Differences in the flight initiation distance near expressways and in open farmland in wintering Common Buzzards (Buteo buteo)

Abstract

The development of road networks has given rise to man-made wildlife habitats. Birds of prey use roadsides and highway infrastructure as attractive foraging sites. In order to measure reactions to human disturbance, the flight initiation distance (FID) of Common Buzzards (Buteo buteo) foraging on expressways and in open farmland was investigated. The GLM model showed that FID was affected by habitat types, distance from the nearest building and distance from the nearest tree. Common Buzzards have a shorter FID on expressway than in farmland. The distance from the nearest building and the distance from the nearest tree were positively correlated with FID. Our results confirmed the behavioural plasticity of these birds and may provide a theoretical basis for measures aimed at reducing the risk of collisions with moving road vehicles.

Keywords:

• Bird of prey
• foraging site
• human disturbance
• perching
• escape distance

Introduction

The human transformation of wildlife habitats is a widespread and ongoing phenomenon in the modern world (Sanderson et al. 2002; Steffen et al. 2015; Elhacham et al. 2020). One of the most attractive scientific issues that researchers are addressing is the impact of roads on wildlife (Pagany 2020). This is due to the development of road networks (especially two-lane expressways and controlled-access highways) expressed in terms of their increased density, area occupied and vehicle traffic. Roads generally have a negative impact on birds, primarily contributing to increased mortality as a result of collisions with vehicles (Fahrig & Rytwinski 2009; Benítez-López et al. 2010). However, some species of birds can adapt to the disturbed environments near roads, especially to the high level of noise pollution caused by the road traffic. These birds are usually consumers of invertebrates, which provide good foraging in the neighbourhood of roads (Wiącek et al. 2015). Birds of prey may also use roadsides and the highway infrastructure as attractive feeding habitats. Especially during poor weather in winter, substantial numbers of buzzards (Kitowski 2000; Wuczyński 2003), kestrels (Mülner 2000; Krasoń & Michalczuk 2019) and black kites (Meunier et al. 2000) can be seen along roads. The conditions in this type of habitat are favourable for the birds: the thinner snow cover, which makes food more accessible (Sonerud 1986; Dobler et al. 1991), more opportunities to obtain food in the form of roadkill (e.g. Orłowski & Nowak 2006; Borkovcová et al. 2012; Garrah et al. 2015), and small mammals living on roadsides (Meunier et al. 1999; Ruiz-Capillas et al. 2013). The road infrastructure also provides plenty of perching sites (Meunier et al. 2000; Wuczyński 2003; Krasoń & Michalczuk 2019), and as birds frequently change them, they can scan a larger area and consequently enhance their hunting success (Kitowski 2000).

Flight initiation distance (FID) is a simple behavioural measure of birds’ reactions to the proximity of humans or other predators (Stankowich & Blumstein 2005; Cooper & Wilson 2007). The decision to initiate escape in response to a disruptive stimulus is determined by a compromise between costs and benefits, where the perceived risk (cost) of an imminent threat becomes greater than the benefits of maintaining regular activity (Cooper & Blumstein 2015). FID offers a clear-cut, easily measurable change in animal behaviour (Beale 2007) and is likely to vary in cost-benefit assessments across individuals. The different responses to disturbances are due not only to biological differences among individuals, but also to the properties of the disturbance itself and the environment in which the interaction takes place. For example, the FID of a particular species relative to a human is shorter in urban than in rural areas (Møller 2008b; Bjørvik et al. 2015; Samia et al. 2015, 2017; Morelli et al. 2022) – this may be mediated by bird feeders (Møller et al. 2015) and cold winters (Møller et al. 2013). In some species, FID in response to vehicles has also been shown to be shorter than in response to pedestrians (McLeod et al. 2013), and the speed of vehicles is negatively correlated with the flight distance of birds (Legagneux & Ducatez 2013).

FID is closely correlated with synanthropization, while the proximity of humans as well as the various types of environmental disturbances they cause shorten the escape distance (Møller 2008b; McLeod et al. 2013; Bjørvik et al. 2015; Samia et al. 2015, 2017; Morelli et al. 2022). The aim of our research was to determine factors affecting the FID of wintering (non-breeding) Common Buzzards Buteo buteo (henceforth: Buzzard(s)) in the vicinity of expressways and in open farmland. This is the most common nesting bird of prey in Poland (Chylarecki et al. 2018), and its breeding season of a single brood begins in April (Cramp & Simmons 1980). During the wintering season, the Buzzard is the dominant species in the raptor assemblage in east-central Poland (Kasprzykowski & Rzępała 2002). It is a sturdily built, medium-sized to large diurnal bird of prey, with a total length of about 51–57 cm and a wingspan of 113–128 cm. Adults typically weigh 800–1200 g, while females are heavier than males by 20–30% on average (Cramp & Simmons 1980). In winter, Common Buzzards feed on small mammals, mainly Common Voles Microtus arvalis (Kowalski & Rzępała 1997).

Our hypothesis is that the FID of Buzzards depends on the habitat type. We anticipate that birds foraging on expressways will have a shorter FID than birds in farmland and that the distance from the hunting site to the nearest building and nearest tree will govern the initiation of escape. We also take into account the starting distance (SD), which is the initial distance between the bird’s position and that of the observer (Blumstein 2010). We believe that comparative research on the FID for this raptor could provide an answer regarding the plasticity of Buzzard behaviour in response to anthropogenic environmental disturbances. It will also be the basis for planning measures aimed at reducing the risk of vehicle collisions with birds of prey, which could have potentially catastrophic consequences for road users.

Material and methods

Study area

The research was carried out in north-eastern Poland, in two types of habitat characterized by different levels of human impact. The former areas were situated in the immediate vicinity of expressway S8 between the towns of Wyszków and Zambrów, a distance of about 60 km. The studied section of the expressway was a two-lane dual carriageway. Between the carriageways, each 10.5 m wide, there was a central reservation 12 m in width. The expressway was bounded along each side by a 2.5 m tall wire mesh fence supported on metal posts inserted every 3 m. There was also a single-lane service road running parallel to the expressway, along which the observer moved. In 2021, an average of 26 439 vehicles moved along the surveyed section of the expressway per day (GDDKiA 2021). The COVID-19 pandemic presumably had only a limited impact on the volume of motor vehicle traffic on the studied section of the S8. Indirect evidence for this is the average increase of 27% in vehicle traffic on international roads in Poland in 2021 compared to 2015 (GDDKiA 2021).

Along the S8, Buzzards perched on the wire-netting fence posts, also on road signs and lamp-posts, which were most often located near the eight road junctions. The latter areas were located in the open agricultural landscape up to about 15 km from that expressway, and covered a total area of about 280 km². The study area is a mosaic of meadows, pastures and arable land with isolated trees and shrubs. The agricultural landscape here lies almost entirely within the Biała Forest, a fragmented forest complex dominated by pine monocultures.

Data collection

Observations in the two habitat types were conducted from early November to early March in the winter seasons of 2020/2021 and 2021/2022. This period is best suited to recording wintering birds because the number of buzzards increases in the second half of October as a result of their autumn migration, whereas the spring migration, taking place between late February and early March, is less pronounced (Tomiałojć & Stawarczyk 2003). Buzzards were observed at regular, approximately weekly intervals on each habitat type. In 2020/2021, FID was tested in 139 birds (80 by the roadside and 59 on farmland), and in 2021/2022 in 132 birds (76 by the roadside and 56 on farmland). FIDs were always recorded during the same part of the day, i.e. from 08:00 to 14:00 hrs. Buzzards were not individually marked, so multiple observations of the same individual were unavoidable. Nevertheless, all the records were treated equally, i.e. each recorded individual was counted separately (see also Peck et al. 2014; Rothery et al. 2017; Golawski & Sytykiewcz 2021). In addition, individuals from eastern European populations may appear during the winter in north-eastern Poland, especially when the weather conditions in their nominal wintering grounds deteriorate (Kasprzykowski & Rzępała 2002; Wuczyński 2003; Baltag et al. 2018). Waves of such individuals arriving from the east, which mix with birds wintering locally, may also reinforce the independence of observations.

FID was measured in that randomly selected buzzards were approached using a modified version of the technique developed by Blumstein (2006). During the approach, the observer counted paces (of equal length, about 0.8 m) at a steady speed, approximately 1 pace/second. The distances measured in paces were converted to metres. The distance from the observer to the bird when it first took flight was recorded as the FID, while the starting distance SD was the distance between the point from where the observer started walking towards the bird and its position. If the bird was perching above ground, the height of its perch was also recorded to the nearest metre. Birds on the ground were assigned a height of 0. Together with FID, the following parameters were also recorded: date, time, place of observation (including the coordinates of the point from which the bird took off) and habitat type. The distance from the hunting site to the nearest building and the nearest tree was determined from geographical coordinates using QGIS software tools. FID was calculated as the square root of the sum of the squared horizontal distance and the squared height (Blumstein 2006).

Statistical analyses

A general linear model (GLM) with gamma distribution and log link function was used to find the effect of habitat types, SD, distance from the nearest building and distance from the nearest tree on Buzzard FID (Table I). The models were constructed using the glm function in the lme4 package for R (Bates et al. 2015). The two habitat types – road and farmland – were treated as categorical factors. SD, the distance from the nearest building and the distance from the nearest tree were treated as continuous predictors and log-transformed. FID was the continuous dependent variable. An information-theoretic approach (AIC) was used to compare the competing models (Burnham & Anderson 2002) and to analyse model support using the AIC value, which corrects for small sample sizes and evaluates the strength of evidence for each model using normalized weights (AICwt). The models selected with the smallest AIC as being the best of all the models were compared; only the models within a ∆AIC of < 5.00 were taken into consideration. All data were analysed in the R environment (R Core Team 2023). The values reported are the mean ± Cl. Only those results with a probability of α ≤ 0.05 were assumed to be statistically significant.

Table I. Factors for analysing effects on Common Buzzard flight initiation distance.

Code	Description
Habitat	habitat (farmland/expressway)
StartDist	starting distance (SD) – distance from where the observer started walking towards the bird and the bird’s position
BuildDist	distance from the hunting site to the nearest building
TreeDist	distance from the hunting site to the nearest tree

Results

A total of 271 FIDs were recorded for Buzzards in their foraging areas. Most often, the birds initiated flight at FIDs between 100 and 200 m (Figure 1); the mean FID was 203.30 m (SD = 107.55 m, min = 44 m, max = 528 m). Models containing four predictors were created on the basis of Akaike’s information criteria (Table I). The best model of the 16 a priori models with the highest ranking received 57% support (sum of AIC_cwt, Table II) and included all the factors analysed: habitat types, SD, distance from the nearest building and distance from the nearest tree. The GLM model showed that three parameters had a significant influence on FID (Table III). Buzzards took flight from a shorter distance on the expressway than in farmland (Figure 2). The difference in the mean distance between these categories was 219.7 m, a figure less than half of that on sites situated close to the expressway. Distance from the nearest building and distance from the nearest tree were positively correlated with FID. SD did not predict Buzzard escape responses to an approaching person (Table III).

Figure 1. Data distribution of flight initiation distance (m) of Common Buzzards Buteo buteo.

Figure 2. Mean number (dots), 95% confidence limits (vertical lines) and kernel distribution (violins) of recorded flight initiation distances in two types of habitats.

Table II. Results of models describing the influence of different factors on Common Buzzard flight initiation distance. Degrees of freedom (df), model log-likelihood (LL), corrected AIC criterion (AIC), difference between the model and the best model in the data set (Δ AIC), and weight for the model (AICwt) are shown.

Model (fixed effects)	df	LL	AIC	Δ AIC	AICwt
Intercept+Habitat+StartDist+BuildDist+TreeDist	6	−1504.11	3020.5	0.00	0.570
Intercept+Habitat+BuildDist+TreeDist	5	−1505.70	3021.6	1.08	0.332
Intercept+Habitat+StartDist+BuildDist	5	−1507.46	3025.2	4.61	0.057

Table III. Estimates of GLM model coefficients for the best model of factors affecting FID.

Parameters	Estimate	Standard Error	t value	P value
Intercept	5.411	0.165	32.782	<0.001
Habitat: expressway	−0.828	0.038	−21.515	<0.001
BuildDist	0.225	0.053	4.218	<0.001
TreeDist	0.073	0.028	2.611	0.010
StartDist	−0.088	0.048	−1.837	0.067

Discussion

Our study showed that the mean FID was at least several times greater than that of other accipitrids with a body mass comparable to that of Buzzards during the non-breeding season (Livezey et al. 2016). However, it is hard to draw unequivocal conclusions from this fact, because most of the studies that we compared ours to were based on a much smaller research sample. In the case of six Australian accipitrid species, the samples consisted of only a dozen or so birds, and for four of these species only single results were obtained (Weston et al. 2012). The same applies to Europe, where the mean FID for Buzzards was more than four and a half times shorter, a result based on a sample more than 30 times smaller than ours (Møller 2008a). Larger research samples were obtained in the case of Rough-legged Buzzard Buteo lagopus in North America, and here the mean FID was the closest to our results (Holmes et al. 1993). Another factor affecting the results of other researchers could have been the magnitude of environmental disturbances, e.g. the density or presence of passers-by (Weston et al. 2012; McGiffin et al. 2013), or even the possibility of a bird managing to learn to recognize particular persons (Levey et al. 2009).

We also found that the FID of Buzzards in localities close to expressways was less than half the figure for farmland. The explanation for this difference should be sought in the nature of the environment immediately adjoining busy roads. This environment is highly disturbed by human activities: intensive vehicle traffic, which is a source of noise, and the drastically transformed landscape with wide stretches of tarred road, embankments and drainage ditches, all enclosed on either side by a wire-netting fence. Such features of this environment can indirectly lead to a reduction in FID, because the birds perceive them as barriers to the free movements of pedestrians, who are then potentially less of a threat (Weston et al. 2012). Habitats associated with road infrastructure are also attractive foraging areas, where the ready availability of lookout posts means that energy can be saved during hunting, an aspect particularly important for maintaining the energy balance in winter (Meunier et al. 2000). The economic theory of escape postulates that by delaying their escape flight after having perceived a potential threat, birds can devote more of their limited energy resources to other key behaviours, such as foraging (Ydenberg & Dill 1986; Cooper & Frederick 2007). The reduced FID therefore reflects a compromise on the part of the bird between the advantages of remaining in situ in the form of time devoted to foraging and saving energy, and the costs of the risk of predation and death (Cooper & Blumstein 2015). So if another lookout post is available nearby, a Buzzard will delay taking off on seeing a pedestrian approaching, thus saving the maximum amount of energy. The modification of escape behaviours in Buzzards may also be due to the heavy traffic on the road and the consequent high level of noise. Where an environment is subject to strong human pressure, e.g. in urban areas, where conditions may be similar to those described above, birds generally have shorter average FIDs (Møller 2008b; Bjørvik et al. 2015; Samia et al. 2015, 2017; Morelli et al. 2022), especially during frosty winters (Møller et al. 2013).

This study demonstrated a positive link between Buzzard FID and the distance to the nearest building. This may be evidence of individual habituation and greater tolerance among Buzzards of the presence of people, i.e. phenotypic plasticity (Blumstein 2016), owing to the heightened activities of people around such artificial components of the landscape like barns, houses and farm buildings. Similar results were obtained in other bird species, where FIDs on perceiving a human were shorter in anthropogenically more disturbed areas (Møller 2008b; McGiffin et al. 2013; Bjørvik et al. 2015; Samia et al. 2015, 2017; Tryjanowski et al. 2020; Morelli et al. 2022). Changes in escape behaviour may be justified by the possibility of strong selection against timid individuals (Carrete et al. 2016) or the recruitment of less sensitive individuals in more disturbed environments (Weston et al. 2012; Carrete & Tella 2013). In addition, this kind of environmental disturbance may be a source of readily available food and encourage the birds to approach buildings, particularly when snowy winters make finding food difficult (Sonerud 1986; Dobler et al. 1991). In such instances, FID in many other bird species may also fall when a human being comes into sight (Møller et al. 2013, 2015). That such a situation may occur among Buzzards is indicated by their opportunistic manner of feeding (Oro et al. 2021) and the fact that in winter, their diet may contain domestic fowl or other synanthropic bird species (Goszczyński et al. 2005). At such times, one observes an apparent increase in the Buzzards’ tolerance towards disturbances, resulting from the limited availability of food in places distant from human habitations (Owens 1977).

The second positive association, between FID and distance to the nearest tree, can be interpreted in the case of Buzzards with respect to trees as sites providing protection and trees as lookout posts. The link between the escape distance and the distance to the nearest shelter has been widely described in the literature for a variety of animals, including fish, reptiles and mammals. As in our study, escape among various species of vertebrates was initiated the sooner, the greater the distance to a shelter (Dill & Houtman 1989; Dill 1990; Bonenfant & Kramer 1996; Cooper 1997, 2007; Cooper & Wilson 2007). The same relationship has been found in birds. In support of the predictions of the optimal escape theory (Ydenberg & Dill 1986; Cooper & Frederick 2007), Morelli et al. (2022) demonstrated that the FID of birds in urban greenery increases with distance to the nearest accessible shelter. This suggests that birds also base their decisions to escape on the relative time needed to reach the nearest shelter, which in turn affects the assessment of the risk of danger from a potential predator. Hence, Buzzards may be more afraid of a pedestrian when they are farther away from the nearest tree. On the other hand, the shorter FID of Buzzards when a tree (lookout post) is closer to hand, can be interpreted as the consequence of saving energy and taking full advantage of the time available for foraging (Ydenberg & Dill 1986; Cooper & Frederick 2007). Maintaining the proper energy balance is absolutely crucial in winter. The lack of food or its limited availability in combination with inclement winter weather can lead to substantial energy losses in animals and may be the cause of their mortality (Sonerud 1986; Schmidt-Nielsen 1997; Mülner 2000; Blix 2016). Under such conditions, Buzzards, being able to fly to a nearby “safe haven”, i.e. another tree, probably postpone the decision to escape.

In the Buzzards that we studied, SD did not affect the FID. However, most other studies do not corroborate this finding (Blumstein 2003, 2010; Mikula et al. 2018, 2022; Morelli et al. 2022). Nonetheless, when analysing the behavioural decisions of birds, it is very important to take their context into consideration (Beale 2007). Therefore, the lack of a correlation between SD and FID in the Buzzards that we studied may be due to factors that hamper continuous monitoring by the observer approaching them and thus modify their escape behaviour (Blumstein 2010; Samia et al. 2017). Not to be overlooked is also the fact that some of the Buzzards in the population we studied are recruited from eastern Europe (Kasprzykowski & Rzępała 2002; Wuczyński 2003; Baltag et al. 2018), where they will have been inhabiting areas less densely populated by humans and less disturbed by human activities, which could have a bearing on their escape behaviours.

Conclusion

The present study has shown that the readiness of Common Buzzards to take a risk is mediated by the place where they happen to be (expressway vs. farmland), and also by the distances to the nearest building and to the nearest tree. Our results confirmed the behavioural plasticity of these birds and may provide a theoretical basis for measures aimed at reducing the risk of collisions with moving road vehicles.

Ethical standards

The authors declare that this publication complies with the current laws of the Ethical Committee.

Acknowledgments

We thank Peter Senn for the English language editing and Przemysław Obłoza for help in statistical analyses.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Dataset supporting reported results can be found at the Mendeley Data Repository: https://doi.org/10.17632/k7jwk6pgpc.1

Additional information

Funding

This work was financed by the Ministry of Science and Higher Education, Poland [152/23/B].