Biotelemetry provides an approach for identifying movement patterns and post-release fate that can be used to inform conservation strategies, yet presents unique challenges for rare or endangered species. Despite increasingly sophisticated technology, electronic tags designed to communicate with the Argos system can fail to transmit position estimates [1]. Distinguishing these tag failures from animal mortality events is critical when examining populations in peril [2]. Double tagging, or the application of complementary tag types designed to report both position estimates and post-release fate, is one approach for assessing post-release fate in the event of an Argos tag failure.

The great hammerhead (Sphyrna mokarran) is a highly migratory species of conservation concern. Great hammerheads are listed on the International Union for the Conservation of Nature (IUCN) red list as Endangered [3] and thought to be functionally extinct throughout parts of their historical range (e.g., Gulf of California, [4]). Despite a lack of a directed fishery in the northwest Atlantic Ocean, great hammerheads are caught incidentally in longline fisheries, where they suffer high at-vessel [5, 6] and post-release [7, 8] mortality. Given their susceptibility to capture stress and subsequent post-release mortality, the application of capture and tagging techniques must be approached cautiously.

Despite this documented vulnerability to capture stress, recent applications of biotelemetry have advanced our understanding of great hammerhead distribution, particularly in the western Atlantic Ocean. Using a combination of techniques, Guttridge et al. [9] identified large-scale migrations, seasonal residency, and site fidelity in great hammerheads tracked in Florida and the Bahamas. Similarly, Graham et al. [10] identified areas of core habitat use in relationship to marine protected areas along the coast of Florida and the Bahamas, but noted that their findings may not be representative of the entire population.

Less is known about the movements of great hammerheads in the Gulf of Mexico, despite this being a well-established part of their range. Great hammerheads are encountered via fishery-independent sampling in the northern Gulf of Mexico from March to November [11], across a broad range of sizes. Given the conservation status of this species, combined with our opportunistic access to them via ongoing fishery-independent sampling projects, we sought to investigate movements of great hammerheads in the northern Gulf of Mexico. However, their documented vulnerability to capture stress necessitated a cautious approach. Therefore, the objective of the current study was to investigate movement patterns in great hammerheads using a combination of biotelemetry tags; our double-tagging approach provided a means for collecting both position information and post-release survival verification.

Great hammerheads were captured during fishery-independent bottom longline sampling in the northern Gulf of Mexico in 2016 following methods outlined in [11, 12]. Briefly, a bottom longline consisting of 1.85 km of 4-mm monofilament (545-kg test) mainline was set with 100 gangions. Gangions were 3.66 m in length and consisted of a longline snap and a 15/0 circle hook (Mustad model 39960D) baited with Atlantic Mackerel (Scomber scombrus). Each gangion was made of 3.66 m of 3-mm monofilament (320-kg test). All bottom longlines were set (i.e., soaked) for 1 h, during which time a Hydrolab MS5 data sonde and surveyor were used to record surface and bottom values for temperature (°C).

Upon retrieval of the bottom longline, great hammerheads that could be boated safely were removed from the main line, and the circle hook severed with bolt cutters. Once boated, a saltwater hose was inserted into the mouth of the fish to provide ambient seawater across the gills. For each individual, sex, length (fork and stretch total in centimeters) and maturity stage (in males) were recorded. Maturity in males was assessed following [13]. Great hammerheads that were deemed to be in good condition (i.e., active and responsive, little or no visible external damage, [14]) were fitted with three tags: a conventional dart tag (Floy) and two satellite tags. Smart position and temperature transmitting tags (SPOT6, Wildlife Computers) were attached to the dorsal fin. Fin-mounted SPOT tags (hereafter SPOT tags) estimate position via Doppler-shift calculations, which are transmitted to the ARGOS satellite system when the saltwater switch is activated (i.e., when the tag breaks the surface). To examine whether a non-reporting SPOT tag was due to lack of surface events as opposed to post-release mortality, each great hammerhead was also tagged with a survivorship pop-off archival transmitting tag (sPAT tag, Wildlife Computers).

sPAT tags are specifically optimized to determine fate (alive or deceased) on animals post-release. These tags record daily minimum and maximum values for depth and temperature and indicate whether daily light levels varied; they are programmed to record these values for 30 days, then pop-off, and transmit data to the ARGOS satellite system. Raw data are analyzed in-house by Wildlife Computers, who provide a report indicating pop-off date, location, daily values for temperature and depth, an indication of daily changes in light, and fate. Fate (i.e., the reason for tag release) is used to infer survivorship and is classified into one of four categories: (1) completed deployment, (2) sinker, a tag attached to a sinking animal that releases at 1700 m, (3) floater, a tag at the surface, and (4) sitter, a constant depth reading shallower than 1700 m. Completed deployment (category 1) indicates survival, sinker and sitter (categories 2 and 4) indicate mortality, and floater (category 3) could be interpreted as either survival or mortality, based on inspection of depth data [15].

Three great hammerheads (GH 1, 2, and 3) were captured and tagged with conventional dart, SPOT, and sPAT tags in May, August, and September of 2016, respectively (Table 1). Water temperature (top/bottom) during tagging was 25/22, 31/24, and 29/29 °C, respectively. Tagged sharks were 187, 203, and 250 cm TL. All three sharks were tagged in the northern Gulf of Mexico and were released quickly (< 120 s) (Fig. 1). The mean number of days at liberty measured by the sPAT tags was 24 days (ranging from 20 to 30 days), and daily changes in depth and light level were recorded for all three sharks. The SPOT tag attached to GH 2, the smallest individual (187 cm TL), was the only SPOT tag that reported transmissions sufficient to estimate positions and did so over a 41-day period.

GH	Sex	FL	TL	Tagging date (2016)	Pop-off date (2016)	Days at liberty	SPOT transmission days	Tagging Location		Pop-off location		Distance traveled (km)
								Lat	Lon	Lat	Lon
1	M	185	250	5/13	6/5	23	0	29.66	− 88.30	29.65	− 87.94	35
2	F	141	187	8/4	8/24	20	81	30.04	− 87.69	28.62	− 91.33	387
3	F	150	203	9/16	10/16	30	79	30.30	− 88.38	30.24	− 88.60	22

Great hammerhead 1 (GH 1) was an immature male (250 cm TL) tagged on May 13, 2016. No positions were estimated from the SPOT tag. The sPAT deployed prematurely on June 5 after 23 days and was classified as a “floater.” GH 1 made daily vertical movements from the surface (0–2 m) to an average depth of 37.2 m (± 6.3 standard deviations, SD) and traveled a minimum straight-line distance (between the tagging and sPAT pop-off locations) of 35 km. Given the consistent and substantial daily changes in depth over the 23-day period, we infer the fate of GH 1 as survival (Fig. 2a, Table 1).

Great hammerhead 2 (GH 2) was a female (187 cm TL) tagged on August 4. The SPOT tag on GH 2 reported positions to the ARGOS satellite system for 81 days. For the first 3 days post-release, the shark moved northeast ~ 100 km. The next SPOT tag transmission was on August 25, in agreement with the sPAT tag pop-off position. The sPAT deployed prematurely on August 24 after 20 days and was classified as a “sitter.” GH 2 made daily vertical movements from the surface (0 m) to an average depth of 50.8 m (± 21.4 SD), with daily changes in depth up to 98 meters. GH 2 showed the greatest horizontal movement of the three tagged sharks, traveling a straight-line distance between the tagging and sPAT pop-off locations of 387 km. Interestingly, the SPOT tag on GH 2 reported position estimates on September 14, approximately 3 weeks after the sPAT assigned the condition “sitter,” a state typically indicative of mortality. Given the consistent and substantial daily changes in depth over a 20-day period, we infer the fate of GH 2 as survival (Fig. 2b, Table 1).

Great hammerhead 3 (GH 3) was a female (203 cm TL) tagged on September 16. While the SPOT tag transmitted data to ARGOS over a 79-day period, no transmissions were sufficient to estimate a position. The sPAT popped-off on October 16 after 30 days and was classified as a survival based on a complete deployment. GH 3 made daily vertical movements from the surface (0 m) to an average depth of 5.9 m (± 1.9 SD). This individual inhabited depths 6 m or less over 93% of the time, consistent with the bathymetry of Mississippi Sound. GH 3 showed the most restricted horizontal movements, traveling a straight-line distance between the tagging and sPAT pop-off locations of 22 km (Fig. 2c, Table 1).

The application of complementary satellite tags provided insight into vertical and horizontal habitat use for great hammerheads that would not have been evident using a single tag type. Critically, one of the three tagged sharks may have been presumed dead if we had used SPOT tags alone. As such, despite their demonstrated utility on other shark species that frequent the surface, fin-mounted SPOT tags may not be the best option for tracking great hammerheads. Despite a low sample size (n = 3), our findings are noteworthy as they characterize movements from immature animals tagged in an area where we know comparatively little about the movements of this species.

Data from the sPAT tags showed all three great hammerheads made daily vertical movements from the surface to depth. Given the consistent use of surface waters and the exaggerated height of the dorsal fin to which the SPOT tag was attached, why were so few SPOT positions reported? To begin with, it is possible that the SPOT tags simply malfunctioned; for example, no SPOT transmissions were ever received for GH 1. However, the SPOT tag on GH 2 transmitted to the ARGOS system, but not enough to generate a single position estimate, which requires at least two consecutive transmissions. This suggests that great hammerheads may not spend sufficient time (i.e., at least 90 s) on the surface to allow a fin-mounted SPOT tag to communicate with the ARGOS system long enough to estimate positions. Given the previously successful application of towed SPOT tags on this species [10], towed packages may present the best option compared to fin-mounted SPOT transmitters.

An alternative explanation for why our fin-mounted SPOT tags failed to generate position estimates for sharks at the surface invokes a swimming behavior recently described for this species. Using a combination of accelerometer loggers, animal-borne video, and observations of aquarium-held sharks, Payne et al. [16] document repeated and prolonged swimming by great hammerheads at rolled angles between 50 and 75°. Using a hydrodynamic model and wind tunnel simulations, Payne et al. [16] estimate a reduction in drag forces that equates to an energetic savings of ~ 10%. Hence, while all great hammerheads tagged in this study frequented the surface, perhaps they were oriented such that the SPOT tag was still below the surface and thus unable to transmit. Alternatively, biofouling could have precluded the saltwater switch on the SPOT tag from functioning. Recent recoveries of fin-mounted SPOT tags from a scalloped hammerhead (Sphryna lewini) and tiger shark (Galeocerdo cuvier) in the northern Gulf of Mexico at liberty for less than 6 months revealed extensive biofouling which had prevented the tag from transmitting any location estimates (Drymon and Wells unpublished data). Previous studies demonstrate that SPOT tags are highly effective for tiger sharks, with reporting rates as high as 100% [7], but it may be that the highly productive waters of the northern Gulf of Mexico promote rapid biofouling compared to more oligotrophic waters like south Florida and the Caribbean. Regardless of the mechanism, fin-mounted SPOT tags appear to be a poor choice for examining short-term horizontal movements or post-release survivorship for great hammerheads in the northern Gulf of Mexico, despite their frequent use of surface waters.

Our findings add to a body of knowledge highlighting shallow water habitat use by great hammerheads. Over a thirty-day period, GH 3 used waters 6 m or less during 93% of the time at liberty. This finding is similar to the results presented by Hueter and Manire [17], who noted that great hammerheads along the southwest coast of Florida were only seen at depths shallower than ~ 6 m. In the US bottom longline fishery, great hammerheads are most commonly captured in waters less than 20 m [6]. Roemer et al. [18] described six instances of extreme shallow water habitat use by great hammerheads, five of which involved prey capture. Great hammerheads are known to frequent the shallow waters off south Florida in pursuit of tarpon (Megalops atlanticus) [19, 20]. Given the abundance of rays, a preferred prey item for this species [21, 22] in the shallow waters of the northern Gulf of Mexico, consistent use of shallow waters by great hammerheads in this region, may represent foraging behavior, although investigations into the feeding ecology of this species would be required to confirm this.

Generally, larger sizes afford increased vagility; as a consequence, many marine animals have larger home ranges as adults [23]. Great hammerheads are one of the largest-bodied species of predatory sharks [20] and can travel long distances over short periods of time [24], including movements exceeding 1500 km [9]. Acknowledging our relatively short tracking periods, two of the three great hammerheads tagged in this study showed movement less than 35 km between tagging and pop-off locations. Counterintuitively, the smallest shark tagged (GH 2) moved nearly 400 km between tagging and pop-off locations; following what appeared to be clear eastward movement from Alabama into Florida, both fin-mounted SPOT and sPAT tags confirmed GH 2 west of the Mississippi River. Individual variation in habitat use has been demonstrated for other large predatory sharks, including tiger [25] and bull shark (Carcharhinus leucas, [26]). Thus, the individual variability in movement patterns illustrated in the current study appears to be common. Such high variability suggests that effective great hammerhead conservation and management plans will require movement data from individuals of all sizes and throughout the extent of their range. Interestingly, the longest horizontal movement in this study was demonstrated by a 187 cm TL female, likely the smallest satellite-tagged great hammerhead to date.

Mitigating the troubling trends in marine defaunation [23] will require immediate actions coupled with innovative approaches. In the case of great hammerheads, their IUCN status, susceptibility to at-vessel and post-release mortality, and high degree of specialization make them particularly vulnerable to over-exploitation [27]. While this study represents a limited number of fish monitored over a relatively short period, our findings highlight some of the difficulties inherent in understanding movements of highly migratory species. In particular, the application of mortality tags in addition to position tags provided the data necessary to interpret vertical movement, as well as verify post-release survivorship. Our findings provide valuable information to guide future studies seeking to understand movement and migration in this species.