2.1 Study area

Our study has been realized in the furthest northeast limit of Algeria, in Annaba (36°.30 N & 37°.03 N &7°.20 E & 8°.40 E, Fig. 1). This city is located at 600 km from the capital Algiers; it extends at about 80 km on the Mediterranean cost, it covers 1412 km², the climate is typically Mediterranean, with an average annual temperature of 18 °C and an annual rainfall ranging from 650 to 1000 mm with a peak in winter and a deficit in summer. The massive Edough (highest altitude 850 m) borders the city to the north and the west, the Mediterranean Sea to the east, and the alluvial plain of the Seybouse River to the south [16].

2.2 Data collection

2.3 Statistical analyses

The presence of birds was found significantly affected by the interactions between species and season, and between species and environments (Table 1). We reran analyses within each one of the 28 species detected to examine how season and environments affected their presences. The results per species are provided in Table 2 and illustrated in Fig. 2. First, stock and common-wood pigeons were never detected in all sampled sites. For all detected species (n = 26) for which we evidenced an effect of the season, the presence was more pronounced during the breeding season than during the winter one (Table 2). The type of urban environment (greenspaces, intermediate or built-up) did not significantly impact the presence of common swift, western cattle egret, black-headed gull, African blue tit, northern house-martin, barn swallow, yellow-legged gull, and European starling (Table 2). In contrast, the presence of birds was positively linked to the degree of greenspaces for the European greenfinch, blue tits, European robin, common kestrel, spotted flycatcher, Eurasian great tit, black redstart, common chiffchaff, European serin, laughing dove, European turtle-dove, black cap and Eurasian blackbird (greenspaces > intermediate > built-up; Fig. 2). Moreover, the presence of birds was higher in intermediate and greenspace environments (intermediate = greenspace > built-up) for the chaffinch, the house sparrow and the common bulbul (Fig. 2). Interestingly, for the feral pigeon, the presence of birds was significantly lower in greenspaces than in intermediate and built-up environments, but was higher in intermediate than in built-up environment (intermediate > built-up > greenspaces). The Eurasian collared dove was the most spread species in our area of research (present in more than 80% of sampled sites). Its presence was higher in greenspaces than in intermediate and built-up environments (green spaces > intermediate = built spaces).

The abundance of birds was found significantly affected by the interactions between species and environments and between species and season (Table 1). Results per species are provided in Table 2 and show that the feral pigeon, when present, is significantly more abundant in intermediate environment than in greenspaces, but more abundant in greenspaces than in built-up environment (intermediate > greenspace > built-up). Moreover, the abundance of the barn swallow and the Eurasian blackbird was negatively linked to urbanization (greenspaces > intermediate > built-up). For the house sparrow and the Eurasian collared dove, the abundances were lower in built-up environment than in greenspaces and intermediate environments (intermediate = greenspace > built-up). When present, the abundance of the European turtle-dove tended to be positively linked with urbanization (built-up > intermediate > greenspaces; marginally non-significant, Table 2). For the other species, the environment did not affect the abundances (Table 2).

The avian species richness (number of species seen at least once on the sample site) was found significantly affected by the interactions between season and environment (F_2,647 = 107.6, P < 0.0001). This interaction was driven by a more pronounced difference in species richness among environments during the breeding seasons than during the winter season (Fig. 3). Overall, species richness was higher during the breeding season than during the winter season (F_1,647 = 489.1, P < 0.0001), and was higher in greenspaces than in intermediate and built-up environments (F_2,647 = 372.3, P < 0.0001; Fig. 3).

As predicted, the presence of greenspaces significantly increases avian species richness (Fig. 3). This trend was observed for most avian species detected in Annaba's agglomeration (Fig. 2). However, the presence of the common swift, western cattle egret, black-headed gull, African blue tit, northern house-martin, barn swallow, yellow-legged gull, and the European starling seems to be unaffected by the presence of greenspaces in urban tissue or we did not have the statistical power to do it because such species are relatively rare. Unsurprisingly, the presence of the feral pigeon was higher in built-up areas, but its abundance was higher in intermediate environment. This species is emblematic of urban areas, as it is present in all cities all over the world [5], and some studies start to understand why this species is well acclimated or adapted to urban environments [21]. The abundances display the same pattern as the presence of avian species. Overall, our study in a Northern African city reinforces the idea that greenspaces significantly increase species richness [11,22], even if a minority of species is more detected in towns. This increase can be due to the characteristics of species that are adapted to their original environment and still failed to adapt or acclimate to built-up environment within the urban environment, which imposes sometimes very contrasted constraints as compared to the original one [10]. Alternatively, this increase in biodiversity may be the result of a better detection of avian species in open areas such as greenspaces. Future studies should try to understand the different patterns observed among towns all over the world and the role and the shape of greenspaces on biodiversity. Indeed, the size of the greenspaces could be a crucial factor for avian diversity and abundances [23] and, therefore need, to be experimentally investigated in future studies.

One species, the collared dove, is particularly interesting, because it is the predominant species in Annaba's agglomeration, whatever the environment. Indeed, even if it is more present in greenspaces, it is also largely present in urban environments (found in more than 85% of urban sites), with important abundances. Therefore, this species seems to be relatively poorly affected by the built-up environment. This result on the collared dove confirmed the results found in a previous study in Algeria [24], which suggests that this species exhibits an extraordinary hardiness and adaptability. This allows it to cope with all environmental constraints, and colonize all habitats in Southern Europe and Northern Africa. In this latter region, it constitutes a serious competitor of feral pigeon in urban areas. The reasons for its rapid expansion in Algeria are poorly known [24,25]. It is possible that the populations of such species are generalist and have undergone some phenotypic adaptations to cope with the heterogeneous environment. In order to better understand the success of this species in cities, it is now needed to investigate the phenotypic diversity and the population dynamic of this species in northern Africa [26–28].

As predicted, the season of data collection is an important factor to consider when investigating the impact of the avian environment on avian diversity. Indeed, our results show that species richness in one site is lower in the winter season as compared to the breeding season (Table 2 and Fig. 3). This is certainly due to the presence of migratory birds that are not present in the site during the winter season, but also to an increase in local movements during the breeding season, which may increase species detection [29,30]. In our case, even if the difference between environments is more important in the breeding season (Fig. 3), we found a similar pattern on species richness among environments (greenspaces, intermediate and built-up) for the winter and breeding seasons. Therefore, even in winter, we were able to observe that built-up areas have less species in lower abundances than intermediate and greenspaces environments.

In conclusion, our work reports for the first time, in a Northern African city, the importance of greenspaces within urban tissues for avian biodiversity. Indeed, our work first confirmed that greenspaces within urban areas significantly increase species richness and abundance of birds. Second, it shows that the season can profoundly affect species detection, and our study underlines the need for collecting data during the breeding season to increase the statistical power to detect environmental effect on such indicators of avian biodiversity.