21 – Determining trends in irruptive desert species^285had been assumed, a priori, might be important. No clear temporal trends were
obvious in any of the nine populations that were monitored. This discussion first
assesses whether the results are likely to be aberrant or broadly representative for
populations in arid regions and then considers how trends can be discerned for
monitoring and management purposes.
Population irruptions are commonly observed in desert biota, often cover very
large regional areas and can show amplitudes of three to four orders of
magnitude. Although rainfall is the ultimate driver of these irruptions, many
other factors inf luence the shape of the population increases. For example,
irruptions may be muted if populations start from a low numerical base, or if
organisms are subject to competition or predation. The amount and seasonal
timing of rainfall, wildfires, nutrient levels in soil and legacy effects of past
climatic and other environmental events may also be inf luential (Dickman et al.
2014). Interactions among these factors complicate attempts to predict the
magnitude of response to any given rainfall event. Long-term studies of rodents in
the Simpson Desert have recorded different species irrupting variably following
the same rainfall event and the same species showing different irruption patterns
after similar rainfall events at different times (Greenville et al. 2016). Other
studies have documented similarly varied responses to rainfall within and
between species (e.g. McLean 2015) and suggest that the results for D. blythi and
S. youngsoni are not unusual for arid-dwelling mammals. Given the many factors
that can affect the pace and magnitude of irruptions, it is clear that simple
approaches to determining population trends – such as drawing trend lines
through time series data – are unlikely to be informative.
How, then, can trends in irruptive desert species be detected? Some suggestions
fol low.
In the first instance, the timing of monitoring should be sensitive to the
behaviour and life history of target organisms. Thus, ephemeral plants, burrowing
frogs, freshwater crabs, shield shrimps and other animals that require free water
for completion of their life histories may well be present consistently in desert
environments, but will not be recorded above ground during dry periods.
Nomadic birds, similarly, are likely to be absent during such periods. Predictive
modelling of distributions under changing environmental conditions promises to
allow future monitoring to be targeted to times and places where it is likely to be
most effective (Runge and Tulloch 2017). Rainfall should be monitored, with
monitoring for these taxa occurring following rainfall. The amount and timing of
rainfall that should trigger monitoring will depend on the target species. For
poorly known threatened taxa, such information may not be available; it is then
advisable to carry out initial surveys after different amounts of rain have fallen so
that monitoring designs subsequently can be adapted as needed and predictive
models ultimately can be generated.