24 – Summary: monitoring frameworks and monitoring program design^317especially given that threatened species are, by definition, rare and usually highly
cryptic. Failure to correct for imperfect detection can lead to unreliable estimates
of population size and/or falsely concluding that populations are stable when in
fact they are in decline, and possibly near extinction. Chapters 20 and 22
demonstrate how detectability can be accounted for by carefully considering the
type of detection method and the survey effort/intensity at sites. For example,
Chapter 22 uses mark–recapture methods, which involves recording the presence–
absence of individual dolphins during repeated visits to sites, to estimate
population size and emigration while accounting for the proportion of the
population that remained undetected. Similarly, the dugong case study, also in
Chapter 22, demonstrates how a double-observer method can estimate how
detectability varies as a function of environmental conditions, in this case with
increasing depth and sea-state conditions.
Even when detectability biases are accounted for, detecting population trends
in threatened species presents additional challenges. Our ability to detect trends in
a time series of surveys depends on the length of the monitoring period, the rate of
change in the population, the frequency of surveys and the precision of population
estimates. Population estimates for threatened species are often imprecise because
populations are small, which means managers can have very low statistical power
to detect population declines even when declines are severe (as demonstrated in
Chapter 20). Statistical power can be improved by increasing the survey frequency
or precision of population estimates. However, for some species, trend detection
might still be problematic because populations can f luctuate dramatically in ‘boom
and bust cycles’ (Chapter 21). In such cases, simple trend lines cannot be drawn
through time series of data; the challenge is distinguishing population changes due
to environmental drivers (such as rainfall) from changes due to a new and
emerging threats. Comparing population dynamics with historical records,
knowledge about the species’ behaviour and life history, and closely monitoring
threats can help discern whether management is required in these cases.
These difficulties of accounting for biases and detecting trends can sometimes
be overcome using extra effort and resources, or by using monitoring techniques
that have been tailored for the target species. For example, advances in motion-
triggered camera technology have improved the duration and feasibility of
monitoring at remote sites. Fortunately, a range of new and emerging technologies
are, or will soon be, available to managers that make it easier to detect species,
account for biases and improve the precision of population estimates at reduced
cost. Chapter 23 summarises two promising emerging technologies: drones and
eDNA. Using a tree kangaroo species in Northern Australia, Chapter 23
demonstrates how drones can potentially access habitat that, until now, has been
inaccessible or expensive to survey. Similarly, eDNA is a powerful tool for
monitoring the presence or absence of threatened species, which is potentially
much cheaper than live trapping, allowing for more data to be collected. Such