Introduction
Migration is a well-known behaviour in birds (Newton Reference Newton2008) and may involve complex processes and mechanisms. For instance, populations may show partial migration, with some individuals migrating to winter quarters while others remaining in or near the breeding colony year-round (Dingle Reference Dingle1996). In some particular cases, members within a given population may follow distinct migration strategies, which is known as differential migration, occurring between sexes, age groups, or morphs within a population (Ketterson and Nolan Reference Ketterson, Nolan and Johnston1983). These different migration behaviours are usually considered to be a strategic response to spatially seasonal variability in the environment, allowing some individuals in the population to avoid unfavorable conditions or to be in locations with higher food availability compared to their breeding grounds (Cresswell et al. Reference Cresswell, Satterthwaite, Sword, Milner-Gulland, Fryxell and Sinclair2011).
Seabirds are in general migratory or dispersive (Burger and Gochfeld Reference Burger, Gochfeld, del Hoyo, Elliot and Sargatal1996). Gulls (Laridae), in particular, show different patterns of migration, with trips ranging from hundreds of kilometres to long distance journeys of thousands of kilometres (Hatch et al. Reference Hatch, Gill and Mulcahy2011, Klaassen et al. Reference Klaassen, Ens, Shamoun-Baranes, Exo and Bairleind2012, Bustnes et al. Reference Bustnes, Moe, Helberg and Phillips2013). The understanding of bird movements can be assessed through ringing studies and in fact the re-sighting of permanent field-readable markings have been used traditionally in migratory studies. This low-cost methodology has the advantage of allowing the study of large numbers of individuals (Bairlein Reference Bairlein2003), and the use of numbered plastic colour rings may even allow the identification of individuals without the need of recapture. However, over the past three decades, the use of electronic devices to monitor the medium- and large-scale movements of birds has greatly enhanced our understanding of migration patterns, moving from knowing few refueling and wintering spots in the migration route, to currently visualising complex migration behavior and use of foraging and resting areas along the way. Moreover, the size of devices has decreased significantly over time, thus allowing the deployment of devices in smaller species and for longer periods of time (Burger and Shaffer Reference Burger and Shaffer2008).
The Olrog’s Gull Larus atlanticus is an endemic species of the south-western Atlantic coast, only breeding in Argentina (Yorio et al. Reference Yorio, Petracci and Borboroglu2013). The species has been listed on the IUCN Red List as ‘Near Threatened’ (BirdLife International 2019) and included in Appendix I of the Convention on the Conservation of Migratory Species of Wild Animals (CMS 2018). At a local scale the species is listed as ‘Vulnerable’ in Argentina and ‘Endangered’ in Uruguay (Azpiroz and Caballero-Sadi Reference Azpiroz, Caballero-Sadi, Azpiroz, Jiménez and Alfaro2017, MAyDS and AA 2017). Main threats include pollution, habitat degradation and human disturbance mainly linked to interactions with sport fishing activities. Its breeding range is mostly restricted to southern Buenos Aires province, with the number of nesting sites varying between seven and 12 depending on the year, and a global breeding population of less than 8,000 pairs, including breeding pairs in Chubut province (Yorio et al. Reference Yorio, Petracci and Borboroglu2013). During the non-breeding season, individuals of different age classes migrate north, being sighted in coastal areas of northern Buenos Aires province (Favero et al. Reference Favero, Copello, García, Mariano-Jelicich, Ravasi, Seco Pon and Azara2016), Uruguay (Escalante Reference Escalante1970) and southern Brazil (Pacheco et al. Reference Pacheco, Olinto Branco and de Queiroz Piacentini2009, Gava Just et al. Reference Gava Just, Rodrigues Rosoni, Romagna and Zocche2018). Little is known about its wintering movements. In Mar Chiquita Lagoon (Buenos Aires province), a regularly reported wintering area for the species in Argentina (Favero et al. Reference Favero, Bachmann, Copello, Mariano-Jelicich, Silva, Ghys, Khatchikian, Mauco and Iribarne2001, Isacch et al. Reference Isacch, Bó, Vega, Favero, Baladrón, Pretelli, Stellatelli, Cardoni, Copello, Block, Cavalli, Comparatore, Mariano-Jelicich, Biondi, García and Seco Pon2016), individuals tracked with GPS and VHF radio-transmitters were reported to use a restricted area in the vicinity of the mouth of the lagoon (Berón et al. Reference Berón, Favero and Gómez Laich2007, Ravasi et al. Reference Ravasi, Seco Pon, Paz, Favero and Copello2019). However, no information on wider movements connecting breeding and non-breeding areas is yet available for this endemic and threatened gull species. In this context our study was designed to analyse the migration of Olrog’s Gulls considering differences in sex and age, by following individuals from the population breeding at the Bahía San Blas Protected Area (southern Buenos Aires province, Argentina) using with tracking loggers, and complementing this information with sightings of breeding adults and fledglings ringed in this breeding area.
Methods
Bahía San Blas is located in the south-western sector of the marine protected area in southern Buenos Aires Province, Argentina (Figure 1). It includes two Olrog’s Gull colonies, Islote Arroyo Jabalí Oeste (40°33°S, 62°16°W) and Banco Nordeste (40°32’S, 62°10’W). The most-up-to-date estimate accounted 359 Olrog’s Gull breeding pairs nesting in the study area (Suárez et al. Reference Suárez, Marinao, Kasinsky and Yorio2014). The coastal sector is characterised by extensive mudflats and marshes (Spartina spp. and Salicornia ambigua) with crab beds (Zalba et al. Reference Zalba, Nebbia and Fiori2008). Egg-laying starts in late September and chicks fledge in early December (Suárez et al. Reference Suárez, Retana and Yorio2012).
Figure 1. Core and range area (kernel 50, 95%, respectively) of GLS-tracked adult Olrog’s Gulls during the non-breeding period (January-August 2015) and sightings of banded individuals during 2009–2018, including those reported in Seco Pon and Favero (Reference Seco Pon and Favero2011) and Paz et al. (Reference Paz, Seco Pon, Favero, Blanco and Copello2018). URU= Uruguay, PRA= Punta Rasa, MCH= Mar Chiquita, MDP= Mar del Plata, NEC= Necochea, SBL= San Blas and SAO= San Antonio Oeste. J= juvenile, SA= subadult, A= adult. Inset: Distribution of distances from the study colony to sighting locations.
A total of 32 geolocators (GLS; 10 C65 Migrate Technology and 22 MK4093 Biotrack, 1.0 and 1.52 g, respectively) were deployed on adults (18 males and 14 females; in 75% of the cases both members of the breeding pair were tracked) breeding at Islote Arroyo Jabalí Oeste on late October 2014 during incubation and early chick-rearing period (Table S1 in the online Supplementary Material). Each device was attached to a plastic ring specifically designed for this purpose with a cable tie and glue making the total weight of ~3 g including the GLS, representing less than 0.3% of the average body mass of individuals (1,047 g ± 108; n = 32). Birds were captured at active nests using a foot snare. Devices were recovered in early October 2015, with all devices (except one that was operational only until June) gathering data during the entire year (Table S1). The sex of individuals was determined by genetic analysis (Quintana et al. Reference Quintana, López and Somoza2008).
For GLS data analysis IntiProc Geolocation Processing Software (©Migrate Technology) and Geolight package 2.0.0 in R version 3.4.4 (Lisovski and Hahn Reference Lisovski and Hahn2013) were used to estimate the locations. Then ‘adehabitatHR’ package 0.4.15 in R and ArcMap ESRI® software were used for further spatial data analysis. The locations during the non-breeding period (January to August) were selected. Locations acquired during the equinox period in March were not considered in the analysis given the low accuracy of location estimates during this period (Biotrack 2013). Calibration of each device was performed for about a week in a fix position pre- and post-deployment. Sun elevation angle was set in -6.05° and -4.41°for Migrate and Biotrack GLSs devices, respectively. These values were estimated using the average for both calibration periods. Lux threshold was set to 3. To estimate the errors in geolocations, the deviation in latitude, longitude and distance to the fix point were calculated (n = 221, mean ± SD = 179.5 ± 112.3 km, with overall SD of 1.1° and 0.3° of latitude and longitude, respectively). Locations were filtered using a speed filter with a maximum velocity of 40 km/h (Klaassen et al. Reference Klaassen, Ens, Shamoun-Baranes, Exo and Bairleind2012) and a filter over the twilight events using the ‘loessFilter’ function. We generated a kernel analysis on a monthly basis for the whole non-breeding period (January to August, h =180 km), using valid locations to identify core and range areas (kernel 50, 95%, respectively) for the whole dataset and also pooled by sex. The smoothing parameter h approximates the mean accuracy estimated for the tag using the calibration period. Kernel analysis is used commonly to identify important bird areas (Hyrenbach et al. Reference Hyrenbach, Keiper, Allen, Ainley and Anderson2006, Copello et al. Reference Copello, Seco Pon and Favero2013). Areas where GLS fixes are dense are assumed to be important for those individuals being tracked as it is where they spend the majority of their time, while regions with fewer GLS fixes are considered transit zones, i.e. areas where birds could be located, but may not be foraging. The 95% density estimate was considered as the range area and the 50% kernel was defined as the core foraging region (Wood et al. Reference Wood, Naef-Daenzer, Prince and Croxall2000). The overlap index HR was used in order to evaluate spatial differences between sexes (Fieberg and Kochanny Reference Fieberg and Kochanny2005) using the ‘kerneloverlaphr’ function in the package ‘adehabitatHR’ in R. This index ranges from 0 if the two home ranges do not overlap at all, to 1 if both kernel areas show total overlap. A z test was used to test differences in the size of the core areas between males and females (Zar Reference Zar1996).
Additionally, a database of sightings of colour-ringed Olrog´s Gulls was analysed. This database included sightings of individuals ringed as fledglings (n = 398) or breeders (n = 176) at Islote Arroyo Jabalí Oeste and Banco Nordeste colonies between the 2007 and 2010 breeding seasons. Individuals were marked with coloured plastic rings with a three- number code (yellow with black numbers and red with white numbers) and a metallic ring. A total of 41 ringed Olrog’s Gulls were sighted on land within a time frame spanning 10 years (2009–2018). Recoveries of ringed birds were reported to the NGO Aves Argentinas or directly to the authors (NS, PY or JPGB) and additional sightings were obtained by searching the web. Twelve ringed individuals were directly sighted by the authors in the San Blas breeding grounds during the austral winter (May and June). Different age classes were identified by plumage characteristics, grouping the individuals as adult, subadult and juvenile (Harrison Reference Harrison1983).
Results
Twenty-one geolocators with available data (70% of the total deployed devices; Table S1) were recovered (eight females and 13 males). A total of 7,466 locations from January to August 2015 (non-breeding period) were obtained. After breeding, tracked gulls remained in a single core area (kernel 50%) between 38 and 44°S and 60–66°W, including the study colony and also other breeding colonies located up to 300 km to the north, within the Bahía Blanca estuary (Figure 1). The maximum distance of the core area was 400 km, covering an area of c.250,000 km2. The range area (kernel 95%) extended between 30 and 50°S and 55 and 69°W including the study colony and northern coastal areas such as Mar Chiquita Lagoon and Necochea and southern coastal areas up to their southern breeding colonies (Figure 1). The maximum range was c.1,000 km.
Core areas used by males and females during the non-breeding period were similar spatially and in size (HR 0.9 ± 0.1 for males vs. females and 0.8 ± 0.1 for females vs. males; 302,068.4 ± 39,919.2 and 320,689.4 ± 42,809.2 km2 for males and females respectively; z=-0.9, P > 0.05; Figure 2). Core areas varied slightly between months throughout the non-breeding period, with tracked individuals remaining south of the breeding colony during late summer, and moving further away to northern areas including the Bahia Blanca estuary during autumn-winter (Figure 3).
Figure 2. Core area (kernel 50%) of GLS-tracked adult Olrog’s Gull females (n = 8) and males (n = 13) during the non-breeding period (January–August 2015)
Figure 3. Monthly core areas (kernel 50%) of the 21 GLS-tracked Olrog’s Gulls during the non-breeding period. Latitudinal variability observed in April and August is an artifact related to the calendar proximity of March and September equinoxes respectively (see methods).
A total of 41 ringed Olrog’s Gulls were sighted over a 10-year span, of which 19 were adults and 22 immature individuals (21 subadults and one juvenile). Repeated sightings of nine ringed individuals obtained at the same location within the same season were not included in the analysis. At least one ringed gull was sighted every year during 2009–2018, except 2014 and 2017. Sightings of ringed individuals occurred between March and September but predominantly in July. All sightings occurred in coastal areas, specifically San Antonio Oeste, Necochea, Mar del Plata, Mar Chiquita, Punta Rasa and Bahía San Blas in Argentina, and in coastal areas in the vicinity of Arroyo Maldonado in Uruguay (Figure 1). Most of these sightings occurred north of the breeding colony except one adult gull sighted in San Antonio Oeste. In addition, 12 adult individuals were sighted during the winter visit to San Blas breeding grounds on May and June. The average coastline distance from the colony to the sighting location was 709 ± 483 km (n = 41) with most of the ringed birds sighted at less than 200 km and between 800 and 1,200 km from the colony (Figure 1). The average distance was larger in immature than adult individuals (1,029 vs. 338 km respectively: z = -6.2 P < 0.05).
Discussion
Our results combining information from geolocators and ringed birds suggest the occurrence of partial migration behavior in Olrog’s Gulls from Bahía San Blas. Using information retrieved from GLS-tracked individuals it was possible to determine that adults breeding at Islote Arroyo Jabalí Oeste (Bahía San Blas), did not show the expected extensive post-breeding dispersion. In fact, none of the core areas (kernel 50%) of tracked individuals comprised wintering sites previously reported and observed in sightings of ringed adults (e.g. Mar Chiquita, Punta Rasa). Although the range area (kernel 95%) included northern sites up to Mar Chiquita Lagoon, these areas are considered as transit zones. The occurrence of tracked individuals around the breeding area, taking into account the relatively low accuracy of GLS devices (c.180 km error, see Methods), is in line with the presence of Olrog’s Gull adults revealed by counts made during winter seasons 2015–2017 in Bahía San Blas (averaging 20 individuals, maximum = 56, n = 3) (N. Suaréz and S. Copello unpubl. data) and also with the relatively higher number of ringed adults compared with other sites along the coast (see below). Similarly, important numbers of adults Olrog´s Gull were recorded during the winter months at the Bahía Blanca estuary (maximum = 129 individuals), an important breeding site located 300 km north of the study colony (Delhey et al. Reference Delhey, Carrete and Martínez2001), suggesting that some birds breeding in that area remain in the estuary during the non-breeding season and/or some Olrog´s Gulls from Bahía San Blas move relatively short distances north during winter. Immatures were also observed during winter seasons in Bahía San Blas (N. Suaréz and S. Copello unpubl. data) and in Bahía Blanca (Delhey et al. Reference Delhey, Carrete and Martínez2001) although in a lower proportion than adults.
Defined core areas included offshore and inland domains. Although terrestrial areas unlikely represent Olrog’s Gull inland movements given their infrequent use of terrestrial and freshwater environments (Chebez and Yorio Reference Chebez, Yorio and Chebez2008), previous studies report their presence in offshore environments. Adult and subadult Olrog’s Gulls were reported during the austral spring, in association with demersal and pelagic fishing vessels operating within the Argentine EEZ, some 200 km East to the study colony (Figure 1) (Seco Pon and Favero Reference Seco Pon and Favero2011, Paz et al. Reference Paz, Seco Pon, Favero, Blanco and Copello2018).
Unlike the results based on areas of tracked individuals, the sightings of ringed adult and young Olrog’s Gulls from our study colony highlight the use of northern coastal areas of Buenos Aires province and Uruguay as stopover or wintering quarters. As mentioned above, several ringed adults were sighted in Bahía San Blas, but most of the rest of the ringed adults and immature individuals were sighted along the northern coasts of Buenos Aires Province. Ringed subadults from Bahía San Blas were also sighted in Uruguay, and although no ringed adults were reported in that coastal sector, flocks of over 100 adult Olrog’s Gulls have been reported (Azpiroz and Caballero-Sadi Reference Azpiroz, Caballero-Sadi, Azpiroz, Jiménez and Alfaro2017) and thus the possible transboundary migration of breeders from Bahía San Blas cannot be ruled out. The reason for this mismatch between tracking and ringing data could be the result of a small sample size of tracked adults and sightings of ringed gulls in relation to the migrating fraction of the population. Besides, the longer distances from the colony to the sighting sites observed in immatures may be the result of the existence of differential migration by age class, as was also observed in other gull species (Liu et al. Reference Liu, Holt, Lei, Zhang and Zhang2006). However, this result has to be taken with caution, as ringing data may provide information on staging areas and not the final destination used as wintering grounds. Among other non-mutually exclusive hypotheses explaining differential migration, the ‘social dominance hypothesis’ states that dominant individuals, which tend to move shorter distances or remain sedentary, exclude subordinate conspecifics from foraging and resting areas close to breeding grounds, forcing them to migrate further away to access similar resources (Marques et al. Reference Marques, Sowter and Jorge2010, Pérez et al. Reference Pérez, Granadeiro, Dias, Alonso and Catry2014). This is somewhat in line with studies in non-breeding areas from south-eastern Buenos Aires province reporting differences in the dominance between gulls of different age class and gender (Zumpano et al. unpubl. data). Another potential hypothesis might be linked to the fact that non-breeders can afford migrating longer distances in comparison to adults, as they do not have to come back to the colonies to breed earlier (Liu et al. Reference Liu, Holt, Lei, Zhang and Zhang2006).
Ringed Olrog’s Gulls were mostly sighted in coastal wetlands characterised by mudflats or saltmarshes with crab beds, in agreement with the species' rather specialised diet. Olrog’s Gulls prey mainly on Varunidae crabs throughout the year (N. granulata, Cyrtograpsus angulatus and C. altimanus) (Copello and Favero Reference Copello and Favero2001, Delhey et al. Reference Delhey, Carrete and Martínez2001, Berón and Favero Reference Berón and Favero2010, Suárez et al. Reference Suárez, Retana and Yorio2011), although during winter months they may also take advantage of alternative food resources such as fishery waste (Berón et al. Reference Berón, Favero and Gómez Laich2007). This use of anthropogenic food sources may explain the sightings in Necochea and Mar del Plata the main fishing and cereal ports in Argentina, although individuals reported from Necochea may have been still in transit to their final wintering grounds while those from Mar del Plata may have been alternating between anthropogenic feeding habitats and crab environments at the nearby Mar Chiquita lagoon. These results support the hypothesis that Olrog’s Gulls are highly dependent on crab habitats during the non-breeding period, raising concerns given the increasing threats faced by these coastal environments due to development and pollution in Argentina and Uruguay (Azpiroz and Caballero-Sadi Reference Azpiroz, Caballero-Sadi, Azpiroz, Jiménez and Alfaro2017, Blanco et al. Reference Blanco, González Trilla, Yorio, Benzaquen, Blanco, Bo, Kandus, Lingua, Minotti and Quintana2017, González Trilla and Blanco Reference Blanco, González Trilla, Yorio, Benzaquen, Blanco, Bo, Kandus, Lingua, Minotti and Quintana2017).
This study provides information for the first time on the migration pattern of individuals of a known colony origin, showing that some adult and young individuals can migrate long distances even crossing international jurisdictions, while other adults remain in their breeding ground during winter (immatures were also sighted at breeding areas but their origin was unknown). The findings of this study are relevant for the design of a conservation strategy for this endemic gull species. Although Olrog’s Gull has been listed in CMS Appendix 1 since 1997, up to now little was known about the migratory behaviour and use of wintering and refuelling areas. Moreover, the information reported here is of critical importance to update key variables used to assess the IUCN Red List conservation status. The partial migration reported here implies than during the non-breeding season, breeding and wintering areas away from the nesting grounds should be considered as one system in the design of conservation strategies for this regionally threatened gull. However, in order to properly identify wintering grounds and elucidate in detail the migration movements of the species, further studies should be undertaken at a higher spatial resolution (using GPS or PTT devices), including tracking of Olrog’s Gull of different age class and individuals from other colonies.
Supplementary Materials
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S0959270920000234.
Acknowledgements
This study was funded by Wildlife Conservation Society, Agencia Nacional de Promoción Científica y Técnica (PICT 33611, 2012-295, 2017-573) and Universidad Nacional de Mar del Plata. Thanks to Centro para el Estudio de Sistemas Marinos (CCT CONICET-CENPAT) and Instituto de Investigaciones Marinas y Costeras (IIMyC, UNMdP, CONICET) for institutional support and Departamento de Áreas Naturales Protegidas (OPDS), Buenos Aires Province, for the permits to conduct the research. We also thank Tatiana Kasinsky and Cristian Marinao for help with field work. We are grateful to two anonymous reviewers who helped to improve the first version of our manuscript.
