Google Play icon

The sound of science: Recognizing species based on acoustics

Share
Posted December 12, 2019

Scientists are set to use data gained from ‘eavesdropping’ on nature to build a picture of Australia’s animal life.

James Cook University’s Professor Lin Schwarzkopf said scientists are developing new acoustic analysis techniques to recognize different species.

Scientists will soon begin testing the accuracy of data they’ve gathered from state-of-the-art audio gear that’s been recording animal sounds. Professor Schwarzkopf will manage the project, made possible by a grant of more than $550,000 from the Australian Research Council (ARC) and involving investigators from four other universities.

Image credit: Ivan Lojko via peakpx.com, CC0 Public Domain

Brown kangaroo. Image credit: Ivan Lojko via peakpx.com, CC0 Public Domain

“We’re aiming to use this data to better predict what will happen to animal life in Australia and the humans that depend on it.

“Improvements in computing power, storage, and cost-efficiency have opened up the possibility of using acoustics not just for single species monitoring as we do at the moment, but for large-scale biodiversity surveys,” said Professor Schwarzkopf.

The scientists will work with sound recorders at 100 environmental sites run by the ARC-funded Australian Acoustic Observatory.

Professor Schwarzkopf said biodiverse ecosystems (those with the appropriate range of animals and plants for their geographic location) provide us with clean air and water, improved soil fertility, pollination, and pest control, and are critical to human health and well-being.

“Understanding what biodiversity we have at present is critical to understanding the operation of natural systems, including those that benefit humans. It’s vital in determining impacts of human activity, and in predicting what will happen to our ecosystem under different climate scenarios,” she said.

Professor Schwarzkopf said such new methods could be practical ways to survey wildlife in a vast and sparsely populated country such as Australia.

“Traditionally, biodiversity assessment uses prohibitively expensive and time-consuming manual surveys or potentially inaccurate predictions based on secondary observations, such as local climate and vegetation type,” she said.

Professor Schwarzkopf said despite its promise, the technology of automatic audio analysis on a large scale was still rudimentary and still had to be proven against the techniques in use now.

“We’re going to test the technology against existing techniques, then deploy it to examine biodiversity on a large scale so we understand what the audio information can tell us about the whole system. Then we will be in a position to make better predictions on what is happening to biodiversity in Australia under different scenarios,” she said.

The project will run for three years.

Source: James Cook University

Featured news from related categories:

Technology Org App
Google Play icon
86,841 science & technology articles