Colonizers on the road: European bee-eaters shift nest-site selection to roadside habitats

Nest-site selection in birds results from a compromise among nest-predation risk, food availability, and microclimatic requirements (Martin, 1993, 1995). The European bee-eater Merops apiaster, an aerial-foraging insectivorous bird, breeds gregariously by digging tunnel nests at a sufficient depth underground to mitigate the impact of surface temperature (Casas-Crivillé & Valera, 2005; White et al., 1978). Additionally, substrate quality is crucial, requiring soils that are soft enough to be excavated yet firm enough to prevent nest collapse (Heneberg & Šimeček, 2004; White et al., 1978).

Doñana is located in the SW of the Iberian Peninsula (Figure 1a) and is mainly composed of two large flat environments: a seasonal freshwater marshland on clayey soils and an extensive system of eolian dune fields hosting Mediterranean woodlands and scrublands (Valverde, 1963). This area includes Doñana National Park, where the European bee-eater has historically been a common nester along the ecotone between marshes and eolian dune fields (locally known as La Vera), on sandy islands within the marshes, and occasionally in other locations, such as the bottom of dry seasonal ponds (Valverde, 1960; Figure 1b). In all these environments, the mixture of sandy and clay soils provides a suitable substrate for nest excavation. Historical records in the field notebooks of Doñana's ornithologists report bee-eater breeding colonies in the National Park and two Special Protection Areas (known as El Acebuche and La Rocina). During the 30-year period between 1970 and 1999, 17 colonies (ca. 77% of the colonies) and 1543 nests (ca. 87%) were recorded in the ecotone. Interestingly, only one colony was reported in 1993 on a roadside in La Rocina, with approximately 40 nests, representing 2% of the nests recorded in the historical data set (Figures 1b and 2).

Only more recently (in the year 2000), 56 nests were reported for the first time along the verges of the Palace Road, a gravel road leading to the Palace of Doñana, and bee-eaters have continued to breed there every year since then (Figures 1b–d and 3). In the following years, bee-eaters also started nesting increasingly along the verges of other roads, many outside the National Park (Figure 1e). Nesting on roadsides is not entirely novel for this species; however, most records in the scientific literature refer to anecdotal observations (Gyurácz et al., 2013; Urbán et al., 2013). The aim of this note is to describe this change in nest-site selection, suggest possible causes, and discuss the implications for bee-eater populations.

In 2011, 2016, and 2022, we carried out three complete censuses of bee-eater active nests across the ecotone and the eolian dune fields (Figure 1c–e). Within the National Park and the Special Protection Areas, in 2011, we recorded 327 roadside nests from a total of 533 nests counted for the whole area (ca. 61%), and in 2016, we documented 190 roadside nests from a total of 202 nests (ca. 94%), and in 2022 we identified 546 roadside nests from a total of 561 nests (ca. 97%; Figures 1c–e and 2). Additionally, in those same years, we counted a total of 357 nests outside the National Park and the Special Protection Areas, of which 343 (ca. 96%) were located along roads (note that we do not have historical data outside the park, Figure 1e).

Most roads in the region were built between the late 1960s and early 1970s, and thus, they have been available as nesting sites since the beginning of the data series collected in the Field notebooks of Doñana's ornithologists. However, European bee-eaters did not start breeding on the Palace Road until 2000, coinciding with an increase in the local livestock pressure in the ecotone, which had been the main historical breeding area for bee-eaters. In Doñana, the absence of topographic slopes implies that European bee-eaters excavate their nests directly in the ground, which are easily collapsed when trampled by livestock, including both free-ranging horses Equus caballus and cattle Bos taurus. Nest trampling by livestock is a well-documented cause of breeding failure in European bee-eaters (Arbeiter et al., 2014; Thiollay, 2006), and we have observed this issue in our study area as well. In Doñana, different areas of the National Park are divided by livestock fences (Botting et al., 2023), and livestock density is artificially managed within these areas (Soriguer et al., 2001). In the ecotone area, this increase was remarkable in the years leading up to 2000, with the annual mean number of cattle rising 3.3 times over a short time period: from an average of 293 individuals (SD = 83) in 1982–1986 to 973 (SD = 185) in 1995–1999, peaking at 1203 in 1996 (Unpublished data from Livestock Management Plan of Doñana). This high density has persisted in recent years. However, such an increase in livestock did not occur in the area around the Palace Road (mean = 94, SD = 19 in 1982–1986; mean = 95, SD = 30 in 1995–1999). This difference in livestock pressure between the ecotone and the area around the Palace Road, along with the potentially differing rates in nesting failure due to trampling, could underlie the reported change in nest-site selection by local bee-eaters. On roadsides, the mixture of loose eolian sand with more compact road-construction materials provided a very suitable substrate for nest excavation by bee-eaters in an otherwise sandy environment. In addition, large herbivores in Doñana have been reported to avoid roads consistently (D'Amico et al., 2016; Suárez-Esteban et al., 2014).

Additionally, another potential factor favoring this shift in nest-site selection may be the difference in predation pressure between roadsides and natural habitats. In Doñana, we have observed that the main predators of bee-eater nests are primarily the European badger Meles meles, but also the red fox Vulpes vulpes and the large gray mongoose Herpestes ichneumon. At least the first two species have been described as common predators of bee-eaters in other regions (Gyurácz et al., 2013), and we know that, at least in Doñana, badgers tend to avoid roads (Revilla et al., 2001; Suárez-Esteban et al., 2013).

Considering that nest failure is a powerful selective pressure on bird breeding habits (Blackmer et al., 2004; Martin, 1995) and that previous experience, along with the success of conspecifics, serves as a key cue for nest-site selection in subsequent years (Danchin et al., 1998; Hatchwell et al., 1999; Ibáñez-Álamo et al., 2015), this factor may have promoted the expansion of bee-eaters into other roads of the region. In addition, natal philopatry has been described in this species (Lessells et al., 1994; Lessells & Krebs, 1989) and may have contributed to the consolidation of roadside-nesting behavior.

Interestingly, European bee-eaters have not just started nesting on every roadside in our study area. In Doñana National Park, there are more than 1100 km of sandy roads with a traffic intensity of less than one vehicle/day, whereas there are only a few gravel or paved roads, all of them with medium or high traffic volumes (i.e., ca. 50–400 and >3000 vehicles/day, respectively; Román et al., 2010). To date, we have only found bee-eater nests on both gravel and paved roads with medium traffic volume, but not on high-traffic paved roads or sandy roads. It is important to highlight that the local road network has remained largely unchanged during the period in which this shift in nesting-site selection has occurred (Román et al., 2010).

Breeding on the roadside has its consequences. In a previous study (Blas et al., 2016), we observed that roadside bee-eaters responded to the passing vehicles as if they were predators and therefore had higher frequencies of alarm calls and flock flushes than individuals nesting on the ecotone. The risk of being road-killed is also greater when nesting on the roadside and, in fact, every year several individuals are found dead in roadside colonies (Blas et al., 2016). However, an increase of more than double in the nestling-feeding rates was observed in the roadside nests, which could result in a higher reproductive success (Blas et al., 2016). This cost/benefit balance may have a threshold, so it is possible that this is the reason behind the absence of colonies on paved roads with high traffic volume.

Transportation infrastructure, and especially roads, plays a crucial role in shaping landscapes and connecting human populations (Laurance, 2015; Laurance et al., 2014). The impact of these infrastructures on wildlife, and especially bird populations, has become a topic of increasing concern in recent years (Johnson et al., 2022; Kroeger et al., 2022). The presence of roads alters the original state of the landscape by creating physical barriers, edge-effect zones, and modifying surrounding environments (Forman et al., 2003; van Der Ree et al., 2015), but it also provides nesting opportunities on roadside habitats (Morelli et al., 2014). In this note, we reported a radical change in nest-site selection from natural environments toward roadsides. Long-term monitoring of this type of change is key to developing effective conservation strategies in order to maximize benefits and mitigate impacts of transportation infrastructure on avian communities.

Jacinto Román and Marcello D'Amico conceptualized the study. Jacinto Román led the methodology and investigation. Giulia Bastianelli, Miguel A. Suárez-Couselo, and Marcello D'Amico performed the censuses. Miguel A. Suárez-Couselo helped with the collection of historical data. Jacinto Román and Marcello D'Amico drafted the original manuscript. All authors revised and approved the final version of the manuscript. Julio Blas, Eloy Revilla, and Marcello D'Amico secured funding for the censuses.

The authors declare no conflicts of interest.