprises utilising remote is not as critical. For example, mapping the relative
sensing of land condition as a management tool. It measures of biomass across a crop paddock can
is much cheaper than UAV or other technologies, provide the basis for a variable rate application of
but generally requires ground truthing or UAV herbicide to control a specific problem species. A
calibration in order to verify and quantify crop commercial imagery service provider, Satamap, has
attributes. The potential exists to calibrate a field been established to sell satellite derived biomass
against a reference field of similar characteristics maps of paddocks to farmers (see Figure 12). The
crop type, soil, weather etc. However, ground farmer is able to use these maps, in combination
truthing of the original field is generally required to with knowledge of the growth cycle of particular
confirm whether factors such as weeds, disease and weed species, to create a variable rate map for use
plant stress are distorting sensory data. when spraying weeds. The result can achieve quite
important savings in herbicide costs, and a reduced nutrient status, climate and livestock data creates
future risk of herbicide resistance (Boughton 2016). the potential for integrated management software
platforms for extensive livestock production that
are similar to those currently available for cropping,
3.5 Productivity implications with the potential for significant productivity gains
The available evidence arising from the use of even from available knowledge and technology.
digital technologies and information systems in the
agriculture sector is that, while in many instances The dairy industry is probably the most advanced
the systems are still undergoing development and in this regard at present, with automated milking
have not reached their full potential, most sectors of and cow management systems currently operating
agriculture have the potential to secure productivity on a proportion of dairy farms and likely to expand
gains from the adoption of these systems. in the future. The ability to monitor and manage
pastures based on digital information is not yet
The observed gains in the broadacre cropping sector technically feasible, but once this is possible it
are of the order of 1015%, with approximately will constitute a major advance and open up the
half of these achieved through yield improvements, potential for important productivity gains.
and half through input cost savings. There is
a dilemma associated with these gains in that Further into the future the potential arises for the
they are generally not able to be obtained until development of autonomous farm equipment that
comprehensive, multi-year datasets are available, could carry out most routine management activities
and farmers obtain little value from the data during including cultivation, spraying, seeding, fertilising,
this initial phase. harvesting, mustering and drafting. Autonomous
tractors are currently being trialled for a number
The scale of potential gains available to the of different crops, and the cotton industry has
broadacre livestock sector are more varied, and already introduced harvesters that constantly
specific to particular production systems and stream machine and production data to cloud-based
geographic locations. There are already major storage facilities, and which can subsequently be
cost savings being achieved in the pastoral region used to guide different management decisions. The
through the use of remote water monitoring step from current systems to fully autonomous, or
systems, and walk-over drafting systems have been remotely-controlled machinery operations is not
demonstrated to create the potential for a 10% large, and already has a precedent in the mining
reduction in mustering costs, at a minimum. The industry.
potential productivity gains from the use of these
technologies in higher rainfall zones is likely to be The ability to change broadacre farm management
less as stock monitoring and mustering costs are from paddock or herd averages to management at
relatively minor in comparison to those incurred in the square metre or individual animal level brings
the pastoral regions. the promise of important productivity gains, even
with the use of existing production knowledge.
The ability to monitor pasture growth and Digital information systems are a critical part of
availability remotely and objectively in either that transition.
pastoral or high rainfall production zones has the
potential to generate important productivity gains Technologies are generally available to enable
for the broadacre livestock industries, as this would these changes to occur, or are feasible given
facilitate better pasture and grazing management. existing technologies operating in agriculture or
Systems to enable this to occur are still in their other sectors. This transition will bring with it
early development phases, and may be a decade or the need for a new generation of technologically
more away from becoming commercially available. savvy agricultural managers, an agricultural
Incorporating digital information about pasture service sector with new knowledge and skills, and
quality and availability with soil moisture, soil telecommunications and related infrastructure to
enable these systems to operate throughout rural must provide information across the value chain to
Australia. ensure traceability and compliance with accepted
aquaculture and wild catch specifications.
3.6 Compliance and supply chain ALDI recently partnered with Sedex (Supplier
implications Ethical Data Exchange) which operates a platform
Digital information systems are already widely through which suppliers can opt to share data
used within supply chains after the farm gate. pertaining to ethical practices in labour standards,
The development of on-farm digital systems that health and safety, environment and business ethics
are integrated with post-farm systems provide with customers (Sedex 2015).
opportunities for cost savings.
The expected benefits include improved supply
In post-farm supply chains, pre-shipment inspection chain risk management as well as reputational
data, quality analysis of products and supply chain benefits. It is expected that consumer preference
operational data are already well developed and can for sustainable practices and efficiency gains will
be expected to become increasingly sophisticated. push suppliers and vendors alike into stricter supply
The greater control afforded post-farmgate has chain regulations and more transparency.
already had a discernible impact on wastage.
Digital agriculture and data analysis is also being
A good example of this is the post-farm supply employed by the Queensland Government to
chain for horticultural exports. Digitised pre- reduce nutrient run-off into the Great Barrier Reef
shipment inspection data has, in specific cases, Marine Park from sugarcane operations. A crucial
largely eradicated the rejection of horticultural component of this initiative is encouraging farmers
produce shipments into China, to the extent that the to comply with Smartcane BMP (Best Management
market is rapidly incorporating such inspection data Practices). The aim of this approach is to showcase
into its standard operating procedures for exporting productivity and input efficiency delivered through
fruit. This includes a chemical analysis of fruit and Smartcane BMP to incentivise uptake by growers.
assessment of physical qualities such as firmness By implementing these practices farmers reduce
maturity, temperature, weight etc against vendor nutrient run-off and associated damage to the reef.
requirements. The success of the pre-shipment
Currently, compliance with the BMP program is
assessment and data collection is such that many
voluntary. However, recent announcements made
insurers of Australian horticultural exports now
by Canegrowers Australia suggest that unaccredited
require the grower to provide this pre-shipment
growers will face increased scrutiny by the
data as a condition of their insurance cover (Hortus
Queensland Government (Sparkes 2015).
Technical Services 2016).
Smartcane BMP ensures that farmers comply with
It is expected that insurers may soon require the
existing regulations. The Queensland Government
full life cycle of production input data. Irrespective
summarises the obligations for cane farmers in the
of contemporary requisites, the insurance
Wet Tropics, Burdekin and Mackay-Whitsundays
market would be expected to move toward data
regions as follows:
transparency as insurer risk is alleviated.
Supermarkets are also beginning to build supply

keep records of their use of fertilisers and agricultural
chemicals
chain monitoring capabilities into their business
models. ALDI enforces its own specifications for undertake soil tests
sustainable farming which effectively requires use the results of soil tests, and the regulated method,
input data from producers. The seafood supply to calculate nutrient requirements and apply no more
chain control is particularly notable. Suppliers than the optimum amount of fertiliser (nitrogen and
phosphorus)
follow product label instructions when using requirements from soil tests, benchmarks, analyses
agricultural chemicals such as herbicides and nutrient trend data against climatic events as well
insecticides as functioning as a management tool for farmers to
follow specific controls when using herbicide products record and monitor costs (Schroeder et al. 2014).
containing atrazine, ametryn, hexazinone and diuron
(including prescribed user training qualifications, Data and information obtained under the
spray-droplet size restrictions, no-spray windows,
BMP program will be used to refine nutrient
and restrictions on use prior to rainfall and near
waterbodies). (Queensland Government 2016) prescriptions, farmer reporting and ultimately the
direction of the project. It is hoped that ongoing
The program employs NutriCalc, an online reporting and collection of data will deliver better
nutrient management tool incorporated into the outcomes and become embedded in farming
BMP program. The tool itself calculates nutrient operations (Canegrowers Australia 2013).
27
4. The market for farm data storage

X.0 XXXXX
and analysis
As is the case in any other economic sector where and chemicals, and decisions about the amount of
digital information and management systems have fertiliser that would be applied, based on soil test
developed, the emergence of digital agriculture results. The software systems essentially facilitated
has brought with it a support industry providing a the decision-making processes, automated the
variety of services. These services include software ordering of required inputs, and also provided
development, data storage and retrieval services, access to technical information about specific inputs.
data integration services, digital information system
development and implementation, data analytics As variable rate seeding and fertiliser applications
services and related training and advisory services. became more common, digital control systems
The digital agriculture service industry is in its were developed, some of which were specific to a
infancy in Australia and its structure and likely particular machinery brand, and others which were
development is as yet uncertain, although the more generic and could be used with a number
history and development of the digital agriculture of different brands. These could be used to define
service sector in the US provides some indications different zones within a field, and to vary seeding
of likely developments in Australia. and fertiliser applications rates for each zone. At
the same time, the capacity of harvesters to record
The initial impetus for the development of digital and map crop yields was further developed, and
information systems in the cropping sector in machinery manufacturers developed proprietary
the US in the late 1980s was the development of digital information systems, which meant that
variable rate fertiliser machinery. These initially yield maps produced by one harvester company
relied on manually-prepared maps of soil conditions were incompatible with those produced by another,
at a sub-field level based on gridded soil testing. and each required different software systems.
As GPS became available in the early 1990s they Machinery manufacturers essentially developed
were adopted for both variable rate fertiliser and digital information systems initially as a loyalty
seeding equipment, and also in harvesters to create service, which had the objective of locking
yield maps. As the volume of data and information farmers into one specific brand of machinery.
increased, software platforms were developed to
help farmers and their advisors manage and store it. A number of factors have changed this situation
over the past five years. Farmers and their advisors
In the US, initial developments in the provision of were unhappy being locked into a single machinery
digital services for crop farmers involved either brand in order to make use of digital information
input suppliers (seed and chemical retailers), or generated on their farm. They commonly buy one
software companies which developed products brand of harvester, a different brand of tractor, and
that removed some of the administration and a different brand of seeder. US corn farmers also
paperwork from the services provided by these typically have external contractors apply in-crop
companies. A typical software application enabled fertilisers. Proprietary digital information systems
a retail crop advisor to visit a farm and to map imposed major limitations on the use of digital
and plan a cropping program with a farmer client. information for different farm operations.
This involved the selection and mapping of fields
to be cropped, the selection of seed varieties to be A competitive market also developed in the
purchased and used, the selection of soil treatments provision of software platforms, with independent
28 CHAPTER 4: The market for farm data storage and analysis
providers developing comprehensive systems

which incorporated a variety of mapping and other
functions that extended well beyond machinery
operations. It became an increasing challenge
for machinery companies to incorporate all the
functionality that farmers were demanding into
theirsystems.
Technological developments also created

opportunities for new players in the digital
agriculture space. For example, a US company
called Farmobile has developed a passive uplink
connection (PUC). It is designed to be a neutral
data pipe that collects and centralises data from Figure 13: The farm data ecosystem.
multiple farm machinery brands and models using Source: Open Agriculture Data Alliance (2014).
the ISO11783 communication protocol. Data can be
shared between application programming interfaces
(APIs) and made accessible to agronomists or can then choose to allow an agronomist or any other
for analysis at the farmers discretion (Farmobile outside party access to that data.
2015).
They believe the data should belong to the farmer
The development of a competitive market for digital to control as they see fit, and that by having open
agriculture software platforms and applications access arrangements and development tools, this
has been further enhanced by initiatives such as will provide farmers a choice in who they can
the Open Agriculture Data Alliance (OADA). The usefully grant access to their data. They also believe
Open Agriculture Data Alliance is an initiative by a that there should be arrangements associated with
group of researchers based at Purdue University in levels of trust which enable farmers to decide who
Indiana, who are promoting the concept of making has access to what level of data.
all agricultural data open source so that it can be
OADA is making some progress with major
more easily integrated, and so that a competitive
machinery manufacturers and agrichemical
market can develop for data application software
companies, who have committed to participate in
and smartphone apps.
the alliance and make data available in open source
The motivation for OADA came from the format (Open Agriculture Data Alliance 2014).
frustration felt by farmers who had multiple sources The open data concept is based on the premise
of data from a range of different applications and that the data produced by each different machine/
were unable to integrate that data into a single technology is able to be accessed by software tools
useful format that could be used for different and converted from one form to another not
purposes. The objective is to provide open source necessarily that they are in a standard format. Some
software from competitive suppliers that can be challenges arise because conversions from one data
used by farm service providers for a range of format to another are not usually loss-less and can
different applications. result in information being lost in thetranslation.
OADA is developing secure data exchange Suppliers have begun to build interoperability into
protocols through APIs and developer libraries. their business models. For example, SST Software
It allows for datasets uploaded from various provides widely used software for agronomists and
on-farm sources to communicate, synchronise and retailers based on a range of data types including
be accessed within the farmers chosen software yield data from seed suppliers across industry,
platform or data storage facility (cloud). The farmer weather and soils data. Data utilised by the SST
CHAPTER 4: The market for farm data storage and analysis 29
Figure 14: A map of the interoperability network that has developed amongst digital agriculture systems
and platforms in the US.
Source: SST Software (2015).
Software is stored on a central hub referred the majority of users would be agronomists and
to as the agX platform. The platform involves retailers capable of building a tractor-ready product
standardised data protocols, a store of competing (for example a variable rate planting prescription)
software applications (APIs), data storage and using data sourced from multiple service providers.
retrieval capacity, and permission-based data
sharing capacity to enable data to be exchanged Naturally, open standards facilitate competition
between farmers, their advisors, and machinery and downstream of equipment manufactures and data
equipment suppliers (agX 2015). collection products. Without transferable datasets,
companies that control the initial creation could
From the perspective of software companies like potentially monopolise downstream data services.
SST Software, interoperability achieved through Encouragingly, a number of large agribusiness
either APIs or a standardised central repository suppliers including the Climate Corporation are
provides a pan-industry data repository with embracing open standards. Larger data silos are
potential to partner with complementary service inherently more valuable where data aggregation
providers. At the time of writing this report agX isrequired.
is building a central store of agX-compliant
applications and services. The hope is that It is the belief of many within the industry that
such a platform will facilitate cross company data storage will gravitate toward either open or
collaboration. For example, modules owned and universal standards, essentially reducing the number
operated by water modelling companies would of data silos to one. It remains to be seen whether
generally be complementary with nutrient modules this will be an effective strategy.
of different companies. These in turn could be
utilised by or combined by a third party with With sufficient interoperability, datasets originating
weather and soil data. Functional interoperability from different sources can be aggregated and used
would allow smaller companies to compete in niche for benchmarking and research purposes at the
corners of an integrated analytics package delivered behest of the farmers. The issue of ownership is
to a grower (agX 2015). discussed further in Chapter 5. It is generally agreed
that farmers or their contractors own the data
On evidence from interviews with farmers and generated on-farm and will be assumed as such for
industry professionals it is unrealistic to expect the time being. Nevertheless, issues of privacy and
most farmers to spend time selecting service appropriating value need to be addressed to ensure
providers piecemeal as part of wider application easy proliferation of data to key stakeholders.
of big data. Rather, it would be expected that
Open source philosophy for data diminishes the governments, agricultural technology providers and
opportunity for profiting from basic storage and researchers, but as yet have not been able to realise
retrieval services. agX anticipates that compliant any direct revenue from those seeking access to
data collectors may be able to sell their data on thedata.
the platform to those wishing to incorporate wider
datasets. It is unknown how large the market for The most direct way to appropriate value is simply
third-party data will become. It is expected that for users to pay farmers (data owners) for access
most of the value will be derived from downstream to their data files. One avenue suggested would be
products including algorithms underpinning the a cooperative agreement where third parties would
analytics and farm management products. purchase a collective set of farm data or pay an
amount for a specific set of farm data provided by a
If effect, what has emerged in the US is actually data cooperative. While attractive from a farmers
three interconnected markets. The first is the market perspective, support for this arrangement appears to
for machinery control and monitoring systems, in be diminishing. Data markets are moving toward a
which the participants are the major machinery model where data warehousing is open source and
manufacturers, plus independent suppliers such the value for farmers is derived from algorithms and
as Trimble, Raven and AgLeader. The second is associated downstream products.
the market for data storage and retrieval services,
in which the participants are major equipment This doesnt necessarily preclude the emergence
suppliers such as John Deere with the myJohnDeere of a data cooperative to warehouse data. As the
platform, the agX platform supported by SST owners of data, farmers could choose to share
Software, and a large number of other cloud storage data only with the cooperative which would then
and retrieval services offered by major companies grant third parties access to the master dataset
such as Amazon and Apple, or smaller regional of farm data. However, the ability to control
farmer cooperatives or crop advisory groups. The data would be contingent on farmers not sharing
third market is the market for APIs computer similar data with other storage platforms. At the
programs and applications on smartphones or tablet same time withholding data from an agricultural
computers that can be used to retrieve, synchronise, technology provider (ATP) may reduce some of the
coordinate and utilise data for the purposes of farm functionality of the ATP equipment or services as
operations or as a guide to decision-making. per licensing agreements (see Section 5.3 below).
It is almost certain that most individual farm data
The adoption of open source data or open access would be provided to ATPs. The exclusivity of such
data platforms has facilitated the development of a cooperative dataset relies on a large proportion of
these markets, and has also enabled suppliers to farmers not allowing the ATPs to share aggregated
generate revenue from value-adding services, rather data outside the cooperative arrangement. It is not
than simply providing them as an extra service to certain that this could be effectively prevented.
secure customer loyalty. Even if it was possible to establish a fence around
an industry-wide dataset, it is unlikely to be
Appropriating value to the farmer desirable. Open source data removes financial
barriers to entry and thereby facilitates competition
While the emergence of competitive service
between downstream software providers delivering
providers has facilitated the rapid development of a
products and services to farmers, and from the use
user-pays digital agriculture service in the US corn
of which farmers ultimately benefit.
industry, which farmers are now paying to access,
one of the most common issues raised by farmers is The simplest and seemingly most likely scenario
whether the data they generate has the potential to is farmers forgoing monetary payment for data and
deliver any value beyond its use for farm decision- indirectly receiving some of the benefits of industry
making. Farmers have been told for some time that research/services. This arrangement would have
their farm data is valuable and potentially useful to the advantage of encouraging research as the cost
of the datas value would be invested by farmers It would be expected that historical data would
who would expect to share in the benefits. Benefits be bundled with the physical real estate. From the
could take the form of research dissemination, perspective of the seller, transferring electronic
additional services or simply improved data quality records to the buyer would require little effort or
underpinning service providers. Competition within expenditure. Provided the data contained no trade
the private sector would also be crucial to driving secrets or sensitive information, a competitive
product improvement for the farmer. The quality of marketplace should see the price of data approach
services would be expected to increase iteratively as zero. In a more likely scenario with limited sellers,
providers seek new customers. particularly within some enterprise categories, there
is the potential to put a price on data transfer, which
Reflecting this approach, the Sense-T project would be an amount less than the expected cost of
operating in Tasmania has been established on acquiring that data from later operations.
the premise that we all benefit by sharing data,
so long as privacy is protected. For example, the The Australian market
Sense-T Water Management program uses sensors
to improve water management in the South Esk and The relative immaturity of the Australian digital
Ringarooma catchments. It hopes that: agriculture service market makes it difficult to
anticipate how this market may develop, although
[B]y providing irrigators with real-time information there are a number of factors that dictate probable
about river flows, weather and water quality, better developments.
decisions about water management can be made for the
benefit of farmers, regulators and the environment.
Australian farmers use farm machinery that has
(University of Tasmania 2015) largely been designed and developed overseas, and
it is highly unlikely that overseas manufacturers will
From interviews with farmers in Australia and
develop unique systems for the Australian market,
industry representatives in the US, it is apparent
given its relatively small size. Consequently, the
that attitudes toward data sharing vary significantly
machinery-related systems and products that will
among farmers. A large portion are content to
be available for use by Australian farmers are
share with industry stakeholders if they can see
likely to be the same as, or only marginally adapted
wider benefit. Many more seemed relaxed about
from existing products and services currently
privacy and trade secret issues but expressed a
being utilised by US farmers. This means that it is
wish to receive payment or at least some form of
likely that the same open source and open-access
return from provision of their data. A small portion
protocols will apply, creating the opportunity for
expressed extreme reluctance toward giving up
a competitive market for farm data storage and
their data.
software applications, and making it likely that
interoperability between different systems and
Purchases of properties with historical data
platforms will be the norm.
Historical data for a property has a much greater
value for the purchaser of the property compared to There is also a strong likelihood that software and
data scientists looking to aggregate the data for the applications that have been developed in the US
purposes of product improvement. Data can provide market by organisations that are independent of
the property buyer with immediate knowledge the major machinery firms will also be adapted
pertaining to farm characteristics and guidance for the Australian market, and made available to
on optimal practice. The same dataset amongst farmers and advisors. Current examples include
many may not provide a large benefit to ATPs and Agworld and SST Software (see Figure 15, over
software developers. Value in this case is defined by page), both of which have already established
use rather than the data itself. Chapter 5 discusses Australian operations and have products that are
methods for quarantining usage of data and by available for a range of different uses associated
extension its value. with both cropping and livestock production.
Figure 15: Australian and New Zealand

croplands mapped using SST
Software.
Source: SST Software (2015).
SST Software products are currently used by

Australian crop advisors employed by several large
companies supplying inputs and advisory services Figure 16: A
n example of the 1:25000 soil maps
to the Australian grains industry. and data accessible at the individual
field level from the USDA.
Independent suppliers of machinery consoles, Source: USDA (2015).
such as Trimble and Raven, also have a presence
in Australia, and are used in conjunction with a soil nutrient analysis in the US costs of the order
number of different machinery brands. of US$810 per sample, compared to costs in
Australia of between $60 and $100 per sample.)
What is less certain about the future Australian
digital agriculture market is whether it is large In the absence of this detailed information, the
enough, and has the underlying infrastructure and process of developing zone maps for VRA seeders
public data stores that have been critical to the or spreaders in Australia will depend on gridded soil
development of this market in the US. It is very test data at a level of intensity that is far lower than
obvious in talking to those involved in the digital that available in the US or farmers or crop advisors
agriculture market in the US that the availability knowledge about the specific soil types present in
of very detailed and extensive climatic data from a a particular paddock. Available climatic data may
dense weather radar and reporting station network, have to be extrapolated from the nearest weather
in combination with detailed soil maps and datasets stations which may be up to 100 kilometres distant
published by the United States Department of from the specific paddock. This set of circumstances
Agriculture (USDA), plus a range of publicly- clearly imposes limitations on the potential utility of
available GPS and cadastral data services has digital agricultural systems, unless initiatives can be
facilitated the development of the digital agriculture implemented to overcome some of these issues.
service sector in the US (see Figure 16).
While it is tempting to think that the Australian
Digital agriculture in the US is also reliant on cropping sector might simply be able to ride on the
very intensive soil testing (at rates of samples back of digital agricultural developments that have
per acre, rather than acres per sample as is the already occurred in the US, there are some very
case in Australia) and relatively inexpensive soil significant limitations in Australia that mean that
testing costs compared to Australia. (A standard this may not be the case.
The most advanced cropping products tend to be generated inside the farm gate, and do not need
targeted at US corn and soybean produces. This to access public soil and climatic data. The inputs
reflects market scale and higher revenue per hectare utilised in intensive livestock production systems
in comparison to other markets such as Australia. are also largely controlled and measured, which
Nevertheless, Climate Corporation has flagged greatly reduces the complexity of data needs and
expansion of its products into wheat and canola analysis. Many of these systems are already in
and is launching platforms in Canada and Brazil use in Australia, or the Australian businesses have
(Stern 2015). Modifying products for other crops developed proprietary digital information systems
is non-trivial. For example, soil nitrogen levels which have been used for some time.
depend partially on last years crop residue. Climate
Corporation uses extensive data on nitrogen and There has not been the same degree of development
carbon retention of corn and soybean residue to of digital information systems for extensive
estimate soil nitrogen levels (Climate Corporation livestock production systems. Those systems
2015). For commodities more common in Australia, that are available tend to operate as an electronic
there may not be comparable repositories of data paddock diary, enabling users to record changes
and field tests to underpin nutrient models. Higher in livestock numbers, grazing arrangements and
costs per soil test and other data inputs further animal weights. These systems have not been
erode potential profitability within Australia. developed to the point where they are used to
Nevertheless, digital agriculture developments in determine the need for specific management
the US represent at the very least, proof of concept changes, or to project the outcome of particular
if not always transferable business models. management decisions. The National Livestock
Identification System (NLIS) in Australia provides
In the case of the livestock industries, there have an important base data infrastructure that may play
been digital systems developed in the US for the a facilitative role in the development of digital
intensive sectors such as pork, poultry, dairy and information systems in the extensive livestock
beef feedlots. These systems utilise data largely industries in the future.
34
5. Agricultural data and privacy
For many farmers, while the potential opportunity provides the following overview on the application
to achieve productivity gains through the adoption of confidentiality agreements:
of digital agriculture is attractive, there are nagging
concerns about making available detailed farm Confidentiality agreements can be made with anyone
(employees, business partners, business associates,
production and other data on electronic platforms research academics and so on) whom you wish to
that are accessible to unknown third parties, and impose an obligation of confidence on, regarding the use
which may be used for purposes over which the and disclosure of your confidential information.
original provider of the data has no control. The (IP Australia 2013)
following section of the report analyses these issues
by examining the current legal framework, by Farmers or data owners who feel data contains
considering some relevant international models, information about the farm business that they do not
and by examining a number of case studies. wish to disclose can request it be kept confidential
via confidentiality clauses in a contract. Even in the
event there has not been a confidentiality agreement
5.1 Legal framework relevant negotiated, under Australian law there is a default
to agricultural data obligation to maintain confidentiality.
There are potentially three areas of Australian law
The Australian Law Reform Commission Report
that may be relevant to the providers and holders
108 provides an explanation for Common Law
of digital information generated from Australian
and equitable duties of confidence. In a section
farms. These are the legal frameworks associated
detailing obligations of confidence, it describes
with:
circumstances where confidentiality obligations
1. Confidentiality and trade secrets may arise through equity and without a prior
contractual agreement.
2. The Australian Privacy Act
15.126 A contractual obligation of confidence can
3. the guidelines and legislation under which arise from express terms in a contract, but also by
Australian statistical agencies operate. implication. The nature of the obligation will depend on
the terms of the contract. Remedies for threatened and
actual breach of the contractual obligations to maintain
Confidentiality confidence include injunctions and damages.
The person who has generated data and who is
15.127 An equitable obligation of confidence can arise
the owner of that data (in this instance generally where the formalities for the formation of a contract are
assumed to be the farmer) can legally apply not present. The obligation arises where information
conditions of use and disclosure on those who are with the necessary quality of confidence is imparted in
allowed to access the data. This is in the form of circumstances importing an obligation of confidence.
Such circumstances will exist where the information is
a legally-binding contract, which is enforceable imparted on the understanding that it is to be treated by
through the courts should that contract be breached. the confidant on a limited basis, or where the confidant
This applies in general to information owned by an ought to have realised that in all the circumstances the
individual, and this arrangement is generally given information was to be treated in such a way. Breach of
the obligation occurs where there is an unauthorised use,
effect via a confidentiality agreement. IP Australia
CHAPTER 5: Agricultural data and privacy 35
not only where there is unauthorised disclosure, of the of the product or the process of manufacturing the
information. product when reverse engineering is unlikely.
15.128 Unlike the position in contract, where loss is The best known example of a trade secret is that of the
the basis of a claim for damages, the plaintiff in a suit Coca-Cola recipe. The company has used trade secrets to
for breach of the equitable obligation does not need to keep its formula from becoming public over a period of
show any damage. Remedies for breach of the equitable decades. It never applied for patent protection, so it was
obligation include compensation or an account of profits, never required to disclose the formula. One disadvantage
an injunction and a declaration. is that trade secrets do not provide any legal security
(Australian Law Reform Commission 2008) against an independent competitor inventing an identical
object. (IP Australia 2013)
Crucially, obligations of confidence apply to Data collected on-farm would assist in developing
both the unauthorised use as well as disclosure of improved processes, some of which may already be
information. implemented by individual farmers and considered
a trade secret. For example, a farmer who believes
This means that in the absence of an agreement
they have identified an optimal summer legume
explicitly waiving the rights of the owner of
for their particular location through iterative
farm data to confidentiality, there is a default
seasonal experimentation may not wish to have this
confidentiality obligation on the provider of
information made available to other farmers in the
the software platform or data storage service to
same region who have not expended the sameeffort.
maintain the confidentiality of the data, and to not
make it available to a third party or use it for any This information might be considered a trade
purposes which the data owner does not agree with. secret and the farmer could therefore seek to have
it remain confidential. However, whether farm
In effect, however, all users of farm software
information constitutes and remains a trade secret
platforms and data storage services (and any
will be influenced by factors such as:
computer software and cloud storage services)
generally sign a user agreement which waives some the extent to which the information is known by
or all of the confidentiality requirements associated others
with the data that is made available by the user. the extent of measures taken by the farmer to
guard the secrecy of the information
Trade secrets
the value of the information to the farmer and
A related area of the law that may be relevant to potential competitors
privacy issues associated with farm data is the
legal framework for the protection of what are the amount of effort or money expended by the
termed trade secrets. There are four basic forms of farmer in developing the information
intellectual property (IP) that are recognised under the ease or difficulty with which the information
law. These are patents, copyrights, trademarks and could properly be acquired or duplicated by
trade secrets. Of the four, trade secrets is the only others.
form that could be relevant to farm data (Janzen
2015a). It is probably reasonable to conclude that, except
in very specific situations, it is highly unlikely
IP Australia describes trade secrets as follows: that farm data would be considered to constitute a
trade secret, and would therefore automatically be
A trade secret is both a type of IP and a strategy for
protecting your IP. It can provide effective protection for subject to protection and confidentiality provisions
some technologies, proprietary knowledge (know-how), on that basis. It is also worth noting that even in
confidential information and other forms of IP. limited situations where this might be considered to
A trade secret is appropriate when its difficult to copy a
apply, it would be likely that this protection would
product. This may include the construction or formulation diminish over time if the information or the practice
became more widely known or adopted.
36 CHAPTER 5: Agricultural data and privacy
Australian Privacy Act Statistical agencies

A second area of law that may be considered to be A third area of law that may have some relevance
applicable in relation to the protection of farm data in relation to farm data that is provided to service
is the Australian Privacy Act 1988 (Privacy Act). providers is the law that governs the collection and
use of information by the Australian Government
The Privacy Act regulates how personal information for statistical purposes. Conceivably, in the event
is handled. The Privacy Act defines personal that digital agricultural applications became widely
information as, information or an opinion, whether used by farmers, the Australian Bureau of Statistics
true or not, and whether recorded in a material (ABS) might decide to require digital agriculture
form or not, about an identified individual, or an service providers to release information to the
individual who is reasonably identifiable. ABS that could subsequently be used to produce
official statistics. For example, in the event that
Common examples are an individuals name, digital agricultural applications were widely used
signature, address, telephone number, date of by crop farmers, the information held by those
birth, medical records, bank account details service providers detailing the area planted to crops
and commentary or opinion about a person by farmers could provide a very useful source
(Commonwealth of Australia 2015a). of information about changes in national crop
production.
The Privacy Act incorporates a series of legally
binding principles Australian Privacy Principles The Census and Statistics Act 1905 governs
(APPs) which regulate the handling of personal the collection national statistics. Section 12
information by the Australian Government, the ACT paragraph(2) of the Act is of particular relevance
Government and private sector organisations. The to data privacy. The paragraph dictates that the
Privacy Act requires that persons whose personal release of information which can be reasonably
information is being collected have a right to: used to identify an individual determine person
or organisation is prohibited, The results shall
know why personal information is being not be published or disseminated in a manner that
collected, how it will be used and who it will be is likely to enable the identification of a particular
disclosed to person or organisation. (Commonwealth of
have the option of not being identified Australia 2006)
ask for access to their personal information Concerns about privacy in relation to national
ask to stop receiving unwanted direct marketing statistics databases are not new, nor limited to
agricultural data. The ABS has collected data from
ask that personal information that is incorrect be households and business long before the era of big
corrected
data and routinely uses various measures to prevent
make a complaint about an entity covered by the identification of individuals. Additionally the
the Privacy Act, if it is considered that personal ABS has internal administrative arrangements
information has been mishandled. in place to ensure that employees cannot access
both the personal identifiers and the content of
The common view expressed in interviews with statistical information that it requires individuals or
lawyers was that the Privacy Act would have little businesses to provide. As the agency explains:
relevance to farm data, as most of the data involved
would not be considered to be personal information. The ABS separates identifying variables from content
variables as part of its suite of strategies to protect the
Nevertheless, for some data, individual farmers
identities of individuals and organisations in datasets.
owning the data may be able to assert protection This means that no-one can see the identifying or
under the Privacy Act. demographic information, used to identify which records
relate to the same person or organisation (eg name,
address, date of birth), in conjunction with the content
data (eg clinical information, benefit information, to the software providers (ATPs) to use the data
company profits). Instead, staff can see only the in many different ways, including via the sale or
information they need to do the linking or analysis. So,
rather than someone being able to see that John Smith transfer of the data to a third party.
has a rare medical condition, or the profits earned by
Company X, the person doing the linking sees only the The US and New Zealand agriculture sectors in
information needed to do the linking (eg John Smiths particular have taken steps to clarify data ownership
name and address) and the analyst just sees a record,
rights, via the introduction of voluntary industry
with no identifying information, showing that a person
has a rare medical condition together with any other standards. The aim of both standards is to establish
variables needed for analysis (eg broad age group, sex). a common understanding between users and service
(Australian Bureau of Statistics 2013) providers about data ownership and protection,
with the hope that this approach will prevent or
In conclusion, while the development of digital discourage misuse of farm data, while avoiding the
agriculture may provide government statistical heavy hand of regulation and its likely negative
agencies with new and administratively efficient impact on innovation.
opportunities to collect data for statistical purposes,
there is no obvious reason for farmers using digital Privacy and security principles
agriculture services to be concerned that they for farm data (US)
are exposed to additional risks in relation to the
misuse of their farm data by government statistical The US Farm Bureaus Privacy and Security
agencies. Principles for Farm Data are a set principles
to be upheld in contracts between compliant
organisations and farmers. The specific standards
5.2 International digital agriculture contained in the US principles are as follows:
standards Education: Grower education is valuable to ensure
Unsurprisingly, farmers in overseas locations where clarity between all parties and stakeholders. Grower
digital agricultural services are more developed organizations and industry should work to develop
programs, which help to create educated customers who
than is the case in Australia have raised concerns understand their rights and responsibilities. ATPs should
about data ownership rights and privacy, and strive to draft contracts using simple, easy to understand
service providers have taken steps to give greater language.
reassurance to users about the protection that is
Ownership: We believe farmers own information
available for their farm information. generated on their farming operations. However, it is the
responsibility of the farmer to agree upon data use and
Digital agriculture service providers typically sharing with the other stakeholders with an economic
address privacy and other issues arising from interest, such as the tenant, landowner, cooperative,
owner of the precision agriculture system hardware,
national legislation through written contracts, and/or ATP etc. The farmer contracting with the ATP is
which users are required to agree to before using responsible for ensuring that only the data they own or
the software. The requirement to agree to the have permission to use is included in the account with
contract or terms of use prior to using the software the ATP.
is not confined to digital agriculture applications, Collection, Access and Control: An ATPs collection,
and in fact is almost ubiquitous for any software access and use of farm data should be granted only with
or computer applications. In effect, most of these the affirmative and explicit consent of the farmer. This
contracts or terms of use require the user to will be by contract agreements, whether signed or digital.
acknowledge a set of conditions, which usually Notice: Farmers must be notified that their data is being
means that the user agrees to waive specific legal collected and about how the farm data will be disclosed
rights in return for being able to use the software. and used. This notice must be provided in an easily
located and readily accessible format.
Alternatively, the terms of use may provide
superficial reassurance about the privacy or use of Transparency and Consistency: ATPs shall notify
the information, but may also bury exclusions deep farmers about the purposes for which they collect and
in the document which in effect give free reign use farm data. They should provide information about
how farmers can contact the ATP with any inquiries data retention period. ATPs should document personally
or complaints, the types of third parties to which they identifiable data retention and availability policies and
disclose the data and the choices the ATP offers for disposal procedures, and specify requirements of data
limiting its use and disclosure. under policies and procedures.
An ATPs principles, policies and practices should Contract Termination: Farmers should be allowed to
be transparent and fully consistent with the terms and discontinue a service or halt the collection of data at
conditions in their legal contracts. An ATP will not any time subject to appropriate ongoing obligations.
change the customers contract without his or her Procedures for termination of services should be clearly
agreement. defined in the contract.
Choice: ATPs should explain the effects and abilities Unlawful or Anti-Competitive Activities: ATPs
of a farmers decision to opt in, opt out or disable the should not use the data for unlawful or anticompetitive
availability of services and features offered by the ATP. activities, such as the use of farm data by the ATP to
If multiple options are offered, farmers should be able to speculate in commodity markets.
choose some, all, or none of the options offered. ATPs
should provide farmers with a clear understanding of Liability & Security Safeguards: The ATP should
what services and features may or may not be enabled clearly define terms of liability. Farm data should be
when they make certain choices. protected with reasonable security safeguards against
risks such as loss or unauthorized access, destruction,
Portability: Within the context of the agreement and use, modification or disclosure. Polices for notification
retention policy, farmers should be able to retrieve and response in the event of a breach should be
their data for storage or use in other systems, with the established. (US Farm Bureau 2014)
exception of the data that has been made anonymous
or aggregated and is no longer specifically identifiable. This set of principles was agreed to and signed by
Non-anonymized or non-aggregated data should be easy 34 ATPs, including some of the largest corporations,
for farmers to receive their data back at their discretion.
in January, 2015.
Terms and Definitions: Farmers should know with
whom they are contracting if the ATP contract involves The principles are aimed at ensuring the terms and
sharing with third parties, partners, business partners, conditions which farmers sign up to are transparent,
ATP partners, or affiliates. ATPs should clearly explain
the following definitions in a consistent manner in all
that the data is owned by farmers, and that farmers
of their respective agreements: (1) farm data; (2) third are advised in the event that their data will be sold
party; (3) partner; (4) business partner; (5) ATP partners; to a third party. Farmers also retain the right to
(6) affiliate; (7) data account holder; (8) original prevent the sale of the data which they own. While
customer data. If these definitions are not used, ATPs
this is an understandable approach, in practice
should define each alternative term in the contract and
privacy policy. ATPs should strive to use clear language the nature of the notification and the response
for their terms, conditions and agreements. timeframe provided to farmers could undermine the
intent of this principle.
Disclosure, Use and Sale Limitation: An ATP will not
sell and/or disclose non-aggregated farm data to a third
party without first securing a legally binding commitment The extent to which these principles will curtail the
to be bound by the same terms and conditions as the ATP practice of burying exception clauses in detailed
has with the farmer. Farmers must be notified if such a contract agreements is unknown. The sheer volume
sale is going to take place and have the option to opt out of terms of use agreements for computers and
or have their data removed prior to that sale. An ATP will
not share or disclose original farm data with a third party software users generally, and the ambiguity about
in any manner that is inconsistent with the contract with what constitutes clear and user-friendly language
the farmer. If the agreement with the third party is not versus needlessly obfuscating legalese has allowed
the same as the agreement with the ATP, farmers must companies to hide exclusion clauses in a manner
be presented with the third partys terms for agreement
or rejection.
which notionally fulfils obligations of notification
and even transparency.
Data Retention and Availability: Each ATP should
provide for the removal, secure destruction and return The principle of data portability is an important
of original farm data from the farmers account upon
the request of the farmer or after a pre-agreed period of
feature of these principles. It should help to ensure
time. The ATP should include a requirement that farmers that farmers dont become tied to a particular
have access to the data that an ATP holds during that service provider because of the potential loss of
farm data that would occur in the event that a in the event that the data is sought by
decision was made to change to another provider. government, that measures are in place to render
data anonymous, and that the farmer is notified
New Zealand Farm Data Code of the request to access the information.
ofPractice
The Development of the New Zealand Farm Data
Code of Practice was funded by New Zealand
dairy farmers through DairyNZ, and also the New
Zealand Ministry for Primary Industries (MPI) and
FarmIQ, a farm management software company
in New Zealand. The development of the code
involved 60 industry and commercial organisations
operating in the agriculture sector of New Zealand,
as well as individuals involved in farming and the
provision of advisory services. Unlike the US code,
the main focus of the New Zealand code appears to
be on-farm data associated with livestock and dairy
production.
The code was developed around a set of principles,

which are as follows:
Compliance with the Code of Practice will be Figure 17: T

rademark available to be used by
voluntary. New Zealand companies compliant
The Code of Practice will offer visible with the New Zealand Code of
credibility for approved agencies. Practice.
Source: Dairy New Zealand (2014).
The Code of Practice will encourage open,
transparent communication and management The New Zealand Farm Data Code outlines a
of data on behalf of primary producers and process and an industry structure through which
endusers. the code is to be reviewed and amended, and
The Code of Practice will respect intellectual establishes a Code of Practice Authority, which
property rights and encourage innovation. has the task of regularly reviewing the code and
determining whether an organisation is compliant
The Code of Practice will raise awareness about with it. Organisations pay of cost of NZ$1400 when
the availability of data. they initially seek accreditation under the code,
and the right to display its logo in association with
The code incorporates some similar principles to their products. In subsequent years, their annual
those included in the US code. These include: accreditation renewal cost is NZ$900 (Dairy New
Zealand 2014).
that farmers are provided with a data access and
storage agreement by service providers A related and parallel development are the New
that service providers ensure that data can be Zealand Farm Data Standards, which are a set
accessed and disseminated at the discretion of of data standards that are being developed for
the farmer recording data associated with different aspects
of farm businesses. The objective is to promote
that the data is securely stored and protected data exchange and increase the opportunity for
innovation in the utilisation of farm data in New
Zealand. The data standards currently under Case study 1: Data generated on-farm
development include the following: about farm resources and operations
Animal Data Standard The data in question in this case study is farm
Land Application Data Standard productivity data generated by equipment used by
the farmer as part of normal operations or through
Stock Reconciliation Data Standard technology supplied and possibly installed by a
Grazing and Feed Data Standard third party with the consent of the farmer. In this
case the farmer/landowner is fully aware data is
Irrigation and Effluent Data Standard being generated.
Financial Data Standard
The requirement that ATPs would ensure this
Chart of Accounts Data Standard. data remained confidential is predicated on the
understanding that it is the farmer who owns
As the development process for these standards is
the data generated through farming operations,
still underway, it is not possible to comment on the
although of course a Conditions of Use
extent of industry support or likely benefits that
requirement by the ATP could be that the farmer
might arise from these New Zealand initiatives.
cedes any ownership rights over the data. The
position taken by both the US Farm Bureaus
5.3 Case studies of different data Privacy and Security Principles and the New
flow models Zealand Farm Data Code of Practice is that the
Conditions of Use should specify that the farmer
One of the challenges associated with the ever- retains ownership of the data and rights to control
growing flood of data being generated in associated its use, and it seems that ATPs in both jurisdictions
with farm production is that not all of the data is are prepared to reflect that in their conditions of use
generated by or under the control of the farmer, agreements with users, irrespective of what may be
and there is often a need to transfer farm data to the actual legal situation (Dairy New Zealand 2014;
external parties which may not necessarily have any US Farm Bureau 2014).
direct contact with the farmer or the farm business.
To gain a perspective of the different types of data This is perhaps a tacit recognition that irrespective
being generated by farm businesses and some of of the legal position, the market appeal of the
the implications of this when considering data services provided by each ATP is likely to be
confidentiality and privacy issues, three case-study affected by any loss of confidence amongst users
scenarios have been developed and are detailed about the confidentiality and security of farm data.
below.
While this case study appears relatively
The three data case studies have been developed straightforward, care should be taken when defining
based on the different flows of data ownership and the farmer. Ambiguity may arise when the farmer
access identified in recent research relevant to these is not the same as the landowner. Ideally ownership
issues (Janzen 2015b). The three case studies are as of any data generated should be specified in any
follows: contract between the landowner and a sharefarmer
or contractor, although it is probably reasonable
1. agronomic data generated on-farm about farm to presume that this is rarely the case at present in
resources and operations Australia.
2. machine data pertaining to performance of farm
machinery There is some uncertainty about the default position
that would apply in this situation. Australian trade
3. drone and remote sensing data collected secret legislation is thought likely to be interpreted
remotely (for example by satellite) by third to mean that whoever farms the paddock owns the
parties.
data, as they ultimately execute the processes and Contracts also need to make clear how the farmer
usually own the machinery that generates the data. can access and change their data and the provisions
Alternatively, it could be claimed that farming in place for opting-out of the agreement (Office of
processes are executed on behalf of a landowner the Australian Information Commisioner 2015).
and as such data are owned or at least jointly owned
by the landowner, particularly where a contract In the event of a legal case, it is understood that any
harvester is employed (Janzen 2015c). data collected on-farm (either digital or manual) is
susceptible to a request for information provided
As a default arrangement, the landowner should it is relevant to a specific case based on criminal
have the ultimate ownership control over the data law. While a person may attempt to claim a public
including the right to share the data at his or her interest immunity in such situations (on the basis
discretion. However, a form of agreement would that the release of the information would be against
be required with, for example a sowing contractor, the public or national interest) it is difficult to
agronomist or sharefarmer to grant limited access conceive of a case where public interest immunity
to the data while under contract and for that person would apply. Farmers should be made aware in
to be able to choose to grant access to third-party any Contract of Use provided by an ATP that
agronomists or other service providers (for example a subpoena on the farm data could be enforced
to generate a variable rate zone map) on the (New South Wales Young Lawyers Civil Litigation
condition that they themselves do not share data Committee 2010).
with individual property identifiers without the
consent of the landowner. Freedom of Information (FOI) requests may apply
where farm data is collected by a government
At the conclusion of the contract, the contractors agency. This could mean, for example, that an
access should be revoked and the landowner free ATP which agrees to make farm data that it holds
to employ a different contractor under similar available to a government agency for research
arrangements and with access to historical data purposes may, unwittingly, create a situation where
from prior contractors. that data could be made public as a consequence of
an FOI request to the relevant government agency.
Ultimately the philosophy should be to allow An exception to this situation would be in the case
unrestricted access to data for those involved in where the data was made available to a government
production while preserving privacy and ownership statistical agency, which usually has immunity
for the landowner to the greatest degree possible. from FOI requirements. In the event the data was
made available to a government agency other than
Once the farmer or owner is established, a statistics agency, it may be possible that the farm
conditions of use can be written into contracts. data would be judged to contain commercially
Two major avenues of privacy protection are sensitive information and could be considered
available to the farmer. Laws of confidentiality exempt from FOI requests.
should be incorporated into contracts which
prohibit information including data being shared Freedom of Information (FOI) request exemption
unnecessarily by the contractor. Confidentiality based on personal information disclosure may apply
agreements are expected to form the parameters in cases where the farm is both business and home.
of data exchange between owner, contractor In the US, which has similar FOI laws, the US Farm
and service providers. Use preferences or use Bureau is appealing against the Environmental
agreements directly negotiated in the contract are Protection Agencys public release to environmental
expected to govern use of data as it bounces from groups of personal details about the home locations
owner (farmer), service provides, input suppliers and contact information of tens of thousands of
and contractors. farm and ranch families (Rodgers & Thornton
2015).
Section 5.4 of the New Zealand Farm Data Code from making market-sensitive farm data available
of Practice provides general guidelines for ATPs to commodity speculators. The US Farm Bureaus
(see below) when faced with disclosure requests. Privacy and Security Principles for Farm Data
Broadly speaking, data collectors should endeavour has an Unlawful or Anti-competitive Activities
to keep data private unless there is a compelling section as follows, ATPs should not use the data
legal obligation, in which case the primary producer for unlawful or anticompetitive activities, such
is to be notified. as the use of farm data by the ATP to speculate in
commodity markets (US Farm Bureau 2014).
This section recognises that all organisations have forms
of legal compliance, but that some organisations have
John Deeres farm data policy includes a clause that
additional regulatory responsibilities. Where information
is required by law or regulation to be provided to holds a similar sentiment.
other parties (for instance, an Official Information Act
request), an organisation that complies with this Code John Deere will NOT use internally or share anonymized
of Practice shall: data to external parties who John Deere believes intend
to use it to influence markets, provide an advantage to
avoid disclosing information that identifies an commodity traders or support supply hedging by food
individual primary producer; or companies.
(Deere & Company 2015)
notify the primary producer if individually identifying
information must be disclosed.
(Dairy New Zealand 2014)
It remains to be seen whether this policy will be
effectively enforced. Alternatively there could be a
Unless the information is required by law, third mandated requirement that any release of such data
parties are expected to adhere to obligations of could only occur in the form of a controlled release
confidence whether written into contracts or arising of anonymised farm data to the entire market, a
through implication (equitable duty of confidence). requirement that is similar to that which applies to
Third parties must notify primary producers about market sensitive releases from listed organisations.
any mandatory information disclosure.
Liability arising from poor data
On-selling data to farm quality control
commodity traders The liability of ATPs in the case where incorrect or
It is not clear whether on-selling farm data (either faulty data has been generated is not clear. It seems
in its entirety or in an aggregated and anonymised likely that there are two potential causes of this type
form) to farm commodity traders would be legal, of problem, one being an equipment fault, and the
irrespective of contract notification or farmer other being an operator error.
approval. An example may be an ATP which has
available a large volume of yield data arising from In the first case, embedded equipment software
harvesters on many farms as the annual harvest may be corrupted and downstream analyses and
progresses. This information could be of strategic services derived from the data may result in losses
commercial value to farm commodity traders, for the farmer. It would be anticipated that in such
enabling them to take market positions before the a situation the manufacturer or software supplier
rest of the market was aware of that information. would be at fault, and liable to make good any
This might be judged to give farm commodity damage. A challenge for farmers in this situation
traders an unlawful or anticompetitive advantage, would be to prove that the error was the fault of the
particularly in markets with a large agricultural manufacturer or supplier.
futures exchange or markets where over the counter
In the second case where an operator error has
(OTC) swaps predominate such as Australia.
resulted in incorrect data (for example through
One solution is to prohibit ATPs from using the incorrect calibration) the cost of any loss would
farm data to speculate on commodity markets, or naturally reside with the machine operator or
software user, although in the case of a contractor
(such as a contract harvester) it is possible that the and disclosure of personal information (Office of
liability for resulting losses could reside with that the Australian Information Commisioner 2015):
contractor.
6.2 An (entity subject to Australian Privacy Principles
legislation) that holds personal information about an
Calibration errors have the potential to be
individual can only use or disclose the information for a
significant, especially in the case where data particular purpose for which it was collected (known as
derived from several different machines are being the primary purpose of collection), unless an exception
combined together. Discussions with industry applies. Where an exception applies the entity may use
personnel in the US revealed that yield sensor or disclose personal information for another purpose
(known as the secondary purpose). Exceptions include:
calibration errors of 10% have been identified
in harvesters caused by, amongst other things, The secondary use or disclosure of the personal
incorrect settings, dirty sensors or GPS errors. One information is required or authorised by or under an
Australian law or a court/tribunal order.
ATP has developed a quality assurance system to
standardise harvester calibration, and adjusts data The APP entity reasonably believes that the secondary
post-harvest by cross-referencing harvester data use or disclosure is reasonably necessary for one or more
with storage delivery data. enforcement related activities conducted by, or on behalf
of, an enforcement body (APP 6.2(e)).
Legal recourse against losses caused by (Office of the Australian Information Commisioner 2015)
downstream analytics service providers remains
unclear. With the prevalence of cloud storage and Regardless, it may be in the wider interest to
open application programming interfaces (APIs) include a clause in crop insurance contracts which
or even standardisation of data, the market will permits sharing of data between farmers and
be open to third-party service providers (Open insurers. It is likely that the market will demand
Agriculture Data Alliance 2014). It would be open access to investigators if crop insurance
expected that indemnity claims would mirror those premiums are set with a reduced risk of fraudulent
that have been made against traditional agronomy payouts.
service providers. However, to date it appears
that no claims have been made against third-party Case study 2: Data relating to the
agronomy software providers. performance of farm machinery
This case study refers to data automatically
Farmers would be expected to accept fault for
recorded about the performance and use of
inaccurate data if ATP proprietary software has
machinery on a farm. This data may include the
been modified. In any case, modifying licensed
engine running time, engine speed and temperature,
on-board software would generally be prohibited by
GPS location, and data detailing the performance of
copyright law (Janzen 2015d).
the engine plus related systems such as the steering,
hydraulics, gearbox and electrical systems. Recent
Fraud prevention
model farm machinery often has in-built capability
The role of ATPs in preventing fraud and supplying to transmit this data automatically to a cloud storage
farm data to fraud investigations will need to be facility or to a computer system owned by the
defined. While ATPs would generally be expected manufacturer.
to uphold privacy, they would be obligated to
provide information where crop insurance fraud and Data of this kind may be used by the machinery
other illicit activities are suspected. manufacturer for further product development and
telematics services for example to alert the owner
Even if a farmer who is suspected of criminal when a service is due or a fault is detected. The
activity asserted that their data was personal question of ownership is less obvious in the case
information, the ATP would still be obliged to of telemetric data. Manufacturers may lay claim to
disclose data in accordance with Chapter 6 of the machine data as a proprietary trade secret (Janzen
Australian Privacy Principles which dictate the use 2015b).
Irrespective of software agreements, it appears would also be expected to apply to machine data. If
reasonable that farmers should retain the right to the machine data did not carry personal identifiers
opt out of sharing machine data, as is currently it would very likely be susceptible to FOI requests
the case with John Deere machinery. Practical if the data were held by a government body
implementation of such arrangements, however, (Australian Government Solicitor 2009).
is not always feasible. Equipment manufactures
can embed licensed on-board software that Fraud prevention
requires upgrades and exclusive servicing by the Machine data may be treated in a similar way to
manufacturer and affiliates. Machinery owners can production data in its use for fraud prevention
be discouraged from opting out of data sharing, as provided it assisted detection. Manufacturers with
this may mean that software updates and upgrades access to machine data either through ownership
are not received. Some reduced functionality rights or purchaser permission may be obligated to
may be an unavoidable by-product of a desire to hand over data to enforcement authorities. As in the
maintain greater privacy, but some may also be case of production data, sharing of machinery data
artificially imposed by the manufacturer or dealer between farmers and insurers could be included
to reduce of minimise support costs, or to maximise within crop insurance contracts (Office of the
the volume of machine data available. Picking Australian Information Commisioner 2015).
apart which is which may not always be possible.
Uncertainty around machine data ownership is not

Case study 3: Drone and remote sensing
unique to agriculture (Pinsent Masons 2014). Tesla data collected by third parties
Motors Customer Privacy Policy adheres to the There are two possible scenarios relating to remote
basic tenet that the vehicle owner owns and has surveillance data captured for a specific farm. Such
control over the dissemination and use of engine data may be obtained from satellite imagery and
data. At the same time, Tesla advises owners that data retailers, or generated via the use of an UAV or
they will experience reduced functionality should drone.
they opt out of sharing telemetric data with the
company: In the case of satellite imagery or data, there are
currently a number of different service providers,
Please note that, if you opt out from the collection of and imagery is available on virtually a weekly
Telematics Log Data or any other data from your Tesla
vehicle, we will not be able to notify you of issues basis from some of these. There are no restrictions
applicable to your vehicle in real time, and this may result preventing any person from purchasing this
in your vehicle suffering from reduced functionality, data (with some minor exceptions for military
serious damage, or interoperability, and it may also installations). Anyone who is prepared to pay for
disable many features of your vehicle including periodic
software and firmware updates, remote services, and
the cost of the service can purchase high resolution
interactivity with mobile applications and in-car features imagery (at sub-metre precision) and data relevant
such as location search, internet radio, voice commands, to a particular farm or specific area of land.
and web browser functionality. (Tesla Motors 2015) Generally, providers of these services also have
available a library of images through time, enabling
Court ordered disclosure changes in land use to be tracked over time. While
As with data collected on-farm, machinery data is a farmer holding such imagery or data may resist
eligible to be subpoenaed provided it is relevant to making their own specific copy of it available to
a specific case (New South Wales Young Lawyers third parties, there is nothing to prevent those third
Civil Litigation Committee 2010). parties, including governments, simply purchasing
their own copy of the same imagery or data.
Legally obligated disclosure
In the case of farm data or surveillance imagery
Confidentiality laws and legal obligations captured using a drone, if the farmer has
applicable to production data (discussed earlier) commissioned the surveillance, then the resulting
imagery or data would be considered to be owned In the case of an individual operating the drone in
by the farmer, and confidentiality and privacy laws their private capacity, it seems that the provisions of
would apply. One area of uncertainty may be a case the Privacy Act do not apply.
where the data is obtained by a contractor providing
a service. In this instance, a question may arise This is by no means certain, as it is an area of law
about whether the contractor has any ownership that is yet to be tested. It has also been emphasised
rights over the data and imagery collected, and that the difference, and therefore the legal
whether the contractor could subsequently use that distinction between imagery from a satellite and
imagery or data for another purpose, or sell it to a imagery from a drone is not obvious. As a result,
third party. It would be anticipated that this question differentiating between drone surveillance and
would be dealt with in the wording of any contract remote satellite surveillance appears to be quite
between the service provider and the farmer, and difficult from a policy perspective.
in that respect the situation is no different to that
involving a contract harvester whose machinery has The Australian Governments Attorney Generals
the capacity to create yield maps. In both instances, Department informed a House of Representatives
it will be important for farmers to clarify not just Committee that remotely piloted aircraft fall within
the ownership of the data they are provided with, the definition of an optical surveillance device of
but also the limitations that may apply to any the Commonwealth Surveillance Devices Act 2004
subsequent use of that data by the contractor. (House of Representatives Standing Committee
on Social Policy and Legal Affairs 2014). As
In a situation where drone surveillance has been such, existing Australian surveillance laws could
carried out which the farmer has not commissioned apply. However, the Committees report also
or authorised, the ownership and privacy issues are acknowledged that the surveillance laws were
much less clear. designed for devices physically attached to a
property, and therefore their application to drones
Farm data and remote imaging captured by drones required review.
might be considered personal information. That
could be the case in the event that data collected Should Australian surveillance laws be extended
was considered to be attached to private property to drones, it is difficult to identify a reason why
and related income generation and farming satellite imaging would not also be treated similarly.
practices. In that situation, the provisions of However, the collection and sale of satellite
Australian privacy legislation would be considered imagery is already commercialised, and condoned
to apply, but only to governments and those by Governments which are regular users of such
organisations to which the Privacy Act applies services. Even in the event that Australian law
(Commonwealth of Australia 2015a). restricted the availability of satellite imagery for
Australian users, the same imagery can readily be
Under this legislation, notification of the collection obtained from international vendors via the internet,
of personal information requirements would which means any restriction on the availability of
determine much of the process by which third- satellite imagery within Australia would be largely
party drone operators could collect farm data. ineffective.
Specifically, consent would be required from the
farmer before farm data could be obtained. It is The Australian Law Reform Commission has put
likely that emerging drone technology will provide forward the following Uniform Surveillance Law
more accurate data, however such improvements proposals.
would not lie outside the scope of current privacy
Proposal 133 Offences in surveillance device laws
principles relating to data collection (Office of the
should include an offence proscribing the surveillance or
Australian Information Commisioner 2015). recording of private conversations or activities without
the consent of the participants. This offence should
apply regardless of whether the person carrying out
the surveillance is a participant to the conversation or based equipment owned by a farmer will be
activity, and regardless of whether the monitoring or controlled by the farmer, except in the case of
recording takes place on private property.
machinery operating data, which the equipment
(Australian Law Reform Commission 2014)
manufacturers may reserve ownership rights over.
The complexity and uncertainty associated with
In the event that this information is transmitted to a
these issues is highlighted by the use of drones
cloud storage facility or farm software platform, the
by animal rights activists to covertly film alleged
subsequent ownership rights will be determined by
mistreatment of animals on farms. Prosecution of
the Conditions of Use provisions of that facility
these groups has proven to be difficult. Animal
or software platform, a number of which now have
rights groups have exploited the band of airspace
arrangements which enable farmers to control the
above the limits of private property (30 metres) and
use of their information, and who has access to it.
below commercial airspace (122 metres) in order
to avoid trespassing laws (Murphy 2013; House Data ownership and use rights become more opaque
of Representatives Standing Committee on Social in the event that the cloud storage or farm software
Policy and Legal Affairs 2014). facility transfers the data to a third party even
in the event that such a transfer is permitted by
In the event that remote surveillance was
the farmer on whose land the data was originally
determined to be illegal regardless of whether
generated. Ideally, data use and access permissions
physical trespassing laws have been breached,
should travel with the data, although whether this
activists may still find legal recourse to pursue
is legally enforceable by the original data owner
drone surveillance as a consequence of potential
is unclear, and the extent to which this applies to
exceptions to proposed laws. For example, proposal
anonymised data is also unclear.
134 in the Serious Invasions of Privacy in the
Digital Era discussion paper by the Australian Law Ownership rights over farm data generated by a
Reform Commission proposes the following as contractor (such as a contract harvester) remain
a valid defence in response to prosecution under unclear, unless the conditions of the contracting
surveillance laws: agreement between the farmer and the contractor
Proposal 134 Defences in surveillance device laws
specify both the ownership rights over any data
should include a defence of responsible journalism, for generated, and the permitted uses of that data by the
surveillance in some limited circumstances by journalists contractor.
investigating matters of public concern and importance,
such as corruption. There are only quite limited ownership rights
(Australian Law Reform Commission 2014) available over remote surveillance data or imagery
(either satellite or drone) associated with a
This may provide an avenue for animal rights particular farm or area of land. High resolution
activists to sustain a defence in a case involving satellite imagery can be and is purchased routinely
prosecution for undertaking covert surveillance of by corporations, individuals and government
a farm. agencies, and the owner of the land from which the
imagery or data was generated has no say in who
5.4 The privacy of farm data could or should have access to that data.
The preceding case studies highlight the Ownership of data generated via drone surveillance
complexities of the privacy issues associated commissioned by the farmer resides with the
with the generation and collection of digital farm farmer. It is not legally possible at present to
information, and the legal uncertainty associated prevent other individuals from using a drone to
with these issues. obtain surveillance data from a privately-owned
area of land or a farm.
Generally speaking, ownership rights and
subsequent use of data generated using ground-
As a general rule, all farm data is susceptible to agency covered by relevant legislation that aims
a subpoena issued by a court of law, assuming to ensure information contributed by individuals
that the data is relevant to a specific legal action. or businesses for statistical purposes remains
This applies irrespective of whether the data is in confidential.
digital or written form, and irrespective of whether
it is stored on the farm or by a service provider A later section of this report provides further
(New South Wales Young Lawyers Civil Litigation discussion of these issues, including proposed
Committee 2010). policy responses to a number of these where it
is judged they may be an impediment to the full
Farm data provided to a government agency is realisation of the benefits available to the farm
likely to be susceptible to an FOI request lodged sector arising from digital agriculture.
on that agency, except in the case of a statistical
48
6. Technological limitations
X.0 XXXXX
of digital agriculture
New technologies invariably come with limitations 3. errors that are caused by unexpected or
or teething problems. This section explores some unmeasured changes in the data environment
specific issues which are ubiquitous across digital (Reimsbach-Kounatze 2015).
agriculture in Australia. The first is the quality of
data. The maxim garbage-in equals garbage-out These categories provided a useful framework in
applies just as much to digital agriculture as it which to identify vulnerabilities in agricultural data.
does more universally to any area of scientific
investigation or use of data for production Poor data quality
decisions. Data quality is essential to generating The Quality Framework and Guidelines for OECD
value and technological acceptance. Statistical Activities define seven dimensions
of data quality; relevance, accuracy, credibility,
The second limitation in Australia is regional timeliness, accessibility, interpretability and
telecommunications connectedness. Most digital coherence (OECD 2011). The ABS Data Quality
agriculture applications are only fully functional Framework also has a similar set of seven
in situations where wireless or mobile internet dimensions. In the case of the ABS, institutional
access is available, and unfortunately this is not environment is included in place of credibility
the case for many regions of Australia where with the other six dimensions listed being identical.
digital agriculture applications might conceivably The ABS notes that their framework is based on the
be deployed. The extent to which ATPs can adapt framework adopted by Statistics Canada and the
products to data restrictive environments will European Statistics Code of Practice.
be critical to the future development of digital
agriculture in Australia. Of these, accuracy and coherence are probably the
greatest limitations to data quality relevant to digital
6.1 Data errors agriculture.
While one of the strengths of true big data Accuracy is defined by the OECD as:
applications is that the sheer volume of information
means that the impact of low-frequency data [T]he degree to which the data correctly estimate or
errors can be minimised, many digital agriculture describe the quantities or characteristics they are designed
to measure. Accuracy refers to the closeness between the
applications at an individual farm level are not
values provided and the (unknown) true values.
true big data applications in reality, and data errors
(OECD 2011)
can be an important limitation to the utility of any
system. Accuracy is determined by the functionality of
the technology generating the data, including the
Data errors can arise from a number of different
calibration processes and collecting techniques.
sources. These include:
Some sense of the challenges associated with data
accuracy can be gained from a consideration of the
1. poor data quality
operation of yield monitors in grain harvesters.
2. errors associated with the inappropriate use of
data and analytics
CHAPTER 6: Technological limitations of digital agriculture 49
Generally, harvester yield monitors rely on some their farm equipment needs and also require the
form of flow meter installed in the grain elevator same of any contractors in order to fully adopt
which transfers the grain to the storage bin on the digital agriculture. This has now changed with the
harvester. The flow meter is normally an electronic development of more sophisticated farm software
load cell, and may be subject to errors depending platforms and the adoption of open access data
on how well the machinery has been maintained protocols.
and the sensor cleaned. Data accuracy can also be
affected by the moisture content of the grain being The open data concept is based on the preference
harvested, which is also monitored by sensors in that the data produced by each different machine/
the grain harvester. The flow meter is normally technology is able to be accessed by software tools
calibrated by harvesting an initial volume of grain, and converted from one form to another not
which is then accurately weighed and used to adjust necessarily that all data are in a standard format.
the flow meter to correctly estimate the weight of Unfortunately, at the present time data conversions
grain being harvested. are not always loss-less and can result in
information being lost in the translation, although
An additional source of data error can arise from this is steadily improving.
the way the yield data is collected, because the
harvester is travelling across a paddock and there The remaining five data quality dimensions are of
is a delay between the standing crop being cut by less immediate concern.
the front of the harvester, and the threshed grain
subsequently flowing into the storage bin. An Relevance:
adjustment or delay factor is required in order to The relevance of data is an assessment of the value
accurately associate the flow of grain into the bin contributed by these data. Value is characterised by the
with the particular part of the paddock from which degree to which the data serves to address the purposes
that grain was harvested. for which they are sought by users. (OECD 2011)
Discussions with industry personnel in the US Somewhat inevitably, the ability to utilise and
revealed that yield monitor errors of up to 10% derive value from the data has not kept up with
have been reported, arising from a combination of the volume of data being generated. Large
factors including faulty flow meters, calibration stores of farm data in the form of yield maps are
errors, and errors in GPS systems which mean that currently cluttering up the hard drives of numerous
the harvester does not accurately sense when driver farm computers, but are not being used and are
error has resulted in the header front overlapping a essentially valueless. This is not a significant
previous run. problem as the cost of data collection and storage is
relatively cheap, although it may foster an attitude
Coherence refers to the compatibility of data amongst farmers that digital farm data is of little
obtained from different sources. In terms of data relevance to their business.
quality, a higher level of coherence allows data
to be compared, aggregated and analysed with Credibility: Credibility as defined by the OECD
greater confidence The ability to merge and transfer (2011) refers to, the confidence that users place in
datasets is pivotal to facilitating a competitive those products based simply on their image of the
market for data analysis and avoiding separate data data producer, ie the brand image. Confidence by
silos that have diminished utility. users is built over time.
During its initial stages, coherence was a major The infancy of many digital agricultural
challenge for farmers adopting digital agriculture. applications make credibility difficult to assess.
Data from different machinery manufacturers were Credibility is largely governed by the reputation
incompatible, and farmers faced the prospect of of the institutions collecting the data. In the case
having to rely on a single machinery brand for all of digital agriculture many of the private data
50 CHAPTER 6: Technological limitations of digital agriculture
collectors and digital agriculture service providers to the public are anticipated to be data-driven
already have longstanding relationships with applications and software rather than publications.
customers through other non-data driven products. The data products (publications) considered as part
Much of the initial credibility of digital agricultural of data quality are destined for specialised service
services will be influenced by company reputation providers who then face challenges associated with
in regard to on-farm services. Over the medium to user interfaces and automated systems compliance.
long term credibility in the eyes of the customer
base will be determined by whether farmers believe Interpretability: The interpretability of data
that digital information is generating returns for products reflects the ease with which the user can
farmers, rather than just scientific publications. understand and properly use and analyse the data.
Timeliness essentially refers to the availability of The magnitude of interpretability issues depends on
data in an appropriate form at the time it is needed the targeted user. Digital agricultural companies are
in order to aid decision-making. This is generally becoming increasingly sophisticated, and what is
not an issue in relation to digital agricultural and isnt interpretable varies significantly between
applications, as most generate data instantaneously. different organisations. Relative inaccessibility and
Perhaps the only timeliness issues that arise are in complexity at the data collection stage may not be
relation to the availability of soil test data, which a limitation provided it can massaged by software
require the services of an external laboratory. systems into user-friendly products (Australian
Generally speaking, however, the timeliness of Bureau of Statistics 2010).
agricultural data is not a major limitation.
Errors associated with
Accessibility: The ABS evaluates data accessibility inappropriate data use
under two key aspects.
Having access to good data and analytic tools is no
guarantee that sensible insights will be obtained.
Accessibility to the public: the extent to which
The sheer volume of data now able to be generated
the data are publicly available, or the level of
in many industries has tempted many to forgo
access restrictions. Additionally, special data
scientific method in favour of population analytics.
services may include the availability of special
Using analytics, the reason for a relationship
or non-standard groupings of data items or
between two variables does not matter in so far as
outputs, if required.
it holds true if analytics can show that an increase
Data products available: this refers to the of A causes an increase in B, then a user of the
specific products available (eg publications, information need not understand the reason for this.
spreadsheets), the formats of these products,
their cost, and the available data items which The first problem is that modelling interactions
they contain. between all relevant variables using data analytics
is rarely possible. Correlations can often be
These aspects are tailored to the role of the ABS discovered, but with no causal relationship. An
as a national statistical organisation, however increase in A may in fact be correlated with an
the underlying concepts remain important to increase in B, but not caused by it. For example,
agricultural data accessibility. In the case of farm increases in both A and B may be the result of
data, accessibility for the public is not as important an increase in variable C or the data may simply
as accessibility to research organisations and be implying a statistical significance within
service providers. random fluctuation. Relying on analytics in lieu of
understanding the underlying relationship leaves
The second aspect of accessibility as defined by decision-makers vulnerable to a change in the wider
the ABS is concerned with the data products environment in which these variables are observed.
available. In a sense this is less of a problem If it is suddenly observed that changes in A and B
with agricultural data. The end products available
CHAPTER 6: Technological limitations of digital agriculture 51
are no longer correlated, it can be very difficult to use option. As a general rule, internet connections
understand why relying on analytics alone. with data transfer speeds of less than 1.5 megabits
per second (Mbps) are not able to transmit even
Data has been utilised to reach erroneous relatively small yield monitor data files (Griffith et
conclusions in a range of different industries, al. 2013).
and sometimes for underhanded purposes. Even
qualified data researchers have been guilty of
cherry-picking spurious correlations for their own
benefit. Spurious correlations can be discovered
when large numbers of variables find their way into
big datasets (Taleb 2013).
Errors that are caused by unexpected

changes in the data environment
Errors occur when data becomes distorted due
to unforeseen or unmeasured changes within the
data environment. For example, the confluence
of weather systems, pathogens and human
management can make agricultural data analysis
particularly susceptible to errors arising from a
changing external or unmeasured environment. This Figure 18: 3G mobile telephone coverage
can result in distorted results which are otherwise provided by Telstra.
accurately modelled under normal conditions. Source: Telstra Corporation Limited (2015).
Compounding these problems are potential As the map in Figure 18 shows, there are large areas
incentives for data tampering, particularly related of regional Australia that do not have any mobile
to farm input regulation (Reimsbach-Kounatze phone coverage, and large areas that only have very
2015). For example, farmers may have an incentive low rates of access speed.
to obfuscate data when restrictions are imposed on
the use of water or nutrients. Subsequent analysis To mitigate poor internet access speeds on farms,
of resulting data may result in conclusions being ATPs design applications so that a user can access
reached about improvements in water and nutrient all important farm information with or without a
efficiency that are not justifiable. mobile telephone signal. An example is Farmobiles
passive uplink connection (PUC) which can store
data until an internet connection is available, and
6.2 Digital agriculture can transfer farm data from multiple different
and internet access machines to a single storage site when a mobile
Many machinery companies now offer a wide range connection is available. This permits data to be
of in-built precision agricultural technologies that collected from machinery without major loss in
involve collecting data which helps monitor the functionality, albeit with slightly delayed upload to
real-time performance of a machine, field or crop. the cloud (Farmobile 2015).
However to fully utilise this type of technology,
For in field applications that require an internet
a mobile data connection is required in order to
connection such as those which incorporate
transfer information to and from the machine to a
weather data a lack of mobile access to transfer
website or digital agriculture platform.
data is an issue. Applications and products
If internet speeds are too slow then most data are generally designed to run as close to full
applications are simply an expensive and limited- functionality as possible without an internet
connection in order to accommodate variable data
52 CHAPTER 6: Technological limitations of digital agriculture
transfer speeds in rural areas. For example, variable available in other nations that compete for similar
rate planting maps are often created with the aid agricultural export market opportunities, such as
of online analytics, but are then exported to the Denmark, the US, New Zealand and Canada. As a
on-board controllers which can execute a crop result, there are many Australian farmers who could
planting or fertiliser prescription without an internet remain years behind global peers in technology
connection. adoption due to inadequate internet access.
Nearly all businesses in Australia access the internet

Regional Australia is scheduled to receive faster
using a broadband connection. However, businesses
and more reliable internet from the first half of
in the agriculture, forestry and fishing industry lag
2016 with the roll out of the NBN. The first of two
behind other industries when it comes to accessing
satellites was launched in October 2015 as part
broadband, building a web presence, placing orders
of regional satellite broadband network designed
via the internet and receiving orders via the internet.
to enable wholesale download speeds of 25 Mps.
There is no doubt that the relatively low quality
Data will be transmitted to and from fixed receivers
broadband coverage in rural and remote regions
attached to regional homes and businesses. This is
has been a major reason for the slow adoption of
expected to improve internet access for many farm
internet functions by these businesses.
businesses, however fixed receivers will not be able
In many cases, Australian farm businesses are to transfer real-time data from operations in field
operating with broadband connections that are (NBN Co. Limited 2015).
slower, more costly and less reliable than those
53
7. Discussion and recommendations
Digital agriculture is developing at a very rapid at a sub-field level in response to evident variations
pace, driven by technological developments in soil and other variables, once data had been
occurring external to agriculture which are accumulated over a number of years.
leading to an inexorable reduction in the cost of
computer power and digital technologies, very During these early stages, farm input and machinery
rapid growth in the use of smartphones and mobile suppliers provided digital agricultural systems as
computer technology, and the global growth of a loyalty incentive for users of their products, and
the internet. Some sense of the rapidity of change many developed unique systems in order to make it
can be obtained from the observation that the first difficult for farmers to swap to competing suppliers
smartphone was released by Apple in June 2007, without losing their farms digital assets.
and since that date there have been over 700 million
iphones sold by Apple worldwide, as well as However, alternative software and digital platform
countless smartphones sold by other manufacturers. suppliers quickly emerged and provided farmers
(Ingraham 2015). with the ability to divert farm digital information
away from proprietary platforms, and also to
The rapidity of these changes makes it difficult translate the digital information to formats that
to project likely future digital agriculture were compatible with alternative systems. At the
developments with any certainty, and hence to same time, it seems that owners of the proprietary
propose industry initiatives or government policy systems came to the realisation that it would be
settings that may have beneficial impacts. The pace difficult to service all the software needs and desires
of change makes it highly likely developments of farmers, and that it was strategically smarter to
will overtake industry or government initiatives, provide platforms with open access arrangements
even before they are implemented. An added that encouraged a competitive market to develop for
complication from an Australian perspective is that specialist software applications.
sub-sectors of Australian agriculture are likely to
inherit digital agricultural platforms and systems Consequently, over recent years digital agriculture
that have been developed in the US, and these will systems in the US have been democratised. Farmers
have been developed within a policy framework and can now choose from competing platforms and
science base associated with that market, meaning systems, and can store information arising from
the opportunity to develop unique Australian a range of different machinery and technology
versions of these systems may be limited. brands on a single platform. They can also choose
preferred software applications to manipulate or
These qualifications noted, it is instructive to analyse that data from a competitive software
review some of the strategic developments that marketplace, and are largely able to switch
have occurred in digital agriculture in the US as between different platforms and systems without a
it has evolved. The first developments arose in significant loss of data.
response to the possibility of more precise crop
management brought about by the incorporation of Australian crop producers, in particular, have
GPS technology into the control systems of farm adopted elements of digital agriculture (often
machinery during the 1990s. Harvester yield maps referred to as precision agriculture in the
created the opportunity to plant and manage crops cropping context) to a greater degree than their US
54 CHAPTER 7: Discussion and recommendations
counterparts, especially controlled-traffic farming, There are existing organisations and industry bodies
minimum or no tillage and GPS guidance systems, already operating in this space to some degree,
but have not adopted variable rate cropping such as controlled traffic farming groups and
applications to the same degree, and do not have precision agriculture groups. These are generally
available the software and digital platforms focused on the broadacre cropping sectors. Given
available in the US market. It is evident that a the rapid development and likely future growth
commercial digital agriculture software market has of digital agricultural applications in the livestock
not emerged in Australia to the same extent as has and horticulture sectors and the desirability of
occurred in the US. having broad, cross-agriculture engagement in these
issues, it is felt that the label digital agriculture
The smaller scale and greater variety of Australian is preferable to precision agriculture, in order to
cropping industry sub-sectors is undoubtedly a ensure there is a clear understanding that the forum
factor limiting software development, as is the encompasses the full breadth of agriculture, and
much lower level of private sector involvement in not just the cropping sector. It also recognises the
seed breeding in the cereals industries compared to reality that Australian farms are more likely to be
the US corn or soybean industries. A further factor mixed enterprise farms than is the case in the US,
that may be inhibiting software development is the and that there is therefore likely to be advantages in
lack of detailed soil and climate data in Australia, the development of systems and platforms that have
recognising that while the US soil data may be the potential to accommodate multiple different
somewhat imprecise within fields, it does at least farm enterprises, rather than just focus on a single
provide a platform to build from. A further limiting enterprise.
factor in Australia appears to be that Australian
university researchers and the CSIRO do not have Recommendation 1:
good track records in engaging with industry and
Australian agricultural industries, Australian
transferring knowledge to the private sector in order
agricultural research agencies and relevant IT,
to enable it to be commercialised and developed.
telecommunications and software organisations
One thing that is very evident from observing should collaborate in the establishment of the
developments in the US is that it is the private Australian Digital Agriculture Forum, with the
sector which has taken the lead in developing broad objective of advancing the development
software systems and platforms that are user- and adoption of digital agricultural applications
friendly and commercially appealing, not public- and systems in Australia.
sector researchers. What is also very evident is
Convening such a forum on a regular basis has the
that it is data analytics and software specialists that
potential to assist in the development of networks
have been at the forefront of commercial digital
between the various disparate groups that might
agricultural developments in the US, not plant and
have an interest in advancing these developments,
animal scientists.
and in particular creating avenues to commercialise
To facilitate faster development of commercial the delivery of digital agricultural applications in
digital agriculture applications in Australia Australia.
(encompassing livestock, cropping and horticulture
The establishment of a forum involving all relevant
sectors) there appears to be merit in creating a
interests across the entire agriculture sector creates
regular, structured forum involving agricultural
the potential to develop, where necessary, industry-
scientists with appropriate technical knowledge and
wide standards. This will assist in ensuring that
experience, software and information technology
a competitive digital agriculture service market
specialists, and farm input suppliers or service
develops, that there is industry-wide agreement
organisations with an interest in developing digital
on issues such as interoperability, transferability,
agriculture platforms or specific applications.
data ownership and privacy. It will ensure that
CHAPTER 7: Discussion and recommendations 55
the potential exists for concerted industry action research conducted as part of the project reported
on issues such as telecommunications network here has led to a very strong recommendation that
infrastructure, and policy issues that may impact on both these principles should be adopted, and are
the development of digital agriculture in Australia. necessary in order to provide a solid foundation
for the development of digital agricultural systems
While developments in digital agriculture in in both the cropping and livestock industries in
Australia or certainly in the development of Australia.
integrated software platforms and systems lag
those in the corn industry in the US, this provides Recommendation 2:
an opportunity to learn some lessons from
the developments that have occurred, and the
industries, agricultural technology providers
subsequent evolution that has occurred in the digital
and digital agriculture platforms and software
agriculture system of the US.
system providers should adopt as a key principle
that the farmers who own the land or livestock
Data ownership and access
from which digital agricultural production
Perhaps the first lesson to emerge has been that information is obtained retain ownership rights
attempts to develop digital agriculture systems over that data. This includes the ability to
as proprietary systems limited to one particular determine the uses to which that information
machinery manufacturer or seed company have can be put, and the persons or organisations
not been successful, and both technological which can obtain access to that data. Where
developments and the desire of farmer clients to contractors and sharefarmers are employed,
have access to interoperable systems has meant that it is recommended that a standard contract be
open access data arrangements quickly evolved. developed that defines data access protocols for
This means that farmers using a number of different each party.
machinery brands on their farm can still have all
their farm data located on a single storage platform, Recommendation 3:
and that the same information can be utilised by
a number of different applications. It also means
that farmers are not tied to a particular machinery
and digital agriculture platforms and software
brand or software system and can choose to move
system providers should commit to open access
to an alternative system without a subsequent loss
data protocols, modelled on the standards
of accumulated farm data.
adopted by the Open Agriculture Data Alliance
Open access data arrangements are based on established in the US.
the intrinsic understanding that farmers who
Data privacy
generate digital farm production data retain
ownership of that data, and have a right to dictate One of the key issues that has the potential to limit
the purposes for which that data can be utilised. the development of digital agriculture in Australia is
These two principles farm data ownership and concern about the privacy of farm data. The concept
the requirement that data should be able to be of privacy in the digital era requires a re-evaluation
seamlessly transferred between different systems of existing concepts and arrangements. As earlier
and providers have been fundamental to the discussion has highlighted, control over farm data
development of digital agricultural systems in the collection and information dissemination has been
US, and there seems very strong logic that similar highly diminished by technological developments.
principles should be adopted in Australia. These It is now entirely feasible for satellite imagery or
principles are potentially the subject of discussion machine telemetry data to be utilised to obtain
and ideally agreement by members of the proposed detailed information about farm operations without
Australian Digital Agriculture Forum, although the a farmer having any awareness this information has
been collected. This data can be combined and fed recommendation from the report was that, Policy
into increasingly sophisticated algorithms which attention should focus more on the actual uses of
allow for private information to be inferred well big data and less on its collection and analysis.
beyond the scope of original datasets.
Focusing on the use of data rather than collection
Attempting to apply privacy controls over the and analysis has the added benefit of allowing
collection of data is increasingly untenable. legislation, regulations or guidelines to be
For example, the information contained within developed without referencing technology. As
telematics data captured by farm machinery may be a general principle, policy measures should be
collected for the purpose of monitoring machinery technologically neutral wherever possible so as to
performance and diagnostics. This data combined preserve relevance as the technology evolves.
with other public and private data (for example
from soil testing laboratories) has the potential to Privacy rights are most commonly determined
reveal information including soil type, operator through a use agreement which consumers would
competence, operational processes and crop yield. recognise as a software licensing agreement that is
Similar issues arise for data derived from livestock ticked and flicked when accessing new computer
industries. National Livestock Identification software or other applications.
Scheme (NLIS) information combined with post-
slaughter abattoir data can provide very detailed In theory, this approach makes sense. The user of
information about a livestock business. Similarly, the data collecting product or service agrees to
the data collected and held by dairy processors what, with whom, and how their personal data will
provides a very detailed picture of the operations of be used. In practice however, these agreements
the businesses of their farmer suppliers. are frequently written in a way that obfuscates key
exception clauses which the user must accept as a
Put simply, the volume of data and the analytics condition of use.
at the disposal of third parties makes quarantining
farm digital information increasingly problematic. The privacy exceptions hidden within the legalese
of user agreements for agricultural machinery may
It should also be recognised that imposing onerous not be as complex as some others, but they are
restrictions on the collection and use of farm data still likely to be significant. It has been recognised
runs the risk that such measures will seriously internationally that even these are impractical
curtail the necessary research and development for farmers to read, comprehend and potentially
that has the potential to provide insights that have negotiate with manufactures and suppliers handling
the very real potential of delivering much-needed their data.
productivity gains for Australian farmers.
The US Farm Bureau, in conjunction with major
This issue was addressed in a general sense in a industry service providers, developed a set of
recent report prepared for the US Government Privacy and Security Principles for Farm Data
entitled, Big data and privacy: a technological to which the signatories agreed to be bound.
perspective (Presidents Council of Advisors on Similarly, The New Zealand Farm Data Code of
Science and Technology 2014). In seeking to Practice establishes industry practices for compliant
address concerns about data privacy, the report organisations, and provides accreditation for
questions the legislative wisdom of focusing on compliant services or products (Dairy New Zealand
data collection and algorithms in isolation. Neither 2014).
data, nor the accompanying algorithms can be
considered harmful to society. Information derived This requirement to satisfy a code of practice places
from combining the two can, however, be used to a degree of faith in participating organisations to
infringe upon individuals rights to privacy. A key adhere to the agreed standards, but also leaves open
the potential for differing interpretations of that code.
For example, does a requirement not to disclose data use option table with standardised use options
information to third parties for commercial benefit similar to those in Table 3.
apply to all data, anonymised data or aggregated
data? It would not be surprising if there were Table 3: Example of standard use options that
different answers to this question, depending on the may be made available for farmers.
person asked. The report by the Presidents Council
User
on Advisors on Science and Technology referenced ATP data use options
preference
earlier recognised this issue, and proposed the
appointment of an intermediary (a data ombudsman) Telematics machine monitoring services Yes only
to digest the terms of use associated with software Private product development Y/N
and provide an assessment to the marketplace (for
Private customer research Y/N
example via accreditation) of whether the product
or service adheres to a set of overarching agreed Public and higher education research Y/N
principles similar to the US Farm Bureaus Privacy Direct marketing (software provider only) Y/N
and Security Principles for Farm Data or the New
Direct marketing (external company) Y/N
Zealand Farm Data Code of Practice.
Real estate property valuation Y/N
The belief underlying this proposal is that as the
Commodity trading No only
market starts thinking about approval from the
intermediary (the data ombudsman) as part of a
purchase decision, software and service providers The ombudsman would be expected to have the
will recognise the commercial imperative and capability to monitor data usage, ensuring that
comply with these standards, otherwise potential it was within the bounds of the pre-selected user
purchasers will seek out other suppliers. preferences.
The role of an ombudsman could be extended Tracking use preferences post- as well as
to afford farmers greater control over the use of pre- analysis represents a formidable challenge.
data by technology suppliers and third parties, Algorithms operating on data will need to be vetted
recognising that not all farmers will make the according to output properties. The Presidents
same privacy demands of their software providers. Council of Advisors on Science and Technology
An ombudsman might not only accredit different report highlights the use of data tags and attributes
software products against an industry standard, to ensure fidelity to the original purpose as data
but may also facilitate the ability of farmers to output proliferates through analysis.
personalise the uses and access arrangements The privacy policies of the output data must be computed
for their data. This could be a way of avoiding a from the policies associated with the inputs, the policies
situation where farmers are faced with a binary associated with the code, and the intended use of the
(yes/no) decision about the use of their data by third outputs (ie the context).These privacy properties are
a kind of metadata. To achieve a reasonable level of
parties, and automatically opt to prevent access
reliability, their implementation must be tamper-proof
by third parties, irrespective of how benevolent or and sticky when data are copied.
useful that third-party access may be. (Presidents Council of Advisors on Science and
Technology 2014)
For example, an intermediary accreditation body
could define different use categories for farm data, The report acknowledges that formalisation
such as direct marketing by third parties, private and implementation of these policies is very
research, public and higher education research, much a recent phenomenon and much of the
farm financial analysis, and property valuations etc, target technology is still in the research phase.
which the farmer could select at their discretion. Nevertheless, demarcating the use of data
On the purchase of software or the establishment throughout larger organisational structures has been
of a user account, a farmer may be presented with a shown to be workable.
Recommendation 4: new technology and more recently, arm themselves

with knowledge to help protect themselves against
privacy violations. The privacy rights of farmers
will potentially be curtailed to some extent under
and digital agriculture platforms and
contracts with digital agriculture providers, and as
software system providers should support the
such, it would be desirable that education programs
appointment of a Farm Data Ombudsman to
on data privacy be developed for farmers. The
oversee data privacy standards, to establish
Australian Centre for Intellectual Property in
data use categories, and to audit compliance by
Agriculture (ACIPA) is one group which provides
providers with industry data privacy standards.
research and extension services which educate
An ombudsman would function as a supplement to farmers on contract law and application (Australian
agreed privacy principles. Farmers should still be Centre for Intellectual Property in Agriculture
able to exercise the right to address and negotiate 2015). This organisation and others could work
contracts directly and be encouraged to do so. It with the proposed Data Ombudsman to deliver
would still also be highly recommended that farmers education programs to farmers.
review their contracts and are educated in their
Empowering farmers to negotiate better outcomes
privacy rights where possible. However, it is also
or at the very least encouraging them to read and
recognised that farmers cannot be expected to be
comprehend privacy clauses is a laudable goal,
IP lawyers, data scientists and an IT department.
however, this should not become the first line of
Serviceable default privacy settings and rights for
defence against privacy violation for all farmers.
the many who simply dont have the time or capacity
Between the turmoil of big data development and
to scrutinise privacy agreements are a necessity.
time constraints of farmers, care should be taken to
The office of the data ombudsman would be avoid overburdening farmers with legal obligations.
responsible for setting a list of prohibited and Education should be seen as a supplement to
mandatory clauses which would (ideally) be industry privacy standards.
developed in conjunction with major ATPs and
Data anonymisation
other stakeholders. The list would require periodic
review as well as mechanisms for triggering Data anonymisation has been useful in the past,
reactive assessment of clauses similar to those particularly for research purposes whereby
within the New Zealand farm data code of practice. personal information is detached from the data,
while sufficient classifiers remain for analysis. For
Signatories to the resulting data privacy standard example, few farmers would be concerned if data
could receive certification in much the same way detailing attributes such as soil, weather, geographic
as those who comply with an industry code of region, or crop yield were accessed for public
practice. Alternatively the ombudsman would research purposes, if the data was anonymised
simply maintain an updated list of compliant and could not be attached to individuals or their
organisations available to the public. property.
The position of the data ombudsman would ideally Anonymisation has served to allow the use of
be reviewed every five years to determine whether private data while removing privacy concerns.
the position remains necessary, as it would be Unfortunately there is growing doubt as to whether
anticipated that user understanding of agreements anonymisation techniques can protect privacy as big
should become normalised over such a period. data analytics evolve and proliferate. Reservoirs of
historical data and analytics software can take new
Farmer education anonymised data and infer information not apparent
Education has long been seen as a necessary in the dataset alone. The Presidents Council
measure to ensure farmers can effectively adopt on Advisors on Science and Technology report
concluded data anonymisation would be better
suited for use as additional safeguards rather than would ideally possess knowledge of design and
in the policy framework. application of sensing technologies, and big data
analytics in addition to specialised knowledge of
Data analytic skills farming systems. It would be expected that much
In a recent report into big data and Australian of the syllabus would be aligned with Agricultural
agriculture, it was identified that a shortage of data Engineering degrees already taught in Europe, and
analytics talent within the Australian agricultural North and South America.
sector could be a major limiting factor in the
Role of national statistics services
development of digital agriculture in Australia
(Allen 2015). Agriculture will to have to recruit The role of national statistics organisations (NSO)
data analytics talent from a highly contested, such as the Australian Bureau of Statistics (ABS)
external labour pool. Indeed, as one senior will need to be considered as digital agriculture
industry participant in North America pointed proliferates.
out, the last people that should be appointed to
these roles are plant or animal scientists. They The OECD raised the possibility that an NSO
are trained to analyse very specific results arising could function as a clearing house whereby they
from tightly controlled trials, usually where would set standards for the data and guidance for
only one or two independent factor are subject its usage and associated limitations. There is a
to variation. By contrast, data analytics involves possibility that a clearing house of this sort may
searching for trends or relationships between be useful, but whether the ABS is best placed
different factors utilising very large sets of data, in to take on this role is debateable. It would be
which all the variables may be uncontrolled. Plant preferable for the agriculture sector to manage
and animal scientists, as a consequence of their and control the industry-good elements of digital
training, are generally not proficient in big data agriculture (as has occurred in the US), and for the
analytics. ABS to consider opportunities to take advantage
of the resulting datasets, as and when it might be
The establishment of an Australian Digital appropriate.
Agriculture Forum, as proposed earlier, may
provide a very useful opportunity to develop Cross-referencing data between NSOs and private
programs that will assist in attracting skilled data datasets could have the effect of improving the
personnel to the Australian agricultural sector. quality of both if managed correctly. An example
is the US company Farmlink, which has the
New or modified university courses and degrees capacity to use harvester yield data from a large
should play a role in meeting the potential number of harvesters to calibrate remote imagery
demand. One avenue is the creation of an elective and obtain quite accurate regional and nation
agricultural bridging course embedded within data crop production estimates (Farmlink 2015) at
science/mathematics degrees. This would equip obviously much reduced cost in comparison with
data scientists with an understanding of potential methods of estimation that are currently used to
applications within the agricultural sector, generate official statistics. Collaboration between
thereby reducing the training burden incurred by this organisation and the USDA could result in
agribusiness in bringing data scientists across. It important accuracy and efficiency gains, which
may also become an important tool in attracting would benefit both.
data science talent.
Role of government, researchers
From an agricultural science perspective, there and industry
may be scope for a new cross-disciplinary degree.
The evolution of digital agriculture in the US
ICT and data science courses would form a large
provides some very useful insights into the
component of the syllabus along with farming
respective roles that can usefully be played by
systems. Students who complete the degree
different groups and organisations in facilitating the There are, however, alternative public/private
development of these technologies and systems. models emerging that may have the result of
producing the desired climate and soil resources,
The US Government or respective state but at a reduced cost. In the case of soil data, there
governments do not appear to have a direct would appear to be merit in exploring the potential
role in digital agriculture in the US, yet have for the results of farmers routine soil tests to be
played a very significant role in enabling these combined with available public data to fill in the
developments to occur. Private sector participants gaps in the national soil inventory. It is understood
acknowledge that there are a number of different that there may be issues in relation to variations in
pieces of infrastructure that have been provided the quality of soil testing by private laboratories,
by governments, without which digital agriculture but mechanisms are available to improve the
would not have evolved to the extent that it has. consistency of different soil testing laboratories, and
These include in particular: this would greatly benefit government and industry.
The detailed 1:25,000 soil maps and associated In the case of climate data, fully automated and
data that provides an important basis of many of highly accurate weather recording stations can now
the digital agriculture software systems used in be purchased by individuals for between $5000 and
the cropping sector of the US. $15,000. These are highly accurate and can be set
The comprehensive climate data that is available up to transmit data to a website which can integrate
from a network of high density weather recording available public and private climate data at specific
stations and weather radar sites in theUS. locations, creating a much better and more localised
climate dataset. Incentivising farmers to install
The comprehensive mobile telephone coverage these could have the desired result of substantially
and GPS correction networks that are available, improving the quality of Australian climate datasets
combined with ready access to cadastral data at less cost than would be the case if relying solely
and remote imagery. on public funding to enhance this infrastructure.
Australian governments and research agencies Recommendation 5:

have begun to augment available soil maps and to
combine these with soil information obtained from ustralian governments should increase
A
recent soil carbon projects, but it is understood that available funding for soil mapping and weather
the quality and completeness of Australian soil recording stations, and actively investigate the
maps is greatly inferior to those available in the US. potential for public/private investment models
There have also been some recent initiatives aimed and private-sector collaboration as a mean of
at improving the coverage of weather recording improving the soil and climate datasets that are
stations and weather radar sites, but again these are an essential foundation of digital agricultural
considered by most to result in climate datasets that systems.
fall well short of the comprehensiveness of datasets
available in the US, and which have formed part of A key factor in the adoption of digital agricultural
the essential underpinning infrastructure for digital systems in the US is access to mobile and data
agriculture systems. networks. Some digital agriculture systems and
platforms are able to function in the absence of data
The recent efforts by governments in Australia in coverage, but at reduced functionality and in ways
relation to soil and climate data are welcomed, but that add to the complexity of operation of these
will require much increased and sustained funding systems, and hence slow adoption. Data coverage
to bring these datasets up to a standard required is deficient in many parts of rural and regional
for digital agriculture applications. Whether this Australia, and this has the potential to be a key
is feasible in Australia, given other government limiting factor which will delay adoption and stall
budgetary pressures, is a challenging question. potential productivity gains that might otherwise
beavailable.
The Australian Government has acknowledged the digital agriculture in the US. The research findings
deficiencies in mobile and data coverage in rural arising from public agricultural research provide
and regional Australia with recent announcements the fundamental knowledge about plant and animal
about added funding to augment coverage in mobile growth that ultimately underpin probabilistic
telephone blackspots, however, as the recently decision-support tools that will be at the core of
released Regional Telecommunications Review digital agricultural systems in the future. The data
(2015) noted: arising from numerous crop, pasture and livestock
production trials augment that knowledge bank over
Despite these gains, and the fact that Australians enjoy time, and enable computer models to be constantly
among the highest penetration of mobile broadband in
the world, the low population density over the remaining improved.
geography means that new approaches are needed
to assess the priorities of those in the 70 per cent of Public research agencies can facilitate the
Australias land mass that has no mobile coverage, and development of digital agriculture by adopting
to improve poor coverage elsewhere. open access data protocols when publishing
(Commonwealth of Australia 2015b) research findings, and by including standardised
information (such as geographical location, soil
The Regional Telecommunications Review type, seasonal conditions, soil test data, pasture
considered strategies such as the utilisation of type and availability, livestock feeding regimes etc)
infrastructure for emergency services and mining and appropriate metadata as part of the published
as ways of augmenting existing mobile telephone datasets associated with research outcomes.
coverage, recognising that public funding alone is
not likely to result in adequate coverage. A plan to implement such a system for USDA
funded agricultural research was developed in the
New technologies are now emerging that make US in 2014, and is currently in the process of being
private investment in telecommunications adopted. Australian governments and rural research
infrastructure more feasible than was the case in and development corporations have the capacity to
the past, and create the potential for cooperative make this a formal requirement of research grants
models between governments, landholders and for relevant research projects they fund, and should
telecommunications companies to expand the implement a collaborative program to adopt this
mobile broadband network in rural and regional approach.
Australia. These cooperative models need to
be developed, and utilised alongside increased Recommendation 7:
funding for mobile telephone coverage to remove
a major impediment to the rapid adoption of digital Australian governments and rural research and
agricultural systems in Australian agriculture. development corporations should collaboratively
develop a strategy to make the detailed data
Recommendation 6: and relevant metadata associated with publicly
funded research available in accordance with
ack of access to mobile and data coverage is
L an open access data protocol, and work to
a major impediment to the adoption of digital standardise the availability of other relevant
agricultural systems. Australian governments information about research trials.
should increase available funding to augment
access to networks in rural and regional In many instances, research organisations such as
Australia, and actively investigate the potential universities and the CSIRO have resisted providing
for public/private investment models as a means open access to research outcomes because of a
of further enhancing data coverage. concern that this would diminish the potential to
generate revenue from intellectual property rights.
While it can sometimes be overlooked, public In reality, however, there are few agricultural
agricultural research agencies have played a very research outcomes that have ever resulted in the
important facilitating role in the development of
generation of significant intellectual property rights, agriculture platforms for adoption by farm service
with the exception being specific cases (such as organisations or farmers. The private sector is much
the development of genetically modified cotton better equipped to perform this role.
varieties in Australia) where there was a clear
intent to commercialise the outcome, and the work Recommendation 8:
was carried out in collaboration with commercial Australian publicly funded agricultural research
partners. organisations have a fundamental role in the
generation of knowledge to underpin digital
In most instances, it would be preferable to
agriculture applications, models and algorithms,
make publicly funded research outcomes more
but should not be involved in the development
freely available to industry including potential
of commercial software programs or digital
commercial developers, in the hope of speeding
agriculture platforms that will be used by farm
up any potential adoption opportunities. The
service organisations or farmers.
Easy IP model initially developed by the
University of NSW (UNSW Innovations 2016) Digital agriculture has the potential to
and subsequently adopted by a number of other fundamentally change agriculture in Australia, as
Australian universities provides an example of this it is likely over time to result in completely new
approach, making much of the intellectual property and novel ways for farmers to access information,
generated by university research available for free to record farm performance, and to integrate
for commercial applications utilising a simple, objective but complex farm data in ways that
single-page agreement. support decision-making. In many respects,
digital agriculture represents a new information
It is also essential to recognise that public research
supply chain to and from farm businesses, and its
agencies are unlikely to have the skills and
adoption will constitute a dramatic change in the
commercial adoption pathways available to develop
processes that have been collectively referred to as
and maintain digital agriculture applications.
agricultural extension.
It is notable that, while public research findings
The software applications and platforms that enable
provide the fundamental underpinnings of many
farmers to more easily manage functions such as
of the digital agriculture systems in operation in
soil testing, crop and animal input ordering, farm
the US, the successful applications have all been
record keeping, the development and management
developed by the private sector, and many are
of cropping plans and grazing rotations, and
utilised by farm service providers, rather than by
many other routine farm planning and operational
farmers themselves. Rather than direct adoption by
functions will also have the potential to supply
farmers, the adoption pathway for these systems has
highly specific and targeted extension information
been via commercial service providers.
relevant to each of those functions.
Publicly-funded agricultural research is
A farmer (often in conjunction with a professional
fundamental to the future success of digital
advisor) using a particular platform or software
agricultural applications in the US and
application for ordering cropping inputs such
Australia, but publicly funded research agencies
as fertiliser or herbicides or contemplating the
should not be involved in the development of
purchase of livestock genetic material will have
commercial software applications or digital
the capacity to instantly compare the performance Recommendation 9:

and price of alternative products, check the likely
Private-sector digital agriculture applications
delivery timing and price, obtain a quotation for the
and platforms have the potential to dramatically
cost of delivery, and check the availability and price
change the way in which farmers access
of any required contractors associated with their
production and other information relevant to
use. At the same time, all the relevant technical
farm management decisions. These systems
information including comparative trial results,
will become the principal information supply
withholding periods and export slaughter intervals
chain for farmers in the future, and public-
will be able to be made available and accessible to
sector agricultural research agencies will need
the farmer at the click of an icon.
to develop new strategies that recognise these
These systems will not develop instantaneously, and systems as the principal extension pathways of
there will be at least a decade or more during which the future.
they will be progressively developed and adopted.
The explosion of digital information and computer
This means that they will not necessarily provide
processing capacity is rapidly changing the way
a new, universal extension pathway in the near
in which businesses in all sectors of an economy
future, but will rapidly increase in importance as an
operate, and some of the best resourced businesses
extension pathway in the not too distant future.
(major media companies, for example) have
Given that the successful digital agriculture systems struggled to respond to the challenges arising as
and platforms will almost certainly be those a consequence of digital disruption. It would be a
developed commercially, rather than by government mistake to imagine the farm sector is immune to, or
agencies or rural research and development isolated from, these changes.
corporations, there will need to be some carefully
The opportunity is available to take advantage
considered strategies developed by rural research
of these changes, or to ignore them and become
and development corporations and state government
victims. Australian agriculture is at a critical point
agriculture agencies in order to optimise the
at present, and will require sound strategic planning
opportunities that digital agriculture provides to
and clear thinking in order to take full advantage of
become a new and extremely efficient agricultural
the changes that are rapidly unfolding.
extension system.
64
8. References
AgInnovators (2016), Precision agriculture, Guidelines%20on%20exemptions%20in%20

available at: http://www.aginnovators.org.au/ the%20Act%20PDF.pdf
initiatives/technology/themes/precision-agriculture.
Australian Law Reform Commission (2008), For
agX (2015), agX Store, available at: your information: Australian Privacy Law and
http://www.agxplatform.com/the-agx-store/ Practice (ALRC Report 108), available at:
http://www.alrc.gov.au/publications/report-108
Allen, R (2015), Turning data into information:
maximising the benefit of digital data technology, Australian Law Reform Commission (2014),
available at: http://nuffieldinternational.org/ ALRC Discussion Paper 80: Serious Invasions
rep_pdf/1431204756Robert-Allen-report-2014.pdf of Privacy in the Digital Era, available at:
http://www.alrc.gov.au/publications/serious-
Australian Bureau of Statistics (ABS) (2010), invasions-privacy-dp-80.
The ABS Data Quality Framework, available at:
http://www.abs.gov.au/websitedbs/D3310114.nsf/ Awty, I (2009), Taking the guess work out of
home/Quality:+The+ABS+Data+Quality+Framework feeding cows, available at: http://sydney.edu.au/
agriculture/pal/documents/Symposium_2009.pdf.
Australian Bureau of Statistics (2013), Statistical
Data Integration, available at: Bem, C (2013), Automated cattle system nears
http://www.abs.gov.au/websitedbs/D3310114.nsf/ automation, available at: http://www.nintione.com.
Home/Statistical+Data+Integration au/sites/default/files/upload/automated_cattle_
management_system_nears_commercialisation_-_
Australian Centre for Intellectual Property in engineers_australia_-_february_2013.pdf
Agriculture (2015), ACIPA, available at:
http://www.acipa.edu.au/events.html Boughton, B (2016), Using data to take proactive
action on barnyard grass, Satamap Blog, available
Australian Computer Society (2015), Australias at: http://satamap.com.au/blog/
digital pulse: key challenges for our nation digital
skills, jobs and education, available at: Brown, DJ, Savage, DB, Hinch, GN & Semple, SJ
http://www.acs.org.au/__data/assets/ (2012), Mob-based walk-over weights: similar to
pdf_file/0006/69720/02062015-Australias-Digital- the average of individual static weights?, Animal
Pulse-FINAL.PDF Production Science, vol. 52, pp. 6138.
Australian Government Solicitor (2009), Canegrowers Australia (2013), About Smartcane

Freedom of Information Guidelines: Exemption BMP, available at: https://www.smartcane.com.au/
Section in the FOI Act, available at: aboutBMP.aspx
https://www.ag.gov.au/RightsAndProtections/
FOI/Documents/Disc%20from%20PMC%20 Climate Corporation, The (2015), Tillage and crop
website%20info%20-%20FOI%20-%20 residue, available at: http://www.climate-help.com/
hc/en-us/articles/204356695-Tillage-and-crop-
residue
CHAPTER 8: References 65
Commonwealth of Australia (2006), Davies, B (2015), Seizing the economic potential

Census and Statistics Act 1905, available at: of big data, available at: http://www.bhert.com/
https://www.comlaw.gov.au/Details/C2015C00247 events/2015-02-17/Big-Data%20Round-Table%20
-Report-Feb-2015.pdf
Commonwealth of Australia (2015a), Privacy Act
1988, available at: https://www.comlaw.gov.au/ Deere & Company (2015), Privacy & Data,
Details/C2014C00076 available at: https://www.deere.com.au/privacy_
and_data/privacy_and_data_us.page
Commonwealth of Australia (2015b), Regional
Telecommunications Review, available at: Delaney, K (2015), The digital homestead taking
http://www.rtirc.gov.au/wp-content/uploads/ grazing into the future, available at:
sites/2/2015/10/RTIRC-Independent-Committee- https://futurebeef.com.au/resources/projects/
Review-2015-FINAL-Low-res-version-for-website. completed-projects/the-digital-homestead-taking-
pd grazing-into-the-future/
Cook, SE & Bramley, RGV (2000), Coping with Deloitte Access Economics (2015), The connected
variability in agricultural production implications continent II How digital technology is
for soil testing and fertiliser management, available transforming the australian economy, available at:
at: http://thirdworld.nl/coping-with-variability- http://www2.deloitte.com/au/en/pages/economics/
in-agricultural-production-implications-for-soil- articles/connected-continent.html
testing-and-fertiliser-management
Drone Deploy (2015), Instant Aerial Data for
Dairy Australia Limited (2014a), Precision dairy Agriculture, available at:
technology: automatic milking systems, available https://www.dronedeploy.com/ag
at: http://www.dairyaustralia.com.au/~/media/
Documents/Animal%20management/Technologies/ Eastwood, C (2008), Innovative precision
Dairy%20shed%20technologies/Precision%20 dairy systems: a case study of farmer learning
dairy%20technology%20factsheet%20%20 and technology co-development, available
automatic%20milking%20systems%20%20final.pdf at: https://minerva-access.unimelb.edu.au/
handle/11343/35094
Dairy Australia Limited (2014b), Precision dairy
technology: walkover weigh scale systems, Edwards, J, Umbers, A & Wentworth, S (2012),
available at: http://www.dairyaustralia.com.au/~/ GRDC Farm Practices Survey Report 2012,
media/Documents/Animal%20management/ available at: http://www.grdc.com.au/Resources/
Technologies/Animal%20technologies/ Publications/2012/11/GRDC-Farm-Practices-
Walk-over%20weigh%20scales%20case%20 Survey-2012
study_Mark%20Billing.pdf
Farmlink (2015), True Harvest by Farmlink, available
Dairy Australia Limited (2015), How precision at: http://www.farmlink.com/Home/TrueHarvest
dairy technologies save farmers time and money,
available at: http://www.dairyaustralia.com.au/ Farmobile (2015), Simplicity PUC, available at:
Standard-Items/Media-Releases/2015/March/11-03- https://www.farmobile.com/
How-precision-dairy-technologies-save-farmers-
time-and-money.aspx Future Dairy (2009), Milking robots: making
the most of money invested, available at:
Dairy New Zealand (2014), New Zealand Farm http://www.futuredairy.com.au/media/NEWS_-_
Data Code of Practice, available at: Milking_robots_-_DEC_08.pdf
http://www.farmdatacode.org.nz/wp-content/
uploads/2014/06/farm-data-code-of-practice-final.pdf Grain Growers Limited (2015), Agriculture
Technology Survey 2015, available at:
66 CHAPTER 8: References
http://www.graingrowers.com.au/policy/micro- IP Australia (2013), Confidentiality/Trade Secrets,

surveys Other Types of IP, available at:
http://www.ipaustralia.gov.au/get-the-right-ip/
Griffith, C, Heydon, G, Lamb, D, Lefort, L, Taylor, other-types-of-IP/confidentiality-trade-secrets/
K & Trotter, M (2013), Smart farming: leveraging
the impact of broadband and the digital economy, Janzen, T (2015a), Big data on the farm: legal
CSIRO and the University of New England, issues surrounding farm data ownership, transfer,
available at: https://publications.csiro.au/rpr/ and control, available at:
download?pid=csiro:EP137659&dsid=DS2 https://westlegaledcenter.com/program_guide/
course_detail.jsf?videoCourseId=100071078
Hale Group, The (2014), The digital transformation
of row crop agriculture, available at: Janzen, T (2015b), Defining big data in
http://www.iowacorn.org/documents/filelibrary/ agriculture, Janzen Ag Law Blog, available at:
membership/agstate/AgState_Executive_ http://janzenlaw.blogspot.com.au/2015/02/defining-
Summary_0A58D2A59DBD3.pdf big-data-in-agriculture.html
Henry, D, Shovelton, J, Charles de Fegely, C, Janzen, T (2015c), Does your co-op own your
Rod Manning, R, Lee Beattie, L & Trotter, M data?, Janzen Ag Law Blog, available at:
(2012), Potential for information technologies to http://www.aglaw.us/janzenaglaw/2015/11/27/
improve decision making for the southern livestock does-your-co-op-own-your-farm-data
industries, available at: http://www.mla.com.au/
CustomControls/PaymentGateway/ Janzen, T (2015d), John Deere does not own
ViewFile.aspx?s47X+13EFCGleRCu7R7N your tractor, Janzen Ag Law Blog, available at:
KoIRqa7hq8ZhkVN5mvcwK2A9sQMPX/ http://janzenlaw.blogspot.com.au/
GH5agC29L6MCGd3EYMKKAfsht7d1Tnt3BqiA
== Jochinke, DC, Noonon, BJ, Wachsmann, NG &
Norton, RM (2007), The adoption of precision
Hortus Technical Services (2016), Services, agriculture in an Australian broadacre cropping
available at: http://public.hortus.net.au/Services.aspx system challenges and opportunities, available
at: https://www.google.com.au/url?sa=t&rct=j&q=
House of Representatives Standing Committee &esrc=s&source=web&cd=2&ved=0ahUKEwiTu
on Social Policy and Legal Affairs (2014), vqvlNHLAhWk2KYKHZ1nBNEQFggiMAE&url
Eyes in the sky: Inquiry into drones and the =https%3A%2F%2Fwww.researchgate.net%2Ffile.
regulation of air safety and privacy, available at: PostFileLoader.html%3Fid%3D54d26eebd11b8b28
http://www.aph.gov.au/Parliamentary_Business/ 588b4644%26assetKey%3DAS%253A2736978756
Committees/House/Social_Policy_and_Legal_ 60817%25401442266066771&usg=AFQjCNGet1v
Affairs/Drones/Report 74zzAhMpF-ZnlhEGA9SHMhg&cad=rja
Jouven, M, Leroy, H, Ickowicz, A & Lapeyronie,
IHS Inc. (2013), Big data in the drivers seat of
P (2012), Can virtual fences be used to control
connected car technological advances, available at:
grazing sheep?, available at:
http://press.ihs.com/press-release/country-industry-
http://www.publish.csiro.au/paper/RJ11044.htm
forecasting/big-data-drivers-seat-connected-car-
technological-advance.
Lamb, D, Trotter, M & Schneider, D (2008),
GPS cattle tracking for understanding the impact
Ingraham, N (2015), Apple has sold 700 million
of grazing on grain and graze rotations and for
iPhones, 25 million Apple TVs, available at: http://
improving pasture utility, available at:
www.theverge.com/2015/3/9/8164357/apple-watch-
http://sydney.edu.au/agriculture/pal/documents/
event-700-million-iphones-sold
Symposium_2008.pdf
CHAPTER 8: References 67
Murphy, S (2013), Animal Liberation activists Presidents Council of Advisors on Science and
launch spy drone to test free-range claims, available Technology (2014), Big Data and Privacy:
at: http://www.abc.net.au/news/2013-08-30/drone- A Technology Perspective, available at:
used-to-record-intensive-farm-production/4921814 https://www.whitehouse.gov/sites/default/files/
microsites/ostp/PCAST/pcast_big_data_and_
NBN Co. Limited (2015), NBN Satellite privacy_-_may_2014.pdf
Broadband, available at: http://www.nbnco.com.au/
connect-home-or-business/information-for-home- Queensland Government (2016), Reef protection
or-business/satellite.html initiatives for cane farmers and graziers, available
at: https://www.qld.gov.au/environment/agriculture/
New South Wales Young Lawyers Civil Litigation sustainable-farming/cane-farmers/
Committee (2010), The Practitioners Guide
to Civil Litigation, 3rd edition, available at: Reimsbach-Kounatze, C (2015), The proliferation
https://www.lawsociety.com.au/cs/groups/public/ of big data and implications for official statistics
documents/internetyounglawyers/026375.pdf. and statistical agencies: a preliminary analysis,
OECD Digital Economy Papers, No. 245, OECD
Ninti One (2013), Precision Pastoral, available Publishing, http://dx.doi.org/10.1787/5js7t9wqzvg8-
at: http://www.nintione.com.au/our-work/ en, available at: http://www.oecd-ilibrary.org/
commercialisation-and-ip-management/precision- science-and-technology/the-proliferation-of-big-
pastoral-pty-ltd data-and-implications-for-official-statistics-and-
statistical-agencies_5js7t9wqzvg8-en
OECD (2011), Quality Framework and Guidelines
for OECD Statistical Activities, available at: Robertson, MJ, Llewellyn, RS, Mandell, R,
http://www.oecd.org/officialdocuments/ Lawes, R, Bramley, RGV, Swift, L, Metz, N &
publicdisplaydocumentpdf/?cote=std/ OCallaghan (2011), Adoption of variable rate
qfs(2011)1&doclanguage=en fertiliser application in the Australian grains
industry: status, issues and prospects, Precision
OECD (2013), Exploring data-driven innovation Agriculture: An International Journal on Advances
as a new source of growth: mapping the policy in Precision Agriculture, April, vol, 13, issue 2,
issues raised by big data, OECD Digital Economy pp.18199, available at: http://link.springer.com/
Papers, No. 222, OECD Publishing, article/10.1007%2Fs11119-011-9236-3
http://dx.doi.org/10.1787/5k47zw3fcp43-en,
available at: http://www.oecd-ilibrary.org/science- Rodgers, W & Thornton, M (2015), American Farm
and-technology/exploring-data-driven-innovation- Bureau Appeals Flawed Privacy Decision, available
as-a-new-source-of-growth_5k47zw3fcp43-en at: http://www.fb.org/newsroom/news_article/286/
Office of the Australian Information Commisioner
Rowe, J & Banks, R (2015), Sheep industry
(2015), Australian Privacy Principles Guidelines,
productivity the role of genomics and digital
available at: https://www.oaic.gov.au/agencies-and-
data, Farm Policy Journal, vol. 12, no. 1, Autumn
organisations/app-guidelines/
Quarter, Australian Farm Institute, pp. 2131,
Open Agriculture Data Alliance (2014), Principals available at: http://www.farminstitute.org.au/
& Use Cases, available at: http://openag.io/ publications-1/farm-policy-journals/2015-autumn-
about-us/principals-use-cases/ from-little-data-big-data-grow/sheep-industry-
productivity-the-role-of-genomics-and-digital-data
Pinsent Masons (2014), Connected cars
phenomenon raises data ownership and liability Schimmelpfennig, D & Ebel, R (2011), On the
issues, says expert, Out-Law.com, available at: doorstep of the information age, recent adoption
http://www.out-law.com/en/articles/2014/february/ of precision agriculture, Economic information
connected-cars-phenomenon-raises-data-ownership- bulletin, EIB-80, Economic research service,
and-liability-issues-says-expert/ United States Department of Agriculture, available
68 CHAPTER 8: References
at: http://www.ers.usda.gov/publications/eib- Telstra Corporation Limited (2015), Our coverage,

economic-information-bulletin/eib80.aspx available at: https://www.telstra.com.au/coverage-
networks/our-coverage
Schroeder, B, Panitz, J & Baillie, C (2014),
Nutricalc: A web-based nutrient management Tesla Motors (2015), Customer Privacy Policy,
decision support system using the Six Easy Steps available at: https://www.teslamotors.com/en_AU/
approach, available at: about/legal
http://www.sugarresearch.com.au/icms_
docs/164357_NutriCalc_IS13020.pdf Trotter, MG, Lamb, DW, Donald, GE & Schneider,
DA (2010), Evaluating an active optical sensor for
Sedex (2015), Empowering sustainable
quantifying and mapping green herbage mass and
and ethical supply chains, available at:
growth in a perennial grass pasture, available at:
http://www.sedexglobal.com/wp-content/
http://www.publish.csiro.au/paper/CP10019.htm
uploads/2012/03/Sedex-information-brochure-
Feb-2012-13.pdf Tru-Test Limited (2016), Walk over weighing,
available at: http://livestock.tru-test.com/en-nz/
Snijders, C, Uwe Matzat, U & Reips, U-D (2012),
walk-over-weighing
Editorial, in Big data: big gaps of knowledge in the
field of internet science, International Journal of UNSW Innovations (2016), Easy Access IP:
Internet Science, vol. 7, issue 1, pp. 15, available Valuable research discoveries, inventions and
at: http://www.ijis.net/ijis7_1/ijis7_1_editorial.pdf intellectual property for free, available at: http://
www.innovations.unsw.edu.au/sites/default/files/
Sparkes, D (2015), Cane growers warned to join
documents/guidelines/Easy%20Access%20IP%20
Smart Cane BMP or face tough scrutiny from
for%20Business_Final2.pdf
Government over water quality, ABC Rural,
available at: http://www.abc.net.au/news/2015-12- University of Sydney: Faculty of Engineering
04/cane-growers-warned-government-will-target- & Information Technologies (2015), Robotics
those-not-in-bmp/7000846 and intelligent systems: the key to future growth
in farming, available at: http://sydney.edu.au/
SST Software (2015), Presentation to InfoAg 2015
engineering/research/robotics/agricultural.shtml
Conference, St Louis, Missouri, July, available at:
http://past.infoag.org/ University of Tasmania (2015), Sense-T:
Agriculture, available at: http://www.sense-t.org.au/
Stern, M (2015), Digital agriculture, available
at: http://www.monsanto.com/investors/ USDA (2015), USDA Natural Resources
documents/2015/2015-11-17_bie/digital-ag- Conservation Service web soil survey, available
stern_2015.11.17.pdf at: http://websoilsurvey.sc.egov.usda.gov/App/
HomePage.htm
Swain, D, Gregg, D, Bishop-Hurley, G, Trotter, M
& Petty, S (2013), On property benefits of precision US Farm Bureau (2014), Privacy &
livestock management, Meat and Livestock security principles for farm data, available
Australia, Final Report, available at: at: http://www.fb.org/tmp/uploads/
http://www.mla.com.au/Research-and-development/ PrivacyAndSecurityPrinciplesForFarmData.pdf
Search-RD-reports/RD-report-details/Productivity-
On-Farm/On-property-benefits-of-Precision- Zwaenepoel, P & Bars, JL (1997), Precision
Livestock-Management/408 Farming, available at: https://translate.google.com.
au/translate?hl=en&sl=fr&u=https://hal.archives-
Taleb, N (2013), Beware the big errors of big data, ouvertes.fr/hal-00461080/document&prev=search
available at: http://www.wired.com/2013/02/big-
data-means-big-errors-people/
The AFI Capital Fund
To secure the long-term future of the Australian Farm Institute
the Board has established a Capital Fund, the earnings of which
will be used to provide a stable funding base to the Institute.
Supporters of the Institute are invited to contribute to the fund.
The Australian Farm Institute Board of Directors has set a target

of $4 million for the Capital Fund. Achieving that target would
mean that sufficient financial resources would be available
annually to support a continuing program of agricultural trade
research, conferences and seminars to heighten awareness of the
need to achieve better trade outcomes for Australian agriculture.
The Australian Farm Institute is recognised by the Australian

Tax Office as an accredited research institute, which means
that contributions to the Institutes Capital Fund are fully tax-
deductible. The Capital Fund will be invested securely, and the
aim will be to grow the fund over time. The investment returns
arising from the Capital Fund will be used solely to fund multi-year
Institute research programs, under the direction and guidance of
the Institutes Research Advisory Committee.
Contributors to the Capital Fund will be permanently

acknowledged in all Institute publications, presentations and
on the Institutes website. Contributors to the Capital Fund in
excess of specific thresholds will also be recognised in other
ways, including complimentary copies of Institute reports
and publications, and complimentary attendance at Institute
conferences and events.
Contributing to the AFI Capital Fund is one very concrete way of

ensuring that credible and objective information is available on
these issues, and can be used to ensure policy-makers understand
that improved agricultural trade access is an urgent priority.
If you or your organisation can help with a financial contribution,

please contact the Institute.
Suite 73, 61 Marlborough St

Surry Hills NSW 2010
T 61 2 9690 1388
F 61 2 9699 7270
E info@farminstitute.org.au
W www.farminstitute.org.au
Membership of Australian Farm Institute
I would like to support the activities of the Australian Farm Institute by becoming a Member
Entitlements:
Farm Policy Journal subscription, Institute papers, Research reports, Institute newsletter and access to Institute seminars
(Please mark the appropriate category one only)
Individual Membership A$330* annually (Incl. GST)
Corporate Membership (Please contact the Institute)
Student Membership A$176* annually (Incl. GST)
Co-funders and Sponsors (Please contact the Institute)
Subscription to the Farm Policy Journal Purchase Online

Simply log on to
I would like to Subscribe to the Farm Policy Journal www.farminstitute.org.au
using your user name and
Entitlements:
password to purchase
Four editions of the Farm Policy Journal and Institute newsletter a membership or
subscription
(Please mark the appropriate category one only)
Hard copy A$220* annually, four issues (Incl. GST)
Electronic copy [PDF] A$187* annually, four issues (Incl. GST)
Institutional Publication Subscription from A$1210* annually (Please contact the Institute)
Institutional Journal Subscription from A$649* annually (Please contact the Institute)
* Subscription charges and membership fees are subject to change. Please contact the Institute for pricing of International memberships and subscriptions.
Title Mr / Mrs / Ms / Other: ________________________
Name __________________________________________________________________________________________________________
Position __________________________________________________________________________________________________________
Organisation __________________________________________________________________________________________________________
Address __________________________________________________________________________________________________________
Town/City __________________________________________ State ______________________________________________________
Country __________________________________________ Postcode ______________________________________________________
Telephone __________________________________________ Fax ______________________________________________________
Email __________________________________________________________________________________________________________
Occupation:
Farmer Student Agribusiness Academic Government Consultant Media R&D Other
Payment Options:
1. I have transferred the amount of $_______________ by EFT to: BSB: 032 002 Account number: 41 2830
Account Name: Australian Farm Institute Limited Bank: Westpac, Royal Exchange branch
Please use full name or invoice number (if applicable) as reference
2. Charge my credit card for the amount of $_______________ Visa Mastercard Diners Club American Express
Card No: ____ ____ ____ ____ / ____ ____ ____ ____ / ____ ____ ____ ____ / ____ ____ ____ ____ Expiry Date: ____ / ____
Card Holders Name: _________________________________________ Signature: ___________________________________________________

3. I enclose my cheque for $_______________ made payable to Australian Farm Institute Limited ABN 29 107 483 661
Return this order form and payment to:

Mail: Australian Farm Institute Limited, Suite 73, 61 Marlborough Street, Surry Hills NSW 2010 AUSTRALIA
Fax: 61 2 9699 7270
Apply online: www.farminstitute.org.au
Australian Farm Institute Publications
The Australian Farm Institute publishes research reports and discussion papers arising from the annual research
program. These publications aim to identify strategic farm sector issues and promote potential solutions to policy-
makers and agribusiness leaders. Research topics are initially formulated at the Institutes annual Agriculture
Roundtable Conference, which involves Australian farm and agribusiness leaders, researchers and policy-makers,
members of the Institute and guest speakers, as well as members of the Institutes Board, Research Advisory
Committee and staff.
Research Reports
The Implications of Digital Agriculture and Big Data for Australian Agriculture
Designing Balanced and Effective Farm Animal Welfare Policies for Australia
Optimising Future Extension Systems in the Australian Grains Industry:
Part 1: Background, Literature and Demographics of the Australian Grain Production Sector
Part 2: The Public and Private Sector Grain Advisory Systems in Australia
Part 3: I nternational Grains Extension Models and Future Directions for the Australian Grains Industry Extension System
Opportunities to Improve the Effectiveness of Australian Farmers Advocacy Groups
Assessing the Opportunities for Achieving Future Productivity Growth in Australian Agriculture
Is Counting Farmers Harder than Counting Sheep? A Comparison of the Agricultural Statistical Systems of Australia,
the United States and France
Does Australia Need a National Policy to Preserve Agricultural Land?
Transport Costs for Australian Agriculture
Farm Level Modelling of Greenhouse Emission Mitigation and Sequestration Options for the Australian Wool Industry
The Implications of the Australian Governments Carbon Farming Initiative for Beef Producers Private Sector
Investment in Agricultural Research and Development (R&D) in Australia
Growing Regional NSW: Policies to Revitalise the Non-Metropolitan Regions of NSW
The Implications of Greenhouse Mitigation Policies on the Demand for Agricultural Land
Towards a Better Understanding of Current and Future Human Resource Needs of Australian Agriculture
Making Decisions About Environmental Water Allocations
Essential Services in Urban and Regional Australia a Quantitative Comparison
Farm Policy Journal

Switching on farm innovation Vol. 13 | No. 1 | Autumn quarter 2016
Australian agriculture, pursuing premium markets or productivity Vol. 12 | No. 4 | Summer quarter 2015
Will consumers stop agricultural technology? Vol. 12 | No. 3 | Spring quarter 2015
Labour matters in Australian agriculture Vol. 12 | No. 2 | Winter quarter 2015
From little data big data grow Vol. 12 | No. 1 | Autumn quarter 2015
Brand Australia: does Australian agriculture need a national brand? Vol. 11 | No. 4 | Summer quarter 2014
Setting the score for agricultural competitiveness Vol. 11 | No. 3 | Spring quarter 2014
Can agriculture grow if food processing dies? Vol. 11 | No. 2 | Winter quarter 2014
Farmers fare well with better animal welfare Vol. 11 | No. 1 | Autumn quarter 2014
Do community perceptions of Australian agriculture really matter? Vol. 10 | No. 4 | Summer quarter 2013
Financing the future of farming Vol. 10 | No. 3 | Spring quarter 2013
Will the Murray-Darling Basin Plan improve with age? Vol. 10 | No. 2 | Winter quarter 2013
Will supermarkets save or enslave agriculture? Vol. 10 | No. 1 | Autumn quarter 2013
Can Australia become the food bowl of Asia? Vol. 9 | No. 4 | Summer quarter 2012
Farm subsidies to continue, despite the crisis? Vol. 9 | No. 3 | Spring quarter 2012
Will corporate agriculture swallow the family farm? Vol. 9 | No. 2 | Winter quarter 2012
Managing uncertainty in the worlds riskiest business Vol. 9 | No. 1 | Autumn quarter 2012
Food, fuel and climate change mitigation: how can agriculture do it all? Vol. 8 | No. 4 | Summer quarter 2011
A private future for food and fibre quality Vol. 8 | No. 3 | Spring quarter 2011
Foreign investment in Australian agriculture: myths and realities Vol. 8 | No. 2 | Winter quarter 2011
Can agriculture manage a genetically modified future? Vol. 8 | No. 1 | Autumn quarter 2011
Are farmers and miners saving the day? Vol. 7 | No. 4 | November quarter 2010
Growing more with less Vol. 7 | No. 3 | August quarter 2010
If I were the Australian Minister for Agriculture... Vol. 7 | No. 2 | May quarter 2010
Bushfire policy: Do we need more than just an ounce of prevention? Vol. 7 | No. 1 | February quarter 2010
New thinking about agriculture
ABN 29 107 483 661
Suite 73, 61 Marlborough Street, Surry Hills NSW 2010

T: 61 2 9690 1388 F: 61 2 9699 7270
E: info@farminstitute.org.au W: www.farminstitute.org.au

Big Data Report Web

Uploaded by

Copyright:

Available Formats

Big Data Report Web

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Big Data Report Web

Uploaded by

Copyright:

Available Formats

research report

The Implications of Digital

2016 Australian Farm Institute

A report prepared by the Australian Farm Institute

2016 Australian Farm Institute

Australian Farm Institute Limited

Suite 73, 61 Marlborough Street

All rights reserved

ISBN 978-1-921808-38-8 (Print and Web)

Design and Production: Australian Farm Institute

Printing: J.A. Wales Printers, Alexandria

List of boxes, figures and tables

Figure 1: Percentage of regional cropping areas on which variable rate

ABS Australian Bureau of Statistics NLIS National Livestock Identification System

GPS global positioning systems TDM total dry matter

GRDC Grains Research and Development UAV unmanned aerial vehicle

NDVI Normalised Difference Vegetation Index

2. Big data in the economy

NSW / VIC Slopes

SA / VIC Btown Wimmera

Yield gains achieved were between

Figure 2: Harvester fuel use map.

and employing data scientists to interrogate them

An opinion common to US industry representatives

To a degree the rate of progress in the initial

3.2 Extensive livestock 49% for animal production monitoring

A recent report by Meat & Livestock Australia Animal location

benefit for beef and sheep livestock enterprises Proximity logging

arising from the use of digital technologies (Henry Intra-animal monitoring

areas with the highest productivity benefits: Animal liveweight

of stock in a timely manner)

Branding percentage 65% 68%

Carrying capacity for cell (hd) 1000

Margin per kg ($) $0.30/kg $0.40/kg

Helicopter mustering costs ($) $30,000 $128,000 $96,000

Labour costs ($) $ 5600 $ 22,500

Source: Swain et al. (2013).

Figure 6: Remote livestock management

as branding percentages, carrying capacity, weight

indicators which can display weight and drafting

collection and its effective use have the potential

Remote and proximal sensors

Index (NDVI), an index of vegetation greenness

Virtual fences information services, for use on-farm, within

This is highlighted in the data displayed in Table 2

Australian pasture Australian pasture

Average milking machine

Milk (litres) harvested/

and livestock surveillance, and uses have even

The use of both fixed wing and hover drones is

An alternative is the use of high speed internet

Currently, high resolution satellite imagery is

Satellite imagery is expected to remain an

Supermarkets are also beginning to build supply

4. The market for farm data storage

providers developing comprehensive systems

Technological developments also created

Figure 15: Australian and New Zealand

SST Software products are currently used by

5. Agricultural data and privacy

Figure 1: Percentage of regional cropping areas on which variable rate

Figure 6: Remote livestock management

4. The market for farm data storage

Figure 15: Australian and New Zealand

5. Agricultural data and privacy

Town/City State ____________

Country Postcode ____________

Telephone Fax ____________

Card Holders Name: _ Signature: ___________