Evaluating coral reef benthic communities in remote Caribbean atolls (Quitasueno, Serrana, and Roncador Banks) to recommend marine-protected areas for the seaflower biosphere reserve

2005·23.8 MB·English

by  Juan Armando Sanchez

ATOLL RESEARCH BULLETIN
NO. 531
EVALUATING CORAL REEF BENTHIC COMMUNITIES IN REMOTE
CARIBBEAN ATOLLS (QUITASUENO, SERRANA,AND RONCADORBANKS)
TO RECOMMEND MARINE-PROTECTED AREAS FOR THE SEAFLOWER
BIOSPHERE RESERVE
BY
JUANARMANDO SANCHEZ, VALERIA PIZARRO,ALISON R. ACOSTA-
DE- SANCHEZ, PAULA A. CASTILLO,PILAR HERRON,JUAN CAMILO
MARTINEZ, PHANORMONTOYA,AND CARLOS OROZCO
ISSUEDBY
NATIONALMUSEUMOFNATURALHISTORY
SMITHSONIANINSTITUTION
WASHINGTON,D.C.,U.S.A.
NOVEMBER2005
Figure 1.SouthwesternCaribbean showingthe loeationofthe Northernbanks(Colombia).
,
EVALUATING CORAL REEF BENTHIC COMMUNITIES IN REMOTE
CARIBBEAN ATOLLS (QUITASUENO, SERRANA, AND RONCADOR BANKS)
TO RECOMMEND MARINE-PROTECTED AREAS FOR THE SEAFLOWER
BIOSPHERE RESERVE
BY
JUANARMANDO SANCHEZ,1 VALERIAPIZARRO,2ALISON R.ACOSTA-
DE-SANCHEZ,1 PAULAA. CASTILLO,3 PILARHERRON,2 JUAN CAMILO
MARTINEZ,2 PHANORMONTOYA 3AND CARLOS OROZCO2
ABSTRACT
Three remote oceanic atolls, theNorthern Cays (Quitasueno, Serrana, and
RoncadorBanks,Archipelago ofSanAndres andProvidencia, Colombia) were
intensively surveyed duringApril and May, 2003 through amultilevel quantitative study
ofthebenthic communities (substrate cover: coral-benthos; densities: octocorals and
urchins; frequencies: coral diseases). This study is the firstapproachtowards the design
ofanetworkofMarine ProtectedAreas (MPAs) in these Caribbean reefs. We found
between 66 and 73 categories ofsubstrates inthe benthic communities, including43-to-
46 species ofreef-building corals and38 species ofoctocorals.Ample ranges ofcoral
cover(2-52 %) andoctocoral densities (0-15 col. nr2) were observed among stations. The
community structure correspondedto shared species-specific environmentalpreferences,
perceptible as minorgradients such as windward-leeward and/or shallow-deep, and
tophysical/topographical characteristics, which in mosthabitats were products ofthe
reef-building corals themselves (e.g., 'Montastraea'or 'Acroporapalmata-DiploricC
habitats).
Depthwas the majorstructuring force. Species abundance distribution followed
the patternthat few species are dominantwhereas mostofthemare rare. Common
species wereMontastraea spp. (reef-building corals), Pseudopterogorgia bipinnata (an
octocoral) andLobophora variegata (fleshymacroalgae). Seaurchins and coral diseases
had aheterogeneous distributionbeing absentin many stations and frequent in a few.
Higherdensities ofseaurchins {Diademaantillarum andEchinometra viridis) were
encountered in Quitasueno Bank. There were areas ofhigh coral abundance and diversity
in the three atolls which can be importantforthe MPAdesign. Theparticularassemblages
hadmeanvalues ofcoral cover> 32% (mostofthem>38%), coral species >16, octocoral
densities >0.6 col. nr2 (usually>2.9 col. nr2), andoctocoral diversity>4 (usually>8),
which are withintheranges ofthebest-conservedreefs inthe Caribbean. Some areas of
special concern were identifiedthatcouldneed special management.
1 Dept.CienciasBiologicas,U. delosAndes,P.O.Box4976,Bogota,Colombia. Email:juansanc@
uniandes.edu.co.
2CorporationparaeldesarrollosostenibledelArchipielagodeSanAndresyProvidencia-CORALINA,
SanAndresisla,Colombia.
3InstitutodeInvestigacionesMarinasyCosteras-INVEMAR,SantaMarta,Colombia.
Manuscriptreceived22September2003;revised16August2004.
Difficult management challenges are the high density/diversity areas of
endangered andor rare octocorals (Roncadorand Serrana Banks: Gorgonia spp. and
P. elisabethae) as well as heterogeneous shallow patch reefs ofendangered corals and
high!) productive coralline algae (./. palmata and Porolithonpachydermum) intermingled
uitfa Montastraea spp. reefs exhibiting disease outbreaks (white plague, bleaching, and
dark spots) in Serrana Bank.
Figure2.Coral reefsattheMontastraeaspp. habitat in RoncadorBank, SouthwesternCaribbean,
Colombia. Upper: Organismswith mostoftheirbiomassabovethesubstrate suchasgorgonianoctocorals
(Pseudopterogorgiaspp.). Below: Organismscoveringthereefsubstrate; dominantcoral in thephotois
Montastraeafranksi.
INTRODUCTION
Quitasueno, SerranaandRoncadorBanks (Northern Cays) are some ofthe
fewtrue oceanic atolls oftheAtlantic surroundedby depths ofthousands ofmeters
(Diazetal., 1996a). The coral-reefbanksbelongto theArchipelago ofSanAndres
and Providencia (Colombia) in the southwesternCaribbean (Fig. 1). This areabeen
declared abiosphere reserve byUNESCO (SeaflowerBiosphere Reserve) andthe local
environmental authority(CORALINA) isproceeding with the design ofanetwork of
MPAs inthe area includingtheNorthern Cays (e.g., Friedlanderetal., 2003). However,
very little information onthebenthic communities ofthese atolls is available to design
a series ofMPAs. This study comprises the firstdetailedandquantitative studyofthe
benthic communities ofthe coral reefs on Quitasueno, Serrana and RoncadorBanks that
will contribute tothe zoning and design ofthe MPAs. Ultimately, local authorities will
have these recommendations, in additionto parallel studies onfish communities and
Queen Conchpopulations (Strombusgigas), to confrontwith stake-holderneeds and
finally design anetworkofMPAs inthese banks.
TheArchipelago ofSanAndres and OldProvidence comprises a series of
oceanic islands, barrier-reefcomplexes, atolls andcoral shoals on avolcanic basement
(Fig. 1) aligned inanorth-northeasterlydirection overnearly 500 km alongthe Lower
NicaraguanRise offthe CentralAmerican continental shelf(Geister, 1983, 1992; Diaz
etal, 1995; Diaz etal., 1996a, 1996b, 2000). Both geomorphological andhabitat
distributionpatterns are similaramongthese reefcomplexes mainlydue to the strong
wave action generatedbytradewinds. Theupperplatformofthese atolls is surrounded
byarathereven 20m-deepterrace. On its windward, fore-reefside ithas low relief
due tothe strong abrasionwhichallows only sparse growth ofencrusting andmound-
shaped corals butpermitsprofuse gorgonian development (Sanchez etal., 1997, 1998).
Thiswindward, fore-reefterrace gradually slopes upwards emerging on areefcrest
which isvariablydevelopedas aspur-and-groove system ofhydrocorals {Millepora
spp.), zoanthids and crustose coralline algae (corallinaceae). The crest forms an almost
continuousbarriertothe waves and fringes the eastern, northern and southern internal
margins ofthe upperplatform.
Behindthe barrierthere is an extensive, sinuous and shallow (1-3 m) sand-and-
rubble lagoonal terracethat is connectedto the lagoonalbasinby an abrupt sandy slope.
There is no reefformationbehindthe crest (e.g., backreef) as inthetraditional Caribbean
fringingreef(e.g., Jamaica). The mid-depth (12-18 m) lagoonalbasin is usually covered
byadensepatch-reefnetwork (e.g., Diaz etal., 1996b). Thesepatchreefs are mostly
dominatedbymassive scleractinian corals, especiallyMontastraeaannularis andM.
faveolata. Thereefs are irregularintopographywith sandpatches occurring amongthe
coral heads (e.g., large massive reef-building corals). Bordering the westernmargin of
the lagoonthere is a leewardterrace, oftentimes crownedwith adiscontinuousband of
peripheral shallowreefs, innersandbars and algalridges thatpartially enclose the lagoon
(e.g., Diaz etal., 1997). The leewardterrace gradually slopes downtowards the 20-30m
deepbreakonthe outer-slope margin supporting moderate reliefreefs withvariable live-
coral cover(see more details inMilliman, 1969; Diaz etal., 1996a, 2000). Thebenthic
community structure ofthese oceanic reefcomplexes, including Serrana and Roncador, is
considerably different than coastal coral reefs (Diaz-Pulido et al. submitted).
The most conspicuous sessile benthic organisms ofCaribbean coral reefs are
localized at two overlapping layers (Fig. 2). One layercomprises organisms that are
covering the substrate, principally scleractinian and milleporinan corals, zoanthids,
sponges, and algae, referred to here as the benthic community. The other layer is
composed oferect orbushy organisms, presentjust above the substrate, such as
gorgonians, octocorals and black corals, which we referto as the octocoral community.
Typically, these layers have been studied separately and little is known about thejoint
distribution patterns ofthe two layers. With a few exceptions (Florida: Goldberg 1973;
\\ heaton and Jaap, 1989) little is known about the distribution patterns ofthe benthic and
octocoral communities, information which could provide valuable clues on the processes
governing the community structure ofCaribbean reefs.
Ecological communities orassemblages are composed ofindividuals ofdifferent
species that co-occurand potentially interact with one anotheras determined by past
and continuing evolutionary processes (McPeek and Miller, 1996). In the coral-reef
benthic community there are several species guilds that coexist overa wide range of
environmental conditions. The most conspicuous guilds ofbenthic sessile organisms
show particularmorphological adaptations related to precise life-history strategies
(Jackson, 1977). Many scleractinian coral species occurovera wide bathymetric range
but with similarwave-energy conditions (e.g., Geister, 1977; Graus and Macintyre,
1989). On the other hand, gorgonian octocorals, due to their bushy physiognomy and
position in the upper layer, have advantages such as a nearabsence ofcompetition for
space orcolonial expansion such as occurs in the coral-reefcommunity. However, they
are more affected by drag forces associated with water movement (Jackson, 1977) and
thus the octocoral community patterns more closely reflect the influence ofwave energy
(Kinzie 1973; Alcolado, 1981; Jordan, 1989; Yoshioka andYoshioka, 1989; Sanchez et
al., 1997). This study presents information collected at the same time on both benthic
and octocoral communities as two semi-independent sources ofcommunity structure
information.
The aim ofthis study was to survey the community/habitat structure and reef
condition on Serrana, Roncadorand Quitasueno Banks. The specific goals included
obtaining quantitative information on: (1) the community oforganisms that coverthe reef
substrate orbenthic community; (2) the community ofgorgonian octocorals that colonize
space above the substrate oroctocoral community (a part ofthe benthic community
that is underrepresented using benthic covermethods); (3) the populations ofkeystone
urchin species; (4) the topographical complexity orrugosity ofthe habitat; and (5) marine
diseases affecting reef-building corals.
METHODS
The Northern Cays were intensively surveyed duringApril 24 to May 1 1, 2003
on board the M/V Spree and M/N Anglonamar II using enriched air SCUBAtechniques
(EANx-Nitrox, up to 32%). Atotal of 118 stations were sampled (48 on Quitasueno,
2
48 on Serrana, and22 on Roncador). We triedto survey as manyhabitats as possible,
thoroughlycoveringthe reefplatformwiththe guidance ofexistinghabitatmaps (Diaz et
al. 1996a, 2000). The surveyedhabitats included shallow lagoonal sites (e.g.,Acropora-
DiploriaorMontastraea spppatchreefs and algal mat habitats), leeward terrace sites
(e.g., Montastraea spp, scattered corals ormixed-coral habitats), and fore-reefterrace
sites (e.g., gorgonian octocorals and scattered corals habitat) (see details inresults
section). The fore-reefterrace, however, was surveyed on Quitasueno Bank. Bad weather
preventedaccess to the habitats on Serrana andRoncador.
The abundance ofthe organisms coveringthe reefsubstrate was estimatedas
percentcoverusingtheplanarpoint-intercept method(PPI) (Dodge etal., 1982). This
method is amodification ofthe point-interceptmethod ofLoya (1978) in which the
points are surveyedwithin a grid instead ofalong a line. Ithas the advantages ofboth
plotandplotless methods as well as typicallyproducing less inter-observervariation than
occurs withtraditional quadrantmethods. Sampling stations consisted of10randomly
chosen one-square-meterplots out of50 potentialpositions alongboth sides ofa25
m- lineartransect. At eachposition we surveyedthebenthic species, genera (for some
common algae), functional groups (forparticularalgae types and sponges), and types of
inert substrate on agrid of64points. Wetriedto record 640 pointsperstation, with 500
pointsbeing consideredthe minimumnumberrequiredto obtain arepresentative survey
ofthe station. Due to diving safety limits, the numberofpoints varied slightly among
stations with amean of628 (sd=44, min>500) andno differences detected amongthe
means forthe threebanks (ANOVA, P=0.25, df=2, 115). Atotal of74,127 points were
surveyedatthe 118 stations whichwas equivalentto 7412.7 linearmeters oftransect
withpoints every 10 cmconcentratedwithin 1158 square meters ofgridusing 64 points
persquare meter. Environmental variables recorded ateach station were depth and
topographic complexityorbottomrugosity. Rugosity, which is an indicatorofwave-
motion energy andthe chronic disturbance effectofwaves (Aronson and Precht, 1995),
was estimated fromthe ratio oflinearlength ofa chain thatwas laid out ina straight line
alongthe bottom following allthe vertical reliefto its lengthwhen stretched out (10 m;
e.g., Sanchez etal., 1997).
The transect linewas establishedby ateam ofthree divers surveyingfish
populations (Dalgreen etal., inprep.) thatprecededthe benthic team at each station. Each
benthic surveywas conductedbyone oftwoteams offourobservers thatalternated dives.
Within each team, two divers estimated substrate coverandthe othertwoperformedthe
remainingtasks. One surveyedgorgonian octocorals, urchins, and, ifpossible, substrate
cover. The fourth divermeasuredbottom rugosity, looked forrare speciesto estimate
more completelycoral species diversity (alpha diversity) and examineddiseasedcoral
colonies visually identifyingthe coral species, the disease andestimating the percentage
ofaffectedtissue. Gorgonian octocoral densityperspecies was estimatedusing the same
10 square-meterplots in which all colonies were counted and identifiedto the species or
genus level (e.g., Sanchez, 1999). We countedandidentified4,828 colonies ofoctocorals
atthe 118 stations. Seaurchin densities (1,710 counted atthe 118 stations), particularly
ofDiademaantillarum andEchinometra viridis, were estimatedbycountingthe number
ofurchins within Im along both sides ofthe 25 m-long transect (50m:). All divers were
fully trained previously for underwater identifications and data collection.
The species distribution patterns and community structure were analyzed using
a variety ofmultivariable methods. Classification methods identify the similarity of
cases (e.g.. stations) according to theirvariable composition (e.g.. species). Ordination
methods allow the variation from several variables to be concentrated orexpressed in
a tew composite, variance-rich variables to enable the detection ofwhole-community
patterns. Classification and ordination methods can be combined to identify both
structure and gradient patterns in the community (Flury & Riedwyl, 1988). Initially the
database was organized into species-per-station matrices. Forboth substrate coverand
gorgonian density, stations were classified by normal (Q-mode) cluster analysis using
species cover, same as substrate cover, ordensity data (log [X+l] transformed), the
|n
Bray-Curtis dissimilarity index, and the Unweighted PairGroup Method (UPGMA) for
building dendrograms (Field et al.. 1982). Ordination analyses with the same data were
obtained by a multivariate eigen-vectorprocedure, Detrended CorrespondenceAnalysis
(DCA) (Ter Brak 1986), which identifies gradients among stations and/orenvironmental
variables. Potential environmental interpretations ofthe DCAaxis values were found
through correlation with depth and rugosity (e.g., Sanchez et al., 1997).
Previously, this suite ofmethods has been applied successfully to similar benthic
assessments within the studiedArchipelago (Friedlanderet al., in press, in prep.). To
further understand the relationship between groups orclusters ofstations and community
variables (e.g., coral cover, octocoral density, urchin density, frequency ofcoral diseases,
etc.) orenvironmental variables (depth and rugosity), the magnitudes and mean location
ofthese variables were superimposed on DCA 1 vs DCA2 station ordination plots.
Finally, the characteristic species ofeach successively nested clusterwithin a dendrogram
were determined by recalculating substrate cover or density for each minimal cluster
and then accumulating theirabundance and successively marking those minimal clusters
which contained at least 70% ofthe species' total abundance (R-mode). Species were
then reordered and grouped according to membership within a given cluster or nest of
clusters (Kaandorp, 1986; e.g., Sanchez et al., 1998).
RESULTS
Quitasueno Bank
Atotal of73 benthic categories among species, genera, and functional groups of
sessile organisms were recorded in Quitasueno Bank. There were a total of46 species
ofmilleporids and scleractinian corals observed qualitatively, 12-to-30 (min-max) coral
species recorded quantitatively per station, and a coral cover ranging from 18-to-50%.
Overall, analysis ofthe major functional groups showed that hard corals (Scleractinia
and Milleporidae) were the most abundant organisms (18-50% cover), followed byfleshy
macroalgae (4-39%), calcareous macroalgae (4-25%), encrusting algae (0.8-10%), and
octocoral bases (0-10%), with lesseramounts offilamentous algae and invertebrates
being recorded (Fig. 3). The classification and clusteranalyses showed adivision ofthe
48 stations in two minimal groups with less than40% dissimilarity (I and II), andwith
two and fourmajor sub-clusters respectively (Figs. 4-5). Station 22 didnot clusterwith
any otherstation. This stationwas situated in atypical rubble and algae (Avrainvillea
spp.) patch in the leeward margin (e.g., Fig. 6 C-D). The most importantresult from
this clusteranalysis is thatthe classification ofthe stations revealed significant spatial
structure on Quitasueno where apparentlyhomogeneous habitats (e.g., habitatmap:
Fig. 5) haddifferent sub-clusters thatcorrespondto the reefwindward-leeward edges
(Fig. 5A: clusters D and C). Similarly, some lagoonal-leeward habitats had a spatial
differentiation depending onwhetherthey were north or south withrespect tothe
lagoonal basin (Fig. 5A: clustersAandB respectively). Therefore, a characterization
ofthe benthic communities based onthe habitat mapping, as observedonly withaerial
photographs (Diaz et al., 1996, 2000), is not entirely arealistic view ofthe quantitative
benthic community structure. The latter, nevertheless, is not completely disparatewith
respectto the majorhabitats (e.g., clusterF: Fig. 5A); these are hypotheses ofcommunity
structure and forzoning andmanagementpurposes the results fromthe classification
analysis and consequent ordination andinverse analyses shouldbe taken in account (see
discussion).
«i m-M
Corals • •
>••
Macroalgae
Calcareous Macroalgae 4flfr>
FilamentousAlgae
EncrustingCorallineAlgae
Octocoral Bases
EncrustingSponges
UprightSponges
OtherfleshyInv. if*'
10 20 30 40 50 60
Cover (%)
Figure3.Boxplotsfromthedatadistributionperstationofthemajorfunctionalbenthicgroupsinthe
benthiccommunity.Themedianlineisinsidethe25thand75thpercentileswithexternalerrorbarsatthe
10thand90thpercentiles. Dotsshowstationsoutsidetheerrorbars.QuitasuenoBank.
Ofthe two main station clusters, clusterI is made up ofleewardreefstations
including sites in the lagoon and onthe leewardmarginthatalways had aprotected
positionwith respectto the reefcrest (Fig. 5: clustersAandB; Fig. 4).As presented
in Table 1 andFigures4-5, sub-clusterAfrom the 'protectedclusterI' is composedof
shallow and intermediate depth stations in the lagoon, and sub-clusterB consists ofsome
intermediate depth leewardreefs.
10
0.4 0.3
Bray Curtis
Figure4. Dissimilarity(Bray-Curtis)dendrogramofthe48 benthicstationsonQuitasuenoBank.The
dashed lineshowsthe40%dissimilaritydivision amonggroups.
Both sub-clustersAand B had stations placed indiscriminately in three different
habitats, 'Acropora-Diploria', 'Montastraea\ and mixed corals. Cluster II is composed
ofa mixture offore-reefand leeward stations that are citherwindward or well leeward
at a greaterdistance from reefcrest compared to cluster I stations. The 'terraces cluster-
IP had sub-clusters ofstations exclusively from the fore reefand leeward terraces (Fig.
5A [cluster E]; Fig. 4; Fig. 7 A-B) and two sub-clusters (C and D) with stations located