Habitat Preferences and Site Fidelity of the Ornate Wobbegong Shark (Orectolobus Ornatus) on Rocky Reefs of New South Wales1

Home , Orectolobus, Ornate wobbegong

Habitat Preferences and Site Fidelity of the Ornate Wobbegong Shark (Orectolobus ornatus) on Rocky Reefs of New South Wales1

Robert Carraro2 and William Gladstone2,3

Abstract: Habitat and microhabitat preferences and site fidelity of Orectolobus ornatus were assessed between September 2002 and August 2003 to assess potential suitability of marine reserves for its conservation. Of six rocky reef habitats available in the study area (sponge gardens, artificial structures, barren boulders, sand, sea grass, macroalgae), O. ornatus exhibited a significant preference for sponge gardens, artificial structures, and barren boulders habitats. Habitat preferences of males and females, and individuals <1 m and >1 m, did not differ. Orectolobus ornatus selected daytime resting positions with a high topographic complexity and crevice volume and did not select on the basis of prey availability. Habitat and microhabitat preferences may be related to the need for predator avoidance. Regular monitoring of 40 individually identified O. ornatus revealed that none was a permanent resident of the study area. Seven individuals exhibited short-term temporary fidelity to the study area; they were resighted frequently for part of an intensive 100-day survey. Remaining individuals were temporary visitors; they were resighted at most once after initial identification or returning after extended absences. Monthly population surveys confirmed the turnover of O. ornatus in the study area. The lack of long-term site fidelity suggests that small marine reserves will be ineffective as a conservation strategy for O. ornatus.

A primary requirement for conserving ma- processes such as competition, predation, rine species is an understanding of the pro- and recruitment ( Jones and Syms 1998). The cesses underlying patterns in distribution and importance of habitat has received increas- abundance (Bell et al. 1991). One such factor ing attention from conservation biologists that has received considerable attention is because anthropogenic degradation of habi- habitat variation. The patchy distribution of tat is a primary cause of global declines in habitats inﬂuences distribution and abun- biodiversity and as spatial approaches to con- dance at many spatial scales (Syms 1995). servation and management (such as marine Habitat structure or quality of habitats may protected areas) are increasingly utilized limit populations through the availability of (Ray and McCormick-Ray 2004). Under- critical resources and may modify ecological standing the importance of habitat for species requires the separation of habitat usage and 1 This research was funded by grants from the Uni- preference, where habitat usage includes the versity of Newcastle and is a contribution from the habitats in which individuals occur and habi- Centre for Sustainable Use of Coasts and Catchments. tat preference is a species’ use of a habitat in Manuscript accepted 2 June 2005. relation to its relative availability (Manly et al. 2 Centre for Sustainable Use of Coasts and Catch- 1993). ments, School of Applied Sciences, University of New- castle (Ourimbah Campus), P.O. Box 127, Ourimbah, Although elasmobranch nursery habitats New South Wales 2258, Australia. have been well studied, there are few studies 3 Corresponding author (e-mail: william.gladstone@ of usage and preference for other life stages newcastle.edu.au). and habitats, despite the critical importance of habitat in determining distribution and Paciﬁc Science (2006), vol. 60, no. 2:207–223 abundance and the application of this in- : 2006 by University of Hawai‘i Press formation for conservation planning (Simp- All rights reserved fendorfer and Heupel 2004). The available

207 208 PACIFIC SCIENCE . April 2006 quantitative studies of habitat preference between 1990 and 2000 raised concerns about demonstrate fine-scale discrimination among the status of wobbegong populations (NSW available habitats. Juvenile lemon sharks Fisheries 2002a). Other sources of anthropo- (Negaprion brevirostris) prefer habitats less genic mortality include recreational fishing than 50 cm depth, with water temperature and protective beach meshing programs (Po- greater than 29 C, and consisting of a mix- gonoski et al. 2002). Due to recent declines ture of rock and sand (Morrisey and Gruber in populations, O. ornatus has been classified 1993a). Tiger sharks (Galeocerdo cuvier)in as vulnerable in New South Wales and near- Shark Bay, Western Australia, prefer shallow threatened throughout its range (Cavanagh sea-grass beds, where their prey is more et al. 2003). Marine reserves have been advo- abundant (Heithaus et al. 2002). With the ex- cated as a strategy to conserve wobbegong ception of general habitat descriptions little sharks (Otway and Parker 2000, NSW Fish- data have been obtained for the habitat usage eries 2002b). However, the ecological under- of wobbegong sharks. There is, therefore, standing about wobbegongs necessary for considerable scope for studies of habitat evaluating the potential usefulness of marine preference in this and other shark species of reserves is lacking (Pogonoski et al. 2002). conservation significance whose habitat is af- This lack of information is a critical gap given fected by human activities. the conservation and management concerns The related concept of site fidelity ex- for this species. The aims of this study were presses the temporal attachment of an indi- to determine (1) habitat preferences and (2) vidual to a space in its habitat. The degree site fidelity of O. ornatus and to use this infor- of site fidelity varies in relation to the avail- mation to assess the potential usefulness of ability and defensibility of critical resources marine reserves as a conservation and man- and ranges between strict territoriality, shared agement strategy. home ranges, and nomadism, although territoriality has not been demonstrated in sharks (Heithaus 2004). Site fidelity is known to oc- materials and methods cur at some stage in the life cycle of a number Study Species of sharks (McKibben and Nelson 1986, Hol- land et al. 1993, Morrissey and Gruber 1993b, Orectolobus ornatus is a nocturnal benthic Goldman and Anderson 1999, Rechisky and predator that occurs in all Australian states Wetherbee 2003). The existence of site fidel- and Papua New Guinea to a depth of at least ity will determine the likely success of spa- 100 m. During daylight hours it rests in pro- tially based management strategies such as tected parts of reefs. Orectolobus ornatus is marine protected areas (Kramer and Chap- ovoviviparous and pups are born at @20 cm man 1999). Further understanding of the ex- length, with adults growing to a maximum istence of site fidelity in sharks is important length of @300 cm. Sexual maturity is be- in assessing the potential usefulness of marine lieved to occur at 175 cm length (Last and protected areas for conservation and manage- Stevens 1994). ment. Two species of wobbegong sharks (family Study Area Orectolobidae) occur in southeastern Austra- lia: Orectolobus ornatus (De Vis, 1883) and O. This study occurred between September maculatus (Bonnaterre, 1788). Both species 2002 and August 2003 in the Fly Point– are demersal, inhabit temperate rocky reefs, Halifax Park Aquatic Reserve (hereafter and are likely to have an important functional ‘‘the reserve’’) in Port Stephens, New South significance because of their role as top pred- Wales, Australia, at 32 42 0 50 00 S, 152 90 ators (Last and Stevens 1994). Wobbegongs 20 00 E (Figure 1). Habitats within the reserve are caught by commercial fishers in New were typical of those present in adjacent South Wales as a target group and as by- nearshore areas (Underwood et al. 1991) and catch. Declines in annual catches of >60% included fringe (dominated by foliose macro- Figure 1. Location of study areas (a) within Port Stephens and (b) within the Fly Point–Halifax Park Aquatic Reserve. The boundary of the reserve is shown within the waters of Port Stephens and the land is shaded gray. 210 PACIFIC SCIENCE . April 2006 algae in depths of 1–8 m), barren boulders Habitat preferences were determined from (with high coverage of crustose coralline al- resource selection ratios (Manly et al. 1993) gae and high abundance of the sea urchin using the formula wî ¼ oi/pI , where oi is Centrostephanus rodgersii in depths of 2–12 the proportion of habitats used that are in m), sponge gardens (high coverage of en- category i, pI is the proportional availability crusting and foliose sponges and limited cov- of habitat i, and wî is the preference score erage of algae in depths of 10–20 m), sand, for habitat category i. The standard error sea grass (Posidonia australis, Halophila ovalis, (SE)pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi of wî was calculated by SEðwîÞ¼ and Zostera capricorni), and artificial structures wî f1/ui 1/uþ þ 1/mi 1/mþg, where ui is (small sunken boats, refuse, and artificial reefs the number of sharks in habitat i, uþ is the constructed of metal frames in depths of 6– total number of observations of sharks, mi 12 m). Observations reported here were un- is the frequency of occurrence of habitat i, dertaken at three sites in the reserve: Fly and mþ is the total occurrence of all habitats. Point, Little Beach, and Halifax Park. Fixed wî 95% confidence intervals were calculated G transects of 500-m length by 5-m width that from wî Za/ðIÞSEðwîÞ, where Za/ðIÞ is the covered all habitats and the depth range of 100a/ðIÞ percentage point of the standard each site were established at the beginning normal distribution and I is the number of of the study. The boundaries of the transects habitat groups. A Bonferroni correction was were noted in relation to prominent substra- applied by dividing the a significance level tum features. (0.05) by the number of habitat groups to allow for multiple comparisons between all Habitat Availability and Preferences habitat categories. Confidence intervals were used to determine the significance of prefer- The relative availability of each habitat in the ence scores. When the upper confidence in- reserve was determined by divers using the terval was <1, the habitat was significantly point transect method (Choat and Bellwood avoided. When the confidence interval fell 1985). Three 150-m line transects were laid between <1 and >1, the habitat was used in across the reef and perpendicular to the shore proportion to its availability (i.e., no prefer- in each of the three sites. A length of 150 m ence or avoidance was exhibited). When the was used because this was the distance from lower confidence interval was >1, the habitat the shore at which the rocky reef ended on was significantly preferred. sand (at depths of 18–25 m). At 20-m inter- Production of a detailed underwater map vals along the transect the habitat below the of the Halifax Park site provided depth data tape measure was recorded, giving a total of that was used, in the same manner as the cal- 63 point counts of habitat type in the reserve. culation of habitat preferences, to determine Underwater visual surveys of the fixed depth preferences of O. ornatus observed at transects were used to locate O. ornatus and this site. to record the habitat in which each individual occurred. All crevices and overhangs on the Microhabitat Attributes of Resting Positions transects were inspected for O. ornatus. For each observed O. ornatus the habitat in which The microhabitat attributes of each site were it occurred, its sex, and its total length were assessed from 21 quadrats (2 by 2 m) ran- noted on waterproof paper. Total length was domly positioned over the depth range and measured with a marked fiberglass pole in all habitats. A 2 by 2 m quadrat was used placed parallel to the shark’s body on the to simulate the area occupied by a resting O. substratum. Replicate surveys of each site to ornatus. Microhabitat attributes quantified in determine habitat preferences of O. ornatus each quadrat included depth (in m) at the were done in October 2002 (n ¼ 4 surveys), center of the quadrat, slope of the substratum January (n ¼ 4 surveys), April (n ¼ 2 surveys), within the quadrat (scored as none ¼ 0, 1– and June 2003 (n ¼ 3 surveys) and the results 15 ¼ 1, 15–30 ¼ 2, 30–45 ¼ 3, 45–60 ¼ combined for analysis. 4, 60–75 ¼ 5, 75–90 ¼ 6), and average vol- Habitat Preferences and Site Fidelity of Orectolobus ornatus . Carraro and Gladstone 211 ume of all crevices/holes in the quadrat Fidell 1989). One-way analysis of similarities (based on measurements of longest width, (ANOSIM) was used to test whether the mul- breadth, and depth of each hole or crevice tivariate set of microhabitat attributes of rest- in the quadrat). Substrate rugosity within the ing positions differed significantly from the quadrat was calculated from the length of a multivariate set of microhabitat attributes tape measure (in cm) required to follow the of the randomly selected positions. PCA and substrate contours from one side of the quad- ANOSIM were undertaken with PRIMER5 rat to the other side and expressed as a ratio software (Primer-E Ltd., Plymouth [Clarke to the straight-line distance (200 cm) (Mc- and Warwick 2001]). Cormick 1994). The average value from three replicate ratios was used. Orectolobus ornatus is Site Fidelity a piscivore (Last and Stevens 1994), and the possibility that O. ornatus was selecting rest- The degree of site fidelity by O. ornatus to the ing positions in response to prey availability area of the reserve was determined by moni- was tested by recording fish abundance be- toring the movements and positions of 40 fore placing the quadrat. As the observer individually identifiable sharks for 100 days approached the position where the quadrat (following the date of their identification) be- was to be placed, the abundance of all fish oc- tween September 2002 and April 2003. Orec- curring on the substratum and in the water tolobus ornatus were individually identified by column for a distance of 3 m above the sub- tagging, photography, and drawing. A dart stratum was estimated and scored in the fol- tag with a unique identification number was lowing abundance classes: no fish ¼ 0, 1–10 implanted with a handspear into the dorsal fish ¼ 1, 11–25 fish ¼ 2, 26–50 fish ¼ 3, musculature anterior to the first dorsal fin. >50 fish ¼ 4. Dart tags implanted in this position on other The microhabitat attributes of the diurnal species produced only localized tissue disrup- resting positions of O. ornatus were recorded tion (Heupel and Bennett 1997). The same for individuals observed in the October 2002 result was observed for O. ornatus and sug- surveys of habitat preferences ðn ¼ 49Þ.A gested that the health and behavior of the small weighted, subsurface marker buoy was tagged individuals were unaffected by the placed on the substratum adjacent to individ- procedure. Only sharks >1 m total length uals observed during the habitat-preference (TL) ðn ¼ 7Þ were tagged to reduce possible surveys to identify the resting position and it adverse effects from tagging smaller sharks was revisited within 24 hr to record the mi- (Kohler and Turner 2001). For sharks <1m crohabitat attributes. Sharks still present at TL photographs ðn ¼ 16Þ and drawings the resting position were gently displaced to ðn ¼ 17Þ were used to record the individually allow placement of the 2 by 2 m quadrat. unique pattern of dots and other markings on Principal Components Analysis (PCA) was the head and distinctive fin injuries on the used to determine the relative importance of dorsal or caudal fins. The use of both pat- microhabitat attributes in resting site selec- terns in external markings and fin injuries to tion. Substrate rugosity and crevice volume identify individuals, and the duration of the were log transformed and depth was square- observation period (100 days), meant that root transformed before analysis and PCA individuals identified by these means were performed using Euclidean distance as the unlikely to be mistakenly identified. In addi- distance measure following normalization. tion, all identifications of individual sharks Microhabitat attributes of random quadrats were done by the same observer throughout and shark resting positions were displayed the study. Resightings of these individually on a PCA ordination and differences between identified sharks were noted during the the relative positions of points interpreted habitat-preference surveys, monthly popula- from principal components 1 and 2. Compo- tion surveys (see later in this section), and nent loadings with an absolute value >0.50 specific surveys undertaken to search for were regarded as important (Tabachnick and these individuals. Resighting frequency was 212 PACIFIC SCIENCE . April 2006 thus determined from 12 surveys of the re- variances. A randomization procedure (Manly serve over the 100-day period, with an aver- 1997, Legendre and Legendre 1998) was used age interval of 8 days between successive to obtain a significance level for the observed surveys. F value because it is likely that the monthly The frequency distribution of resightings samples were not independent due to a de- of individually identified O. ornatus was used gree of site fidelity by the sharks (see Re- to indicate the degree of site fidelity. The to- sults). The randomization procedure works tal numbers of resightings of each individual by calculating an observed F value from the were recorded and the difference between the gathered data and then randomly assigning resighting frequency distribution and a Pois- observations to each month (while maintain- son distribution was tested by G-test (Sokal ing the same monthly sample size). To obtain and Rohlf 1995). A Poisson distribution was a frequency distribution of F values 5,000 used because no significant difference be- randomizations of the data were used, and tween it and the resighting frequency distri- the significance level of the observed F value bution would indicate that resightings were was the percentage of randomized F values randomly distributed and the majority of in- equal to or larger than it (Manly 1997). A dividuals were unlikely to be present in the chi-square contingency test (Sokal and Rohlf reserve for the duration of the observation 1995) was used to test the null hypothesis that period (Pielou 1977, Samoilys 1997). sex ratio did not vary through time, with ran- Fidelity by O. ornatus to specific resting domization also used to determine the signif- positions was tested at the Halifax Park site icance of the obtained Pearson chi-square by plotting the location of all O. ornatus ob- statistic. served during the habitat preference and monthly population surveys (see following paragraph) on a detailed underwater map of results the site. Habitat Preferences Further information on the possible existence of long-term site fidelity was gathered The relative availability of habitats in the re- from data on temporal patterns in abundance, serve was sand 34.5%, sponge gardens 20.3%, sex ratio, and length of O. ornatus. The ab- barren boulders 15.5%, fringe 19.6%, sea sence of a stable pattern in these three vari- grass 7.7%, and artificial structures 2.4%. A ables over the study period would indicate total of 283 observations of O. ornatus was that the population of O. ornatus in the study recorded during the habitat-preference sur- area did not exhibit high site fidelity. Sur- veys, and the habitats occupied included veys of the fixed transects were undertaken sand (3.0% observations), sponge gardens monthly by divers between September 2002 (37.0%), barren boulders (24.2%), fringe and August 2003 in each site. Data from each (18.6%), and artificial structures (17.2%). site were combined for analyses. Data were No O. ornatus were observed in sea-grass collected from repeated monthly surveys of a beds. Significance tests of preference scores fixed transect, and so the replicate scores were indicated that O. ornatus preferred sponge not independent. Repeated measures analysis gardens, barren boulders, and artificial struc- of variance (ANOVA) was used, with time as tures habitats; avoided sand habitat; and ex- the repeated measures factor, to test the null hibited no preference for fringe habitat hypothesis that abundance of O. ornatus did (Figure 2a). The same pattern of habitat pref- not vary through time. Huynh-Feldt’s epsilon erences was shown by males and females and was used to test the assumption of sphericity by sharks <1 m and >1 m TL (Figures 2b (Quinn and Keough 2002). A one-way and 3a). Orectolobus ornatus at the Halifax ANOVA was used to test the null hypothesis Park site preferred depths of 4–12 and 16– that mean length of sharks did not vary be- 20 m, showed no selection for depths <4m, tween months, and Levene’s test was used and avoided depths of 12–16 and >20 m to test the assumption of homogeneity of (Figure 3b). Figure 2. Habitat preferences of (a) all Orectolobus ornatus,(b) males and females. Values shown are preference scores and their 95% confidence intervals. Habitat preference scores with lower confidence limit >1 indicate a significant preference; confidence intervals that range from <1to>1 indicate no significant preference; an upper confidence limit <1 indicates significant avoidance. Figure 3. Habitat preferences of Orectolobus ornatus: (a) two classes of total length, and (b) depth preferences. Values shown are preference scores and their 95% confidence intervals. Habitat preference scores with lower confidence limit >1 indicate a significant preference; confidence intervals that range from <1to>1 indicate no significant preference; an upper confidence limit <1 indicates significant avoidance. Habitat Preferences and Site Fidelity of Orectolobus ornatus . Carraro and Gladstone 215

TABLE 1 Microhabitat Attributes of Resting Positions Summary of Principal Component Axis Loadings (Only Orectolobus ornatus used resting positions that the First Three PCs Are Shown) differed significantly in their multivariate set of microhabitat attributes from randomly se- Variable PC1 PC2 PC3 lected positions on the reef that were not Slope 0.23 0.63 0.67 used as resting positions (R ¼ 0:18, P ¼ Fish abundance 0.41 0.26 0.07 0:001). In the principal components analysis Depth 0.01 0.73 0.64 used to determine the specific attributes that Rugosity 0.63 0.03 0.13 Crevice volume 0.61 0.09 0.34 were selecting in their resting positions the Cumulative % total variation 36.7 57.7 77.6 first three PCs explained 77.6% of the variation in the raw data (Table 1, Figure 4). The Note: Components with an absolute value of >0.5 (in bold) are important (Tabachnick and Fidell 1989). PC1 axis represented an increasing gradient for two features of substrate complexity (rugosity and crevice volume), and the PC2

Figure 4. Principal components plots comparing microhabitat attributes of a random selection of positions (b) and microhabitat attributes of resting positions of Orectolobus ornatus (f). 216 PACIFIC SCIENCE . April 2006

Figure 5. Relative sighting frequency of Orectolobus ornatus at the Halifax Park site, as indicated by the size of the shaded circles. Values in bold refer to the number of sightings in a 20 by 20 m grid cell between September 2002 and August 2003, and grid cells without numbers indicate no sightings. axis represented an increasing gradient for termined by randomization because of non- depth. The topographic complexity and crev- independence of observations), indicating ice volume of resting positions were generally that the observed patterns reflect selection greater than randomly selected positions, and by O. ornatus rather than survey effort. Over the depth range of resting positions (3–21 m) time, different individuals successively used was smaller than the depth range of the ran- the same crevice. A crevice at D1 was used domly selected positions (0.5–25 m). The successively by four different sharks over 150 PC3 axis (not shown in Figure 4) represented days (October 2002 to March 2003), suggest- increasing reef slope and indicated that most ing that some resting positions were more O. ornatus resting positions occurred where suitable than others. reef slope was low. Orectolobus ornatus used some parts of Hal- Site Fidelity ifax Park for resting positions more frequently than others (Figure 5), in particular The 40 identified O. ornatus were observed grid positions D1, D6, F4–F7, B5, and C5. on 133 occasions over the 100-day observa- D1, where the highest frequency of sightings tion period. Five individuals were not re- occurred, is an area of a steep slope within sighted after their first identification, and 10 the barren boulders habitat comprising individuals were resighted only once after boulders 1–2 m in height and with numerous their first identification (Figure 6). The great- crevices and recesses. The frequency of occu- est resighting frequency was 11. The fre- pation of resting positions was unrelated to quency distribution of resightings did not the number of times each position was follow a Poisson distribution (G ¼ 37:39, surveyed (r ¼ 0:25, P ¼ 0:17, significance de- df ¼ 10, P < 0:001): the frequency of 0 and Habitat Preferences and Site Fidelity of Orectolobus ornatus . Carraro and Gladstone 217

Figure 6. Frequency distribution of resightings of individually identiﬁed Orectolobus ornatus over a 100-day period and a Poisson distribution assuming resightings are randomly distributed.

7–11 resightings was greater than expected dividuals, but the difference between the and the frequency of 2–5 resightings was less three identification methods was not signifi- than expected. The frequency distribution cant (Kruskal-Wallis test, w2 ¼ 3:63, df ¼ 2, of resightings (Figure 6) suggests that indi- P > 0:05). There was no significant differ- viduals fell into two groups: those that were ence between the resighting frequencies of resighted frequently (7–11 resightings) and males and females (Kruskal-Wallis test, w2 ¼ those resighted infrequently (0–5 resight- 0:001, df ¼ 1, P > 0:05) or of sharks <1m ings). The latter group included an individual TL and b1 m TL (Kruskal-Wallis test, w2 ¼ resighted after an absence of 88 days and 13 0:002, df ¼ 1, P > 0:05). individuals (eight males and five females of Individuals displayed fidelity to specific 0.8–1.3 m TL) with resighting intervals of resting positions. One individual at Little >20 days. Despite regular, biweekly surveys Beach was observed in the same position on of the study area only two of the identified 10 surveys of the 12 surveys conducted there individuals were resighted after the 100-day over 100 days. Another individual was ob- period: a tagged 1.25-m male was resighted served in the same resting position at Halifax on single occasions 117 days and 172 days Park on 10 consecutive surveys over 59 days. after tagging, and a tagged 1.05-m male was An individual that was absent from the re- resighted 211 days after tagging. On all occa- serve for 95 days returned to the same resting sions these individuals occupied a resting po- position at Fly Point it had occupied before sition within 10 m of the position where they its departure. Of resighted sharks 76.8% were tagged. were observed within 20 m of their previous Tagged individuals were resighted more sighting. Two sharks moved between areas: frequently (5:7 G 1:4 resightings) than photo- a 0.98-m female moved from Little Beach to graphed ð3:1 G 0:7Þ or drawn ð2:6 G 0:6Þ in- Halifax Park (a distance of 300 m); and a 218 PACIFIC SCIENCE . April 2006

Figure 7. Mean (G SE) abundance of Orectolobus ornatus in the study area. Mean values are based on the three study sites.

1.05-m male moved from Fly Point to Halifax w2 ¼ 19:08, df ¼ 11, P ¼ 0:05). There were Park (a distance of >1,000 m), where it was more males than females in the transects on seen on a further nine occasions within a most survey occasions, with the exception of @20-m radius. April and May 2003 when numbers of females exceeded the numbers of males (Figure 8). The sex of a number of sharks in each Abundance, Sex Ratio, and Length survey could not be determined because of Abundance of O. ornatus increased between an obstructed view; however, their numbers September 2002 and January 2003, remained were low relative to the numbers of con- relatively stable until April 2003, and then de- firmed males and females. clined until the last survey in August 2003 The TL of O. ornatus observed during the (Figure 7). Average shark abundance in Janu- monthly surveys ranged from 0.5 to 1.3 m ary 2003 was double the abundance in Sep- (mean G SE ¼ 0:98 G 0:01 m, n ¼ 215). tember 2002 and three times that in August Three smaller sharks (0.35, 0.23, and 0.40 m) 2003. At the time of peak abundance 13 were observed on other occasions. Mean TL sharks were recorded in the transect at Hali- varied significantly between monthly surveys fax Park, seven in the transect at Fly Point, (F ¼ 3:46; df ¼ 11;203; P < 0:001; Levene’s and six in the transect at Little Beach. At the statistic ¼ 0:70, P ¼ 0:65) (Figure 9). Mean last survey in August 2003 there were two TL decreased from 1:1 G 0:04 m in Novem- sharks at Halifax Park, three at Fly Point, ber to 0:9 G 0:04 m in January and then grad- and three at Little Beach. Overall, the abun- ually increased over the remainder of the dance of O. ornatus varied significantly study period. The smaller TL in December through time (Huynh-Feldt epsilon ¼ 0:99; 2002 and January 2003 was due to an in- F ¼ 2:61; df ¼ 11;22; P < 0:05). creased abundance of smaller females and a Sex ratio varied through time (Pearson reduced abundance of larger males. Habitat Preferences and Site Fidelity of Orectolobus ornatus . Carraro and Gladstone 219

Figure 8. Numbers of males, females, and individuals of unknown sex of Orectolobus ornatus observed between September 2002 and August 2003.

discussion volume. It therefore appears that O. ornatus preferred sponge gardens, barren boulders, The preferred habitats of O. ornatus were and artificial structures because of the struc- sponge gardens, barren boulders, and artifi- tural complexity they provided. cial structures. Although these habitats The depth preferences of O. ornatus at differed in biogenic composition they over- Halifax Park correspond with areas of greater lapped in depth distribution and were the structural complexity. The 4- to 12-m range most topographically complex habitats. All includes the steep slope and large boulders habitat types occurred at the Little Beach of grid reference D1 and the 12-m ridge site, but almost all O. ornatus there occupied (D6); the 16- to 20-m contour contains the artificial structures. This probably occurred twin bommies (patch reefs) (F4 and F5) and because the sponge gardens and barren other smaller features (Figure 5). These boulders occurred on gently sloping shelves depth ranges contain a number of crevices of and, unlike the other sites in the study area, potential resting positions and were strongly provided no shelter or protection. Artificial utilized by O. ornatus. By comparison the structures are used by fishes for shelter and depth ranges avoided or not selected consist foraging (Hair and Bell 1992, Carr and of shallow water (0–4 m), broad sponge pla- Hixon 1997), but their importance for elas- teaus (12–16 m), and flat areas of sand with mobranchs is less well known. Multivariate few structural elements. analysis of the biophysical attributes of rest- Orectolobus ornatus may have selected struc- ing positions found that the most consistent turally complex habitats and resting positions features distinguishing resting positions from to avoid predators. Predators such as the tiger a random selection of positions were high shark (Galeocerdo cuvier), bull shark (Carchar- topographic complexity and greater crevice hinus leucas), and the great white shark (Car- 220 PACIFIC SCIENCE . April 2006

Figure 9. Mean total length (G SE) of Orectolobus ornatus in the study area between September 2002 and August 2003.

charodon carcharius) are occasionally seen of higher predation rates in other habitats. within Port Stephens, and large numbers of An alternative explanation for the observed sharks were fished from the area in the early habitat and microhabitat preferences of O. or- twentieth century (Roughley 1955), suggest- natus is that existence of strong tidal currents ing that predation by larger sharks may be a (three to four times daily) within the study risk for O. ornatus in the study area. All O. or- area may encourage the use of crevices for natus observed were in the size range 50 to shelter. However, O. ornatus has also been 140 cm TL (including two pregnant females observed offshore residing in crevices where of 120 cm and 140 cm TL), and there was no current was present (pers. obs.). no difference in the habitat preferences of in- The absence of a relationship between the dividuals <1 m and >1 m. This indicates that small-scale distribution and abundance of O. predation risk is similar across the size range ornatus and fish abundance is surprising, given observed in the study. Reduction of predation that O. ornatus is piscivorous (Last and Ste- risk has been hypothesized to be responsible vens 1994). However, predatory attacks upon for habitat preferences in the nursery grounds fishes from a position within a topographi- of lemon sharks, Negaprion brevirostris (Mor- cally complex shelter site may be physically risey and Gruber 1993a); blacktip sharks, difficult. Limited nocturnal observations in Carcharhinus limbatus (Heupel and Heuter the study area of predation by O. ornatus on 2002); and sandbar sharks, Carcharhinus plum- fishes showed that this occurred away from beus (Rechisky and Wetherbee 2003). Dem- shelter and as ambush predation. Many shark onstration of the value of preferred habitats species feed infrequently and follow this with for predator avoidance requires experimental long periods of inactivity (Heithaus 2004). evaluation, predator surveys, or observations Nocturnal feeding, the use of an ambush Habitat Preferences and Site Fidelity of Orectolobus ornatus . Carraro and Gladstone 221 feeding strategy, and infrequent feeding were sighted within 5–15 m of their previous would decouple the relationship with diurnal resting locations. The resighting of some in- fish abundance. Quantitative studies of pred- dividuals following an absence suggests that ator distribution and abundance and prey the reefs in the reserve may be a part of a density in elasmobranches are uncommon larger home range. The possibility that indi- (Heithaus 2004), and such studies have shown vidual sharks did not leave the reserve but that prey availability was influential in the remained hidden and undetected during the habitat preferences of tiger sharks (Heithaus intensive surveys is unlikely. The monthly et al. 2002) but was not correlated with the population surveys confirmed significant habitat preferences of blacktip sharks (Heu- changes over time in the abundance, sex ratio, pel and Heuter 2002). and length of O. ornatus in the study area, This study used direct observations by indicating a change in composition of the divers to determine the daytime habitat and population. The study area was surveyed microhabitat preferences of O. ornatus. This frequently using an intensive searching tech- was feasible because underwater visibility is nique for tagged and recognizable individuals. 7–15 m throughout the year, large sections The reserve is also very popular with recre- of the study area are able to be searched by ational divers and at the start of the tagging divers in a single day (thus avoiding temporal period an awareness program was mounted confounding of observations of individual with local dive shops and at the entry points sharks), and all of the study area is accessible to the dive sites advising divers about the from the shore. Acoustic telemetry may have presence of tagged sharks in the reserve and been impractical for daytime monitoring of requesting information about sightings. The O. ornatus because of its habit of sheltering sightings of tagged sharks reported by divers in physically complex habitats that would coincided with our observations at the same have impeded acoustic signals. Acoustic te- time. lemetry at night will provide information on Changes in abundance, length, and sex nocturnal habitat preferences when O. ornatus ratio and the relative lack of movement be- is foraging in the open; however, of the 37 O. tween sites within the study area suggest that ornatus observed at night 14 were in the same the population of O. ornatus that utilized the shelter sites occupied during the day. Further study area was dynamic and reflected move- research is required to study movement pat- ments by individuals at a much larger spatial terns at a larger scale and to identify home scale. Individual O. ornatus were temporary ranges or migratory patterns within and be- residents of the study area, and some individ- yond Port Stephens. uals returned for short periods of time after None of the individually identified O. or- long absences. Although advocated for a natus was a permanent resident of the study range of conservation and management issues area over the time scale of the intensive sur- (Halpern and Warner 2002, Gladstone et al. veys (100 days). Some individuals exhibited 2003), marine reserves will only be effective short-term temporary fidelity to the study for species’ conservation when they are lo- area, being resighted frequently for part of cated and designed in accordance with life the intensive surveys. Other individuals were history requirements and habitat preferences. temporary visitors, being resighted only on Reserves suitable for O. ornatus conservation 0–1 occasions after initial identification or will need to include their preferred habitats returning after extended absences. Only two (barren boulders, sponge, artificial structures) male individuals were resighted in the study with a high structural complexity. However, area within 6 months of the end of the inten- the lack of long-term site fidelity suggests sive surveys, indicating that some temporary that reserves may not be effective when used visitors use the reserve regularly but only for as the sole management strategy. Further re- short periods of time. Subsequent sightings search is needed to determine the nature of of the same two males occurred 12 and 24 O. ornatus site fidelity at much larger spatial months after completion of the study. Both scales. 222 PACIFIC SCIENCE . April 2006

acknowledgments Hair, C. A., and J. D. Bell. 1992. Effects of enhancing pontoons on abundance of fish: The design of this study benefited from Initial experiments in estuaries. Bull. Mar. initial discussions with N. Otway and D. Har- Sci. 51:30–36. asti. D. Pollard, M. Lincoln Smith, D. Har- Halpern, B. S., and R. R. Warner. 2002. Ma- asti, D. Powter, J. Morton, and N. Otway rine reserves have rapid and long lasting and two anonymous reviewers provided help- effects. Ecol. Lett. 5:361–366. ful comments on early versions of the manu- Heithaus, M. R. 2004. Predator-prey interac- script. Thanks to G. Lewis and N. McCarthy tions. Pages 487–521 in J. C. Carrier, J. A. for assistance with diving and fieldwork and Musick, and M. R. Heithaus, eds. Biology the divers of Nelson Bay for providing infor- of sharks and their relatives. CRC Press, mation. Tagging was done under permits Boca Raton. P01/0064 from NSW Fisheries and ACEC Heithaus, M. R., L. M. Dill, G. J. Marshall, 8100803 from the University of Newcastle’s and B. Buhleier. 2002. Habitat use and Animal Care and Ethics Committee. behavior of tiger sharks (Galeocerdo cuvier) in a seagrass ecosystem. Mar. Biol. (Berl.) Literature Cited 140:237–249. Heupel, M. R., and M. B. Bennett. 1997. His- Bell, S. S., E. D. McCoy, and H. R. Mushin- tology of dart insertion sites in the epau- sky. 1991. Habitat structure: The physical lette shark. J. Fish Biol. 50:1034–1041. arrangement of objects in space. Chapman Heupel, M. R., and R. E. Hueter. 2002. The and Hall, London. importance of prey density in relation to Carr, M. H., and M. A. Hixon. 1997. Artifi- the movement patterns of juvenile sharks cial reefs: The importance of comparisons within a coastal nursery area. Mar. Fresh- with natural reefs. Fisheries (Bethesda) water Res. 53:543–550. 22:28–33. Holland, K. M., B. M. Wetherbee, J. D. Cavanagh, R. D., P. M. Kyne, S. L. Fowler, Peterson, and C. G. Lowe. 1993. Move- J. A. Musick, and M. B. Bennett. 2003. ments and distribution of hammerhead The conservation status of Australian shark pups on their natal grounds. Copeia chondrichthyans. Report of the IUCN 1993:495–502. Shark Specialist Group Australia and Oce- Jones, G. P., and C. Syms. 1998. Disturbance, ania Regional Red List workshop, Univer- habitat structure and the ecology of fishes sity of Queensland, Brisbane. on coral reefs. Aust. J. Ecol. 23:287–297. Choat, J. H., and D. R. Bellwood. 1985. In- Kohler, N. E., and P. A. Turner. 2001. Shark teractions amongst herbivorous fishes on a tagging: A review of conventional methods coral reef: Influence of spatial variation. and studies. Environ. Biol. Fishes 60:191– Mar. Biol. (Berl.) 89:221–234. 223. Clarke, K. R., and R. M. Warwick. 2001. Kramer, D. L., and M. R. Chapman. 1999. Change in marine communities: An Implications of fish home range size and approach to statistical analysis and inter- relocation for marine reserve function. En- pretation, 2nd ed. PRIMER-E, Plymouth. viron. Biol. Fishes 55:67–79. Gladstone, W., F. Krupp, and M. Younis. Last, P. R., and J. D. Stevens. 1994. Sharks 2003. Development and management of a and rays of Australia, CSIRO, Australia. network of marine protected areas in the Legendre, P., and L. Legendre. 1998. Nu- Red Sea and Gulf of Aden region. Ocean merical ecology, 2nd ed. Elsevier Science, Coast. Manage. 46:741–761. Amsterdam. Goldman, K. J., and S. D. Anderson. 1999. Manly, B. F. 1997. Randomization, bootstrap, Space utilization and swimming depth of and monte carlo methods in biology, 2nd white sharks, Carcharodon carcharias, at the ed. Chapman and Hall, London. South Farallon Islands, central California. Manly, B. F., L. L. MacDonald, and D. L. Environ. Biol. Fishes 56:351–364. Thomas. 1993. Resource selection by ani- Habitat Preferences and Site Fidelity of Orectolobus ornatus . Carraro and Gladstone 223

mals: Statistical design and analysis for Quinn, G. P., and M. J. Keough. 2002. Ex- field studies. Chapman and Hall, London. perimental design and data analysis for McCormick, M. I. 1994. Comparison of field biologists. Cambridge University Press, methods for measuring surface topography Cambridge. and their associations with a tropical reef Ray, G. C., and J. McCormick-Ray. 2004. fish assemblage. Mar. Ecol. Prog. Ser. Coastal-marine conservation: Science and 112:87–96. policy, Blackwell, Malden. McKibben, J. N., and D. R. Nelson. 1986. Rechisky, E. L., and B. M. Wetherbee. 2003. Patterns of movement and grouping of Short-term movements of juvenile and grey reef sharks, Carcharhinus amblyrhyn- neonate sandbar sharks, Carcharhinus chos, at Enewetak, Marshall Islands. Bull. plumbeus, on their nursery grounds in Del- Mar. Sci. 38:89–110. aware Bay. Environ. Biol. Fishes 68:113– Morrissey, J. F., and S. H. Gruber. 1993a. 128. Habitat selection by juvenile lemon sharks, Roughley, T. C. 1955. Fish and fisheries of Negaprion brevirostris. Environ. Biol. Fishes Australia. Angus and Robertson, Sydney. 38:311–319. Samoilys, M. 1997. Movement in a large ———. 1993b. Home range of juvenile lem- predatory fish: Coral trout, Plectropomus on sharks. Copeia 1993:425–434. leopardus (Pisces: Serranidae), on Heron NSW Fisheries. 2002a. Discussion paper: reef, Australia. Coral Reefs 16:151–158. Management of wobbegong sharks in Simpfendorfer, C., and M. R. Heupel. 2004. NSW. NSW Fisheries, Sydney. Assessing habitat use and movement. ———. 2002b. Grey nurse shark (Carcharias Pages 553–572 in J. C. Carrier, J. A. taurus) draft recovery plan. NSW Fish- Musick, and M. R. Heithaus, eds. Biology eries, Sydney. of sharks and their relatives. CRC Press, Otway, N. M., and P. C. Parker. 2000. The Boca Raton. biology, ecology, distribution and abun- Sokal, R. R., and F. J. Rohlf. 1995. Biometry, dance, and identification of marine pro- 3rd ed. W. H. Freeman, New York. tected areas for the conservation of the Syms, C. 1995. Multi-scale analysis of habitat threatened grey nurse sharks in South East association in a guild of blennioid fishes. Australian waters. NSW Fisheries Final Mar. Ecol. Prog. Ser. 125:31–43. Report Series No. 19. NSW Fisheries, Tabachnick, B., and L. Fidell. 1989. Using Sydney. multivariate statistics. Harper and Row, Pielou, E. C. 1977. Mathematical ecology. New York. John Wiley and Sons, New York. Underwood, A. J., M. J. Kingsford, and N. C. Pogonoski, J. J., D. A. Pollard, and J. R. Andrew. 1991. Patterns in shallow subtidal Paxton. 2002. Conservation overview and marine assemblages along the coast of action plan for Australian threatened and New South Wales. Aust. J. Ecol. 6:231– potentially threatened marine and estuarine 249. fishes. Environment Australia, Canberra.