INTRODUCTION
Understanding animal movement patterns is not only important for understanding their biology, but is also relevant to conservation planning. This information is vital for species listed in Annex II of the EU Habitats Directive (Council Directive 92/43/EEC), for which Natura 2000 sites need to be designated. However, in aquatic and wide-ranging species such as cetaceans, this is often difficult. One way of tracking them and studying their movements and ranging patterns is to equip them with tracking devices such as satellite transmitters (Ropert-Coudert et al., Reference Ropert-Coudert, Beaulieu, Hanuise and Kato2009; Bograd et al., Reference Bograd, Block, Costa and Godley2010; Block et al., Reference Block, Jonsen, Jorgensen, Winship, Shaffer, Bograd, Hazen, Foley, Breed, Harrison, Ganong, Swithenbank, Castleton, Dewar, Mate, Shillinger, Schaefer, Benson, Weise, Henry and Costa2011). However, this is not always practicable or possible. Alternatively, cetaceans can be identified by natural markings (Hammond et al., Reference Hammond, Mizroch and Donovan1990) and their movement patterns inferred through re-sightings of identified individuals.
The common bottlenose dolphin (Tursiops truncatus, hereafter ‘bottlenose dolphin’) is the most common and widespread cetacean in the northern and central Adriatic Sea (Bearzi et al., Reference Bearzi, Holcer and Notarbartolo di Sciara2004; Fortuna et al., Reference Fortuna, Holcer, Filidei, Donovan and Tunesi2011) and has been the focus of long-term studies in some areas. These include the Cres-Lošinj (Kvarnerić) archipelago in Croatia, northern Adriatic (since 1987; Bearzi et al., Reference Bearzi, Notarbartolo di Sciara and Politi1997; Fortuna, Reference Fortuna2006); the Slovenian, Italian and Croatian waters of the Gulf of Trieste, northern Adriatic (since 2002, Genov et al., Reference Genov, Kotnjek, Lesjak, Hace and Fortuna2008); and the Vis-Lastovo archipelago in Croatia, central Adriatic (since 2007; Holcer, Reference Holcer2012). All of these local populations show relatively high degree of site fidelity, but their movements and ranging patterns are not well understood. Available evidence based on comparison of photo-identification data (Genov et al., Reference Genov, Wiemann and Fortuna2009; Pleslić et al., Reference Pleslić, Rako Gospić, Mackelworth, Wiemann, Holcer and Fortuna2015) and molecular markers (Gaspari et al., Reference Gaspari, Holcer, Mackelworth, Fortuna, Frantzis, Genov, Vighi, Natali, Rako, Banchi, Chelazzi and Ciofi2015a, Reference Gaspari, Scheinin, Holcer, Fortuna, Natali, Genov, Frantzis, Chelazzi and Mourab) suggests some of them may be fairly distinct units.
Bottlenose dolphins inhabiting the Gulf of Trieste and adjacent waters (Figure 1) have been studied since 2002 (Genov et al., Reference Genov, Kotnjek, Lesjak, Hace and Fortuna2008), primarily through photo-identification (Würsig & Würsig, Reference Würsig and Würsig1977; Würsig & Jefferson, Reference Würsig, Jefferson, Hammond, Mizroch and Donovan1990; hereafter ‘photo-ID’). Over 150 individuals have been photo-identified and about half have been encountered on a regular basis over the past decade (Genov et al., Reference Genov, Kotnjek, Lesjak, Hace and Fortuna2008; Genov, Reference Genov2011). The entire range of this local population remains unknown.
The sighting (circles) and stranding (square) locations of the photo-identified bottlenose dolphin, with some locations cited in the text.
In Italy, a volunteer-based cetacean stranding network has been operating since 1986 (Cagnolaro, Reference Cagnolaro1996; Cagnolaro et al., Reference Cagnolaro, Podestà, Affronte, Agnelli, Cancelli, Capanna, Carlini, Cataldini, Cozzi, Insacco, Maio, Marsili, Nicolosi, Olivieri, Poggi, Tommaso Renieri and Wurtz2012), and from 2005 onwards, an institutional monitoring network for stranded marine mammals was established (Pavan et al., Reference Pavan, Bernuzzi, Cozzi and Podestà2013). In recent years, most reported carcasses are recovered and examined following standard necropsy procedures (Geraci & Lounsbury, Reference Geraci and Lounsbury2005). On average, about 10 carcasses are recovered annually in the northern Adriatic Sea, with two per year fresh enough to warrant photo-ID comparisons with existing photo-ID catalogues (University of Padova, unpublished data).
In this paper, we report a re-sighting of a bottlenose dolphin observed alive along the Slovenian coast and found dead on the coast of Italy, about 130 km from the original sighting location.
METHODS AND RESULTS
On 24 April 2014 we observed and photographed a single adult bottlenose dolphin off the Slovenian coast (45°33.086′N 13°29.246′E; Figure 1). The individual was not photographed in the area previously, suggesting it was not part of the ‘resident’ local population. It was also not part of existing photo-ID catalogues in Croatia (D. Holcer, G. Pleslić and N. Rako, personal communication). The dorsal fin of the animal was well-marked (Figure 2), allowing easy identification. In particular, the individual featured a deep but healed cut on the anterior edge of the dorsal fin, suggesting a previous entanglement in fishing gear. In addition, it had an open wound on its left flank, below and posterior to the dorsal fin (Figure 2), roughly 2–3 cm in diameter. Finally, the animal appeared slightly emaciated, but its surfacing movements appeared normal as compared with other conspecifics observed in the area. The mean dive duration was 19.2 s (SD = 4.5, range = 11–28, N = 33). The focal follow of the animal lasted for about 40 min, after which it was terminated to avoid potential disturbance.
Live bottlenose dolphin in Slovenian waters, with a healed lesion in the dorsal fin, suggestive of entanglement in fishing gear. The wound on the flank is of unknown origin. (Photo: T. Genov).
Two days later, on 26 April 2014, we observed the same individual again, 2.8 km from the previous location (45°33.244′N 13°31.383′E; Figure 1). As in the previous encounter, its surfacing movements appeared normal. The mean dive duration was 24.7 s (SD = 44.2, range = 4–195, N = 46). We also observed and photographed the dolphin with a fish in its mouth, suggesting successful feeding. While observing the individual for about 50 min, a sub-group of three other bottlenose dolphins appeared in the area. This sub-group was part of a large group containing 30–40 individuals. Although the sub-group and the single animal passed each other at a distance of about 200 m, the single individual continued to swim on its own, in its original direction, and moved away. As we obtained a sufficient amount of photographs for photo-ID of the first individual, we moved on to photograph the new group.
On 5 May 2014, the same dolphin was found stranded on the shores of Goro (the Po River delta, Italy; 44°49.260′N 12°19.920′E; Figure 1), about 130 km from the two sighting locations. The well-marked dorsal fin, in combination with the flank wound, made the identification straightforward (Figure 3). The post-mortem examination showed that the animal was a 3.1 m male. The animal was found freshly dead, in condition code 2, according to carcass decomposition evaluation criteria (Geraci & Lounsbury, Reference Geraci and Lounsbury2005) and as evaluated by an expert veterinarian. The necropsy results and associated analyses of samples showed that the animal was in poor body condition, and likely died due to a severe endotoxic shock caused by haemolytic E. coli. The animal was apparently old, based on age-related lesions observed during necropsy. Moderately digested remains of an angler fish (Lophius piscatorius) of about 15 cm were found in its gastric chambers and oesophagus.
Dead bottlenose dolphin found stranded on the Italian coast. (Photo: C. Centelleghe.)
To assess the potential effects of sea currents on passive drift of the carcass, we ran simulations using Langrange's drift trajectories (Griffa et al., Reference Griffa, Kirwan, Mariano, Özgökmen and Rossby2007) in a prognostic numeric model. Based on the date, time and position of the last confirmed live sighting, and the date, position and freshness of the carcass, we ran two simulations of drift: forward (from the last sighting) and backward (from the stranding). Specifically, we used two models (Figure 4). The first was the Adriatic Forecasting System, AFS (Oddo et al., Reference Oddo, Pinardi and Zavatarelli2005, Reference Oddo, Pinardi, Zavatarelli and Coluccelli2006; Guarnieri et al. Reference Guarnieri, Oddo, Bortoluzzi, Pastore, Pinardi, Ravaioli, Dahlin, Fleming and Petersson2008, http://oceanlab.cmcc.it/afs/), with a horizontal resolution of 1/45° (approximately 2.2 km). The model utilizes atmospheric data from the European Centre for Medium Range Weather Forecast (ECMWF) and the Mediterranean Forecasting System, MFS (Pinardi et al., Reference Pinardi, Allen, Demirov, De Mey, Korres, Lascaratos, Le Traon, Maillard, Manzella and Tziavos2003; Tonani et al., Reference Tonani, Pinardi, Dobricic, Pujol and Fratianni2008; http://medforecast.bo.ingv.it/), resulting in daily mean velocities of sea currents, sea surface height, temperature and salinity. The second model was the North Adriatic Princeton Ocean Model, NAPOM (Malačič et al., Reference Malačič, Petelin and Vodopivec2012) with a horizontal resolution of about 600 m. The model includes tidal dynamics and utilizes atmospheric data from the ALADIN/SI model by the Slovenian Environmental Agency, and the aforementioned AFS model. NAPOM results are composed of hourly values of sea currents, sea surface height, temperature and salinity. To calculate Lagrange's trajectories we used a procedure developed by Jarle Berntsen (Institute of Marine Research, Norway, http://www.ccpo.odu.edu/POMWEB/contrib-code/tracer.f), assuming that objects only drift at the surface (depth 0–1 m). Data on wind conditions in the period of interest were derived from the oceanographic buoy Vida (Marine Biology Station Piran, http://www.nib.si/mbp/en/buoy).
Clusters of Lagrange's trajectories of the two numerical models: A, Adriatic Forecasting System (AFS); B, North Adriatic Princeton Ocean Model (NAPOM). The upper right cluster represents ‘forward’ trajectories from 26 April 2014, while the lower left cluster represents ‘backward’ trajectories from 5 May 2014. Boxes show the dates of the position on the trajectories (‘dd.m’, where ‘dd’ is the day in month and ‘m’ is the month). See main text for details.
Both models produced similar trajectories and showed that the reported movement was highly unlikely to be a result of passive drift (Figure 4). The backward trajectories, together with the carcass condition, suggest that the animal was likely less than 30 km from the Po River Delta at the time of death (Figure 4). Likewise, the forward trajectories show that a dolphin carcass would not reach the shores of Goro by the stranding date, assuming it died at the time and location of the initial sighting (Figure 4).
DISCUSSION
In north-western Greece, Bearzi et al. (Reference Bearzi, Bonizzoni and Gonzalvo2011) reported several round-trips of bottlenose dolphins between areas up to 265 km apart, showing that despite their relatively high degree of site fidelity, the Mediterranean bottlenose dolphins are capable of considerable movement. Moreover, Gnone et al. (Reference Gnone, Bellingeri, Dhermain, Dupraz, Nuti, Bedocchi, Moulins, Rosso, Alessi, Mccrea, Azzellino, Airoldi, Portunato, Laran, David, Di Meglio, Bonelli, Montesi, Trucchi, Fossa and Wurtz2011) showed that some bottlenose dolphins in the Ligurian Sea make movements between locations as far as 427 km apart. Therefore, although often philopatric (Natoli et al., Reference Natoli, Birkun, Aguilar, Lopez and Hoelzel2005), Mediterranean bottlenose dolphins can make substantial movements over short periods of time, as has been shown elsewhere for this species (Wood, Reference Wood1998; Wells et al., Reference Wells, Rhinehart, Cunningham, Whaley, Baran, Koberna and Costa1999; Silva et al., Reference Silva, Prieto, Magalhães, Seabra, Santos and Hammond2008; O'Brien et al., Reference O'Brien, Berrow, Ryan, McGrath, O'Connor, Pesante, Burrows, Masset, Klötzer and Whooley2010; Robinson et al., Reference Robinson, O'Brien, Berrow, Cheney, Costa, Eisfeld, Haberlin, Mandleberg, O'Donovan, Oudejans, Ryan, Stevick, Thompson and Whooley2012; Cheney et al., Reference Cheney, Thompson, Ingram, Hammond, Stevick, Durban, Culloch, Elwen, Mandleberg, Janik, Quick, Islas-Villanueva, Robinson, Costa, Eisfeld, Walters, Phillips, Weir, Evans, Anderwald, Reid, Reid and Wilson2013).
Knowledge on movements of Adriatic bottlenose dolphins is still limited at best. Available photo-ID data suggest that there is no or little mixing between dolphins from the Gulf of Trieste and those from the Cres-Lošinj archipelago, about 150 km apart (Genov et al., Reference Genov, Wiemann and Fortuna2009). Similarly, Pleslić et al. (Reference Pleslić, Rako Gospić, Mackelworth, Wiemann, Holcer and Fortuna2015) reported only five matches (0.8%) between the Cres-Lošinj archipelago (594 identified individuals) and the Kornati archipelago (81 individuals), about 80 km apart. All of the above suggests a reasonably strong site fidelity in these local populations, a notion supported also by genetic evidence (Gaspari et al., Reference Gaspari, Holcer, Mackelworth, Fortuna, Frantzis, Genov, Vighi, Natali, Rako, Banchi, Chelazzi and Ciofi2015a, Reference Gaspari, Scheinin, Holcer, Fortuna, Natali, Genov, Frantzis, Chelazzi and Mourab). However, the entire ranges of these local populations remain unknown (Genov et al., Reference Genov, Kotnjek, Lesjak, Hace and Fortuna2008; Pleslić et al., Reference Pleslić, Rako Gospić, Mackelworth, Wiemann, Holcer and Fortuna2015). Temporary emigration (Fortuna, Reference Fortuna2006) and the occurrence of transient or visitor animals (Genov et al., Reference Genov, Kotnjek, Lesjak, Hace and Fortuna2008) have been reported, but there are currently no data that would allow photo-ID comparisons with other, unstudied areas within the northern and central Adriatic Sea. The area surrounding the Po River delta is one such unstudied area. This area and the Gulf of Trieste are separated by about 130 km, with no islands or geographic barriers between them, and minimal variability in depth. Bottlenose dolphin habitat use (Cañadas & Hammond, Reference Cañadas and Hammond2006; Torres et al., Reference Torres, Read and Halpin2008), foraging specializations (Connor et al., Reference Connor, Wells, Mann, Read, Mann, Connor, Tyack and Whitehead2000; Bearzi et al., Reference Bearzi, Fortuna and Reeves2008) and genetic population structure (Gaspari et al., Reference Gaspari, Holcer, Mackelworth, Fortuna, Frantzis, Genov, Vighi, Natali, Rako, Banchi, Chelazzi and Ciofi2015a; Louis et al., Reference Louis, Viricel, Lucas, Peltier, Alfonsi, Berrow, Brownlow, Covelo, Dabin, Deaville, De Stephanis, Gally, Gauffier, Penrose, Silva, Guinet and Simon-Bouhet2014) are often strongly linked to particular habitat characteristics. It is therefore possible that movements of bottlenose dolphins between these two areas are more common than is currently known.
Information on stranded animals, if not used cautiously, can bias inferences on their biology, including movements (Williams et al., Reference Williams, Gero, Bejder, Calambokidis, Kraus, Lusseau, Read and Robbins2011). We considered this here. In particular, the main currents in the northern Adriatic move in a cyclonic fashion, going northward along the eastern shores of the basin, turning counter-clockwise and then returning southward along the western side (Mauri & Poulain, Reference Mauri and Poulain2001). As bottlenose dolphins have a positive buoyancy at the surface (Williams et al., Reference Williams, Davis, Fuiman, Francis, Le Boeuf, Horning, Calambokidis and Croll2000; Kipps et al., Reference Kipps, McLellan, Rommel and Pabst2002), the animal might have died close to its initial sighting location, and then drifted with sea currents until reaching the final stranding location. However, the dolphin was found freshly dead, in condition code 2 (Geraci & Lounsbury, Reference Geraci and Lounsbury2005). This indicates that it died very recently prior to being found (24–48 h) and suggests that the reported movement is a real one, rather than an effect of currents. The simulations (Figure 4) strongly support this. Next, the single case presented here may indeed reflect the rare nature of such movements, and the apparently poor health of the individual may have affected its behaviour. Hence, the reported movement may not be generally representative of free-ranging bottlenose dolphins in the northern Adriatic Sea. However, the rareness of similar records may simply indicate that such information, for the large part, is still generally missing, and that such movements may be occurring undetected.
The aim of this paper is not to draw conclusions on bottlenose dolphin movement patterns. Instead, it highlights the need for a better understanding of these patterns and the need for collaboration and data sharing. The power of photo-ID is sometimes underestimated, but this simple method can provide valuable insight into a number of important issues, including regional and long-distance movements of cetaceans (Frantzis et al., Reference Frantzis, Airoldi, Notarbartolo di Sciara, Johnson and Mazzariol2011; Stevick et al., Reference Stevick, Neves, Johansen, Engel, Allen, Marcondes and Carlson2011; Genov et al., Reference Genov, Bearzi, Bonizzoni and Tempesta2012), particularly in species with high mark rates, such as the bottlenose dolphin. Photo-ID comparisons, both of live and dead animals, can provide highly useful information and should therefore be encouraged. Furthermore, additional photo-ID data should be collected in previously unsurveyed areas, to facilitate a better understanding of movement patterns in Adriatic bottlenose dolphins. This has implications for the conservation of mobile marine predators such as the northern Adriatic bottlenose dolphin population, which ranges across borders of three countries.
ACKNOWLEDGEMENTS
We are grateful to Aljaž Malek, Joëlle De Weerdt and Danijel Mrvoš for their assistance in the field, and to Draško Holcer, Nikolina Rako and Grgur Pleslić for feedback on a potential match with other photo-ID catalogues. Caterina Fortuna provided constructive comments and encouragement to publish this information, while comments by Giovanni Bearzi and Joan Gonzalvo helped improve the manuscript.
FINANCIAL SUPPORT
This work was partially funded through the project NETCET (Network for the Conservation of Cetaceans and Sea Turtles in the Adriatic) through the IPA Adriatic Cross-Border Cooperation Programme, and through support by OceanCare.
