Agritech Summit Palmerston North New Zealand 8 December

Agritech. Summit Palmerston North, New Zealand 8 December 2017 #LPAg. Tech Every. THING is some. WHERE on the planet in space and in time

Jim Wilson - Ag Gateway • Digital agriculture • Perspective – • farmers want to profit, feed people, sustain environment, use historical data, actionable insight, equipment that works together • Developers want clear context, semantics, purpose. Standards help, but which ones. • Ag Gateway – build stuff and develop standards from the practice • Resources – ADAPT models for standardizing import data • Plugins to ”adapt” proprietary formats to ADAPT models. • Github repository for ADAPT framework

Jim Wilson • Collaboration – Ag Gateway with AEF – Ag Electronics Foundation • CSIRO collaboration to use O&M • Data claims • • Farmer -> 3 rd party, party analyzes data and delivers insight back 3 rd party profits from data by selling insights back. Farmers don’t trust 3 rd parties to deliver value and protect data Social dynamics, legal frameworks, tech challenges

Andrew Cooke - Rezare • Software product and service development, mathematical models for ag organizations in NZ, AUS, UK • Data use – on-farm strategic, tactical decisions from the data • Inputs – fertilizer, feed, seeds needs, timing, placement • Outputs – provenance, timing, productivity, quality • Initiatives • Code of Practice – transparency on rights and policies, security practices, accreditation (similar low uptake to AFB transparency calculator) • Farm Data Standards – common vocabulary targeting developers, analysts • Data. Linker Framework – datalinker. org “handshake” • Based on Oauth 2, data access agreement standards for easier data sharing connections, JSON-LD standard schemas & message API’s • INSPIRE catalog definitions

Peter Dalhaus • Given transparency, where do farmers see the value of sharing data to receive valuable insight that makes money? • Some ag industries are more connected to data than others (cotton, cane

Ian Yule – Massey University Centre Precision Ag • Information from data • E. g. yield effects from treatments – sensors have come a long ways • E. g. fertilizer sensor “Green. Seeker” not widespread adoption – needs path to insight • “Pasturemeter” optical sensor becoming more robotic to better fit farmer need • Cow monitoring – more comprehensive for individual animals – big boost from 5 g networking • Data and geospatial scale: PA is zone – paddock – farm level now but data collected at many different scales and needed at many scales for particular value • Food security & logistics & value chains, e. g. beef • Tech example – hyperspectral imaging for hill country soil fertility analysis and pasture PA at meter scale instead of extrapolating from handful of samples

Matt Flowerday – GPS-IT Drones • Agritech changes and advances • Tech firms moving into ag space • Data ownership a hot topic, but the issues (enforcement vs value) are not very clear.

Owen Dance – GS 1 NZ • • • Global barcode / RFID standards organization Item identification – data capture – data sharing EPC electronic product code – EPCIS data sharing system GTIN is a globally unique identifier with specific code lengths (12, 13, . . ) GS 1 -128 as much information as needed (ASCII encoding). Serial shipping container code – GS 1 Logistic Label GLN – Global Location Number (uses GTIN-13) Capture – e. g. fruit labels Databar: track individual produce, meat items globally All codes can be represented in RFID as well EU 1169 rules for “sold at a distance” food information transparency

Frank Bollen – Zespri International • • • Kiwi fruit exporter for NZ – 230 K tons / yr to 60 countries Spatial information in Kiwi fruit supply chain Fruit – tree – block (of shade cover) – orchard features Fruit -> bin -> box -> pallet tracking / quality checking through long-term cool storage then transport then distribution Goal to manage variability through supply chain to different market windows (e. g. Dry Matter or size) Biosecurity: e. g. spatial analysis of PSA infestations Provenance of fruit – how to connect producers and consumers, so far linking to grower regions rather than individual growers / orchards Some spatial analysis of correlations between size, yield, profit, other parameters such as canopy size Challenge still to link supply chain “performance” in detail back to orchard – case for consumer “feedback” not yet clear

Mark Neal – Dairy. NZ Big Data • Payoff curves for precision agriculture – hard to see some value in profits • Millions of cows generate millions of data points -> breeding improvements for new traits such as heat tolerance, milk type, urine concentration, grazing preference (hill vs flat) • Internet of Cow Things – GPS, accelerometers, mesh RFID, imaging • Virtual fencing and herding • Pasture condition vs yield (paddock performance) • Pixel level data doesn’t necessarily express the pasture potential yield, e. g. with more fertilizer. Need detailed actual yields vs observable properties. • Data landscape of funders, innovators, farmers, regulators, etc. • Data middleware – integration, fusion, analysis, metrics (e. g. performance gap) • Need for developers / innovators to provide farmer value (but that can be hard to determine or predict)

Jochen - NIWA • High resolution farm forecasts (1. 5 km) • System integrations, e. g. irrigation • Farm connectivity • Profitability of irrigation decisions • “Co-innovation”

Peter Dalhaus – Federation University • Online Farm Trials – federating farm trial data for all Australia • Soil health e. Library – community contributed soil test data (shared in de-identified form) • Interest in soil change over time

Sean Hodges – Horizons Regional Council • Hats on – public safety, regional planning, regulatory compliance, environmental monitoring • Lightbulb – regional data sharing • Councils need to use carrots more than sticks for standards compliance. Collaboration (e. g. between councils) is effective. • Value of FAIR principles for better outcomes, more reuse, e. g. can use common broker • Farm-scale data: land erosion management, farm mapping of nutrients, etc. so much farm data collected. Agreements affect what can be shared and with whom (e. g. real estate agents).

So, what about standards…

The Role of Standards • Standards are (persistent) agreements among people and their tools] • Shared value, compromise, consensus, trust • Information standards are agreements for exchanging / sharing data • Where is sharing needed? By whom? Under what conditions? • Who benefits, who loses from sharing data? • Where and when can agreements persist (self-interest, public interest, regulation)? • Agreements between friends, competitors, enemies. • Enforcement, compliance not foregone conclusions.

The Context of Standards - Sharing • Standards are necessary but not sufficient for interoperability • Physical sharing versus shared understanding – syntax is not enough, but machine understanding (expected behavior) hard to evaluate. • Sharing vs open sharing • Value of data – innate or comes from integration / application to decisions -> sharing • Competitive advantage of proprietary knowledge • Standards vs business value • What value does a farmer get from participation in “learning”? View of risk / reward applies to contributing to and/or using standards

The Role of Science and Technology • Geospatial information science matters • • MAUP Autocorrelation Geometry Projections • Lack of standards vs lack of technology • Technologically infeasible or not interoperable • E. g. yield uncertainty: failure to exchange yield measurement quality or infeasibility of calculating uncertainty itself. • Separation of concerns – architecture matters • • Interchangeability - interfaces Technological resilience - layers Functionality – distributed processing Transparency - comparability

The Social (Standards) Network • Agreements are made by people even if implemented by machines • Consensus “for” standards precedes consensus “on” standards • Playing fields may or may not be leveled (e. g. ease of implementation, regulatory hurdles) • “Coopetition” makes strange bedfellows • Perceived value may be a matter of time horizon • Perceived value also depends on legal / regulatory context • Consultant’s guide to standards: “sell the standard, then break the standard” • Hard to get innovation credit for following someone else’s standard • Hard to get academic credit for developing a (consensus) standard • Standards work done well, like diplomacy, is usually worth trying, whether or not specifically successful

Now what (for standards) • Pick a standards development organization… • For geospatial, geographic standards, agricultural features, various OGC domain working groups that host specific standards activities: • • Agriculture Domain Working Group (DWG) Sensor Web Enablement DWG / Sensor Model activity Ux. S DWG Hydrology DWG Geosciences DWG Met. Ocean DWG Geosemantics DWG • Ag Gateway ADAPT: • PAIL – irrigation data standards • SPADE – fertilization data standards • PICS – imagery tagging