Office for Environmental Programs - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 1 of 1
  • Item
    Thumbnail Image
    Finding the last Hawkweed plant: how detection rate scales with cluster size
    Kozel, Ben ( 2018)
    When surveying or monitoring a species, the assumption that all the individuals living in the sampling area will be detected rarely holds. Even for plants, detection probabilities are typically less than 1 (where 1 represents perfect detection). Detectability estimates can be affected by ecological processes, observational processes, and by the interactions between the two. For invasive species control programs understanding the factors affecting detectability is critical to determining the effort needed to reduce population density to a desired level, or prevent their establishment in new areas. The influence of cluster size on detection rate was investigated in the context of Orange Hawkweed and King-Devil Hawkweed, herbaceous species with the potential to cause significant environmental harm in Australian alpine habitats. Detection rate was modelled using a well-recognised time-to-detection modelling approach. In this model, detection rate is proportional to a power function of Hawkweed cluster size, with a scaling exponent that depends on the nature of this relationship. All other factors being equal, rate of detection is expected to increase less-than-proportionally with increasing cluster size; specifically, it should rise with the square root of cluster size, owing to rudimentary mathematics governing apparent size. The scaling exponent, and other model parameters, were estimated using data from search experiments involving six different Hawkweed cluster sizes and 15 searchers. The data support the validity of the model. High inter-observer variability in detection rates generated substantial credible intervals for the scaling exponent estimates. Nevertheless, these estimates suggest a deviation from the expectation that detection rate varies with the square root of cluster size. Reasons for this discrepancy are proposed, centring on the influence of perception psychology in the formation of Hawkweed ‘images’ by searchers. The findings may couple well with distance sampling methodology to refine the systematic monitoring, or eradication, of Hawkweed within a defined area. Quantifying the Detection rate – Cluster-size relationship yields better estimates of the minimum effort required to find solitary plants, which, in turn, improves resource allocation.