Diel trends in the mesopelagic biomass community of the Northwestern Hawaiian Islands observed acoustically

2006·7.5 MB·English

by  Marc O Lammers

DIELTRENDS IN THE MESOPELAGIC BIOMASS COMMUNITY OFTHE
NORTHWESTERN HAWAHAN ISLANDS OBSERVEDACOUSTICALLY
BY
MARC O. LAMMERS', RUSSELL E. BRAINARD-,AND WHITLOW W.L. AU'
ABSTRACT
The nighttime mesopelagic biomass occurring on and near six banlcs in the
NorthwesternHawaiian Islands was investigatedusing a ship-based EK60 scientific
echosounder. The locations investigated included French Frigate Shoals, Maro Reef,
Lisianksi Island/Neva Shoals, Pearl and HerniesAtoll, KureAtoll, and MidwayAtoll.
Surveys were designed to sample parallel and/orshore-normal at each site during
differenttimes ofthe night and during the day. Astrong diel trend exists in the presence
ofmidwatersound-scattering biota at all six locations visited. Dense communities of
organisms accumulate on the edges ofeach island and the associated banks at night. The
highest densities oforganisms tendto occur in waters 30 meters ordeeper, but significant
increases in biomass were also observed at shallowerdepths. There was considerable
temporal and spatial heterogeneity in the occurrence ofthe biota observed both between
and within locations sampled. The biological composition ofthe observed biota is
presently unclearbut it resembles the mesopelagic boundary community that occurs in
neritic waters offthe Main Hawaiian Islands. The nightly influx ofthis biota into shallow
waters is likely a significant, thoughpoorlyunderstood, component ofthese islands' reefs
and nearshore ecosystems.
INTRODUCTION
Sound-scattering layers (SSLs) are communities oforganisms composed of
various combinations ofzooplankton, planktonic larvae, and micronekton. SSLs are
found in manyparts ofthe world's oceans and are characterized by a diel vertical
migration from daytime subphotic habitats into surface waters at night. Vertically
migrating SSLs are an importanttrophic link in pelagic food webs because they promote
a downwardtransferofenergy from epipelagic waters into the deeper(>500 m),
mesopelagic layers ofthe ocean (Rogerand Grandperrin, 1976).
Inthe Main Hawaiian Islands (MHI), an island-associated SSL occurs that is
known as the Mesopelagic Boundary Community (MBC) (Reid et al., 1991). This is
acommunity ofmicronekton specifically adaptedtotheboundaryregionbetweenthe
neritic and oceanic habitats. This community is made up ofat least 23 species offish.
'HawaiiInstituteofMarineBiology,P. O. Box 1106.Kailua,HI 96834USA, E-mail: lammers(a;ha\vaii.edu
-NOAAPacificIslandsFisheriesScienceCenter, 1125BAlaMoanaBoulevard, Honolulu, HI 96814USA
392
MBC
shrimp, and squid in 12 families. During the day, the is found in waters 400-
m m
700 deep along the slopes ofthe islands, while atnight it rises to within 10 ofthe
surface.
Recent work on the MBC has demonstratedthat, in addition to migrating
vertically atnight, this community also moves horizontally towards shallower inshore
waters (Benoit-Bird et al., 2001). This net diagonal movement begins shortly before
MBC
sunset and reaches its shallowestpoint at aroundthe midpoint ofthe night. The
then reverses its movement so that it is back in deeperoffshore waters by sunrise. The
shallowest depth reached by this migration is presently still unknown, but acoustic
MBC m
observations have detected the in waters with abottom as shallow as 30
(Benoit-Bird andAu, 2004), and it is thoughtthat itenters even shallowerdepths.
The influx ofboundary community biomass into coastal waters on anightly
basis remains apoorly understood component ofHawaii's neritic habitat. Stomach
analyses oftuna (He et al, 1997), billfish (Skillman, 1998), bottomfish (Haight et al.,
1993), and spinnerdolphins (Norris et al., 1994; Benoit-Bird, 2004) have shown that
boundary communitypreyrepresent an important component oftheirdiets. In addition,
MBC
the occurrence ofthe in waters shallow enough to overlap with coral reefs raises
thepossibilitythat a significanttrophic relationship may also existwiththis community
between these two communities.
Reid et al. (1991) suggested thatboundary communities are likely to occur
globally in regions where land-associatedmesopelagic species are found and that,
consequently, an important, but still poorly understood ecological relationship exists
between oceanic and island-associated near-shore habitats. To examine this possibility
overabroadgeographic scale, the occurrence ofmesopelagic biomass nearislands and
atolls in theNorthwestern HawaiianArchipelago was investigatedusing ship-based
echosounders. The objectives ofthis work were to establish whetherdiel migrations of
biota into neritic waters are as common around the atolls oftheNorthwestern Hawaiian
Islands (NWHI) as they are aroundthe main Hawaiian Islands (MHI), and secondly,
whetherthis nightly influx overlapswith coral reefhabitat.
METHODS
The occurrence ofthe MBC and othermesopelagic biota was investigated
acousticallyusing two Simrad EK60 echosounders operating at 38 kHz and 120 kHz.
Surveys were conducted using theNOAAship OscarElton Sette during the 2003 NWHI
ReefAssessment and Monitoring Program (RAMP) cruise, between 12 Julyand 17
August. Both EK60 frequencies were setto operate atthe maximum ping rate relative
to the detectedbottom, apulse duration of0.256 ms, atransmit powerof 1000 Watts, a
beam angle (-3 dB) of7.1 degrees, and atransducergain of24 dB at 38 kHz and25.1 dB
at 120 kHz. Both sounders were calibrated in Septemberof2004 andagain inMarch of
2005. Calibration values remained consistentwithin 0.5 dB.
Six locations in the NWHIArchipelago were examined. These were: French
Frigate Shoals (N23°45'N latitude, W166°10'W longitude), Maro Reef(N25°25'N
393
W
latitude, W170°35' longitude). Neva Shoal/Lisianski island (N26"04'N laliUidc.
W173^'58' W longitude). Pearl and Hermes Atoll (N27''50' N latitude. W175"S()" W
W
longitude), MidwayAtoll {N28"12' N latitude, Wi77"22' longitude), and KureAtoll
(N28°25'N latitude, W 78"2()" W longitude). Acoustie surveys were conducted during
1
a 6-hourtime window at night and a 1-2 hour opportunistic time window during the
day. Each location was surveyed overa period ofeither 2, 3, or4 days and nights. A
set of2 to4 systematic transect lines ranging in length from 3 to 6 nautical miles were
acoustically sampled at a speed of5-6 knots three times during the night (Fig. 1 ). The
placement oftransect lines was balanced between the study's objectives, cruise logistics
relatedto daytime diver-based operations, and local weatherand sea conditions.
Differenttimes relative to the middle ofthe night were examined to establish whether a
net movement ofbiomass across and/oralong transects took place. The middle ofthe
nightwas definedas the halfwaypoint between sunsetand sunrise, the time ofwhich
changed with increasing longitude. Each transect was also sampled once during the day
as cruise logistics permitted to give a day/night comparison. In addition, to provide a
comparison between data obtained in the NWHI and the boundary community known
to occur in the MHI, a single nighttime transect along the leeward coast ofthe island of
Oahu was conducted during a separate ciiiise on 9April, 2004 with the same vessel.
Data were analyzed using Echoview 2.25. To examine the relative abundance and
distribution ofbiomass between and along transect lines, the sampled watercolumn was
m m
divided into cells 100 long by 5 or 10 deep (deeperwaters were divided into deeper
cells). Twenty percent ofall the cells from each transect down to a depth of 180 m were
randomly selected as the basis for statistical comparison between times. Cells deeper
m
than 180 were excluded from the analysis due to the presence oftime-varying gain-
related noise with increasing depth.
The mean volume backscattering strength (Sv) foreach cell was used as a relative
measure ofbiomass (Throne, 1971; MacLennan and Simmonds, 1992). Changes in Sv
values were used as indicators thatthe total biomass and/orthe relative composition of
biomass had changed overtime and space. Larger(less negative) Sv values are indicative
ofan increase in biomass density, a shift in the species composition towards those
with highertarget strength, oran increase in target strengths due to changes in animal
orientation orswimbladdervolumes, ora combination ofthese (Deemerand Hewitt,
1995) (Fig. 2). To represent the relative occurrence ofbiomass as a function oftime,
depth, and location along transects, each transect was divided into 3-5 segments, and the
cells for each segmentwere averaged into depth bins of5, 10, or 30 m.
Priorto calculating Sv, the data were visually inspected and pre-processed using
Echoview's data exclusion utilityto remove extraneous noise artifacts, such as false
echoes arising fromwaterturbulence related to the ship's motion. In addition, the top 10
m
ofeach transectwas rejected from the analysis to avoid the confounding influence of
m
wave-induced surfacebubbles. Volume backscatterwas calculated only forwaters 2
ormore above thebottom.
394
RESULTS
In Table 1 we describe the statistical relationship between the daytime and
nighttime occurrence ofmid-waterbiota at the six locations surveyed. In all cases,
more biomass occurred in the watercolumn at nightthan during the day. There was
considerable variability inthe diel occurrence ofbiomassboth temporally and spatially.
This was the case within as well as between locations. More backscattered acoustic
energy was consistently receivedwith the 38- kHz echosounderthan the 120- kHz
system. Consequently, the summaryfindings detailedbelow foreach locationreflect only
the 38- kHz data. Acomparative analysis ofthe results obtainedusingboth frequencies
will be the subject ofa future publication.
French Frigate Shoals
Three acoustic transects were conducted at French Frigate Shoals (FFS) (Fig.
1A). TransectAwas adjacent and parallel to the reefflat, transect B was centrally
located on the main bank ofthe shoals, and transect C was placed parallel to the slope
ofthe bank. The diel difference was greatest along transect C, where dense layers of
biota accumulatedthroughoutthe night (Table 1). The relative diel difference was
approximately equal alongtransectsAand B, butAhad a greaterabsolute density of
biomass during both daytime and nighttime.
Forfurtheranalysis, eachtransect was divided into five equidistant 1.9- km
segments. Table 2 reveals where biomass occurred as a function ofdepth and time.
Along transectsAand B, the occurrence ofnighttime biota increased throughoutthe
watercolumn, but especially nearthe bottom 5-10 m. Increases were nothomogeneous,
but rather occurred in localized maxima or 'patches.' The densestpatches alongboth
transects occurred during theperiodpreceding the middle ofthe night (2200h) and began
to dissipate by 0300h. The increases inbiotaobserved alongtransect C differed inthat
patches ofbiomass occurred as localized layers in the water column. Between two
and three distinct layers occurred simultaneously during the early (2200h) andmiddle
(0030h) periods ofthe nightbetween 50 and 150 m. During the late period (0300h),
distinct layers were still presentbut occurred deeper. Throughoutthe night, the densest
aggregations occurred where a layerwould come into contact with the bottom along the
edge ofthe slope (Table 2, transectC, segment 'Edg').
To determine whethernoctumally present biotamigrate horizontally from
the slopes ofthe bank onto the shallows nearthe reefflat, we consideredthe relative
occurrence ofbiomass inrelationto the time ofnight. We expected that, ifhorizontal
migration across the banktakes place through the night, two roughly equivalent local
maxima ofrelative abundance would occuralong transects B and C during the firstand
third quarters ofthe night (2200h and 0300h, respectively), and a local maximumwould
be observed along transectAduring the midpoint ofthe night (0030h) (Fig. 3A). This
was notthe case, however(Fig. 3B). Asimilaranalysis ofrelative biomass occurring
along (ratherthan across) eachtransect as a function oftime also did not match the
predictions oflarge- scale horizontal movement, at least not withinthe time frame
examined.
395
Maro Reef
Maro Reefwas surveyed over the course oftwo days and nights, during uhich
two shoal-normal transects were systematically sampled (Fig. B). Both transects
1
initiated adjacent to the shallow reefflat and extended past the slope ofthe bank. For
analysis, each transect was divided into five segments based on depth and distance from
thereefflat: three 'shallow- bank' segments (length= 2.1 km for transectA, 2.9 km
fortransect B), an 'edge- of- bank' segment (length = 1.6 km for transectA, 1.2 km for
transect B), and a 'slope- of- bank' segment (length = 1.5 km for transectA, 0.8 km for
transect B). Biomass increases occurred throughout the length ofboth transects, but
the highest densities accumulated on the 'edge- of- bank" segment in water between 30
and 90 m deep (Table 2). The layerwas densest between 30 and 60 m deep where it
impinged on the rising slope ofthe bank, but it extended well onto the bank along the 1
m closest to the bottom. This distribution pattern was relatively consistent throughout
the three nighttimeperiods sampled, suggesting that only limited, ifany, net horizontal
movement normal to the reeffiat tookplace within the time frame examined.
Neva Shoal/Lisianski Island
Three transects parallel to the reefflatwere sampled atNeva Shoal overthree
days and nights (Fig. IC). As at FFS, transectAwas adjacent and parallel to the reef
flat, transect B was located centrally on the main bank ofthe shoal, and transect C was
placedparallel to the slope ofthebank. Significant nightly increases in biomass were
measured on transectsAand C (Table 1). Transect B was not sampled duringthe day due
to operational restrictions with the ship.
The difference in daytime vs. nighttimebiomass densitywas considerably greater
along the slope ofthebank than nearthe reefflat. As was observed at FFS, there were
predominant increases in biomass towards the bottom halfofthe water column nearthe
reefflat(transectA), comparatively lessbiotaalong the middle ofthe bank (transectB),
m
and adistinct layering ofbiomass centeredbetween 30 and 90 deep along the slope
ofthe bank (transect C). As at Maro Reef, the layers found along the slope were densest
where they impinged on the rising slope ofthe bank. An examination ofthe occurrence
ofbiomass between and along transects in relation to the time ofnight, as described for
FFS, also didnotyield any clear evidence ofnet horizontal movement across the bank
within the time frame considered. During the daytime, most ofthe remainingbiota
occurred inthe middle ofthe water column, towards the southern ends ofboth transectsA
and C.
Pearl andHermesAtoll
Pearl andHermesAtoll was surveyed during four days and nights. Cruise
logistics and favorable weatherallowed fourtransects to be conducted on three sides
oftheAtoll (Fig. ID). TransectsAand B were shore-normal onthe northeastern and
southwesterncomers oftheAtoll, respectively. Bothtransects initiated adjacent to the
396
shallowreefflat and extendedpastthe slope ofthe bank. Foranalysis, each transectwas
m
divided into five 1,400- segments representing different depth strata and distances from
the reefflat. These were labeledusingthe same nomenclature employed at Maro Reef
Transects C and D were both on the southern side oftheAtoll, parallel to shore and
to one another. Transect C extended overa long segment ofboth declining and inclining
m
slope, dropping to a depth ofapproximately 650 in between and leveling offinto abank
on the western end. Foranalysis, transect C was divided into fourdepth strata: a 'slope-
of- bank' ('Slo') segment, an 'edge- of- bank' ('Edg') segment, andtwo 'shallow- bank'
('Shb') segments. Transect D was offshore ofC andwas mostly overwatergreaterthan
1,000 m deep. Foranalysis, itwas divided into a 'deep- water' segment, a 'slope- of- bank'
segment, and an 'edge- of- bank' segment.
TransectAhad the lowest nighttime Sv values ofthe four, but exhibited adistinct
m
layerofbiomass centered atthe 31-60- depthrange throughoutthe night (Table 2). This
layeroccuiTed along the entire transect, butwas densestmid-wateralong the 'edge'and
'slope' segments. Asecond, more localized layerwas associated with the bottombelow
approximately 60 m, primarily along the middle 'shallow- bank' segment(Table 2, transect
A, Shb2).
Transect B exhibited a similardistribution pattern as transectA, butwith a
considerably higherdensity ofbiomass along the shallowesttwo segments ofthe bank
(Table 2, transect B, Shbl and Shb2). In addition, the dense patch ofbiota occurring along
the middle 'shallow- bank' segment (Shb2) persisted into the day, although it disassociated
itselffrom the bottom andbecame concentrated in a layercentered approximately 20 m
above the bottom.
Adense layer ofbiota centered between 31 and 60 m deep occurred along the length
oftransect C. This layerwas densestnearthe bottom ofthe 'edge' andfirst 'shallow-bank'
('Edg' & Shbl) segments during the middle ofthe night. The layer scattered somewhat and
descended deeper as the night wore on, but a notable density ofbiomass remained inboth
&
'shallow- bank' segments (Shbl Shb2) during the pre-dawn hours and persistedthere
during the day.
Transect D was dominatedby deeperwaters than transectsA, B and C. However, as
m
with the otherthree transects, a layerofbiota centered between 31 and 60 deep occurred
there during the majority ofthe night. Also consistent with the otherthree transects was the
higherconcentration ofbiomass at the lowerdepths ofthe 'edge' segment. In contrastwith
transects C and B, however, low densities of'edge'-associatedbiota remained duringthe
day.
MidwayAtoll
Two transects parallel to the southern slope ofMidwayAtoll were sampled
during two days and nights (Fig. IE). TransectAextended from the centeroftheAtoll's
southwesternbankto nearthe entrychannel into the lagoon. Thebottom alongthis transect
gradually sloped upward from a maximum depth of97 m on the western endto aminimum
depth of46 m on the eastern end. Transect B ran parallel toAalong the edge and slope of
theAtoll. The depth along transect B varied widely between 423 m and 91 m.
397
TransectAexhibited a nocturnal increase in biomass througiioul the water
column, but especially towards the eastern end below 20 m. The abundance ofbiota
remained high throughout most ofthe night and began lo decrease prior lo sunrise
(0530h). It persisted the longest towards the eastern end ofihc transect. \\hich was
adjacent to a steeper slope than the western end and was therefore characterislic ofthe
'edge' bathymetry described forother locations in theArchipelago.
Transect B differed in the distribution ofbiomass between the western and eastern
end. The western end was characterized by a distinct biomass layer near the surface and
an accumulation nearthe bottom, separated by low densities in the middle ofthe water
column, particularly atOBOOh. On the eastern end, the surface layerbecame denser and
reached deeper, but no accumulation nearthe bottom was observed. Toward sunrise,
the distribution pattern changed considerably, with the bulk ofthe biota occurring at the
lowerdepths ofthe western end, most likely representing the downward phase ofthe diel
migration cycle.
KureAtoll
KureAtoll was surveyedduringtwo days andnights. Two transects were sampled
parallel to theAtoll's western slope (Fig. IF). TransectAwas the shallowerofthe two
with nearly homogeneous depths. Transect B was parallel toA, approximately 2.3 km
furtheroffshore. The bottom oftransect B sloped upward on the northern end, but was
roughly constant in depth towards the southern end.
Anocturnal increase in biomass occurred throughoutthe watercolumn along
transectAduring the middle ofthe night (0030h) and gradually decreased in density
as the night wore on, particularly along the bottom halfofthe watercolumn (Table 2).
TransectB was characterizedby dense aggregations below approximately 100 m and a
secondary, more diffuse layertowards the surface. The densest patches observed along
transect B occurred during the latterpart ofthe night (0300h) along the northern end. A
distinctpatch persistedthere into the lastphase ofthe night(0530), but was entirely gone
by daytime.
Waianae, Oahu
Asingle 6.3- km transect was conductedparallel to the northern Waianae coast
ofOahuduring the middle ofthe night in waters between45 m and 120 m deep. The
average volume backscattering strength calculated forWaianae was nearthe median
ofthe distribution ofall theNWHI transects forboth frequencies (Table 1). The
concentration ofbiomass was patchy, with a distinct mid-water layeroccurringtowards
the northern end and amore bottom-associated layertowards the southern end. The
highestobserved densitywas found along the edge ofa descending slope towards the
southern end. The density and distributionpattern ofthe biota encounteredoffWaianae
was not distinct in anynotable way from the range ofthepatterns observed in theNWHI.
398
DISCUSSION
The study's primary objective was to answerthe question: are die! migrations
ofbiota into neritic waters common in the NWHI? We indicate that they are. Increases
in nocturnal mid-waterbiomass were noted at all locations and along each transect
surveyed. However, considerable spatial andtemporal heterogeneity characterizedthe
occurrence ofthis biota. Each site exhibited localized maxima in densities thattended
to peak during the middle ofthe night and gradually subside priorto sunrise. During the
day, most locations, regardless ofdepth, exhibited a substantial decrease and even total
absence ofthe sound-scatteringbiota observed atnight.
Although there was much variability, certain spatial patterns in the occurrence
ofthis nighttime biota did emerge. The most consistent and dense aggregations were
observed on andnearthe edges ofthe slopes ofthe atolls and shoals visited. Theband
ofwaterbetween 30 m and 90 m deep nearly always had one ormore distinct layers
associated with it, usually throughout the night. These layers typically extended well
beyond the slope, both offshore and towards the shore orreefflat. Interestingly, the
layers often had well-defined upperboundaries, usually below 20 m deep. This maybe
tied to avoidance oflightreflected from the moon, which can leadto greaterpredation
(Gliwicz, 1986; Gal et al, 1999). Conversely, there appearedtobe no avoidance ofthe
benthos, although this changed during the day when, on the few occasions where a layer
did persist into daytime hours, there was always a clear separation from thebottom (e.g.,
Pearl & HennesAtoll transect B).
The second majorobjective was to determine whetherthe migratorybiota
observednearthe slopes ofatolls enters coral reefhabitat. This point remains
unresolved. There was clearly anocturnal influx ofbiota into the watercolumn at sites
with depths commonly associated with coral reefhabitat (-20-40 m), such as transectA
atFrenchFrigate Shoals (FFS). However, the data obtained didnotreveal an identifiable,
horizontallymigrating 'front' oforganisms that might account forthis biomass, as
has been observed in the MHI (Benoit-Bird et al 2001; Benoit-Bird andAu, 2004).
Therefore, we presently cannot exclude the possibility that at least some ofthe biota
observed arose fromwithin ornearthe bottom locally. Benoit-Bird andAu (2004) have
reported that the average horizontal migratory rate ofmicronekton offOahu is 1.7 kmh'.
So, it is possible that biota observed overreefhabitats within 2-3 km ofan atoll's slope
migrated there from deep waters quickly following sunset, and therefore did not appear
as a moving front during the time frame we sampled. However, this is unlikely forthe
interioroflarge banks such as FFS andMaro Reef, where the reefflat is more than 10 km
from thebank's slope.
Regardless oforigin, the finding ofconsistently highernocturnal biomass
densities overreefhabitat is importantbecause it suggests that traditional daytime
biological assessments may not capture all the trophic relationships present on the reef
andmayunder-represent certain groups. This is relevant to efforts aimed at creating
ecological models forthe NWHI. Forexample, Friedlanderand DeMartini (2002)
reportedthat over 54% ofthe total fish biomass observed on reefs in theNWHI consists
ofapexpredators, raisingthe intriguing question ofhow so manytop-level consumers
399
are trophically supported. A part ofthe answer may lie with the noctunial iiitlux ofbiota
reported here.
Another unresolved issue is the ta.xonomie makeup ofthe biota obser\ed.
Logistical restrictions did notallow trawling forsamples during acoustic data collection.
Consequently, we can say relatively little about the identity ofthe organisms that occur
atthe various locations sampled. The fact that more backscattered acoustic energy
was consistently receivedwith the 38- kHz echosounder (vs. 120-kHz)than 120 kHz
is suggestive ofmicronekton ratherthan zooplankton, since the smallerzooplankters
wouldbe expected to reflect more acoustic energy at the higherofthe two frequencies
(MacLennan and Simmonds, 1992). In addition, the factthat the limited data obtained
offthe Waianae coast ofOahu fell in line with mean Sv values from the NWHl further
points to biota related to the mesopelagic boundary community.
In summary, it is reasonable to conclude that the nocturnal composition of
biota in neritic waters offatolls and islands in the NWHIArchipelago is substantially
different fromwhat is observed there in the daytime. This should be carefully considered
when planning ecosystem assessments ortrying to model trophic relationships based
on observed biomass. To betterunderstandthe ecological importance ofthese diel
migrations, future surveys will need to resolve questions aboutthe biological makeup of
biota at different sites. In addition, engaging in long-term monitoring ofmigration trends
andcorrelatedoceanographic conditionswillyield important insights into the dynamics
ofthese communities and possibly provide information on long-term patterns in the
health ofneritic ecosystems.
ACKNOWLEDGEMENTS
Wewould like to thankthe officers and crew ofthe OscarElton Settewho made
every effortto accommodate the scientific needs ofthe study. Phil White in particular
provided invaluable assistance. An anonymous reviewerprovidedhelpful commentary
on an earlierdraft ofthe manuscript. This is HIMB publication #1201.
400
-m
1^^^^^^^^^^
D
,:|P^
P^
1
4
W
B
Southeast
Island
Seal-Kittery
Island
Figure 1. AcousticallysampledtransectlinesatFrenchFrigate Shoals(A),MaroReef(B),LisianskiIsland'T^levaShoal
(C),PearlandHenTiesAtoll (D),MidwayAtoll(E),andKureAtoll(F).