288 Monitoring Threatened Species and Ecological Communities
MARSS models are a good way to carry out these latter steps because of their
f lexibility, ability to analyse trends in multiple time series datasets, handle missing
data, partition variation due to process and observation errors, and utility in
incorporating the effects of covariates. Further modelling approaches can also be
contemplated to increase the options available for management interventions. For
example, when interaction strengths between different system components have
been quantified, structural equation models (SEMs) can be constructed that allow
managers to predict the effects of manipulating one component (e.g. culling a pest,
reducing grazing pressure) on remaining components of the ecosystem. SEMs are
currently under construction that predict the effects of climate change and predator
management on desert small mammals; similar research on irruptive species in arid
regions elsewhere suggests that this is a fruitful approach (Deguines et al. 2017).
At present, effective conservation monitoring in arid Australia is woefully
inadequate. In large part, this is because the region is remote, little known and ‘out
of sight, out of mind’ for most of the Australian population. Because of the lack of
infrastructure and capacity, arid regions also pose considerable cost and logistic
problems for remotely based management agencies that seek to conduct
monitoring. A vastly increased allocation of resources and much greater
involvement of local people, managers and researchers is essential to improve this
situation (Dickman and Wardle 2012). Although there is currently little indication
that this will occur, some form of monitoring may be required if proposed
irrigation schemes and extractive industries are allowed to proceed in the northern
and eastern parts of the arid zone (Dickman et al. 2017). Monitoring the effects of
such developments may not constitute conservation monitoring as discussed in
this chapter, but could help to focus a little more attention on the presently
forgotten landscapes of arid Australia.
Lessons learned
● (^) Species can exhibit very different spatial population dynamics, even within the
same family. A ‘one size fits all’ management approach therefore cannot be
employed for threatened species conservation.
● (^) Monitoring needs to be done strategically with respect to the life history of the
target species, and the timing, number and placement of monitoring sites to
increase the likelihood of reliable detection.
● (^) For many threatened irruptive species, information on natural history will
have to be gathered in initial surveys to guide the design of subsequent
monitoring; adaptive changes may need to be made to optimise the monitoring
design in light of this information.
● (^) Environmental factors and known or putative threats to the target species
should be monitored simultaneously so that drivers of irruption can be better
understood.