Scientists Turn 4.4 Million Minutes of Sound Into a New Picture of Life in Croatia’s Iconic Wetland
Unveiled as Kopački Rit marks 50 years of protection, the three-year study identified 164 vocal species and established a new baseline for understanding how the wetland is changing.
KOPAČKI RIT, Croatia, October 1, 2026 — Biodiversity scientists have transformed more than 4.4 million minutes of sound recorded across Kopački Rit Nature Park into a new picture of wildlife within one of Europe’s most important wetland landscapes.
Led in partnership by the Faculty of Agrobiotechnical Sciences Osijek, University of Osijek, WildMon, and Huawei Technologies via its TECH4ALL Nature Guardian program, the project is a demonstration of how ecological research, field knowledge, and artificial intelligence can help protected-area managers understand changes in biodiversity over time.
Listening for ecological change
Kopački Rit supports a rich community of species closely connected to its wetlands, floodplain forests, and waterways. More than one-fifth of the species classified in the study depend on aquatic or semi-aquatic food resources, highlighting the importance of the park’s wetland habitats.
Between 2023 and 2025, park scientists deployed passive acoustic recorders at 98 locations across Kopački Rit’s wetland and forest ecosystems, producing 4,442,539 one-minute recordings, or equivalent to more than eight years of continuous sound. WildMon’s biodiversity team analysed the recordings using artificial intelligence, acoustic matching, ecological models and expert validation, helping them identify sounds of 164 species, including 154 birds, 5 mammals and 5 amphibians.
Among the species detected was the European turtle dove, classified as Vulnerable on the IUCN Red List following major population declines across Europe. Researchers also recorded four Near Threatened species: the Eurasian curlew, northern lapwing, ferruginous duck and European otter.
The monitoring also focused on species of particular ecological or conservation interest to the park, including the white-tailed eagle, European tree frog, great spotted woodpecker, red deer and golden jackal. The white-tailed eagle was consistently detected throughout the three-year study. Closely associated with large wetlands, riparian forests and abundant fish and waterbird populations, the species has become an important symbol of Kopački Rit’s conservation value.
“For 50 years, Kopački Rit’s protected status has helped safeguard an extraordinary living landscape,” said Prof. Ivana Majić, PhD, Faculty of Agrobiotechnical Sciences Osijek, University of Osijek. “This research gives us a new way to study the landscape and strengthen the evidence available to guide its protection in the years ahead.”
Turning millions of recordings into usable evidence
Processing a sound archive of this scale manually would be extremely difficult. To make the recordings usable, researchers combined established sound-recognition tools with acoustic similarity searches and manual expert validation.
The project also produced a regional acoustic classification model trained on 154 species. In testing, the model achieved a weighted average precision of 98.9%, with 150 of the 154 species classes achieving precision of at least 90%. This performance enables large volumes of acoustic data to be processed rapidly and reliably, allowing conservation teams, park staff, academic partners, and decision-makers to track species occurrence and activity year after year without manually reviewing thousands of hours of recordings.
Identifying species, however, is only the first step towards understanding ecological change. Researchers also combined the acoustic observations with information about water, vegetation, terrain, and forest structure. Occupancy models were then used to estimate where species were likely to occur while accounting for the fact that an animal may be present without being detected during every recording period.
Together, these tools create a more complete ecological baseline than a species list alone. They can help researchers examine how wildlife distribution relates to changing water conditions, vegetation productivity, and habitat structure across multiple years. For a protected area of international significance such as Kopački Rit, this transforms passive acoustic monitoring from a time-limited research project into a practical system for detecting changes in species and ecological communities over time.
“Artificial intelligence is allowing our team at WildMon to process ecoacoustic information at a scale that would otherwise be impossible, but technology is only one part of the solution,” said José Wagner Ribeiro Junior, Quantitative Ecology Lead at WildMon. “The strength of this project comes from combining technology with the experience of rangers, local researchers and biodiversity scientists to build a sustainable monitoring system.”
A new baseline for understanding change
The three-year study establishes a consistent ecological baseline across Kopački Rit’s wetland and forest habitats. By monitoring the same landscape over multiple seasons and years, researchers can begin to understand where focal species occur, how their activity fluctuates and how these patterns relate to water conditions, vegetation and forest structure.
Future surveys can build on this baseline by repeating the same monitoring approach and comparing new observations with the data collected between 2023 and 2025. Over time, this could help researchers distinguish normal seasonal variation from longer-term ecological change—including potential responses to climate change, shifting water dynamics and changes in habitat conditions.
The findings can also help identify emerging signals that warrant closer investigation. Changes in the acoustic activity and occupancy of species such as the European tree frog, for example, could guide more targeted research into freshwater ecosystem health.
Beyond Kopački Rit, the project offers an approach that could be adapted across other protected wetlands. By combining ranger-led data collection, artificial intelligence and ecological analysis, protected areas can establish comparable baselines and strengthen their ability to understand and respond to environmental change.
The complete findings will be presented at the 15th International Symposium KOPAČKI RIT from October 1-2, 2026 at the Faculty of Agrobiotechnical Sciences, Vladimira Preloga 1, Osijek, Aula Magna.
About WildMon
WildMon is a nonprofit organization led by global experts in ecology and conservation technology. By combining multi-source biodiversity data and AI-powered analytics, WildMon transforms ecological information into practical conservation insights. Wildmon tools reduce technical barriers to biodiversity monitoring and help frontline communities, practitioners, and organizations make informed decisions for nature.
About the Faculty of Agrobiotechnical Sciences Osijek
The Faculty of Agrobiotechnical Sciences Osijek is one of the leading research and higher education institutions in the field of biotechnical sciences in Croatia with the aim of developing a research and educational profile recognized at the international level.
About Huawei Technologies
The Huawei TECH4ALL Nature Guardian program is a conservation initiative that uses acoustic monitoring, cloud computing, and artificial intelligence to protect endangered wildlife and ecosystems
Media Contacts
WildMon
Adlai Salcedo
Communications Lead
University of Osijek: Faculty of Agrobiotechnical Sciences
Ivana Majić
Project Leader
