Introduction

The breeding distribution of Northern Gannets (Morus bassanus) ranges from eastern Canada to North-West Europe, with Iceland and Norway holding the northernmost colonies and France and the Channel Islands (UK) the southernmost colonies. 75% of the breeding population is concentrated in the North-East Atlantic (Burnell et al. 2023). After cessation of intense human exploitation in the nineteenth century, the numbers of Northern Gannets increased strongly, especially in the second half of the twentieth century (Nelson 2002; Mitchell et al. 2004). Continuous increases (ca. 3% annually) have led to the foundation of new colonies, and population sizes in the southerly colonies have only recently stabilised or even declined, likely in response to the warming of ocean waters, as well as the outbreak of avian influenza (Guillemette et al. 2018; Le Bot et al. 2019; Lane et al. 2023). The island of Helgoland in the south-eastern North Sea is the site of a newly established colony of Northern Gannets. The first brood (2 pairs) was recorded in 1991, followed by an exponential population increase (2000: 93 pairs, 2010: 443 pairs, 2022: 1485 pairs; Dierschke et al. 2023; J. Dierschke unpubl. data). Recoveries of ringed birds revealed that the Helgoland gannets originated from the colonies on Les Etacs and Ortac in the Channel Islands (UK; 9 individuals) and from Great Saltee in South-East Ireland (3 individuals) (Bairlein et al. 2014).

Northern Gannets breeding in the North-East Atlantic spend the winter between northern Europe and North-West Africa, with birds from different colonies partly overlap** in their distributions (Fort et al. 2012; Garthe et al. 2016). Some individuals winter near their breeding sites, while others migrate to areas > 4000 km south of their breeding sites (Kubetzki et al. 2009; Fort et al. 2012). There is currently no information on the wintering distribution of Northern Gannets breeding on Helgoland. The western Baltic Sea has historically only recorded irregular, occasional observations of Northern Gannets (Berndt and Drenckhahn 1974; Klafs and Stübs 1987; Grell 1998; Sonntag et al. 2006), but there has been an apparent recent increase in the numbers wintering in the western Baltic Sea (based on local bird reports).

This study aimed to analyse the migratory pathways and winter destinations of adult Northern Gannets breeding on Helgoland, which were tracked successfully for the full non-breeding period in 2016–2017. In addition, we analysed data from the online databases dofbasen.dk (Denmark) and ornitho.de (Germany), and from the German Seabirds at Sea database, to explore the extent to which Northern Gannets used the western Baltic Sea, as well as the Kattegat and Skagerrak, during the winter months. We further compared the movements and wintering regions of Helgoland gannets with those of conspecifics studied at other colonies in the North-East Atlantic.

Materials and methods

Study concept

Fieldwork was conducted at the Northern Gannet colony on Helgoland (54.1859° N, 7.8746° E) in the southern North Sea, Germany. Ten incubating and five chick-rearing adults were equipped with geolocation devices (see below) on 12 May, 7–8 June, and 16–17 August 2016, respectively. Twelve birds with loggers were recaptured in summer 2017, and one bird each in summer 2018 and summer 2021, resulting in a recapture rate of 93%. Twelve devices delivered full datasets from summer 2016 to late spring 2017, covering the Northern Gannets’ entire migratory periods. A breast feather was sampled for molecular sexing in the laboratory (Tauros Diagnostics, Berlin, Germany).

Devices

We used geolocation data loggers (GeoLT) from Earth & Ocean Technologies (Kiel, Germany) to estimate the daily positions of the birds. The loggers were enclosed in a pressure-tight seawater-resistant casing (diameter, 14 mm; length, 38/45 mm [2 versions]); weight, 8.2 g, equivalent to approximately 0.3% of the bird’s body mass) and attached to a custom-built leg band. The main sensor of the device was a light sensor, which allowed the geographic position to be calculated based on the day and night lengths and time of local midday and midnight (Wilson et al. 1992; Kubetzki et al. 2009; Fifield et al. 2014). Light levels were measured every 30 s, allowing the device to operate for 1 year while providing two positional fixes per day.

The devices also recorded ambient temperature (i.e. air or sea surface temperature [SST], depending on its position) every 120 s throughout deployment, using a temperature sensor (range 0–32 °C, resolution 0.125 K, accuracy 0.2 K, 90% response time < 4 min).

Geolocator position estimation

Geographic positions were calculated from raw light curves using MultiTrace Geolocation (Jensen Software Systems, Laboe, Germany). The data were analysed largely as described by Kubetzki et al. (2009), Fifield et al. (2014), and Garthe et al. (2016).

Light levels were calibrated for sunrise and sunset from known locations (colony) and then applied to the whole dataset. Because our focus was on migratory movements rather than at-colony activity, the analysis was restricted to the period from colony departure to colony return. Colony attendance was derived from logger temperature profiles (for details see e.g. Wilson et al. 1995; Garthe et al. 2003). The error in the light-based latitude can be more than twice the corresponding longitude error estimated by geolocators (Phillips et al. 2004; Teo et al. 2004; Shaffer et al. 2005). Light-based latitude cannot be estimated during the solar equinoxes, when day length is the same at all latitudes (Hill 1994); however, light-based latitude estimates can be improved (or recovered during equinoxes) by reconciling geolocator-measured SST with remotely sensed satellite SST (Teo et al. 2004; Shaffer et al. 2005). The algorithm for this procedure was implemented in Matlab (Mathworks, Natick, MA, USA) and used to adjust the daily gannet latitude estimates accordingly (Fifield et al. 2014). The SST-corrected positions were filtered to remove positions requiring unreasonable speeds. When calculating the phenology and sea area use (Skagerrak, Kattegat, Baltic Sea), missing positions (e.g. due to failure of light-based geolocation during equinoxes and/or failure of the SST correction algorithm) were interpolated linearly between surrounding positions (Garthe et al. 2016). The proportion of missing positions to all positions varied between 8 and 31% per track (overall mean 15 ± 7%). For calculations of the maximum distance of each bird from the breeding colony, positions were smoothed to avoid unrepresentative positions using the weighted mean (1:2:1 ratio) of the previous (× 1), current (× 2), and subsequent (× 1) position’s coordinates.

Further data logger analysis

Latitude was chosen as the best indicator of migratory movements, given that all birds migrated largely south. A smoothed latitude value was determined for each bird, for each 5-day period from the date of colony departure until the date of colony return. December has previously been determined as the most suitable time to define Gannet ‘winter home range’ in the eastern Atlantic, because most larger-scale movements of Northern Gannets had ceased by then (Kubetzki et al. 2009; Fort et al. 2012). The mean winter region position was thus defined as the centroid (mean latitude and longitude) of all locations obtained during the period from 1 to 31 December.

Seabird counts on Helgoland

Attendance of Northern Gannets at the cliffs was recorded every morning from 1 November 2018 to 1 March 2019, and from 22 to 27 March 2019, and then again from 11 October 2019 to 2 April 2020. Only birds visible from public tracks were counted, but these included large parts of the gannet colony. Counts were ceased when the weather conditions did not permit accurate counts, mainly due to fog (3 days in winter 2018–2019 and 1 day in winter 2019–2020).

Seabirds at sea data

All Seabirds at Sea data were collected by ship-based surveys in German Baltic Sea waters from 2000 to 2022, according to the internationally standardized Seabirds at Sea method (Tasker et al. 1984; Garthe et al. 2002; Camphuysen et al. 2004). Birds were counted by human observers at high temporal resolution (1-min intervals). Gannets were recorded inside and outside the transect and numbers were then set into relation to the distance travelled by the observation platform (i.e. birds/km travelled). Analyses are based on all data in the joint Seabirds at Sea database of the Federation of German Avifaunists and the FTZ, Kiel University, as of December 2022, covering the main winter period from November to February.

Bird observation databases

Sightings of Northern Gannets were obtained at Danish and German online databases for birds. These sightings were not based on systematic counting efforts, but included a wide range of opportunistic observations by volunteers. Overall, these databases provide a solid base to determine the distribution and relative abundance of birds, and changes over time (Nyegaard et al. 2012; Wahl and König 2021).

Observations in Denmark were taken from the online database DOFbasen, run by Dansk Ornitologisk Forening (DOF)/BirdLife Denmark. The observations were first pooled for each winter period, defined as November to February, and the maximum number per cell in a 10 × 10 km grid site per winter was further processed to account for the increasing observation effort (Heldbjerg et al. 2024). Two types of analyses were carried out. For trend analysis, all maximum numbers per 10 × 10 km grid cell were allocated to regions (North Sea, Skagerrak, Kattegat, Baltic Sea) and then summed for each winter per region. An index was applied to facilitate comparisons of trends among the different regions, setting the first winter (2003–2004) for the whole data series for each region to 1 and thus adjusting all subsequent winters relative to 2003–2004. To visualise distributions and their changes over time, the data were grouped into blocks of four consecutive winters, and only the maximum number per 10 × 10 km grid cell per four-winter period was retained and plotted in the map.

Observations in Germany were taken from the online database ornitho.de, run by the Federation of German Avifaunists (DDA). Data were analysed for the winters 2009–2010 to 2021–2022, summarizing the months from November to February. Only the offshore areas and counties adjacent to the Baltic Sea were taken into account. To correct for increasing observation efforts over the years, the data are presented as sightings of gannets divided by all sightings in the database.

Results

Winter destinations

Northern Gannets from Helgoland moved east and south-west to a broad range of wintering sites, ranging from the western Baltic Sea to North-West Africa (Table 1, Fig. 1). Three of the 12 birds with complete data (25%) spent the winter in Africa, while 9 birds (75%) wintered in Europe, with the primary wintering sites located in the North Sea. The direct distance from the colony to the most distant point during migration ranged from 516 to 4,852 km (Table 1). Females tended to travel further than males (but note the small sample size: Table 1). Most individuals moved around a lot, but showed generally relatively straight migratory routes with no loo** tracks (see Fig. 2 for three individuals with different winter destinations).

Table 1 Summary of data for the 12 adult Northern Gannets from Helgoland with complete datasets
Full size table
Fig. 1
figure 1

Average December positions for all tracked Northern Gannets (for details see text). Yellow squares, females; black circles, males; blue asterisk, colony on Helgoland. Other colonies listed in Table 2 indicated by numbers: 1 = Skrúður, 2 = Bass Rock, 3 = Grassholm, 4 = Great Saltee, 5 = Rouzic

Full size image
Fig. 2
figure 2

Examples of individual tracks of three Northern Gannets tracked over the winter 2016–2017. A: individual 146 (♂), B: individual 116 (♀), C: individual 141 (♀). Table 1 provides more information on these birds. Different colours indicate positions of birds at different periods of winter: yellow, (August +) September; orange, October; light red, November; dark red, December; purple, January; black, February + March

Full size image

Phenology

Northern Gannets departed from the colony between 28 August and 16 September (median last night in the colony, 5 September). Migration trips and over-winter colony absence (the ‘non-breeding period’) were remarkably similar among individuals, ranging from 190 to 215 days (Table 1). Notably however, most birds remained in the wider North Sea, i.e. relatively close to the colony, for some weeks, with some birds remaining for the whole non-breeding period, despite not returning to land overnight. The timing and duration of migration and the wintering periods varied considerably among individuals (Fig. 3). The three birds that headed to North-West Africa had different temporal schedules, arriving at their most-southerly position between the end of October and mid-December and staying for 4–12 weeks (Fig. 3). The birds returned to the colony for the first time overnight from 15 March to 7 April (median, 29 March).

Fig. 3
figure 3

Timing of movements of individual Northern Gannets showing mean latitude for each bird during standardized 5-day periods. Dotted line indicates latitude of the Helgoland colony

Full size image

Although the birds tagged in this study did not stay on the island overnight until mid-March, count data from the breeding colony site on Helgoland in 2 other years showed that Northern Gannets started to attend the colony from the end of January, with high numbers from mid-February 2019 and from early March 2020 (Fig. 4).

Fig. 4
figure 4

Attendance of Northern Gannets at the breeding colony at Helgoland in the winters 2018–2019 and 2019–2020. Question marks indicate missing counts

Full size image

Use of Baltic Sea waters

All but one tagged Northern Gannet spent some time in the Baltic Sea or in the transitional waters between the North Sea and the Baltic Sea (i.e. Skagerrak and Kattegat) in the winter months (Fig. 5). One bird in particular spent 85% of the four winter months in these waters. Overall, five gannets (42%) visited the Baltic Sea (Fig. 5).

Fig. 5
figure 5

Proportions of time spent in Skagerrak, Kattegat, and western Baltic Sea by all 12 Northern Gannets. Data summarised for period November 2016–February 2017. Individuals sorted according to percentages. Bird_IDs are identical to Table 1. Map insert shows the three marine regions (SK Skagerrak, KT Kattegat, BS Baltic Sea) and the country names

Full size image

Records of Northern Gannets in the Danish ornithological online database dofbasen.dk have increased substantially over the last 18 winters, with the greatest increases in Baltic Sea waters, followed by the Kattegat and Skagerrak regions (Fig. 6). These increases were also evident when comparing multi-year maps (Fig. 7), especially in relation to the southern part of the Kattegat and the east coasts of Sjaelland and Falster.

Fig. 6
figure 6

Abundance of Northern Gannets along Danish coasts from November to February in the winters 2003–2004 to 2020–2021. Data based on sightings in DOFbasen. Totals per region summed per winter and transformed into an index with the first winter 2003–2004 set to 1.0. Note logarithmic scale

Full size image
Fig. 7
figure 7

Distribution of Northern Gannets along Danish coasts from November to February in four four-winter periods covering the winters 2005–2006 to 2020–2021. Maximum counts per 10 × 10 km grid cells per four-winter period shown. Data based on sightings in DOFbasen

Full size image

There was also an increase in Northern Gannet sightings in German Baltic Sea waters, but this occurred later than in the more northerly Danish waters. The online database ornitho.de showed a strong increase from the winter 2017–2018 onward (Fig. 8a). Systematic Seabirds at Sea counts from ships revealed regular sightings of Northern Gannets since the winter 2015–2016, with relatively high abundances (effort-corrected) in the last four winters to date (Fig. 8b).

Fig. 8
figure 8

a) Abundance of Northern Gannets along German Baltic Sea coast from November to February in winters 2011–2012 to 2023–2024. Data based on sightings in ornitho.de; number of observations summed per winter and divided by number of all bird observations in the same time period and area. b) Abundance of Northern Gannets in German Baltic Sea waters from November to February in winters 2000–2001 to 2021–2022 (shown by blue columns). Data based on Seabird at Sea surveys using ships. In addition to abundance, survey effort is also shown (red line). Metrics presented as number of individuals per km sailed.

Full size image

Discussion

Northern Gannets breeding at Helgoland fit into the North-East Atlantic ‘system’ of winter destinations, with wintering sites ranging from local waters near Helgoland to North-West Africa (Fort et al. 2012). The proportion of Helgoland gannets wintering in the North Sea was higher than for any other colony with published data so far, while the proportion of birds wintering in Africa was the lowest (Table 2; but note the relatively small sample sizes). Females tended to travel further than males, as also found for Northern Gannets breeding in Scotland and Wales (Deakin et al. 2019). Birds left the breeding colony about 1 month earlier than those from the Bass Rock (UK; Garthe et al. 2012) and 2 weeks earlier than those from Skrúður (Iceland; Garthe et al. 2016). Interestingly, Helgoland gannets returned roughly 2 weeks later than those from the Bass Rock and also later than those from Iceland; however, counts at the Helgoland breeding cliff suggest that birds started to return to the vicinity of the island much earlier, with many being within reach of the colony from February onward. More detailed information on the pre-breeding spatial patterns of the birds will require tracking by GPS or other similarly precise methods.

Table 2 Proportions of Northern Gannets wintering in the Baltic Sea, North Sea, North-West Africa, and other regions
Full size table

In contrast to previous studies from other colonies (Table 2), all the Helgoland birds visited the waters east of the North Sea (Skagerrak, Kattegat, western Baltic Sea) during their migration, and almost all of them during the core winter period (November to February). This is in accord with the distribution patterns derived from the Danish database dofbasen.dk, which also shows clear increases in winter sightings of Northern Gannets in the Kattegat. Danish and German online databases, as well as long-term Seabirds at Sea data, indicate the establishment of a regular winter occurrence of Northern Gannets in the western Baltic Sea. This is corroborated by strongly increasing numbers during migration counts at Falsterbo in South-West Sweden (www.falsterbofagelstation.se) in autumn. Furthermore, we repeatedly recorded a few long-lasting foraging trips reaching the south-western Skagerrak region during chick rearing on Helgoland (Garthe et al. 2017; Peschko et al. 2021). The clear spatial link between the Northern Gannets tagged at Helgoland with the Kattegat and the western Baltic Sea, as well as the temporal overlap of the increases in the western Baltic Sea with the population increase on Helgoland (Dierschke et al. 2023), suggest a strong connection.

Regarding the potential prey source for birds in these new wintering areas, Northern Gannets generally feed almost exclusively on pelagic fish, taking almost all sizes from small sandeels to large Atlantic Mackerel (Scomber scombrus; Lewis et al. 2003; Hamer et al. 2007). Skov et al. (2000) found significant correlations between several piscivorous birds and Atlantic herring (Clupea harengus) abundance at coarse spatial scales (> 20 km) in the Skagerrak–Kattegat. Atlantic herring is a well-known staple food species in the diet of Northern Gannets (Garthe et al. 2014; Barrett 2016) and was by far the most numerous fish species in the sound between Kattegat and the western Baltic Sea (Nielsen et al. 2001). However, the population of western Baltic spring-spawning herring has declined hugely in recent years, while European sprat (Sprattus sprattus) numbers have increased (ICES 2023). Nielsen et al. (2023) reported regular aggregations of razorbills in the western Kattegat in October/November and related their occurrence to the availability of sprats. They also highlighted the fact that the razorbill aggregations attracted Northern Gannets and other seabirds, implying a major prey resource available to multiple species. It is therefore plausible that both Atlantic herring and European sprat could be key food items for Northern Gannets wintering in or at least visiting the Skagerrak, Kattegat, and western Baltic Sea areas.

These findings demonstrate that Northern Gannets explore the western part of the Baltic Sea, as well as the Kattegat and Skagerrak, increasingly intensively, leading to speculations about possible future breeding sites. There was one incident of a Northern Gannet defending a nest in the harbour on the Danish island of Christiansø (Lyngs 2015), while in Germany, gannets seem to be attracted to the area adjacent to the cliffs of Rügen (K. Borkenhagen pers. obs.). Overall, the Baltic Sea appears to be a suitable marine region, at least for migrating and wintering Northern Gannets. The recent increase in sightings is in line with the establishment of exponential increase in breeding Northern Gannets at the nearest breeding colony on Helgoland. It is also possible that Northern Gannets from other colonies, most likely Norway and Scotland, use this area more often than in the past (Fort et al. 2012). Another reason for the observed increases of Northern Gannets in the Kattegat and western Baltic Sea may be related to climate change. Traditional wintering areas may become less suitable and/or the western Baltic Sea more suitable for gannets ranging from multiple colonies. For instance, the Baltic Sea may become more suitable for fish species such as European pilchard (Sardina pilchardus), European sprat and European anchovy (Engraulis encrasicolus; Schickele et al. 2021), representing very suitable prey items for Northern Gannets.