ACT proof-of-concept collaboration
CSIRO partnered with the Australian Capital Territory (ACT) Government and the University of Canberra to develop a proof-of-concept for a digital solution that supports evidence-based weed management in the ACT. This research was enabled by multiple long-term datasets that had been collected by the ACT Government at weed control sites as well as at hundreds of vegetation monitoring plots.
(Left) The ArcGIS Fieldmaps app enables the ACT Parks and Conservation Services to record all their weed control operations and create a detailed spatial database. (Right) The ACT Conservation and Offsets teams annually visit fixed monitoring plots to independently record the density of target weeds at each location.
Bringing together experts and data sources
We convened a series of workshops, bringing together researchers, land managers, policy makers, and technical specialists from CSIRO, the ACT Government, and the University of Canberra. These workshops focused on harmonising and integrating more than a decade of weed control records with five years of vegetation monitoring data, alongside environmental and spatial datasets. Through iterative engagement and field visits, we developed well-documented and repeatable workflows to clean, join, and contextualise these datasets, ensuring they were fit for robust modelling and evaluation.
CSIRO researcher Dr Kyana Pike (left) learns from Parks and Conservation Services Senior Manager Mr Steve Taylor (right) about the decisions practitioners must make in the field when mapping control activities.
Developing robust data analysis methods
The new integrated datasets enabled robust analytical models for estimating historical weed recurrence rates after control (that is, the increase in weed density each year following a control event by replenishment from the seed bank or by other mechanisms) and landscape-scale habitat suitability (that is, the ability of a location in the landscape to support relatively lower or higher weed densities) for four priority weeds: African lovegrass, Chilean needle grass, serrated tussock and St John’s wort. This was complemented by a structured process for eliciting estimates of weed control efficacy (that is, the proportion of weeds that are typically removed at a control site using existing methods) from expert practitioners in the ACT.
These three factors (control efficacy, recurrence and habitat suitability) interact with each other in complex ways to determine weed population trends in actively managed landscapes such as the ACT. The relationship between them forms the backbone of our Weed Invasion & Management Simulator (WIMS) tool.
Through data integration, we generated new information that had never been available before in the ACT, such as the history of weed control at long-term vegetation monitoring plots. This enabled us to explore how weed control efforts and the ecological condition of plots were associated with native plant biodiversity across ACT grasslands and woodlands. This research is currently written up for publication in a scientific journal (read a brief summary here).
Funding and Acknowledgements
The research conducted through the ACT proof-of-concept collaboration was generously supported through the CSIRO APaIR investment: Future Technologies for Australia’s Biosecurity.
The CSIRO project team consisted of Ben Gooden (Senior Research Scientist), Jens Froese (Senior Research Scientist), Kyana Pike (Postdoctoral Fellow), Wen-Hsi Yang (Research Scientist), Felipe Albornoz Ramirez (Research Scientist) and Will Hore-Lacy (Data Visualisation & Analytics Specialist). It was supported by Richard Duncan from the University of Canberra. A large number of representatives from the ACT Government were actively involved in co-designing this collaborative research, including: Luke O’Loughlin, Steve Taylor, Megan Wyllie, Richard Milner, Brett Howland, Cameron Gallagher, Harley Baker, Nic Jario, Anthony Hart, Thea O’Loughlin, Laura Rayner and Emma Carlson.
The research was enabled thanks to the ACT Government’s investment in long-term monitoring, and the many ACT Government staff and contractors that have contributed to collecting field data over many years.