Co-occurrence with Greylag Geese Anser anser and other factors affecting vigilance in White‑fronted Geese Anser albifrons during their spring staging in Central Europe

Vigilance, a crucial factor affecting the behaviour of birds, is key to the safety of migratory animal flocks against predators and other threats. We examined the vigilance of White-fronted Geese in the Biebrza Basin (NE Poland), one of their major stopover sites in Europe. The study took place during their spring staging, from the end of February until the beginning of April. We found that the vigilance of White-fronted Geese decreased in the centre of mixed flocks with Greylag Geese. This may be associated with the larger size and wider field of view of Greylag Geese, as well as their territorial behaviour, as some of them belong to the local breeding population. White-fronted Geese can benefit from the group-size effect when they form larger flocks. In the presence of a vigilance stimulus, the individuals in the centre of a flock raised their alertness levels. As the probability of detecting danger is higher in large groups, flock size negatively affected the birds’ vigilance. Owing to their different experience and roles (adults care for their broods by being vigilant), adult birds appeared to be vigilant on both grassland and arable land, whereas immatures were vigilant only on arable land, where exposure to predatory pressure was greater. We believe that the lower level of vigilance of White-fronted Geese in large and mixed flocks may give them more time for foraging and, therefore, improve their chances of survival and breeding success during the oncoming spring.


Introduction
Vigilance is an important factor associated with predation, which affects the survival of potential prey (Stephens et al. 2008). Foraging behaviour can be viewed as a trade-off between the risk of starvation, when too little time is spent foraging, and the risk of being killed by a predator, when too little time is spent being vigilant (Siekiera et al. 2020). Studies on vigilance are associated mainly with anti-predator behaviour during breeding (Kotrschal et al. 1998), but fewer analyses of this behavioural aspect have been carried out in relation to flocking migratory birds, such as waders (Beauchamp 2008;Beauchamp 2015). Flocks detect threats faster and more effectively than individual animals (FitzGibbon 1989;Lima 1994;Panday et al. 2021), yet this takes up time that could otherwise be devoted to foraging or resting. Flock formation can be beneficial, as more individuals are then vigilant (Stephens et al. 2008). Living in a group with other animals and the benefits accruing from such behaviour are referred to as the "many eyes" hypothesis (Lima 1995), described in some studies on mammals and birds (Dehn 1990;Quenette 1990;McNamara and Houston 1992;Roberts 1996). In fact, animals can increase their chances of avoiding danger by monitoring their surroundings or waiting for signals from other flock members (Treves 2000;Panday et al. 2021).
Safety considerations are paramount in the functioning of migratory birds and to a large extent determine their behaviour (Alerstam and Lindström 1990). Numerous studies have examined the vigilance of nesting geese (e.g. Lazarus and Inglis 1978;Elgar 1989;Gauthier and Tardif 1991;Kristiansen et al. 2000), but fewer have focused on their behaviour during migration (e.g. Beauchamp 2015). The issue of vigilance during spring migration is a highly fascinating one, but as yet insufficiently understood. Birds need to forage more in preparation for the oncoming breeding season, so the time available for vigilance may be constrained (e.g. Polakowski et al. 2019). It can be expected that foraging in large, often mixed flocks, when birds benefit from each other's vigilance, may affect the amount of time spent foraging.
One of the most numerous Arctic goose species, migrating in spring from its western European wintering grounds to its breeding areas in Siberia, is the Greater White-fronted Goose Anser albifrons (hereafter White-fronted Goose) (Madsen et al. 1999;Fox and Leafloor 2018). According to estimates from the wintering areas, numbers of these birds are currently increasing (Fox et al. 2010). One of their major spring stopover sites is the Biebrza Basin in NE Poland, hosting up to 10% of the overall wintering population in Europe and about 70% of the country's population (Polakowski et al. 2011;Polakowski and Kasprzykowski 2016). Flocks are relatively numerous and usually mixed with other goose species, i.e. Tundra Bean Goose Anser serrirostris and Greylag Goose Anser anser. While some of the latter are migratory, others are breeding residents (Polakowski et al. 2011;Polakowski and Kasprzykowski 2016). Although more has been discovered about the behaviour of White-fronted Geese and Greylag Geese independently (Kotrschal et al. 1998;Nilsson et al. 2002;Shimada and Shimada 2003;Olsson et al. 2017;Polakowski et al. 2021), little is known about the behaviour of mixed-species flocks, or, among others, their coexistence, vigilance and safety (e.g. Polakowski and Kasprzykowski 2016).
Our objective was to test the hypothesis that Whitefronted Geese benefit from the company of Greylag Geese (see Kristiansen et al. 2000). We hypothesized that the presence of Greylag Geese in a mixed flock could reduce the vigilance of White-fronted Geese. Furthermore, according to the "many eyes" hypothesis, we anticipated that flock size would influence the behaviour of geese-with the time spent by an individual being vigilant decreasing as the flock size increases (Lima 1995) due to higher probability of detecting threats in larger groups. We expected that the flock size effect would also be noticeable in Greylags' absence in the flock. Additionally, we expected that the position of an individual bird in a flock and the foraging habitat would influence its vigilance due to different levels of exposure to predators in such conditions. Our final hypothesis was that the immatures, due to their lack of experience and often being under parental care, would be less vigilant than adults.

Study area
The study was conducted in the Biebrza Basin (NE Poland), in the valleys of the rivers Biebrza and partially Narew. The research area has been described in detail and depicted on maps in several papers (Polakowski and Kasprzykowski 2016;Polakowski et al. 2018Polakowski et al. , 2019Polakowski et al. , 2021. It covers an area of ca 260,000 ha, most of which is protected as a national park. Grassland covers 52.0% of the Basin and arable land 25.2% (Polakowski and Kasprzykowski 2016). The fields are smaller than the grasslands and both habitats are mostly open, with only a few clumps of bushes or single trees locally. The other habitats are mainly villages and forests. The Biebrza Basin is one of the most important spring staging areas for White-fronted Greese in central Europe. The numbers of geese staging in the Biebrza Basin during the spring migration have increased significantly in recent decades and now reach 100,000 individuals. As many as ca 8‚000 Greylag Geese on average may be staging in the Basin per season (Polakowski et al. 2011) and about 20% of all individuals occurring here in the spring migration period breed in this part of Poland (Polakowski and Kasprzykowski 2016).

Data collection
The field work was done during the spring migration period (from the end of February until the beginning of April), whereas the annual migration peak depended on the harshness of the winter. The methods used in this study were described in detail by Polakowski et al. (2021). 271 Whitefronted Geese (181 adults and 90 immatures) in 66 flocks were filmed using the digiscoping method (see Leary  . We did not determine sex of the studied geese due to considerable complication of such identification in the analysed movie clips. Greylag Geese were recorded in 38 flocks of Whitefronted Geese. The mean number of White-fronted Geese per flock was 1015 birds. Each flock occupied a definite space, separate from and at some distance to other groups. Filming took place at an average distance of a few hundred metres with the use of a Nikon Coolpix A10 camera combined with a Swarovski ATS 80HD spotting scope fitted with a 25-50 × zoom lens (Fig. 1). Each film was 180 s long and all were recorded by the same person (MP). 173 films were made on low-lying and seasonally flooded meadows or pastures (hereafter grasslands, either mown or grazed) and 98 on arable lands. To avoid selectivity bias, birds were recorded at random in different habitats, at different points in the flock, periods and sites within the study area. All the material was digitally analysed by the same person (MB) and the results were entered on a spreadsheet. The edge zone of the flock was determined as being up to 10 birds from the outermost individuals, while the others in the middle were classified as being in the centre. Such observations were performed and noted in the field as it was not always possible to determine this from the recordings; the films showed a zoomed, small part of the flock and usually just a few, occasionally up to a dozen birds. The measure of White-fronted Goose vigilance was the number of times the focal individual raised its head during one recording (alertness frequency). Birds usually raised their heads for a short time (mostly a few seconds), then quickly returned to their previous activity. We usually analysed 3-4 geese from each film (the range was 1-10 individuals). Using this indicator, i.e. the alertness frequency, one can show the extent to which various factors can interfere with foraging. Duration of vigilance (total amount of time with head raised) had been analysed before in the time budget of White-fronted Goose at the same stopover site (see Polakowski et al. 2021). The main predators affecting goose alertness in the Biebrza Basin were Domestic Dogs Canis familiaris, Red Foxes Vulpes vulpes, White-tailed Eagles Haliaeetus albicilla and above all humans. Dogs both caused alarm when approaching a flock and, occasionally, they chased geese.

Statistical procedure
All the data were analysed in the R environment (R Development Core Team 2018). The influence of habitat type (arable land or grassland), bird age (adult or immature), position in the flock (centre or edge), vigilance stimulus (predator or human present or absent), Greylag Goose (present or absent) and flock size on the alertness frequency were analysed using a generalized linear mixed model (GLMM; model 1) with a negative binomial distribution and logit link function (we checked the Poisson distribution and found overdispersion to be a problem). The vigilance stimulus was the factor that made the geese restless. Flock size was treated as a numerical factor (after logarithmic transformation), five other fixed effects were categorical factors, and random effects included flock identity nested within year. We also tested the following interactions: bird age × habitat type, bird age × vigilance stimulus, position × Greylag Goose, position × vigilance stimulus, habitat × Greylag Goose, vigilance stimulus × Greylag Goose. We also performed a separate analysis for White-fronted Geese alertness with the size of a flock in which Greylag Geese were absent (homogenous White-fronted Geese flocks, n = 108; model 2). Furthermore, we analysed White-fronted Geese alertness in the subset of data where Greylag Geese were present (mixed flocks, n = 163) and independent variables were the number of White-fronted Geese and the number of Greylag Geese (model 3). The model terms were tested by single term deletions using the likelihood ratio test (LRT), which compares the full model, according to the Akaike Information Criterion (AIC), to a reduced model, where the target variable has been dropped (drop1 function in R software) (R Development Core Team 2018). The models were constructed using the glmer.nb function in the MASS package for R (Venables and Ripley 2002). Multiple comparisons were performed with post-hoc Sidak tests (package emmeans in R) (Lenth 2020). The values were reported with 95% confidence limits. Only the results with a p value ≤ ɑ (0.05) were assumed to be statistically significant.
In the model 1 we tested interaction position × Greylag Goose and it was statistically significant (Fig. 2D, df = 1, LRT = 4.184, p = 0.041). Detailed interaction analysis (Sidak post hoc tests) showed that the presence of Greylag Geese did not affect vigilance at the edge of the flock (Fig. 2D, p = 0.794), but White-fronted Geese in the centre raised their heads less frequently when Greylag Geese were present (Fig. 2D, p = 0.006). Comparison of the centre and edge of the flock indicated a higher level of alertness on the edge only if Greylag Geese were present (p = 0.005). We found no such effect when Greylag Geese were not present (p = 0.999). The effect of a vigilance stimulus on the vigilance of birds also depended on the position of the bird in the flock ( Fig. 2E; the position × vigilance stimulus interaction was significant: df = 1, LRT = 4.517, p = 0.034). In the presence of a vigilance stimulus, White-fronted Geese exhibited a higher alertness frequency in the centre (Fig. 2E, p = 0.043) but not in the flock edge zone (Fig. 2E, p = 0.916). We also found the bird age × habitat interaction to be significant (Fig. 2F, df = 1, LRT = 4.593, p = 0.032). While there were no differences in vigilance between arable land and grassland in adult birds (Fig. 2F, p = 0.973), immature geese raised their heads less often in grassland than on arable land (Fig. 2F, marginally insignificant p = 0.076). Immature White-fronted Geese displayed a lower level of alertness in comparison to adult birds on the grasslands (Fig. 2F, p < 0.001), but we did not find any difference between the age groups on arable land (Fig. 2F, p = 0.864). The bird age × vigilance stimulus (df = 1, LRT = 1.496, p = 0.221), habitat × Greylag Goose (df = 1, LRT = 1.390,

Discussion
Our results showed that the vigilance of White-fronted Geese was connected with the size of a flock regardless of the cooccurrence with Greylag Geese. However, the influence of the Greylag Geese was evident in relation to the position of a bird in the flock. White-fronted Geese that were foraging in the centre of the flock were calmer in the presence of Greylag Geese than when they were absent. Some of these Greylag Geese probably belong to the local breeding population (as reported by Polakowski and Kasprzykowski 2016) and at this time of the year they are beginning to form pairs, with males displaying more aggressive territorial behaviour. Greylag Geese spend most of their time being alert and protecting their mates from predation and other males trying to mate with them (Cramp and Simmons 1977;Kristiansen et al. 2000). Additionally, due to usually lack of aggressive behaviour of Greylags towards other geese species in the flock, White-fronted Geese may benefit from the presence of persistently alert Greylag Geese, allowing them to spend more time foraging and resting during their migration (Polakowski et al. 2021). The physical characteristics of Greylag Geese, such as their larger body size (about 15% taller than White-fronted Geese) and wider field of view (Cramp and Simmons 1977), may also play an important role, as the Greylags can detect threats earlier, to the whole flock's benefit. Our results support the findings of Kristiansen et al. (2000), who described a similar phenomenon: the presence of Greylag Geese reduced vigilance levels in Greenland White-fronted Geese Anser albifrons flavirostris. Moreover, in such conditions the Greylags could also benefit from the proximity of the Greenland White-fronted Geese due to the large size of their flocks, which were able to warn of potential predators more effectively. Randler (2004a) showed that the Eurasian Coot Fulica atra benefitted from the company of feral Swan Geese Anser cygnoides in much the same way that we found. Other examples of co-occurrence advantages can be found among breeding birds nesting close to potential predators, which inadvertently protect their territories against bigger predators. Such relationships are observed in Snowy Owls Bubo scandiacus and Dark-bellied Geese Branta b. bernicla as well as in Woodpigeons Columba palumbus and Eurasian Hobby Falco subboteo (van Kleef et al. 2007;Bogliani et al. 1999).
We also noted that when flocks were larger, birds were less alert and could spend more time on other activities. This can be explained by the lower probability of being killed by a predator when a flock is bigger and a higher number of birds detect danger more quickly (Treves 2000;Beauchamp 2008;Panday et al. 2021). Studies on migratory finches showed the opposite effect: the success of a predator's attack increased with finch flock size, whereas the risk of an individual bird being hunted was not correlated with flock size (Lindström 1989). However, Cestari et al. (2020) obtained results similar to ours. Studying waders in Brazil, they found that congregating in large mixed-species flocks was beneficial for safety reasons: the larger the flock size, the lower the vigilance level of the birds (Lilleyman et al. 2016). Moreover, in Germany, a negative correlation between group size and alertness was also found in Mallards Anas platyrhynchos (Randler 2004b), which likewise supports our findings.
The behaviour of geese is also influenced by potential vigilance stimuli. Our results show that geese are more vigilant when they are at the edge of a flock, but when a predator approaches, birds in the usually safer central part of the group raise their alertness levels, too. This could also explain the lack of effect of Greylag Geese on the alertness of Whitefronted Geese at the edge of a flock. Furthermore, it is likely that the centre-edge effect may be predator-dependent. Aerial predators like White-tailed Eagle Haliaeetus albicilla may attack the middle of the flock, although we did not observe such behaviour during this study. Nonetheless, birds at the edges are the first to be exposed to pressure from mammals, like foxes or dogs, whereas the chances of gaining time for escaping and surviving are greater in the centre. Therefore, birds in the centre of a flock can spend more time feeding. These findings support the general statement that it is safer to be in the middle rather than at the edge of a flock (Ward and Webster 2016). Similar results showing such an edge effect in the Scaled Dove Columbina squammata were described by Dias (2006), where the peripheral individuals were more vigilant and spent less time foraging than the central ones. Also Black et al. (1992) described the relatively high costs of living at the edge of a flock in Barnacle Geese Branta leucopsis. However, the cited research also showed that the geese at the edge of the flock obtained more food because of the higher biomass of grass and lower level of competition there than in the safer centre of the flock.
The adult birds in the flocks we studied were equally vigilant in both habitats, whereas immatures were more vigilant on arable land than on grasslands. Jónsson and Afton (2009) reached similar conclusions for Snow Geese Chen caerulescens and Ross's Geese Chen rossii: adults spent more time being alert than immatures. This can be explained by the different experience of birds from the two age classes, as immature birds are less aware of potential dangers (Caro 2005). In our study, representatives of both age classes were similarly vigilant on arable land, as the fields at the edge of the Biebrza Basin are subject to frequent disturbance by humans and other predators, which is an important factor affecting the behaviour of geese (Polakowski et al. 2021). In contrast, most of the grasslands are situated in less accessible, usually protected areas, where conditions are safer. This could be why the less experienced immatures, usually in the care of their parents (Ely 1993;Kear 2005), reduced their vigilance there. Adults are similarly vigilant owing to their greater experience and expectation of predators everywhere, as well as parental care, whereas alertness levels in immature birds decrease in safer habitats.
In conclusion, several factors influence the safety of White-fronted Geese. In mixed flocks, White-fronted Geese benefit from the presence of Greylag Geese, probably due to both the territorial behaviour and the larger size of the latter. The age-related parental status and experience of Whitefronted Greese also influence their behaviour. In the context of exposure to predatory pressure, such factors as flock size, feeding habitat and an individual bird's position all affect alertness levels. We suggest that the lower level of vigilance of White-fronted Geese at spring stopover sites may result in a greater amount of time that they can spend foraging, which increases their chances of survival and breeding success during the oncoming spring.