Occupancy dynamics of the mottled owl Strix virgata using object-based image analysis along an urbanized Neotropical gradient

• Published in: Global ecology and conservation 2024-11, Vol.55, p.e03243, Article e03243
• Main Authors: San Martín-Cruz, Miguel A., Villegas-Patraca, Rafael, Dáttilo, Wesley, Enriquez, Paula L., Ruelas Inzunza, Ernesto
• Format: Article
• Language: English
• Subjects: Habitat fragmentation; Habitat selection; Modeling; Playback surveys; Urban ecology; Urban wildlife
• ISSN: 2351-9894; 2351-9894
• DOI: 10.1016/j.gecco.2024.e03243

Urbanization has profound effects on wildlife. Although some species benefit or even thrive in urban environments, most species respond differently to the varying degrees of disturbance that can be found across an urbanized landscape. Quantifying the effects of urbanization in wildlife distributions, however, is complicated: species vary in their patterns of presence/absence, abundance, and detectability across spatial and temporal scales, e.g., daily, seasonal, or throughout the annual cycle. Here, we use occupancy models to offer a realistic approach to the study of populations of urban owls. Most studies on owl population dynamics have not considered temporal variation in occupancy between seasons or assess the uneven effects of urbanization along a habitat gradient. We investigated the seasonal habitat associations of mottled owls along an urban gradient in the Neotropical city of Xalapa, Veracruz, Mexico. Using high-resolution satellite images and object-based image classification techniques, were analyzed the relationship between different vegetation and environmental characteristics with the occupancy of mottled owls. We employed different sampling techniques, including playback surveys and silent listening periods, to detect the presence or absence of owls along a gradient from highest to lowest urbanization. Environmental data and different vegetation types were used to analyze the habitat associations of mottled owls during January (late non-breeding season) and May (late breeding season) of 2023. In total, we detected 68 mottled owls during the non-breeding season and 102 during the breeding season, with higher detection rates in areas with >28 % forest surface. Our results revealed that the best occupancy model included forest and forest division (occupancy), ambient noise and moonlight (detection) for the non-breeding season, as well as urban, forest, grass, and forest division (occupancy), and noise (detection) for the breeding season. The percentage of forested and grass areas positively influenced mottled owl occupancy while the percentage of urbanization and forest division influenced it negatively. Moonlight was positively related to mottled owl detection, while ambient noise had a negative effect on detection probabilities of mottled owls during both seasons. Forested areas emerged as pivotal for owl occupancy, indicating their sensitivity to forest changes along the urban gradient. With urban areas increasing, the interplay between fore