Starlings, pest species for agriculture, monitored with new technology
Starlings could cause an estimated $42.8 million per year in damage to Western Australia’s grain and horticulture industries. In addition, starlings aggressively push native birds from their nests and cause damage to property, incurring costs for cleanup and the reestablishment of native birds.
In an effort to prevent the spread of one of the world’s most invasive species, a $156,440 project by the Government of Western Australia’s Department of Agriculture and Food, in collaboration with Queensland University of Technology (QUT) and DKB Solutions, aims to refine starling surveillance activities throughout the Esperance region in Western Australia.
The WA Government is taking action to prevent the starlings becoming established. Bird watching groups record sightings and the government is working to eradicate the birds with lures and traps. However, the distribution of the birds makes spotting and trapping them problematic. A solution is to automatically monitor environmental sounds and identify the calls of starlings.
As a result of this year-long project, “Tweetfinder2,” the newly-revised software will be capable of rapidly and more effectively detecting starling calls, assisting in the location and tracking of starling invasions. Specifically, the accurate monitoring of this pest species will enable the department to better identify where the birds are distributed, and the habitat corridors that they favor as they travel inland across Western Australia towards South Australia.
Professor Stuart Parsons of QUT’s Science and Engineering Faculty and Victor Obolonkin of Auckland company BellSoft, designed and built the original “Tweetfinder” software, which has been trialed by the WA government. This new collaboration makes it possible to refine this unique software, based on feedback from the first version, and to incorporate it into the Electronic Acoustic Recognition Sensors (EARSv2) hardware developed by DKB Solutions of Western Australia. This will allow starling recognition to occur ‘on-board’ the units so no post-processing is needed. The recognition algorithms can be trained to detect any acoustic signal, and so could be used to detect other invasive species, or to monitor those of conservation concern.
Professor Stuart Parsons of Queensland University of Technology’s Science and Engineering Faculty said, “Tweetfinder allows managers and conservationists to analyse huge volumes of acoustic data in faster that real-time. With the changes we plan to the algorithms, and their incorporation into EARS, identification of starlings will happen in real-time onboard the acoustic monitors. This will greatly decrease the amount of time taken to detect invasions.”
Contact Professor Stuart Parsons to learn more about this project.
Find out more about research at QUT.
