Home - Bleaching Part 2 - References

REVIEW ON CORAL REEF BLEACHING (Part 1)

BY

MARTIN PECHEUX (1997)

(martinpecheux@minitel.net)

TABLE OF CONTENTS

ABSTRACT
INTRODUCTION
FIRST PART : FIELD KNOWLEDGE
 

I) IN SITU OBSERVATIONS ON MASS BLEACHING

A) SPECIFIC AND INDIVIDUAL PATTERNS

Organisms involved
Interspecific differences
Intraspecific differences
Bleaching at individual scale

B) THE BLEACHING PHENOMENON AND RELATED REACTIONS

Bleaching: loss of symbionts or of pigments ?
In situ observations of release of symbionts
Zooxanthellae and histopathological observations
Diseases
Calcification of corals
Shell abnormalities in large foraminifers
Unusual colors and secondary pigments
Polyps behaviors

C) SPATIO-TEMPORAL PATTERNS

Global time patterns
Local time dynamic of bleaching
Spatial patterns
Mass bleaching is a recent phenomenon, becoming chronic

D) CONSEQUENCES

Mortality, recovery and long-term consequences
Adaptation and/or selection of host and symbionts
Extinction
Recruitment
Diversity
Ecological interactions
Erosion
Geological comparaison

2) CAUSAL FACTORS

A) IN SITU OBSERVATIONS

Diseases
Temperature
The East Pacific 1982-1983 El Niño event
The Great Barrier Reef event in early 1982
In situ data
Satellite data
Dolldrum time
Light and UV
Others factors
In situ bleaching with obvious local causes

B) LOCAL REEF FACTORS

Water characteristics
Extreme warm temperatures and high salinities
Acidity
Water agitation
Water stratification
Warm oceanic patches and rings

C) LINKS WITH GLOBAL CHANGES

Warming
Other climatic parameters
El Niño-Southern Oscillation
Ultra-violets
Carbon dioxide
Nutrients
 
 
 

ABSTRACT

 
Since the early 80's, it has been observed a global and massive dysfunction of major reef organisms. It affects not only corals but also all the other animal-algae symbioses. The loss of symbionts and/or their photosynthetic pigments brings a discoloration, hence the common denomination of "bleaching".
Hereunder is a bibliographic synthesis on this badly understood recent mass bleaching of reef photosynthetic symbioses, inattended to be the most comprehensive possible, from biochemical aspects to global perspectives. Meanwhile it is advocated that CO2 rise is the cause of this phenomenon through symbionts photoinhibition mechanism.
 
Key-words : Bleaching, reef, bibliography, CO2, photoinhinbition.
 
INTRODUCTION
 
This bibliographic synthesis is aimed at i) giving an overview of the problem in its globality; ii) providing support to reflexion and design of future research; iii) making easier access to informations for scientists who will join people already working on the subject, as it is clearly needed. Updated versions will be regularly delivered by electronic media (http://www.essi.fr/~sander/articles/Misc/Coral_Reef.html, also from the Internet reef site, http://coral.aoml.noaa.gov).
 
Observations of mass bleaching made in situ were collected by Williams and Bunkley-Williams and already excellently summarized by them in 1990. Their synthesis is only completed by more recent and related works judged noteworthy to be treated. Relevant informations on local and global factors are examined. All the physiological and biological aspects of the question are specifically developed. It is attempted not only to treats direct data on the bleaching phenomenon, but also to signalle works which are potentially relevant to it, in spite of the danger of this approach. This synthesis may appear as a frustrating fragmentary puzzle, as our state of knowledge is. It is rather a collection of details, each of which could be important for future understanding. A special attention is given to carbon process, even though at the edge of our current knowledge.
 
Mass reef bleaching is unique amongst the many alarming threats on earth ecosystems: it is the only one where the primary level of an ecosystem may collapse all over the world, and for reasons which are yet unclear, but certainly global. This makes mass bleaching a prioritary concern.
 

FIRST PART : FIELD KNOWLEDGE

 

I) IN SITU OBSERVATIONS ON MASS BLEACHING

A) SPECIFIC AND INDIVIDUAL PATTERNS

Organisms involved
The most stricking feature of recent mass bleaching is that it affects probably all and quasi-exclusively photosynthetic symbioses in reefs (last review by Williams and Bunkley-Williams, 1990a, thereafter W90). It is known with certainty that bleaching affects symbioses of hosts ranging from cnidarians, sponges, mollusks to foraminifers associated with symbionts which are either dinoflagellates, diatoms or cyanobacteria:
 
- all symbiotic cnidarians: hard corals (scleratinian) and fire corals (Millepora), black, blue and soft corals, anemones, gorgonians, alcyonaceans and zooanthids. There was even once a "marked discolouration of many small individuals" of the free-floating zooxanthellate shyphomedusan Phylloriza punctata. This was observed in Puerto Rico in 1990, in an inshore coastal lagoon, for the first time since they were studied in 1986 (J. R. Garciá, in Goenaga and Canals, 1990). Cnidarian'symbionts belong to the dinoflagellate genera Symbiodinium ("zooxanthellae"). They live in so-called perialgal vacuoles inside the endodermic host cells, and also, in the particularly bleaching-sensitive zoanthid genera Palythoa (as well as in Protopalythoa and Isaurus) in their outer epidermal cells (Trench, 1971, 1974);
 
-Tridacna bivalves, although not very currently (W90, Yellowless et al., in press). Their zooxanthellae live in diverticula of the digestive tract, which is well irrigated by the blood system (Norton et al., 1992, abstract, Rees et al., 1992, abstract). No case of bleaching of the less current symbiotic bivalve Hippopus have been yet reported, nor of other molluscs/algae cryptic symbioses such as those of nudibranchs (Crossland and Kempf, 1985) or Strombus gigas and other gasteropods (Berner et al., 1986a, 1986b);
 
- sponges belonging to seven different orders (W90, in annex), and associated either with cyanobacteria (the red-colored phycoerythrin-bearing cyanobacteria Aphanocapsa living in the mixotroph Xestospongia muta, and Petrosia pellasarca, Aplysina sp. Spheciospongia vesparia) or zooxanthellae (paling noted in Anthosigmella varians) (Vincente, 1990, Dennis and Wicklund, 1993). Aphanocapsa and zooxanthellae are the only sponge symbionts living intracellularly, inside specialized cells called cyanocytes and in "bubble-shaped" cells (Wilkinson, 1982, 1992, Rützler, 1990, and ref. herein). Aphanocapsa are also found in the intercellular collagenous matrix (mesohyl), as do others symbionts (chlorophytes, diatoms or red macroalgae), in sponges not yet known to bleach. Bleaching of sponges appears rather uncommon: the survey of Vicente (1990) reported a frequency of 10-30% of bleached X. muta on hard ground at 4-15m depth. Some specimens of the symbiotic sponge Mycale laevis (or rather the symbiotic M. laxissima, V. Vicente, com. pers.) in Puerto Rico in 1987 changed of color from bright orange to dull grey, apparently not due to zooxanthellae pigment change (?), and they survived (Bunkley Williams and Williams, 1988, Bunkley Williams et al., 1991). Another sponge, Aplysilla sp., changed from yellow to blue in Bahamas, 1987, but this was interpreted as a mere consequence of their death (R. Wicklung in Williams and Bunkley-Williams, 1989). This genera lives associated with Aphanocapsa (Rützler, 1990), and Prochloron has been observed as an incidental symbiont on the outside of this sponge (Parry, 1986). In addition, all sponges contain non-photosynthetic bacterial symbionts, the status of which remains unknown during bleaching episodes;
 
- at least one genera of a large foraminifer, Amphistegina, with diatom endosymbionts, extensively bleached in Florida, 1991 till 1996, and worldwide in 1992 with the four species A. gibbosa, A. lobifera, A. lessonii and A. radiata in vicinity of many bleached cnidarians (Hallock and Talge, 1993, Hallock et al., 1995, Talge et al., in press). A first observation of bleaching was made in Bahamas, 1988. Amphistegina are seemingly more prone to bleaching than adjacent corals (Hallock and Talge, 1993, tab. 4) and bleached in 1992 while corals not (Hallock et al., 1995). Unexpectedly, it appears that the peak frequency of bleaching is decreasing, from 20-80% in the 1991 to 20-50% in 1995, though with greater variability (Talge et al., in press). Standing crop fell sharply, also in the Gulf of Aqaba (J. Erez, com. pers.). A great shift in dominance from large symbiotic foraminifers to small heterotrophic ones has been observed between the decades 1960 to 1990 off Florida, rather related incorrectly in our opinion to eutrophication rather (Cockey et al., 1996), as between 1974 and 1989 in Mauritius (Hottinger and Pêcheux, 1991). There, about six months after mass bleaching, white but living, moving, Heterocyclina tuberculata were collected from 70m depth, and many other areas were devoid of living large foraminifers while fresh shells were abundant in sediments (pers. obs.). Informations on other groups of large foraminifers would be interesting as they harbour various, sometimes more than one, symbiont groups (dinoflagellates among them Symbiodinium, rhodophytes, chlorophytes or isolated chloroplasts, Leutenneger, 1984, Lee, 1990, 1992a, 1992b, Lee et al, 1980, 1982). Many show nowadays shell abnormalities (see below). Cyclorbiculina, with chlorophyte symbionts, displayed abnormal orange colors (P. Hallock, com. pers.). Mass bleaching of symbiotic planktonic foraminifers must also be logically envisaged (Hallock and Talge, 1993, Hallock et al., 1995, J. Erez, com. pers.). Large foraminifers are "giant" unicellulars of a few millimeter size at most and it is not surprising if they have been overlooked in census, although they may produce more calcium carbonate than corals themselves (Smith et al., 1985, Tudhope and Scoffin, 1988, and pers., in prep.). Their symbionts are, as in corals, harboured inside host vacuoles, or free in cytoplasm in the cases of Peneroplis associated with rhodophytes and Elphidium with isolated chloroplast ;
 
- no observations are yet available on the last but ecologically minor photosymbiosis, ascidian associated with cyanobacteria or with the so particular cyanobacterial-eukaryotic intermediate Prochloron, living extracellularly. In Curaçao, ascidians increased by nine fold over 1978-1993, perhaps because of pollution (Bak et al., 1996) but they are probably non-symbiotic ones (R. Bak, com. pers.).
 
One non-symbiotic coral, Stylaster roseus, was observed to discolour in Mona Island (Puerto Rico). This population was thereafter drastically reduced (C. Kontos, in Williams and Bunkley-Williams, 1989, W90).
Whereas Merlen (1985) remarks that in Galapagos, July 1983, the only points of color were from "few beautiful yellow-orange Tubastrea" (thus probably T. tagusensis), this endemic non-zooxanthellate species discoloured and may even have diseappered (Glynn and De Weerdt, 1991). The closely related but non colored T. coccinea was unaffected (Robinson, 1985, Glynn et al., 1985a). It must already be emphasized that a wide range of organisms were affected in Galapagos, 1983, with water characterized by temperature of at least 2-3°C higher than normal. Associated with mass bleaching, but restricted to heated reef flats, death of urchins and mollusks is also reported by Tsuchiya et al. (1987) in Okinawa, 1986, and together with some mortality of Strombus gigas, and abnormal behavior of mollusks, of echinoderms and of polychaetes in Florida, 1987 (Berg in Jaap, 1988). Other reports noted mortalities or unusual behavior of mollusks and urchins, and occasionally of bryozoans, coralline algae, tunicates and polychaetes, in Florida, 1983, California and Kenya, 1987, and Okinawa, 1988, maybe also in inshore area (W90).
 
Almost no discoloration of photosynthetic non-symbiotic organisms, i.e. algae, are signalled. Mention of severe damage to Thalassia near a thermal effluent was done by Jokiel and Coles (1974), and also of death and/or bleaching (?) of sea grasses in Florida Keys, 1983 (Causey, in W90). Mastaller (1979, Diss., in Mergner, 1981) described a "quick massive expulsion of brownish dyes" by an entire brown algae population, subsequently dying off, after a daily increase of 8.5°C over the reef flat in the Gulf of Aqaba.
 
Interspecific differences
Williams and Bunkley-Williams (1989) and W90 listed around 89 species that bleached, and no new information of interest is to be added to their analysis. There is great variability between sites and events. To summarize it briefly, the most currently affected coral species in the Caribbean zone are Montastrea annularis (more than M. cavernosa), the 2 Millepora species (M. alcicornis more than M. complanata, but see Sandeman, 1988, Losada, 1988), and the various Agaricia sp. in deeper water. Acropora palmata is commonly cited but is rather resistant, more than A. cervicornis. The zoanthid Palythoa caribbea emerges as the most sensitive species. Alcyonarians are also strongly affected (Glynn, 1993). All those species have high cover in reef. Glynn (1988a) remarked that in Atlantic realm, massive corals seem more affected than branched ones (but see W90), whereas the reverse appears to be true in the Indo-Pacific, in particular with the bleaching of the fast-growing Pocillopora and Millepora. Newton (in W90) pointed out that the most extensively bleached host Agaricia lamarcki in Bonaire, 1987, is a coral of cooler water origin.
 
Available data confirm the general rule that there are crude correlations between the various indicators of species sensitivity: apparition and speed of bleaching, percentage of affected colonies, percentage of bleached surface of colonies, percentage of mortality and time of recovery. Exceptions exist: in Panama, Millepora was the most extensively bleached but recovered well therafter (Lasker et al., 1989). Noticeable deviations concern mostly time dynamics : Acropora are often those which bleach the first but are not the most affected (W90), Porites panamensis may show delayed response (Glynn, 1984), Pocillopora had immediate, but lower mortality rate than Acropora in French Polynesia, 1991 (Salvat, 1992).
 
Intraspecific differences
Variability between individuals within the same species is very great: often two similar specimens side by side, one bleaching, the other not, are observed. It is supposed to result from the genetic variability of the symbionts (Gladfelter, 1988), even whithin a single colony (Sandeman, 1988) or of the host, as showed by clone neighbour analysis (Edmunds, 1994). Genetic studies suggest that bleached heads of Porites compressa in Kanehoe Bay, summer 1988, were clone-mates of a genotype that is extremely sensitive to higher temperature (Hunter and Kinzie in Jokiel and Coles, 1990). Sensitivity to high light (500-1300µE/m2.s) of healthy zone of partially bleached Agaricia tenuifolia was greater than in normal colonies, as measured by chlorophyll fluorescence (Lovelock et al., 1996). In Jamaica, Goreau and Macfarlane (1991) suggest that dark colonies of Montastrea annularis are more resistant (this is opposed to the finding of Coles and Jokiel, 1978, in experiments, see below). Thin walled, fused tridents Porites lutea sub-group did not bleached in summer 1991 in Thailand (Tudhope et al., 1992). Clear examples of difference in intensity of bleaching of the same species among localities are given by W90 for the sponge Cliona and the gorgonian Briarium (cf. also Lang, 1988).
 
Harriott (1985) states that mortality of smaller individuals was higher. Survivorship was greater in large massive coral colonies six years after the El Niño 1983 event in East Pacific region (Glynn, 1989). In Moorea, French Polynesia, during bleaching in 1991, it was observed diseappearance of more juvenile Pocillopora than adults, but this has perhaps other causes (Gleason, 1993). The opposite pattern of size-dependent bleaching of fungus corals in Indonesia, 1983, was explained by their spatial distribution, as big specimens migrate to the more affected shallow outer reef (Hoeksema, 1991). Preferential death of small corals was also observed at first after a hurricane (Knowlton et al., 1981). In the large foraminifer Amphistegina, bleaching was more frequent in large adults, and small forms less then 0.6 mm were seldom affected (Hallock et al., 1995, Talge et al., in press).
 
In Jamaica, the same M. annularis colonies bleach from years to years, with others in addition (Goreau, 1990); in Rosario, Colombia, the few shallow Acropora cervicornis survivors of the 80's events seemed unaffected in 1987 (Lang, 1988), whereas in East Pacific, coral still alive after the 1983 El Niño bleached again in 1987, though they recovered in 4-6 weeks (Glynn, 1989). In the Flower Garden Banks, Gulf of Mexico, the same colonies often bleached each summers 1989-1991, and with the same pattern (Hagman and Gittings, 1992). The same sponge colonies affected in Bahamas, 1987, bleached again in 1990, and colonies of A. cervicornis and Porites, also bleached in 1987, died in 1990 (Dennis and Wicklung, 1993). After 4 years of weekly survey in Oahu, Hawaii, C. Hunter (Internet reef site) recognizes some colonies of deeply pigmented Porites and encrusting Montipora which will bleach for few weeks every spring and fall.
Another interesting observation is that individuals of M. annularis, A. agaricites, Mycetophyllia and other species with a particular orange-red color at Negril, Jamaica, a color unusual in other parts of Jamaica, did not appear to be affected by bleaching (T. Goreau, unpublished). This suggests that the presence of some pigments, as carotenoids in symbionts, would make them resistant to bleaching. It seems that no short-term adaptation exists, as would be expected with (i) existence of resistant symbionts and (ii) their selection either from previous host population or free living ones during bleaching and recovery.
Bleaching affects also the coral larvae : in Bonaire, September 1995, Morse (1996) observed discolouration of Agariciid larvae, which was correlated with parent one.
 
Bleaching at individual scale
It is quite variable, from paling, bleached blotches, to total bleaching. The surface area bleached ranged generally from 20% to 100% (W90). The most coherent pattern is in relation to light, as bleaching affects very preferentially the upper faces of corals and/or let shaded parts intact (Glynn, 1983, Jaap, 1985, Harriott, 1985, Fisk and Done, 1985, Robinson, 1985, Woodley, 1988, Lang et al., 1989, Williams and Bunkley-Williams, 1989, Goreau and Macfarlane, 1990, Goenaga and Canals, 1990, Gates, 1990, Yap et al., 1992, Tudhope et al., 1992). Fine stripes of dead area along ramets of Acropora after a bleaching event in Mauritius, 1989, indicated very precisely the zenithal direction (pers. obs.). But opposite pattern also exists (Woodley, 1988, Tudhope et al., 1992). Goenaga et al. (1990) counted 64% of corals bleached on upper part, and 5% on lateral faces. Shaded colonies bleached only at the end of the 1983 event (Glynn, 1989) or when transplanted to illuminated places (in Ogden and Wicklung, 1988). Notable exceptions to this rule is bleaching on the undersides of Agaricia lamarcki from recessed cave environments (Porter et al., 1989), and of the gorgonian Pacifigorgia under boulders, in a surge channel (Robinson, 1985). Gorgonians were also sometimes white-striped above and below their ramets (W90).
 
Branched corals may have their peripheral ramets more affected (Porites elegans, Glynn, 1989) or bleaching can be variable between branches (Woodley, 1988). The tips are often the first and the most affected (Glynn, 1984, Lasker et al., 1984, 1989, Glynn et al., 1985b, Sandeman, 1988, Gates, 1990, Choquette, Reyes-B. in W90, Rougerie, 1992, Drollet et al., 1995), or verrucae in P. eydouxi (Drollet et al., 1994, 1995), but Acropora in Andaman Island bleached from their base toward their tips (Wood in W90). A colony of A. cervicornis in a public aquarium in St Thomas bleached in one day from bottom to top (Nunn in W90). A. palmata has often large, white, irregular blotches (Williams and Williams, 1988). Tabular Acropora bleached from outside to inside, A. valida in patch, and A. gemmifera only in the coenosteum and not at the branches (Drollet et al., 1994, 1995). Death of the upper part of Acropora was also reported following a cold event in Arabian Gulf (Coles and Fadlallah, 1991).
 
In massive corals, M. annularis bleached from base to upper face at San Blas in 1983 (Lasker et al., 1984). Less bleaching was noted by Glynn (1984) and Hof (in W90) in fissures and or depressions but Jaap (1988, and in Holling, 1988) indicates an oppposite pattern. Bleaching patterns of M. annularis differed with depth in Jamaica (Sandeman, 1988): shallow ones had bleached areas either in ridges or valleys, on top or side of pillars. Some colonies had a white 2cm ring at their edge. Below 10 meters, they were affected only at top, and below 20 meters no bleached colonies were seen. M. annularis in Mexico after the fall 1995 event were still bleached on top six months later whereas sides had recovered (R.E. Rodriguez, Internet reef site).
 
Flat corals as Agaricia and less commonly Leptoseris may become white from both the edges and the center (Faure et al., 1984, Bunkley Williams and Williams, 1988) or from edges toward center (in Williams and Bunkley-Williams, 1989), or in ridges (Gates, 1990). In Florida, 1987, as well as in Jamaica, deeper colonies were found striped (Jaap, in Holling, 1988, Jaap, 1988, Goreau, 1991). Bicolor patterns in recovering corals is signalled by Newton (in W90).
 
In Bahamas, among two P. asteroides which bleached in 1991 and again in 1992, one bleached at the opposite side of previous year (Lang et al., 1993). A strange pattern of bleaching was carefully reported by Kobluk and Lysenko (1994), after a cold event in Bonaire, June 1992, with temperature lowering only from 27°C to 25°C, in Agaricia agaricites with "ring" bleaching circling unbleached area of centimeter scale, somewhat located on ridges.
 
The large foraminifer Amphistegina were "mottled" to some degree, ranging from one or more anomalous white spots to nearly bleached with a few remaining browns areas (Hallock and Talge, 1993). Bleached chambers were not reoccupied by symbionts (as in DCMU experiments) but new added chambers were colored (Hallock et al., 1995).
 
Some bivalvesTridacna gigas bleached only in the central portion of their mantle (Goggin in W90).
Bleaching in the sponge Xestospongia muta in Belize in 1996 usually starts from the base of the sponge and gradually works itself up, until the whole sponge is completely bleached out. Then the sponge just crumbles apart (S. Paz, Internet reef site).
 

B) THE BLEACHING PHENOMENON AND RELATED REACTIONS

Bleaching: loss of symbionts or of pigments ?
Loss of zooxanthellae was observed under microscope (Glynn, 1983, Faure et al., 1984, Lasker et al., 1984, Glynn et al., 1985b). Jaap (1985) observed 2 specimens virtually devoided of pigments. Some symbionts are always still present except in extremely necrotic coral (Glynn, 1989), and even in the most highly bleached corals, Jokiel and Coles (1990) could always detect a few live zooxanthellae through fluorescence microscopy. Bleached part of one M. annularis colony from Cayman had only 55%-36% of zooxanthellae in June 1988, 9-12 months after bleaching event, and 88% in December 1988 (Hayes and Bush, 1990).
 
Different cases were observed:
- both reduction of zooxanthellae density together with a decrease of pigment content per zooxanthellae : only 10% of symbionts in bleached M. annularis and Diplora strigosa of St Croix, 1987, with less chlorophyll a and c per symbionts in M. annularis (Gladfelter, 1988). Four colonies of M. annularis (from Florida, collected 5-6 months after bleaching in Dec. 87), had 27% symbionts, themshelves with a lower level of pigments (chlorophyll a 36%, chlorophyll c 9%, peridinin 20%, diadinoxanthin 16%) (Kleppel and al., 1989). Porter et al. (1989) measured in bleached specimens of M. annularis and A. lamarcki from Florida, 1987, collected 5 months after bleaching, a reduction to respectively 14% and 43% of zooxanthellae number, together with a reduced content of chlorophyll a at 52% and 34% of normal level. Coral biomass decreased to respectively 39% and 73%, photosynthetic rate to 17% and 74% at saturating light level, but respiration decresed only slightly. One surprise of their results is that, on a chlorophyll a basis, photosynthesis was multiplied by about 2.5 in bleached M. annularis and by 5 in A. lamarcki. In the study of Brown et al. (1995) of six species just 5 weeks after bleaching in Thailand, 1991, the counts of zooxanthellae number were very low, few percent to at most 50% the normal level, and particularly marked in oral tissue. Pigments analysis 4 months later measured less chlorophyll per symbionts (0.4-2.1 versus 1.5-4.6 pg chl);
 
- only a loss of zooxanthellae, down to 25% of normal level, in 2 colonies of M. annularis of Florida, 1987, collected 9-12 months after bleaching (Szmant and Gassman, 1990);
 
- more zooxanthellae (about twice) with less pigment (one tenth) in a third colony;
 
- less zooxanthellae (19%) with more pigment (+47%, n=8, p<0.05) in Seriatopora hystrix of Lizard Island, 1987, collected maybe up to 3 months after bleaching (Hoegh-Guldberg and Smith, 1989). Bleached Stylophora pistillata from the same place were more "orthodox" with only loss of symbionts down to 29%.
 
Lovelock et al. (1996) measured chlorophyll fluorescence of bleached Agaricia tenuifolia, which was only 2% normal level. A surprise came from the quantum efficiency (Fv/Fm) which was identical of that of brown colonies. Jaap (1985) observed an increase of the chlorophyll a/b ratio in corals bleached in Florida, 1983. At the same place, in 1987, this ratio rose from about 2.6 to 10 (Kleppel and al., 1989, if chlorophyll a of endoliths is not taken into account, as they represent generally a contamination of less than 10%). Peridinin declined from 34% to 18% of chlorophyll a on a weight basis, whereas the slight increase of one of the xanthophyll cycle pigments (involved in photoprotection, see below) is to be confirmed (ratio of diadinoxanthin/chlorophyll a+c from 11% to 16%).
 
In addition, whitening of some corals in the field may be the result of strong tissue retraction, as seen after subaerial exposure with 10% reduction of absorbance though aspect (Brown et al., 1994b). One morph of Montastrea annularis with naturally pale polyps may have been classified as bleached according to Knowlton et al. (1992). Digitization of color photographies of colonies in Bahamas showed that bleaching increases the brightness but does not change the color spectrum (Lang et al., 1993).
 
In conclusion, the data are quite controversial. Loss of either symbionts or pigments may be secondary phenomenons, but the long time lag between the bleaching events and the analysis are more probably responsible for these discrepencies, as already advocated by Hoegh-Guldberg and Smith (1989). They consider the possibility of an initial decline in both the pigment content and population density of zooxanthellae, followed at first by a recovery of only the pigment content. In addition, natural daily variation in pigments may be important, of the order of 50%, according to Titlyanov et al. (1991), and there is more color in winter (Lang et al., 1993).
 
 
Table 1: Zooxanthellae loss and/or pigment loss in bleached corals in situ.

Zoox. number

Chlo /Zoox.

Species
Place and date
of bleaching
Delay of analysis
(months)

References

10%

less
M. annularis
D.strigosa

St Croix, 1987

?

Gladfelter, 1988

36%

?

M.annularis

Cayman, 1987

9-12

Hayes and Bush, 1990

27%

9%-36%

M. annularis

Florida, 1987

5-6

Kleppel and al., 1989

14%

52%

M. annularis

Florida, 1987

5

Porter et al.,1989

43%

34%

A. agaricites

Florida, 1987

5

Porter et al.,1989

27%

normal

M. annularis

Florida, 1987

9-12

Szmant and Gassman, 1990

200%

10%

M. annularis

Florida, 1987

9-12

Szmant and Gassman, 1990

19%

147%

S. hystrix

Great Barrier

3

Hoegh-Guldberg and Smith

29%

normal

S. pistillata

Reef, 1987

3

1989

50%

few

6 species

Thailand, 1991

1

Brown et al., 1995

1/4-1/2

6 species

Thailand, 1991

1

Brown et al., 1995

In situ observations of release of zooxanthellae
Massive expulsion of zooxanthellae is at least proved in some few in situ cases. The first description was made by Fankboner and Reid (1981), during several incoming very warm tides, giving a burning sensation, on inshore area of Eniwetak Atoll, Aug. 1979, in the form of a cloudy-green vertical front, about 2 m deep. Microscopic examinations revealed numerous zooxanthellae along with bits of filamentous algae, miscellaneous protozoa, polychaete setae and molted crustacean cuticules. Samples of a less opaque cloud contained 5100 zooxanthellae per liter. Large dense yellowish brown clouds, up to 2 m thick, or golden yellow mucus above corals, were reported just prior mass bleaching events, and are surely expulsed zooxanthellae (Bunkley-Williams and Williams, 1987, 1988, 1989, Goenaga and Canals, 1990). They were said "dead and dying" by Bunkley-Williams and Williams (1990b), without any further information.
 
Zooxanthellae and histopathological observations
Coral tissue loss or sloughing has been rarely seen: in Uva Island, 1983 (Glynn et al., 1985b), in Montastrea annularis and Diploria spp. in St. Croix, 1988 (Gladfelter, in W90), as well as in the former species in Florida Keys and Puerto Rico in 1987 (Hudson, 1988, Goenaga et al., 1988), and in Porites evermanni in Hawaii, 1987 (Jokiel and Coles, 1990).
 
Desappression of thylakoids is the first observable event according to M.D.A. Le Tissier (com. pers., 1994). Degenerated zooxanthellae are commonly observed, but, according to Glynn et al. (1985b), only after animal necrosis, at contrast to Brown et al. (1995) observations. They are described as vacuolized (Glynn, 1989), with abnormal empty appearance (Lang, in Holling, 1988), or as "empty, devoid", with unstained vacuoles, without densely stained granules, some with a central condensation and peripheral cytoplasmic clumps (Hayes and Bush, 1990), with loss of circularity, vacuolization around them, and many empty vacuoles (Brown et al., 1995). Wide perialgal space was also observed in "solar bleaching" (Le Tissier and Brown, 1994). Pyrenoids and assimilation bodies were still visible, except in a few ones of the oral disk (Szmant and Gassman, 1990). In the transition zone from bleached to unbleached parts, accumulation bodies were hypertrophied (Faure et al., 1984). Chloroplasts were reduced or absent, lipids lacking and membranes ruptured (Jaap, 1985). Host lipid reserves were seen to increase in mesenterial area (Brown et al., 1995). Symbionts may be localized in the base of the polyp and in mesenterial filaments, either as "refugees" or in course of expulsion (Szmant and Gassman, 1990), or simply due to more reduction in oral than mesenterial or basal areas (Brown et al., 1995). Those authors also observed zooxantellae in mesoglea ; and, as in some laboratory experiments, symbionts were often in dividing state. No damage of perialgal membrane is reported.
 
Tissue of bleached corals shows general atrophy and necrosis. There is 30-50% less tissue per surface, with a normal C:N ratio (Szmant and Gassman, 1990). Invasion of cell by ovoid dark inclusions at the base of endoderm was described by Faure et al. (1984). Jaap (1985) observed abnormal mitochondria in coral cytoplasm, with mucoid-polysaccharide material. Mucus secretory cells increased in number and size (Brown et al., 1995), even in healthy appearing Agaricia (Lasker et al., 1984) and in Pavona clavus, but decreased in P. gigantea and P. varians (Glynn et al., 1985b, Glynn, 1989). Based on staining with pentachrome, a change of mucus composition and of pH in Porites damicornis, P. panamensis, Psammocora stellata, and Pavona clavus must have occurred, together with formation of basophilic "blobs" (Glynn et al., 1985b). One "healthy"-looking Pocillopora was in early stage of necrosis (loss of architecture, basophilic tinge in mesoglea), from which it was concluded that the problem is on the animal side, with maybe thereafter nutrient-starvation of zooxanthellae (Glynn et al., 1985b). Gonads are reduced and reproduction is generally impaired, according to health state (Glynn et al., 1985b, Glynn, 1989, Szmant and Gassman, 1990). Bleached corals had half normal lipid level (Glynn et al., 1985a). Phenoloxidase, a biomarker of immune capability, was found to have lower activity in bleached and semi-bleached M. annularis (Smith, 1992, abstract). Large foraminifers diplayed abnormal number of large vacuoles and lysosomes surrounding deteriorating symbionts (Hallock and Talge, 1993, Talge and Hallock, 1993).
 
Diseases
Secondary parasites were observed in a few cases, with massive invasion of fungal hyphae in M. complanata, but not in Acropora nor Palythoa (Te Starke et al., 1988) or in one colony with ovoid granular basophilic bodies similar to those of the "White Band Disease" (Lasker et al., 1984). Coccoid and rod-shaped bacteria-like objects were observed in the gastrodermis of bleached corals of East Pacific, 1983, sometimes present in the vacuoles vacated by zooxanthellae (Glynn, 1989). There is no transmission of bleaching following iso-, allo- and xenografts, i.e. between parts of the same colony, between specimens of the same species or different ones (Glynn, 1983, Glynn et al., 1985b).
Recently, this hypothesis gains support from the finding of Kushmaro et al. (1996, and oral com., Panama 1996) : in the Israel mediterranean coast, bleaching was observed in summer in Oculina patagonensis (a very strange coral, see Zibrowius and Ramos, 1983). Bleached rim zone was infected with Vibrio-like bacteria, which were cultivated. They revealed to be potent bleaching agent within one week to one month according to aquarium inoculation. Antibiotics protected from bleaching. Moreover bleaching was effective at 25°C but not at 16°C, in resonance with worldwide "warming" bleaching. Such a peculiar restricted phenomenon was not unexpected. Although impressive, these data can be taken in account only when some similar finding are found somewhere else, and more crucially if infection can be carried out on other bleaching taxons.
 
Calcification of corals
During bleaching, there is no visible calcification, as seen with alizarine marks (Glynn, 1983, Glynn, 1989), nail technique (Goreau and Macfarlane, 1991), band analysis (no stress band, Leder et al., 1991, missing or great reduction of deposit, Porter et al., 1989, Reese et al., 1988) or direct caliper measurements (at most very reduced growth in coral transplanted to nearshore area, Shinn, 1966). At contrast, Tudhope et al. (1992) still noted calcification in 3 weeks Alizarine-stained bleached corals, albeit extremely limited. Risk and Pearse (1992) were able to observe daily growth bands in one of the few corals of the west coast of Costa Rica which survived the El Niño 1983 bleaching event. It grew 22µm per day before the event, 9µm at maximum in 1983-1984 with possible total cessation of growth, and recovered to 18µm after 1985. Rougerie et al. (1992) suggest that fluorescent colonies may continue to calcify. Calcification can still be negligible even 6 months after recovery of normal color, suggesting that bleaching inhibits calcification more than photosynthesis (Goreau and Macfarlane, 1991). About one year after bleaching event, growth rate of M. annularis was still null for bleached ones (or 37% for the colony n°67 with twice more zooxanthellae), 67%-93% for recovering ones, and 81%-98% for those that did not bleach (Leder et al., 1991).
 
Thus no clear signals was given by isotopic carbon and oxygen deviations in skeletal bands. d18O from Porites lobata of Isla del Caño records a +2°C temperature anomaly during the 1982-1983 El Niño, corresponding to 32°C, whereas in situ temperature in January 1983 was only 30°C in shallow areas and 29°C at 15m depth. Salinity interference was postulated (Carriquiry et al., 1988). In Florida, 1987, the 0.25-0.5°% lighter d18O, corresponding to 0.5-1°C warming, was seen very only in bleached colonies and unexpectedly not in the unbleached ones (Porter et al., 1989). Leder et al., (1991), studying also M. annularis colonies of Florida, 1987, have not seen a d18O temperature signal in normal colonies except a little one in three affected colonies, but again perhaps because of a salinity effect. A 1°% lighter d13C in bleached specimens can be interpreted as an effect consecutive to lower growth rate (Porter et al., 1989), whereas Leder et al. (1991) interpreted the absence of d13C signal to opposite effects of photosynthesis reduction and of lower growth rate. The colony n°67 (with twice more zooxanthellae) has a d13C 1-2°% heavier than all others. Isotopic deviation and growth rate were good but not perfect indicators of past ENSO events, because of would-be cessation of calcification and of other isotope excursions not linked to ENSO (Druffel et al., 1989, and ref. therein, Quinn et al., 1993, Dunbar et al., 1994). Ten growth discontinuities were identified in a record since 1587 (not always correlated with ENSO), but included the one associated with the 1973 ENSO without bleaching, and none presented bioerosion nor encrustation observed after the 1983 mass bleaching (Dunbar et al., 1994).
 
Shell abnormalities in large foraminifers
At contrast to corals, large foraminifers enregister bleaching event in their calcification. Spectacular shell abnormalities were encountred in various genera in Mauritius, 1989, six months after a bleaching event (cf. Hottinger and Pêcheux, 1991) (see Annex V). In Florida 1991 to 1996, Amphistegina showed calcification defects, with extensive shell damages and deformities: breakage (3-29%), margin chipping, loss of outer chambers, twisted, elongated or conjoined tests, tests with two apertures or distended embryo (6% or less) (Hallock and Talge, 1993, Hallock et al., 1995, Talge et al., in press). Strong abnormalities were also observed in Amphistegina lessoni and A. bicirculata from Palau on the front reef, 30m-60m depth, in October 1995 (J. Hohenegger, com. pers.) and in a small sample, in Amphistegina, Heterostegina and Sorites from Caribbean Panama in few meters depth in July 1996 (pers. obs.). We are currently monitoring a population of S. variabilis near Nice, France. They display 17.9%-27.3% of abnormal forms (supplementary plans, irregular periphery, twisted disk, twin specimens), but we could not find old samples to be sure it is a new phenomenon (in prep.).
 
Those observations are of utmost importance. Minor irregularities in chamber formation were quantified in normal population of Operculina (Pêcheux, 1995b), and they appears as quite distinct phenomenon. Shell abnormalities of symbiotic foraminifers were known in tidal pools or very shallow water with temperature, salinity, oxygen and pH excursions (cf. Reiss and Hottinger, 1984), in culture (Röttger and Berger, 1972, pers. obs.) or in small foraminifers in heavily polluted areas (Venec-Peyre, 1981, Yanko et al., 1994, see reviews in Boltovskoy and Wright, 1976, Alve, 1995) and in planktonic foraminifers believed to be under stress (Berger, 1972), but never in the recent past in pristine areas bathed by stable oceanic open waters. It prooves that bleaching is a new phenomenon.
 
Moreover, to my knowledge and of many consulted specialists as well as extensive bibliographic search, such high frequency of deformities was never observed in geological time. For example, we checked our collection of Mexican upper Cretaceous-lower Tertiary large foraminifers, containing circa 20 000 representative specimens (Pêcheux, 1984). Abnormalities accounted for about 4 per mil, almost always from special facies (very shallow, deltaic, transgression or emersion ones) and never comparable to certain monsters today observed. The only similar record is found just after the Cretaceous/Tertiary catastrophe, not in reef foraminifers which all disappeared but in planctonic ones, half of which was abberant during about 50 000 years (Gerstel et al., 1986), a time of considerable CO2 rise. These abnormalities indicate that bleaching is truly an unprecedent stress at planetary evolution scale.
 
Unusual colors and secondary pigments
Many observers report unusual blue, green, yellow or pink colors of "bleached" colonies (Goreau, 1990, Salvat, 1992), which colors were reported before mass bleaching only by Jokiel and Coles (1974) for local bleaching in the flume of a thermal effluent, mostly in Pocillopora meandrina. In particular Siderastrea sidera become lavander, lilac or blue (Jaap, 1988, Hudson, 1988, Williams and Bunkley-Williams, 1989, Bunkley-Williams et al., 1991, Rougerie, 1992). Montastrea annularis turned blue before bleaching, Porites asteroides and P. porites yellow (Williams and Bunkley-Williams,1989), and Montastrea cavernosa grey (Goreau, 1991). Some anemons turned bright yellow or rust, whereas the mouths of Agaricia and Leptoseris remained yellow (Bunkley Williams and Williams, 1988). There were some instabilities in this phenomenon, with changes of color from pink to blue or yellow to pink within weeks in Acropora (Rougerie et al., 1992).
 
Those colors are sometime characterized as "fluorescent" or "iridescent" (Goenaga et al., 1990, Rougerie, 1992). In Moorea, 1991, T. Goreau (unpublished) has observed that unbleached colonies, and "white" or yellow bleached ones have the same fluorescence, whereas none was visible in rose-pink, blue or violet bleached colonies. The fluorescent pigments were localized in the animal, and were soluble in water but not in acetone and are probably the mycosporine-like amino-acids. In bleaching due to thermal effluent in Hawaii, an increase of 330nm-absorbing pigments have been observed in bleached Montipora verrucosa but not in other two studied species and was said to be merely due to the amount of tissue loss (Jokiel and Coles, 1974). In Indonesia, 1983, among the mushroom corals, Heliofungia actiniformis was unaffected, perhaps because of its green fluorescent pigment, or alternatively because of its thick gastrodermis (Hoeksema, 1991).
 
Polyps behavior
Polyps response is variable, from normal behavior, to extention without prey capure, to total retractation (Williams and Bunkley-Williams, 1989, Lang in Holling, 1988) probably depending on the intensity of the perturbation. Fire corals (Millepora) may or may not inflict pain (Jaap, 1979, 1985, 1988, Losada, 1988, Goenaga and Canals, 1990). Abnormal expansion of polyps (Faure et al., 1984), or continuous extension night and day of normally night species (Glynn, 1989, Williams and Bunkley-Williams, 1989) were also noticed. Before mass bleaching, it has been observed that white Montastrea cavernosa (due to reduced light) expanded their polyps abnormally in the day (Wells et al., 1973). At opposite, Lasker (1979) found in 1979 one "diurnal" morph of M. cavernosa with a small bleached region. This region expanded its polyps at night instead of day, and recovered the normal behavior of the rest of the colony progressively, in parallel with pigmentation.
 

C) SPATIO-TEMPORAL PATTERNS

The remarkable work of W90 summarizes all informations made available till 1988, and the reader is invited to refer to it. Recent bleaching events are now signalized on the Internet reef site (http://coral.aoml.noaa.gov). Table 2 summarizes published works. An overview is given here for completude.
 
Time patterns
The first described mass bleaching event sensu stricto (with no obvious local cause) began in Bonaire (Leeward Islands, Caribbean) in June 1979 and ended in February 1980 (Hof, in W90). It occurred extensively on all but the windward coast of the island from 10 to 40 m depth. Bleaching occurred in summer 1980 in several areas in the Florida Keys (W90), and in the Great Barrier Reef (Oliver, 1985). Therafter, mass bleaching events are well reported, in early 1982 in the Great Barrier Reef, and in 1983 mostly in the Eastern Pacific, clearly associated with El Niño, and also in Indonesia, Japan, Mayotte, and Caribbean. Some mass bleaching occurred in 1986 in Hawaii, Mayotte, Puerto Rico, Barbados and perhaps Bahamas. In 1987-1988, the bleaching occurred worldwide, and was the "most severe, extensive and long-term bleaching ever [previously] recorded" (W90). Although a detailed synthesis is not yet available for more recent years, mass bleaching was reported in Caribbean and other Indo-Pacific regions from 1989 to 1995, 1990 being considered as major events in Carribean. Severe bleaching in Society Islands in 1991 is clearly related to the 1991-1992 El Niño. From all evidences, mass bleaching is more and more frequent and widespread, as in 1995 (Panama 1996 congress abstracts and Internet reef site).
W90 emphasized the cyclical pattern of mass bleaching in 1979-80, 1982-1983 and 1987-1988, and divided bleaching complex in preceding, main and following events. It appears that the 1979-80 cycle is far from evident, and that bleaching is becoming more continuous in the recent years. Once the connection with the El Niño phenomenon is discarded (W90, Atwood et al., 1992, 1996 and see below), the existence of worldwide cycles in 1982-1983 and 1987-1988 raises more interrogations on their origin than provides light on the cause of recent bleaching.
 
Local time dynamic of bleaching
With local heating on reef flats during low tides, the time response of bleaching may be as fast as a few hours. In most mass bleaching cases, it ranged from one week to months. In Looe Keys, Florida, bleaching occurs within a week or so once dolldrum weather sets (Causey, in Atwood et al. 1992). In Puerto Rico, September 1990, seas became unusually calm one week before bleaching (Goenaga and Canals, 1990). According to Dennis and Wicklung (1993), bleaching appears in Bahamas, 1990, more than a couple of days, but less than 18 days, after the raise of temperature. The rate of increase was lower than in other years, about 3.1°C in 14 days.
 
If one takes warming as the cause of bleaching, it is difficult to know when the temperature threshold, if there is one, is reached during a slowly warming period. Temperature records before bleaching are rare. The warming may be as long as 17 months (Glynn, 1989). In Florida, 1987, Cook (in Porter et al., 1989) recorded warming up to 0.5°C per day, and temperature was above 30.2°C during two weeks, and bleaching started 5 days later. (Cook et al., 1990). In East Pacific, 1983, bleaching began about one month after warming according to Glynn (1984), and after 30°C during three weeks and 32°C during one week at San Blas (Coffroth et al., 1989). In Indonesia, April 1983, bleaching took place 4-6 weeks after the end of the wet season and of the warming (Brown and Suharsono, 1990), and between one and two months after warming in Society Islands, April 1991 (Rougerie et al., 1992). In Jamaica, 30°C was reached at least in early August 1989, and bleaching becames apparent in early October (Goreau, 1990). The data of Kato (1987) show a warming from 25°C to 31°C in two weeks in a shallow area of Okinawa in 1986. Bleaching affected 30-40% of the colonies at the end of this period as well as one month later, but curiously not in between.
 
Bleaching may be sudden or progressive (W90), lasting many months (1 1/2 month in Australia, 1982, Oliver 1985, or increasing from July to November 1987, Goreau and Macfarlane, 1990) or lasting more than a year. Glynn (1989) observed a delay of 2-4 weeks between bleaching and death.
Increase of bleaching has been observed after its initiation, with delay of 4-7 months for Colpophyllia and of 8-11 months of Palythoa (Lang et al., 1992). Some corals have been seen to recover while others began to bleach, as Millepora in Bermuda, 1988 (Cook et al., 1990, and see W90). This could be interpreted as an acclimatation of the corals bleached during the first phase, or rather as the fact that they were already bleached and thus unaffected by the conditions which generate the second phase of bleaching.
 
Spatial patterns
Mass bleaching affects all regions. The only main exception today is the Gulf of Akaba, perhaps because of its low temperature. At lower spatial scale, the perplexing variability of mass bleaching was well analyzed by W90, and, as they stated, "few if any trends can be assembled which are not contradicted by some reports. The patterns and extent of bleaching seem to suggest a large number of local, unrelated, pratically unique events but they are too highly coordinated to be coincidental".
At reef scale, bleaching affects large continuous tracts, spotty areas or isolated colonies (Glynn, 1983, Lang et al., 1984, W90). There is no trends with depth. Bleaching occurs from near surface down to 90 m depth (although at a low level of 3%). It may be more intense in shallow waters, intermediate ones (12-15m) or deeper ones (10-55m). Opposite depth trends has been even observed in two adjacent sites in Jamaica (also CARICOMP, 1996). In some events, bleaching moves deeper with time, but toward shallower areas in others. In 1994, bleaching was more frequent in shallow waters in the North-West of Moorea, Society islands, but more in deeper waters in the North-East, 10 km apart. It seems to be induced by previous species distribution, in particular of the resistant Porites (Hoegh-Guldberg and Salvat, 1994). Bleaching is as often reported in inshore than in offshore zones. The few coherent indications of bleaching in grooves and channels is discussed later in relation to dense water formation.
 
Preferential bleaching on one side of islands is often reported, but with no emerging pattern, suggesting local circulation factors. Relative intensity of bleaching at Looe Key and at Key West in Florida was reversed from 1983 to 1987 (Jaap, 1988). No regional trend is clear, apart for the El Niño East Pacific 1982-83 event, but even in this region some places escaped bleaching (Glynn, 1989). W90 indicated that in January-November 1987, bleaching was centered in northern Caribbean, Bahamas and south Florida, perhaps because they are large platforms and/or with poor water circulation. Bleaching later expanded with less intensity to include all of the greater Caribbean Islands and eastern Pacific in November 1987-January 1988.
 
Mass bleaching is a recent phenomenon, becoming chronic
There is almost no doubt that mass bleaching events have appeared since the early 80's and tremedously increased in magnitude and frequency. One may argue that there were fewer observers at the begining of the century, but this objection certainly does not hold for the 60's and the 70's. It is significant that almost no mention of bleaching was made in the reviews on coral reef stress by Johannes (1975), Endean (1976), Pearson (1981), Peters (1984) and Rogers (1985). A special chapter first appeared in Brown and Howard (1985). One must be confident that since the begining of the century, and probably a little sooner, mass bleaching events similar to the present ones would have been at least sometimes described, as were reported local events (see table 2; and below).
 
Before historical reports, halt of coral calcification during bleaching and absence of specific mark in the growth bands have prevented so far a retrospective long-term analysis. A contrario, in Florida, the apparent lack of missing years in the growth bands of Solenastrea since 1880, and Montastrea annularis since 1861 (Hudson et al., 1976, 1989) tends to indicate that there was no widespread previous bleaching event. The maximum reduction of growth in those two corals was respectively 50% and 30% from the mean, with only stress bands due to cold winter. Sclerochronological data are more ambiguous for the east and central Pacific (Druffel, 1985, Druffel et al., 1989). Growth discontinuities were identified in a record since 1587, but none presented bioerosion nor encrustation observed after the 1983 mass bleaching (Dunbar et al., 1994).
 
Death of centuries old colonies, at least 200 years, and perhaps more than 500 years in East Pacific (Glynn, 1985, 1989, Robinson, 1985) and in Puerto Rico (Goenaga in Goenaga et al., 1989) reinforces the conviction that mass bleaching is a new phenomenon. Frequent abnormalities of shell of large foraminifers, never observed before, support this view.
 
Is there a normal "back ground" level of bleaching in reefs ? Natural aposymbioses are known in the temperate coral Astrangia danae (Peters and Pilson, 1985) and the sea anemone Anthopleura along the Pacific coast of North America (Muscatine, 1974), and for sponges in deep waters (Vicente, 1990). Most reefs would contain a few bleached colonies normally (Williams and Bunkley-Williams, 1989). Lastly, we found no old reports of the phenomenon. Goreau and Goreau (1959) judged an opportunity for experiments their finding of in situ bleached colonies of Manicina areolata under a large coral head in semi-darkness. Goreau (1964) has observed temporary and reversible bleaching on the fore-reef slopes at depths below 30 m, which was described more precisely by Goreau et al. (1970): "For reasons not yet understood, almost complete bleaching, i.e. loss of zooxanthellae, is often observed in deep water reef corals that appears to be otherwise normal. It is perhaps noteworthy that such corals regain their normal complement of zooxanthellae within a few days whereas in severely stressed corals the recovery of the zooxanthellae takes several months". This casts doubt whether true symbiont expulsion occurred, and at least indicates an important difference with recent bleaching phenomenon. Yonge in 1973 stated that "colourless colonies of hermatypes are from time to time encountred in deep shade, usually under some man-made erection". During a monitoring of Montastrea cavernosa at 3 sites on Panama's Caibbean coast over 2 years, Lasker (1979) found only one colony with a small bleached region, at time of heavy waves in December 1979. Muscatine et al. (1979) stated: "Most desirable are naturally-occurring aposymbiotic corals of the same species [for experiments]. Unfortunately these are difficult to obtain, and for most species may be virtually non-existent. During the course of investigations (...) we discovered naturally-occurring aposymbiotic colonies of Madracis mirabilis (...) in Discovery Bay, Jamaica", due to sediment covering. Upton and Peters (1986) examined 3 species of corals (A. agaricia, M. cavernosa, M. meandrites) from Puerto Rico which were found in 1980 with partial or patchy bleaching and necrosis, and often, but not always, infected. No frequency of the phenomenon is given but the largest white patch measured 4x5cm. In 1981, some corals in Puerto Rico were bleached but supposely because of a ciliate attack (Vicente and Goenaga in Williams et al., 1981, Williams et al., 1987, unpublished).
 
Fisk and Done (1985) stated that isolated bleached corals are commonly observed in the Great Barrier Reef, especially on nearshore reefs. Very low level of bleaching for long periods was also said to exist in Maldives and Fiji (Wood, 1988, Beckman in W90). "Chronic level of bleaching since early 80's" is said to occurs in Colombia (Zea and Duque Tobon, 1989). Bleaching occured at 1-2% level in Bonaire in the years 1977-1981 and 1985-1992, at 10-30 m depth, outside main events (in part due to cooling also) (Kobluk and Lysenko, 1994). Deep Agaricia in Jamaica partially bleached in October months before mass bleaching in 1987 (Porter in Woodley, 1988). Up to 4.5% bleached colonies in sites of Florida Keys in July-September 1985 were observed by Glynn et al. (1989a). Low level of bleaching have occurred in Caribean throughout 1987-1988 (W90). In winter 1986-1987 in Jamaica, between summer bleaching events, minimum level of bleaching of M. annularis and Agaricia sp. was between 5% and 25% in the fore reef (Gates, 1990). One third of the Favia fragum and Palythoa ?mammilosa was found in a bleached state in Bermuda in spring 1988, between main events (Cook et al., 1988). This must be hardly considered as a "normal" background level. One is forced to conclude that a chronic bleaching (Glynn, 1993) had appeared.
 

D) CONSEQUENCES

Mortality, recovery and long-term consequence
Coral mortality after mass bleaching is highly variable, affecting up to 97% of corals in Uva Island after 1983 (Glynn, 1989), but in a number of cases recovery seems total (W90, and Table 2). General discussion of recovery after various environmental perturbations are found in Johannes (1975), Endean (1976), Pearson (1981), and Brown and Howard (1985). Recovery time of reef coral commonities was evaluated to be between 4 to 100 years (Coles, 1984). Long-term consequences of mass bleaching were particularly discussed by Coffroth et al. (1989), Glynn (1989, 1993), Glynn et al. (1991), Smith and Buddemeier (1992). Given that causes and mechanisms of mass bleaching are not understood, few predictions can be made apart of gloomy subjectives guesses. At ecosystem level, stability is a complex question quite far beyond current knowledge. It can only be affirmed that mass bleaching has revealed coral reefs may be surprinsingly fragile. The 1980-85 period was catastrophic in the Great Barrier Reef but somewhat better in 1985-90 (Done, 1992). The implications of bleaching for the carbon cycle is treated in annex IV. A few more points can be quoted:
 

Adaptation and/or selection of host and symbionts

Adaptation of corals to chronic or repeated bleaching stress conditions is probably very slow as they are long-lived species (Glynn, 1989, 1993, Glynn et al., 1991). Selection of resistant strains of symbionts, if it occurs, would be important for fast adaptation. Unbleaching of orange-colored corals in Negril, Jamaica, is an argument in favor of the existence such strains (Goreau, 1991). Buddmeier and Fautin (1993) posit that bleaching allows a host to be repopulated with a different partner. In a laboratory experiments, Franzisket (1970) described a recovery of a bleached coral tip with zooxanthellae slowly spreading from a contact with another tip, and Aiptasia was reinfected with more resitant zooxanthellae from the sun-loving Cassiopea (Jokiel and York, 1982). In situ, the observation of recovery from unbleached area at a rate of 12 mm per month in Cayman M. annularis suggests an absence of selected residual symbiont population (Hayes and Bush, 1990). Bleaching of the same colonies every year seems to infirm change of algal partner.
 

Extinction

Two species of Millepora (Hydrozoa) were first believed extinct in the eastern Pacific after the 1983 event, M. boschmai (described by de Weerdt and Glynn, 1992) and M. platyphylla (known also in Marquesas Islands), as well as perhaps also Acropora valida and the non-zooxanthellate coral Tubastrea tagusensis, a Galapagos endemic (Glynn and De Weerdt, 1991). However, five little live colonies of M. boschmai of 4-7 years old were discovered in 1992 at Uva Island (Glynn and Feingold, 1992).
 

Recruitment

A first step of recovery after coral mortality is recruitment. Following the East Pacific 1983 El Niño event, the "low rates of larval recruitment suggest that local populations have capacity to recover" (Glynn et al., 1991, 1996a). Robinson (1985) noted high rate of recruitment of black corals. Species can recover through recruitment in places were they all died, as Acropora from 1983 to 1988 in Indonesia (Brown and Suharsono, 1990) or Millepora intricata in Uva Island (Glynn and De Weerdt, 1991). Bleached populations of the large foraminifer Amphistegina gibbosa observed in culture within months after field collection showed both reduction of clonal reproduction by large adults and attempt of reproduction by small ones, producing reduced number of young, frequently non-viable or deformed. Size histogram of Floridan population in 1993, as well as double normal size of A. radiata from Caroline Islands, 1992, indicated also partial suppression of asexual reproduction (Hallock and Talge, 1993, Hallock et al., 1995, Talge et al., in press).
 

Diversity

Recovery of diversity, and more fundamentally of the integrity of the ecological system for which no definition nor measurement can be proposed, is hardly predictable. Diversity decreased but not significantly, with the dominant species, tabulate Acropora, less affected, according to the study of Fisk and Done (1985), whereas Jaap (1985) observed long term reduction of diversity and change of dominance. Variability of the effects on diversity measurements are emphasized by Brown and Suharsono (1990). The results of the long-term biotic monitoring of Warwick et al. (1990) in Indonesia after the 1983 event are complex but promising. They found an approach to recovery till 1985, then a halt perhaps under an "unidentified stress". At one place (Pari), the maximum of diversity between quadrats was in 1984 because of a higher mortality and/or a previous variability in the dominant species. Shift in coral population structure, different than dominance or diversity, was also shown up. This shift was roughly similar at species and genera level at one site, but not at the other (op. cit., fig. 6).
 

Ecological interactions

Reefs are basically dominated by coral competition against algae. It is stricking that in Jamaica, coral cover reduced from 50-80% in the 1970' to less than 5% in 1990, and in the same time algal cover changed from 1-3% to 60-95%, not only because of mass bleaching but also because of overfishing, Diadema mortalities and cyclones (Hugues, 1994), and perhaps eutrophisation as well (Goreau, 1992). There is rapid algae overgrowth on bleached and dying corals (W90). After recovery, when healthy, corals dominates those algae with formation of calcified rims and walls till a complete enclosure of the algae. Rim formation by brown living part can be as rapid as one week after bleached tissue is gone, and recovered corals sometimes showed many nodosities, sometimes with enclosed H2S-smelling dark liquid (pers. obs.). In Florida, numerous growth protuberance were also observed on A. palmata, without identification of the origin (Porter and Meier, 1992)
 
At an upper ecological level, relative foraging pressure of the sea urchin Eucidaris increased after bleaching (Robinson, 1985), as well as predation from the sea star Acanthaster and the Arothron after the 70-95% coral mortality in East Pacific, 1983, but gasteropod Jeneria population declined (Glynn, 1985a). The urchin population increase dramatically, from 3-5 to 50-80 individuals/m2 (Glynn and Colgan, 1992). The most beautiful example of the complexity of the ecological disturbance is the story of the corals, their crustacean guards, and the sea star Acanthaster in Uva Island (Glynn, 1985b, Glynn et al., 1985a): the crustacean Trapezia and Alpheus, obligate partners of Pocillopora corals, feed on the lipid-rich mucus and in exchange, protect them efficiently against predation of Acanthaster. Once the corals bleached and died, the crustacean guards migrated or died. Acanthaster could go through the circle of pocilloporids surrounding the lagoon and devasted the area. Another example is the colonization of algal mats on partially bleached corals after El Niño 1983: algae were grazed by damselfishes which inflicted additional coral bite and mortality, thus providing further substrates for algal mats (Glynn, 1990). Damselfish interactions are complex as they also protect against Acanthaster (Glynn and Colgan, 1992).
 
Erosion
Reef framework destruction was evaluated in the very detailed analysis of bioerosion conducted by Glynn (1988b), Reaka-Kula et al. (1996) in Panama and Galapagos. The bioerosion increased to respectively 10-20 kgCaCO3/m2.year and 20-40 kgCaCO3/m2.year, with a net erosion of somewhat less than 10 kg/m2.year, or around 7mm/year. In Uva Island, the erosion is now of 22 mm/year, and even 44 mm/y at the walls (Eakin, 1992, 1996).
 
Geological comparaison
From a geological point of vue, reef ecosystems recovered slowly from extinction events (Cowen, 1988, and ref. therein). The present degradation of coral reefs, with its magnitude and speed, due both to local and global reasons, has no known equivalent in the past time, put aside the Cretaceous/Tertiary event. Present foraminifer shell abnormalities strongly support this comparaison (see above). Corals survived this probable bolide impact (Grigg, 1992), but the reconstitution of the ecosystem, with reappearance of hermatypic corals evolving from deep sea ones and of larger foraminifera from little benthics was a process which needed about 10 millions years (Pêcheux, 1995, Pêcheux and Michaud, 1997 ; Int. Geol. Correl. Prog. 286 "Early Paleogen Benthos", Jaca, Spain, 1991, verbal conclusion).

(See Part 2)


MARS © Copyright 1997 - All Rights Reserved