Our research

Project supported by the NSW Government’s Environmental Trust (Stage 3 Implementation Plan 2 – Biocontrol research for weed management, 1 July 2020 to 30 June 2023)

This project focused on the releases and evaluation of the approved biological control agent for wandering trad, the leaf-smut fungus Kordyana brasiliensis, in New South Wales (July 2020 to June 2023).

Key activities of the project were to:

• Supply material of the agent for community releases in NSW via post or at workshops.
• Establish monitoring plots throughout the range of the weed in NSW and gather baseline vegetation data.
• Assess establishment, spread and impact of the agent at monitoring plots one, two and three years post-release.

Establishing monitoring plots and baseline vegetation surveys

Study sites

Between July and October 2020, we established a total of 97 monitoring plots (dimensions: 5 m x 2 m) across 14 sites along a ~750 km latitudinal gradient in NSW, located between the townships of Dorrigo in the north and Bega in the south. Plots were located in subtropical and warm temperate wet sclerophyll forest communities on the coastal plains, foothills, and escarpment of the Great Dividing Range.

Monitoring plots were assigned to one of three experimental treatments: (1) biocontrol release plots dominated by wandering trad (50 plots); (2) experimental control plots dominated by wandering trad where the biocontrol agent was not released (27 plots); (3) plots dominated by native vegetation with only minimal level of wandering trad invasion, acting as a benchmark against which the recovery of desirable native vegetation in the biocontrol release plots will be assessed (20 plots).

Vegetation surveys

The composition of resident vegetation with each 10 m² plot was censused between July and October 2020, mostly when each plot was first established. Vegetation was identified to species level following the NSW Flora available through PlantNET. Abundance (% cover) for each species was estimated visually to the nearest 1 %. A total of 249 vascular plant species, consisting of 191 native and 58 alien species, were identified across the 97 monitoring plots prior to release of the biocontrol agent.

Native species richness was strongly negatively associated with wandering trad cover (see graphs below). However, there was an apparent negative threshold change in native species richness, such that richness values remained high as wandering trad cover increased to about 60 %, then decline dramatically when wandering trad cover exceeded 60 %. There was no significant difference in native species richness between low and medium wandering trad cover, yet richness declined significantly by 80 % on average where wandering trad cover exceed 60 %. There was also a significant decline in native species abundance with increasing wandering trad. However, the richness and abundance of other alien plant species were not associated with wandering trad cover.

Example of (a) plot arrangement at Kangaroo River, (b) plot configuration at Macquarie Pass and (c) a native reference plot at Yarranabee.

Effects of wandering trad cover on (a) native and alien species richness and (c) native and alien cover abundance (n = 93). Blue curves represent non-linear (i.e. negative quadrat function) lines-of-best fit for significant effects. The vertical black arrow denotes likely position of an invasion threshold (~ 60%) above which native species richness declines precipitously with increasing wandering trad cover. The histogram (plot b) represents mean (± SD) native species richness in three categories of wandering trad cover: low (< 20%, n = 16), medium (20 ≤ x < 60 %, n = 23), high (60 ≤ x ≤ 100 %, n = 54); letters above each bar denote significant differences amongst means based on Tukey’s Honestly Significant Difference test.

Biocontrol agent release

Releasing the biocontrol agent at monitoring plots

Between August and October 2020, wandering trad stems infected by the biocontrol agent were harvested from a nursery release site established in March 2019 at Brogers Creek, near the township of Berry on the South Coast of NSW. Each plot was divided into 10 contiguous 1 m² subplots, in which we planted five infected wandering trad stems (one in the centre, and four in each corner) by inserting at least one basal stem node into a small hole in the soil. This resulted in a total of 50 infected stems planted per plot.

Monitoring the biocontrol agent

We quantified the incidence and severity of disease symptoms caused by the biocontrol agent at each release plot approximately six months after the infected stems were first. Symptoms initially appeared as small chlorotic flecks that enlarge with time, forming whitish lesions on the underside of leaves and yellowish areas on the upper surface of leaves. As the lesions aged, they became reddish-brown and coalesce, eventually resulting in leaf death.

Disease incidence was calculated in two ways: (1) as the percentage of 50 stems showing signs of infection; and (2) the percentage of infected leaves per stem, as an average of the 50 randomly selected stems per plot. Disease incidence and severity, wandering trad cover and volume, and associated vegetation richness and cover were assessed on several occasions up to 2. 5 years post-release of the fungus.

Variation in disease severity over time

Kordyana brasiliensis disease severity varied significantly across time and was dependent on region and treatment. Maximum levels of disease severity were observed across the three northern regions after 18 months (stabilising at 24 months). The rate of increase in disease severity was significantly higher at the three northern regions, with maximum severity at the southern sites only reached after 24 months. Disease severity was always significantly lower in the south coast region. By 30 months, disease severity began to decline to a lower level in all regions, commensurate to a decrease in the cover of wandering trad (see below).

Kordyana brasiliensis disease severity over 30-months across four different regions (north to south: Coffs Harbour, Port Macquarie, Illawarra, and the South Coast). Points represent mean values and error bars represent standard errors from the predictions from the generalised linear mixed model.

Infection of wandering trad by Kordyana brasiliensis across monitoring sites 18-months post release. (A) Infection of a single leaf, (B, C, D) different levels of infection of stems across sites, (E) significant leaf death on stems at Bega Anabranch, (F) examples of leaf death and missing leaves observed on stems across sites, predominantly at sites in Port Macquarie and Coffs Harbour where disease infection was greatest.

Variation in wandering trad abundance over time

The change in wandering trad abundance over time varied significantly amongst the four regions. In this analysis, abundance was represented by wandering trad volume. In the two northern regions, wandering trad volume declined dramatically (approaching zero) in the biocontrol-release plots between 0 – 18-months post release, before stabilising between 18- 30- months after release of the fungus. In the two southern regions (Illawarra, South Coast), wandering trad volume increased significantly in the biocontrol release plots between 0 and 6-months after release before declining significantly in the Illawarra region. In the south coast region of NSW, wandering trad volume never decreased below the pre-release level.

This result highlights that host plant responses to infection by a biocontrol agent are context specific and can vary spatially, likely as a result of broad regional climate context (warm, humid in northern NSW and relatively cool and dry in the southern region).

Variation in wandering trad volume over 30 months in each of four regions; Coffs Harbour (n = 102), Port Macquarie (n = 86), Shoalhaven/Illawarra (n = 140) and the South Coast (n = 87). Points represent mean values and error bars represent standard errors.

Photo point monitoring showing a reduction in wandering trad cover at three plots, approximately 18-months following release of K. brasiliensis. Values represent the percentage reduction in wandering trad cover over time.

Relationship between wandering trad volume and disease severity

The percentage change in wandering trad volume was calculated between each monitoring period and the baseline volume that was measured immediately prior to release of the fungus. The magnitude of decline in wandering trad volume increased with increasing disease severity – i.e., the higher the disease severity, the greater the reduction in wandering trad abundance.

The effects of disease severity varied regionally, with the greatest reduction in wandering trad volume observed in the northern regions of NSW and the lowest rate of decline in the south coast region. There were impact thresholds of infection on wandering trad volume in the Illawarra and south coast regions, where wandering trad volume only declined where disease severity exceeded approximately 20 % and 40 %, respectively.

Relationship between percentage change in wandering trad volume and Kordyana brasiliensis disease severity over 30 months in each of four regions, Coffs Harbour (n = 63), Port Macquarie (n = 51), Shoalhaven/Illawarra (n = 88) and South Coast (n = 69). Points represent raw data and lines represent the predictions from the linear mixed model bound by 95 % confidence intervals. Zero values along the y-axis represent no change, positive values represent an increase and negative values represent a decline in wandering trad volume between each time point.

Community engagement and large-scale release of the biocontrol agent

Biocontrol agent release methods

Between October 2020 and June 2021, wandering trad stems infected by the biocontrol agent were delivered to registered participants. Stems were stored in plastic bags and posted to registered participants along with a booklet containing information about the biocontrol program and instructions on how to release the agent in a wandering trad infestation (instructions published online here: https://research.csiro.au/wandering-trad/wp-content/uploads/sites/173/2020/09/Wandering-Trad-Biocontrol_Release-Information_NSW-Enviro-Trust_07.09.20.pdf).

Participants were instructed to make a small hole in the soil, insert a couple of inches of one of the infected wandering trad stems received and gently press the hole closed. We advised participants to plant at least five stems in a single patch of wandering trad about 1 m apart, in areas with dense cover of the weed, usually in deeply shaded, moist environments.

In exchange for the agent, we requested that participants return the following information to CSIRO researchers via email: name of participant and associated community group, release date, release location (in the form of a GPS coordinate), number of stems planted at each location, a general description of habitat and weather conditions during the week in which the stems were planted.

Results of biocontrol releases across NSW

Altogether, between October 2020 and June 2023, we distributed > 11,000 healthy, viable units of the biocontrol agent to a total of 248 participating entities for release at 425 unique sites. Releases were targeted to areas of highest frequency and abundance of wandering trad, especially in areas of high conservation value (e.g., national parks, reserves, remnant rainforests with high potential for natural recovery from the seed bank), ensuring that releases were made across the entire range of the weed in NSW.

Participants were drawn from 73 community groups, usually for release in reserves and national parks, and 175 private citizens for release on private properties. Participation in the biocontrol release program remained steady over the three years, with generally high demand for releases in the peak weeding seasons of spring and autumn. All levels of government and the volunteer sector were invited to participate in the program; predominantly Landcare and Bushcare groups, key local government partners and NPWS.

This represented one of the largest community-based mass-release, monitoring and evaluation programs for any biocontrol agent across Australia. The CSIRO is sincerely grateful to all participants for engaging with us on this high-impact project to help tackle this damaging weed. Results have contributed significantly to further optimising the release of the biocontrol agent in potential future projects.

Geographical distribution of 425 sites (blue dots) where the biocontrol agent Kordyana brasiliensis was released on infestations of wandering trad throughout NSW over three years, between October 2020 and June 2023. Orange dots represent major urban centres; large grey dots represent verified wandering trad occurrence records obtained from Atlas of Living Australia in July 2021.

Next steps

The CSIRO is sincerely grateful to all participants for engagement on this biocontrol agent release program.

The CSIRO acknowledges that it was not possible to meet face to face with most participants between2020-2022, during the establishment phase of this project, due to COVID-19 travel restrictions. The CSIRO also acknowledges that many participants who expressed interest in participating in the program could not be provided with fungal material, due to very high demand for participation. Managing very high demand for participation in the midst of ongoing lockdowns to our laboratories and inability to travel in the field also meant that some participants to whom we promised the fungus were not provided with a biocontrol agent release kit.

In terms of next steps, the CSIRO welcomes ongoing interactions with community members who retain interest in the wandering trad biocontrol program. Participants are invited to engage as follows:

1. Participants who were promised to be sent a biocontrol agent release kit, but were missed during COVID-19, are very welcome to reach out to Dr Ben Gooden via email for assistance (Ben.Gooden@csiro). The CSIRO will consider sending such participants the fungus for release into the environment over spring and early summer 2023 and autumn 2024. As our stocks of the fungus are limited, we will only consider providing the fungus to participants who were missed during the main project.
2. The biocontrol agent can be collected from the field and redistributed to new locations on the proviso that:
   1. 1. 1. 1. Participants seek and are granted permission from site custodians to enter areas to collect the fungus for redistribution (note that the CSIRO does not provide advice or hold liability for on-ground activities pertaining to collection or redistribution of the fungus or any other on-ground activities related to this biocontrol program). The location of all known established populations of the fungus located on public land (private land excluded for privacy reasons) will be uploaded to the Atlas of Living Australia’s Biocontrol Hub in December 2023. https://biocollect.ala.org.au/biocontrolhub
            2. Participants are encouraged to record any new biocontrol agent release locations. This can be done via iNaturalist (preferred by most community members). The iNaturalist platform already has > 260 recorded populations of the fungus (search for Kordyana brasiliensis), along with photographs etc. Verified observations of the fungus uploaded to iNaturalist will eventually be uploaded to the Atlas of Living Australia Hub.
            3. The fungus can be released following the previous methods described here: Wandering Trad Biocontrol_Release Information_NSW Enviro Trust_07.09.20

Projects supported by the Australian Government as part of the National Environmental Research Program — Emerging priorities (Phase 1 project: 1 July 2014 to 1 May 2015) and the National Environmental Science Program — Emerging priorities (Phase 2 project: 15 May 2016 to 22 May 2017), administered by the Department of Environment and Energy (now Department of Agriculture, Water and the Environment).

Collaborator: Agriculture Victoria Research Division, Department of Jobs, Precincts and Regions (previously Department of Economic Development, Jobs, Transport and Resources)

Research performed by CSIRO focused on assessing the risk that the fungus Kordyana brasiliensis may pose to non-target native and ornamental plants closely related to wandering trad that are present in Australia. Additional host-specificity tests, to complement testing performed as part of the New Zealand program, were required because Australia has several native species in the Commelinaceae family to which wandering trad belongs to. Tests were undertaken in a quarantine facility in Australia.

Results obtained in these tests further demonstrated that K. brasiliensis is highly host specific towards wandering trad. All Australian populations of wandering trad tested were attacked by the fungus. Non-target plant species tested developed no symptoms or a limited number of small flecks on some leaves – an indication of resistance to infection by the fungus.

An application to release K. brasiliensis into Australia was submitted in November 2016 to the Department of Agriculture and Water Resources. Following a comprehensive risk assessment performed by the Department and consultation with states and territories as well as general public, the fungus was deemed to pose very low (negligible) risk to non-target plants and was approved for release from quarantine on 14 December 2018 (https://www.agriculture.gov.au/biosecurity/risk-analysis/memos/ba2018-32).

As part of these projects, scientists from Agriculture Victoria Research Division also performed host-specificity tests with the Tradescantia leaf beetle, Neolema ogloblini in a quarantine facility to assess risks to non-target plants relevant for the Australian context.

In the Phase 1 project, choice-minus-target testing using leaf beetle adults was undertaken with 16 native and ornamental plants closely related to wandering trad. Oviposition and adult feeding damage by the leaf beetle were lower on the non-target species than on wandering trad, and none of the eggs hatched on any of the non-target plants.

In the Phase 2 project, a no-choice experiment was conducted by placing equal numbers of leaf beetle eggs onto the underside of wandering trad and the non-target species tested during Phase 1. Results showed that eggs hatched on all species and developed to adults on wandering trad and to various life-stages on seven of the Australian native Commelinaceae species tested: Aneilema acuminatum, A. biflorum, Murdannia graminea, Commelina ciliata, C. cyanea, C. diffusa and C. lanceolata. Higher numbers of individuals at all life-stages and more larval feeding damage (skeletonised leaves) were recorded on wandering trad.

Based on these results, a further no-choice adult oviposition experiment was set up using all test plants on which larval development had been recorded. After 3 months, only a single pupa was recovered from the cage containing A. acuminatum, confirming oviposition and development of the beetle to at least this stage on that species. With remaining resources in the project, one further experiment was undertaken to compare utilisation of A. acuminatum and wandering trad by the leaf beetle, by exposing plants to mating adult pairs within enclosed gauze sleeves. Results confirmed that the leaf beetle can oviposit on A. acuminatum but at rates nearly 20 times lower to those recorded on wandering trad. Furthermore, oviposition was only recorded on 2 of the 15 A. acuminatum replicate plants tested. After 14 days, numerous first and second instar larvae were present on wandering trad, but none were observed on A. acuminatum plants.

These experiments showed that the leaf beetle N. ogloblini is potentially not restricted in host range to wandering trad and may be able to develop on a few related Australian native plant species. Additional no-choice adult oviposition and developmental experiments performed over a longer period to allow multiple generations of the beetle to develop are required to fully assess risks before an application for release in Australia can be considered.

Project supported by the Australian Government as part of the Improving Your Local Parks and Environment Program (2017 to 2020) administered by the Department of Industry, Innovation and Science on behalf of the Department of the Environment and Energy (now Department of Agriculture, Water and the Environment).

This project focused on releasing and monitoring progress of the fungus Kordyana brasiliensis on wandering trad in the Dandenong Ranges (2019-20) and throughout Victoria more broadly (2020), in partnership with local community stakeholders (in particular the Community Weed Alliance of the Dandenongs [CWAD], coordinated by Bill Incoll).

Developing release methods

In March 2019, after permission was granted to release the fungus from quarantine, an efficient system was developed to produce wandering trad plants infected with the fungus to have a continuous supply of material for delivery to community members. Following initial trials, two methods were selected for large-scale releases: the ‘bin-inversion’ and ‘planted-stem’ methods.

For the ‘bin-inversion’ method, between 10 and 15 wandering trad leaves infected with the fungus are carefully stuck using Vaseline to the inside-bottom of a large plastic tub (e.g. garbage bin), with the underside of each leaf facing upwards towards the opening. The bin is then inverted over a dense patch of wandering trad growing in cool, moist, shaded habitats for at least 24 hours to allow spores to be discharged from the infected leaves, land on the healthy wandering trad leaves below and infect them.

For the ‘planted-stem’ method, a single stem of wandering trad with several leaves infected with the fungus is planted into moist soil in the centre of a patch of wandering trad, thereby allowing spores to be passively discharged over several weeks and transferred to nearby healthy leaves.

Establishing long-term monitoring plots to measure impacts of the fungus on wandering trad and associated vegetation

In May 2019, the fungus was released using the ‘bin-inversion method’ at each of 15 long-term monitoring plots (6m x 10m in size) established in moist forests of Victoria. Baseline data on the abundance of wandering trad and diversity of native vegetation were collected prior to release of the fungus are each site. Monitoring undertaken between October and December 2019 –approximately five months after the fungus was first released – recorded infection of wandering trad at 60% of the monitoring plots. The fungus was re-released at 10 monitoring plots using a combination of the ‘bin-inversion’ and ‘planted stem’ methods in October 2019 to boost infection levels.

Unfortunately, due to covid-19 travel restrictions, it was not possible for CSIRO field ecologists to revisit the monitoring plots to evaluate establishment and spread of the fungus in 2020. However, visits by Bill Incoll from CWAD indicated that the fungus is beginning to spread widely at several sites.

Releases of the fungus throughout Victoria in partnership with local community organisations

CSIRO held three information sessions in 2018 to showcase the project to local community stakeholders: one hosted by the CWAD (at Birdsland, 31^st January 2018), one hosted by Yarra Ranges Council (Lilydale, 5^th June 2018) and another by Yarra Ranges Landcare Network (Monbulk, 5^th June 2018). Attendance was open to any interested local community stakeholder, weed control contractor or government officer. During these meetings, CSIRO provided attendees with information about the wandering trad fungus, a background to research underpinning its development, progress regarding the risk assessment and application to release the fungus from quarantine, its anticipated post-release impacts on wandering trad and benefits to native vegetation, and how community members would be able to participate in the release program once permission to release the fungus from quarantine was granted.

In 2019, five face-to-face workshops on the biocontrol project were delivered to local community stakeholders across the Dandenong Ranges region of Victoria. During these workshops, attendees were trained in release and monitoring techniques for the fungus. At least one bag containing approximately 10 wandering trad stems infected by the fungus were given to each of workshop participants to make releases at their trad-infested sites. Two attendees were provided with bags of at least 100 infected stems each, allowing for several releases to be made. Overall, we estimate that approximately 830 infected stems were provided to 63 unique attendees at these face-to-face workshops.

Between September 2019 and October 2020, CSIRO posted approximately 940 additional wandering trad stems infected by the fungus to 22 community participants, many of whom were unable to attend the face-to-face workshops in 2019. During 2020, CWAD set up 13 nursery sites with 990 infected stems sent by CSIRO, to provide sources of infected material in the future to make additional releases of the fungus in the region. The fungus has successfully established at these nursery sites (Bill Incoll, pers. comm.).