Send message to Author

Poster Short Abstract

A Digital Twin integrates citizen science and an automated bird sound classifier to deliver real-time biodiversity predictions. The DT approach enables scalable and dynamic bird monitoring and empowers public participation.

Poster Abstract

Accurate environmental policies rely on timely and reliable data. Citizen science offers vast biodiversity observations, yet predictive power is often limited by data quality variability. We present a Digital Twin (DT) approach that integrates a citizen science bird sound app, machine learning, and high-performance computing to generate real-time biodiversity predictions. The DT stores raw audio to continuously refine species classifications and mitigate detection errors. Our campaign addressed spatiotemporal sampling biases via interval recordings and a permanent point count network. Over two years, 5% of Finland’s population contributed 14.5 million bird detections. Independent test data show that the DT significantly improves predictions of bird distributions. The scalable system enhances biodiversity monitoring, supports environmental policy, and fosters citizen engagement in science.