Earth’s doppelganger: scientists make $15m digital twin

Jennifer Dudley-Nicholson |

Satellite images and climate data will help create a digital replica of the agricultural landscape.
Satellite images and climate data will help create a digital replica of the agricultural landscape.

It will look and act like earth, but not exactly as we know it. 

The Australasian Space Innovation Institute announced plans to build a $15 million digital replica of the nation’s agricultural landscape on Tuesday, fuelled by satellite images, sensor information, and climate data. 

The project, backed by a university and agricultural firms, will be designed to cut millions of dollars from the cost of field trials and help to research issues such as biosecurity, climate disasters and water management.

Australia’s virtual copy will also use artificial intelligence to test scenarios and comes two months after the federal government released its National AI Plan encouraging greater investment in the field. 

A field of canola crops (file image)
The digital project’s tasks could include stress-testing crops and simulating natural disasters. (Mick Tsikas/AAP PHOTOS)

The National Digital Twin for Australian Agriculture, the first project overseen by the institute, will focus on using information from above to help agricultural projects below. 

Satellite data and space research collected in Australia was typically fragmented and handled on a per-project basis, institute chief executive Andy Koronios said, but could be collated for greater insights in the digital twin project.   

“Australia is investing more and more in its own space capabilities, satellite capabilities and so on, but we don’t do well in bringing that data for special applications on earth,” he told AAP. 

“What we need is a continuously learning environment that tests, that validates and scales a lot of the data that is available, not only from satellites but also from drones, from sensors on the ground, as well as climate data.”

The digital twin project will combine satellite images, farm models, weather, soil and water information to test innovations and scenarios in a virtual environment before they are tested in the real world. 

Scenarios could include stress-testing crops, mapping biosecurity outbreaks, simulating floods and droughts, and monitoring harvest strategies. 

Other nations had begun work on similar digital environments to support agricultural research, Mr Koronios said, and the project could help Australia remain competitive, 

“No country has a full digital twin system for agriculture yet, but countries like The Netherlands and the United States are working (on) them,” he said.

“We want to be building the sovereign capability here for the next generation of agricultural technology rather than depending on others.”

Cattle feeding (file image)
Meat and Livestock Australia believes the project may help save years of research time. (David Mariuz/AAP PHOTOS)

Charles Sturt University, the agricultural business Elders, and Meat and Livestock Australia will support the project’s development, which is expected to deliver research findings on a yearly basis. 

Being able to test hypotheses before committing to expensive trials could deliver significant savings in food production, Meat and Livestock Australia managing director Mick Crowley said. 

“This means we can test livestock management options and research questions faster, refine trials before we invest in large-scale field trials, adoption or commercialisation,” he said.

“Done well, this approach can save millions of dollars and years of research time compared with traditional methods.”

AAP