Automatic acoustic monitoring and inventorying of .. (AMIBIO)
Automatic acoustic monitoring and inventorying of biodiversity
Start date: Feb 1, 2010,
End date: Jun 30, 2013
Animals use acoustic vocalisation as a very efficient means of communication because the sound does not require visual contact between emitter and receiver, it can travel over long distances and can carry the information content under low visibility conditions. Animals use sound for mate attraction, territorial defence and for early warning of other members of the species of the presence of dangerous predators. Acoustic surveying can be used for rapid assessment programmes (RAP), which quickly assess the biodiversity of specific regions. This enables estimates of biodiversity to be made more quickly. Acoustic surveying is useful because many animals (such as birds, bats, frogs) and stridulating insects are heard more often than they are seen or trapped.
The AMIBIO Biodiversity project aimed to develop an automated system for remote analysis of biodiversity using acoustic technologies. It specifically sought to demonstrate efficient and effective monitoring of specific groups of acoustically active biota, including insects, amphibians, birds and bats in the area of a forest in a Natura 2000 site near Athens: Hymettus area with forest Kaisariani and lake Vouliagmeni.
The project planned to develop and deploy a prototype hardware system based on small multi-sensor monitoring stations and a remote central station for automatic signal processing and analysis. The system would gather audio and environmental information from the field in an improved and non-obtrusive way. It would then transmit the data to the central station to enable automatic assessment and analysis.
The project thus aimed to conduct an inventory of biodiversity in the project site and enable the monitoring o biodiversity trends over time. It would provide an index of biodiversity based on the complexity of calls recorded within a region. Finally, the system was expected to enable detection of specific atypical sound events, including natural and human-induced disasters - such as fires and storms - as well as those related to potentially hazardous human activities â such as firing guns or tree-cutting.
The AmiBio Biodiversity project developed a particularly innovative integrated technological system to enable remote, quick and non-obtrusive monitoring of biodiversity, proving the reliability, robustness and efficiency of the system in field conditions. It contributed directly to the biodiversity assessment of the project site and demonstrated great transferability potential for the rest of Europe.
The project team developed and deployed a network of 17 monitoring stations across the project site. Each station was less than 0.5 m3 to limit environmental and aesthetic impact. They were progressively enhanced as the project developed more sophisticated equipment. The final monitoring stations were totally autonomous, able to be controlled remotely from the central station at the University of Patras, could operate 24/7 and were powered by solar panels.
Each station had microphones able to make audio recordings of animals vocalising within the human hearing range, between 100 and 20 000 Hz. In addition, they had ultrasound recording equipment, enabling detection of bats and insects with very high-pitched songs. Stations also took numerous environmental and climatic measurements. They were able wirelessly to transmit all the information collected to the central station using 3G technology and prioritising âalarm eventsâ. They were also equipped with local storage to ensure integrity of sound recordings and station autonomy in case of network malfunction.
During its course, the project performed 11 biodiversity surveys: three for insects; seven for birds; and one for bats. By the end, the project had generated a reference sound library of around 28 000 identified sound recordings of insect, avian, mammalian and amphibian species. This work increased the list of known sound-producing animals in the study area to 180 species, including 26 insects, 1 toad, 11 reptiles and 15 mammals. Of the 127 birds recorded, 15 had not previously been recorded in the area by researchers. The number of bat species identified increased from two to eight. The system offered a rough estimation of population density through mapping of presence at the monitoring locations and showed that most threatened species in the area depend on open and semi-open habitats, not forest.
The project developed sound detectors to automatically identify and classify sounds. âLightâ ones were installed in the monitoring stations able to classify basic sounds and send priority messages about âalarm eventsâ for operators to investigate. More computationally demanding sound detectors were run in the central station able to find and extract sound recordings of specific species from the large amount of recorded data in the systemâs central database. Each species requires its own sound recogniser. Based on its own recordings, the project itself programmed, trained and tested detectors to recognise 70 of the 180 animals in the inventory, including rare species. It won international awards for its new insect sound detectors.
The biodiversity assessment produced by the project will serve as the baseline for the future development of an official conservation plan for Hymettus that should be developed with stakeholder participation. Meanwhile, excellent dissemination and networking work helped the project to showcase the high transferability of the projectâs innovative system internationally.
Further information on the project can be found in the project's layman report and After-LIFE Conservation Plan (see "Read more" section).
Get Access to the 1st Network for European Cooperation