Avian metapopulation dynamics in an urbanizing landscape
Benjamin D. Padilla, MS
Advisor: Amanda Rodewald
Although habitat loss and fragmentation have troubled the Midwest for centuries, landscapes are once again changing profoundly as historically dominant agriculture is replaced by urban development. Urbanization alters both composition and structure of avian communities, yet little is known of the landscape level structure of bird populations. Because urban development may restrict movements of many species, changes in land use should affect connectivity and persistence of local populations that are distributed among remnant habitats. Metapopulations, which are comprised of demographically independent local sub-populations connected via dispersal, have the potential to promote long-term viability of small and isolated urban populations through rescue of sub-populations that are in decline. Such metapopulation dynamics require asynchrony in recruitment and survival rates among sub-populations. To understand the extent to which avian populations in urbanizing landscapes operate as metapopulations, I analyzed the population synchrony (spatial autocorrelation) of densities, rates of local extinction and recolonization for 16 species, and demographic growth rates (!) for 2 species breeding in 14 mature riparian forest sites in central Ohio, USA from 2005-2011. Two scenarios were considered to be consistent with metapopulation structure, cases with (1) high negative spatial autocorrelation and high rates of extinction and recolonization, or (2) low spatial autocorrelation with high rates of extinction and recolonization. Only 3 of & 16 species (19%) exhibited significant positive spatiwal autocorrelation across the entire study area. Only Acadian flycatchers in urban sites showed significant positive spatial autocorrelation of demographic growth rates. Furthermore, migrants and residents showed similar degrees of spatial autocorrelation (migrants r = 0.014 ± 0.0585 SE vs. residents r = 0.067 ± 0.0475 SE). Although autocorrelation did not differ significantly between resident and migratory species, patterns of autocorrelation differed between rural and urban landscapes, suggesting that urbanization affected connectivity among remnant patches. Probability of extinction ranged from 0.0 – 0.789 (mean = 0.166 ± 0.0522 SE), whereas probability of recolonization ranged from 0.0 – 1.0 (mean 0.652 ± 0.0815 SE). Low prevalence of positive spatial autocorrelation combined with high rates population turnover provide evidence that spatially subdivided avian populations in my system may function as metapopulations. The possibility that metapopulation dynamics may operate within urbanizing landscapes is particularly important given that demographic analyses indicated that the majority of subpopulations of Acadian flycatchers (Empidonax virescens) and northern cardinals (Cardinalis cardinalis) had negative growth rates and, thus, acted as population sinks. My results suggest that long-term persistence of birds breeding in Midwestern riparian forest fragments may be facilitated by metapopulation dynamics.