Dichtheid, reproductie en sekse-specifieke verschillen in ruimtegebruik en voedselkeus van Wespendieven Pernis apivorus in ZO-Nederland en aangrenzend België in 2013-15

Diermen J. van, van Manen W. & van Rijn S. 2024. Density, breeding and sex-specific range, habitat use and diet of Honey Buzzards Pernis apivorus in SE-Netherlands and adjacent Belgium in 2013-15. De Takkeling 32: 189-259. In 2013-15 we studied Honey Buzzards in Het Groene Woud (51.54 N, 5.34 E) in the province of Noord-Brabant (85 km2) and region Kempen-Broek (51.22 N, 5.67 E) in the border zone of the provinces of Noord-Brabant and Dutch and Belgian Limburg (115 km2). Woodland covers 23-30% of the study areas and is relatively often on richer (loamy) and more humid soils as compared to other regions in The Netherlands (mostly sandy soils). Territories were mapped mainly by observing from tree tops (Douglas Fir and Norway Spruce) and from the ground in open spaces in July and August, with on average three hours per observation session. With notes on behaviour, plumage (sex) and simultaneously observed birds or pairs we outlined territories. Birds with offspring showed typical straight flights above the canopy between nest and foraging site, enabling us to find nests via triangulation of food transportations. In July and August, non-breeding pairs consistently showed different behaviour than nesting pairs, especially spending time ‘hanging around’ above the canopy, soaring together or wing-clapping at great heights. We trapped 22 birds (using a live Eagle Owl as decoy, and mistnet). Seventeen (11 males, 6 females) were equipped with GPS-trackers (UVA-Bits and Ecotone Skua), of which ten collected data in more than one season. The GPS fixes were homogenised to a 10 min interval and interpreted using an algorithm. Birds at <200 m from the nest were regarded as involved in parental care at the nest. Fixes >30 min after sunset and <30 min before sunrise were regarded as roosting birds. For other behaviour we used fixes with an interval between next and previous fixes of less than 35 minutes. Birds were regarded flying when the distance between previous and next fix was >200 m or at an altitude >150 m, and foraging was identified as the distance between previous and next fix was <200m. Assigning landscape features to GPS fixes was done from 1:10.000 digital maps. Preferences were calculated by comparing fixes with 150 m grid characteristics in the four 90o-directions. Woodland structure preference was recorded by describing and measuring 27 paired habitat variables in a 10 m radius at 72 locations showing or lacking concentrations of GPS fixes that indicated foraging. We used three approaches to determine diet composition. (1) Close range nest surveys with camera in activity sensor mode. This yielded unbiased data, but we only managed to get complete hatch to fledge coverage in one nest. Other sessions were either ceased by chick predation and battery limits or started halfway the chick period following a later nest finding date. By scrolling through pictures, most prey could be identified and the majority assigned to parent sex. (2) Prey remains were identified and quantified during nest checks. At each nest visit to determine brood size, chick condition and retarded growth, we kept record of prey remains, almost exclusively combs. (3) Interpretation of GPS tracks helped in finding soil-nesting social wasps (n=498) and bumblebees (n=5). However, this method likely missed hunting on vertebrates (too short-lived an action to be traced via GPS-data) and predation of smaller Dolichovespula nests hanging in bushes as well as Polistes nests in herbs (ditto). Wasp abundance was measured using juice + beer as a bait in 0.33 litre bottle traps, provided in pairs in various habitats (3 types of woodland, 2 types of rural landscape, open natural habitat, adding heath and heathland edge in 2015). Traps were set for 24 h each ten days from the end of June, but not every season yielded complete trapping series in both study areas. We consider the results as biologically meaningful for the three main prey species, i.e. Common Wasp Vespula vulgaris, German Wasp V. germanica and European Hornet Vespa crabro. Frogs were counted along all-purpose walking lines within and outside woodland from April to October. The focus was on Rana temporaria in terrestrial habitat away from water bodies used for reproduction. Frogs from the Pelophylax genus preferred water edges but were hardly preyed upon and therefore omitted from the survey. Birds (individuals) were counted using 5 min point counts twice in May-June, before noon and skipping the hour around sunrise. Numbers of bird species known as potential Honey Buzzard prey were transformed into biomass. Comparing nine NW-European studies, the Honey Buzzard density of 0.19 territories per 100 ha of landscape and 0.83/100 ha woodland in Het Groene Woud and the slightly lower density of 0.17/100 ha of landscape and 0.58/100 woodland in Kempen-Broek were relatively high. Nests (n=56) were found in a wide variety of woodland types and the chosen nest tree species did not differ from what was available. On average 24% of pairs bred successfully, raising 1.73 young per nest, which amounts to 0.41 young per territorial pair. This means that around half the pairs that started breeding did not succeed because of loss of progeny. Reproduction rate nevertheless is within the range of Dutch studies during the last decades. Breeding attempts rarely failed during the egg stage, yet losses in the chick-stage were substantial and mainly due to Northern Goshawks of different age and status, preying on chicks of any age. Variation in breeding success between years was considerable and both the number of young in successful nests and the proportion of successful breeding attempts were correlated with wasp abundance as recorded via the wasp trapping scheme. Poor wasp abundance, on the other hand, negatively affected Honey Buzzard’s nest attendance through intensified hunting effort over larger areas and longer absences at the nest, increasing predation risk. Birds equipped with GPS trackers (n=17) initially foraged in the nest vicinity in the early part of the breeding cycle, females ranging further from the nest in the egg stage than males (sexes share breeding duties and forage for themselves). Both sexes strayed farther from the nest as the chicks matured, but in the early chick stage the birds primarily foraged in woodland. As the breeding season progressed, a shift to rural areas became apparent. This tendency was strongest in females, heading for plots far outside their partners’ range. Feeding pattern and prey type were registered using camera traps at twelve nests during 193 effective days. Diet was studied with three methods, of which trap cameras at the nest probably was the most reliable (71.2% wasps combs, 3.3% vertebrates, 25.2% amphibians; n=1470), compared to prey recorded during nest visits of the observers (95.3% wasp combs, 0.6% vertebrates, 3.7% amphibians; n=486) and extracted from GPS-data (99% wasps, 1% bumblebees; n=503). The foraging method of Honey Buzzards did not allow to register the capture of vertebrates in GPS-data. Males started foraging in the early morning hours, mainly on frogs as these were most active in the morning and became more obtrusive in the course of the day. Females typically would spend the night on the nest and started the day adding fresh greeneries to the nest. Females started foraging after the first male prey delivery, which accounted for the female’s later daily onset in chick feeding. Females almost exclusively supplied their chicks with combs containing wasp brood (>90% of items). Apart from their different timing, the female’s focus on wasp brood can be explained by their ‘free ranging’ outside woodland where the preferred prey (wasp brood) was more abundant in the years of study (2013-15). Wasp trapping showed that social wasp activity in woodland dropped by mid-July but at the same time increased outside woodland in farmland. Nesting males on the other hand demonstrated a stronger tendency to prey on vertebrates (40% of items), probably because they had to cope with a limited supply of wasp nests within their more restricted feeding range in woodland. The GPS-data clearly showed sex-specific feeding ranges, reducing interference within the partnership. Describing and measuring woodland characteristics at foraging and reference locations showed that Honey Buzzards preferred well-structured and species-rich woodland types, predominantly in moist conditions. This also turned out to be the type of woodland where we recorded the highest density of breeding birds within the food spectre of Honey Buzzards. Wasp nests are still in the foundation state when Honey Buzzards arrive in May. Queen activity in spring is no guarantee to harvestable larvae in July (usurpation, nest failure in adverse weather). The insect traps in humid study areas showed relatively high insect (including wasp) abundances compared to dry areas nearby, probably attributing to the higher density of Honey Buzzards in more humid woodland. In the course of July and August small-scale farmland showed a higher abundance of German and Common Wasps (species with longer colony cycles than Dolichovespula species). Accordingly, male and especially female Honey Buzzards increasingly shifted to foraging in small-scale farmland. Breeding success was strongly related to wasp supply. The availability of wasp nests in farmland seemed crucial in the latter part of the chick-rearing stage (August). The shift to foraging in farmland as the season progresses hints at depletion of social wasps in woodland (at least in our study). This brings new inferences, viz. between Honey Buzzards, social wasps and the massive use of insecticides and herbicides in modern agri-business (two instances of ‘pinching- off’ were recorded in Dutch Honey Buzzard chicks, possibly related to the use of tembotrione in maize, a herbicide known to cause ochnorosis and probably linked with ‘pinching-off’).