Machine learning is being used to help farmers gain access to near real-time airborne fungal data.
BioScout technology helps provide proactive disease management by combining automated microscopy with machine learning.
The company's Head of Science Michelle Demers was a speaker at the 2024 GRDC Grains Research Update held in Adelaide, and was part of an Ag Tech Startups Forum.
Ms Demers, a trained plant pathologist, said disease was costly to farmers worldwide.
"On average, diseases destroy 20 per cent to 40pc of crops globally and this costs us over $220 billion annually on a global scale," she said.
"Unfortunately this is expected to get worse from numbers coming out of the CSIRO.
"Fungicide use is increasing and yet disease pressure is also increasing."
Ms Demers said disease in crops was a complex problem.
"One aspect of the problem is deciding when to apply fungicide," she said.
"Typically one of the things we'll do is apply at a key growth stage. The other time we'll use it is when the weather is favourable for disease.
"In order to have disease you need to have three elements happening at once: a susceptible host, the right environment or weather and you also need to have a pathogen present.
"So when you're spraying at key growth stages, you've covering that host element; when you're deciding to spray based on weather variables, you're covering the weather element.
"But we're missing something. We're effectively trying to manage disease based on only one or two elements.
"You're trying to make a decision on a complex problem while missing a huge chunk of information."
Ms Demers said a key issue with disease management was knowing which diseases were present in a given area before plants were symptomatic, particularly with spores of disease-causing fungi being largely invisible.
"The problem is this can lead to overuse of chemicals, spraying when you don't need to or spraying at the wrong time," she said.
Ms Demers said spray misuse could lead to loss of yield value, increased chemistry costs and reduced sustainability.
"BioScout's technology can detect pathogens in the air at almost real-time," she said.
"What we do is allow for fully-informed disease management."
SporeScout units photograph microscopic airborne particulates, analyse the imagery to identify and quantify fungal spores and scales the process through machine learning.
The unit is powered by a solar panel and a wind vane keeps the intake nozzle consistently pointed into the wind for optimal air sampling.
"It's a fully automated system so rain or shine it will give you data," Ms Demers said.
"The units don't need any servicing by you (the farmer) but we ideally need to come out about once a year."
Automated disease surveillance is currently available for broadacre pathogens including general rust, blackleg, general Alternaria, powdery mildew and Botrytis.
Other diseases like ascochyta will be added in the future, as well as a text messaging function.
"We train machine learning algorithms to act as pathologists," Ms Demers said.
"We'll have programs that will go through and identify certain types of spores, quantify them and put that data onto the dashboard."
The SporeScout units' dashboard has a landing page with a GPS view of a property, allowing users to see where the units are, and showing the concentration of spores for the different diseases being tracked over the last 24 hours.
The units are also complete weather stations, showing what weather variables happened during that time.
Ms Demers said it was highly scaleable technology.
"The more machine learning models we have for more spores, we can cover multiple different disease types in parallel, with essentially no additional time cost," she said.
Surveillance network
Starting in April, BioScout will be part of the nation's first dedicated airborne fungal pathogen surveillance network.
The network project involves deploying 60 SporeScout units across all three GRDC regions - north, south and west.
The strategic deployment of SporeScout units will be augmented by 4 iMapPESTS sentinels for DNA validation.
This work will lead to near real-time data on the presence and concentration of airborne fungal spores being available across vast grain growing regions.
There is also a research component, with several Spore Scout units in the network being placed within existing disease field research trials. The data generated from these trials will help provide recommendations on how to best incorporate BioScout data into existing integrated disease management practices.
Participation in the initiative will be available to GRDC-approved users free of charge until April 2026.
To express your interest in the project, visit www.bioscout.com.au/grdc to register.
Paula Thompson
Paula Thompson is ACM's National Machinery and Ag Tech writer. She has been an agricultural journalist for more than 20 years. Prior to her current role, she was a reporter with the Stock Journal, covering livestock markets and the grains industry.
Paula Thompson is ACM's National Machinery and Ag Tech writer. She has been an agricultural journalist for more than 20 years. Prior to her current role, she was a reporter with the Stock Journal, covering livestock markets and the grains industry.
