16 – Why, what, how much, and is it worth it?^221more investment, they could substantially increase the reliability. An argument
for more funding is more convincing if accompanied by evidence such as this,
rather than one based on a vague notion that more monitoring is better than less.
Power analysis is a good way to decide whether to give up, seek more funding, or
stay calm.
Most threatened species monitoring programs are low-powered, meaning their
ability to detect change is very poor. So are these programs really a waste of
money? From a purely statistical perspective, the answer is probably yes. However,
there are good reasons to persist with low-powered monitoring, as long as
stakeholders and managers are aware of the situation and are not taking substantial
comfort from the lack of bad news arising from their monitoring program. Low-
powered monitoring can provide good public engagement for species, help secure
resources for higher powered monitoring, and provide pilot data for improving
future monitoring design. Either way, it makes sense to know the power of your
monitoring program.
Lessons learned
● (^) There are several reasons to monitor threatened species, but it is necessary to
be clear about why a particular monitoring program is needed so that the
benefits of monitoring can be quantified and monitoring options compared.
● (^) In many instances it is as important to monitor threats and other covariates, as
it is to monitor threatened species. It may be sufficient to only monitor
threatened species, or a surrogate, when ecological processes are very well
understood.
Fig. 16.4. The relationship between budget and statistical power of the Kakadu monitoring program to
detect a 40% reduction in the distribution of rufous whistler Pachycephala rufiventris (solid line) and spangled
drongo Dicrurus bracteatus (dashed line). The blue zone indicates monitoring effort that is insufficient to
achieve acceptable statistical power to detect 40% declines. The pink zone indicates budgets that achieve
very high power and the yellow zone indicates adequate and near adequate statistical power, depending on
the costs of failing to detect change.