To provide a comprehensive view of a species ecology, it is important to investigate all periods of its life cycle. While many studies investigate habitat selection during the breeding season, fewer studies investigate the non-breeding period [1]. In this study we showed that small mammal distribution in different habitats varies over the annual cycle, which affects barn owl prey availability and consequently also barn owl habitat preference. Barn owl breeding and non-breeding home-range size were similar, but there was a small shift in the location of home-ranges which was more pronounced in females than males. The changes in prey availability led to a mainly grassland-oriented habitat selection during the non-breeding period. Further, our results showed the importance of biodiversity promotion areas and border structures within the intensively managed agricultural landscape.
Home range size, location, and composition
With an average size of 6.8 km2 barn owl non-breeding home range size was similar to its breeding home range size and comparable to breeding home range sizes found in previous studies in barn owls [37, 38, 50, 53]. Overall, females had larger home ranges than males in both seasons. During breeding, home ranges of both sexes are closely bound to the breeding site, while the distance enlarged for females but not for males during non-breeding period. Such a pattern could emerge through a difference in nest box fidelity of the sexes during the non-breeding period. Previous studies have shown that in case of divorce (change of partner between years) males stay at the initial nest box while almost all females change the nest box [54]. Our data showed a similar pattern, as females are more prone to change the nest box between breeding seasons. Additionally, own and unpublished findings showed that males visit nest boxes more frequently and regularly than females outside the reproductive period. These findings together indicate that males exhibit higher nest box fidelity resulting in regular visits of their nest box during the non-breeding period. This would indicate that males are more regionally bound and thus might be less prone and less flexible in exploitation of surrounding areas. Overlap of breeding and non-breeding range was higher for males than for females, indicating a shift in spatial placement of the home ranges during the non-breeding period in females. Habitat composition of non-breeding and breeding home ranges[37] is similar, with crop rotation covering the biggest proportion of area.
Habitat selection
Our data suggested, the following, decreasing, order in habitat preference during non-breeding season: urban area > intensive grassland > extensive grassland > crop rotation > forest. The reasons for such a strong selection towards urban area cannot be entirely answered in the frame of this study, but we formulate two hypotheses. First, owls roosting in buildings can save energy compared to roosting in trees especially in cold periods [55]. During the non-breeding period the pressure to save energy might be especially high, resulting in prolonged roosting behaviour during the night and thus select locations dominated by urban area. Additionally, owls only need few prey items during the non-breeding period, as they only need to feed themselves. As catching fewer prey might be less time-consuming, there would be more time available to roost. Second, microstructures around houses, such as gardens, hedges, or small orchards, might be suitable habitats of sufficient size for hunting during the non-breeding period. Being close to buildings and thus less exposed to weather conditions compared to hunting in open area might further favour this behaviour during periods of rough weather. However, these explanations remain to be tested.
The selection of intensive grassland is not very surprising, as previous studies showed that different types of grassland (pastures or meadows, intensively or extensively managed) are preferred by birds of prey during breeding [9, 37, 56, 57] and our analysis showed that this preference seems to persist into the non-breeding period. Even though we see a positive selection of extensive grassland compared to crop rotation, the effect is very weak, which could be partly explained by the scarcity of these structures within the landscape (rareness effect). Further, the presence of biodiversity structures (hedges, trees, wildflower strips, rotational fallows) seemed to positively influence selection of grassland dominated areas. This effect was pronounced in extensive grassland dominated areas and weaker in intensive grassland. However, the presence of biodiversity structures does not seem to influence the selection of the other habitat types. Again, there are different explanations. While grasslands may provide suitable habitat for hunting, the neighbouring biodiversity structures could provide perching opportunities, allowing owls to use a sit-and-wait hunting strategy instead of hunting on the wing [58]. Own, but yet unpublished data suggest that hunting success in barn owls is higher if owls hunt from perches compared to hunting on the wings, possibly leading to a strong selection towards perch hunting in periods with high energetic needs such as the colder non-breeding period. Another possibility is that nearby biodiversity structures could cause a spill-over of small mammals into adjacent grasslands. Previous studies showed that semi natural habitats harbour high density of small mammals and can indeed act as a refuge and even as a source for reuse or recolonisation of intensively managed area in spring and summer [14, 15, 17, 18, 23–25, 27, 59].
Surface dedicated to crop rotation, which represents the most abundant habitat category within the home ranges of our birds, showed a lower selection compared to intensive grasslands. Crop rotations, especially cereal fields, are used by hunting owls during breeding [9, 37], however crop rotations fields look totally different in terms of vegetation coverage and food availability during the non-breeding and the breeding period. In our analysis we did not find any preference for crop-rotations (cereal, colza, catch crop, maize, sugar-beet) over grassland during non-breeding period. Looking at our data we see that small mammal index in crop-rotations drop drastically in late summer. This drop is most probable induced by harvesting which has been shown to affect small mammal populations either directly due to mortality caused by the harvesters, or due to a change in habitat suitability (reduced cover and food availability) for small mammals [14, 15, 18, 60]. While birds of prey might profit from the increased availability of prey in fields dedicated to crop rotation during breeding, this effect may only persist till shortly after harvest when this habitat instantly become unattractive for small mammals and thus also for their predators.
Habitat selection, biodiversity structures and prey
We showed that biodiversity structures (such as hedges, single trees, wildflower strips and rotational fallows) boost the preference for grassland dominated areas during the non-breeding period. This could be either due to an elevated abundance of perching opportunities when biodiversity structures are present, or due to high prey availability within these structures. We thus investigated whether habitat selection varies depending on prey index and the presence or absence of biodiversity structures. We found a preference for areas with a high vole index, which was even stronger when biodiversity structures were present. For overall small mammal index the presence of biodiversity structures has a small opposite effect. Knowing that the two methods used in the small mammal monitoring represent a different spectrum of prey species, there are different possible explanations for the observed results.
The vole index represents almost exclusively the availability (presence, abundance and activity) of Microtus and Arvicola species (∼70% of barn owl prey, [33]). Selection for areas with a high vole availability may be due to the altered accessibility of this prey type during the non-breeding period. While high, dense, and fast re-growing vegetation reduce accessibility of prey during breeding [9, 35, 56] lower and less dense vegetation during the non-breeding period may increase accessibility of this prey. Indeed, the increase in vole traces observed from the breeding to the non-breeding period doesn’t only represent an increase in vole abundance but also partially reflects increased visibility of the traces for the observer during non-breeding period. While one could argue that the vole index is biased regarding true abundance, the increased visibility for the observer also indicates an increased accessibility of the prey for predators. Even though the prey might be more accessible, hunting during the non-breeding period could be very time and energy consuming. Common voles, for example, are living a rather stealthy live, partly underground and covering small areas (home range: 200 m2, [15]). In addition, low temperatures and decreased vegetation cover further decrease vole activity [60, 61], which may result in very time-consuming hunting of this prey species. Perch hunting in areas with high vole availability, could be energetically beneficial and might explain the selection for areas with high vole availability and presence of biodiversity structures, which provide perching opportunities. These circumstances could result in a seasonal shift of hunting behaviour resulting in elevated rates of perch hunting during non-breeding period. That seasons can influence hunting behaviour has been shown for kestrels which switched from mainly hunting on the wing during summer, to perch hunting during winter [62, 63].
While the presence of biodiversity structures seems beneficial at low small mammal indices, the positive effect of the structures vanishes towards higher indices. Sampling mainly above ground movement of several species, this index tends to also reflect the availability of rather mobile prey (wood mice home range: 4’900 up to 14’400 m2, [18]). The overall lower impact on selection of small mammal index compared to vole index could result from the seasonal shift of prey availability. While indices tend to be high among different habitat types and especially in crop-rotation during the breeding period, this pattern changes during the non-breeding period. In agreement with previous studies on different small mammal species [14, 15, 17, 18] we observe a decrease of presence, abundance and activity within crop plantations after harvest. Hunting mobile rather than stealthy prey might be advantageous during the breeding period when availability is high, but after harvest the circumstances might change drastically. Prey availability in crop-rotation drops and activity only remains high in habitat which is unsuitable for hunting (forest).
Looking at the different foraging habitat categories separately, we see that focal areas with high prey indices within grasslands and border structures get selected. As we already know that owls prefer intensive grassland and areas with overall high vole indices, the selection of high overall prey indices within intensive and extensive grassland is thus consistent with the previous findings. This further strengthens our conclusion that especially grasslands are providing crucial food sources during the non-breeding period. The preference for high indices in border structures is quite interesting as these structures usually only cover small areas within focal areas. That border structures are attractive habitat for voles [26] and consequently attract birds of prey has already been shown for other species [35, 58]. The preference for border structures could be a combination of two effects, firstly elevated availability of prey within border structures, secondly, they are often found next to roads equipped with road sings and poles. These poles and sign could act as perching opportunities and could have a similar effect on selection as we hypothesized for presence of biodiversity structures, allowing to save energy by perch hunting [58, 62]. The negative effect of high prey availability in biodiversity structures might seem contradictory at first. However not only prey abundance but also its accessibility is important for the selection of certain habitat structures [9, 35, 56]. High prey availability per se thus does not necessarily favour selection as those prey items are not accessible and are only reachable at the edge of those structures or when entering the adjacent grassland [35, 56].