Journal Article
Providing regular and frequent maps of losses and gains of farmland birds based on European monitoring data
Sergi Herrando; Guillem Pocull; Sara Fraixedas; Anna Gamero; David Martí; Oriol Solà; Dani Villero; Verena Keller; Petr Voříšek; Alena Klvaňová; Gabriel Gargallo; Vitalie Ajder; Marc Anton; Ainars Aunins; Dawn Balmer; Mattia Brambilla; Tomasz Chodkiewicz; Przemysław Chylarecki; Cristi Domșa; Vlatka Dumbović Mazal; Virginia Escandell; Néstor Fernández; Ruud Foppen; Juan Gallego‐Zamorano; Carlos Godinho; Irene Guerrero; Christina Ieronymidou; Frédéric Jiguet; Aleksi Lehikoinen; Primož Kmecl; Peter Knaus; Lechosław Kuczyński; Åke Lindström; Qenan Maxhuni; Blas Molina; Jean‐Yves Paquet; Maria Luisa Paracchini; Danae Portolou; Draženko Z. Rajković; Dimitrije Radišić; Jiří Reif; Paul Shimmings; Henk Sierdsema; Jovica Sjeničić; Karel Šťastný; Bård G. Stokke; Nicolas Strebel; Zoltán D. Szabó; Tibor Szép; Norbert Teufelbauer; Nicolas Titeux; Sven Trautmann; Judit Veres‐Szászkac; Thomas Vikstrøm; Lluís Brotons
Conservation Biology · Vol. 40, Issue 4 · 2026
Abstract
Knowledge of species distributions is essential for informing policies on nature conservation and restoration. However, updating them on a regular basis and doing so in a harmonized manner at the international level is difficult. The European Bird Census Council integrated national monitoring data covering 5 years to update farmland bird distributions and assessed how they changed. We used these data on 50 farmland bird species to generate 10×10 km maps showing their probability of occurrence from 2018 to 2022. We produced these maps with weighted ensemble species distribution models. We also developed models for the previous 5 years and plotted the differences in probabilities of occurrence per 10 × 10‐km area between the two periods as calibrated change maps. We evaluated model performance at continental and regional levels and interpreted changes in probability of occurrence in relation to known abundance trends. Models showed good predictive performance (mean AUC ≈ 0.84; mean squared error ≈ 0.13). Change estimates were reliable for 43 species (high accuracy, low bias), and distribution changes were positively correlated with independent abundance trends (Pearson's r ≈ 0.50). Thus, the distribution maps for the two periods accurately captured species’ distribution patterns and their temporal changes for all species at the European scale and for the majority of species in all regions except southeastern Europe. Among the 43 species with reliable estimates, predicted occurrences declined for 33 species, increased for nine, and were the same for one species. For most species, the direction of change in distribution was consistent with changes in species overall abundance in the same period, except for four species. Overall, our results indicated a recent contraction of farmland bird distributions in Europe, highlighting the strong capacity of existing bird monitoring networks to provide continent‐wide species maps that can be updated regularly.