Examining the functional role of current area closures used for the conservation of an overexploited and highly mobile fishery species

Abstract

Protecting essential habitats through the implementation of area closures has been recognized as a useful management tool for rebuilding overfished populations and minimizing habitat degradation. School shark (Galeorhinus galeus) have suffered significant stock declines in Australia; however, recent stock assessments suggest the population may have stabilized and the protection of closed nursery areas has been identified as a key management strategy to rebuilding their numbers. Young-of-the-year (YOY) and juvenile G. galeus were acoustically tagged and monitored to determine ontogenetic differences in residency and seasonal use of an important protected nursery area (Shark Refuge Area or SRA) in southeastern Tasmania. Both YOY and juvenile G. galeus showed a distinct seasonal pattern of occurrence in the SRA with most departing the area during winter and only a small proportion of YOY (33%) and no juveniles returning the following spring, suggesting areas outside the SRA may also be important during these early life-history stages. While these behaviors confirm SRAs continue to function as essential habitat during G. galeus early life history, evidence of YOY and juveniles emigrating from these areas within their first 1–2 years and the fact that few YOY return suggest that these areas may only afford protection for a more limited amount of time than previously thought. Determining the importance of neighbouring coastal waters and maintaining the use of traditional fisheries management tools are therefore required to ensure effective conservation of G. galeus during early life history.

acoustic telemetry, functional role, Galeorhinus galeus, nursery area

Introduction

Anthropogenic impacts on marine life such as overfishing, pollution, and coastal development have resulted in widespread collapse of many global fisheries and degradation of marine habitats (Pauly et al., 1998; Hutchings, 2000; Myers and Worm, 2005). Management tools such as catch quotas and fishing gear restrictions have traditionally been used to curb overfishing; however, in more recent times, area closures and implementation of marine reserves have gained popularity as a supplementary management strategy for enhancing stock recovery at the same time protecting important marine habitats (Roberts et al., 2001; Gell and Roberts, 2003; Hilborn et al., 2004; Gaylord et al., 2005; Thorpe et al., 2011).

Area closures typically involve protecting areas regularly used by key species for feeding, predator avoidance, or reproduction by restricting access to activities such as fishing, mining, or other human disturbances (Gell and Roberts, 2003). Area closures are considered most beneficial to species such as invertebrates and reef fish that are relatively sedentary and spend a significant proportion of their time within the protected area (Gell and Roberts, 2003; Gerber et al., 2003; Hilborn et al., 2004). However, for more highly migratory species such as sharks and pelagic fish, the protection afforded by area closures may be limited given the difficulties of protecting expansive areas that cover their large-scale migrations and movement patterns (Kramer and Chapman, 1999; Shipp, 2003; Heupel et al., 2007; Kinney and Simpfendorfer, 2009).

Spatial closures for sharks have typically focused on protecting areas used during vulnerable stages such as mating, pupping, and early life-history stages when their movements are often confined to discrete areas (Heupel et al., 2007). Although some shark nursery areas have been protected, many of these have been protected with very little prior understanding of how and when sharks use such areas (Heupel and Simpfendorfer, 2005a; Heupel et al., 2007). There is also a view that protecting nursery areas may have limited value to shark populations if more traditional fishery management is not used to protect juveniles and adults outside the closed area (Kinney and Simpfendorfer, 2009). Therefore, determining the movement behaviors of juvenile sharks and the nature of their association with particular areas during early life history is crucial to implementing more effective area closures or evaluating their effectiveness for protecting these life-history stages (Speed et al., 2010).

The school shark (Galeorhinus galeus) is found circumglobally in temperate waters and has been subject to intense fishing pressure resulting in significant stock declines, particularly in Australia (Walker et al., 2006). As a strategy to aid rebuilding the population, recognized nursery grounds around Tasmania and Victoria were declared as Shark Refuge Areas (SRAs) as early as 1954. In such areas, the taking of any species of shark is prohibited (Walker, 1999). However, despite 40 years of protection, surveys of SRAs conducted during the 1990s found that young-of-the-year (YOY) and juveniles were either absent from or occurred in very small numbers in many of these refuges (Stevens and West, 1997; Walker, 1999). Overfishing is considered the main reason for the population collapse (Walker, 1999); however, it has also been suggested that habitat degradation, particularly the loss of seagrass meadows in Victorian nursery areas may have contributed to this decline (AFMA, 2009). Subsequently, further fishing restrictions, gear modifications, and fishing ground closures have been implemented in an effort to reduce fishery captures (McLoughlin, 2008) and a School Shark Rebuilding Strategy (SSRS) has been in place since 2008 to promote stock recovery (AFMA, 2009). Recent stock assessments estimate G. galeus populations in southern Australia are at 9–14% of virgin biomass, although modelling indicates that the population may have stabilized (AFMA, 2009; SharkRAG, 2010). Furthermore, recent research fishing in southern Tasmanian SRAs has recorded significantly higher catches of YOY and juvenile G. galeus than reported in similar surveys conducted in the 1990s (authors unpublished data). While the relative contributions of the various management measures to this stabilization are unclear, identifying additional nurseries, and protecting current nursery areas has been recognized as the priority in the SSRS (SharkRAG, 2010).

Despite the high importance placed on SRAs as a management measure, little is known about the role that they play in supporting G. galeus during the early life-history phases, with much of the previous research on SRAs in southeast Australia assessing patterns in relative abundance and movement based on conventional tag-recapture methods (Olsen, 1954; Stevens and West, 1997). Knowledge of how G. galeus utilize these areas in space and time is required to evaluate the effectiveness of the SRAs in the overall conservation of this species. Furthermore, understanding how G. galeus use these areas may also provide greater insight into this species' functional role in these areas, particularly in relation to predator–prey dynamics (Barnett and Semmens, 2012). In this study, acoustic telemetry was applied to describe spatial and temporal movement behaviors, including seasonality of residency of YOY and juvenile G. galeus, within a long-established nursery ground off southeastern Tasmania. Duration of residency and patterns of emigration from the protected area are also examined to inform an assessment of the effectiveness of the protected area in the management and conservation of this species.

Material and methods

Ethics statement

This study was conducted under Department of Primary Industries Parks Water and Environment permit #11055 and with approval from the University of Tasmania Animal Ethics Committee (# A0011882).

Study site

The main study area was located in an SRA in southeast Tasmania, Australia (42° 53.710′S 147° 34.228′E) that incorporates Pitt Water (PW), Frederick Henry Bay (FHB), and Norfolk Bay (NB) (Figure 1). PW is a shallow estuary (average depth 4 m; maximum depth 9 m) originating from the Coal River and is made up of mostly intertidal sand flats and a narrow tidal entrance that connects to the deeper waters of Frederick Henry and NB (average depth 15 m; maximum depth 44 m).

Figure 1.

Map of main study area showing position of acoustic receivers (filled circles) in southern Tasmania (a and b). Note that receivers without a unique identification have been pooled to form a resource (UPW and LPW) or curtain (E and H). Dotted line represents Shark Refuge Area boundary and grey lines are the 5 m and 10 m depth contours. Triangles represent the approximate capture and tagging sites in UPW and FHB. Map (a) shows the receiver curtains MARIA and CAPE BARREN extending to the continental shelf break (200 m).

Fifty-eight acoustic receivers (VEMCO Ltd, Halifax, Canada) were deployed either as stand-alone units (FHB1–20) or in a line with overlapping detection ranges (E and H) so as to form a gate or curtain through which a tagged animal would need to pass to confirm it had entered a given area (Heupel et al., 2006) (Figure 1). In PW, receivers were deployed in a Vemco Positioning System (VPS) array (Figure 1) to examine fine-scale movements and behaviour of G. galeus (authors unpublished data), and as stand-alone units located throughout the area but in particular in a main channel. Previous range testing in NB (42° 59.943′S 147° 47.153′E), which adjoins FHB (curtain E; Figure 1), using the same transmitters as in this study had determined that 100% of tag transmissions could be detected at a range of 400–500 m (Barnett et al., 2011). Therefore curtain receivers were deployed at a maximum distance of 800 m apart. The entire array was deployed on 6 January 2012 and recorded data until 22 May 2013.

In addition to the main study area, an array of 66 acoustic receivers was also deployed off the east coast of Tasmania by the Australian Animal Tagging and Monitoring System and Ocean Tracking Network forming two curtains which extended from mainland Tasmania to Maria Island then to the continental shelf break (MARIA—27 receivers) (42° 40.874′S 148° 15.101′E), and from the most easterly point of Cape Barren Island to the continental shelf break (CAPE BARREN—39 receivers (Figure 1) (40° 28.901′S 148° 39.189′E).

Acoustic tagging

Sharks were captured using bottom-set baited longlines and were measured from the snout to the tip of the tail to the nearest mm (total length or TL mm) and their sex recorded. Sharks were categorized as either YOY (<500 mm TL) or juveniles (>500 mm but typically >600 mm TL and <1000 mm TL) in January and February 2012 (austral summer). Previous seasonal sampling from the same area suggests that during summer individuals >500 mm TL are typically 1+ years of age, with most YOY (0+) typically 350–450 mm TL (Stevens and West, 1997).

YOY were fitted with either VEMCO V9 2L (n = 6) or V13 1L (n = 26) coded acoustic transmitters (transmission off times: random between 120 and 180 s; predicted battery life: 2 and 5 years, respectively). Also, one YOY and seven juveniles were implanted with V13P 1L acoustic-sensor tags (transmission off times: random between 60 and 180 s; predicted battery life: 5 years) (VEMCO Ltd., Halifax, Canada) (Table 1). Sharks were held ventral side up on a piece of foam with running seawater pumped over the gills. The acoustic tag was surgically inserted in the peritoneal cavity by making a 1–2 cm incision in the abdominal wall, and closing the incision using surgical sutures (Braun Safil^® HS26s). Aseptic techniques were used during all stages of the surgery, taking no >2–5 min to complete, after which the animal was released back into the water. Animals were held in the water boat side until they could swim unassisted before being released. Sharks were tagged in PW (32 YOY and one juvenile) and FHB (seven juveniles) between January 2012 and May 2012 (Figure 1). In PW, sharks were captured and released near the VPS array, and between receivers FHB3 and FHB4 in FHB (Figure 1).

Table 1.

Details of YOY (N) and juvenile (J) G. galeus fitted with acoustic or sensor (P) tags.

ID	Location	Tagging date	Sex	TL	Age	Tag type	Predicted battery life	Last detection date	Days detected	% detected	Days present	% present	Return
4,061	UPW	1/02/2012	M	385	N	V9-2L	633	19/03/2012	45	9	477	100
4,062	UPW	1/02/2012	M	390	N	V9-2L	633	16/02/2012	16	3	477	100
4,063	UPW	11/01/2012	F	350	N	V9-2L	633	14/01/2012	4	1	498	100
4,064	UPW	11/01/2012	F	355	N	V9-2L	633	15/01/2012	5	1	498	100
4,067	UPW	4/02/2012	M	395	N	V9-2L	633	16/02/2012	13	3	474	100
4,069	UPW	1/02/2012	F	380	N	V9-2L	633	15/02/2012	15	3	477	100
14,714	UPW	11/03/2012	M	460	N	V13P-1L	879	24/07/2012	75	17	135	31	**
31,159	UPW	15/01/2012	F	365	N	V13-1L	1336	11/04/2013	247	50	452	91	***
31,160	UPW	15/01/2012	M	370	N	V13-1L	1336	15/06/2012	131	27	152	31	*‡†
31,161	UPW	15/01/2012	F	355	N	V13-1L	1336	12/02/2012	26	5	494	100
31,162	UPW	15/01/2012	M	333	N	V13-1L	1336	9/02/2012	18	4	494	100
31,163	UPW	11/01/2012	F	370	N	V13-1L	1336	5/02/2012	12	2	498	100
31,164	UPW	15/01/2012	M	370	N	V13-1L	1336	6/07/2012	156	32	494	100
31,165	UPW	15/01/2012	M	388	N	V13-1L	1336	6/02/2012	14	3	494	100
31,166	UPW	15/01/2012	M	340	N	V13-1L	1336	15/11/2012	127	26	305	62	**
31,167	UPW	15/01/2012	F	358	N	V13-1L	1336	10/04/2013	304	62	451	91	**
31,169	UPW	18/01/2012	M	405	N	V13-1L	1336	21/06/2012	159	32	155	32	*†
31,170	UPW	18/01/2012	F	400	N	V13-1L	1336	8/02/2012	15	3	491	100
31,171	UPW	18/01/2012	M	350	N	V13-1L	1336	20/01/2012	3	1	491	100
31,172	UPW	18/01/2012	F	385	N	V13-1L	1336	19/04/2013	320	65	457	93	**†
31,173	UPW	15/01/2012	F	442	N	V13-1L	1336	10/07/2012	156	32	177	36	**
31,174	UPW	1/02/2012	M	395	N	V13-1L	1336	12/02/2012	8	2	477	100
31,175	UPW	27/01/2012	F	385	N	V13-1L	1336	21/07/2012	144	30	176	37	**†
31,176	UPW	20/01/2012	M	395	N	V13-1L	1336	27/05/2012	124	25	489	100
31,178	UPW	18/01/2012	M	370	N	V13-1L	1336	7/02/2012	9	2	491	100
31,181	UPW	1/02/2012	F	405	N	V13-1L	1336	21/06/2012	114	24	141	30	**†
31,183	UPW	1/02/2012	M	380	N	V13-1L	1336	14/05/2012	104	22	103	22	*
31,184	UPW	1/02/2012	F	392	N	V13-1L	1336	7/06/2012	106	22	127	27	**
31,194	UPW	11/01/2012	F	365	N	V13-1L	1336	Not detected
31,196	UPW	11/01/2012	F	360	N	V13-1L	1336	17/01/2012	7	1	498	100
31,197	UPW	11/01/2012	F	365	N	V13-1L	1336	15/05/2013	181	36	498	100
31,198	UPW	11/01/2012	F	370	N	V13-1L	1336	9/02/2012	24	5	498	100
14,727	UPW	19/01/2012	M	675	J	V13P-1L	879	18/04/2012	91	19	90	18	*
14,713	FHB	29/02/2012	M	510	J	V13P-1L	879	21/04/2013	286	64	417	93	**
14,718	FHB	29/02/2012	F	680	J	V13P-1L	879	12/03/2012	13	3	12	3	*
14,722	FHB	28/02/2012	M	730	J	V13P-1L	879	12/03/2012	14	3	13	3	*
14,723	FHB	28/02/2012	F	745	J	V13P-1L	879	12/03/2012	13	3	13	3	*
14,728	FHB	29/02/2012	F	630	J	V13P-1L	879	13/04/2012	45	10	44	10	*
14,729	FHB	29/02/2012	F	710	J	V13P-1L	879	11/04/2012	38	8	42	9	**
14,733	FHB	29/02/2012	F	685	J	V13P-1L	879	13/03/2012	11	2	13	3	**

Date is the day animal was tagged. TL is total length in millimetre. Predicted battery life is the estimated number of days the tag was expected to transmit. Days detected are the number of days an individual was detected by the acoustic array; % detected represents the percentage of days detected from the date of tagging until the end of the study (i.e. 22/5/2013); Days present represents the number of days the animal was considered inside the SRA (i.e. had not passed the H-curtain); % present represents the percentage of days the animal was considered inside the SRA during its time at liberty. The symbols in the Return column indicate whether an animal departed the SRA and did not return (*), departed and returned to FHB (**), or departed and returned to UPW (***). † Detected by MARIA curtain. ‡ Detected by CAPE BARREN curtain.

Data analysis

Seasonal residency and use of SRA

Seasonal use of the SRA (% of time animals spent within the SRA) was determined by examining a visual plot of daily detections and by dividing the total number of days an individual was detected in the SRA by the total number of days that animal had been at liberty since tagging. An animal was considered present in the SRA if it was detected by any receiver within PW or FHB more than once per day. FHB comprised receivers FHB1-20 and curtains E and H. An animal was considered to have departed the SRA if it was detected on the H-curtain, and subsequently went undetected by any receiver in the SRA for >1 d.

Seasonal distribution within SRA

Receivers were grouped into five locations and the number of days an animal was detected at each location used to compare the seasonal distribution of YOY and juveniles throughout the SRA. The five locations were: Upper Pitt Water (UPW) (receivers to the north of the causeway), Lower Pitt Water (LPW) (receivers to the south of causeway), FHB (receivers FHB1-20), and curtains E and H (Figure 1). The number of sharks detected per day for each location for each month was analysed using circular statistics (Oriana 4 software, Kovach Computing Services). Rao's Spacing Test was used to test for uniformity in detections over a year. For this purpose, we used a 12-month subsample of the data, collected between January and December 2012.

Fine-scale ontogenetic utilization and spatial overlap within SRA

Fine-scale utilization of the SRA by YOY and juveniles was determined by examining the total number of hours each animal was detected at the geographical location of stand-alone receivers (FHB1-20) or groups of receivers (UPW, LPW, E and H) in a given day. If an animal was detected at least once in a given hour for that day then it was considered as being present during that hour. Using the stand-alone and grouped receivers, spatial overlap between YOY and juveniles was then compared using niche overlap analysis in the EcoSimR package (Gotelli and Ellison, 2013) with R statistics software (R Development Core Team, 2013). Pianka's index (O) was selected and permutated 1000 times using the RA3 algorithm (Meyer et al., 2009). The degree of overlap is presented by values between 0 and 1, where 0 = no overlap and 1 = 100% overlap. In addition, a log-likelihood test (χ²) was performed using the adehabitatHS package (Calenge, 2006) in R statistics software to test for individual selection (w_i) for particular receivers. Selection ratios >1 indicate a preference for a particular receiver, whereas values <1 indicate avoidance (Manly et al., 2003). Kernel utilization densities (KUDs) were also used to visually estimate the preference for each receiver and approximate area used, the 95% fixed kernel representing the overall use of available receivers and the 50% fixed kernel the receivers most often used. KUDs were estimated using the Hawth's Analysis Tools for ArcGIS 9.3. Minimum movement paths between each possible combination of two stand-alone and/or grouped receivers made by each animal were also summed and mapped to show approximate travel paths.

Receivers at E, H, and FHB were also assigned a depth zone based on the average depth covered by the receiver (<10, 10–15, 15–20, or >20 m) to calculate selection (w_i) and overlap in the use of particular depths between YOY and juveniles in these locations. Circular statistics were also used to determine the diel use of each depth zone.

Long-distance movement patterns

Additional acoustic detection data from the MARIA and CAPE BARREN curtain, and the recapture of an acoustically tagged individual was also mapped to describe the long-distance movement patterns of G. galeus once leaving the shark refuge area.

Results

Of the 40 sharks that were tagged, one YOY went undetected and was excluded from the analysis. Therefore, a total of 31 YOY and 8 juvenile sharks were monitored for seasonal occurrence in the SRA over the duration of the study. On average individual YOY and juveniles were detected for 91 (SE ± 16) and 93 (SE ± 32) d, representing 19 and 21% of their time at liberty, respectively (Table 1).

Seasonal residency and use of SRA

Overall, 19 (62%) YOY remained within the SRA for the duration of the study, three (9%) departed the SRA and were not detected again in the SRA, and nine (29%) periodically departed and returned to the SRA between May and September 2012. The general trend was for YOY (and the single juvenile) tagged in UPW to spend the summer (December—February) in UPW and progressively migrate to LPW in autumn (March—May) and, then into FHB before either leaving the SRA (i.e. past the H-curtain) in late autumn (May—June), remaining within FHB, or leaving then returning to the SRA at a later date (Figures 2 and 3). This latter group was absent from the SRA for most of winter (July—September), and not detected in FHB again until mid-spring (October) (Figures 2 and 3).

On average, YOY were present in the SRA for 392 d representing 80% (range: 22–100% d) of their time at liberty (Table 1). Eighteen YOY (58%) remained in PW, whereas the others (42%) moved out of the estuary towards the end of May 2012 (Figure 2). Of the YOY that remained in PW, one was exclusively detected moving throughout UPW for the entire study duration and 16 were detected moving for up to 47 d (average = 15 d) after being released but then went undetected thereafter. Of the 14 YOY that departed PW, only two returned to UPW, one remaining there for the rest of the study (Figure 2). All but one of these 14 YOY departed the SRA around late autumn 2012 after moving from UPW, of which nine returned to FHB around spring 2012, one remaining there for the rest of the study, the others departing the SRA again soon after or in winter the following year (Figure 2).

Figure 2.

Abacus plot showing the detection dates for YOY and juvenile (J) G. galeus tagged in UPW and FHB between January 2012 and May 2013. Each line represents an individual shark.

All tagged juveniles departed the SRA (Figure 2). On average juveniles were present in the SRA for 81 d representing 18% (range: 3–100% present) of their time at liberty (Table 1). The juvenile tagged in UPW moved to FHB in early autumn (i.e. March 2012), passed the H-curtain in late autumn (i.e. May 2012) and was not detected again (Figures 2 and 3). None of the juveniles tagged in FHB entered PW and all remained in FHB up until April 2012 (autumn), after which they all moved out of the SRA with only one individual returning in spring before departing again in late autumn the following year (i.e. May 2013) (Figure 2).

Fine-scale movements and spatial overlap within the SRA

Overall, YOY tagged in UPW were detected by UPW receivers for 88% of their time at liberty and were detected at most receivers in LPW and FHB upon leaving UPW (Figure 3). Given juveniles were not detected in PW spatial overlap with YOY was only compared in FHB. Overall, there was a significant overlap in the use of receivers in FHB between YOY and juveniles (O = 0.8, p < 0.01); however, YOY tended to utilize a larger proportion of FHB (KUD₉₅ = 132.76 km²), compared with juveniles (KUD₉₅ = 90.95 km²) (Figure 4). Selectivity analysis revealed a strong preference for YOY to remain near the entrance to PW and select receivers FHB1, FHB4-5, E and H once leaving UPW (χ² = 75712.7, d.f. = 205, p < 0.01) (Figure 4). In contrast, juveniles selected receivers in the middle of FHB (FHB4, 5, 7, 12, 17) (χ² = 12583.1, d.f. = 93, p < 0.01) and to a lesser extent receivers near shore (FHB1, 4, and 5) (Figures 4 and 5). YOY displayed more movements between receivers than juveniles and appeared to move mostly between receivers closer to land in contrast to juveniles which appeared to occupy and move mostly between receivers located in the middle of FHB (Figure 5).

Figure 3.

Circular plots showing the number of days YOY and juvenile G. galeus spent in each area for each month. Plots represent 1 year of data from January 2012 until December 2012. Note the number of days individuals were detected varies on the axis between YOY and juveniles, and between areas. Different greyscale colours represent the maximum number of individuals detected in a given day for that month.

Figure 4.

Spatial use of the SRA, excluding UPW, by YOY G. galeus tagged in UPW (a) and juveniles in FHB (b). Size of circles indicate the percentage of total detections at each receiver. Dashed and solid line represents the overall 50 and 95% KUD, respectively.

Figure 5.

Movement patterns of YOY (a) and juvenile (b) G. galeus showing minimum transition paths between receivers. Thickness of line represents the number of times the transition was made. Circles represent position of acoustic receiver or midpoint of pooled receivers in UPW, LPW, E and H.

Receiver preference was also reflected in depth use with YOY preferring shallower areas (<10 m) in FHB (χ² = 11953, d.f. = 37, p < 0.01) (Figure 6). YOY used largely the same depths over a 24-h period; however, they had a slight preference for deeper waters >20 m at night (χ² = 5267.8, d.f. = 35, p < 0.01). In comparison, juveniles showed a strong preference for depths 10–15 m during the day (χ² = 2591.2, d.f. = 27, p < 0.01) and the shallower areas <10 m during the night (χ² = 3152.95, d.f. = 28, p < 0.01).

Figure 6.

Circular plots showing the number of YOY and juvenile G. galeus detected at each hour of the day (% in total) at different depth contours in FHB. Different shading represents the number of sharks detected in a given hour.

Long-range movements

Five YOY were detected at the MARIA curtain (receiver depth: 88–113 m), which represents a minimum travel distance of 155 km, 10–358 d (mean = 154 ± 70 d) after leaving the SRA (Figures 2 and 7). One YOY (Tag ID# 31160) was detected by the CAPE BARREN curtain (bottom depth: 105 m), representing a minimum travel distance of 280 km, 348 d after leaving the SRA and was then detected 9 d later at the MARIA curtain (Figures 2 and 7). This animal was not detected at the MARIA curtain before the initial detection on the CAPE BARREN curtain. In addition, one acoustic tagged juvenile was recaptured in waters near Robe, South Australia 383 d after leaving the SRA representing a minimum distance of 1200 km (Figures 2 and 7).

Figure 7.

Long-distance movements and presumed travel paths of G. galeus detected at the MARIA and CAPE BARREN curtains, and recapture location of an acoustically tagged juvenile.

Discussion

Residency in the SRA

This study has confirmed that YOY and juvenile G. galeus seasonally use SRAs in southeastern Tasmania. However, evidence of YOY and juveniles emigrating from these areas within their first 1–2 years and the fact that few YOY (33%) and none of the juveniles returned suggests these areas may only afford protection for a limited amount of time, much <3–4 years that was estimated from previous mark-recapture results (Olsen, 1954; Stevens and West, 1997). Despite the high percentage of YOY that appeared to remain in the SRA (62%) for the duration of the study (80% of days spent in SRA), it must be emphasized to note that most of these animals (15 of 19) were detected for only a small percentage of their time at liberty (<5% of days) upon being released in UPW, and this needs to beconsidered in evaluating neonatal residency in the SRA. While perhaps these YOY were still present in the SRA but residing in areas of low receiver coverage, which seems unlikely given the number of receivers in UPW (n = 17) covering a relatively small area (∼20 km²) and the positioning of several receivers in areas forming gates through which animals would need to pass, it is more likely these animals may have died outside receiver coverage. Given natural mortality rates tend to be comparatively high in juvenile sharks (Bush and Holland, 2002; Heupel and Simpfendorfer, 2002) then removing the likely deceased YOY from the analysis would imply that the relative proportion of surviving YOY leaving the protection of the SRA in their first year could be as high as 75%. However, further research may be required to validate these residency behaviors given that tag-induced mortality in large pelagic fish including sharks can also be comparatively high (Skomal, 2007) and may therefore have been responsible for mortality in this study.

Similarly, all juveniles that emigrated from the SRA within 12 months of being tagged were present in the SRA for only 18% of their time at liberty and only one individual returned to the SRA the following spring after leaving for winter. Considering these juveniles were most likely only >1+ based on size at the time of emigration (Stevens and West, 1997), they are spending considerable time in other areas outside the protection of the SRA at a young age. In fact, three YOY were detected by acoustic receivers up to 280 km from the SRA and one juvenile was re-captured 1200 km away demonstrating that animals <2+ years are moving considerable distances outside the SRA during their early life development.

The northward movements of some YOY and juveniles are also consistent with previous mark-recapture studies demonstrating that juveniles typically moved from southern Tasmania nurseries to areas of Bass Strait and South Australia (Stevens and West, 1997). However, in contrast to these earlier findings, our study indicates that YOY are migrating much greater distances in their first few months of life than previously shown and could explain why there are fewer YOY returning to their natal areas. Whether these individuals are migrating to other known nursery areas around Tasmania such as Georges Bay (42° 19′0S 148° 14′0E) or utilizing neighbouring waters of Storm Bay (Stevens and West, 1997) remains unclear. Our results would suggest that a proportion of YOY are moving long distances to locations such as Cape Barren Island (Bass Strait), which coincides with Bass Strait being a traditional hotspot for juvenile abundance (Olsen, 1954; Stevens and West, 1997; Walker et al., 1999).

The movement of young G. galeus from the SRA within their first 1–2 years and the ability to migrate long distances suggest that once individuals find suitable habitats elsewhere there is little need to return to their natal origins. Similar observations have been made in other coastal elasmobranchs such as grey smooth-hound sharks Mustelus californicus, leopard sharks Triakis semifasciata (Carlisle and Starr, 2009), and blacktip sharks Carcharhinus limbatus (Heupel and Simpfendorfer, 2005b). These studies reported that most young sharks spent their first 12 months in estuarine and inshore nursery areas before moving into adjacent coastal waters, with only a small portion of the population exhibiting philopatry to their natal origins. Therefore, determining where and how these additional areas are being used will be essential to ongoing recovery efforts for G. galeus as many of these areas remain un-protected from exploitation activities.

Fine-scale movements and spatial overlap within the SRA

YOY typically resided in the shallow estuary of PW for most of the summer before emigrating to FHB during autumn. After entering FHB only two individuals returned to UPW, while none of the juveniles tagged in FHB were detected in UPW. This is somewhat contrary to previous work using conventional fisheries tags, which suggests that most YOY G. galeus are philopatric and return to their former estuarine nurseries (i.e. UPW) in the following spring (Olsen, 1954). Our data showed that having left UPW most individuals either then left the SRA entirely or only returned to FHB. Delayed tag-induced mortality seems unlikely given that some YOY returned to FHB or were detected elsewhere outside the SRA at a later date. One possible explanation why YOY did not return to UPW is that there may have been some behavioural changes in the utilization of this area by G. galeus during later life stages given the habitat degradation that has occurred throughout the estuary such as the loss of seagrass meadows (Rees, 1993). Philopatry to nursery grounds is common among sharks yet habitat degradation is thought to have been responsible for the demise of this behaviour in many shark species (Hueter et al., 2005). Continuing to monitor UPW and the remaining SRA with acoustic receivers may therefore be useful in determining if philopatric behaviour is still occurring but at a later stage in G. galeus life history.

YOY and juveniles showed a distinct ontogenetic disparity in their use of habitats in the SRA. YOY upon leaving UPW spent time in LPW before moving into and widely dispersing throughout mostly the shallower margins of FHB. In contrast juveniles were typically associated with habitats towards the middle of FHB. Ontogenetic resource partitioning by way of feeding on different prey or occupying different habitats is a common strategy among chondrichthyans that occupy similar spatial areas (Bethea et al., 2004; Papastamatiou et al., 2006; Taylor and Bennett, 2008; Grubbs, 2010). Crustaceans and cephalopods are important prey for YOY G. galeus, whereas teleosts become increasingly important in their diet with age (Stevens and West, 1997), suggesting that individuals may be selecting habitats based on the presence of their preferred prey. However, without data on the abundance and distribution of these prey types in the area it is not possible to ascertain whether this represents a key separation.

Juvenile use of deeper parts of FHB, particularly during the day, and expansion of their range and use of shallower areas at night may also represent a strategy to avoid predation. Juvenile G. galeus and gummy shark (Mustelus antarcticus) also preferred deeper waters of nearby NB (Figure 1) (Barnett and Semmens 2012). Using the growing literature on behavioural responses of prey to the threat of predation and relevant theory as a guiding framework for interpretation (e.g. Wirsing and Ripple, 2011), Barnett and Semmens (2012) suggested that deeper water may enable a greater escape probability from, or allow easier detection of, the dominant predator in the area, the broadnose sevengill shark (Notorynchus cepedianus) in a relatively featureless environment (i.e. there is a lack of complex habitats in which to hide). Similarly, FHB is also a fairly featureless environment and suggests the use of deeper waters may be a common tactic used by juvenile G. galeus in these inshore waters to avoid predation during the day. Conversely, juveniles may then use the cover of darkness to move into shallower, potentially riskier foraging areas, at night (Barnett and Semmens, 2012). In contrast, YOY were less selective in their habitat choice during the day as evidenced by their exploratory behaviour and broad use of FHB, and may represent more naive behavior as they have not yet learned to move into deeper areas where they are potentially less vulnerable to predation. Previous studies on lemon sharks (Negaprion brevirostris) have shown that naive individuals become more efficient at foraging as they mature as they learn to feed when prey is easily targeted and predation risks are lower (Guttridge et al., 2009; Guttridge et al., 2013). In addition, N. cepedianus are rarely found in UPW compared with FHB and NB (Stevens and West, 1997; Barnett et al., 2011; Barnett and Semmens, 2012); therefore, the higher fidelity and residency times that YOY spend in UPW may be an innate mechanism to avoid predation from N. cepedianus. Conversely, juveniles can trade-off the risk of predation in FHB for potentially increased resources (food, etc.) because they are more adept at avoiding predators.

Implications for conservation and management

This study has demonstrated that southeast Tasmania SRAs continue to represent important nursery habitats for G. galeus; however, they may only play a temporary role in their overall conservation given YOY and juveniles spend a considerable time outside of the SRAs. In addition to the protection afforded by SRAs during the early life-history phase, these regions also have a role in protecting pregnant females as they move in to the inshore areas to pup (Olsen, 1954; Walker, 1999). However, the protection of nursery areas such as with the implementation of SRAs alone are not sufficient to ensure rebuilding and sustainability of the populations (Kinney and Simpfendorfer, 2009). Of great importance are fisheries management measures such as minimum size limits that protect prerecruits from fishing pressure when outside the protection of SRAs and total catch limits. In essence, a combination of SRAs and fisheries management has likely been key to stabilization in G. galeus stocks providing a good example of how overfished populations can be stabilized using multiple management strategies. Therefore, maintaining the function of SRAs and fisheries management measures will be essential to recovery of this species.

The strong affiliation of YOY with UPW and shallow areas of FHB in the SRA reinforces that continued protection of these areas is important. Although current management measures prevent the take of sharks in the SRA, recreational fishing practices such as gillnetting are still permitted in shallow waters of FHB (i.e. out to 200 m beyond the low tide mark). Given that YOY and to a lesser extent juveniles utilize a large proportion of these shallow habitats and the fact that gillnetting may cause significant incidental mortality of young G. galeus (Williams and Schaap, 1992; Lyle et al., 2014), re-assessing the use of gillnets in these areas is warranted.

Fishery closed areas (in the absence of take) can provide an opportunity to monitor the recovery of overexploited species by acting as control sites (Gell and Roberts, 2003; Hilborn et al., 2004). Given there are currently no fishery-independent surveys or other appropriate means of monitoring G. galeus stock sizes as the fishery is now managed as incidental bycatch and fishers no longer target G. galeus (Huveneers et al., 2013), examining the long-term use of closed areas by acoustically tagged animals in this study may therefore be helpful in monitoring the recovery of the G. galeus population. For example, acoustic monitoring of YOY G. galeus may provide estimates of natural mortality (Heupel and Simpfendorfer, 2002) which could be used to refine current recruitment and stock assessment modelling. Expansion of the acoustic array to cover areas outside of SRAs may also provide further insight to the importance of un-protected areas to G. galeus during early life history, providing critical empirical evidence needed to refine and enhance current management and conservation strategies such as closed area boundaries.

Acknowledgements

We thank E. Forbes, D. Jones, and R. French for their dedicated assistance with fieldwork particularly the long, cold days diving to maintain the acoustic receiver array. The Australian Animal Tagging and Monitoring System (AATAMS), in collaboration with the Ocean Tracking Network (OTN), Institute for Marine and Antarctic Studies (IMAS), and Commonwealth Scientific and Industrial Research Organisation (CSIRO) Marine and Atmospheric Research, deploy and maintain the Maria and Flinders Island arrays. AATAMS also provided the data from the detections of our sharks on these arrays. This study was supported by grants from the Winifred Violet Scott Foundation to JMS and Save Our Seas Foundation to J.M.S. and A.B. and the Holsworth Wildlife Research Endowment to J.D.M.

Author notes

Handling editor: Howard Browman