Bleaching and hurricane disturbances to populations of coral recruits in Belize

Page created by Jessie Hodges
 
CONTINUE READING
MARINE ECOLOGY PROGRESS SERIES
                                                                                                         h b l i s h e d December 14
                                                     Mar Ecol Prog Ser

               Bleaching and hurricane disturbances to
                populations of coral recruits in Belize
                                                      Peter J. Mumby*
         Centre for Tropical Coastal Management Studies, Department of Marine Sciences and Coastal Management,
                  Ridley Building. University of Newcastle, Newcastle upon Tyne NE1 7RU. United Kingdom

              ABSTRACT: In 1998, coral populations in Belize were disturbed sirnultaneously by a severe coral
              bleaching event and Hurricane Mitch. The impact of these disturbances was assessed for naturally
              occurring populations of coral recruits (2 to 20 mm diameter), at a depth of 8 to 10 m on the forereef of
              Glovers Atoll. Bleaching took place at aii 4 study sites but the hurricane only affected 2 sites, enabling
              the effects of bleaching to be compared to those arising from bleaching plus hurncane damage. Pre-
              disturbance recruit density, size-frequency distribution, and cornmunity structure were similar be-
              tween sites (at kiiometre scales). The bleaching event lasted ca 3.5 mo. From 70 to 90% of adult
              colonies bleached and at least 25% of recruits exhibited signs of bleaching. A month after adult colonies
              had regained usual colouration, only 1 % of recruits showed even partial bleaching. Surprisingly, coral
              bleaching alone had no measurable effect on either recruit density or community structure. The com-
              bination of bleaching and hurricane disturbance reduced total recruit densities to 20% of pre-&stur-
              bance levels. Effects of bleaching/hurricane disturbance on community structure were spatially patchy,
              and I suggest that such patchiness may arise from variable Cover of protective rnicrohabitat and/or dif-
              ferent storm conditions mediated by proximity to reef cuts (breaks in the reef crest).

              KEY WORDS: Coral recruitment . Bleaching . Hurricane Mitch . Mortality . Cornmunity structure

                     INTRODUCTION                                    severest hurricane event at Glovers Atoll since Hurri-
                                                                     canes Greta in 1978 and Hattie in 1961 (see Stoddart
   From September to December 1998, coral popula-                     1963, Stoddart et al. 1982).
tions in Belize expenenced two of the most severe dis-                  Here, I document the effects of bleaching and hurri-
turbances of recent decades: massive coral bleaching                 cane events on coral recruits occurnng naturally on the
and Hurncane Mitch. Coral bleaching refers to the pal-               forereefs at Glovers Atoll. Recruits are defined as juve-
ing of coral tissue brought about by a loss of coral's               nile corals at the size of first sighting in the adult habi-
zooxanthellate symbionts and/or a breakdown of zoo-                  tat (Caley et al. 1996),which encompasses diameters of
xanthellate pigments (Glynn 1993, Brown 1997). The                   2 to 20 mm in this study (see also Hughes 1985). Corals
1997/8 bleaching event in Belize was Part of the most                on the seaward side of the atoll expenenced both
widespread bleaching event ever recorded and was                     bleaching and hurricane events, whereas populations
probably attributable to anomalies in maxirnum sea-                  on the leeward side were sheltered and expenenced
water temperatures and solar irradiance (International               bleaching but no apparent mechanical damage from
Society for Reef Studies 1998). In late October (during              the hurricane. By surveying recruits shortly before and
the bleaching), the eye of Hurricane Mitch passed                    after these disturbance events, and on opposite sides of
approximately 200 km south-east of Glovers Reef. Hur-                the atoll, I compare the impacts of bleaching with the
ricane Mitch was a category 5 storm and was the                      combination of bleaching and Hurncane Mitch. As far
                                                                     as I am aware, this is the first study to examine the
                                                                     effects of bleaching and hurncanes on juvenile Carib-
                                                                     bean corals. Specificaliy, I ask the following questions

Q Inter-Research 1999
Resale of full artide not permitted
28                                              Mar Ecol Prog Ser 190: 27-35, 1999

with respect to an offshore atoll in Belize: (1) Does a              ardly. Care was taken to inspect cryptic habitats (e.g.
severe bleaching event (as defined by the degree of                  between Montastraea knobs) in which recruits were
bleaching on adult colonies) lead to net mortality of                often found. Each site was located near the drop-off
recruits? (2) What is the net post-recruitment mortality             and was dominated by Montastraea annularis and M.
rate ansing from the combination of bleaching and                    faveolata. Sites measured approximately 60 by 30 m
Hurncane Mitch? (3) How do bleaching and 'bleaching                  and depth ranged between 8 and 10 m. The maxi-
plus hurricane' events affect the community structure                mum diameter of each recruit was measured using a
of recruits?                                                         clear rule and species identification was made when
  Synergistic effects between bleaching and hurricane                possible. Recruits ansing from asexual fragmentation
disturbance tvere not exarnined because no site expe-                of branching colonies (Hughes 1985) were ignored as
nenced hurncane damage but no bleaching.                             were (rare) cases where larger juveniles (>20 mm dia-
                                                                     meter) of massive species had fragmented into recruit-
                                                                     sized daughter colonies. The diameter of the smallest
                          METHODS                                    observable recruit was 2 mm but species designation
                                                                     was not possible, and therefore data are reported here
   Surveys were conducted at 4 sites on Glovers Atoll                by genus. Replication differed during pre-disturbance
(Fig. 1). Sites on the seaward forereef (East 1 [El] and             sunreys: 2 transects were surveyed at W2, 6 at E2, ? at
East 2 [E2]) experienced bleaching and hurricane da-                 W1, and 10 at El. Sample size for post-disturbance
mage whereas leeward sites (West 1 [Wl] and West 2                   surveys (14 transects, except W2 which had 13) was
[W2]) only experienced bleaching. Sampling for re-                   determined for a 90% power of detecting a 2-sample
cruits was conducted in late July 1998 and January                   differente in density of 6 recruits (sample standard de-
1999, which equates to ca 5 wk before the bleaching                  viation) based on the pooled standard deviation of
and 4 wk after the end of bleaching, and 3 mo before                 pre-disturbance data. Measurements of adult coral
the hurricane and 2 mo after the hurncane. In situ                   Cover where made using 18 replicate 1 m2 quadrats
bleaching observations were made in early October.                   per site.
  Recruitment was measured by the same diver visu-                     Overall differences in recruit density were tested
ally inspecting replicate 10 by 0.5 m belt transects ori-            using a 2-factor unbalanced general linear model
entated parallel to the reef crest and located haphaz-               ANOVA (Minitab 1997) with sites and tirnes as fixed

                 MEXICO

                                                    -   ..
                                                  ---    ...
                                                -:-     ..-

 -    17"N
   GUATEMALA

      -      IOD km
                              -  -
                               - --
                                 --8g0W
                                               -..:
                                                                                                   Prevailing direction of
                                                                                                        Hurricane Mitch

Fig. 1. Distribution of sampling sites at Glovers Atoll (= Glovers Reef) showing the location of reef cuts (breaks in the reef crest)
                                                  and the path of Hurncane Mitch
Mumby: Bleaching and hiirricane damage to coral recruits                                   29

factors. Multiple cornparisons were defined a priori               tially bleached d u n n g October (n = 130) but only
and examined using Bonferroni tests. The size-fre-                 5 recruits (of 374 censused) were partially bleached by
quency distribution of recruits was examined using                 January 1999 and none were fully bleached (see
4 classes: 2-5, 6-1.0, 11-15 and 16-20 mm. To investi-             McField 1999 for discussions of full and partial bleach-
gate the effects of disturbance on the structure of                ing).
recruit cornrnunities at each site, comrnunity distance               With 7 m seas for several days, Hurricane Mitch
was measured using the Bray-Curt1.s similarity co-                 caused extensive damage to shallow reefs removing ca
efficient and represented using non-metric multidi-                90% of living Acropora palmata at some sites. Direct
mensional scaling (MDS). The Bray-Curtis similarity                darnage to adult corals at Sites E1 and E2 (8 to 10 rn)
coefficient has a number of desirable properties for               will be reported elsewhere but ca 85% of large Mon-
measuring community similarities, such as ignoring                 tastraea annularis colonies showed recent partial mor-
joint absences of species (Mumby et al. 1996),and. is a            tality.
robust measure of ecological distance (Faith et al.
1987). MDS is a widely used method of presenting
multivariate differences in community structure be-                             Spatial patterns of recruitment
tween samples in 2 dimensions (Clarke 1993). Data
were averaged per site without transforrnation. Differ-               Pre-disturbance densities of recruits were similar
ences in community structure were examined using                   among sites (Fig. 2): ca 10 recruits per 5 rn2 transect.
similarity percentage (SIMPER) analysis of the contri-             Unbalanced ANOVA gave highly significant main
bution made by each genus to mean inter-site dissimi-              effects of site and time, and site-time interactions
larity (See Clarke 1993).                                          (Table 1).Bonferroni multiple cornparisons (not shown)
                                                                   revealed that recruit density did not vary significantly
                                                                   between sites before the disturbances. Overall re-
                         RESULTS                                   cruitment levels after the bleaching event (at Sites W1
                                                                   and W2) were not significantly different to pre-
  General impacts of coral bleaching a n d Hurricane               bleaching levels. Significant ANOVA effects were
              Mitch at Glovers Atoll                               attributable to the reduction of recruits at eastern sites
                                                                   following the combination of bleaching and Hurricane
    The 1998 coral bleaching event began in early Sep-             Mitch ( p < 0.001) The reduction of recruitment den-
tember following a rnonth of calm weather and in-                  sity at eastern sites was remarkably consistent: 80 and
 creasing water temperature (to 29 to 32"C, depending              81:,0 of pre-disturbance levels at Sites E1 and E2
 on the site, T. J . Bright pers. comm.). Most adult corals        respectively. Conversely, recruitment density showed
showed recovery of colouration by late December 1998               only a 1"/o decrease at Site W1 and a 3 % increase at
 (i.e. 3 to 4 mo later). Early signs of bleaching-induced          Site W2.
partial mortality (large patches of
recently dead tissue overgrown
with filamentous algae) on adult
                                              IWr    -        --

                                                            Bleach ing only                          Bleaching & Hurricane
                                              16 -
corals at Sites \V1 a n d W2 sug-
gested that Agaricia spp. experi-
                                              1.1 -
enced the greatest partial mortal-
ity (ca 20%, Mumby unpubl.
                                            -
data) . Acropora cervicornis is not
common ( 1 2 % total benthic CO-
ver), and ca 20% of total colony
area has died recently. Recent             2
rnortality on adult Montastraea
spp. and Porites spp. was ca 10
a n d 5 % respectively (Mumby
unpubl. data). During the bleach-
ing event, 70 to 90% of adult
colonies a w ~ e a r e dto be either
            1   L

fully or partially bleached at all                     West I               West 2              East I                  Easi 2

sites (Mumby unpubl' data' M' D'           Fig. 2.Overall patterns in recruitment density for aii genera. Open bars: pre-distur-
McField Pers.                *pproxi-      bance (July 1998) data, solid bars: post-disturbance (January 1999) data. Error bars:
mately 25 % of recruits were par-                                              standard error
l o n a plepuels :sieq i o i i g . e p p (6661A~enuey)
a~ueqrnlsrp-lsod:sleq pqos ' e p p (8661 hpy) asueqinls!p-ald :sleq uado .aseqd 6uqdures qsea Buunp ' a l ~ sqJea Je slrnl~alne loj suognqulsrp K>uanbaij-az~s.E '61d
                           (uiui) ssep laiauie!a
        -
      OZ 91             S I - II             01   -9        S-2
                           (uiui) ssep Ialauie!a
Mumby: Bleaching and hurncane damage to coral recruits

Table 1 Unbalanced general Linear model ANOVA comparing recruit density             change community structure. HOW-
            between sites, sampiing times, and their interaction                   ever, the bleaching/hurricane impacts
                                                                                   were site-specific resulting in different
  Source        df Sequential SS Adjusted SS Adjusted MS F               p         communities at Sites E1 and E2 (Fig. 4,
                                                                                   Table 2).
  Site           3        675          263             87      5.67     0.002
  Time
                                                                                      Table 2 shows how the combination
                 1        623          316            316     20.42
32                                            Mar Ecol Prog Ser 190: 27-35, 1999

Table 2 . Differences in community structure of eastern sites, E1 and E2, before and after     settlement mortality, (2) popu-
the combined bleachng and hurricane disturbances. Contnb.: absolute contnbution of             lations are not strongly recruit-

                                                                              -1
each genus to the overaii dissirnilarity between cornmunities E1 and E2 in a given             ment limited but settlement
               sampling period and Sums to total dissimiiarity between sites
                                                                                               space and post-settlement mor-
                                                                                               tality are similar between sites,
  Genus             Pre-disturbance communities         Post-disturbance cornrnunities
                                                                                               and (3) populations are not
                    Mean colony density Contnb.         Mean colony density Contrib.
                           (~m-~)                             (5m-')                           recruitment limited but post-
                      E1         E2                       E1         E2                        settlement mortality is highly

                                                                     ;
                                                                                               density-dependent. These ex-
  Agaricia                                                                                     planations cannot b e elimi-
  Pontes
  Stephanocoenia                                                                               nated at present and further
  Siderastrea                                                                                  studies are needed to examine
  i'vlontastraea                                                     0.07                      spatio-temporal dynamics of
  Leptoseris                                                         0.50                      larval supply, settlement space,
  Scolymia                                                           0.07
                                                                                               and post-settlement mortality
  Madracis                                                           0.00
  Mycetophyilia                                                      0.07          2         at different sampling grains
  Colpophyllia                                                       0.14          4         (sample unit sizes) as well as
                                                                                            spatial distances (Caley et al.
                                                                                             1996, Mumby 1999).
    The density, size-frequency distribution, and com-                 It is tempting to suggest that the dorninance of
 munity structure of recruits were found to be highly               recruits in the 6 to 10 mm size category may represent
 consistent between sites before the bleaching and                  a large-scale early summer recruitment event. How-
 hurricane disturbances began. This was surprising                  ever, densities of the sinaiiest recruits in this size class
 considering that sites were drawn from opposite sides              (2 and 3 mm diameter) were probably underestimated
 of Glovers Atoil and the spatio-temporal Patterns of               during visual sampling even though cryptic habitats
 recruitment usually show great vanation. The scale of              were scrutinised carefully. The 2 to 5 mm class is prob-
 the study may provide some insight into the apparent               ably a better record of 4 to 5 rnrn recruits. The propor-
 uniforrnity of community characteristics. Edmunds et               tion of 2 to 3 mrn corals ranged from 0 to 4 % of total
 al. (1998) described an appropriate sampling strategy              recruit abundance, so undersampling had Little impact
 for juvenile corals (0.4 to 5.0 cm diameter) to consist of         on the conclusions reached by this study, although
 seventeen 0.25 m2 quadrats placed along 4 transects                very recent (
Mumby: Bleaching and hurricane damage to coral recruits                              33

the population dynamics of juvenile corals. I suggest at       have been compensated by recruitment and growth.
least 4 possible explanations for the apparent absence         Future studies are needed to examine population
of bleaching-induced mortality. Experimental tests a r e       dynamics at various temporal scales and with manipu-
necessary to discriminate among them.                          lations in recruit density. Empirical studies are also
                                                               needed on the effects of bleaching on juvenile growth
                                                               rate, but, as I point out next, local conditions may miti-
 (1)Did juveniles experience a severe bleaching event?         gate the influence of bleaching on growth.

   Few studies of in situ bleaching (if any) have specifi-
cally examined whether juvenile corals expenence                (3) Did heterotrophic feeding on sediments mitigate
similar bleaching responses to those exhibited by adult           reduced autotrophic nutrition during bleaching?
colonies I did not conduct detailed recruitment sur-
veys during the bleaching event but visits to each               The western side of Glovers Atoll experiences con-
study site suggested that approximately 25",, of re-          tinual sedirnent deposition from the seaward lagoon
cruits showed either partial or (occasionally) full           (Gischler 1994). Whilst sediment can reduce the
bleaching. These observations were made a third of            growth rate, smother and kill juvenile corals (Birke-
the way into the bleaching event a n d cannot be con-         land 1977, Hunte & Wittenberg 1992, Babcock &
sidered representative since a n unknown proportion of        Mundy 1996), low levels of sediment may perrnit hete-
recruits may have bleached earlier and subsequently           rotrophic feeding on epipelic micro-organisms (Toma-
recovered or died, or bleached later (the Same is true of     scik & Sander 1987).Since the levels of suspcnded sed-
snapshot estimates of bleaching intensity in adult            iment deposition at Glovers a r e not high enough to
corals). It is safest to conclude that at least 2 5 % of      deem the reef uninhabitable to juvenile corals, it is fea-
recruits expenenced some bleaching during the event           sible (but not tested) that sedimentation levels aug-
a n d that 1 mo after the adults had regained normal          ment the feeding and therefore survivorship of recruits
colour only 1 "" of recruits exhibited even partial           on this atoll.
bleach.ing. The difficulties of quantifying the severity
of the bleaching event notwithstanding, it appears
unlikely that recruits survived a bleaching event              (4) Do juvenile corals receive less irradiance by virtue
purely by avoiding becoming bleached; it is more                            of their cryptic microhabitat?
likely that the incidence of bleaching in juvenile corals
was less than that in adults. It is not clear whether            High levels of ultraviolet and photosynthetically
reduced bleaching in juvenile corals might result from        active radiation (PAR) have been implicated to stimu-
a n ontogenic physiological change in zooxanthellate          late coral bleaching (see review by Brown 1997).
susceptibility to bleaching, or simply a n external con-      Whether the cryptic location of many juvenile corals
sequence of juvenile microhabitat (see suggestions            on, for example, under-sides and vertical surfaces
under Question 4 below).                                      (Babcock & Mundy 1996) confers a refuge from sus-
                                                              tained direct exposure to i.rradiance remains to be
                                                              Seen. Furtherrnore, whilst some juvenile corals bleach,
(2) Was bleaching-induced stress insufficient to create       exposure of the (bleached) coral tissue to PAR may
                  colony mortality?                           partly determine the fate of the coral. Studies are
                                                              needed into the scattering and attenuation of irradi-
    Did bleaching stress reduce growth rates effectively      ance at the scales of recruit microhabitat and into the
shifting juveniles into a dormant state? Dormancy             direct effects of irradiance on bleached coral tissue.
would help explain the consistent size-frequency dis-
tributions at Sites W1 and W2 before a n d after bleach-
ing, but for population Parameters to remain un-                 Effects of coral bleaching plus Hurricane Mitch
changed, chronic levels of mortality would have to be
absent or balanced by recruitment and growth. Mor-              Coral bleaching was present throughout the Belize
tality rates are unlikely to be Zero over 6 mo (Bak &         Barrier Reef and atolls making it impossible to distin-
Engel 1979) and therefore recruitment and growth              guish the effects of hurricane disturbance on bleached
must have occurred. It is not clear whether mortality         and non-bleached coral populations. Given the ob-
rates are simply low and balanced by low levels of            served absence of a bleaching-induced disturbance to
recruitment and growth, or whether population sizes           juvenile corals at western sites, however, it seems
a r e tightly regulated (e.g. by density-dependent mor-       unlikely that bleaching would have significantly exac-
tality) such that even a marked mortality event could         erbated the damage created by the hurricane. Hurri-
34                                         Mar Ecol Prog Ser

  cane events lead to mortality of recruits by sediment         Canbbean reefs which also expenenced severe
  scouring, direct mechanical breakage, and the re-             bleaching. The responses of juvenile coral populations
  moval of substratum. Post-hurncane events such as an          to large-scale phenomena such as coral bleaching
  ephemeral bloom of blue-green and filamentous green           need to be categonsed. For each type of bleaching
  algae (R. B. Aronson & T. J. Bnght pers. comm.) may           event (e.g. duration, percent of adults affected in each
  also create further stress: Van Moorsel (1985) found          bleaching class), classifications should attempt to
 filamentous algae to be the greatest cause of mortality        relate the physical and biological conditions of the site
 in juvenile Agancia humilis.                                   (e.g. depth, physical exposure, herbivore grazing in-
    Spatially patchy responses of adult cornmunities to         tensity, adult coral cover) to the responses of juveniles
 hurncane damage are weli established (Woodley et al.           at the scales of comrnunity, life history strategy, and
  1981, Edmunds & Witrnan 1991, Witman 1992, Bythell            individual populations. If patterns emerge at large spa-
 et al. 1993a,b,Rogers 1993),and, based on this study, it       tial scales, such classifications will be instrumental in
 appears that this premise extends to juvenile commu-           modelling the irnpacts of bleaching events on the
 nities at a depth of 8 to 10 m. The proportional reduc-        dynamics of coral metapopulations (Mumby 1999) and
 tion of Agancia and Leptoseris recruits was less at Site       therefore help bndge the gap between the large scale
 E1 than Site E2. Site-specific mortality patterns may          of many population processes and the often inade-
 offer a simple explanation of the differences in relative      quate scale at which they are studied.
 agancid abundance. First, survivorship of agancid
 recruits may have been higher at Site E1 because the
 cover of Montastraea was greater at Site E1 (ca 60 % vs        Acknowledgements. I would iike to thank Elizabeth Glynn,
                                                                Ben Hadfield, and Karen Jarnes for their assistance in the
 39% at Site E2; note cover of Montastraea is excep-
                                                                field. Dr Tom Bright, Cindy Liles, and the staff of Glovers Reef
 tionally high at these sites). Most agancid recruits sur-      Marine Research Station provided invaluable logistical sup-
 veyed in January were found deep between Montas-               port. I thank the Belize Departrnent of Fisheries and the
 traea knobs, which may have provided shelter from              Wildlife Conservation Society for permission to work at
 scouring and direct mechanical stress. Second, prox-           Glovers A t d . The comments of 4 referees improved a n ear-
                                                                Lier version o m e manuscript. This study was funded by a
 irnity to reef cuts may affect the velocity and surge of       NERC Fellowship under the Earth Observation Science Initia-
water. Site E1 is near a reef cut whereas Site E2 is mid-       tive. This is contribution number 5 of Glovers Reef Marine
way along a continuous reef (Fig. 1).Mechanical dam-            Research Station.
age to adult colonies appeared to be more severe at
Site E2, which might be because dispersal of stornl
surge is irnpaired where connectance to the lagoon is                                 LITERATURE CITED
impaired by a reef crest. It is possible, but perhaps
unlikely, that differences in agancid abundance were            Babcock RC (1989) Fine-scale spatial and temporal patterns
actually attnbutable to (or modified by) recruitrnent. If           in coral settlement. Proc 6th Int Symp Coral Reef 2:
                                                                    635-639
agancid planulae settled immediately after the hurn-            Babcock R, Mundy C (1996) Coral recruitrnent: consequences
cane and grew at their maximum reported linear                      of settlement choice for early growth and s u ~ v o r s h i pin
extension rate of ca 80 pm d-' (Van Moorsel 1985),they             two scleractinians. J Exp Mar Biol Ecol206:1?9-201
would have reached 6 mm diameter by the January                 Bak RPM, Engel MS (1979) Distribution, abundance and sur-
census, placing them in the size class of greatest abun-           vival of juvenile hermatypic corals (scleractinia) and the
                                                                   importance of hfe history strategies in the parent coral
dance at Site E l . However, it is doubfful whether max-           community. Mar Biol 54:341-352
imum growth rates would have been realised. High                Birkeland C (1977) The importance of rate of biomass accu-
water temperatures would have favoured high growth                 mulation in early successional stages of benthic communi-
rates but reduced autotrophic productivity dunng                   ties to the survival of coral recruits. Proc 3rd Int Symp
                                                                   Coral Reef 1:15-21
bleaching and the abundance of fouling filamentous              Brown BE (1997) Adaptations of reef corals to physical envi-
green algae would lower growth rate (Van Moorsel                   ronmental stress. Adv Mar Biol 31:221-299
1985).                                                          Bythell JC, BytheU M, Gladfelter EH (1993a) Initial results of
                                                                   a long-term coral reef monitoring progranl: impact of Hur-
                                                                   ncane Hugo at Buck Island Reef National Monument, s t .
                                                                   Croix, US Virgin Islands. J Exp Mar Biol Ecol 172:171-183
            Large-scale context of results                      Bytheli JC, Gladfelter EH, Bythell M (199313)Chronic and cat-
                                                                   astrophic natural mortality of three cornmon Caribbean
  The absence of significant mortality in populations of           reef corals. Coral Reefs 12:143-152
coral recruits during a severe bleaching event was              Caley MJ, Carr MH, Hixon MA, Hughes TP, Jones GP,
                                                                   Menge BA (1996) Recruitrnent and the local dynaniics of
unexpected, especially given the sensitivity of juve-              Open manne populations. Annu Rev Ecol Syst 27:4?7-500
niles to some other Stresses (Bak & Engel 1979).It is too       Clarke KR (1993) Non-parametnc multivariate analyses of
early to comment on the generality of this result among            changes in community structure. Aust J Ecol 18:11?-143
Mumby: Bleaching and hurncane damage to coral recruits                                   35

Edmunds PJ, Witman JD (1991) Effect of hurricane Hugo on              disturbance hypothesis revisited Coral Reefs 12:127-137
    the primary framework of a reef along the south shore of      Rogers CS, Fitz HC 111, Giinack M, Beets J, Hardin J (1984)
    St. John, US Virgin Islands. Mar Ecol Prog Ser 78:201-204         Scleractinian coral recruitment patterns at Salt River Sub-
Edmunds PJ, Aronson RB, Swanson DW, Levitan DR, Precht WF            manne Canyon, St. Croix, U.S. Virgin Islands. Coral Reefs
    (1998) Photographic versus visual census techniques for the       3.69-76
    quantification of juvenile corals. Bull Mar Sci 62:93?-946    Rylaarsdam KW (1983) L ~ f ehistones and abundance patterns
Faith DP, Minchin PR, Belbin L (1987) Compositional dissimi-         of colonial corals on the Jarnaican reefs. iMar Ecol Prog Ser
    larity as a robust measure of ecological distance. Vegeta-        13:249-260
    tio 69:57-68                                                  Smith SR (1992) Patterns of coral recruitment and post-settle-
Gischler E (1994) Sedimentation on three Caribbean atoils:           ment mortality on Bermuda's reefs: comparisons to the
    Glovers Reef, Lighthouse Reef and Turneffe Islands,              Canbbean and Pacific reefs. Am Zool 32:663-673
    Belize. Facies 31 :243-254                                    Smith SR (1997) Patterns of coral Settlement, recruitment and
Glynn PW (1993) Coral reef bleaching: ecological perspec-            juvenile mortality with depth at Conch Reef, Flonda. Proc
    tives. Coral Reefs 12:1-17                                       8th Int Symp Coral Reef 2:1197-1202
Hughes TP (1985) Life histones and population dynamics of         Stoddart DR (1963) Effects of Hurricane Hattie on the British
    early successional corals. Proc 5th Int Symp Coral Reef 5:       Honduras reefs and cays. October 30-31, 1961 Atoll Res
    101-106                                                          Bull No 95
Hughes TP, Jackson JBC (1985) Population dynamics and hfe         Stoddart DR, Fosberg FR, Sachet MH (1982) Ten years of
    histones of foliaceous corals. Ecol Monogr 55:14 1- 166          change on the Glover's reef cays. Atoll Res Bull No. 257
Hughes TP, Baird AH, Dinsdale EA, Moltschaniwskyj NA,             Szmant ALM (1986) Reproductive ecology of Canbbean reef
    Pratchett MS, Tanner JE, WilLis BL (1999) Patterns of            corals. Coral Reefs 5:43-54
    recruitment and abundance of corals along the Great Bar-      Tomascik T, Sander F (1987) Effects of eutrophication on reef-
   rier Reef. Nature 397:59-63                                       building corals. 111. Reproduction of the reef-building coral
Hunte W, Wittenberg M (1992) Effects of eutrophication and           Pontesporites. Mar Biol94:77-94
    sedimentation on juvenile corals: 11. Settlement. Mar Biol    Van Moorsel GWNM (1983) Reproductive strategies in t ~ v o
    114:625-631                                                      closely related stony corals (Agaricja, Scleractinia). Mar
International Society for Reef Studies (1998) ISRS Statement         Ecol Prog Ser 13:273-283
   on bleaching. Reef Encounter 24:19-20                          Van Moorsel GWNM (1985) Disturbance and growth of juve-
McField MD (1999) Coral response during and after rnass              nile corals (Agaricia humjljs and Agaricia agaricites, Scle-
   bleaching in Belize. Bull Mar Sci 64:155-172                      ractinia) in natural habitats on the reef of Curacao Mar
Minitab (1997) Minitab user's guide. Minitab Inc. PA                 Ecol Prog Ser 24:99-112
Mumby PJ (1999) Can Canbbean coral populations be mod-            Wiens JA (1989) Spatial scaling in ecology. Funct Ecol 3:
   elled at metapopulation scales? Mar Ecol Prog Ser 180:            385-397
   275-288                                                        Witman JD (1992) Physical disturbance and community struc-
Mumby PJ, Clarke KR, Harborne AR (1996) Weighting spe-               ture of exposed and protected reefs: a case study from St.
   cies abundance estimates for manne resource assessment.           John, US Virgin Islands. Am Zool 32:641-654
   Aquat Conserv Mar Freshw Ecosys 6:115-120                      Woodley JD and 19 others (1981) Huncane Ailen's impact on
Rogers CS (1993) Hurricanes and coral reefs: the intermediate        Jamaican coral reefs. Science 214:749-755

Editorial responsibility: Otto Kinne (Editor),                    Subrnitted: March 15, 1999; Accepted: July 16, 1999
Oldendorf/Luhe, Germany                                           Proofs received from author(s): November 25, 1999
You can also read