Rangelands NRM Blog, News & Resources

Combined ground and aerial searches close to eradicating rubber vine

[Sept-Oct 2016]

The annual search for invasive rubber vine plants in the West Kimberley combining aerial and land searches has resulted in 1500 flowering vines being detected and destroyed.

The search is the continuation of the largest weed eradication program in Western Australia, combining ground searches, aerial scanning and the innovative use of crowdsourcing where volunteers detect plants in aerial images.

The project, led by John Szymanski together with Indigenous locals Kelvin Mitchelson and Dougie McCasker is supported by many partners, including the Department of Agriculture and Food WA (DAFWA), the Department of Parks and Wildlife, Kimberley Rangelands Biosecurity Association (RBG), Rangelands NRM through funding from the National Landcare Programme, local businesses and land managers and State NRM.

Through the Kimberley dry season, the team, including DAFWA’s Mick Everett, scanned the heritage-listed Fitzroy River Valley near Willare via helicopter as part of the ‘Astro Mission’, flying 265 square kilometres to detect unchecked rubber vine (Cryptostegia grandifloria) that has in the past become a monoculture with resultant loss of flora and habitat for fauna.

In addition to this, the ‘Aquila Project’ used a combination of high resolution imagery, advanced software processing and Internet volunteers to scan and identify the white flowers of mature rubber vine.

“With assistance from 244 Internet volunteers, we’ve proven that using cutting edge technologies can achieve really precise results in detecting plants in such a vast area,” project manager, John Szymanski said.

Following this, a total of 180 person days on the ground was completed, walking 4000 kilometres while searching 60 square kilometres at 10 and 15 metre spacing, including a grid searching of all return sites

Mr Szymanski said the field work has been a hard slog with long hours, walking an average of 25km a day.

“We are conducting precise grid searches 10 and 15 metres apart, adhering to a standard practice developed and proven over many years,” he said.

In 2015, the same search process by the Astro Mission returned 60 flowering plants, compared to only 15 vines detected this year. With the addition of the Aquila project, an additional 10 vines were identified, previously missed by human observers during the aerial search.

“A total of 1500 vines were found and destroyed, mostly by ground searching and almost entirely from known return sites,” Mr Szymanski said.

The project has established that the Probability of Detection (POD) by human observers in choppers is only 35 per cent, with a POD of Aquila using a single 29 megapixel camera is 60 per cent, indicating Aquila is more reliable than human observers.

Mr Szymanski is confident that fewer than 200 vines exist in the system of all plant classes, seedlings, juveniles, adults and breeders.

“We have two more years of ground searching in 2017 and 2018, prior to purely aerial monitoring from 2019, with the funding requirement reducing to a third from 2019 on,” he said.

The transition from human observers in choppers to using Aquila Methodology only from 2018 will result in halving chopper costs.

Mr Szymanski adds that doubling image resolution by adding a second camera will increase the POD of Aquila from 60 per cent to a predicted 86 per cent.

“Basically we are almost there—the end is in sight,” Mr Szymanski said.

Images: L: Graph showing Astro progress over the last 4 years / R: Only 6.30 am and already soaked (©Jenny Coleman)