image: Muuttolintujen Kevät (Migration Birds Spring) is a mobile application developed at the University of Jyväskylä, enabling users to record bird songs and make bird observations using state-of-the-art artificial intelligence.
Credit: University of Jyväskylä
New cutting-edge research, led by Academy Professor Otso Ovaskainen of the University of Jyväskylä and David Dunson at Duke University, combines citizen bird observations with artificial intelligence and the computing power of supercomputers at CSC – IT Center for Science. The international and multidisciplinary research team has developed the world’s most accurate prediction model, capable of anticipating even small shifts in bird occurrence almost in real time.
Around seven million bird observations are collected in Finland every year using the Muuttolintujen Kevät (Spring of Migratory Birds) mobile app, which has more than 300,000 users. The research team has now published a prediction model that utilises these citizen observations alongside cutting-edge information technology.
“When combined with long-term bird survey data and land use and forest structure data, the continuous flow of the new citizen science data through our mobile app creates an exceptionally valuable data resource that is processed daily by CSC’s supercomputers,” says Academy Professor Otso Ovaskainen from the University of Jyväskylä.
“The new model we have developed is updated in real time and provides the most accurate predictions yet of bird movements and singing activity.”, notes Arts and Sciences Distinguished Professor David Dunson from Duke University. “The predictions rely on novel algorithms for rapidly combining different types of data in a Bayesian learning procedure based on a spatio-temporal process.", he continues.
Citizens’ observations boost research
Previously, predictions about where birds would be found relied heavily on observations made by skilled birdwatchers, and updates to species distributions based on new data were done only periodically. Now the new AI-based model of the Muuttolintujen Kevät mobile app enables anyone, regardless of their knowledge of birds, to participate in research. Mobile data transfer and new statistical models also mean that the predictions can be updated daily.
“This new approach opens up a wide range of possibilities for investigating various factors affecting the occurrence of species as well as predicting dynamic ecological phenomena,” says Ovaskainen.
The study was published in the prestigious Nature Ecology & Evolution journal. It was carried out as an international, multidisciplinary collaboration involving the University of Jyväskylä, Duke University, the CSC – IT Center for Science, the University of Helsinki, and a network of Finnish research stations.
Information back to users: towards a shared understanding of the environment
In the future, predictive information can be safely and ethically directed back to the Muuttolintujen Kevät app and its users. Sharing this information with users fosters a shared understanding of the environmental situation.
“Finns have been enthusiastic users of the Muuttolintujen Kevät app, thereby helping to achieve significant research results,” says Head of Development Ari Lehtiö at the University of Jyväskylä. “The goal is to develop an even better app with more features for public use by spring 2026.”
The Muuttolintujen Kevät app is available to download from smartphone app stores. It currently recognises over 250 Finnish bird species, and the recognition algorithm is constantly being developed.
Further information:
-
Muuttolintujen Kevät - Research for JYU Mobile: https://www.jyu.fi/en/research/muuttolintujen-kevat
Journal
Nature Ecology & Evolution
Method of Research
Data/statistical analysis
Subject of Research
Animals
Article Title
A digital twin for real-time biodiversity forecasting with citizen science data
Article Publication Date
27-Jan-2026