Grapevine trunk disease in the sights of researchers

A tool that can be used by growers or viticulturists to map and monitor grapevine trunk disease as part of normal vineyard operations is in the sights of an Australian research team.

The project – which is being undertaken by the South Australian Research and Development Institute (SARDI) in conjunction with the University of New South Wales – is investigating the use of in-vineyard imagery combined with algorithms to visually monitor and map eutypa dieback.

Eutypa dieback, previously known as ‘dead arm,’ is a fungal disease that eventually causes grapevine death – particularly in Shiraz, Cabernet Sauvignon and Grenache. The disease is spread by airborne spores that are released from dead and infected wood following rainfall. Grapevines can become infected by spores through fresh cuts in the vine – such as those made during pruning. The disease costs the Australian wine sector millions of dollars each year in lost production and additional vineyard management.

Dr Mark Sosnowski and his team at SARDI – who are international experts in trunk disease assessment and treatment – visually assessed 13,000 individual vines in the 10 vineyard study blocks in the McLaren Vale and Clare Valley regions of South Australia. Image data was then collected from all of the study blocks. The work was conducted in collaboration with agronomy business, DJ’s Growers in McLaren Vale and the Clare Valley Wine and Grape Association.

The first set of surveys were conducted using both traditional manual assessment and proximal imaging using a camera mounted on a tractor in October 2020. The researchers are now analysing the images.

‘We are currently matching an image of every vine with the manual assessment and are training algorithms to automatically assess the level of dieback in the vines', said lead researcher, Dr Paul Petrie (SARDI).

‘For the vines where symptoms are visually present, we can assign individual vine ratings and estimate the severity as well.’

Dr Mark Whitty, who leads the Smart Robotic Viticulture Group at the University of New South Wales, said the results would then now be transferred into a smart phone app.

‘This app is capable of continuously capturing data and generates a georeferenced map of disease incidence and severity across a block in real-time', he said.

The aim is to generate a tool that can be used by viticulturists or growers to rapidly quantify and locate dieback symptoms as part of routine vineyard operations, to better guide treatment regimes.