Data stewardship in the age of precision agriculture
At the University of Kansas in Lawrence from 23 to 26 May 2017, IASSIST 2017 treated data as a common language for research. Its archived programme, schedule, plenary sessions and talks on big data, deep learning, digital agriculture and global food security showed how disciplines can connect when terms, methods and responsibilities are clear. The same principle shapes precision agriculture, where useful decisions depend on evidence drawn from fields, machines, markets and laboratories.
Precision agriculture combines remote sensing, machinery telematics, soil probes, weather records, farm-management software and machine-learning models. Collecting those inputs is simpler than governing them. Stewardship covers the full data lifecycle: how observations are captured, described, validated, shared, protected, reused and retired. It asks who may use the information, under which conditions, and whether the claimed benefit is credible. Those choices determine whether a model earns trust or becomes another black box between the field and the manager.
For Australian producers, this is practical rather than abstract. Rainfall variability, long distances, export scrutiny and a mix of family-run and large-scale operations create different obligations. A stewardship approach turns scattered records into accountable data assets while preserving the local knowledge that no sensor fully captures. The goal is not to place every record in one national database. It is to make each decision traceable while allowing useful data to move at the speed of the farm.
Stewardship as a working discipline
A data steward acts as custodian of information that no single organisation can manage alone. A contractor may own the header, an agronomist may interpret the results, a farmer may own the land, and a processor may require traceability. The steward maps those roles, records who contributes what, and establishes decision rights for access, correction, publication and deletion. This role does not need to be a separate full-time job on a small farm; it can be a named responsibility shared by the grower, adviser and supplier.
Every stream needs a documented owner, purpose and basis for use. Metadata should identify the source, location, time, units, calibration, transformations and consent conditions. Provenance then records how a raw observation became a map, forecast or recommendation. This chain of evidence protects quality, supports reproducibility and lets an agronomist challenge an unusual reading instead of treating the dashboard as unquestionable.
Australian conditions make a single approach unsuitable. A wheat enterprise near Ginninderra, a cotton farm outside Moree, a horticulture business near Mildura and a sugar grower around Bundaberg may use different machines and buyers, even within one national market. Australian growers tend to be blunt about tools that add little: a service has to pull its weight. Clear data contracts and simple interfaces are part of stewardship because they determine whether information is used at all.
Standards that make agricultural data useful
Interoperability begins with the FAIR principles: data and metadata should be findable, accessible in an appropriate format, interoperable and reusable. Agricultural data often fails the first test in practice, not theory, because a field file carries cryptic names and a processor report carries no map reference. Minimum metadata, stable identifiers and documented quality limits make records intelligible across organisations and research projects.
Common APIs, controlled vocabularies and standard schemas allow machinery, weather services and farm-management platforms to exchange information without repeated manual work. A yield monitor, weather station and spray log can exchange records through documented services rather than email attachments. Standards such as OGC formats for spatial data and ISO metadata guidance provide useful foundations, while controlled lists for crops, soil types and treatments prevent “pasture” and “grazing pasture” becoming different answers to the same question. Semantic interoperability also gives machine-learning teams dependable inputs.
Work spanning agriculture and artificial intelligence benefits from shared agendas that connect domain experts with data scientists. Events represented by the conference programme can help expose vocabulary gaps before systems are built. This matters for global food security: a model that cannot explain which regions, seasons and management practices it covers may transfer poorly from a research site to a commercial farm.
Ownership, consent and fair value
Ownership and consent are related but distinct. A farmer may own raw observations, a contractor may own telemetry and a processor may hold derived quality records. Australian contract law and state privacy duties therefore matter, but there is no single farm-data rule that answers every sharing case. Agreements should state purposes, permitted users, duration, security, breach handling, export rights and deletion. A click on a service-terms page is rarely enough to settle those commercial and ethical questions.
Australian grain, cotton, sugar and livestock markets are strongly connected to export buyers, handlers and processors. Traceability requests can create value, yet contracts should also make clear who bears compliance costs and who can reuse insights. Family-run holdings can have less bargaining power than large corporate operations, so collective purchasing, cooperatives and research and development organisations can help turn fragmented demand into fair terms. As Australian growers often say, if a vendor cannot answer a straight question about data use, it is better to give it a miss.
Fair value covers more than price. It includes affordable access, the right to correct errors, portability, a meaningful role in benefit sharing and protection for commercially sensitive or culturally important information. Aboriginal and Torres Strait Islander communities also require recognition of Indigenous data sovereignty, including authority over knowledge about Country and culturally significant species. Open-data publication is useful only when consent, cultural rights and commercial obligations are respected.
From collection to accountable practice
Operational stewardship starts when data arrives. The archived IASSIST 2017 conference check-in guidance offers a useful parallel: attendees need to know where to go, what happens next and whom to contact, just as a dataset needs an agreed intake path. A practical intake record should name the supplier, purpose, format, owner, permitted uses and any restrictions before automated processing begins.
Quality assurance then checks missing values, duplicate observations, sensor drift, coordinate accuracy and unusual weather readings. Version control preserves transformations, while role-based access, encryption, audit logs and tested recovery routines protect the system. Incident plans should also identify who can suspend a feed or model during a cyber-security event. Machine-learning services need additional evidence: representative training data, known limitations, monitored drift and a route for human review. A deep-learning model that works across one wheatbelt may fail after a machinery change alters image quality or a dry season shifts the landscape.
Reliability must also account for distance. Remote properties in western Queensland or the Northern Territory can pass through mobile “black spots”, so a field application may need offline validation and edge processing rather than depend on a constant cloud connection. Time sensitivity adds pressure during planting, spraying and harvest, making clear escalation paths and retained audit records essential. The same discipline applies to model output: a recommendation needs its input version, model version, confidence and responsible decision-maker attached.
A stewardship model that fits Australian farm businesses
A workable model starts with an inventory of relevant records, then classifies each by sensitivity, ownership and intended benefit. Data-sharing agreements translate those classifications into access rules. Common formats and metadata make exchange possible, while consent logs, quality dashboards and security controls make that exchange accountable. An independent review can test whether stated controls match actual practice. The final step is a retention schedule that removes obsolete personal or operational data without erasing evidence needed for audits, research or legal compliance.
Governance should bring growers, agronomists, contractors, data engineers, legal advisers and cyber-security specialists into the same room. In Australia, state agriculture departments, universities, grower networks and organisations such as Grains Research and Development Corporation can provide regional context and extension support. A regional dairy or grain cooperative may have more influence than an individual farm when negotiating with a platform vendor, while specialist advisers can translate legal clauses into everyday operating choices.
Performance measures should cover more than hectares enrolled or dashboards delivered. Useful indicators include metadata completeness, duplicate rates, time to resolve access requests, model error by region, cyber incidents, farmer confidence and the share of value returned to data contributors. These measures connect technical stewardship to productivity, resilience and global food security, because trustworthy evidence improves decisions without shifting risks onto those least able to absorb them.
| Dimension | Conventional record-keeping | Stewardship-led precision agriculture |
|---|---|---|
| Purpose | Preserve information for later use | Support traceable operational and research decisions |
| Metadata and provenance | Minimal labels attached after collection | Source, time, units, transformations and quality recorded continuously |
| Access | Broad account permissions or vendor discretion | Role-based, purpose-limited and reviewable permissions |
| Consent | Implied through use of the service | Documented by data type, purpose and participant |
| Interoperability | Manual exports in proprietary formats | Standard schemas, APIs and controlled vocabularies |
| AI assurance | Vendor accuracy claim with little context | Validation, bias checks, drift monitoring and human review |
| Commercial control | Platform sets most conditions | Contracts cover reuse, portability, benefit sharing and exit |
| Retention | Data kept indefinitely by default | Time-limited retention linked to legal and operational needs |
Precision agriculture will keep adding sensors, autonomous machines and predictive models. Its durable value will come from evidence that can be traced, challenged and shared under agreed rules, making accountable data an agricultural asset as practical as the machinery that produces it.
At the Conference
What attendees experienced in Lawrence
Plenary Sessions
Keynotes from Daniel Reed on data, technology, and culture, and Jennifer Clarke on digital agriculture and the Midwest Big Data Hub.
Workshops
Full-day technical workshops on Tuesday, May 23. Morning sessions ran 9:00–12:00 and afternoon sessions 13:00–16:00. Laptops were required.
Social Events
An opening reception at The Oread, a banquet, and an optional post-conference tour of Kansas City including Crescent Moon Winery.
Venue & Accommodations
Where the conference took place
Kansas Union
University of Kansas campus, Lawrence. Main conference venue with check-in on the 4th and 5th floor lobbies.
The Oread
1200 Oread Avenue, Lawrence. Hosted the opening reception and offered a room block for attendees.
The Eldridge
701 Massachusetts Street, Lawrence. A partner hotel with a reserved room block for conference guests.
Springhill & TownePlace Suites
Marriott properties in Lawrence with room blocks reserved under the "KU IASSIST Conference" name.
Program Highlights
Sessions and activities
Getting Here
Lawrence, Kansas
Kansas Union · University of Kansas
1301 Jayhawk Blvd, Lawrence, KS 66045
Kansas City International Airport (MCI) is approximately 50 minutes by car. Lawrence Transit Routes 10 and 11 served the area ($1 exact change).
Plan Your Stay
Accommodation options that were available
The Eldridge
701 Massachusetts Street, Lawrence, KS 66044. Room block now closed.
The Oread
1200 Oread Avenue, Lawrence, KS 66044. Room block now closed.
Springhill Suites
Marriott property. Room block reserved under "KU IASSIST Conference."
TownePlace Suites
Marriott property. Room block reserved under "KU IASSIST Conference."