Coral reefs are the most diverse ofall marine ecosystems, and they
are rivaled in biodiversity by few ter-restrial ecosystems. They support peo-ple directly and indirectly by buildingislands and atolls. They protect shore-lines from coastal erosion, support fish-eries of economic and cultural value,provide diving-related tourism andserve as habitats for organisms thatproduce natural products of biomed-ical interest. They are also museums ofthe planet’s natural wealth and placesof incredible natural beauty.
Despite their recognized biological,economic and aesthetic value, coralreefs are being destroyed at an alarm-ing rate throughout the world. Somecountries have seen 50 percent of theircoral reefs destroyed by human activi-ties in the past 15 years. Some humaninfluences are acute—for example,mining reefs for limestone, dumping
mine tailings on them, fishing with ex-plosives and cyanide, and land recla-mation. Reefs that experience such in-sults often die; those that deterioratebut survive cannot recover to theiroriginal health as long as the distur-bances continue. In other countries thedisturbances are more chronic thanacute. Reefs are assaulted by muddyrunoff, nutrients and pesticides fromadjacent river catchments, overfishingand global-warming effects. These dis-turbances affect the key parameterspermitting reef resilience: water andsubstratum quality. As a result, coralsfail to reproduce successfully, and thecoral larvae arriving from more pris-tine reefs are unable to settle and thriveon substrata covered by mud, cyano-bacteria or fleshy algae. Coral popula-tions thus fail to recover or reestablishthemselves.
Can science help save coral reefs?Despite much talk about managingcoral reefs, the potential role of scienceis limited. But it is important: Scientistscan demonstrate the key processes con-trolling the health of coral reefs andhow human activities damage them.Then, we can hope, land-use managersand marine-resources managers will beable to modify human behavior to re-duce or reverse damage to coral reefs.Toward this end we have developed alarge-scale model for illuminating reefdegradation and predicting the impactof future human activity.
The Coral Reef EcosystemThe ecological functioning of a coralreef relies on the symbiotic associationbetween corals and dinoflagellate algae(zooxanthellae). In this system, the di-noflagellates reside as symbionts with-in the cells of the coral host; the sym-bionts take in nutrients and produce
metabolites from which the corals de-rive much of their energy. (Corals con-struct their hard habitat the way mol-lusks grow their shells, by accretingcalcium carbonate.) The main function-al components of a coral reef ecosys-tem include the hard corals, corallinealgae, filamentous and fleshy algae,blue-green algae (cyanobacteria), anda host of invertebrates and fishes.Coralline algae are essential to ahealthy reef because they cement reefstructures and contain chemicals thatinduce metamorphosis in coral planulalarvae. Filamentous and fleshy algaecan be very abundant; indeed, themost telltale sign of a degraded coralreef is the replacement of corals by al-gae (Figure 2).
Three genera of corals dominate Pa-cific reefs: Acropora, Porites and Pocillo-pora. Acropora corals are the most spec-tacular; they are also frameworkbuilders, providing habitat for a vari-ety of fishes and other reef organisms.They include the table, elkhorn,staghorn and fast-growing branchingspecies. Porites corals include boulderor massive corals. Pocillopora corals in-clude both coarsely and finely branch-ing species, widely distributed acrossthe Pacific and into the Red Sea.
Natural disturbances—includinghurricanes (tropical cyclones or ty-phoons), river floods (Figure 4), earth-quakes and lava flows—have affectedcoral reefs for millions of years; they aretypically acute and have short-lived ef-fects. Reef areas away from human in-fluences often recover within a fewyears if water and substratum qualityremain high. Acute, natural distur-bances thus help maintain diversity oncoral reefs by knocking back dominantspecies and allowing less competitivespecies to reestablish themselves.
44 American Scientist, Volume 91
Mud, Marine Snow and Coral Reefs
The survival of coral reefs requires integrated watershed-based management activities and marine conservation
Eric Wolanski, Robert Richmond, Laurence McCook and Hugh Sweatman
Eric Wolanski received his Ph.D. in environmentalengineering from the Johns Hopkins University in1972. He is a leading scientist at the AustralianInstitute of Marine Science, where he studies trop-ical coastal oceanography and its biological impli-cations for mangroves and coral reefs. Robert Rich-mond received his Ph.D. in biology from the StateUniversity of New York at Stony Brook in 1983.He is a professor of marine biology at the Universi-ty of Guam Marine Laboratory. His research inter-ests include sublethal stresses on coral reefs. Lau-rence McCook received his Ph.D. in biology fromDalhousie University in 1992; he is a research sci-entist specializing in the ecology of algae and reefdegradation at the Australian Institute of MarineScience, working with the Cooperative ResearchCentre for the Great Barrier Reef World HeritageArea. Hugh Sweatman received his Ph.D. fromMacquarie University in 1985; he is a research sci-entist at the Australian Institute of Marine Sci-ence where he leads the long-term reef-monitoringprogram. Address for Wolanski: AIMS, PMB No.3, Townsville MC, Qld. 4810, Australia. E-mail:[email protected]
The synergistic and cumulative effectsof human disturbances superimposedover natural disturbances make recov-ery less likely and, in some cases, resultin stable states dominated by algae.
Marine SnowThe water around coral reefs some-times looks clear, but it can contain avariety of suspended matter, startingwith inorganic particles such as resus-pended calcareous material, fecal ma-terial, organic detrital particles, andmucus secreted by plankton, algae andbacteria. Corals themselves secrete mu-cus to cleanse their colony surfaces andas a metabolic by-product. This mucusprevents corals from becoming cloggedwith sediment particles and alsoshields against desiccation if they areexposed to air at low tide. Particles insuspension in water are rapidly aggre-gated by flocculation as well as by ad-
hesive bridges of exopolymer (mucus).The aggregates resemble snowflakesand hence are called “marine snow”(Figure 5).
Human activities on land often re-sult in increased nutrient concentra-tions in coastal coral reef waters, whichenhances the growth of algae. In turn,increased nutrients also result in an in-creased prevalence of marine snow. Inthe nutrient-enriched coastal waters ofGuam and the Great Barrier Reef, ma-rine snow flocs can exceed several cen-timeters in diameter. Wave-exposedwaters have smaller flocs than do moresheltered areas.
Marine snow is almost neutrallybuoyant and can remain in suspensionfor hours in turbulent reef waters.Near-shore waters, however, may con-tain additional suspended fine clayparticles, rich in nutrients and detritusderived from land runoff. This mud in
suspension readily attaches to thesticky marine snow, forming muddymarine snow. The clay particles act asballast that makes the flocs settle ontocoral reefs. Muddy marine snow flocssettle fast, typically at a speed of about5 centimeters per minute—about 1,000times faster than individual mud parti-cles settle. The settled muddy marinesnow has detrimental or even lethal ef-fects on small coral reef organisms.
Corals and reef organisms such asbarnacles are able to clean themselvesof small settling flocs as long as the siltcontent remains low—less than 0.5 mil-ligrams per square centimeter. At highsiltation levels (4 to 5 milligrams persquare centimeter) or when flocs arelarge (particles 200 to 2,000 micrometersin diameter), the coral polyps initiallyexude thick layers of mucus and die af-ter less than one hour of exposure—ashort time compared with the rate of
Figure 1. Healthy coral reefs such as this section of Australia’s Great Barrier Reef not only are aesthetic treasures but also possess the greatestbiodiversity of any aquatic ecosystem. In particular, reefs in which various species of the fast-growing, branching Acropora corals dominate pro-vide habitat for a spectacular array of life both large and small. Despite the recognized value of reefs, worldwide they have been degraded dra-matically over the past several decades owing to anthropogenic effects, and little has been done to halt their decline. A large part of the prob-lem has been a lack of tools that would allow land- and marine-resource managers to estimate the relative benefits to reef health of variousregulatory actions. The authors have developed a model that permits managers to estimate the effects on reefs of regulating various human ac-tivities. (Except where noted, photographs by the authors.)
natural marine snow deposition oncoral reefs. Katharina Fabricius, at theAustralian Institute of Marine Science,found, from laboratory experiments,that settling muddy marine snow flocsare far more hazardous to young thanto older corals; the mortality rate is 10times greater for young coral recruits(newly established animals) than foradult corals. Coral recruits typically dieafter 43 hours of exposure to muddymarine snow, a threshold that is rou-tinely exceeded in nutrient-enrichedcoastal waters of the Great Barrier Reefbut not in areas farther offshore. Thepresence of muddy marine snow oncoral is often short-lived, resulting froma river flood or from resuspension bywind-generated waves in a storm. Thesettled muddy marine snow, havingdone its dirty work, is consumed byplankton and reef organisms, or flushedout by waves and currents. It usuallyleaves no “smoking gun,” just a de-graded reef.
Pollutants, including pesticides,heavy metals and hydrocarbons, alsodegrade coastal reefs. They can inter-fere with the chemically sensitiveprocesses of reproduction and recruit-ment in corals and other reef organ-isms, such as synchronization ofspawning, egg-sperm interactions, fer-tilization, embryological development,larval settlement, larval metamorpho-sis and acquisition of symbiotic zoo-xanthellae by young corals followingrecruitment.
46 American Scientist, Volume 91
Figure 2. Most coral-reef decline, including this example at Palau, is tied in subtle but pre-dictable ways to sediment-laden runoff from human activities within adjacent watersheds. Suchrunoff may not initiate the decline of a reef; more often, in fact, it prevents a reef from recoveringfrom an acute shock such as a tropical cyclone. The effects are multiple. First, coral larvae, whichmay have drifted from a distant healthy reef, are unable to colonize because the reef has becomecovered in a muddy algal mat. Second, those larvae that do establish a foothold may be smoth-ered by newly deposited mud. Third, and at least as significant, the prevalence of herbivorousfish, which feed on the fleshy algae that damage coral larvae, is inversely proportional to waterturbidity. In much the same way that regular mowing maintains a fast-growing suburban lawn,herbivorous fish control filamentous algae and enable coral growth. A reef with adult corals butno recruitment of juveniles is in reality dead, but just doesn’t know it yet.
Cape Grenville
Princess Charlotte Bay
Cooktown/Lizard Island
Cairns
Innisfail
Townsville
Cape Upstart
Whitsunday
Pompey
Swain
CapricornBunker
145°E 150°E
kilometers
0 100 200 300 400 500
20°S
15°S
10°S
Figure 3. Model area included 261 reefs in sev-eral zones (black type) of Australia’s Great Bar-rier Reef. Coral cover on 48 of these reefs hasbeen monitored annually for more than adecade and shows great variability in time andspace. This variability is characterized by a ma-jor decrease in coral cover following acute dis-turbances, such as a tropical cyclone or a riverflood, and slow recovery thereafter. For in-stance, tropical cyclone Ivor in 1990 affectedthe coral cover on all the outer shelf reefs inthe Cooktown/Lizard Island sector; 10 yearslater, the coral had recovered to cover morethan 50 percent of the reef surface. Inshorereefs in the Cairns sector recovered slowly afterlosing coral cover through the combined ef-fects of bleaching, an infestion of coral-eatingcrown-of-thorns starfish and tropical cycloneRona in 1999. Among midshelf reefs nearTownsville, Rib Reef is dominated by fast-growing tabulate Acropora corals. The reef cov-er was severely reduced by starfish infestationin the 1980s. Recovery was reversed by tropicalcyclone Justin in 1997 and by another outbreakof starfish. Coral cover declined sharply on twomidshelf reefs in the Whitsunday sector in1997 as a result of tropical cyclone Justin.
The TollIf, through increased muddiness or pol-lution, a single link in the coral repro-ductive chain is broken, the system cas-cades toward eventual demise. Onehundred percent successful fertilizationfollowed by 0 percent recruitment hasthe same outcome as 100 percent fertil-ization failure. Therefore many reefsnear human population centers simplydo not recover from disturbances. Lessattractive Porites and Pocillopora coralsreplace the habitat-building Acroporacorals, and, quite often, the benthos be-comes covered by fleshy algae, spongesand worms. This in turn causes a shiftin resident fish populations. Environ-
mental degradation can also result fromoverexploitation of populations of her-bivorous fishes, such as parrotfish andsurgeonfish, which feed on fleshy al-gae. In overfished systems, fleshy algaecan overgrow corals and prevent corallarvae from recruiting.
Human activity will continue to in-crease. As a result, coral reefs will in-creasingly degrade—as long as humanactivities on land and in coral reef wa-ters continue to be managed indepen-dently. In both developing and devel-oped countries, including Australia,the U.S., Japan and the French overseasterritories, different government agen-cies deal with land-based issues and
with marine and reef issues. In Aus-tralia, about four agencies deal withland-based issues and two with reefs(not counting fishing). It is as if landand sea were not interconnectedecosystems. This disconnect betweenwatershed-based activities and marineconservation has resulted in serious en-vironmental degradation throughoutthe world.
Science can help save coral reefs byproviding land- and marine-resourcemanagers with accurate and adequatedata on key threats and the synergismsinvolved, specific indicators of reef re-silience, as well as science-based mod-els to predict the impact of various de-
Figure 4. Muddy plume of the Burdekin Riv-er during a river flood, in the Great BarrierReef near Townsville, can be seen in thisoblique aerial photograph. The plume at thislocation spans 5 kilometers in width and ismade readily visible by its high turbidity.Both the entrained mud and reduced salinitycan cause acute damage to coral reefs. Themap (above) shows the salinity distributionin a 400-kilometer-long section of the GreatBarrier Reef for the 1991 river flood, predict-ed by Brian King using a three-dimensionalhydrodynamic model. (By contrast, duringtimes of normal river outflow, no reductionsin salinity would be evident.) The modeluses historical data on daily river discharges,as well as wind speed and direction data, tocalculate the movement of river flood plumesfrom 1969 onwards.
cisions about land use and reef fish-eries on reef health. We have devel-oped such a mathematical model forthe Great Barrier Reef, for which exten-sive physical and biological data areavailable.
Modeling Reef HealthAustralia’s Great Barrier Reef stretchesalong 2,600 kilometers of the east coastof Australia from 25�S to 10�S. The do-main of our model comprises 261 reefs
in a 400-kilometer-long swath that ex-tends from Lizard Island in the north tothe Whitsunday Islands in the south.We wished to model the region believedto be most susceptible to anthropogenicimpacts from land runoff (Figure 4).Data to develop the model came fromthe Long-term Monitoring Program atthe Australian Institute of Marine Sci-ence, which has surveyed 48 reefs an-nually since 1992 for assemblages of reeffishes and communities of benthic or-
ganisms in the upper northeast reefslope. Reefs for the program were cho-sen from three shelf positions (inshore,midshelf and outer shelf) at six lati-tudes, four of which fall within themodel domain. Coastal reefs are themost affected by runoff but are notmonitored because of poor visibilityand the presence of crocodiles. The dataon algal and coral cover (Figures 6 and 9)present evidence of large changes in thereefs over time in each region; thechanges vary among regions. The datashow a strong linear relationship be-tween the abundance of herbivorousfish and water visibility (Figure 7)—asmight be expected, since the fish keepalgae populations down—and this ob-servation is central to the formulationof a reef-health model.
The model uses algal cover as aproxy measure of reef health. The eco-logical components of the model in-clude hard corals in two age groups,juvenile and adult, together with algaeand herbivorous fish. Corals and fleshyalgae compete for space, and herbivo-rous fish consume algae. Reef distur-
48 American Scientist, Volume 91
Figure 5. Marine snow flocs such as the one at left—photographed by Katharina Fabricius in reefal waters of Palau, Micronesia—may reach 30centimeters in length but are typically smaller—0.5 to 5 millimeters in size. Flocs are formed by colonies of phytoplankton, fecal pellets, mucussecreted by bacteria and plankton, macroscopic aggregates of Thalassiosira nana, large diatoms, dinoflagellates and tintinnid ciliates, as well asa variety of other plankton and their remains. Additional mucus is produced by diatoms and microbes that colonize the nutrient-rich clay par-ticles derived from erosion in river catchments. Mud readily aggregates on marine snow because it is sticky, forming muddy marine snow flocs,shown spanning 0.8 millimeters in the photograph at right. The mud acts as ballast, forcing the marine snow to settle on the corals, which theysmother, particularly the juvenile recruits.
alga
e (m
odel
), p
erce
nt
algae (observed), percent
100
80
60
40
20
0100806040200
Michaelmas Chicken
Opal (2)
Hastings
Dip
Davies
St. Crispin
Myrmidon
AgincourtNo. 1
No Name
Carter
LizardIsland
Martin(14123)
Mackay
NorthDirection
IslandJohn
Brewer
Rib
FitzroyIsland Low
Isles
Green Island(16049)
HavannahIsland
Figure 6. Algal cover, which displaces corals,serves in the model as a parameter for thehealth of the Great Barrier Reef. Its distribu-tion has been monitored yearly since 1992 forthe 48 reefs. This scatter plot of observed andpredicted algal cover suggests that the modelhas promise. The outlier points (red) are reefsinfested with the coral-eating crown-of-thornsstarfish, Acanthaster planci. Such infestationsare not incorporated into the model becausethe ecology of the starfish is insufficiently un-derstood—in particular the ability of thestarfish to migrate within a reef and betweenreefs. Masahra Ogura at Tokai University,Japan, has measured crown-of-thorns starfishmigrating at speeds of up to 546 meters per day.
bances are of two types: acute andchronic. Acute, natural disturbances in-clude river plumes, tropical cyclonesand warm-water events resulting inbleaching. These events kill coral,thereby providing free space that israpidly colonized by filamentous andthen fleshy algae. Chronic disturbancescome in the form of increased nutrientconcentration and the influence of in-creased turbidity—at least 90 percentof the nutrients from land runoff arriveattached to mud particles. Corals canslowly recover from tissue remnants orthrough recruitment of larvae fromhealthier reefs. The recovery rate de-pends on nutrient concentration andthe number of herbivorous fish avail-able to consume algae.
It is not necessary to calculate her-bivorous fish dynamics because peopledo not target herbivorous fish in theGreat Barrier Reef; their abundance ispredicted by water visibility. It is thussufficient for the model to set the pre-vailing visibility conditions from fieldobservations. The mean visibility hasapparently halved since 1927—that is,turbidity has doubled—in the LowIsles area near Port Douglas.
Corals spawn each year at night inthe early summer, on a date set by themoon phase. In the model, the spawnmaterial is carried by water currentsfor about 10 days, a conservative esti-mated time of survival for coral larvaecapable of recruitment, by which timemost of the coral larvae have left theirnatal reef. The currents are affected bytides, forcing by the Coral Sea andwind, resulting in connectivity amongreefs (Figure 8). Recruitment rates werecalculated for every spawning yearsince 1969, when reliable wind data be-came available. Outside of spawningperiods, coral populations on individ-ual reefs are isolated from each other.
Natural disturbances were deter-mined from historical records of riverfloods and of the trajectory and inten-sity of tropical cyclones, for which dataalso are available since 1969. The coraldie-off from tropical cyclones was cal-culated from the trajectory and intensi-ty of the storms, using an empiricalfunction derived from surveys of coralcover immediately before and after thepassage of a storm. From models ofriver plumes, we extracted two key pa-rameters: the minimum salinity andthe duration of the river plume at eachreef. These parameters in turn wereused to calculate the die-off of coral.
The model predicts reef health as pa-rameterized by algal cover in a 400-kilometer-long stretch of the Great Bar-rier Reef (Figure 9). Without humaninfluences, coastal runoff degrades thereef in a zone whose width and degreeof impact vary with latitude, with max-imum damage in the Cairns region. Im-pacts within the zone vary considerablyowing to the passage of tropical cy-clones; as a result, reefs outside thecoastal zone are occasionally coveredwith algae. The model further predictsthat, with human activities on land, thezone of damage has already grownmuch larger than the natural state andwill increase in size and intensity in thefuture, unless human influences are
curtailed. At least as important, themodel enables one to quantify the ef-fects of various scenarios for control ofland-use activities—anywhere from“do nothing” to “strict control.” It of-fers decision makers and the public ascience-based tool to decide what activ-ities should be allowed, and how theyshould be controlled, on land and atsea, in order to produce a desired stateof health for coral reefs. The modelcould presumably also be used to testthe impact on reef health of various lev-els of fishing for herbivorous fish; thisdoes not apply to Australia’s Great Bar-rier Reef at present, but the question isrelevant to most reefs elsewhere in theworld, including Micronesia and the
mea
n he
rbiv
orou
s fis
h pe
r he
ctar
e
1,000
800
600
400
200
00 5 10 15 20 25 30 35 40
mean visibility (meters)
Figure 7. In Australia’s Great Barrier Reef, a relation exists between visibility and the abun-dance of herbivorous fish. These data were averaged for 10 years of surveys at 48 reefs scat-tered along a 400-kilometer stretch of the coast. In the model, visibility is used as a proxy forherbivorous fish prevalence.
4,500
4,000
3,500
3,000
2,500
2,000
1,500
1,000
500
0
num
ber
of p
artic
les
50 hours 100 hours
Figure 8. As visualized from the oceanographic model for Bowden Reef in the Great BarrierReef, mass spawning of corals creates a plume (left) of coral larvae over the reef. This plumeslowly mixes and is diluted by ambient oceanic waters while at the same time being carriedaway by the oceanic currents (right). Larvae from one reef can settle on other reefs. This processenables degraded reefs to recruit coral larvae from healthier reefs. The distribution of sourceand sink reefs varies yearly with the wind after annual coral spawning.
Figure 9. Model of coral health in this 400-kilometer-long stretch of the Great Barrier Reef was run with and without human influences. The pa-rameter for reef degradation is algal cover. These results suggest that a wide swath is degraded by human activities on land, particularly in near-shore and mid-shelf reefs. The immediate cause of coral death at some sites may be natural, acute disturbances such as hurricanes and riverfloods. Anthropogenic effects, via land runoff, on water quality appear responsible for the failure of reefs to recover after disturbance. This re-sults in a long-term decline in reef health.
Figure 10. Annual surveys of hard corals at Rib Reef on the Great Barrier Reef show the effects of disturbances such as tropical storms and out-breaks of the coral-eating, crown-of-thorns starfish, Acanthaster planci. Such disturbances have different effects on corals with different growthforms. The horizontal plates of fast-growing table Acropora spp. rapidly overgrow other corals but are easily broken during tropical cyclones andare a preferred food for crown-of-thorns starfish. Massive corals grow slowly, their round shape helps limit storm damage, and their perforateskeleton allows coral tissue to be sequestered during episodes of stress from reduced salinity, sedimentation and eutrophication associated withagricultural runoff and sewer outfalls. After such disturbances, the dead corals are colonized by filamentous or fleshy algae. Algae also replacecorals by direct overgrowth. In the photograph (right), the brown seaweed Lobophora variegata is growing up around the branches of Porites cylin-drica, smothering the coral tissue (shown by the dead, white skeleton where the researchers removed the algae covering the coral tissue).
U.S. Mariana Islands, where these fishare relentlessly pursued.
Limitations to ActionLike all ecosystem models, our modelhas limitations. These include un-known or poorly understood ecosys-tem processes, the lack of sufficientdata on predisturbance water qualityor habitat status, the lack of data fromundisturbed sites and inadequacies inthe measurement of water-quality para-meters. Additional ecological processescould be added to the model, but thisdoes not make it more useful becausethese new processes require additionalparameters for which data are unavail-able. Prediction of the response of areef to human influences is thus inher-ently uncertain.
Unfortunately, uncertainty in ecosys-tem models cannot readily be quanti-fied; too often, this is used as an excusefor inactivity, citing that “more re-search is needed before a sound deci-sion can be made.” Rather, the majorimpediment at this point appears to bepolitical will. Scientists could do muchmore to exert influence—for example,by translating existing scientific datainto the social and economic costs ofinaction and making this informationavailable to stakeholders and the broad-er community. Otherwise, “proof” ofimpacts will have to wait for seriousdegradation—and in comparison tocurrent data, rather than to the undoc-umented predisturbance state. Clearlythis outcome must be avoided. The ob-
jective should be prevention, notdemonstration, of extensive degrada-tion. Effective monitoring programsmust go beyond documentation ofcoral reef demise and be used as toolsto guide responses that prevent out-right mortality.
Examples from Guam and Hawaiishow that once a reef has been killed, itcannot be restored, even by importingoutside corals, unless the underlyingcause—for example, soil erosion in theadjoining catchment—is first ad-dressed. The most logical approach tocoral reef restoration is to alleviatethose conditions that caused the de-cline and allow natural recovery to oc-cur. More specifically, restore the wa-ter and substratum quality that allowscorals and other reef organisms to suc-cessfully reproduce and recruit. Thismeans controlling poor land-use prac-tices that spill mud, nutrients and pes-ticides into coral reef waters; manag-ing fisheries through quotas andfishing-gear restrictions; reducingtourism impacts; and establishing ma-rine protected areas. Science has a cru-cial role to play in demonstrating theconnections between land and reefecosystems and the profound effectsthose connections can have.
AcknowledgmentsThe authors thank the Australian Instituteof Marine Science, the University of GuamMarine Laboratory, the STAR program ofthe U.S. Environmental ProtectionAgency, National Oceanic and Atmos-
pheric Administration, IBM-Australia,Katie Marshall, Simon Spagnol, RichardBrinkman, Brian King, Katharina Fabri-cius, Angus Thompson, Greg Coleman andWilliam M. Hamner.
BibliographyAyukai, T., and E. Wolanski. 1997. Importance
of biologically mediated removal of finesediments from the Fly River plume, PapuaNew Guinea. Estuarine, Coastal and Shelf Sci-ence 44:629–639.
Birkeland, C. E. 1997. Life and Death of CoralReefs. New York: Chapman and Hall.
Done, T. J. 1992. Phase shifts in coral reef com-munities and their ecological significance.Hydrobiologia 247:121–132.
Fabricius, K., and E. Wolanski. 2000. Rapidsmothering of coral reef organisms by mud-dy marine snow. Estuarine, Coastal and ShelfScience 50:115–120.
McCook, L. J. 1999. Macroalgae, nutrients andphase shifts on coral reefs: Scientific issuesand management consequences for theGreat Barrier Reef. Coral Reef 18:357–367.
Richmond, R. H. 1993. Coral reefs: Presentproblems and future concerns resultingfrom anthropogenic disturbance. AmericanZoologist 33:524–53
Wolanski, E. 2001. Oceanographic Processes of CoralReefs. Physical and Biological Links in the GreatBarrier Reef. Boca Raton, Fla.: CRC Press.
Links to Internet resources for “Mud,Marine Snow and Coral Reefs” are
