The slow start of a digital revolution is underway for rural real estate owners, with new technology from CSIRO bringing artificial intelligence to rural property sales, financial lending and public payments for landscape management.
To commercialise its new technology CSIRO has partnered with start-up company DAS to launch the the Rural Intelligence Platform, a subscription service that essentially uses modelling and machine learning to provide insights into a range of land values.
The platform uses satellite imagery to assess and monitor rural land all over the country, with modelling developed by CSIRO to detect changes over time in the images.
The artificial intelligence in the platform provides a risk rating for drought, frost, heat stress for livestock and so on. It can also monitor development of infrastructure on a property.
The platform uses satellite imagery, climate information and Australia’s digital soil map to track paddocks and their performance over time. It draws on publicly-available data to assess productivity, water access, yield, land use, crop type, rainfall and drought impact.
It could enable investors to assess property values, or by financial institutions to calculate the risk rating for loans to farmers, or even to monitor the outcomes for schemes where farmers get paid for things like soil carbon capture or revegetation.
CSIRO Agriculture and Food deputy director Dr Michael Robertson said it was unlikely investors, buyers and other users would rely solely on the platform, and traditional know-how, agronomy and other tools would be used as well.
He explained that machine learning is one the platform’s key components.
The platform’s proprietary model, dubbed the automated valuation model, makes mathematical estimations through machine learning to calculate the sale value of a farm.
“We’ve used historical sale value of farms and lined that up with the features of a farm, such as location, soil, climate, land use, production performance. Throw those into a computer and we can see which variable predict sale prices,” Mr Robertson said.
“We tested the accuracy of the model against actual sale prices and found its 90pc accurate, which is the same as the models for urban real estate, and that’s about as accurate as any machine learning can get.”
The platform could also be used by government or carbon reduction schemes to calculate payments for ecosystems services like soil carbon capture or re-vegetation.
“We’ve already had conversations with the market about the potential for this tool to support carbon sequestration or or greenhouse gas reduction schemes,” Mr Robertson said.
“We’ve also spoken to government about using it to target drought assistance, assessing needs down to a property-by property level.”
Australian Farm Institute executive director Richard Heath said while he was advocating for ecosystem services, technology like the Rural Intelligence Platform would be needed to measure land management outcomes.
“There are a lot of policy changes needed in Australia before we see the sorts of ecosystem-farm schemes like there are in Europe,” Mr Heath said.
“But if those programs did eventuate in Australia then this is the sort of platform needed to demonstrate outcomes to get payments or funding.”
Mr Heath said it was critical platforms such as DAS use good quality data.
“There is caveat over how good the data being collected is. If the platform is using erroneous information, or not truly reflecting the state on the ground, that could lead to negative sustainability outcomes.”
Mr Robertson said the platform was targeted at investors and buyers, not farm managers. However, it creates an opportunity to build trust between farmers and financial institutions by creating objective performance-measuring data.
“This platform is not designed for agronomic management, there are other tools for that,” Mr Robertson said.
“However, DAS and CSIRO are hoping to identify a business model to provide farmers access to the platform.”
Mr Heath said it is critical that technology like the Rural Intelligence Platform is accessible to all users to ensure the insights are not harnessed to favour lenders, buyers or even sellers over other parties.
Farm owners could look at this nervously, if it is used by funders and affects their willingness to lend and/or impose sustainability metrics on farms.
“It can and should be equal access and democratic so everyone gets access to it. Farmers need to make sure they use this data to their advantage, as well as other players,” Mr Heath said.
Melbourne-based DAS secured $4.25 million in funding from CSIRO, Ruralco and private investors.
“Digital agriculture is far more than just on-farm technology, it’s also about improving off-farm decision making and this platform lays the foundation,” said DAS chief executive Anthony Willmott.
