Conclusion and future directions
In this study, we have demonstrated the possibility of building high
quality single call recognisers for monitoring ecosystem restoration.
While testing the recognisers on data collected before and after
watering would determine their precision in responding to actual
restoration, we expect the response to be sharp as precision and recall
scored high for most species. If sharp responses are found, we encourage
testing transferability of recognisers to other locations as site
specificity of recognisers has rarely been investigated. Amphibians are
also not the only soniferous taxonomic group that responds to
environmental watering. Birds were responsible for the bulk of the
manually annotated response curves described by Linke & Deretic (2020).
We strongly encourage additional studies that build call recognisers for
water-dependent birds or alternatively trial the performance of a
pre-built recogniser such as BirdNet (Kahl et al., 2021). This has the
potential to lay the foundation for a real-time monitoring system that
can detect the ecological outcomes of environmental water allocations in
the Murray-Darling Basin – a milestone in the management of a stressed,
but recovering, river.