Using acoustic telemetry monitoring techniques to quantify movement patterns and site fidelity of sharks and giant trevally around French frigate shoals and midway atoll

2006·8.9 MB·English

by  Christopher G Lowe

USINGACOUSTIC TELEMETRY MONITORING TECHNIQUES TO
QUANTIFY MOVEMENT PATTERNSAND SITE FIDELIT\ OK
SHARKSAND GIANTTREVALLYAROUND
FRENCH FRIGATE SHOALSAND MIDWAYATOLL
BY
CHRISTOPHER G. LOWE', BRADLEYM. WETHERBEE-,AND CARL G. MEYERS
ABSTRACT
TheNorthwestern Hawaiian Islands (NWHI) host a variety oflarge vertebrate
animals including seabirds, green sea turtles (Chelonia mydas), Hawaiian monk seals
{Monanchusschauislandi), and large teleost fish such as trevally (Family Carangidae)
and several species ofsharks. The air-breathing vertebrates have been the subjects of
relatively continuous and well-funded research programs overthe past several decades,
and many aspects oftheirbiology in theNWHI have been documented fairly well.
However, studies directed atunderstanding the biology and ecology oflarge teleost fishes
and sharks in the NWHI have lagged substantially behind research conducted on birds,
turtles and seals. Inthe summer of2000, an array ofautonomous acoustic receivers was
deployed at French Frigate Shoals (FFS) in theNWHI as part ofa project investigating
the movement patterns oftiger sharks {Galeocerdo ciivier) within the atoll, particularly
in relationto the high seasonal abundance ofpotential prey (birds, turtles, seals). Shortly
afterthe establishment ofthe initial array ofmonitors in 2000, additional monitors were
deployed in an effort to monitorthe movements ofGalapagos sharks (Carcharhinm
galapagensis) at FFS, particularly at locations where monk seal pups had been preyed
upon by these sharks. The scope ofthe monitoring study was furtherexpanded to
MidwayAtoll during summerof2001 to monitormovements ofGalapagos sharks near
seal haul-outbeaches andto examine survivorship and behaviorofgianttrevally {Caranx
ignobilis) captured and released in a commercial sport fishing operation conducted
within the MidwayNational Wildlife Refuge. Foreach study, experimental animals were
captured and surgically fittedwith long-life, individually-coded acoustic transmitters.
Duringnearly4 years ofacoustic monitoring at FFS and 2 years ofmonitoring
atMidway, a total ofover45,000 detections ofsharks andfish with transmitters
were recorded on acoustic monitors. These data enable an assessmentoflong-term
movementpatterns ofthese large predators within theNWHI. Each species investigated
demonstrated somewhatrepeated and predictable behavioral patterns that provide a basis
for improved understanding ofdeterminants ofbehaviorand forenhanced management
ofthese animals andprey (birds, seals, turtles) with which they may interact.
'Dept.ofBiologicalSciences,CaliforniaStateUniversity,LongBeach, 1250BellflowerBlvd.,LongBeach,
CA90840USA,E-mail: [email protected]
^Dept. ofBiological Sciences,Univ. ofRhodeIsland, 100FlaggRd,Kingston, RI02881 USA
'HawaiiInstituteofMarineBiology,POBox 1346,Kaneohe, HI 96744USA
282
INTRODUCTION
TheNorthwestern Hawaiian Islands (NWHI) support awide variety oflarge
marine vertebrates and are awell knownbreeding grounds forseabirds, green seaturtles
{Che/oniamydas), andthe endangered Hawaiian monk seal {Monanchiisschaiiislandi)
(Gerrodette andGilmartin, 1990; Gilmartinand Eberhardt, 1995). The nearshore waters
surroundingthese islands are also hometo several species oflarge, predatoryfishes and
sharks. Concern overnegative human impacts on NWHI seabird, seaturtle, andmonk
seal populations has resulted in substantial efforts to monitorand rebuildpopulations of
these animals (Gilmartin andEberhardt, 1995). Establishment ofNWHI field camps and
permanent field stations has enabledlong-term studies ofthese populations, andmany
aspects ofthebehavior, feeding, reproduction, andpopulation dynamics ofthese species
havebeen characterized (Rice and Kenyon, 1962; Harrison etal, 1984; Gilmartin and
Eberhardt, 1995).
Despite theirabundance (FriedlanderandDeMartini, 2002), importance in trophic
interactions as apexpredators (Polovina, 1984), andpossible impact on protectedand
endangered species populations (Balazs andWhittow, 1979;Alcorn andKam, 1986;
Lowe etal., 1996), studies on thebiology and ecology ofthe largepredatoryfishes
(sharks andtrevally) ofthe NWHI have lagged considerablybehindthose ofseabirds,
turtles and seals. Much ofthe research thathas been conducted on large marine fishes in
theNWHI hasbeen limitedto islands with sufficient infrastructure (i.e., field stations,
small boats, andready access) to support seasonal or short-termfieldwork (French
Frigate Shoals and Midway), orhasbeen conducted fromresearch shipsbrieflyvisiting
various islands within the NWHI (Tricas et al., 1981; Sudekumetal., 1991). Because
oftheir solely aquatic nature, these fishes cannotbe observed, captured, ormonitored
as easily as air-breathing vertebrates that spendperiods oftime eitheron land oratthe
surface.
Standardtechniques typicallyusedto assess and monitorfishpopulations in other
NWHI
locations are not effective in the for several reasons: 1) the remoteness ofthe
NWHI adds greatly to the cost offieldwork andtransportation to study sites and reduces
the effectiveness ofmethods that rely typicallyon local recreational or commercial
fisheries; 2) the limited availability ofsuitable boating facilities within theNWHI and
the often difficult sea conditions severelyrestrictuse ofsmall boats that are neededto
access thesefishes; 3) there are extensivefishing restrictions withintheboundaries ofthe
NWHI and MidwayAtoll National Wildlife Refugebecause ofpotential interactions with
endangered monk seals; and 4) diver surveys are limited to only daytime observations
and are oftenbiasedbecause divers tendto attract some ofthe largepredatoryfishes and
mayrepel others.
Because ofthe limitations ofvarious fisherytechniques, telemetryhas become
increasinglypopularforremotemonitoring offishpopulations (Voegeli et al., 2001;
Simpfendorferet al., 2002; Heupel et al., 2004; Lowe and Bray, 2006). Acoustic
telemetry monitoringutilizes autonomous receivers to continuously"listen" forthe
presence orabsence oforganisms fittedwithuniquely coded transmitters, andto store
these data for long periods oftime. Placement ofautonomous receivers along a coastline,
in channels, orin arrays can allow forrelatively long-term (>1 year) monitoring of
283
movement patterns and fidelity to an area. Unlike conventional tag and recapture
methods, acoustic monitoring allows for repeated "electronic" recaptures without the
need forcontinuous fishing efforts and in some instances may be a more effective tool for
monitoring population dynamics ofspecies such as sharks and trevally that are difficult to
study (Voegeli et al., 2001).
We used an array ofautonomous acoustic receivers to monitor the movement
patterns and site fidelity oftiger sharks {Galeocerdo ciivier), Galapagos sharks
{Carcharhimisgalapagemis), and giant trevally (Caranx ignobilis) around specific
islands at FFS and MidwayAtoll from 2000 to 2004. The objectives ofthis paperare to
demonstrate whetherthese large predatory fishes show any affinity to islands containing
common semi-ten-estrial prey (i.e.. seabirds, sea turtles, and monk seals) and to illustrate
the utility ofacoustic monitoring for studying the movement patterns oflarge fishes in
remote locations overvarying spatial scales.
METHODS
Study Sites
This study was conducted at two atolls within theNWHI: French Frigate Shoals
(FFS) from 2000 to 2004, located midwayalongthe HawaiianArchipelago (23° 52.3'
N latitude, 166° 14.4'W longitude); and MidwayAtoll from 2001 to 2003, nearthe
northwestern end ofthe chain (28° 15'N latitude, 177° 20'W longitude). At FFS, our
base ofoperation was the U.S. Fish andWildlife Service (USFWS)fieldstation on Tern
Island, and at Midway operations were conducted in cooperation with USFWS and
Midway Phoenix Corporation from Sand Island.
Fishing andTagging
Sharks were caughtusinghandlines baitedwithdeadbirds orfish. Handlines
were monitored continuously during all fishing efforts. Ourfishing methods used large
hooks (14/0) and large baits in orderto target larger sharks, although several species of
smaller sharks (gray reefsharks- Carcharhimis amblyrhyiichos and whitetip reefsharks
{Triaenodon obesus) were occasionally caught at FFS. All tigerand Galapagos sharks
caughtwere brought along side ofthe 6-m boat, and a rope was placed around theirtail.
Once sharks were restrained, theywere inverted andplaced in tonic immobility, atwhich
point each was measured, sexed, taggedwith an external identification tag (M-capsule
tags or spaghetti type dart tags) in the dorsal musculature, and fitted with a coded acoustic
transmitter.
At FFS the majority offishing fortiger sharks was conducted nearthe centerof
the atoll atEast Island, whereas Galapagos sharks were targetedprimarily at Trig Island,
along theperimeterofthe atoll (Fig. 3). Duringthe final 2 years ofoperations at FFS,
we were notpermitted to fish within 800 m ofTrig Island orto use chum in attempts to
attract sharks to baitedhooks. The same methods used to fish for Galapagos andtiger
sharks at FFS were employed at MidwayAtoll; however, giant trevally were caught via
trolling orby dunking fresh bait from a boat.
284
Transmitters andAutonomousAcoustic Receivers
To determine longer-term site fidelity ofsharks and trevallyto islets atFFS and
Midway, individuals werefittedwith codedacoustic transmitters (V16-R256 random
coded, 69.0 kHz, Vemco). Sharks caught onhandlines werebrought along side the boat
andplaced intonic immobility (Fig. la, b). Codedtransmitterswere implanted surgically
into thebody cavity ofsharks through a small incision (4 cm), and the woundwas closed
with 4-5 interrupted sutures. Transmitters were coated with acombination ofbeeswax
(30%) andparaffin wax (70%) to reduce immune response (Holland etal., 1999). Each
transmitter emitted a uniquely coded acoustic signal at random intervals between 40-70
seconds andhadbatteiy lives ofup to4 years.
Gianttrevallywere anaesthetizedwithMS-222 (0.2 g/L, 30 to45 s immersion
time), placed on a foam pad and measured (fork length (FL) in cm).Acodedtransmitter
(V16-R256 randomcoded, 69.0 kHz) coatedwithbeeswax/paraffinwas implanted
surgically into thebody cavity ofeachfish (Fig. Ic). Before surgery the scalpelblade
and transmitterwere immersed in iodine solution, andthe incision site was swabbed
with iodine solution.Asmall (20 mm) incision was made throughthe peritoneal wall
into the posteriorregion ofthe body cavity. This site was chosento avoid damage to
internal organs from transmitter insertion. The transmitterwas inserted into thebody
cavity through the incision, which then was sutured closed. Each fish was also tagged
externally with a seriallynumbered, 10-cmplastic dart identificationtag (Hallprint,
SouthAustralia), resuscitatedbytowingorswimming italongside the boatuntil fully
responsive, and then released (Fig. 2).
An array ofautonomous acoustic receivers (VRl model, Vemco) wasplaced at
locations aroundvarious islandswithinFFS andMidway. Thesereceivers are designed
to listen forcoded transmitters and to recordthe date and time ofarrival and departure
ofindividual sharks andtrevally.AtFFS, 10 receivers were placed around Tern, Trig,
Round, East, Shark, and Gin Islands at depths easily reachedby free diving (average
depth ofmonitors was 2.5 mbelow the surface) (Fig. 3a). At Midway, five receivers were
placed adjacentto Sand and Eastern Islands, in the mainboat channel and onthe outer
reefat a dive site named "FishHole" (Fig. 3b). USFWS personnel recoveredthree of
these receivers in summer 2004, but were unable to relocate the receiverfrom FishHole.
All receivers were securedto the benthos using sand screws and swiveling
stainless steel rods. Foam floats were used to buoy acoustic receivers andattachment
gear(Fig. 4). This design was chosen to reduce the risk ofmonk seal entanglement in the
equipment an'ays. The majority ofreceivers remained inplace formanyyears withthis
design, although severalfloats were lost, and all floats that were still attachedtomonitors
showed evidence ofshark bites.
Acoustic range ofeachreceivervaried depending onwaterdepth, tide, and
neighboring reefstructure. Range tests at several sites indicatedtransmitterdetection
m
ranges ofup to 400 m; however, atmost locations the range was onthe orderof20-50
due to shallow depth andproximity ofareeforan island. Receivers were downloaded
every4 to 7 monthsby the researchteam orby USFWS personnel.
2X5
Site Fidelity and MovementAnalysis
Degree ofsite fidelity and extent ofuse ofa particular area was determined hy
the amount oftime a fish spent in proximity to a particular receiver and by the number
ofdetections at each location. Annual catch rales (CPUE) and recapture rales were
determined foreach island. Extent ofmovementwithin the acoustic receiverarray at al
islands was determined by measuring the lineardistance between the two most distant
receivers where tagged sharks orgiant trevally were detected.
RESULTS
French Frigate Shoals
Catch Data. During four summers (2000-2003) and one fall (2002), a total of
477 h were spent fishing at East andTrig Islands, with 190.5 h spent fishing around East
Island. Atotal of34 sharks were caught at FFS, including tiger, Galapagos, whitetip reef,
andgrey reefsharks. Ofthe 34 sharks caught, 4 Galapagos and 13 tigersharks were
fitted with coded acoustic transmitters (Table 1). With the exception ofa few whitetip
reefandgray reefsharks, only tiger sharks were caught at East Island, whereas many
ofthe sharks caught andobservedatTrig Islandwere Galapagos sharks. The CPUE for
tiger sharks in all fishing at East Island was 0.052 sharks h'. In 2002 and 2003, very little
time was spentfishing at East Island (7.5 h), andno tiger sharks were caught. In previous
years, tigersharks were fi"equently observedpreying on fledging albatross chicks in the
mornings, when the winds appeared to provide the best opportunities forthe young birds
to fly. In 2003, we sightedvery few tiger sharks at East Island, although this trip was
conductedduringAugust, whennearly all albatross have fledged from East Island. No
Galapagos sharks were seen or caught at East Island.
During 2002-2003, the majority offishing effort was focused in the vicinity of
Trig Island in an attempt to target Galapagos sharks. Atotal of274 hwas spent fishing
nearTrig Island. Although tiger sharks were rarely seen atTrig Island, over all years
we caught one small, one medium and two large-sized tiger sharks (178, 259, 394,
and 397 cmTL), three ofwhich were captured in Octoberof2002 (Table 1).Atotal of
fourGalapagos sharks were also captured atTrig Island. CPUEs fortiger sharks and
Galapagos were identical (0.015 sharks h'). Galapagos sharks were the most common
large sharks observed atTrig Island; however, theiroccurrence appeared to vary widely
on both a daily and annual basis.
The total fishing effort in all years ofthis study resulted in the capture, tagging,
and instrumentation with transmitters of 13 tiger sharks and 4 Galapagos sharks. Ten gray
reefsharks were also caught during this time period but were onlytaggedwith standard
identificationtags, andnone ofthe whitetip reefs sharks caughtwere tagged. All tiger
sharks caughtwere females, ofwhich-70% appeared notably rotund and may have been
pregnant. The averagetotal lengthoftigersharks caught was 350± 7 cm (± sd), and,
based on available reproductive data, it is likelythat all except two sharks were mature
(Wetherbee et al., 1994). The fourGalapagos sharks captured atTrig were relatively large
andhadan average total length of248 + 2 cm (Table 1).
286
Acoustic Monitoring. All ofthe 13 tiger sharks tagged at FFS were detectedby
acoustic receivers. Tiger sharks were detected atotal of38,886 times during the course of
this project. Two tiger sharks (ID tag #005 and #011) werenot detectedonreceivers until
26 and 11 months, respectively, following tagging andrelease. Ofthe nine tigersharks
tagged at East Island, all were detected at East Island as well as at islands otherthan East
Island (Trig, Gin, Round, Shark, andTern Island) throughoutthe yearatFFS. Based on
the numberofacoustic detections (hits) recordedby differentreceivers, the amountof
A
time sharks spent in proximityto certain islands varied considerably. vastmajorityof
the hits from tiger sharks were recorded inJune andJuly at East Island, whereas tiger
sharks spentproportionally moretime aroundTern Island in the wintermonths (Fig.
5). With the exception ofthe monitors at East Island, detections were usuallybrief,
suggesting that sharks were passing through an areawhen detected. Tigersharks also
showed distinct temporal patterns ofvisits tothe various islands, particularly at East
Island, where they were typically detected during summermonths in the mornings. One
tiger shark (#005) tagged at East Island, FFS in July 2000 was detected by an array of
acoustic receivers offthe Konacoast (approx. 1,190 km straight-line distance) from
January-March 2003. Anothertigershark (#008) tagged at East Island, FFS inJuly 2000
was detectedby ouran-ay ofacoustic receivers offMidway (approx. 1,280 km straight-
line distance) from September-December2002 (Table 1).
Ofthe fourGalapagos sharks tagged, three were detectedby acoustic receivers
at FFS, yielding atotal of2,891 detections duringthe entire study. These sharks were
detectedprimarilybymonitors atTrig Island, followedbyTern Island, and only a few
briefdetections at Shark and East Islands. The occurrence ofGalapagos sharks atTrig
Islandvaried seasonally, with fewest detections recorded betweenFebruaryandJuly,
andanelevatednumberofdetections betweenAugust andJanuary (Fig. 6). Detections at
Tern Island, as well as Shark and East Islands, also were highestbetween Septemberand
Febmary (Fig. 6). The numberofdetections atdifferent times ofday forall Galapagos
sharkspooled indicated thatthese sharks visitedTrig throughoutthe day, butmore
frequently atnight. At other islands (Tern and Shark), Galapagos sharks also were
detectedmore frequently during nighttime hours (Fig. 6).
Midway
Acoustic Monitoring. The MidwayAtoll Galapagos shark data are skewedby
VRl receivercoverage due to difficulties in getting to MidwayAtoll in orderto download
andrebattery receivers. The batteries in severalVRl receivers deployedin summer
2001 failed in May2002 andwere notreplaceduntil September2002. Onlythree offive
VRl receivers deployed in September2002 were recovered successfullyby USFWS
personnel. The two VRls thatwere lost (Fish Hole, Main Charmel) were historically
the receivers with the most Galapagos shark detections. The combination ofthese
events meant that no data were available forthe heavilyutilizedFish Hole and Chaimel
locations afterMay 2002.
Six Galapagos sharks were detectedbythe array ofunderwaterreceivers at
MidwayAtoll overperiods ranging from 55 to 749 days (Table 2). Based on detections
atreceivers spread across the atoll, sharks were detected atreceivers ranging from 1 to 9
287
km apart. The movements ofall six sharks overlapped, with each individual being most
frequently detected at the Fish Hole and Channel locations (Fig. 7). Five sharks showed a
day-night habitat shift, with four individuals occupying channel and forereefhabitats by
day and venturing up onto the shallow reefflats at night. One Galapagos shark showed
the reverse pattern (arriving in the channel only at night), while the remaining indi\ idiial
did not show any obvious diel periodicity in movements (Fig. 7).
During September2002, fourgiant trevally ranging in size from 100 to 146 cm
FLwere captured using hook and line (trolling and dunking from a boat) at Midway
Atoll (Table 3). Three ofthe fourgiant trevally tagged at Midway were detected by
the array ofunderwaterreceivers at MidwayAtoll overperiods ranging from 280 to
374 days (Table 3). Two ofthese fish had previously been tagged and released by the
Midway sport fishery. Based on detections at receivers spread across the atoll, giant
trevallywere detected atreceivers ranging from 5 to 9 km apart. The movements of
these three fish overlapped, even though they were captured at different locations up to
9 km apart. The one receiver located on the outside edge ofthe atoll was lost (Fish Hole
-Fig. 2b), but the fourremaining receivers each detected at least two giant trevally on
multiple occasions overa 12-monthperiod (Fig. 8). The diel pattern ofdetections varied
among the giant trevally, with onefish (U2792) showing a day-night habitat shift during
2002, whereas the othertwo lacked obvious diel periodicity(Fig. 8). There was also
some seasonal variation in frequency ofgiant trevally detections, with fewest detections
occurring during the wintermonths (Fig. 9).
DISCUSSION
Acoustic monitoring proved to be an effective method forstudying site fidelity
and movement patterns oflarge marine fishes at French Frigate Shoals and Midway
Atoll. This technologyyielded tens ofthousands ofdetections oftransmitter-equipped
animals, which provided new insight into both general patterns ofbehaviorand distinct
behavioral differences among individuals and among species oflarge fishes at these
locations. Forexample, previous anecdotal observations oftiger sharks at French Frigate
Shoals suggestedthattiger sharks dramatically increase in abundance during summer
andwere perhaps only seasonal visitors to this atoll (Tricas et al., 1981; Lowe et al.,
1996). However, acoustic monitoring data from 13 tagged tiger sharks indicated that at
least 70% ofthese sharks exhibited some degree ofyear-round residence at FFS overa
3-yearperiod. Although some tiger sharks were detected at islands within FFS during
every month ofthe year, many were not detected foras long as 2-month intervals. While
it is possible thatthese individuals could have traveled to neighboring atolls orshoals
during these periods, it is also possible that they simply moved to otherareas in or around
the atoll where there was no receivercoverage. Some ofthe individuals tagged at FFS
were detectedby acoustic receivers at Midway and offthe Kona coast (on the Island
ofHawaii), indicating that individual tiger shark movements can encompass the entire
Archipelago.
Even though tigersharks were detected atFFS throughouttheyear, therewas a
strong seasonal trend in area use through the atoll, with tiger sharks spending more time
288
around East Island in the summermonths, but more time aroundthe northern islands
(Tern, Trig, and Shark Islands) in wintermonths. The one tiger sharktagged at Midway
Atoll (#019) in July 2001 was detectedneartheflats offEasternIsland andnearthe cargo
pieronly during summermonths.
Atotal of38,886 detections were recorded from all receivers placednearsix
islands atPES. The estimated total acoustic detection area ofall 10 acoustic receivers was
approximately 0.031 km-, which accounts for less than 0.004% ofthe shallow lagoon
habitat at FFS. Considering the vast areaofavailable habitat fortiger sharks at FES
and the small detection areas ofacoustic receivers in these shallow reefareas, the high
numbers ofdetections clearly indicate that tiger sharks regularly visitthese islands, in
response to concentration ofimportantprey items at particular islands during summer
months.
Compared to tiger sharks, there is a much smalleramount ofdata available for
analysis ofmovementpatterns ofGalapagos sharks at EFS. Eurthermore, the presence
ofthese sharks at Trig Islandvariedwithin the diel cycle, within annual cycles, and
among individual sharks. Although only fouradult Galapagos sharks were caught and
tagged at EES, acoustic receiver data and visual observations bymany researchers at
EES suggest that Galapagos sharks are most common at islands close to the outerreefof
EES (i.e., Tern, Trig, and Shark) and are not frequentvisitors to the interiorofthe atoll.
This contention is supported byprevious studies which indicate that Galapagos sharks
are typically found along outerreefdrop-offs (DeCrosta et al., 1984; Wetherbee et al.,
1996). Galapagos sharkswere the most common species oflarge shark observed atTrig
Island, possibly attractedby the recent increase in seasonal monk seal pupping atthis site.
Adult Galapagos sharks have been observed cruising very close to the shore (< 2 m) and
occasionallypreying onpre-weanedmonk seal pups atthis location (BakerandJohanos,
2004). Acoustic monitoring indicated high variability in Galapagos sharkactivity atTrig
Island, but these data were primarily derived from only two individuals that each showed
differentpatterns ofactivity aroundTrig. One sharkwas most commonly detected inthe
late afternoon during summermonths, whereas the otherwas most commonly atTrig
during early moming hours in winter. Clearly, more research is requiredto understand
the behaviorofadult Galapagos sharks atTrig Island, andtoprovide sufficient data
for assessing the potential success ofusing shark culling to reduce seal predation.
Nevertheless, itappears that Galapagos sharks do not exhibitthe same islandvisitation
patterns as tiger sharks.
The Galapagos sharks tagged at Midway exhibited different movementpatterns
from those tagged at EES; however, this maybe attributedto differences in size/age
ofsharks tracked. The lagoon and main channel at Midway contained large numbers
ofjuvenile Galapagos sharks, which were not observed orcaught at EFS. Thejuvenile
Galapagos sharks at Midway tended to use the channel areas or forereefduring the day,
but would venture onto flats inside the atoll at night, and some ofthese small sharks
moved at least 10 km between acoustic receivers. Considering the arbitrary positioning
and limited numberofacoustic receivers throughoutthe atoll, the numberofdetections
and individual sharks detected suggest that these young Galapagos sharks move
extensively throughoutthe lagoon habitat at Midway. The differences in Galapagos shark
289
movements and habitat use at FFS and Midway may be related to the ditTcrent size of
sharks. Forexample, in some locations Galapagos sharks use shallow lagoons as nursery
grounds (Kato and Carvallo, 1967) and in the Main Hawaiian Islands Galapagos sharks
segregate by size and sex, but do not appearto use lagoon nurseries (Wetherbec et al.,
1996).
Three ofthe fourgiant trevally equipped with acoustic transmitters at Midway
Atoll were detected by fouracoustic receivers spread across the southern portion ofthe
atoll. Only one ofthe three giant trevally detected at Midway showed any diel pattern
ofarea use; however, all three were found to span at least 10 km between the most
distant receivers. Interestingly, the one trevally that exhibited a diel pattern ofhabitat
use (U2792) exhibitedthat behavioronly forthe first few months. Fish were typically
detected on the flats by Eastern Island or Frigate Point at night, sometimes for many
hours. These observations suggest high plasticity in behavior. Otherfish have been shown
to exhibit diel-habitat shifts, includingbluefin trevally (Caranx melampygiis) andjuvenile
giant trevally in the Main Hawaiian Islands (Holland et al., 1996; Wetherbee et al.,
2004; Meyerand Honebrink, 2005). Two ofthe gianttrevallydetected at Midway were
most common during summerand fall months, but decreased substantially in the winter
months. It is unclearwhether these fish left the atoll during winter ormoved to locations
atMidwaythat lacked receivercoverage. This sort ofseasonal shift in habitat use has
notbeen seen inyounger size classes studied in the Main Hawaiian Islands (Wetherbee
et al., 2004). Nevertheless, seasonal differences in watertemperature between the Main
Hawaiian Islands and Midway may explain these possible seasonal area use patterns
observed among the few giant trevally monitored.
We demonstrate that acoustic monitoring can provide an effective method for
assessing long-term sitefidelity andbehavioroflargefishes inremote areas. Obviously,
more detailed information about movement patterns and habitat use could have been
obtained ifthere were agreaternumberofreceivers spreadthroughout each atoll;
however, the main focus ofthe studies at FFS and Midwaywasto examine shark and
trevally affinity to islands thathold large numbers ofsemi-terrestrial prey. Extensive
fishing, tag andrecapture, andvisual observations conducted continuously overmany
years would havebeen requiredto answerthis question, resulting in amuch highercost
and impactto the environment. While acoustic monitoring provides a far less labor-
intensive method formeasuring site fidelity and movementpatterns oflarge fishes
in remote areas, it still requires a certain degree ofmaintenance to ensure successful
retrieval ofdata. Autonomous acoustic receivers must be periodically downloaded, and
batteries mustbe replaced. Securing ground tackle also needs to be maintained annually,
particularly in areas exposed to high surf. Although this maintenance does not take
long and canbe done by small crews, the remoteness ofthe NWHI makes regulararray
maintenance challenging, as was seen at MidwayAtoll where we were unable to place
personnel to regularly maintain receivers. This resulted in loss ofdata and a receiver.
In addition, autonomous acoustic receivers have the capacityto record and store large
amounts ofdata, which, overtime, requires extensive database management.
With amoderate fishing effort, hundreds oflarge marine apex predators (fishes,
sharks, seals, andturtles) could be tagged, and acoustic receivers could be placed
290
strategically around each ofthe majorislands and shoals throughouttheNWHI to assess
long-temi site fidelity, dispersal potential, and even species interactions. Receiverarrays
canbe maintained quickly andeasilywithmoderate ship support. In fact, the newest
foiTn ofautonomous acoustic receiver (VR3, Vemco Ltd.) now incorporates atethered
surface transmitterthat can relay stored datato a satellite orvia acoustic modem to a ship,
eliminating the needto retrieve andmanually downloadthe receivers. Because ofthe
logistical challenges ofaccess to theNWHI, potential conflicts with endangered species,
and difficulty in studying large marine fishes, acoustic monitoring coupledwith satellite
telemetry may provide the mostcost-effective, environmentally sound means ofstudying
the apex predators ofthe NWHI.
ACKNOWLEDGEMENTS
Conducting studies ofthis nature requires seemingly endless permits and red
tape, and cutting through this process requires the aid ofnumerous individuals at several
different agencies. We thank B. Antonelis (National Marine Fisheries Service), R.
Shallenbergerand B. Flint (USFWS), and the dedicated refiage managers B. Allen and
T. Palermo stationed at FFS and those at Midway fortheirassistance overthe years.
We thank D. Topping,Y. Papastamatiou, J. Vaudo, D. Cartamil,A. Bush, K. Duncan,
andK. Holland fortheirassistance in the field. Funding and support forthese studies
were providedby National Geographic Society (#6577-99), National Marine Fisheries
Service, U.S. Fish & Wildlife Service, and Midway Phoenix Corp. All animal procedures
were approvedby the California State University Long BeachAnimal Care andWelfare
Committee (protocol #142 & 187). This workwas conductedunderU.S. Fish & Wildlife
permits (#12521-00014, #12521-01007; #12521-02020; #12521-03015).