Leopard Sharks and Their Prey: Effect of prey availability on the movement patterns of Triakis semifasciata within the Elkhorn Slough National Estuarine Research Reserve
Elijah S. Woolery Moss Landing Marine Labs December 15, 2003
ABSTRACT
Anthropogenic changes in the last 25 years to the area of Elkhorn Slough in California have brought about many alterations to the habitats contained within the Slough, including increased erosion and scouring which has resulted in a change in geomorphology and reduction in complexity to those habitats. This reduction in complexity may have in turn affected the density of many prey items for top predators. The leopard shark, Triakis semifasciata, is a good example of an Elkhorn Slough part-time resident elasmobranch which uses the reserve in essential parts of its life history, foraging in the slough and bearing its young in the tidal creeks. Leopard sharks are opportunistic benthic predators, and feed on a range of epifaunal and infaunal invertebrates, as well as various fish including sculpins, rockfishes and gobies. To better understand critical shark habitat, early tracking data from a study currently being undertaken by Aaron Carlisle of MLML indicates that the sharks spend longer periods of time in certain habitats, in particular the mudflats near the two lagoons in the northern part of the Elkhorn Slough. The hypotheses that there are differences in prey availability within three distinct habitats (intertidal & subtidal mudflats, deepwater channels, and subtidal lagoons) of Elkhorn Slough National Estuarine Research Reserve and these differences can be related to the daily movement patterns of one leopard shark tracked within the reserve in 2003 was addressed by several methods. First, those areas were sampled with baited traps to obtain an estimate of relative prey abundances. Video surveys were then performed to obtain a qualitative idea of prey abundance in each habitat. These data were then compared to tracking data of one leopard shark within ESNERR in an attempt to reject the null hypothesis that there was no difference in prey availability and no relationship between prey availability and time spent for the shark within each habitat. Organisms caught in all three habitats were predominantly decapod crabs. There was no significant difference in prey abundances between any of the three habitats shown in the results of the ANOVA calculated for the mean CPUE of all species by location. Video surveys illustrated many qualitative differences between the habitats. There appeared to be a significant difference in time spent between the mudflats and the channels, with greater time spent in the mudflats. It is hoped that this study will provide the impetus for further investigation of prey availability and its relation to habitat choice for leopard sharks within ESNERR.
INTRODUCTION
The fragility of estuarine environments often leave them extremely vulnerable to human impact, an observation which has gained much popular attention in the last few decades (http://www.nrdc.org/greengate/wildlife/wetlandsv.asp). Elkhorn Slough in central California is no exception, and anthropogenic changes in the last 25 years have brought about many alterations to the habitats contained within the slough, including increased erosion and scouring which has resulted in a change in geomorphology and reduction in complexity to those habitats (Malzone and Kvitek 1994, Lindquist 1998).
This reduction in complexity may have in turn affected the density of many prey items for top predators, as “prey density is often positively correlated with habitat structure because it provides food and substrate to the prey as well as a relative refuge from predators” (Crowder and Cooper 1982). There has also been a notable reduction in the diversity of some organisms within the slough, specifically fishes and their prey
(Yoklavich et al.. 2002). Elasmobranchs in particular may be vulnerable to such changes in habitat and prey availability or diversity, due to their low fecundity, relatively long lifespan and late maturity, characteristics which make them susceptible to overfishing
(Cailliet 1992).
The leopard shark, Triakis semifasciata, is a good example of an Elkhorn Slough part-time resident elasmobranch which uses the reserve in essential parts of its life history, foraging in the slough and bearing its young in the tidal creeks (Yoklavich et al..
2002). Though the population may not be at immediate risk, declines may have already taken place in areas such as the kelp forests of California, where leopards shark have
2 been heavily impacted by gillnet fishing and have not been seen with the frequencies they
were spotted in the early 1970’s (Dayton et al. 1998). Leopard sharks range from Baja
California northward to Oregon, are common in shallow water from the intertidal to 3 m,
less common to 100 m or deeper in ocean waters (Smith 2001). In northern California the shark frequents muddy bays and sloughs, and “is known to move in and out with the tides to feed over shallow mudflats” (Smith 2001). Maximum recorded and verified total length is approximately 2 m, and the oldest validated age is 26 yrs for a 124 cm female, an average growth rate of 4.57 cm per year (Cailliet 1992). Males mature at 7 years, and females at 10 years, when the sharks are between 100 and 107 cm total length (Cailliet
1992). Female leopard sharks are ovoviviparous, producing from seven to 36 offspring per annual reproductive cycle, with gestation period estimated from 10 to 12 months
(Cailliet 1992).
Leopard sharks are opportunistic benthic predators, and feed on a range of epifaunal and infaunal invertebrates, as well as various fish including sculpins, rockfishes and gobies (Smith 2001). Feeding studies of leopards within Elkhorn Slough in 1999 revealed a diet consisting predominantly of the fat innkeeper worm Urechis caupo, with a
%IRI value of roughly 60 for both small and large sharks, followed in importance by fish for larger sharks and polychaetes for smaller sharks, with %IRI 37.42 and 13.87 respectively, see figure 1 (Kao 2000). This contrasts with a 1973 feeding study of sharks
in the slough, where clams and crabs were of higher relative importance, pointing again
to large-scale physical and biological changes in the estuary (Talent 1976, Kao 2000).
Clearly, it would be useful to understand where these sharks spend their time foraging in
3 order to determine the effects of further anthropogenic changes or the continuing effects of past changes on critical habitat.
To better understand this habitat utilization, a tracking study is currently being undertaken by Aaron Carlisle of Moss Landing Marine Laboratories in Moss Landing,
CA as a part of his Masters thesis. Early tracking data indicates that the sharks spend longer periods of time in certain habitats, in particular the mudflats near the two lagoons in the northern part of the Elkhorn Slough National Estuarine Research Reserve
(ESNERR). There are many reasons why the shark may spend the majority of its time in such areas. Physical factors play an important role in determining suitable habitat for estuarine elasmobranchs such as leopard sharks, and temperature and salinity appear to be the most important environmental variables in determining their distribution and abundance (Hopkins and Cech 2003). Aside from the physical and chemical factors such as salinity, temperature, turbidity, and tidal movements, that determine appropriate habitat for estuarine fish species, biological functions such as feeding, migration, and reproduction are of paramount importance in determining where such fish spend time
(Morin et al. 1992). Also, studies have suggested that “organisms select their habitat on the basis of food resources and alter their behavior as needed to avoid predators,” which has resulted in a “series of complex interactions that involve the use of many different habitat areas” (Heupel and Hueter 2002). In this study, the factor of prey availability will be focused on, and though previous studies of other shark species have not found any correlation between time spent in a particular area and the prey density within that area
(Heupel and Hueter 2002), it is hoped that this investigation will reveal some reasons for
4 the particular habitat utilization within ESNERR, with a focus on the mudflats, channels,
and lagoons of the reserve.
The hypotheses that there are differences in prey availability within three distinct
habitats (intertidal & subtidal mudflats, deepwater channels, and subtidal lagoons) of
Elkhorn Slough National Estuarine Research Reserve and these differences can be related
to the daily movement patterns of one leopard shark tracked within the reserve in 2003
was addressed by several methods. First, those areas were sampled with baited traps to
obtain an estimate of relative prey abundances. Video surveys were then performed to
obtain a qualitative idea of prey abundance in each habitat. These data were then
compared to tracking data of one leopard shark within ESNERR in an attempt to reject
the null hypothesis that there is no difference in prey availability and no relationship
between prey availability and time spent for the shark within each habitat.
METHODS
The leopard shark tracked and all prey items and habitats sampled were located in
the Elkhorn Slough National Estuarine Research reserve, located at approximately 121°
44’ W, 36° 48’ N (fig 2), in the months of October, November, and December of 2003.
The three habitats of interest were the deeper channels (2.5-4.5 m depth), intertidal and subtidal mudflats (0.3-2.5 m depth), and the two lagoon areas in the northern part of the reserve (depth varies, but generally <2.5 m) (fig 3) Originally, three random locations within each of the habitats were to be selected for sampling using random GPS coordinates correlated with GIS maps of the reserve, but due to restrictions on the
5 research permit issued by the reserve and the lack of access to a boat, the three replicate
locations chosen in each habitat were more haphazardly chosen.
Sampling of these locations was accomplished by the use of two trapping
methods, one a larger nylon mesh minnow type trap of which three were utilized, the
other a smaller steel mesh trap of which two were utilized (fig 4). These five traps were
systematically deployed at each of the trapping locations in an effort to sample each location as thoroughly as possible. Traps were baited with two 50 gm pieces of frozen
bait squid (Loligo opalescens) and left for 24 hour blocks of time to encompass all tidal
heights. Upon collection, organisms were identified and counted, then released.
Photographs were taken of those organisms that were not identifiable in the field.
Mean catch per unit effort (CPUE) per 24 hour sampling period for each taxa in each
location was calculated, and a single factor ANOVA was performed for the mean CPUE
of all species at each location. A contingency table for crab catch data was also used to
determine if catch was independent of location between each taxa and the mudflat and
channel locations.
Video dive surveys were originally intended to include extensive transects within
each habitat so that densities of epibenthic prey items and visually identifiable infaunal
burrowing prey items could be calculated. However, pilot dive surveys revealed a paucity
of visible prey items as well as a difficulty in completing transects without severe
reduction in visibility due to disturbance related turbidity. Qualitative video surveys were
undertaken instead, with the use of a Light & Motion Mako housing for the Sony PC120
digital video camera, with lighting provided by Light & Motion High Intensity Discharge
(H.I.D.) lights. Representative sites from each habitat were selected so as to include
6 areas that had been sampled with the trapping studies (fig 5). For shallower sites, several
random initial headings were taken and several 20 to 30 m length surveys filmed. In the
deeper channels where direction was more constrained, a 40 minute dives was completed
and filmed.
Tracking of the leopard shark within ESNERR took place from May through
August of 2003. The shark was caught using tended monofilament gillnets within the
reserve. Upon being caught, the shark was placed in a tub of fresh sea water and
anaesthetized by adding a 0.2 g/L solution of MS-222 to the tub. After being sexed,
measured, and tagged with a conventional visual tag in its dorsal fin, it was placed in a v-
board for surgery. A small 3 cm incision was made into the peritoneal cavity and a
VEMCO V16-5H transmitter gently inserted into the peritoneum. Incision was then
sutured, cleaned with an antibiotic wash and the shark was revived and released.
The shark was then manually tracked from a small flat-bottom boat using a
VEMCO VR60 field receiver. Sharks were located every five minutes, and the GPS position of boat, signal strength, estimated distance & bearing of shark from boat logged, as well as any other relevant notes on location. Tracking was done in four 6-hour blocks
(0000-0600, 0600-1200, etc).
These data were then converted from degrees-minutes-seconds into decimal degrees which were then entered into a GIS map of the area (fig 6). To account for the fact that many of the mudflat areas were only available to the shark when tides were above 1 meter, all data taken from lower tides was ignored for this segment of the study.
From the boat positions, notes, and the information provided by the map, it was determined which habitat the shark was frequenting (as it was only tracked into the
7 lagoon once, only channel and mudflat locations were logged for comparison). These frequencies were then compared using a paired sample t-test
RESULTS
Organisms caught in all three habitats were predominantly decapod crabs. Catch was dominated by the invasive green crab Carcinus maenas (fig 7 ). Next in abundance were grapsid crabs including the yellow mud crab Hemigrapsus orogrensis and the lined shore crab Pachygrapsus crassipes, followed by the graceful crab Cancer gracilis. Fish caught in the traps included representatives from three taxa: most common was the staghorn sculpin Leptocottus armatus , next a juvenile rockfish, last the yellowfin goby
Acanthogobius flavimanus (fig 8 ).
Results from the mean catch per unit effort (CPUE) per 24 hour sampling period for each taxa in each location varied, though all areas were dominated by C. maenas.(fig
9). In the lagoons, the C. maenas. had the highest mean CPUE, followed by the grapsid crabs, L. armatus., C. gracilis., and A. flavimanus. In the mudflats, only three taxa were represented, with C. maenas again dominating, followed by the grapsids and C. gracilis.
In the channels, C. maenas once again had the highest CPUE, though in this case not by such a high margin. It was followed by the grapsids, C. gracilis, S. saxicola, and L. armatus.
There were no significant differences in prey abundances between any of the three habitats shown in the results of the ANOVA calculated for the mean CPUE of all species by location in fig (df=2, α=0.05, p=0.546, fig 10). However, an upwards trend could be
8 noted from channel with the lowest mean CPUE through lagoon with the highest. In the contingency table, crab catch was determined not to be independent of location (Pearson
Chi-square 12.464, df=2.00, α=0.05, p=0.002).
Results from the video surveys, though qualitative in nature, were nonetheless
interesting in the number of obvious differences between each habitat, apparent even
from first glance at video screen captures (fig 11 ). The mudflat survey showed areas
relatively depauperate of epibenthic organisms & identifiable infaunal burrows. The
dominant organism, at least in density of visible burrows, was the amphipod Corophium sp., followed by the polychaete Capitella capitata, 6 of whose 1-1.5 cm fecal mounds can be seen in the screen capture. Crabs, both H. oregensis and C. maenas were also seen, though not in great numbers. Also of note were several depressions of a size and shape that could have been a result of leopard shark foraging, though they also could have been due to the bat ray Myliobatis californica (Greg Cailliet, pers. comm.).
A large number of crabs were seen in the lagoon survey, ranging from high densities of H. oregensis at the tips of Ulva sp. fronds, to individual C. maenas foraging among the mussel beds. Also of note was the high density of the sponge Haloclona sp., especially near the culverts, areas of high tidal flow.
In the channel survey, there was a shift in the benthic composition from soft mud to harder, scoured clay, and the infauna appeared to change accordingly. The dominant infaunal organism was the bivalve Zirfaea pilsbry, evident from exposed siphon, at bottom of channel. In addition, large piles of shell, mostly the clams Tresus sp &
Saxidomus sp, appear to be remnants of otter foraging from the last major recruitment period of these bivalves (John Oliver, personal comm.)
9 There appeared to be a significant difference in time spent between the mudflats
and the channels, as illustrated by the results of the paired t-test, with greater time spent in the mudflats. (fig 12 ) The figure illustrates comparisons made between the frequency
of time spent in each location in each sampling block, with only one outlier where greater
time was spent in the channel than the mudflat.
DISCUSSION
There was no significant difference found in prey densities or abundance in any of
the three habitats, and thus a direct comparison between prey availability and habitat
utilization could not be made. Though we can not reject the null hypothesis, several
intriguing observations were made during the course of the study, and problems with the
methods were noted that could assist with the design of future studies.
One of the first observations of note, though perhaps of no direct consequence to
the leopard shark. was the large number of invasive green crabs found in all sampling
sites. Though C. maenas has been found in the slough as early as 1994, trapping studies
performed as recently as 2000 found far fewer crabs: in 4 24 hour time block days with
66 traps set only 6 crabs were caught (Estelle 2000). This evidence for an increase in the
number of green crabs also has implications for the other prey items of the leopard shark,
as these invasives are intense predators of native bivalves, crustaceans, and polychaetes
(Cohen et al. 1995).
Another reason the numbers of sampled native prey organisms may have been so
low could have been bias in sampling caused by the traps themselves. Because of the
10 relatively long 24 hour sampling period, organisms had time to escape, which would have
been aided by the large mesh size in the 3 large traps. The results from the contingency
table supported the idea that crab catch was not independent of location, which provides
some impetus for further investigation of differences in prey availability between
habitats. Also, constraints on location limited the sampling to areas where prey
abundances may have been lower simply due to natural variation.
This variation could have also been reflected in the video surveys, though they did
tend to give at least a qualitative idea of the difference between habitats. Within the
mudflats, none of the distinct conical burrow entrances of the Inkeeper worm were seen,
which was surprising given their importance in the leopard sharks diet (Kao 2000) and
their historical abundance in the slough (MacGinitie 1935). The abundance and density of
H. oregensis perched on Ulva within the lagoon was another surprising observation, as
similar Ulva patches were seen in the mudflats and channels, but no correspondingly high
numbers of crabs were seen. This could perhaps be due to the high tidal flow near the
culverts, which may in turn be providing plankton for the crabs to feed upon as well as
supporting the high concentration of filter-feeding organisms in the area; similar high flow environments exist near the bridge at the mouth of the slough and also support a diversity of organisms (Yoklavich et al. 2002).
The effects of high tidal flow, and its associated scour, on the composition of the channel habitats was also observed. As noted, the high densities of Z. pilsbry dominated the beds of the channels, probably due to their ability to burrow efficiently in the relatively hard-packed clay substrate (Oliver, personal comm.). The large beds of clam shells point to utilization by otters, whose diet consists primarily of crabs and clams when
11 they are present in the habitat (Kvitek et al.. 1988). This in turn supports the idea that changes within the slough have a biological as well as a physical component (Kao 2000).
Although these changes may not directly affect the leopard shark population of
Elkhorn Slough, and though the greater leopard shark population does not currently
appear to be at risk (Smith 2001), it is nonetheless important to identify critical areas of
their habitat and understand why they need to utilize those parts of the slough in
particular, if for no other reason than to have data that could assist in creating a management plan as a precautionary measure, should the population suffer a significant decline. The preliminary tracking results seem to support a significant difference in the amount of time spent in the mudflats and channels, with a preference for the channels during higher tide periods. Other results of this study may assist in the design of future
studies of prey availability within ESNERR, particularly with respect to sampling
methods and locations. It may be more practical to use photo quadrats rather than video
transects, due to the low-visibility and ease of increase in turbidity due to disturbance.
Also, more efficient trapping methods with respect to mesh size and sampling period, and the use of boats to place traps, could greatly increase their effectiveness. A larger sample size and better attempts at randomization could also lead to a more accurate estimate of prey abundance. With these considerations in mind, future studies could help identify essential habitat for the shark and assist in the management and protection of those areas.
12 Literature Cited
Cailliet GM (1992) Demography of the Central California population of the leopard shark (Triakis semifasciata). Australian Journal of Marine and Freshwater Research 43:183-193
Cohen AN, Carlton JT, Fountain MC (1995) Introduction, dispersal and potential impacts of the green crab Carcinus maenas in San Franscisco Bay, California. Marine Biology 122: 225-237
Crowder LB and Cooper WE (1982) Habitat structural complexity and the interaction between bluegills and their prey. Ecology 63(6):1802-1813
Dayton PK, Tegner MJ, Edwards PB, Riser KL (1998) Sliding baselines, ghosts, and reduced expectations in Kelp Forest Communities. Ecological Applications 8(2):309-322
Estelle VB (2000) Trapping Results of Green Crabs (Carcinus maenas) in Elkhorn Slough. Report to the Monterey Bay Nationial Marine Sanctuary. Contract Number 40ABNC901298.
Heupel MR and Hueter RE (2002) Importance of prey density in relation to the movement patterns of juvenile blacktip sharks (Carcharhinus limbatus) within a coastal nursery area. Mar Freshwater Res 53: 543-550
Hopkins TE and Cech JJ (2003) The influence of environmental variables on the distribution and abundance of three elasmobranches in Tomales Bay, California. Environmental Biology of Fishes 66:279-291
Kvitek RG, Fukuyama AK, Anderson BS, Grimm BK, Sea otter foraging on deep- burrowing bivalves in a California coastal lagoon (1988) Mar. Biol. 98(2):157- 167
Kao JS (2000) Diet, diet, daily ration and gastric evacuation of the leopard shark (Triakis semifasciata). M.S. Thesis, California State University San Jose, California , p 25- 87
Lindquist DC (1998) The effects of erosion on the trophic ecology of fishes in Elkhorn Slough, California. M.S. Thesis, California State University San Jose, California, p 25-29
Morin B, Hudson C, Whoriskey FG (1992) Environmental influences on seasonal distribution of coastal and estuarine fish assemblages at Wemindji, eastern James Bay. Environ. Biol. Fish 35(3):219-229
13 MacGinitie GE (1935) Ecological Aspects of a California Marine Estuary. American Midland Naturalist 16(5):646-650
Malzone C and Kvitek R (1994) Tidal scour, erosion, and habitat loss in the Elkhorn Slough, CA. Report to the Elkhorn Slough Foundation to the National Oceanic and Atmospheric Administration. Award #NA370M0523
Smith SE (2001) Nearshore Ecosystem Fish Resources: Overview: Leopard shark in Leet WS, Dewees CM, Klingbeil R, Larson EJ (eds) California’s Marine Living Resources: A Status Report. California Department of Fish and Game, p 252- 254
Talent LG 1976 (1976) Food habits of the leopard shark (Triakis semifasciata). In: Elkhorn Slough, Monterey Bay, California. Calif Fish and Game 62(4):286-298
Yoklavich MM, Cailliet GM, Oxman DS, Barry JP, Lindquist DC (2002) Fishes. In: Caffrey JP, Brown J, Tylee M, Silbersein M (eds) Changes in a California Estuary: A Profile of Elkhorn Slough. Elkhorn Slough Foundation p 163-185
14 Tables & Figures
70
60
50
40 small
%IRI large 30
20
10
0 Urechis Fish Polychaetes Crabs Clams
Fig 1. %IRI Values of various prey items in large and small leopard sharks Within Elkhorn Slough (Adapted from Kao 2000)
Fig 2. Study area within ESNERR near Moss Landing, California
15
d a o R
n r o h k d l Roa E lley en Va Hidd
d a o r l i a
R
s e u q n a T ia V
g in id S ad Ro vila A Dolan Road
Legend Roads Streams Channels
Mudflats 00.1 0.2 0.4 0.6 0.8 Kilometers
Channels
Fig. 3 Map of selected trapping sites within ESNERR
16
Fig 4. Large (left) and small traps used in study
Fig 5 Map of dive survey locations within ESNERR
17 Legend
Roads Streams
Channels
Mudflats 0513_1218 Events 0522_1800 Events 0721_1218 Events 0722_1218 Events 0724_1218 Events 0725_0612 Events 0728_1800 Events 0731_1800 Events 0801_0006 Events
Fig. 6 Boat locations during tracking sessions by day and 6 hour block
18 Total N of crabs caught per habitat
50 45 40 35 30 Channel 25 Mudflat 20 Lagoon 15 10 Lagoon 5 Mudflat 0 Channel Grapsid Cancer crabs Green crabs crabs
Fig 7
Total N of fish caught per habitat
7
6 Channel 5 Mudf lat 4 Lagoon
3
2
1 Lagoon Mudf lat 0 Channel
s sp. tocottus sp. astes sp. u p b bi o Le Se thog an Ac
Fig 8
19
Mean CPUE in Lagoon
25 20 15 10 5 0 -5 Grapsids Cancer sp. Carcinus sp. Leptocottus sp. Acanthogobius sp.
Mean CPUE in Channel
6 5 4 3 2 1 0 -1 Grapsids Cancer sp. Carcinus sp. Leptocottus sp. Sebastes sp.
Mean CPUE in Mudflat
20
15
10 5 0 -5 Grapsids Cancer sp. Carcinus sp.
Fig. 9 Mean CPUE in each habitat type
20 Least Squares Means
46
39
32
T 25
N
U
O 18
C
11
4
-3 Channel Mudflat Lagoon LOCATION
df=2, α=0.05, p=0.546
Fig. 10. Single factor ANOVA calculated for mean CPUE of all taxa by location
Fig 11. Video capture of dive surveys (from left to right): Mudflat, Lagoon, Channel
21
50
40
y
c 30
n
e
u
q
e r 20
F
10
0 MUDFLAT CHANNEL Index of Case
df=7, p<0.05,α=0.05
Fig 11. Results from paired t-test comparing frequency shark spent in mudflat vs channel
22