In short
|

Alpine Ash forest after fire – photo by DEECA
Alpine Ash in Victoria
Alpine Ash (Eucalyptus delegatensis) is a tall, long-lived eucalypt that provides habitat for wildlife, stores carbon, protects water catchments and is valued for its timber. Native to the high country of Victoria, the Australian Capital Territory and New South Wales, with a subspecies in Tasmania, it typically grows at elevations between 900 and 1500 metres.
The Alpine Ash forests of mainland Australia were listed as an Endangered ecological community under the Commonwealth Environment Protection and Biodiversity Conservation Act 1999 in March 2026.
Alpine Ash forests are highly vulnerable to bushfires. Large, intense fires can kill entire stands of mature trees, while young Alpine Ash can take up to 20 years to reach reproductive maturity and produce seed. The species relies on the release of seed stored in the canopy to regenerate after fire, so the timing of repeat fires is critical. If another fire occurs before young trees can mature and produce enough seed, the forest may be unable to recover naturally.
Climate change is increasing the frequency and severity of bushfires, raising the risk that Alpine Ash forests will fail to regenerate naturally. Since 2000, around half of Victoria’s Alpine ash forests have been affected by fire.
To help secure the future of these forests, efforts are focused on:
- collecting and storing seed for future use
- direct seeding in areas affected by repeat fires
- monitoring flowering events to guide seed collection
- landscape assessments to inform where forest restoration programs are most needed post-fire.
These conservation actions aim to promote Alpine Ash regeneration into the future.
What we're doing
Detecting flowering events using a custom-built machine learning model
ARI has developed a machine learning model that uses satellite imagery to identify flowering events, making seed collection easier and more cost-effective.
Alpine Ash flowering is sporadic and difficult to predict, making seed collection challenging. Rather than flowering across the entire forest at the same time each year, flowering occurs in patches of forest at different times. The environmental conditions that trigger these events are also poorly understood.
Current detection methods rely on on-ground monitoring and tracking, followed by aerial surveys in small planes to manually record locations of flowering that may be suitable for seed collection. On-ground surveys are then needed to search these locations and confirm flowering activity, adding to the cost.
ARI’s machine learning model uses freely available satellite imagery to identify flowering events. With clear imagery (without cloud cover), it can search the entire range of Alpine Ash at intervals as short as five days. In contrast, aerial surveys can cover only a limited area on a small number of selected days which can be further limited by weather, resources and smoke from bushfires or planned burns.
The method can also analyse historical imagery, providing a large dataset to work with. When combined with environmental variables, this information could reveal underlying patterns and drivers that could not previously be detected.
Acknowledgments
This project has been delivered in partnership with the Victorian State Government and Forest Solutions.

Example of a positive detection of flowering during model validation – by Lachlan Francis
Research impact
Alpine Ash seed recovery crews are currently testing the new flowering detection technology and validating its accuracy with on-ground searches. As the team becomes more familiar with the technology, and on-ground searches validate findings, we may be able to find seed across wider areas and rely less on aerial spotting flights.

Alpine Ash flowering near Alexandra – photo by 3FB Aerworx for DEECA
What’s next
Future investigations might combine environmental variables with the extensive flowering dataset to improve predictions and explore how collected seed could be used more effectively to support the species’ recovery.
The flowering detection technology has also been trialled briefly on other eucalypts, including Mountain Ash (Eucalyptus regnans). However, applying it to Mountain Ash is more complex because these forests are much more diverse than Alpine Ash forests.
Rapid detection of eucalypt flowering could transform monitoring of nectar-feeding threatened species such as the Swift Parrot. Frequent, low-cost checks could help conservation surveys target the right places at the right times, when these species are most likely to be present.

Alpine Ash seeding from a helicopter post bushfire – photo by Forest Solutions
Publications
The journal publication of this research will be submitted in late 2026 for peer review.
For more information contact Lachlan Francis (ARI) (lachlan.francis@deeca.vic.gov.au) or Caitlin Cruikshank (DEECA) (caitlin.d.cruikshank@deeca.vic.gov.au)
Page last updated: 17/11/26