<aside> ✋🏼

1 September 2025: this grant round is now closed and we are assessing applications.

</aside>

Wedgetail is a private foundation based in Lutruwita/Tasmania that supports and funds the frontiers of biodiversity regeneration. We back bold efforts to understand, conserve and restore ecosystems.

We’re calling for grant applications to accelerate the accuracy and completeness of bird bioacoustics in Tasmania.

Why birds?

Birds are powerful indicators of ecosystem health. They respond quickly to environmental change, occupy diverse ecological niches, and play critical roles in pollination, seed dispersal, and pest control.

Thanks to advances in passive acoustic monitoring and machine learning, we can now detect and track many species using sound – but only if we have the data.

Right now, many Tasmanian bird species are missing or poorly represented in acoustic models. In some cases, existing recordings haven’t been surfaced or annotated. In others, there may be no recordings at all.

This is a solvable problem. We have the tools, the knowledge and the urgency. What’s needed is a targeted push to gather, improve and share data so Tasmanian birds can be more reliably detected and ecological change more easily tracked.

What we’re looking for

We see an emerging space opening up as the technology matures for recording and identifying bird audio.

Tasmania has a rich environment for birds, but not all of them are easily identifiable with existing machine-learning models. This is the right time to improve species identification by getting more species recordings annotated and into AI recognisers, to accelerate the availability and usefulness of Tasmanian bird audio data so that species can be more accurately identified and monitored.

We are seeking proposals from Tasmanians and Tasmanian organisations capable of contributing to one or more of the following components:

1. Understanding which species we should prioritise

Some species are well-represented and easily identified in existing machine-learning models, and others are not. Some species are more ecologically significant than others, or better biodiversity indicators, or more endangered. We need to identify which species to prioritise.

2. Uncovering, identifying and annotating existing recordings of priority species

There are hard drives in cupboards just waiting for someone (perhaps a machine) to listen to them. We welcome efforts to surface, digitise, and annotate your existing recordings.

3. Making and annotating new recordings of priority species

Some species may require new field recordings – whether via autonomous recording units or focal methods. We’re open to both.

4. Contributing data to open machine-learning models