Abstract
Columbidae (pigeons and doves) are among the most abundant resident birds in agricultural landscapes and recorded as turbine fatalities, but they are rarely included in collision risk assessments. We evaluate nine years (2016–2024) of monitoring data from a wind farm in the Isthmus of Tehuantepec, Mexico, to estimate species-specific collision risk and fatalities for five species using a Bayesian Collision Risk Model (CRM). We integrated long-term visual census data, systematic carcass searches corrected for detection probability and carcass persistence, and biologically informed priors to estimate exposure rates, collision probabilities, and annual fatalities. Results revealed marked interspecific differences in collision risk that were not directly proportional to abundance. The most abundant species, White-winged Dove (Zenaida asiatica), showed high exposure but low collision probability. The opposite was found in less abundant species White-tipped Dove (Leptotila verreauxi). While overall estimated fatalities remained low, our model revealed a steady downward trend in absolute fatalities, projecting continued declines for 2025. Our findings highlight the importance of long-term monitoring and Bayesian approaches that incorporate uncertainty and species-specific traits. We provide parameterized priors that can be applied in future collision risk models and support the conclusion that wind farms may exhibit limited local mortality impacts on the studied avifauna.