Bioclimatic Drivers of Amur Falcon Habitat Dynamics Using Advanced Machine Learning Models

Abstract. The Amur Falcon (Falco amurensis) is a migratory raptor known for its long-distance journeys from eastern Russia and northern China to southern Africa, passing through various stopovers, including northeastern India and Southeast Asia. The Amur Falcon's migration spans diverse habitats and climatic zones, offering insights into its dynamics. However, climate change, habitat loss, and bio-climatic variability increasingly threaten its breeding and stopover sites. To date, no comprehensive study has analyzed how bio-climatic factors influence migration patterns across such a broad range. This study explores the bio-climatic factors influencing the falcon's migration and habitat suitability using remote sensing, GIS, and machine learning models—Maximum Entropy (MaxEnt) and Random Forest (RF). It evaluates 56 bio-climatic variables, such as temperature, precipitation, solar radiation, wind speed & water vapour pressure. Species occurrence data from citizen science is used to train and validate models. RF showed higher accuracy (AUC=0.98) than MaxEnt (AUC=0.96) and identified 6.69% of global land as moderately to highly suitable for the falcon, compared to MaxEnt’s 2.07%. The analysis also revealed potential habitats outside the bird's natural migration route, including parts of North America, South America, and Oceania. Key factors affecting habitat suitability included mean temperature, precipitation, and solar radiation. This research emphasizes the importance of bio-climatic data in understanding species distribution and migration patterns, offering valuable insights for conservation planning and avian ecology.