We found that urbanization affected gene flow differently for both species. Urbanized areas do not represent a barrier for house wren, consequently this species appears to be a single panmictic population in Costa Rica. On the contrary, gene flow is restricted by urban development in white-eared ground-sparrow. Using microsatellites, our analyses showed higher genetic structuring in white-eared ground-sparrow than in house wren. STRUCTURE and DAPC suggested differences in allele frequencies among white-eared ground-sparrow populations, while clustering all house wren individuals into a single group.
Genetic diversity estimates were similar for both species because confident intervals overlapped. These results contradict our initial predictions, since we expected a higher genetic diversity for house wren which is an urban tolerant species that shows great plasticity in habitat and nesting site preferences 27–29. On the other hand, we expected white-eared ground-sparrow to have less genetic diversity since this species is isolated in small populations that are expected to experience higher genetic drift 12,27. Similar results were found in two European bird species. In this study the tolerant blue tit (Cyanistes caeruleus) and less tolerant great tit, both species had similar genetic diversity in an urban area, even when genetic structure was significantly higher for great tit than for blue tit 30. Other studies have reported similar genetic diversity estimates between urban and rural populations of the same bird species (e.g., song sparrow; burrowing owl, Athene cunicularia; house sparrow, Passer domesticus) 16,31,32. Therefore, our results may be explained because both species have large effective population sizes due to the recent urbanization in the study area (< 60 years) 33, which is not sufficient to reduce population sizes and consequently genetic diversity through drift. This will be particularly true given that habitat fragmentation has occurred gradually 33, which probably allowed remnant forest patches to act as stepping-stones for gene flow in the last three decades 1,24,34
Clustering algorithms suggest that urbanization functions as barriers to gene flow in white-eared ground-sparrow, genetically structuring its populations, but in contrast, all populations of the house wren in Costa Rica may be considered a large panmictic population. Habitat loss and fragmentation often limit connectivity and gene flow and, consequently, increase differences in allele frequencies between populations 35. Multiple studies have shown how urban areas increase fragmentation and isolation of avian populations with the corresponding effects on genetic diversity 14,15,18,19,30. However, the magnitude of the effects of urbanization varies among species. For example, differences in sensitivity to urbanization resulted in higher genetic structure for great tits than for blue tits in Poland, in which fragmentation (urban vs. forest) and isolation by urban areas limited gene flow for great tits but did not for blue tits 30. The authors attribute these differences to the greater migration and dispersal ability of blue tits in urban populations 30. Tan et al. 19 also found that the effective population size of the forest-edge tolerant Singaporean yellow-breasted babbler correlates with the habitat contraction caused by urbanization. Genetic differences in urban subpopulations were attributed to small effective population sizes, limited dispersal, and a lack of connectivity between patches 19. The differences we observed on genetic structure between white-eared ground-sparrow and house wren may also be explained by contrasting tolerance to urbanization between these two species 27. White-eared ground-sparrow and house wren differ in how they can benefit from urban habitats. White-eared ground-sparrow feeds in the humid leaf litter and requires thickets to build nests, while house wren feeds on arthropods on a wide array of substrates and can build nests in buildings and human made cavities 22,27−29. A large proportion of the original habitat occupied by white-eared ground-sparrow in Costa Rica is now located within the Greater Metropolitan Area (GAM), characterized by a recent, rapid, and unplanned urban development 36. For the earlier half of the 20th century, a large portion of the GAM was covered by coffee plantations and second growth vegetation, but recent urban development has reduced coffee plantations and natural vegetation cover to small, isolated fragments 1,26. Hence, the preferred habitat of white-eared ground-sparrow has been constantly reduced; threatening the viability of this species’ populations 25,37,38.
House wren inhabits a wider array of habitats and can move between populations across large, urbanized areas 29. Accordingly, we were not able to find evidence of isolation by distance or by habitat resistance. However, we found intermediate levels of genetic structure attributed to its likely low natal dispersal, because young individuals do not need to move long distances to find suitable habitat 39. However, this hypothesis does not explain why we did not find significant IBD. This issue will be discussed below in further detail.
From the four populations of white-eared ground-sparrow, two (HDA and UCR, Fig. 1) were clustered together. These two populations are separated by the shortest distance and likely interconnected through remnant coffee plantations and patches of second growth vegetation. This suggests that gene flow for white-eared ground-sparrow is dispersal limited due possibly to urban areas limiting migration among populations 27. JBL is separated from HDA and UCR by a large, urbanized area of approximately 2280 ha 40, and by a mountainous densely forested terrain which likely limited migration and gene flow. These results are supported by the high correlation between urban resistance and genetic differentiation for white-eared ground-sparrow. MTV is geographically the most distant site, but clustered closer to HDA and UCR, due to large patches of coffee plantations located among these sites 37. Contrary to white-eared ground-sparrow, our results suggest that house wren may be a single panmictic metapopulation. This is supported by the lack of IBD, and the weak correlations between genetic differentiation with urban resistance and climatic resistance. These results support our initial expectations that this species benefits from the recent urban development which increases the availability of nesting sites and habitat use, thus increasing connectivity among populations 27,29. In contrast we expect gene flow to decrease for white-eared ground-sparrow in future generations. This is expected for city populations of an urban avoider that may reach the lowest densities in urban areas 9,27.
All our isolation models suggest that geographic distance, urban resistance, and climatic resistance limit gene flow for white-eared ground-sparrow. Distance accounted for the strongest effect on genetic structure, which suggests that this species has limited dispersal, probably related to its territoriality 41,42, low natal dispersal, and philopatry, though this remains to be tested. The geographic distribution of this species in Costa Rica, which coincides with the largest urbanized area, further limits movement between populations as indicated by the high correlation between the resistance matrix and genetic structure 23,43,44. Another study using circuit theory analysis showed a similar effect from urbanization on population differentiation in song sparrow populations, concluding that urbanization reduces gene flow in a mobile, disturb-tolerant species 15. Other studies have drawn similar conclusions, finding that habitat fragmentation caused by urbanization expansion affects gene flow among isolated populations, particularly for species with a limited adaptive response to such drastic changes 14,15,30.
In contrast, genetic structure did not correlate with distance in house wren, though genetic structure correlates weakly with environmental resistance. These results may be related to its widespread distribution and tolerance to changes imposed by urbanization. The house wren is a generalist species that nests in different man-made cavities and feeds on arthropods from a large variety of substrates 28,29. However, 20 years ago, a study of six house wren populations in Costa Rica found significant IBD and genetic structure 39. The differences findings of both studies may be attributed to the rapid urban expansion that occurred during the last 20 years in Costa Rica, which likely has increased the dispersal and gene flow for house wren. Additionally, most populations studied by Arguedas & Parker 39 at that time were separated by forested areas, which is a barrier of dispersion for this species, producing less connected populations.
Our results provide evidence for the need of conservation strategies and management for non-resilient species whose habitat has been reduced to small fragments immersed in urban matrices. The neotropics is the most avian diverse region, but the rapid growth of overcrowded cities poses a risk for urban avoider birds 1,45,46. This is the case of white-eared ground-sparrow that may struggle to withstand the effects of the uncontrolled sprawl of Costa Rican urbanization 36. Urban expansion will further isolate white-eared ground-sparrow populations and reduce their available habitat, increasing the effect of drift and inbreeding, which may increase the risk of local extinction 47,48. Other species with similar ecology and behavior than white-eared ground-sparrow may be exposed to the same consequences of urbanization. It is then urgent to implement management and conservation strategies that ameliorate the negative effects of urbanization expansion on the rich Costa Rican avifauna 1.
In conclusion, our study shows that the effect of unplanned urbanization sprawl on genetic diversity may strongly depend on a species ecology and behavior. Our results indicate that urbanization limits gene flow and has an effect in the genetic structure of white-eared ground-sparrow, but barely affects house wren. Such differences are associated with differences in tolerance to urban ambient and the resources that these landscapes offer to each species. White-eared ground-sparrow is more sensitive to urbanization and habitat loss than house wren, which may result in lower diversity if habitat availability is further reduced. We are aware that our sample sizes may limit our conclusions, however, differences between species were consistent across all our analyses and our results contribute to the scant information on the vulnerability of city avoider species, such as white-eared ground-sparrow. Studies on gene flow in neotropical cities are fundamental to implement conservation strategies that increase connection between patches of natural vegetation immersed in extensive urban matrices.