Mega- to micro-scale classification and description of bottomfish essential fish habitat on four banks in the Northwestern Hawaiian Islands

2006·5.6 MB·English

by  Christopher Kelley

MEGA-TO MICRO-SCALE CLASSIFICATION AND DESCRIPTION OF
BOTTOMFISH ESSENTIAL FISH HABITAT ON FOUR BANKS IN THE
NORTHWESTERN HAWAIIAN ISLANDS
BY
CHRISTOPHER KELLEY', ROBERT MOFFIir. and JOHN R SMITH'
ABSTRACT
We coupledmultibeam sonardata with submersible and remotely operated vehicle
(ROV) observations to classify and describe bottomfish essential fish habitat (EFH)
on fourbanks in theNorthwestern Hawaiian Islands Coral ReefEcosystem Reserve
(NWHICRER). From 2001 to 2003, a total of22 PiscesIVand Fdives along with 37
RCV-150ROVdives were conducted on Raita Bank, W. St. Rogatien Bank, Brooks Bank,
and Bank 66 to evaluate the impacts ofbottomfishing on these banks. In theprocess of
addressing that issue, extensive datawere collected on the biological communities and
substrate characteristics within the EFH depth range of 100 to 400 meters. Multibeam
mapping was conductedbetween dives from the submersible support ship ""KOIC as
well as during a separate cruise onthe RVKilo Moana.All fourbanks hadrelativelyflat
featureless tops (i.e., <5 % slopes) which extended down to a depth of120 m. ROV dives
revealed that the areabetween 100-120 m was characterized by sediment interspersed
with rhodoliths and carbonate outcrops. At this depth on Raita, W. St. Rogatien, and
Brooks Bank, the slope increased to 25-60 degrees, which continued down to 300-400 m.
The substrate on these slopes was carbonate bedrock interspersed with flats and channels.
Ten sponge, 64 cnidarian, I ctenophore, 49 echinoderm, 15 moUusk, 30 crustacean.
3 tunicate, and 152 fish species were observed during the dives. Adistincttransition
occurredbetween shallow-waterand deep-waterfish families within this depth range that
may be temperature related.
INTRODUCTION
The term EFH was definedby Congress as "those waters and substrate necessary
tofish forspawning, breeding, feeding, orgrowth to maturity" (16 U.S.C. 1802(10).
According to the EFH website maintainedbytheNational Oceanic andAtmospheric
Administration (NOAA), National Marine Fisheries Service (NMFS), "waters" in
the definition refers to the "aquatic areas and theirassociatedphysical, chemical, and
biological properties that are usedby fish." "Substrate" refers to "sediment, hardbottom,
structures underlyingthewaters, and associatedbiological communities," and "spawning,
'HawaiiUnderseaResearchLaboratory, 1000PopeRoad,MSB303,Honolulu,HI 96822USA,
E-mail: [email protected]
=NOAAPacificIslandsFisheriesScienceCenter,2570Dole Street,Honolulu,HI96822USA
320
breeding, feeding, orgrowth to maturity" encompasses the full life cycle ofthe fish.
EFH is therefore atermthatblends togetherthe more basic concepts of"habitaf, which
havetraditionallybeenusedto describejustthe physical aspects ofanenvironment,
with "ecosystem", which has been usedto describe the biological communities andtheir
interactions, andthe physical properties ofan environment. The concept ofEFH was
created in an attemptto advance the application ofecosystem-basedapproaches to fishery
management (Park, 2002). To develop an EFH definition foramanagedfish species, the
task is to describenotonly substrate andhydrological features but also the other living
organisms (e.g., fish, invertebrates, and algae) living in associationwiththat species.
Hawaiianbottomfish are agroup offederallymanaged species, most ofwhich
are commerciallyvaluable deep-slope snappers. TheNMFS is presently engaged in
refining its EFH definition forthis fishery, which foryears has been simplythe 100-400
m
depth zone around each island andbankwithinthe HawaiianArchipelago. Studies on
benthic habitats andtheirbiological communities are typically approachedby coupling
seafloormapping withdirect observations and/orbenthic sampling (Greene et al., 1999).
Thebottomfish EFH depthrangeprecludes opticalmappingtechniques and SCUBA,
requiring insteadthe use ofacoustic mapping techniques coupledwithmannedand/or
unmanned deepwatervehicles. The costs associatedwith these types ofoperations
haveprevented examinationofall but a few specific sites. Furthermore, multibeam
mapping anddirect observations havebeen carried out opportunistically andusually in
conjunctionwith othermission priorities. Even so, valuable datahave been obtainedfor
useincreating amore accurate and specificEFH definition forthis fishery. Inthispaper
we initiate the development ofa mega- to micro-scale classification anddescription of
bottomfish EFHbyproviding a summary ofacoustic mapping dataand submersible/
ROV observations obtained onbottomfishhabitats intheNorthwesternHawaiianIslands
(NWHI).
MATERIALSAND METHODS
During September-November, 2001-2003, three cruises were conductedin
NWHICRER onthe Hawaii Undersea Research Laboratory's (HURL) submersible
support ship, Kaimikai-o-Kanaloa {KOK). These cruises hadtwo tasks: a) tomapthe
100-fathom contouraround Raita Bank, W. St. Rogatien Bank, Brooks Bank, andBank
66 to obtain amore accurate position foreach bank, andb) to obtain insitu observations
ofbottomfishfishing sites foruse in evaluatingthe impacts ofbottomfishing onthe
banks. The firsttaskwas carried outwiththeKOK's SeaBeam 210 multibeam sonar
mapping system while the second was carried outwith HURL's manned and unmanned
deepwatervehicles.
Multibeam SonarData
Mega- (1-10 km) andmeso-scale (10 to 1000 m) features ofthebottomfishEFH
on the fourbanks wererevealedfrommultibeam sonardata obtained in conjunction
321
with submersible operations. The SeaBeam niuUibeam system on board the submersible
support ship KOKwas used to map the 100-fathom contour around Raila and W. St.
Rogatien banks between submersible and ROV dives. During this process, a large portion
ofthe bottomfish EFH was covered. These data, which include only bathymetry, were
processed using the freeware multibeam sonarprocessing and plotting packages MB-
System (Caress and Chayes, 1996) and the generic Mapping Tools [GMT] (Wessel and
Smith, 1991). Manual and/orautomatic bathymetric "ping" editing was carried out on the
data to reduce outliers, followed by gridding ofthe swath data collected in various years.
The optimum grid cell size was used for the target waterdepth, usually 10-20 meters,
along withrunning amedian filterofminimum width overthe grids to furtherreduce
noise while maintaining maximum resolution. The data were converted intoASCII grids
and subsequently imported intoArcGIS where they were layered overdigitized NOAA
nautical charts. The charts provided avisual reference for understanding the multibeam
coverage on each bank.
ROV
In Situ Submersible and Data
In situ data within the 100-400 m depth range were obtained during 22 manned
PiscesIVand f'submersible dives and 37 unmannedRCV-I50ROVdives conducted
on the fourbanks. All vehicles were deployed from the KOK. Each 8-hour submersible
dive was conducted during the daybetween 0830-1630 hrs while each ROV dive was
conducted at nightbetween 1900-0200 hrs. During submersible dives, temperature,
dissolved oxygen (DO), and salinity data were obtained from Seabird CTDs mounted on
the vehicles. Macro- and micro-scale geological observations and biological data were
obtained during 30-min transects (fourper dive) designed to obtain quantitative data on
potential bottomfishing impacts (see Kelley et al., submitted forthis volume). Transects
were conducted at different depths (i.e., Tl: 190-210 m, T2: 240-260 m, T3: 290-310 m,
and T4: 340-360 m) during which substrate observations as well as counts offish and
invertebrates were made. These datawere recorded on the audio tracks ofthePisces
digital video camera systems along with the submersible's GPS positions at 10-minute
intervals. The average length ofeach transect was 1 km and the average visual range from
each side ofthe sub was 10 m. Eachtransecttherefore covered an area ofapproximately
2 hectares while each dive covered approximately 8 hectares.
The ROVwas typically deployed to conduct 1.6-3.2 kilometertransects over
selected survey sites. Two trainedobservers werepresent inthe ROV control roomand
taskedwith making substrate observations and identifications offish and invertebrates
encountered. The video along with the audio remarks from the observers were recorded
throughoutthe dives on mini-DV video cassettes. Afterthe dives, observercounts from
the submersible transects were extracted from the videotapes. However, ROVtransect
videos wereprocessed only by following HURL's standard ROV video-logging protocol
that identifies species encounteredduring the dives with only rough quantification.
Light, an additional physical factor, changes considerably within the bottomfish
EFH depth range. Since we are unaware ofany actual light intensity measurements
beingmade onthese banks, theoretical values were derived fromWetzel's (2001)
322
attenuation equation: I = I,, e""^ , where
I = irradiance atdepth z
Ip= irradiancejust below surface (i.e., z = 0)
e = natural logarithm
k = extinction coefficient (0.033 for clear seawater)
NWHICRERwaters are known to be extremely clear, and therefore it was assumed that
the k value used in this equation wouldbe appropriate.
Bottomfish EFH Classification andDescription
Sonar data coupled with substrate observations made from the submersibles and
ROVwere used to describe the geological aspects ofthe bottomfish EFH aroundthe
banks according to the mega- to micro-scale classification scheme designedby Greene et
al. (1999) fordeep-waterbenthic habitats. Hydrological data were analyzed foreach 100
m
interval. Biological data (i.e., algae, invertebrate, andfish observations) were grouped
into taxonomic categories and by abundance.
RESULTS
Multibeam Sonar Data
The multibeam sonarcoverage ofthe EFH around eachbank is shown in Figure
m
1 between black lines. Multibeam data outside ofthe 100-400 depth range from a
2002 Kilo Moanamapping cruise, as well as single-beam sonardata obtained onthe top
ofRaita Bank (courtesy ofJ. Miller), were included in the Raita andW. St. Rogatien
images (see Miller et al, 2004). No EFH boundaries are shown forBank 66, which is
located entirelywithin the 100-400 m depth range. Forsimplification, eachmap provides
a slope analysis whereby green represents lowerand red represents higher slope values.
The tops ofthe banks were generally flat with slope values below 5°. Withthe exception
ofBank 66, all were above 100-m depth. The "break" occurred at approximately 120 m
where slope values increasedrapidly to over25°, and in some locations offRaita, over
60°. Steep slopes continued down to varying depths, however, in general, not below the
lower400-m boundary ofthe bottomfish EFH. Furthermore, the steepest slopes on Raita,
W. St. Rogatien, and Brooks were found on the southwest sides ofthe banks while the
lowest slope values were found on the northeast sides. The top ofBank 66 came up to
m
approximately 120 with the break generallybeginning at 170 m. Slope valuesbelow
m
thebreak to a depth of250-270 were forthe mostpart between 10-20°.At thatpoint,
the slope flattened outto less than 5°, similarto the top.
The multibeam data did not reveal anyparticularly surprising features on the
banks. All fourhad arelatively homogenous structure consisting ofaflattop with a
moderately steep slope inthe bottomfish EFH thatgenerallyflattenedoutbefore reaching
adepth of400 m. The one exception was the presence ofseveral smallpinnacles found
323
within the northern boundai"y ofthe EFH offRaita. These features extended up from the
seafloorapproximately 40-60 m and it is likely that more will be found \\hen the mapping
ofthe EFH in this area is completed.
Submersible Data
The number ofsubmersible and ROV dives conducted on each bank wilhin the
100- to 400-m depth range are summarized in Table 1. Since more than one dive took
place on some sites, the numberofsites examined on each bank also is provided. Data
from submersible, ROV, orboth vehicles, were obtained during a total of59 dives on 28
different sites.
Observations made during the dives revealed that the substrate within the EFH on
all banks consisted ofcarbonate bedrock interspersed with sediment deposits. The latter
were mostly composed ofcarbonate sand and pebbles with smalleramounts ofgravel and
cobbles. Not surprisingly, bedrock was predominantjust below the break where the slope
was the steepest, whereas sediment was predominant above the break as well as deeper,
nearthe lowerboundaryofthe EFHwhere the slope was flatter(Fig. 2). Low amplitude
sedimentwaves were present even where the sand layerwas relatively thin. In these
cases, the underlyingbedrock was clearly visible in the troughs.
Exposed carbonate bedrock clearly had different levels ofcomplexity (i.e.,
rugosity +porosity). Bottomfish, as well as many otherfish species observed, were
typically found in association with high complexity bedrock ratherthan low complexity
bedrockorsediment. Furthermore, porosity (i.e., the number ofholes in the rock as the
term is usedhere) was clearly a more important factorthan rugosity, presumably because
it offered more effective shelteragainst predators.
Asummary ofthe CTD data obtained within thebottomfishEFH onthebanks
as well as the calculated theoretical light intensity values are presented in Table 2. Due
to technical problems, temperature and salinity measurements were only a\'ailable from
15 ofthe 16 submersible dives conducted in 2001 and2002. Furthermore, only the DO
measurements from 9 ofthe 10 submersible dives in 2002 were considered useable.
Within the 100-400 m EFH depth range, both salinity and DO remained relatively
constantat all sites, varying between 34-35 ppt and 5-6 ml/1, respectively. In contrast,
temperature ranged from a high of23°C at 100 m to a low of 10°C at 400 m, while the
theoretical irradiance values ranged between a low of to a high of4,098 klux (4% ofthe
light intensityjustbelow the surface).
Asummary ofthe biological organisms observedwithin the EFH depth range on
these fourbanks is presented in Table 3. Ofthe invertebrates, atotal of64 cnidarian, 49
echinoderm, 30 crustacean, 15 mollusk, 10 sponge, 3 tunicate, and 1 ctenophore species
were recorded during the dives. Examples ofthese are provided in Figure 3. Anemones
(11 species), seastars (22 species), gastropods (10 species), and crabs (11 species)
were the mostdiverse groups ofcnidarians, echinoderms, mollusks, and crustaceans,
respectively. Mosturchins, seastars, andcrustaceans were identified to species; however,
many ofthe sponges and cnidarians were not, due to the difficulty in making accurate
identifications ofthese organisms without close inspection ofspecimens. Clearly different
324
types were noted, such as small white pennatulids vs. large orange ones, whichwere
assumedto be different species. Small branchinghydrozoans were not routinelyrecorded
because in most cases, they couldnotbe distinguished from small dead antipatharians.
Furthennore, the seven different species ofalgae observed during the dives were not
identifiedpast majordivision. Those observed appearedto beprimarily non-attached
fragments which had originated from the tops ofthe banks andwere subsequently carried
down slope. Therefore, these were not consideredto bepart ofthe natural biotawithin
the bottomfish EFH andwere notcarefullyrecorded, although that assumption should
be more thoroughly investigated. Furthennore, the importance ofalgae to thebottomfish
EFHmaybe understated in this study, because locations at ornearthe 100-m upper
boundaiy where naturally growing algae occurwere unden'epresented.
One hundredandfifty-two different fish species were observedwithin the EFH
on the banks representing fifty-nine families (Table 3). Ofthese, serranids (groupers)
were the most specious (12) followed by lutjanids (snappers, 9), labrids (wrasses,
9), scorpaenids (scorpionfish, 7) andmorids (cods, 7). Twenty-one families had only
one representative and includedaberycid(alfonsin), amullid (goatfish), anapogonid
(cardinal fish), an ammodytid (sandlance), and an argentinid (deep-sea smelt).
Two clearpatterns were evident from the fish identifications and countdata.
First, a diurnal-nocturnal shift inthe fish communities on thebanks was detected
within the EFH depth range. The majority ofthe families shown in Table 3 appearedto
be diurnal; however, there were a numberoffamilies thatwere only observed during
ROV surveys atnight. Mostnotable among thesewerethe morids, carapids (pearlfish),
myctophids (lantern fish), trachichthyids (slimeheads), andnettastomatids (duck-billed
eels). Furthennore, most ofthe congrid (congereels) observations were made at night
aswell. Three types ofbehaviors appearedto be responsible forthis pattern. Morids and
the congrid. Congeroligoponis, appearedto remain in the EFH during the day, hiding in
holes in the rocks until nightwhentheypresumably emerged to feed. In contrast, other
congrids, such asAriosoma marginatus, also hid during the daybut by diggingburrows
inthe sediment instead. The nettastomatid, Saurenchelysstylurus, was enigmatic since
these fish neverwere observed during the day and only observed on sediment substrates
atnight. Unlike the burrowing congrids, this species was not observed digging in
response to the approach ofthe ROV, and, furthermore, ithas adelicate caudal fin that
does not appearto be well adapted forcreating burrows. Third, it is well known that
many myctophids undergo a daily vertical (i.e., from further down the slope) and/or
lateral (i.e., from furtheroffshore) migration atnight. It isbelievedthatthese fishmost
likely leave thebottomfish EFH, orthatportion close to the substrate, duringthe day and
return each night.
The second pattern was a shift in the families observedbetweenthe upperand
lowerboundaries ofthe EFH, clearly indicatingthis depth range is the majortransition
zonebetween shallow and deep-waterfish species. The depth ranges observed onthe
banks for39 ofthe 59 families are shown in Figure 4.Acomplete change takes place
between 100 and 400 meters with the upperend ofthe EFH dominated by shallow-water
families such as acanthurids (surgeonfish), chaetodontids (butterflyfish), pomacentrids
(damselfish), priacanthids (big-eyes), whilethe lower endwas dominatedby deep-water
families such as epigonids (deepwatercardinal fishes), chlorophthalmids (green-eyes),
325
bembrids (deep-waterflat-heads), symphysanodontids (no comnion name), and others.
While this pattern is not surprising given the changes in both water temperature and light.
it is certainly worth noting in any update oi'the bottomiish F.Fl1 detinilion. Similarly,
invertebrate communities showed a considerable change between 100 and 400 m,
although not with such a clear pattern at the family level.
DISCUSSION
EFH definitions are designedto guide managementdecisions on the protection
and sustainable exploitation offishery resources and therefore need to be as complete
and specific as possible. Similar to many otherfisheries in the U.S., the EFH for the
Hawaiian bottomfish fishery has been defined in general terms due to the lack ofavailable
infoiTnation on theirecology (Park, 2002) and therefore does not provide the value it
was intended to provide. This situation is changing, however, with several recent studies
generating multibeam sonardata and insitu obsei'vations useful forcreating a more
specific definition. In the Main Hawaiian Islands (MHI), a bottomfish habitat geographic
information system (GIS) that incoiporates multibeam bathymetiy and sidescan data with
over 5,000 fishing survey records was submitted this past yearto state and federal fishery
management agencies (Kelley, unpublished). Additional ship days have been scheduled
for2005-2006 to complete the mapping ofthe entire MHI 100-400 m EFH depth zone.
Recent submersible dives have been conducted on bottomfish grounds offthe islands
ofOahu, Molokai, and Kahoolawe (Kelley et al. unpublishedreport; Moffitt et al.,
unpublished) which provided macro- and micro-scale geological and biological data. In
the NWHI, multibeam mapping and submersible/ROV dives have also been conducted on
fourbanks, the data from which are summarized inthis paper. In short, a more extensive
archipelago-wide description ofthe EFH is forthcoming which will include multibeam
and insitu data from both the NWHI and MHI.
With respect to the largerpicture, this paperpresents only a brieflook at the
EFH- relevant information obtained on a deep-waterfisheiy during a study examining the
impacts offishing activities in theNWHI. Many studies are being conducted elsewhere,
whichare also accumulating large amounts ofEFH-relevant data for otherfisheries (see
Benaka, 1999). However, awidely accepted data framework forcreating EFH definitions
has notbeen developed, and consequently these efforts are not being conducted in
a coordinated manner. GIS is being commonly used to visualize habitattypes and
boundaries andmay provide the means by which the process can be standardized. All
ofthe various types ofdata summarized in this paper, including multibeam bathymetry,
substrate observations, waterquality parameters, and the various species presentat
different times ofthe day and at different depths, can be converted into GIS layers. One
can imagine many othertypes ofdata layers, such as current vectors, catch data, and life
stage distributions, which would be useful toward achieving more accurate and functional
definitions. Aconsensus needs to be attained as to which layers to include and how each
type ofdata are collected and coded. Once this occurs, the concept ofEFH truly can
beginto achieve its intendedgoal ofecosystem-basedfishery management.
326
ACKNOWLEDGEMENTS
The authors are indebtedto the following individuals and organizations
forproviding the funding forthis study: Robert Smith, National Ocean Survey;
Tom Hourigan, NOAAFisheries; Kitty Simonds, Western Pacific Regional Fishery
Management Council; andNURP. We would like to thank all ofthe observers who
assisted us withthe submersible operations: SeanCorson, Bruce Mundy, Frank Parrish,
Ray Boland, Jane Culp, Terry Kerby, andEric Conklin. Charles Holloway andT Kerby
pilotedthePisces/Fand V,while Chris Taylor, Dan Greeson, Tym Catterson, Keith
Tamburi, andMarkDrewrypilotedthe ROV. Wewouldalso like to thanktheNWHI
fishers whoprovidedus with invaluable information on theirfishing sites: Guy Ohara,
TimTimony, Dane Johnson and Bobby Gomes.
Table 1: Numberofsubmersible and ROV dives conducted on eachbank.
Bank Sub Dives ROV Dives Total Dives # Sites
Raita 10 14 24 9
W. St. Rogatien 8 15 23 12
Brooks 3 5 8 3
Bank 66 1 3 4 4
Total 22 37 59 28
Table 2: Summary ofCTD dataand calculatedlight intensity.
Depth Range (m) Salinity(ppt) DO (ml/1) Temp CO Light(klux)
100-200 34-35 5-6 15-23 38-4098
200-300 34-35 5-6 12-21 1-151
300-400 34-35 5-6 10-17 0-6
100-400 34-35 5-6 10-23 0-4098
'
327
% (NCN CN(N(N(NCNCN
en
-a
>< a-u3a5..-Do(EoecaXn). a5Dc. cCuoc/«5.-CoopcCoa3 -IoccoaS o0^e3n .re2snf •^^Soso^.---EcoeT2^n3 T=oo§e3i«n)J-J3wcHs5 o"IrEe^tn ^eOn •§e>—-nv -eCOnuTo=3'3^ J_OEQSo_!.3-co2e.an OEd0j-0;-^eao^n.^!S o;oCeE-3n =ctn ."-c9O .hoa5i.,•"OO _ oen
en
% )_2 OS ONr-r-vciO'^ •^'^(^rnrnmmr^r'-in-irjfNr^joioir^jfNt^irjoitNfN
- T^ •^- -Ea •g2 S§:2o ?e«n ^3oc« -^rcSt —_=>OcJ.-OD(eoUnO4-•^mO•o=O- J^2CE3c5S^•:-oO-=;,--gSOc-2;^-l.coc.j« "•^eo3c0O-n3 -Xxe;Ea)>-n •ax3fT^l O^cmedn.-ean "-O3 •rnoe3n.t "3e5reOann "ooonrO0eOan-0'zOfOo3^.-^:OO0253 oHO03i3
% m m — ^ \c ^onos ONoocxJt^m (^ ro — 2 <N (N ^- ^^ r^ ^ m ro fN
en
^ 03 03 -a "o en e3n e3n en 23 e3n a en
--PQosS.-S2on^^. "-ooo^^. oX 3 pr-o>3O0o0^t00n-S:r0c30ri-33!3 33I (^§0D3 nce-«n j-2z« oi- ie10feSn>3en •}P3e-^n .^J-fC"3~ 1-•'O5c:3C«= CstOo!<=^U^^ "-00eODSOb3n0.X•^—>ea>)n> 3QOO.^3 Sen DE.J!300230•T00£D30Xte)nn on. .30y0
00 D.
u E
C e
O
O. U
C/3
328
s-<
m
6ofl
oi
I
m
-S
T3
C3I
o
A
I
X
o
o
c m
—
c