This chapter describes recent advances in digital technologies and analyses the drivers of digitalisation in the agriculture sector.
Digital Opportunities for Better Agricultural Policies
Chapter 2. Digital innovations and the growing importance of agricultural data
Abstract
The agriculture sector has a long history of innovating and adopting new technologies to increase productivity, manage risk and improve environmental, social and economic sustainability. The use of digital technologies and related innovation—by farmers and also by policy makers and administrators—is another step in this history, which offers new opportunities but also brings new challenges. The OECD’s Recommendation of the Council on Digital Government Strategies defines “digital technologies” as:
ICTs [information communication technologies], including the Internet, mobile technologies and devices, as well as data analytics used to improve the generation, collection, exchange, aggregation, combination, analysis, access, searchability and presentation of digital content, including for the development of services and apps. (OECD, 2014[1])
The definition encompasses existing information communication technologies (ICTs), many of which have been used in agriculture since their inception – for example, Landsat satellite data has been used to generate soil and land-use land cover maps, for global agricultural production monitoring and for GPS since 1972 (Leslie, Serbina and Miller, 2017[2]). In many cases, recent advances have substantially broadened the breadth, scale and immediacy of what these technologies are able to deliver. Advances in in situ and remote sensing technologies have greatly increased the spatial and temporal resolution of physical measurements, and allowed for low-cost, automated measurement of many aspects of agricultural production that were previously only able to be measured in a limited way – for example at discrete points in time by a human observer conducting a field visit. Advances in massive data acquisition, storage, communication, and processing technologies have enabled the rapid transfer of vast quantities of data which would not have been possible even a decade ago, and have greatly magnified the ability to process large datasets and to automate analytical processes with machine learning.
These technological developments have occurred in the context of evolution of local and global challenges facing the food system, including the increasing need to produce more food with fewer resources, leading to changes in policy objectives. Sustainability is not a new objective of agricultural policies; however, it is an objective which has been difficult to effectively integrate into the agriculture policy mix (OECD, 2017[3]; OECD, 2013[4]).
Agricultural policies co-evolve alongside technological progress; each both drives and is shaped by the other. Earlier waves of technological progress in agriculture introduced mechanisation, higher yielding and more resilient seed varieties, and the first foray into precision agriculture with the adoption of satellite-based GPS for farm machinery guidance. These earlier waves did in some cases make extensive use of data, for example in developing conventional breeding and genetic engineering. Building on these past advances, the current wave of technological progress centres on the creation, use, combination, analysis and sharing of agricultural and other data in digital format to improve the sustainability and productivity of agriculture and food systems.1 This chapter briefly summarises the key technological innovations in this most recent wave, as well as the key drivers for digital technology adoption in the agriculture sector.
2.1. Overview of recent and ongoing digital innovations for agriculture and food
A range of new technologies promise to improve efficiency and significantly impact business models in the agriculture sector. These technologies can be grouped according to their function in relation to data, broadly defined to include any piece of information available in machine language (Table 2.1). Key categories are data collection, data analysis, data storage, data management and data transfer and sharing. The category of data transfer and sharing includes technologies which use data transfer or sharing to facilitate other kinds of transactions, such as transfer of ownership or value, communication (between humans or digital devices) and digitally-delivered services.
Many of these technologies can be used directly by policy makers and administrators (section 2.2). Others (e.g. software for automating agricultural machinery) are unlikely to be directly used by policymakers and administrators, but are nevertheless relevant for improving policy-making because they are capable of producing, sharing, managing (e.g. securely storing) or analysing policy-relevant data. Moreover, policies can be designed with these technologies in mind: while this work does not focus directly on policies aimed at fostering adoption in the agriculture and food sectors2, agricultural and agri-environmental policies may nevertheless alter incentives for farmers and other actors to adopt certain technologies.
Some of the technologies listed in Table 2.1 have existed in some form for many years, but recent advances have greatly improved the ability to obtain, analyse, manage or transfer data that is relevant for agricultural policies, including by reducing the cost and increasing the speed of data collection, analysis and dissemination.
The sub-sections below provide an overview of key recent technological and institutional innovations, and identify some of the factors driving digitalisation in the agriculture and food sectors. Specific ways in which these trends can benefit policy-making, or the agriculture sector more broadly, are identified in subsequent chapters.
Table 2.1. Digital technologies for agriculture and food
a. Advances in sensor technology are comprised not only of advances in digital technologies, and in particular advances in the creation of wireless sensor networks, but also innovations in physics or chemistry. For example, advances in nanotechnology have been critical to the development of the most advanced physical sensors existing today. This project focusses on the digital components of sensor technologies and related services.
b. Serious games are publicly-available apps which seek to employ citizen effort for data collection or data processing. These apps have “a serious purpose but [include] elements of gamification (i.e., the addition of game elements to existing applications) to help motivate the volunteers (Bayas et al., 2016[5]). In the agricultural context, serious games have to date been used primarily for land use and land cover monitoring and classification.
c. Citizen science technologies are technologies which facilitate “public engagement and participation in science and innovation” (Daejeon Declaration, 2015[6])
d. In relation to technologies which automate agricultural machinery, such as automated milking systems, planters and harvesters, and irrigation systems, this project focusses on the sensor and software components and related services of these technologies.
e. See Box 2.2.
2.1.1. Global and local: Recent advances in remote sensing and edge-of-field monitoring
Much recent progress has been made in the use of satellite-based remote sensing to produce higher resolution (both spatial and temporal) and more accurate data products for agriculture. According to Atzberger (2013[7]) “[r]emote sensing data can greatly contribute to the [agricultural] monitoring task by providing timely, synoptic, cost efficient and repetitive information about the status of the Earth’s surface”. It can provide comprehensive information on crop acreage, biomass and yield, monitoring of stressors (e.g. drought) as well as precise information on farm management actions such as crop rotations, and structures such as farm buildings, fencing, conservation buffers etc.
Gholizadeh, Melesse and Reddi (2016[8]) provide a detailed survey (circa 2016) of the evolution of space-borne and airborne sensors which provide data for water quality assessment. Their analysis shows a steadily increasing spatial resolution (including the launch of multiple satellites over the period 2007-2014 which provide sub-metre resolution), as well as a steady decrease in the time between revisits from more than two weeks for most satellites in the period before 2000, to around 1-2 days more recently. Gómez, White and Wulder (2016[9]) similarly explain that, until recently, land cover maps generally were based on relatively coarse resolution data (>1km), but that now there has been a significant increase in capture of medium resolution (10m-100m) data by earth observation satellites. Pettorelli, Safi and Turner (2014[10]) refer to data provided by the EU Sentinel satellites as a “game changer” for global efforts to monitor biodiversity (on both agricultural and other lands). Bégué et al. (2018[11]), reviewing the potential for remote sensing to provide data on cropping practices, similarly note that the Sentinel satellites are expected to overcome previous limitations and constraints and improve the ability to detect small and fragmented land use types (e.g. irrigated areas) and to obtain regional and global data on soil tillage practices. Such advancements pave the way for increased use of satellite-based data products to provide field-level, landscape-level and even global data to improve agricultural policies in a variety of ways. Box 2.1 provides further detail on the EU’s Copernicus programme.
Box 2.1. The use of remote sensing by the European Union Joint Research Centre and the Monitoring Agriculture Resources (MARS) programme
The Sentinel satellites of Europe’s environmental Copernicus programme are used, among other things, to study changes in farming on a weekly basis, with a 10 metre resolution, and with a free and open data policy. The European Union Joint Research Centre (JRC) has been using satellite data for identification of information on crop areas and yields since 1988. Satellite data allow the observation of changes in land use: which crops are being grown, how well they are developing, etc. This data can be used to predict seasonal yield, and to support thinking about how to cope with low harvests in various places in the world. This includes crop yield forecasting, enabling early warnings of crop shortages and failure and to support aid for food insecure countries.
Increased accuracy of satellite data allows more effective and efficient management and monitoring of the Common Agricultural Policy (CAP). The increased capacity of satellites allows improved remote monitoring of agriculture, with measurement of field areas, identification of crop types, geo-location of landscape features and assessment of environmental impacts.
Various agencies throughout Europe (including in Spain, Lithuania, Greece, the United Kingdom, Serbia, Belgium, the Czech Republic, Slovenia, Romania and the Netherlands) are testing the potential of such data to simplify processes and streamline monitoring. Monitoring previously covering only 5% of producers can now be extended to 100%, potentially changing the policy design and implementation of the CAP (see Sen4CAP project for the modernisation and simplification of the CAP in the post-2020 timeframe). For a further discussion of use of remote sensing for CAP administration, (See Box 3.7 in section 3.2.5.
Source: JRC, Copernicus and Sen4CAP projects.
Advances in unmanned aerial vehicles (UAVs, drones) and remote sensor design have also dramatically reduced the cost and improved the efficacy of airborne remote sensing. This has opened up a new field: the use of UAVs for conservation3 (also referred to as “drone conservation”) (Koh and Wich, 2012[12]). Airborne remote sensing is also becoming increasingly important as a source of data for high resolution mapping (e.g. land cover and land use, elevation, soils, watersheds, etc.), particularly for remote areas and for areas with high cloud cover which impede some kinds of satellite-based sensors. UAVs also offer the opportunity to capture better species-specific data relevant for biodiversity policies, by automating wildlife counts and greatly improving the accuracy and level of detail of biodiversity indicators (Hodgson et al., 2018[13]; Arts, van der Wal and Adams, 2015[14]). On-farm uses of UAVs as part of precision agriculture systems are also multiplying: for example, early evidence suggests that farmers are able to use drones to significantly decrease the cost of monitoring crop growth, increase data resolution and identify areas presenting potential problems (e.g. identification of low yield areas, earlier and quicker identification of pests or disease) (Jarman, Vesey and Febvre, 2016[15]; Hunt and Daughtry, 2018[16]).
Rapid technological advances have also occurred in edge-of-field monitoring (EOFM). For application in water quality, Daniels et al. (2018, p. 5[17]) note that within a relatively short time period, EOFM “has evolved as a research concept and tool to a routine practice to document runoff water quality on real, working farms”. Harmel et al. (2018[18]) provide a brief history of EOFM, noting that widespread use of electronic sampling devices began in the 1990s. Automated electronic sampling also emerged, and became more common in the 2000s. Current research is focussing on reducing the cost of sampling systems, making further practical improvements, and devising methods for measuring uncertainty. Daniels et al. (2018[17]) note that cost is still an obstacle preventing widespread adoption of EOFM by farmers and that there may be a role for government to provide financial assistance for EOFM; for example, the authors note that, recognising the value of EOFM for monitoring of the performance of on-farm conservation activities, USDA NRCS now offers cost share assistance for several EOFM activities.
2.1.2. Automating and accelerating analysis: The new capacity to harvest, combine and analyse data in agriculture and food
The use of digital data in agriculture was first introduced as a source of productivity growth through precision farming.4 At first, precision agriculture mostly involved the use of guidance systems, yield monitoring, variable rate application,5 long-distance transmission of computerised information (telematics) and data management (OECD, 2016[19]). A plethora of unrelated systems were developed to gather data about on-farm activities and performance such as yield variation and the characteristics of production assets.
Yet, while a large amount of data was being acquired and used for various specific purposes, much of it was not able to be combined with other data and was not readily re-usable beyond the initial intended purpose. Moreover, much agricultural data accessible by other actors such as governments, researchers and the public has been only in aggregated form; use of data at the level of the individual animal, field or farm has therefore been costly and limited.
One of the key reasons data has not been used to its full potential to date is that farmers often lacked the tools and skills to fully exploit data and use them for decision-making. The inability to link data across systems, each focussed on a specific task, prevented both insights into the relationship between certain management practices and within the farm system, at least in the absence of costly manual data synthesis. A single data point does not make much sense without a context, benchmarks, trends, or causal references. While this data can be individually informative, the insights obtained can be considerably multiplied if data of different types and from different sources6 can be combined.
Several technological innovations have recently significantly increased the capacity to collect, aggregate, process and analyse agricultural data: massive data acquisition, storage, communication, and processing technologies. These innovations allow the digitisation and datafication of agriculture:
Digitisation: the conversion of analogue data and processes into a machine readable format (OECD, 2019[20]). Many types of agricultural data were previously held in paper-based filing systems. Digitisation thus does not create new data, but rather by converting existing data into digital format allows data to be used and transferred in new ways.
Datafication: is the transformation of action into quantified digital data, allowing for real-time tracking and predictive analysis. Datafication takes previously unrecorded processes and activities and produces data that can be monitored, tracked, analysed and optimised (Naimi and Westreich, 2014[21]).
ICTs, including the Internet and the development of connected sensors which transform the analogue world into machine readable data, are increasingly leveraging large volumes of digital data. Datafication and digitisation have together not only rapidly expanded the volume of agricultural data recorded in digital format, but have also expanded data coverage to many aspects of farm production and associated variables of interest, including for public policies (e.g. discharge of waste, nutrients from farms) for which data was not previously available.
These large streams of data, and the capacity to combine them, are referred to as "big data” (OECD, 2015[22]).7 The access and processing of these large volumes, enabled by increased computing power, in turn enabling helps to infer relationships, establish dependencies, and perform predictions of outcomes and behaviours (OECD, 2015[23]), informing real-time decision-making.
Indeed, having more data is not enough. But combined with progress in communication and processing capacity, this data is progressively used to create knowledge and provide advice about production processes, and even to automate some activities on farm. This is referred to as actionable insights8 at the farm level (Figure 2.1): farmers can benefit from the knowledge created over time on their own farm but also by others, either peers or research and development institutions. Nevertheless, turning data into useful information generally requires models and algorithms, as well as knowledge about factors such as data quality and error tolerance for each data source. These provide the basis for new forms of knowledge, and new services and tools, with the potential to deliver significant change in agricultural practices as well as agriculture and food value chains (Wolfert et al., 2017[24]). This combination of precision farming with digitalisation has led to labels such as “farming 4.0” or “smart digital farming”.
The combination of data is further facilitated by cloud computing, which allows computing resources to be accessed in a flexible on-demand way with low management effort (OECD, 2014[25]). Cloud computing offers the capacity for the data to be stored and aggregated in locations other than where it is created or used, which supports big data analytics (OECD, 2016[26]).
Finally, all these innovations have underpinned advances in Artificial Intelligence (AI), defined as the ability of machines and systems to acquire and apply knowledge and to carry out intelligent behaviour (OECD, 2016[26]). AI helps computers interact, reason, and learn like human beings to enable them to perform a broad variety of tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, translation between languages, and demonstrating an ability to move and manipulate objects accordingly. Intelligent systems use a combination of big data analytics, cloud computing, machine-to-machine communication and the Internet of Things (IoT) to operate and learn (OECD, 2017[27]).
The availability of these new tools enable the creation of new information, and in particular, “actionable insights” not only for farmers but also regulators and policy makers who are increasingly demanding data to support policy-making is increasing, as governments move deliver “data driven” policies and services (OECD, 2014[1]), and see Case Study 8 for an example in Estonia).
Figure 2.1. Technological revolution for the production of actionable insights in agriculture
Note: This figure makes use of the term Artificial Intelligence (AI), introduced in 1956 and defined as a broad concept of machines being able to carry out tasks in a way that is considered “smart”. Recent advances in AI have been done through machine learning. Machine Learning is a current application of AI, according to which machines should get access to data to be able to learn for themselves. See Note 8 for a definition of “actionable insights”.
Reflecting the dynamic nature of many factors relevant to land management decisions, there is strong demand for up-to-date information. One particularly beneficial aspect of new data analysis tools is that they are often designed to be dynamic and updatable. These features lessen the need for constant investment in new hardware or software, and better match users’ needs. Therefore, tools that can allow for rapid update of information better match demand for information, and as such are likely to be used more, both now and in the future (source: Part IV, Case Study 1).
2.1.3. Advances in encryption, data protection and data sharing technologies, and institutions for data sharing
Advances in technologies for data access, management and sharing are changing the technical feasibility, costs and risks associated with access to and use of agricultural data. Key developments are:
Confidential computing and multi-party computation: “Confidential Computing” allows access to a proscribed set of analytics functions that are performed over encrypted data that is not disclosed to the data scientist or analyst. This enables a new, low friction, method of doing exploratory linkage and analysis of datasets (source: Case Study 6).
Synthetic data release: A recent advance in privacy technology is known as Differential Privacy. This is a quantifiable measure of the privacy of certain data analytics techniques that involve random perturbation of either the data being analysed or the analysis itself. Researchers are currently working on a variety of differentially private mechanisms to allow the release of synthetic unit record datasets that contain statistically similar data to the original data, but can guarantee that the released data cannot be re‐identified. These methods can allow the release of government datasets with fewer restrictions than are currently needed to ensure confidentiality. These techniques involve adding noise to the data, and so have some impact on the utility of the data for analytics (source: Case Study 6).
Advances in data visualisation software: recent years have seen the release of many different kinds of software which assist users to more easily customise visualisation of data. Many actors in the agriculture sector (including public agencies) are making use of such software to improve the usability of existing datasets and facilitate access by making access more “user friendly”. Examples include:
Institutions for accessing, managing and sharing agricultural data are evolving alongside the technological innovation described above. Institutional innovations are important pathways for ensuring that opportunities offered by technological innovation can be realised in practice. Key developments in recent years are:
Open data principles11
FAIR data principles (Case Study 6)
New arrangements for improving access to agricultural data held by public organisations (Box 2.2 and Case Study 6)
Interoperability and metadata standards (Case Study 1 and Part III)
New partnerships for co-innovation and collaboration in research and governance (Case Studies 1 and 9)
New models of collective governance for agriculture and for data (Case Study 2)
Digital property rights and data access rights.16
Given that policies are themselves institutions, and moreover ones which can in turn shape other institutions (for example by creating or protecting property rights, incentivising collaboration, setting a regulatory framework for data access), many of the examples given in this report, particularly via the case studies, are examples of how institutional innovation is enabling governments to make better use of digital technologies, or enabling others to do so. Further discussion of such institutional innovations are provided throughout this report.
Box 2.2. Advances in arrangements for access to agricultural data held by public organisations (Case Study 6)
Technological solutions have been developed over many years to enable more data to be available for use, such as anonymisation and data obfuscation techniques. There are also a large number of newer approaches to confidentialisation to facilitate data sharing for research while protecting privacy or meeting confidentiality requirements. All of these have been used in successful, large scale implementations in Australia and internationally (O’Keefe and Rubin, 2015[28]; Reiter and Kohnen, 2005[29]):
De-identified open data access – the analyst downloads the data directly (e.g. datasets accessible via the GODAN initiative1)
User agreements for offsite use (licensing), in which users are required to register with a custodian agency, and sign a user agreement, before receiving data to be analysed offsite.
Remote analysis systems, in which the analyst submits statistical queries through an interface, analyses are carried out on the original data in a secure environment and the user then receives the (confidentialised) results of the analyses.
Virtual Data Centres (VDCs), which are similar to remote analysis systems, except that the user has full access to the data, and are similar to on‐site data centres, except that access is over a secure link on the internet from the researcher’s institution (e.g. the USDA-ERS data enclave platform provided by NORC;2 Australian Bureau of Statistics DataLab3). VDCs may also make use of containerisation, where the analyst can access the data in a limited way, on a secure platform through a containerised application (e.g. the SURE platform used by the Sax Institute4).
Secure, on‐site data centres, in which researchers access confidential data in secure, on‐site research data centres (e.g. the Secure Access Data Center, France5).
Each arrangement makes data available at a specified level of detail, where sensitive detail can be reduced by methods including removal of identifying information; confidentialisation of the data by one of a range of methods, including aggregation, suppression or the addition of random “noise”; or replacement of sensitive variables or data with synthetic (“made‐up”) data.
Notes
1. The Global Open Data for Agriculture and Nutrition (GODAN) initiative promotes the “the proactive sharing of open data to make information about agriculture and nutrition available, accessible and usable”. GODAN promotes data sharing both within and across national borders. See https://www.godan.info/, accessed August 2018.
2. “The [United States] Economic Research Service (ERS) and the National Agricultural Statistics Service (NASS), in coordination with the Food and Nutrition Service (FNS) utilise the [university of Chicago’s NORC] Data Enclave to provide authorised researchers secure remote access to data collected as part of the Agriculture Resource Management Survey (ARMS), the primary source of information to the US Department of Agriculture and the public on a broad range of issues about US agricultural resource use, costs, and farm sector financial conditions.” See http://www.norc.org/Research/Projects/Pages/usda-ers-data-enclave.aspx, accessed August 2018.
3. “The DataLab is the data analysis solution for high-end users who want to undertake interactive (real time) complex analysis of microdata. Within the DataLab, users can view and analyse unit record information using up to date analytical software with no code restrictions, while the files remain in the secure ABS environment. All analytical outputs are checked by the ABS before being provided to the researcher.”
http://abs.gov.au/websitedbs/D3310114.nsf/home/CURF:+About+the+ABS+Data+Laboratory+%28ABSDL%29, accessed August 2018.
4. SURE is “Australia’s only remote-access data research laboratory for analysing routinely collected [health-related] data, allowing researchers to log in remotely and securely analyse data from sources such as hospitals, general practice and cancer registries.” See https://www.saxinstitute.org.au/our-work/sure/design-and-functionality/, accessed August 2018.
5. See https://www.casd.eu/en/, accessed September 2018. This is the channel for accessing agricultural micro-level data in France, including FADN data, but also surveys of farm practices. The CASD has been in place since 2012 and contains various types of sensitive data (e.g. health, taxation, business surveys, and administrative data such as agri-environmental measures).
Source: Case Study 6, Part IV.
2.1.4. The drivers of digitalisation of the agriculture and food sectors
The increased capacity to capture, manage and draw insights from data has the potential to disrupt the organisation of the food system, from influencing the supply and use of inputs in agriculture as well as the way agricultural products are supplied and valued downstream in the value chain. Digital data and technologies can enable better management of farms, agricultural productivity and resource use (on-farm drivers). Digitalisation of agriculture and farms is occurring across a broad spectrum, from low-tech solutions using mobile devices and platforms to provide management decisions services, to high-tech “digital farms” making use of integrated systems involving in-field sensors and internet of things (IoT); big data analytics for decision making; and drones, robotics and artificial intelligence (AI) for the automation of processes. The need for investments at the farm level varies enormously depending on the type of services required, which in turn depend on the type of production system and farm: for example, large extensive livestock producers do not have the same constraints and needs as hydroponics fruit and vegetable producers, or subsistence farmers. Regardless, all can benefit from new services. Whether investments in technologies are made on-farm or by service providers, the main reasons why farmers make use of digital technologies is that these technologies reduce costs or answer a new need in a changing environment
This increased capacity benefits both the agriculture sector itself and also upstream and downstream sectors. Agricultural big data can support real time farm management, a range of added-value services, and automation capabilities which in turn further support the improvement of agricultural processes (Sonka and Cheng, 2015[30]).
On-farm drivers for digitalisation of agriculture
The agricultural digital transformation potentially supports:
Improved agricultural productivity and sustainability.
Better risk management, including to adapt to or mitigate the impacts of climate change.
Improved access to markets and business management.
Improved management of administrative processes.
These processes need not require large on-farm investments: a mobile phone and a camera can be enough to provide services such as remote identification of pests. Many initiatives also currently rely on remote sensing, in particular satellite data. Satellite data are increasingly precise and the price of the information they create is decreasing. They also have the advantage of global coverage, homogeneous data and repeated observations creating historical data. Satellite data are already integrated in many systems, meaning that entry costs are low.
However, satellite data has its limits, in particular for provision of local services reaching farmers in an intelligible way. For service provision, satellite data will often have to be combined with other data or sensor systems. In particular as the level of precision is still not refined enough at the farm level. It is also necessary to pre-analyse satellite data to allow their use for analytics services and to reach and be useful to farmers (Case Study 8). An example of a satellite based system is the EU's Copernicus programme (Box 2.1).
Digital innovations can also indirectly affect farms’ sustainable productivity. For instance, big data analytics increase the capacity of scientists to engineer plants resistant to drought or to certain pests, reducing their need for water or the use of chemical inputs, and increasing resilience of farmer’s production to such exogenous events.12 Such indirect pathways are acknowledged but not discussed further in this report.
Off-farm drivers for digitalisation of agriculture
There are also off-farm drivers such as increased demand for information on agri-food products from consumers, and the need to adopt technology on-farm in order to participate or remain competitive in increasingly digitised global value chains. Demand for farm-related data all along the value chain is increasing, both from the public and private sectors. Several downstream factors pushing agricultural producers toward increased digitalisation can be identified.
The first set of factors relate to value chain management and trade requirements. Digital technologies can support the creation and management of a “data cycle”13 from farm to fork, where information is passed on by all actors in the supply chain, allowing for full traceability. Access to farm data can also improve efficiency in the management of trade regulations, particularly when trade systems are administered through the adoption of paperless trade and electronic documents (OECD/WTO, 2017). In particular, automatic recording of farm data (e.g. agriculture practices, provenance etc.) online can provide important information for customs processes and speed up clearance at the border. Overall, this can increase market access for agriculture producers and reduce trade costs.
The second set of factors relate to consumer demand and government implementation of agriculture policies. Newly created information or increased access to information can create new sources of value related to reputation and responding to consumer preferences. Food safety is one of the most important quality attributes for consumers and the effect of an outbreak on the reputation of a food processor or retailer can be lasting (Jouanjean, 2012[31]). The food industry is exploring use of digital technologies, in particular distributed ledger technologies (blockchain), to maintain secure digital records and improve traceability. The objective is to revamp data management processes across a complex network that includes farmers, brokers, distributors, processors, retailers, regulators, and consumers, to facilitate investigations into food-borne illnesses. Investigations can take weeks and can have dramatic consequences; digital technologies such as blockchain could reduce that time to seconds. There are also opportunities related to the use of other quality attributes beyond food safety for the creation of niche markets (see Jouanjean (2019[32])).
2.1.5. Adoption may be hampered by lack of skills; but what and whose skills?
It is often mentioned that farmers and advisors may not have the skills to use digital technologies or the full understanding of their potential uses (OECD, 2018[33]). It is undeniable that there is a difference in accessibility between generations, with a gap between the younger generations raised in the new digital era and older generations. However, the question of adoption is not necessarily a question of farmers’ skills to use digital technologies themselves, or of technical understanding of how technologies work. Many digital tools for agriculture are platforms or applications which rather require an understanding of social media and awareness and trust about all the possibilities offered by such platforms. Such platforms are being used in developing countries by populations with low levels of formal school education.
The issue of what level of understanding is required is also relevant for high-tech digital tools. Digital technologies are an aid to decision-making and may even allow for automation of decisions on farm. This may entail farmers delegating parts of the knowledge and decision making on-farm to the technology, which is to say to those who programmed and created the technology. While farmers need not understand all the technical elements of technologies, they need enough understanding to be able to manage them effectively on their farms. For example, when using precision agriculture machinery, while farmers’ understanding may not necessarily extend to being able to perform maintenance themselves (e.g. on precision agriculture machinery), farmers need to be able to understand the technologies’ functions and how to make use of digital elements such as yield maps, fertiliser or pesticide application regimes produced by precision agriculture machinery. They may also need to know how to use automation programmes (e.g. irrigation schedulers, robotic planters, harvesters). Understanding is also important for acceptance of recommendations: otherwise technologies may appear to be a “black box” and farmers may not act on recommendations due to a lack of confidence or trust.
References
[14] Arts, K., R. van der Wal and W. Adams (2015), “Digital technology and the conservation of nature”, Ambio, http://dx.doi.org/10.1007/s13280-015-0705-1.
[7] Atzberger, C. (2013), “Advances in Remote Sensing of Agriculture: Context Description, Existing Operational Monitoring Systems and Major Information Needs”, Remote Sensing, Vol. 5/2, pp. 949-981, http://dx.doi.org/10.3390/rs5020949.
[5] Bayas, J. et al. (2016), “Crowdsourcing in-situ data on land cover and land use using gamification and mobile technology”, Remote Sensing, http://dx.doi.org/10.3390/rs8110905.
[11] Bégué, A. et al. (2018), “Remote Sensing and Cropping Practices: A Review”, Remote Sensing, Vol. 10/2, p. 99, http://dx.doi.org/10.3390/rs10010099.
[6] Daejeon Declaration (2015), Daejeon Declaration on Science, Technology and Innovation Policies for the Global and Digital Age, https://legalinstruments.oecd.org/Instruments/PrintInstrumentView.aspx?InstrumentID=335&InstrumentPID=389&InstrumentHID=0&Lang=en (accessed on 7 March 2018).
[17] Daniels, M. et al. (2018), “The utilization of edge-of-field monitoring of agricultural runoff in addressing nonpoint source pollution”, Journal of Soil and Water Conservation, Vol. 73/1, http://dx.doi.org/10.2489/jswc.73.1.1.
[36] Directorate-General for Parliamentary Research Services (European Parliament) (2018), Precision agriculture in Europe - Legal, social and ethical considerations, http://dx.doi.org/10.2861/278.
[8] Gholizadeh, M., A. Melesse and L. Reddi (2016), “Spaceborne and airborne sensors in water quality assessment”, International Journal of Remote Sensing, Vol. 37/14, pp. 3143-3180, http://dx.doi.org/10.1080/01431161.2016.1190477.
[9] Gómez, C., J. White and M. Wulder (2016), “Optical remotely sensed time series data for land cover classification: A review”, ISPRS Journal of Photogrammetry and Remote Sensing, Vol. 116, pp. 55-72, http://dx.doi.org/10.1016/j.isprsjprs.2016.03.008.
[18] Harmel, R. et al. (2018), “Measuring edge-of-field water quality: Where we have been and the path forward”, Journal of Soil and Water Conservation, Vol. 73/1, pp. 86-96, http://dx.doi.org/10.2489/jswc.73.1.86.
[13] Hodgson, J. et al. (2018), “Drones count wildlife more accurately and precisely than humans”, Methods in Ecology and Evolution, Vol. 9/5, pp. 1160-1167, http://dx.doi.org/10.1111/2041-210X.12974.
[16] Hunt, E. and C. Daughtry (2018), “What good are unmanned aircraft systems for agricultural remote sensing and precision agriculture?”, International Journal of Remote Sensing, Vol. 39/15-16, pp. 5345-5376, http://dx.doi.org/10.1080/01431161.2017.1410300.
[15] Jarman, M., J. Vesey and P. Febvre (2016), White Paper: Unmanned Aerial Vehicles (UAVs) for UK Agriculture: Creating an Invisible Precision Farming TechnologyWhite Paper, http://aerosociety.com/News/Insight-Blog/3698/RAeS-helps-set-the-drone-agenda (accessed on 20 September 2018).
[35] Jensen, T., R. Vatrapu and N. Bjørn-Andersen (2018), “Avocados crossing borders: The problem of runaway objects and the solution of a shipping information pipeline for improving international trade”, Information Systems Journal, Vol. 28/2, pp. 408-438, http://dx.doi.org/10.1111/isj.12146.
[32] Jouanjean, M. (2019), Digital opportunities for trade in agriculture and food sectors [TAD/TC/CA/WP(2018)4/FINAL], http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=TAD/TC/CA/WP(2018)4/FINAL&docLanguage=En.
[31] Jouanjean, M. (2012), “Standards, reputation, and trade: evidence from US horticultural import refusals”, World Trade Review, Vol. 11/03, pp. 438-461, http://dx.doi.org/10.1017/S1474745612000274.
[12] Koh, L. and S. Wich (2012), “Dawn of drone ecology: low-cost autonomous aerial vehicles for conservation”, Mongabay.com Open Access Journal -Tropical Conservation Science Mongabay.com Open Access Journal -Tropical Conservation Science Tropical Conservation Science, Vol. 55/522, http://www.tropicalconservationscience.org.
[2] Leslie, C., L. Serbina and H. Miller (2017), “Landsat and Agriculture—Case Studies on the Uses and Benefits of Landsat Imagery in Agricultural Monitoring and Production”, https://pubs.usgs.gov/of/2017/1034/ofr20171034.pdf (accessed on 21 February 2018).
[38] Mcfadden, J. et al. (2019), “Development, Adoption, and Management of Drought-Tolerant Corn in the United States”, No. EIB-204, U.S. Department of Agriculture, Economic Research Service, http://www.ers.usda.gov (accessed on 10 May 2019).
[21] Naimi, A. and D. Westreich (2014), “Big Data: A Revolution That Will Transform How We Live, Work, and Think”, American Journal of Epidemiology, Vol. 179/9, pp. 1143-1144, http://dx.doi.org/10.1093/aje/kwu085.
[20] OECD (2019), “Vectors of digital transformation”, OECD Digital Economy Papers, No. 273, OECD Publishing, Paris, https://dx.doi.org/10.1787/5ade2bba-en.
[33] OECD (2018), Summary Record: OECD Global Forum on Agriculture, 2018. Digital technologies in food and agriculture: Reaping the benefits, OECD, http://www.oecd.org/tad/events/gfa-18-summary-record.pdf (accessed on 21 September 2018).
[3] OECD (2017), Agricultural Policy Monitoring and Evaluation 2017, OECD Publishing, Paris, http://dx.doi.org/10.1787/agr_pol-2017-en.
[27] OECD (2017), Going Digital: Making the Transformation Work for Growth and Well-Being.
[37] OECD (2016), Big Data: Bringing competition policy to the digital era, OECD, Paris, http://www.oecd.org/daf/competition/big-data-bringing- (accessed on 21 September 2018).
[19] OECD (2016), Farm Management Practices to Foster Green Growth, OECD Green Growth Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264238657-en.
[26] OECD (2016), “The Internet of Things: Seizing the Benefits and Addressing the Challenges”, OECD Digital Economy Papers, No. 252, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jlwvzz8td0n-en.
[22] OECD (2015), Data-Driven Innovation: Big Data for Growth and Well-Being, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264229358-en.
[23] OECD (2015), The role of new data sources in greening growth - the case of drones, https://www.oecd.org/greengrowth/ggsd-2015.htm#documents.
[25] OECD (2014), “Cloud Computing: The Concept, Impacts and the Role of Government Policy”, OECD Digital Economy Papers, No. 240, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jxzf4lcc7f5-en.
[1] OECD (2014), Recommendation on Digital Government Strategies - OECD, http://www.oecd.org/gov/digital-government/recommendation-on-digital-government-strategies.htm (accessed on 11 September 2018).
[4] OECD (2013), Policy Instruments to Support Green Growth in Agriculture, OECD Green Growth Studies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264203525-en.
[28] O’Keefe, C. and D. Rubin (2015), “Individual privacy versus public good: protecting confidentiality in health research”, Statistics in Medicine, Vol. 34/23, pp. 3081-3103, http://dx.doi.org/10.1002/sim.6543.
[10] Pettorelli, N., K. Safi and W. Turner (2014), “Satellite remote sensing, biodiversity research and conservation of the future”, Philosophical Transactions of the Royal Society B: Biological Sciences, http://dx.doi.org/10.1098/rstb.2013.0190.
[29] Reiter, J. and C. Kohnen (2005), “Categorical data regression diagnostics for remote access servers”, Journal of Statistical Computation and Simulation, Vol. 75/11, pp. 889-903, http://dx.doi.org/10.1080/00949650412331299184.
[34] Solanki, M. and C. Brewster (2013), Consuming Linked data in Supply Chains: Enabling data visibility via Linked Pedigrees, https://pdfs.semanticscholar.org/81ef/499243b9d091dbb9d3c618419ff9fe5230ea.pdf (accessed on 5 March 2018).
[30] Sonka, S. and Y. Cheng (2015), ““Precision Agriculture: Not the Same as Big Data But…”, (5):206,”, Farmdoc Daily, Vol. 5/206.
[39] UNECE (2011), The Data Pipeline: Discussion paper for the Global Trade Facilitation Conference 2011, UNECE, https://www.unece.org/fileadmin/DAM/trade/Trade_Facilitation_Forum/BkgrdDocs/UNPaper_DataPipeline.pdf (accessed on 10 May 2019).
[24] Wolfert, S. et al. (2017), “Big Data in Smart Farming – A review”, Agricultural Systems, Vol. 153, pp. 69-80, http://dx.doi.org/10.1016/j.agsy.2017.01.023.
Notes
← 1. It is acknowledged that the current wave of technological progress also has some very important non-data-centric technological advances (e.g. advances in gene editing technologies). While such advances may rely on data and make use of digital tools, they are not the focus of the report.
← 2. For example, innovation policies for agriculture and digital hard infrastructure policies aimed at bolstering the development of digital technologies in the agriculture sector (e.g. broadband).
← 3. Koh and Wich (2012[12]) principally discuss the use of drones for conservation of endangered species, which is relevant to a variety of landscapes, including agricultural landscapes. For conservation more generally in agricultural lands, UAVs can be used not only for monitoring threatened species (and biodiversity more generally), but also for diverse activities such as monitoring erosion, water bodies in agricultural catchments, spread of invasive species.
← 4. Precision farming uses geographical information systems (GIS) data, soil information, as well as information on weather and environmental conditions at the field level to optimise the management of the production process (this involves the choice of crop, when and how to apply inputs on the crop e.g. pesticides, fertilisers, water management, seeding rates and when to till or harvest the crop).
← 5. Variable Rate Application (VRA) (also “Variable Rate Technology”) refers to the application of a material, such that the rate of application is based on the precise location, or qualities of the area that the material is being applied to. VRA can be Map Based or Sensor Based.
← 6. There are many categories of agricultural data, including: “agronomic data, financial data, compliance data, metrological data, environmental data, machine data, staff data, personal data, and operational data (employee data, usage data related to inputs such as fertiliser, and other mapping, sensor and related data created or needed to operate including raw data, field data and experimental data).” (Directorate-General for Parliamentary Research Services (European Parliament), 2018, pp. 14-15[36])
← 7. While many definitions of “big data” exist, the term generally refers to (1) the large dimension of datasets; and (2) the need to use large scale computing power and non-standard software and methods to extract value from the data in a reasonable amount of time. Big Data is often characterised with respect to the “4 Vs” of volume, velocity (of data collection and dissemination), variety and value. See, for example, OECD (2016[37]).
← 8. According to Technopedia: “Actionable insight is a term in data analytics and big data for information that can be acted upon or information that gives enough insight into the future that the actions that should be taken become clear for decision makers.”
← 9. See Tableau https://www.tableau.com/; Qlik https://www.qlik.com/us; Datawrapper https://www.datawrapper.de/; accessed March 2019.
← 10. See as ArcGIS https://www.arcgis.com/index.html, QGIS https://qgis.org/en/site/; MapInfo® https://www.pitneybowes.com/ca/en/location-intelligence/geographic-information-systems/mapinfo-pro.html; GRASS GIS https://grass.osgeo.org/, accessed March 2019.
← 11. Open data principles and digital property rights and data access rights will be discussed in a forthcoming OECD report on Regulatory aspects of data governance for the digital transformation of agriculture.
← 12. See, for example, Mcfadden et al. (2019[38]), which describes recent development, adoption and management of drought-tolerant corn hybrids in the United States. Most of the current drought-tolerant corn hybrids available in the United States were developed using molecular breeding, which makes heavy use of big data and computer modelling.
← 13. The concept of a “data pipeline” is often used in the context of value chains and cross-border trade logistics, as a way to make sure that information about products moves along with it throughout the value chain (see e.g. Solanki and Brewster (2013[34]); UNECE (2011[39]); Jensen, Vatrapu and Bjørn-Andersen (2018[35])).