From Research to Policy: Data-Driven Decisions for Food Security

For decades, food security has sat at the intersection of paddock realities, lab findings, and cabinet-room trade-offs. In Australia, where the climate swings between flood and drought and where more than two-thirds of exports are tied to agriculture, the path from a research finding to a policy lever is rarely a straight line. The work of turning numbers into decisions is shaped as much by funding cycles, ministerial priorities, and the rhythms of Canberra as by the data themselves. Yet when the plumbing works, the results can be striking: better-targeted subsidies, faster drought relief, smarter water allocations in the Murray-Darling Basin, and more credible advice reaching the desks of people who need it.

The IASSIST 2017 gathering, held at the University of Kansas, brought together the people who build that plumbing. Data managers, social scientists, archivists, and policy analysts spent four days comparing notes on how information moves from sensor to spreadsheet to statute. Their theme, "Data in the Middle: The Common Language of Research," captured something the Australian food security community has been quietly wrestling with for years: the stubborn gap between excellent research and durable policy.

The Long Road from Field Notes to Cabinet Papers

A researcher in the NSW Riverina can spend three seasons collecting soil moisture readings, and a policy advisor in Canberra can ignore all of it because the dataset lives on a laptop in a shared drive nobody else can find. That mismatch is not unique to Australia, but it has a particular flavour here. Federal agencies such as the Australian Bureau of Agricultural and Resource Economics and Sciences publish rigorous forecasts, yet state and territory departments often operate on different cadences, formats, and definitions. A "small farm" in Tasmania is not the same administrative object as a "small farm" in Western Australia, and the language of "drought-affected" can mean anything from a single bad season to multi-year collapse.

What changes the dynamic is when research output is packaged for non-academic readers. Plain-English summaries, briefings of two pages or fewer, and visualisations that show a clear trend line rather than a confidence interval can do more to shift a policy position than a peer-reviewed paper ever will. The shift is partly cultural: researchers must accept that a busy decision-maker, often working after hours and reading on a tablet, will not engage with methodology in the body text. The shift is also structural, requiring dedicated translation roles inside research institutions.

Open Data and the Australian Research Commons

Australia was an early mover on open government data through data.gov.au, and agricultural datasets have been a steady feature of the portal. Yet availability is not the same as usability. A CSV file of rainfall observations across 800 stations is, in raw form, almost useless to a parliamentary committee that wants to know which electorates are most exposed. The data must be cleaned, normalised, and joined with socio-economic indicators before it tells a story. That work is unglamorous, underfunded, and almost always performed by the same handful of data wranglers in agencies such as CSIRO, ACIAR, and the Bureau of Meteorology.

The case for investing in this middle layer is straightforward. When a researcher at the University of Adelaide or the University of Queensland can pull together a regional food security dashboard in days rather than months, the policy cycle shortens. The conference programme in Lawrence dedicated several sessions to these hidden workflows, and the conference social events page captures some of the hallway conversations where data curators swapped notes on schema design and persistent identifiers. Those informal exchanges are often where the real standards get hammered out, before a formal working group is convened.

Bridging the Gap Between Agronomists and Policy Advisors

Agronomists speak in terms of yield gaps, cultivar performance, and nitrogen use efficiency. Policy advisors think in terms of cost-benefit ratios, distributional impacts, and electoral implications. Bridging the two vocabularies is a craft, and it is one that has not always been taught. The food security community in Australia has tried to address this through programmes like the Policy Fellowships run by the Australian Academy of Technological Sciences and Engineering, which place mid-career scientists inside departments for a year. The alumni of such schemes often become the people who can write a brief that survives a round of red-pen editing from the minister's office.

The risk is that translation work is treated as a soft skill, bolted on to a research career rather than central to it. Until data-to-policy translation carries the same weight as a journal publication in academic promotion, the pipeline will leak. Some institutions are starting to adjust: research outputs now include policy briefs, datasets, and code as legitimate deliverables. That shift is slow but visible, and it is one of the more encouraging signs for anyone worried about the slow uptake of evidence in food security planning.

The Role of Indigenous Knowledge and Community Data

Food security in remote Australia is not just a question of yield and trade. Aboriginal and Torres Strait Islander communities hold detailed knowledge of seasonal calendars, fire management, and the productive capacity of country. Too often, this knowledge has been excluded from formal datasets, sometimes because it does not fit Western scientific categories, sometimes because of historic mistrust. Reasonable efforts to address this gap are gaining traction, with bodies such as the Northern Land Council and the Indigenous Land and Sea Corporation supporting community-led monitoring programmes.

The data infrastructure for these programmes is different from a satellite feed or a yield survey. It often relies on relationships, oral history, and place-based indicators. A serious food security policy in Australia needs to make room for both, which means building platforms that can hold qualitative and quantitative data side by side, and that respect data sovereignty principles. The IASSIST community has begun to engage with these questions, partly because the same archival standards that protect census records can be adapted to protect community-held knowledge, provided the consent frameworks are right.

Measuring What Matters: Indicators, Models, and Uncertainty

A useful comparison of the data sources that feed food security policy looks something like this. Each has strengths, each has limits, and each carries a different kind of uncertainty that policy makers must understand before they act.

Source	Typical Use	Strengths	Limits	Uncertainty Profile
Satellite remote sensing (e.g., MODIS, Sentinel)	Crop area, vegetation health, drought monitoring	Wide coverage, frequent revisit, repeatable	Cloud cover, coarse resolution, mixed-pixel effects	Moderate; reduced by multi-sensor fusion
Household surveys (e.g., ABS SIH, FAO approach)	Income, food expenditure, dietary diversity	Direct measure of lived experience	Costly, infrequent, sampling bias	High in small regions, low at national scale
Agricultural census and farm surveys	Production, inputs, structural change	Comprehensive at sector level	Lagging, often voluntary, definitional drift	Low for totals, high for sub-sectors
Sensor networks and IoT (soil moisture, weather stations)	Real-time on-farm decisions, water budgeting	High temporal resolution, localised	Sparse coverage, maintenance burden	Low locally, high when extrapolated
Crowdsourced and community data	Locally observed crop loss, market prices	Granular, fast, contextually rich	Variable quality, sustainability of contributions	Variable; needs curation and validation

The lesson from years of food security research is that no single source is enough. A serious policy brief will draw on two or three of these streams, will name the uncertainty explicitly, and will avoid the temptation to present a model output as a fact. Australians, with their deep scepticism of spin and their preference for a fair go approach to evidence, tend to respond well to this kind of honest framing. Hiding the uncertainty is the fastest way to lose the room.

Putting the Pieces Together: A Practical Path Forward

The practical work of turning research into food security policy happens in small, repeatable steps rather than grand reforms. Three habits tend to separate the projects that influence policy from those that sit on a shelf.

• Build the dataset before the question is asked. Researchers who maintain a clean, well-documented dataset on, say, regional wheat yields over twenty years are ready when a minister's office calls for a quick brief. The opposite, assembling data under deadline pressure, almost always produces a shakier result.
• Write the one-page summary first, the paper second. Starting from the policy-relevant conclusion and working backward forces researchers to identify the single finding that matters. Everything else becomes supporting evidence.
• Cultivate the translator role. Every research team working on food security should include someone whose job is to engage with policy users, attend the right roundtables, and understand the rhythms of government.

A further four habits, often overlooked, make a real difference over a career.

• Keep the data alive. A dataset that is updated annually, with clear versioning, is far more useful than a frozen snapshot.
• Use shared identifiers. Persistent identifiers for places, crops, and indicators make it possible to join datasets years later, which matters more than researchers usually expect.
• Document the limitations openly. A dataset with a clear "what this cannot tell you" note is more trusted than one that overclaims.
• Stay close to the end user. A wheat breeder in Narrabri and a food policy advisor in Canberra need different versions of the same finding, and both versions matter.

The IASSIST 2017 conference did not solve the data-to-policy problem, and nobody expected it to. What it did was bring together the people who keep the data moving: the archivists, the curators, the metadata specialists, the ones who make sure a finding from 2017 in a Kansas seminar room can land cleanly in a brief going to a minister in 2025. For Australia, where the climate is shifting, the population is growing, and the export market for food is both an opportunity and a vulnerability, that invisible work is worth a great deal.