Calculating species detectability as a function of abundance
Surveillance of rare species is a critical part of environmental monitoring that can affect the management of invasive and endangered species. Imperfect detection can influence estimates of abundance in ecological surveys. This can complicate management decisions by causing bias in estimates and falsely declaring the presence and absence of species at a site. For example, determining whether a species is present at a site when it has previously not been detected can be critical for environmental impact assessments (Garrard et al. 2013), as well as for the control of invasive weeds.
There have been a number of models developed to try and account for imperfect detectability of individuals and imperfect detectability of species in ecological surveys. If the probability of detection of individuals can be estimated, then the observed number of individuals can be calibrated to estimate the actual number of individuals. Distance sampling (Buckland et al. 2001) is one of the most well-established statistical methods for estimating detectability of individuals. It based on the distribution of distances at which individuals are detected.
Species detectability has received increasing attention, with McCarthy et al. (2013) demonstrating that species detectability is related to abundance via a model of detection of individuals. Guillera-Arroita (unpub.) has extended the model of McCarthy et al. (2013) to explicitly account for how the probability of detection varies with distance. This leads to a predicted relationship between distance sampling models (based on probability of detection of individuals) and models of the detection of species.
Establishing a link between detection probabilities of individuals and detection probabilities of species is important because it can help to generalize different approaches to imperfect detectability in ecology. While a theoretical link between the two has been developed (Guillera-Arroita, unpub), whether such a relationship holds in reality has yet to be determined. The aim of this research project is to test this relationship between detection probabilities of individuals and detection probabilities of a species by using field experiments.
The project will involve field exercises where targets (e.g. fake Hieracium flower and golf tees) will be placed in prescribed locations. The first set of experiments will involve recording detection rate via distance sampling methods. The detection probabilities of individuals will then be used to predict quadrat-level detection rates, using the model developed by Guillera-Arroita (unpub.). This will be followed by a second set of experiments to estimate the detection rate within quadrats. The values for individual detection rates will be compared to rates of detection of species at a quadrat-level to estimate species-level detection rates. The experiments will involve searchers that will be recruited from members of the Quantitative and Applied Ecology Group in the School of Botany, and from students in the subject Environmental Monitoring and Audit.
Buckland S.T., Anderson D.R., Burnham K.P., Laake J.L., Borchers D.L. and Thomas L. 2001, Introduction to distance sampling: estimating abundance of biological populations, University Press, Oxford.
Garrard, G.E., McCarthy, M.A., Williams, N.S.G., Bekessy, S.A. and Wintle, B.A. 2013, ‘A general model of detectability using species traits’, Methods in Ecology and Evolution, vol. 4, no.1, pp. 45-52.
McCarthy, M.A., Moore J.L., Morris W.K., Parris K.M., Garrard G.E., Vesk P.A., Rumpff L., Giljohann K.M., Camac J.S., Bau S.S., Friend T., Harrison B. and Yue B. 2013, ‘The influence of abundance on detectability’, Oikos, vol. 122, no.5, pp. 717–726.