Victoria’s hawkweed eradication program

An enduring alpine partnership

A volunteer seeks hawkweed in the 2012 detection experiment. (Photo by Roger Cousens)

A volunteer seeks hawkweed in the 2012 detection experiment. (Photo by Roger Cousens)

Three species of hawkweed – orange hawkweed (Hieracium aurantiacum), mouse-ear hawkweed (H. pilosella) and king devil hawkweed (H. praealtum) – have established small populations at the Falls Creek Alpine Resort and surrounding areas of Victoria’s Alpine National Park. This is not good news. These places are of national significance due, in part, to their unique flora and fauna. The hawkweeds’ invasive potential threatens the natural values of these areas. Hawkweeds’ ability to spread rapidly and disrupt floristic communities is a major concern (see the box ‘Invaders par excellence’). In New Zealand, these weeds have already dominated over 500,000 hectares of vegetation cover and here in Australia early modelling estimated the threat to agriculture at $74 million. The challenge of controlling an invasive species like hawkweed is enormous, requiring skills and analysis that lie across many groups. Here we set out a few of the roles that science has played in helping managers get on top of this situation. Bringing together that science has forged an enduring partnership that might serve as a model to future efforts aimed at controlling damaging invasive species. A Project Control Group has been established to oversee eradication programs for all three naturalised species (see ‘Project Control’) in Victoria. The challenge is to find and kill individual hawkweeds over the 900 hectares that is currently at risk of infestation. The terrain is often steep and occasionally dangerous, and even when hawkweeds are found and treated, any root fragments left behind can regrow. Since 2003, the Project Control Group has collected high quality data and this has allowed researchers to focus on answering key questions for eradication.

Where are the hawkweeds?

Given the large risk area for hawkweed infestations, the first research priority was to identify where hawkweeds are most likely to occur within the landscape. Hawkweed managers had the unusual advantage of knowing with some confidence where the main source of seed spread was likely to be: a Falls Creek garden bed in the case of orange hawkweed, and a large and dense infestation adjacent to the Rocky Valley Dam in the case of king devil. Researchers used local wind data to predict where seeds were likely to have travelled with respect to the source populations, initially relying on a single distance observation and later expanding to sophisticated simulations of seed travel tailored to hawkweed seed characteristics and the alpine landscape. Researchers also investigated what habitats would be suitable for hawkweed germination and establishment, via expert opinion and glasshouse experiments. Overlaid with dispersal predictions, managers were now armed with maps to help target their search efforts.

How hard should we look?

So we know where to look. But how much effort should we put into finding individual plants? Hawkweed surveys are typically carried out in teams, with people walking in a line formation, scanning the ground for the species.

 “The challenge of controlling an invasive species like hawkweed is enormous, requiring skills and analysis that lie across many groups.”

Receiving the University of Melbourne Staff Engagement Award (l-r): Roger Cousens, Nicholas Williams, Cindy Hauser, Neil Smith (DEPI) & Iris Curran (Parks Victoria).

Receiving the University of Melbourne Staff Engagement Award (l-r): Roger Cousens, Nicholas Williams, Cindy Hauser, Neil Smith (DEPI) & Iris Curran (Parks Victoria).

Not all ground provides the same search experience – steep slopes can slow searchers down, and hawkweeds will generally be easier to spot in grass than when the searchers are knee-deep in heath. Researchers developed a model to optimise resource allocation across a landscape, making use of the maps described above and this variable relationship between effort and detection This research was previously discussed in Decision Point #31. The research team have been supplying prioritisation maps based on this optimisation method to the Project Control Group annually since the 2011/12 summer search.

How ‘detectable’ are hawkweeds?

While orange hawkweeds have distinctive orange flowers, their non- flowering rosettes and the yellow-flowering king devil and mouse-ear hawkweeds can be mistaken for several other daisy species at first glance. Between identification challenges and the varied terrain of the park, there’s a risk that search teams might miss hawkweeds in spite of their efforts and experience. To learn more about the detection rates of ground searches, researchers and agency staff collaborated on two ‘hide and seek’ field experiments. Researchers hid glasshouse-grown hawkweeds and flower mimics in amongst different vegetation types, then searchers – government agency staff, contractors and volunteers – were asked to find them. Analysis of the data revealed enormous variation in the time-to- detect-hawkweeds depending on the circumstances. Flowering hawkweeds were often detected quickly but the presence of other yellow-flowered species did cause distraction, and non-flowering rosettes were more often missed than found, especially in dense vegetation types. The resulting detection models can be fed directly into the resource allocation optimisation, and are used by the Project Control Group to pre-emptively set and retrospectively understand the level of confidence they can have in the surveys they undertake.

The future of hawkweed eradication

Over the past three seasons the number of new plants found has remained low, despite increased and more strategically targeted surveillance efforts. An ‘eradograph’ confirms that surveys appear to be effectively limiting the orange hawkweed and king devil incursions, while known infestation sites are being controlled and their recurrence is declining.

A king devil hawkweed mimic. Why is it a mimic? See the story on page 16. (Photo by Roger Cousens)

A king devil hawkweed mimic. Why is it a mimic? See the story on page 16. (Photo by Roger Cousens)

As the list of known hawkweed sites has expanded throughout the program, the burden of regularly revisiting sites to ensure ongoing control is increasingly dominating the program’s budget. To address this challenge, researchers have recently developed statistical models predicting infestation recurrence, and are working on optimisations to effectively balance effort spent on revisiting these known sites against effort spent searching for not-yet-discovered infestations. In parallel, the Project Control Group has trialled new chemicals that are extremely effective in killing hawkweed in one application, thus reducing the need for revisits. Now that detection experiments have established that human searchers struggle to find non-flowering hawkweeds in thickly vegetated terrain, the Project Control Group is investigating whether a dog can be trained to detect hawkweeds as part of the surveillance and monitoring program. With their known heightened senses of smell, agility and ability to be trained, it is possible that a hawkweed detection dog could more quickly and effectively find plants at all growth stages, potentially even sniffing out underground rhizomes than human searchers would never find. Ridding the alps of hawkweeds is still some years away, but looks to be on track. The Project Control Group has responded swiftly and responsibly to the incursions, investing their resources and harnessing community support to seek out and destroy hawkweeds across land tenures, and most recently, assessing the potential of detection dogs. Their commitment to quality data collection and support of relevant research has generated new findings that could not be achieved by researchers alone. These findings have been effectively incorporated back into eradication operations to ensure effective targeting of the Project Control Group’s resources. These hawkweed invaders are up against a heck of a team!

More information: Cindy Hauser

We gratefully acknowledge the financial support of Parks Victoria through their Research Partners Program, the Victorian Department of Environment and Primary Industries, the NSW Office of Environment and Heritage, the Australian Weeds Research Centre and the Australian Research Council (via Linkage Project 100100441). You can access a list of this team’s publications at program/