CSIRO PUBLISHING

www.publish.csiro.au/journals/mfr Marine and Freshwater Research, 2008, 59, 703–716

Temporal dynamics in coral bioindicators for water qualityon coastal coral reefs of the Great Barrier Reef

Timothy F. CooperA,B,F,G, Peter V. RiddC, Karin E. UlstrupA,D,Craig HumphreyA, Matthew SlivkoffE and Katharina E. FabriciusA

AAustralian Institute of Marine Science, PMB No. 3, Townsville MC, Townsville Qld 4810, Australia.BSchool of Marine and Tropical Biology, and ARC Centre of Excellence for Coral Reef Studies,James Cook University, Townsville Qld 4811, Australia.

CSchool of Mathematics, Physics and IT, James Cook University, Townsville Qld 4811, Australia.DMarine Biological Laboratory, Department of Biology, University of Copenhagen,Strandpromenaden 5, DK-3000 Helsingør, Denmark.

ESchool of Imaging and Applied Physics, Curtin University of Technology,Kent St, Bentley WA 6102, Australia.

FCurrent address: Australian Institute of Marine Science, Botany Building (M096), Universityof Western Australia, 35 Stirling Highway, Crawley WA 6009, Australia.

GCorresponding author. Email: t.cooper@aims.gov.au

Abstract. There is a need to identify effective coral bioindicators that provide quantifiable links between changes inwater quality and the condition of coastal coral reefs. Temporal variation in a range of coral bioindicators includingsymbiont density, concentration of chlorophyll a, skeletal density and colony brightness of Pocillopora damicornis, aswell as colony brightness and density of macro-bioeroders of massive Porites spp. was examined for 2 years on a coastalcoral reef of the Great Barrier Reef. The specificity to changes in water quality varied among bioindicators. For example,a 2.5-fold variation in symbiont density of P. damicornis was related strongly to mean 14-day sea surface temperature andseasonal changes in water quality, suggesting medium specificity to changes in water quality. In contrast, the density ofmacro-bioeroders in Porites did not vary seasonally but there were consistently more macro-bioeroders at the coastal thanmid-shelf reference locations, suggesting high specificity of spatial differences in water quality. In situ measurementsof benthic irradiance and turbidity allowed the quantification of potential stress thresholds for coastal corals. Our datasuggest long-term turbidity >3 NTU leads to sublethal stress, whereas long-term turbidity >5 NTU corresponds to severestress effects on corals at shallow depths.

Additional keywords: benthic irradiance, bioerosion, sea surface temperature, specificity, symbiont density, turbidity.

Introduction

Responses of reef-building corals and their endosymbioticdinoflagellates (Symbiodinium) to changes in environmentalconditions can be used as bioindicators in monitoring programsof coral reefs.These responses range from the genetic to the com-munity level of organisation and include variation in parameterssuch as RNA/DNA ratio, symbiont density, photo-physiologyand coral growth rates, through to rates of bioerosion in mas-sive corals (e.g. Hoegh-Guldberg and Smith 1989; Lough andBarnes 2000; Risk et al. 2001; Hutchings and Peyrot-Clausade2002; Meesters et al. 2002). For example, the photo-physiologyof Symbiodinium has received attention as a proxy of coral health(e.g. Jones et al. 1998; Jones and Hoegh-Guldberg 2001; Lesserand Gorbunov 2001) because the physiological performance ofthe symbionts is governed by a suite of environmental factors,with light considered one of the key environmental resourcesbecause it drives photosynthesis. Symbiotic photosynthesis incorals results in the translocation of photosynthates to the coral

host, which is essential to maintain high rates of calcification,and hence growth, of coral reefs (Barnes and Chalker 1990;Muscatine 1990). Other factors that influence the physiologicalperformance of corals are sea surface temperature (SST) andwater quality such as exposure to nutrients, suspended particu-late matter and pesticides (see reviews by Hoegh-Guldberg 1999;Fabricius 2005). Environmental conditions that maintain light,SST and water quality within tolerance thresholds are, therefore,vital to the physiological and ecological success of coastal coralreefs.

Environmental conditions on coastal coral reefs are influ-enced by a range of natural and anthropogenic factors that vary intime and space. On short time scales (hours to weeks), corals oncoastal reefs are exposed to variable regimes of hydrodynamics,meteorological patterns and turbidity that influence the amountof light reaching the benthic assemblage (Anthony et al. 2004)and alter the energy balance of corals (Anthony and Fabricius2000). Weather-dependent resuspension of sediments that leads

© CSIRO 2008 10.1071/MF08016 1323-1650/08/080703

704 Marine and Freshwater Research T. F. Cooper et al.

to so-called turbidity events is a natural feature of coastal coralreefs (Orpin et al. 2004). These events may result in suddenchanges in irradiance as well as the remobilisation of particulatenutrients into the water column (Alongi and McKinnon 2005).Comparisons of benthic irradiance (Ez) with the minimum sat-urating irradiance (Ek) showed that corals may switch betweenperiods of light limitation (Ez < Ek) and light stress (Ez > Ek)on time scales of days to weeks in response to changing turbidity(Anthony et al. 2004). The extent of these turbidity events on theGreat Barrier Reef (GBR) is currently poorly understood.

Superimposed on these short-term events are seasonalchanges in daily insolation (Kirk 1994), SST and terrestrialrunoff (Wolanski et al. 2008) that alter environmental conditionson coral reefs. Studies from the Indo-Pacific and Caribbean haveshown that the density and pigment content of Symbiodiniumare regulated by seasonal variations in SST (Brown et al. 1999;Fagoonee et al. 1999; Fitt et al. 2000), irradiance (Stimson 1997;Brown et al. 1999) and water quality (Stimson 1997; Fagooneeet al. 1999). Seasonal differences in photo-physiology have alsobeen reported for corals in the Caribbean and Red Sea (Warneret al. 2002; Winters et al. 2006). Studies from the GBR haveshown that the tissue layer of colonies of Porites is significantlythicker in summer than in winter (Barnes and Lough 1992) andthat lipid content varies seasonally in relation to spawning cyclesin seven coral species (Leuzinger et al. 2003). Pillay et al. (2005)reported that symbiont density increases 2-fold with the onset ofwinter in Acropora millepora from the central GBR and Hill andRalph (2005) showed that daily diel changes in irradiance hasa greater influence on photo-physiology than seasonal variationfor three coral species from the southern GBR. Recently, Ulstrupet al. (2008) demonstrated seasonal regulation of coral photo-physiology, with non-photochemical quenching being higher insummer than in winter. To date, however, no studies have exam-ined seasonal variation in coral bioindicators within the contextof temporal variation in water quality on the GBR.

To use coral bioindicators as response measures of the effectsof changes in water quality requires an understanding of bothspatial and temporal patterns of variation of such measures. Theaim of the present study was to obtain estimates of the seasonaland temporal variability of a range of bioindicators in coastalcorals on the GBR including symbiont density, concentrationsof chlorophyll a, skeletal density and colony brightness of Pocil-lopora damicornis, as well as colony brightness and densityof macro-bioeroders in colonies of massive Porites. Specifi-cally, the objectives of this study were to: (1) quantify temporalchanges in environmental parameters such as turbidity, benthicirradiance, SST and water column nutrients, and develop poten-tial stress thresholds for turbidity; (2) test the hypothesis thatvariations in environmental conditions can be measured as aresponse in coral bioindicators at a coastal coral reef of theGBR; and (3) investigate the significance of the relationshipsbetween bioindicators and environmental parameters to assesstheir specificity to changes in water quality.

Materials and methodsStudy area and sampling designTo quantify temporal variation in coral bioindicators and envi-ronmental parameters, a study location was selected in the

coastal zone of the GBR, i.e. within the 20-m isobath (Alongi andMcKinnon 2005), whereas two reference locations were selectedat nearby mid-shelf reefs. Sampling at the coastal location wasundertaken at Horseshoe Bay (19◦06.5′S, 146◦50.2′E) on thenorthern side of Magnetic Island, a continental island locatedin the northward-facing Cleveland Bay ∼7 km from the city ofTownsville (Queensland, Australia; Fig. 1). The main influenceson water quality in Cleveland Bay are wind-driven resuspen-sion of sediments (Larcombe et al. 1995; Orpin et al. 2004),although several rivers may also influence water quality in thebay. These are the Ross River that flows through Townsville,and the Haughton and Burdekin Rivers located ∼40 km and90 km, respectively, to the south. The rivers differ markedly incatchment size (∼1700 km2, 4000 km2 and 130 000 km2) andestimated annual runoff (0.49 km3, 0.74 km3 and 10.49 km3) forthe Ross, Haughton and Burdekin Rivers respectively (Furnas2003). The Burdekin River is likely to have a greater influ-ence on water quality during episodic runoff events than thesmaller Ross and Haughton Rivers. In addition to these rivers,there are several small creeks, e.g. Endeavour Creek and GorgeCreek, with catchments dominated by natural vegetation andlow-scale urbanisation that flow into Horseshoe Bay.These smallcreeks are likely to have localised effects on water quality inHorseshoe Bay. The reference locations were Davies (18◦49.6′S,147◦37.8′E) and Broadhurst Reefs (18◦52.3′S, 147◦42.4′E),which are both located ∼60 km from the Australian mainlandand thus distant from most terrestrial influences (Fig. 1). Thegeneral sampling design used for the study comprised each ofthree sampling events within each of two seasons: Dry (June toNovember inclusive) and Wet (December to May inclusive) overa 2-year period from 2005 to 2007 at each of the three locations.Coral and water samples were collected in: June, September andOctober 2005 (Year 1, Dry Season); February, March and April2006 (Year 1, Wet Season); July, September and October 2006(Year 2, Dry Season); and March, April and May 2007 (Year 2,Wet Season).

Weather dataTotal monthly rainfall for the study period from May 2005 to May2007 for the Townsville region was obtained from the Bureauof Meteorology (www.bom.gov.au/weather/qld/townsville/observations.shtml). SST, surface irradiance (µmol photonsm−2 s−1) and wind speed (m s−1) recorded at half-hour inter-vals for Cleveland Bay and Davies Reef were obtained fromautomatic weather stations (AWS) operated by the AustralianInstitute of Marine Science (www.aims.gov.au/pages/facilities/weather-stations/weather-data.html).The Cleveland Bay weatherstation (19◦09′S, 147◦53′E) is located ∼6 km from the Horse-shoe Bay study site off the south-eastern side of Magnetic Islandand the weather station at Davies Reef (18◦50′S, 147◦41′E) is∼500 m from the study site. Weather conditions were assumedto be comparable at Broadhurst Reef, situated ∼10 km south ofDavies Reef, hence weather data from the Davies Reef weatherstation were used as explanatory variables for Broadhurst Reef.From the weather data, mean SST, mean total daily surface irra-diance (mol photons m−2 day−1) and mean wind speed (m s−1)were calculated for the 14 days preceding each sampling eventas per Brown et al. (1999). Mean wind speed from the ClevelandBayAWS was used to provide an indication of the effects of wave

Coral bioindicator temporal variability Marine and Freshwater Research 705

Longitude

Magnetic Is.

Broadhurst Rf.

Davies Rf.

Townsville

Burdekin R

50 km

N

146.0°E 146.5°E 147.0°E 147.5°E 148.0°E

20.0°S

19.5°S

19.0°S

18.5°S

18.0°SLa

titud

e

Ross R

Queensland

Great Barrier Reef

Haughton R

Horseshoe Bay

Cleveland Bay

Fig. 1. Map of study locations at Horseshoe Bay on Magnetic Island (coastal) and Davies and Broadhurst Reefs (mid-shelf).× AIMS weather station in Cleveland Bay.

height, and hence resuspension of sediments within ClevelandBay (as described byAnthony et al. 2004), which was consideredto be representative of conditions in nearby Horseshoe Bay.

Turbidity and benthic irradianceTurbidity and benthic irradiance loggers (JCU MK8, James CookUniversity, Townsville) were deployed at Horseshoe Bay (2.0 mbelow lowest astronomical tide, LAT) and Davies Reef (2.8 mbelow LAT) during the study. Turbidity was measured as neph-elometric turbidity units (NTU) with optical backscatter sensors(Ridd and Larcombe 1994). Benthic irradiance was measuredwith a cosine-corrected light sensor measuring photosyntheti-cally active radiation (PAR) between 400 and 700 nm. Both sen-sors were housed in the same unit.The light sensor was calibratedin water against a LI-192 light sensor (LI-COR, Lincoln, NE).The loggers measured turbidity and benthic irradiance every

10 min (comprising 20 measurements integrated over 10 s). Eachlogger had a cleaning mechanism to maintain the surface of thesensors free of bio-fouling (Ridd and Larcombe 1994).

The relationship between turbidity and attenuation coeffi-cient for downward irradiance Kd (PAR) for the coastal locationwas examined using data recorded at noon by the turbidity andirradiance logger at Horseshoe Bay and the automated weatherstation located in Cleveland Bay. The noon Kd values were deter-mined following rearrangement of the equation describing theBeer-Lambert Law:

Kd = ln(Ez/E0)/z (1)

where Ez is noon benthic irradiance recorded by the logger, E0is irradiance beneath the surface determined from the surfaceirradiance at the Cleveland Bay weather station and correctedfor 5% surface reflectance (Kirk 1994) and z is water depth

706 Marine and Freshwater Research T. F. Cooper et al.

above the logger in metres determined as the sum of the loggerdepth below LAT and the corresponding noon tide height. Asdata were being compared from a time series spanning a 2-yearperiod, it was necessary to normalise each Kd value for changesin solar zenith angle following Mobley (1994) using:

SZAuw = arcsin(sin(SZA)/1.34) (2)

where SZA is the solar zenith angle above the water, 1.34 isthe refractive index of seawater and SZAuw is the solar zenithangle underwater. The noon Kd values for each day were thennormalised by:

Kdn = Kdm cos(SZAuw) (3)

where Kdn is the normalised value and Kdm is the measuredvalue.

Water column nutrientsSurface water was sampled at each of two sites (separated by∼200 m) within each location during each sampling event tomeasure chlorophyll a, phaeophytin, particulate nitrogen (PN),particulate phosphorus (PP), particulate organic carbon (POC),dissolved inorganic nitrogen (DIN), dissolved inorganic phos-phorus (DIP), dissolved organic nutrients (DON and DOP) andtotal suspended solids (TSS). Analysis of the water samples fol-lowed standard analytical procedures described in Cooper et al.(2007). A water quality index (WQI) was calculated for eachseason using sites and sampling events as replicates and derivedfrom the sum of z-score transformations for the water columnnutrient variables measured at each of the sampling locations(further details in Fabricius and De’ath 2004; Fabricius et al.2005).

Coral bioindicatorsApical branches (∼6 cm long) of Pocillopora damicornis werecollected from the centre of each of six colonies at 1–3 mbelow LAT during each sampling event within each of thethree locations and frozen at −20◦C pending further analysis.Physiological parameters analysed included determinations ofsymbiont density, chlorophyll a and skeletal density. Each coralsample was placed into a plastic bag with filtered (0.2 µm) sea-water and the tissue was stripped from the skeleton using anair gun. The resulting tissue slurry was then homogenised for30 s using a tissue grinder and divided into sub-samples ofknown volume for determinations of density of symbionts andchlorophyll a. The density of symbiont cells was determinedwith eight replicate counts in a Neubauer improved haemo-cytometer (Brand, Wertheim, Germany). The determination ofchlorophyll a comprised a double extraction using acetone(100%) for 24 h in darkness at 4◦C. The optical propertiesof the combined extracts were measured using a ShimadzuUV-1700 Spectrophotometer (Shimadzu Corporation, Kyoto,Japan). Levels of chlorophyll a were determined using the for-mula from Jeffrey and Humphrey (1975) after correction for thevolume of the homogenate and solvent.The density of symbiontsand concentrations of chlorophyll a were normalised to surfacearea of the coral branch determined by wax dipping (Stimsonand Kinzie 1991). Skeletal density (ρs) of P. damicornis was

determined as the ratio between dry weight and buoyant weightfollowing the procedures described by Anthony et al. (2002).

Colony brightness and density of macro-bioeroders wererecorded for colonies of massive Porites occurring in belt tran-sects (100 m long × 3 m wide) during each sampling event ateach of the three locations described above. Colony bright-ness was estimated visually using a colour chart (Siebecket al. 2006). Macro-bioeroders on each colony were counted onsmall colonies, but sub-sampled with three replicate 25 × 25 cmquadrats for larger (i.e. >1 m diameter) colonies.

Statistical analysesPrincipal components analysis (PCA) was used to examinethe relationships of the physiological variables of P. damicor-nis among the study locations and times of sampling with theenvironmental variables. Data were log-transformed before thePCA. Differences in the concentrations of water quality param-eters and the bioindicators between years (two levels: random)and seasons (two levels: fixed, orthogonal), and among loca-tions (three levels: random, orthogonal) at each of three times(three levels: random, nested) were analysed with a four-factoranalysis of variance (ANOVA). Data were tested for devia-tions from the assumption of homogeneity of variances andtransformed if necessary. Pooling procedures involving elim-ination of terms from the mean square estimates were doneif a term was non-significant at P > 0.25 (Underwood 1997).Means for significant factors in the ANOVA were comparedusing Student–Newman–Keuls (SNK) tests.

Predictive models were used to determine which environ-mental variables best explained the patterns of variation of theresponse variables of P. damicornis and Porites. These analy-ses involved fitting a full linear model using all of the predictors(environmental variables) with terms dropped one at a time fromthe overall model. The best models were then selected based ontheAkaike Information Criterion (AIC) (Akaike 1974; Burnhamand Anderson 2002), which is a measure of the goodness of fit ofthe estimated model to the data. Once the models were selected,comparisons of indicators in P. damicornis and Porites weredone using ANOVA. The statistical analyses were done usingthe statistical software R (R Development Core Team 2006).

ResultsWeather dataThere were several weather events that influenced water qualityat the study locations during the 2-year study period. SevereTropical Cyclone (TC) Larry was classed a Category 4 storm(maximum wind gusts to 240 km h−1) when it crossed the NorthQueensland coast 250 km north of the study area on 20 March2006. The main influence of the cyclone on the Horseshoe Baylocation was strong winds generating swells that resuspendedbottom sediments.

The monthly rainfall record for the Townsville region wasconsistent with a monsoonal influence with greatest rainfalloccurring between January and April compared with June toNovember for both years. In early February 2007, intense rain-fall resulted in major flooding to North Queensland catchmentsbetween Cairns and Mackay. In the Townsville region, totalrainfall for the period between 30 January and 8 February 2007

Coral bioindicator temporal variability Marine and Freshwater Research 707

Townsville

Cairns

Mackay

50 km

N

Burdekin R

Ross R

Haughton R

Grafton Passage

Fig. 2. MODIS-aqua satellite image of a flood event on the Great Barrier Reef taken 10 February 2007. Red line shows study area in Fig. 1, red dots indicatestudy locations. Image downloaded from Ocean Colour Web (Feldman and McClain 2007) and processed by SeaDAS (Baith et al. 2001).

708 Marine and Freshwater Research T. F. Cooper et al.

May Sep Jan May Sep Jan May0

10

20

30

40

50

60

20

22

24

26

28

30

32

SS

T (

°C)

Irradiance (mol photons m

�2 day

�1)

Irradiance (mol photons m

�2 day

�1)

SS

T (

°C)

(a)

(b)

May Sep Jan May Sep Jan May20

22

24

26

28

30

32

0

10

20

30

40

50

60

Year 1 Year 2

2005 2006 2007

Fig. 3. Sea surface temperature (◦C, solid line) and monthly mean (28day ± s.e.) total daily irradiance (mol photons m−2 day−1) at the surface(◦) recorded by an automated weather station, and benthic irradiance (•),recorded at (a) Horseshoe Bay and (b) Davies Reef. Gaps in the time-seriesare missing data owing to instrument malfunction.

was ∼700 mm and major flooding was reported for severalrivers in the region including the Ross, Haughton and BurdekinRivers (Bureau of Meteorology 2007). A satellite image takenon 10 February 2007 clearly shows flood plumes emerging fromrivers along the North Queensland coast (Fig. 2). On reachingthe coastal zone, the flood plumes formed a continuous, north-ward moving plume with an exit point from the GBR lagoonat the Grafton Passage, which is consistent with a recent studythat suggested siliciclastic material is transported along the shelfleaving the GBR through inter-reef passages in the Cairns regionand ultimately deposited in the QueenslandTrough (Francis et al.2007). Records maintained by the Bureau of Meteorology indi-cate that this was a ‘1 in 5 year’ flood event for the BurdekinRiver (P. Baddiley, Hydrologist, Bureau of Meteorology, pers.comm., 2007).

Data of SST and irradiance during the study period are shownin Fig. 3. Sea surface temperatures at Magnetic Island reacheda maximum of 31.43◦C (December 2005) and 30.45◦C (March2007) and a minimum of 20.70◦C (July 2005) and 20.33◦C (July2006) in Year 1 and 2 respectively. The mean summer SST atMagnetic Island inYear 1 was 29.82 ± 0.01◦C, whereas the sum-mer of Year 2 was cooler with mean SST of 28.10 ± 0.01◦C.At Davies Reef, the modulation of SST was not as great as

at Magnetic Island, reaching a maximum of 29.75◦C (January2006) and 29.12◦C (March 2007) and a minimum of 22.72◦C(July 2005) and 22.54◦C (August 2006) in Year 1 and 2 respec-tively. The mean summer SST at Davies Reef in Year 1 was28.64 ± 0.01◦C, whereas the summer of Year 2 was cooler, witha mean SST of 27.55 ± 0.01◦C. In general, SST oscillated withina range of ∼10◦C between seasons at Magnetic Island and ∼7◦Cat Davies Reef during both years (Fig. 3).

Surface irradiance showed seasonally determined variationat both locations. At Magnetic Island, mean total daily surfaceirradiance (28 day mean ± s.e.) reached a maximum during thesummer months at 55.6 ± 1.2 mol photons m−2 day−1 (February2006) and 55.6 ± 2.0 mol photons m−2 day−1 (December 2006)with a minimum of 28.2 ± 1.5 mol photons m−2 day−1 (June2005) and 27.7 ± 1.4 mol photons m−2 day−1 (June 2006) inYear1 and 2 respectively. Similarly, at Davies Reef, mean total dailysurface irradiance reached a maximum of 50.0 ± 1.0 mol pho-tons m−2 day−1 (October 2005) and 51.1 ± 2.5 mol photons m−2

day−1 (November 2006) with a minimum of 30.9 ± 1.5 mol pho-tons m−2 day−1 (June 2005) and 24.6 ± 1.4 mol photons m−2

day−1 (June 2006) in Year 1 and 2 respectively (Fig. 3).The mean wind speed for the study period at Magnetic Island

was 5.53 ± 0.01 m s−1 and only exceeded 15 m s−1 when TCLarry passed to the north of the study area in March 2006 andduring the flood event in February 2007 (Fig. 4). At Davies Reef,the mean wind speed for the study period was 7.40 ± 0.02 m s−1,with winds greater than 20 m s−1 recorded in July 2006 andFebruary 2007.

Turbidity and benthic irradianceThe maximum hourly turbidity recorded during the study was83 NTU when TC Larry passed to the north of the study site atHorseshoe Bay.Turbidity remained above 5 NTU and mean windspeed was 7.03 ± 0.65 m s−1 at this location for a period of ∼10days following the cyclone in March 2006 (Fig. 4). There wereanother three weather-dependent turbidity events at HorseshoeBay where levels remained elevated around 5 NTU for periodsof ≥5 days in March, April and May of Year 1. Turbidity atDavies Reef changed very little during these events in Year 1and rarely exceeded 1–2 NTU. In Year 2, there were at least sixweather-dependent turbidity events where levels remained ele-vated around 5 NTU for periods of ≥5 days in June, September,October and November 2006 and January and March/April 2007(Fig. 4). Interestingly, for a 4-week period in March/April 2007following the flood event in February 2007, turbidity at Horse-shoe Bay averaged 9.2 ± 0.2 NTU, reaching a maximum hourlyturbidity of 67 NTU on 17 April 2007 (Fig. 4). This turbidityevent coincided with a mean wind speed over the four weeks of6.86 ± 0.37 m s−1.

Changes in benthic irradiance at Horseshoe Bay relatedinversely to turbidity (Fig. 4). The mean total daily benthic irra-diance recorded at this location for a 10-day period after TCLarry was 1.3 ± 0.5 mol photons m−2 day−1. There were sim-ilar decreases in irradiance during the turbidity events in bothyears (Fig. 4). For example, during the 4-week turbidity eventin March/April 2007 (Year 2), the mean total daily benthic irra-diance recorded at Horseshoe Bay was 5.8 ± 0.9 mol photonsm−2 day−1 (Fig. 4).

Coral bioindicator temporal variability Marine and Freshwater Research 709

0200400600800

10001200

0

20

40

60

80

100

May 2005

Sep Jan May Sep Jan May0

100200300400500600

0

5

10

15

20

Jul Nov Mar Jul Nov Mar

Year 1 Year 2

TC LarryFlood

(a)

Tur

bidi

ty (

NT

U)

B

enth

ic ir

radi

ance

(µ

mol

pho

tons

m�

2 s�

1 )W

ind

(m s

�1 )

Rai

nfal

l (m

m)

(b)

(c)

(d )

Jun Aug Oct Dec Feb Apr Jun Aug Oct Dec Feb Apr

Dry Season Wet Season Dry Season Wet Season

2006 2007

Fig. 4. (a) Turbidity (NTU), (b) benthic irradiance (µmol photons m−2 s−1), (c) wind speed (m s−1) and (d) total monthly rainfall during study. (a) and (b)recorded by a logger at a shallow depth (2 m below LAT) on the fringing reef at Horseshoe Bay, (c) recorded by an automated weather station in ClevelandBay. Gaps in the time-series during Dry Season of Year 1 for (a) and (b), and Wet Season of Year 1 for (c), are missing data owing to instrument malfunction.Dashed line denotes time series shown in Fig. 5.

During the flood event in February 2007, the mean tur-bidity and daily benthic irradiance (10-day mean ± s.e.) were13.0 ± 0.7 NTU and 0.7 ± 0.5 mol photons m−2 day−1 respec-tively (Fig. 5). Strong winds coincided with the commencementof flooding and the mean wind speed during the flood event was8.69 ± 0.61 m s−1. Thus, the increase in turbidity observed atthe start of the flood event was due in part to resuspension ofsediment adjacent to the reef, in addition to the influence of theflood plume (Fig. 2).

The simultaneous measurement of light and turbidity by alogger provided insight to the effect of changes in turbidity onbenthic irradiance and the attenuation coefficient for downwardirradiance (Kd) on the fringing reef (average depth of 3.5 m) atHorseshoe Bay. Exploratory analysis showed an inverse rela-tionship between noon benthic irradiance and turbidity withthe extinction of benthic irradiance at this depth when turbid-ity exceeded ∼15 NTU at Horseshoe Bay (Fig. 6a). However, asdata were being compared for noon irradiance from a time seriesspanning 2 years, it was necessary to account for differencesin the water depth of the logger, i.e. tidal regime, and also forchanges in solar zenith angle. Comparisons of the normalised Kdand turbidity showed a linear relationship that explained ∼68%of the variance between these parameters (Fig. 6b). From thisrelationship, an increase in turbidity to 3 NTU, a level commonly

reached during this study, resulted in an 88% reduction in irra-diance intensity to levels around 200 µmol photons m−2 s−1 onthe fringing reef at Horseshoe Bay.

Water column nutrientsThere were temporal differences in concentrations of waterquality parameters such as chlorophyll a, particulate nitrogenand phosphorus, particulate organic carbon, dissolved inor-ganic nitrogen and dissolved organic phosphorus; however, thisvariability occurred at small temporal scales with differencesdetected among times of sampling within each of the seasons andyears. For example, there was great variation among samplingtimes for concentrations of chlorophyll a and particulate nutri-ents especially in October 2006, which followed a turbidity eventin Horseshoe Bay whereby turbidity averaged 8.29 ± 0.11 NTUfor a period of 16 days (Fig. 4). Dissolved organic nitrogen variedinconsistently between years and seasons, and among locations.At Broadhurst Reef, concentrations of dissolved organic nitro-gen were greater in the Dry compared with the Wet Season ofYear 1 (F2,16 = 1.12, P = 0.0227), but there were no differencesbetween seasons or years at the other locations. There were nodifferences between years and seasons or among locations forconcentrations of dissolved inorganic phosphorus. In additionto these temporal differences, the patterns of variation for some

710 Marine and Freshwater Research T. F. Cooper et al.27

Jan

07

28 J

an 0

729

Jan

07

30 J

an 0

731

Jan

07

01 F

eb 0

702

Feb

07

03 F

eb 0

704

Feb

07

05 F

eb 0

7

07 F

eb 0

708

Feb

07

09 F

eb 0

710

Feb

07

06 F

eb 0

7

11 F

eb 0

7

12 F

eb 0

713

Feb

07

14 F

eb 0

715

Feb

07

16 F

eb 0

7

Win

d (m

s�

1 )B

enth

ic ir

radi

ance

(µm

ol p

hoto

ns m

�2

s�1 )

0

200

400

600

800

1000

1200

0

5

10

15

20

0

20

40

60

80

100

27 J

an 0

728

Jan

07

29 J

an 0

7

30 J

an 0

731

Jan

07

01 F

eb 0

702

Feb

07

03 F

eb 0

704

Feb

07

05 F

eb 0

7

07 F

eb 0

708

Feb

07

09 F

eb 0

710

Feb

07

06 F

eb 0

7

11 F

eb 0

7

12 F

eb 0

713

Feb

07

14 F

eb 0

715

Feb

07

16 F

eb 0

7

27 J

an 0

728

Jan

07

29 J

an 0

7

30 J

an 0

731

Jan

07

01 F

eb 0

702

Feb

07

03 F

eb 0

704

Feb

07

05 F

eb 0

7

07 F

eb 0

708

Feb

07

09 F

eb 0

710

Feb

07

06 F

eb 0

7

11 F

eb 0

7

12 F

eb 0

713

Feb

07

14 F

eb 0

715

Feb

07

16 F

eb 0

7

Tur

bidi

ty (

NT

U)

(a)

(b)

(c)

Day

Fig. 5. (a) Turbidity (NTU), (b) benthic irradiance (µmol photonsm−2 s−1) and (c) wind speed (m s−1) recorded during the flood event inFebruary 2007. Arrow indicates the commencement of major flooding inthe Haughton River. A peak level of 3.8 m was recorded above the BurdekinFalls Dam spillway on 4 February 2007 (Bureau of Meteorology 2007).

water quality parameters were also characterised by differencesamong locations. For example, concentrations of chlorophyll a(F2,16 = 7.32, P = 0.0055) were ∼2-fold greater at HorseshoeBay (0.66 ± 0.14 µg L−1) than Davies (0.25 ± 0.03 µg L−1)or Broadhurst Reefs (0.29 ± 0.03 µg L−1). A similar pat-tern occurred for particulate organic carbon (F2,36 = 17.34,

Turbidity (NTU)

Kd

(PA

R)

0 5 10 15 20 25 300

1

2

3

4

5

Turbidity (NTU)

Irra

dian

ce (

µmol

pho

tons

m�

2 s�

1 )

(a)

(b)

0 5 10 15 20 25 300

200

400

600

800

1000

1200

y � 0.1009x � 0.2603 r2 � 0.68

Fig. 6. Relationship between (a) benthic irradiance at noon and turbidityand (b) Kd (PAR; photosynthetically active radiation); attenuation coef-ficient for noon downward irradiance corrected for tide and solar zenithangle, and turbidity recorded by a logger deployed at 2 m depth at HorseshoeBay between May 2005 and May 2007. Dashed lines are 95% confidenceintervals.

P = 0.0001), with levels ∼1.7-fold greater at Horseshoe Bay(21.60 ± 3.52 µmol L−1) than Davies (12.96 ± 1.33 µmol L−1)or Broadhurst Reefs (12.91 ± 1.17 µmol L−1).

Coral bioindicatorsThe PCA illustrated relationships of physiological variables ofP. damicornis among each other and their distribution acrossthe locations and times of sampling. Physiological measures ofcorals from Horseshoe Bay, such as concentration of chloro-phyll a, symbiont density and colony brightness of P. damicornis,associated positively with the WQI and negatively with mean14-day benthic irradiance (Fig. 7a). In general, physiologicalmeasures of corals from Horseshoe Bay tended to be spatiallydistinct from those at the mid-shelf reefs, although there wassome overlap of the measures during the study (Fig. 7a). Thetemporal PCA showed the physiological measures of P. dami-cornis grouped together in the Dry Season of Year 1 but therewas no separation among seasons in Year 2 (Fig. 7b).

Coral bioindicator temporal variability Marine and Freshwater Research 711

Chlorophyll a per cell

Benthic irradiance

Dim 1: 83.15%

Dim

2: 1

4.48

%

Skeletal density

Chlorophyll a

Symbiont density

Colony brightness

SST

Turbidity

WQI

Chlorophyll a per cell

Benthic irradiance

Skeletal density

Chlorophyll a

Symbiont density

Colony brightness

SST

Turbidity

WQI

(a)

(b)

Fig. 7. Principal components biplot of physiological variables of P. dam-icornis grouped by (a) locations: Horseshoe Bay (dark grey) and mid-shelfreefs (light grey); and (b) seasons: Dry Season ofYear 1 (light grey), and WetSeason of Year 1 and both Dry and Wet Season of Year 2 (dark grey). Envi-ronmental variables are overlaid on the plot. Abbreviations: WQI = waterquality index, SST = sea surface temperature. Data for benthic irradiance,turbidity and SST are 14-day averages for time preceding each samplingevent.

The patterns of variability of the coral indicators were dom-inated by inconsistent variation between years and seasons andamong locations (Year × Season × Location interactions). AtHorseshoe Bay, the density of symbionts of P. damicornis wasgreater in the Dry than the Wet Season of Year 1 (F2,16 = 7.08,P = 0.0063; Fig. 8), but there were no differences between sea-sons at this location in Year 2 nor between seasons for eitheryear at Davies and Broadhurst Reef. Similarly, the colonies ofP. damicornis at Horseshoe Bay were darker in the Dry than theWet Season ofYear 1 (F2,16 = 7.03, P = 0.0064; Fig. 8), but there

were no differences between seasons at this location inYear 2 norwere there any differences for colony brightness between seasonsfor either year at Davies and Broadhurst Reef. These patternswere similar for concentrations of chlorophyll a (µg cm−2) ofP. damicornis, which were greater in Dry compared with theWet Season of Year 1 in Horseshoe Bay. However, this patternwas reversed in Year 2, when greater amounts of chlorophyll aoccurred in the Wet than Dry Season (F2,16 = 10.82, P = 0.0011;Fig. 8). The content of chlorophyll a per symbiont (pg cell−1)varied inconsistently between years and seasons (Fig. 8). InYear 1, there was more chlorophyll a per symbiont in the Wetthan Dry Season (F1,16 = 8.42, P = 0.0032; Fig. 8), but therewere no differences between seasons in Year 2. The densityof macro-bioeroders of Porites differed among locations withmore macro-bioeroders recorded at Horseshoe Bay than at eitherDavies or Broadhurst Reefs (F2,2 = 104.35, P = 0.0095; Fig. 8).

The model selection showed that there was no clear predic-tor that contributed consistently to the fit of the models forthe response variables of P. damicornis and Porites (Table 1).Location was an important predictor for symbiont density andconcentrations of chlorophyll a of P. damicornis, colony bright-ness for both taxa and the density of macro-bioeroders of Porites.However, changes in turbidity, benthic irradiance, SST and theWQI also needed to be considered in the models, particularlyfor skeletal density and the concentration of chlorophyll a persymbiont of P. damicornis.

Symbiont density of P. damicornis varied inconsistentlyamong locations with changes in benthic irradiance and theWQI (Table 1). However, the best predictor of symbiont den-sity was mean 14-day SST (F1,210 = 327.48, P < 0.0001). Thiswas evidenced by the decrease in symbiont density during theWet Season of Year 1, when summer water temperatures wereelevated on the GBR (Fig. 8).TheWQI and Location were impor-tant predictors for concentrations of chlorophyll a of P. dami-cornis, which increased with increased WQI (F1,210 = 327.48,P < 0.0001) and were greater at Horseshoe Bay than Davies orBroadhurst Reefs (F2,210 = 87.22, P < 0.0001). Colony bright-ness of P. damicornis varied inconsistently among locations withchanges in SST and benthic irradiance.

Changes in SST and benthic irradiance were important pre-dictors for responses of chlorophyll a per symbiont cell ofP. damicornis (Table 1). Interestingly, the amount of chloro-phyll a per symbiont cell increased with increases in SST(F2,213 = 32.94, P < 0.0001) but decreased with increases inbenthic irradiance (F2,213 = 36.96, P < 0.0001). In contrast,changes in the WQI and SST were important predictors forresponses of skeletal density of P. damicornis. Skeletal densitydecreased with increases in the WQI (F2,212 = 8.66, P = 0.0036)and SST (F2,212 = 8.80, P = 0.0034).

For the Porites indicators, changes in turbidity, the WQIand Location were useful predictors of colony brightness(Table 1). Colony brightness increased with increases in turbidity(F1,711 = 89.20, P < 0.0001) and WQI (F1,711 = 337.99,P < 0.0001), with the colonies darker at Horseshoe Bay thanat Davies and Broadhurst Reefs (F2,711 = 72.74, P < 0.0001).Location was a strong predictor of the density of macro-bioeroders in Porites, with greater densities recorded on coloniesfrom Horseshoe Bay compared with either Davies or BroadhurstReefs (F2,717 = 403.96, P < 0.0001).

712 Marine and Freshwater Research T. F. Cooper et al.

Ske

leta

l den

sity

(g

cm�

3 )

Sym

bion

t den

sity

(�

106

cm�

2 )

Chl

orop

hyll

a (µ

g cm

�2 )

Chl

orop

hyll

a pe

r sy

mbi

ont c

ell

(pg

cell�

1 )

(b)(a)

(c) (d )

Year 1 Year 2

May Sep Jan May Sep Jan May0.0

1.0

2.0

3.0

4.0

May Sep Jan May Sep Jan May 0

4

8

12

16

20

May Sep Jan May Sep Jan May 0

2

4

6

8

10

May Sep Jan May Sep Jan May2.2

2.3

2.4

2.5

2.6

2.7

2.8

2.9

May Sep Jan May Sep Jan May 2

3

4

5

6

May Sep Jan May Sep Jan May2

3

4

5

6

May Sep Jan May Sep Jan May 0

50

100

150

200

250

300

350

Col

ony

brig

htne

ss(c

olou

r sc

ore)

(e) (f )

Den

sity

bio

erod

ers

(m�

2 )

(g)

Col

ony

brig

htne

ss

(

colo

ur s

core

)

Year 1 Year 2

2005 2006 2007

2005 2006 2007

Fig. 8. Summary of mean (± s.e.) (a) symbiont density (cells cm−2), (b) chlorophyll a (µg cm−2), (c) chlorophyll a per symbiontcell (pg cell−1), (d) skeletal density (g cm−3), (e) colony brightness of P. damicornis, and ( f ) colony brightness; (g) density ofmacro-bioeroders of Porites sampled at each of Horseshoe Bay (•), and Davies (�) and Broadhurst Reefs (�).

Discussion

The present study is the first to document the temporal dynamicsof coral bioindicators and environmental conditions on a coastalcoral reef on the Great Barrier Reef over a 2-year study period.

The symbionts of P. damicornis and colony brightness of Poritesappeared to be seasonally modulated in Year 1, which is consis-tent with studies in other coral systems (Stimson 1997; Brownet al. 1999; Fagoonee et al. 1999; Fitt et al. 2000). This was not

Coral bioindicator temporal variability Marine and Freshwater Research 713

Table 1. Summary of analyses for model selection to examine the relationship betweenphysiological measures of P. damicornis (a–e) and Porites (f–g) with environmental

variables (predictors)Predictors selected by dropping terms from the full model based on calculation of AkaikeInformation Criterion (AIC). Data for sea surface temperature (SST), benthic irradiance andturbidity averaged over the 14-day period preceding the site visit. WQI = water quality index

Variate d.f. AIC F P

(a) Symbiont density (cells cm−2) −154.91Benthic irradiance 1 −146.69 10.18 0.0016WQI 1 −145.86 11.03 0.0011SST 1 −133.58 23.95 <0.0001Location 2 −52.80 66.61 <0.0001

(b) Chlorophyll a (µg cm−2) −166.95WQI 1 −153.04 16.21 0.0001Location 2 −41.62 86.91 <0.0001

(c) Chlorophyll a per symbiont cell (pg cell−1) −290.38Benthic irradiance 1 −258.05 36.69 <0.0001SST 1 −243.92 53.58 <0.0001

(d) Skeletal density (g cm−3) −1025.30Turbidity 1 −1022.35 4.92 0.0277SST 1 −1018.34 8.98 0.0031WQI 1 −1013.07 14.44 0.0002

(e) Colony brightness: P. damicornis −588.28Benthic irradiance 1 −586.43 3.79 0.0530SST 1 −576.02 14.39 0.0002Location 2 −452.60 95.91 <0.0001

(f) Macro-bioeroders (m−2) 2100.10Turbidity 1 2100.80 2.70 0.1037Location 2 2550.60 315.00 <0.0001

(g) Colony brightness: Porites −1926.31WQI 1 −1919.28 9.03 0.0028Turbidity 1 −1918.32 9.99 0.0016Location 2 −1797.51 72.41 <0.0001

the case in Year 2, however, where there was no clear differencebetween seasons for many of the bioindicators examined. Theinterruption of this seasonality coincided with a major floodevent (Fig. 2) but also with slightly cooler SST than the pre-vious year. Indeed, the predictive models found support for arange of environmental parameters for the bioindicators exam-ined including spatial differences between coastal and mid-shelflocations and changes in water quality, SST, turbidity and benthicirradiance. Thus, our findings are discussed within the context ofseasonal changes in water quality as well as temporal differencesin SST.

Coral bioindicatorsResponses of symbiont density, concentrations of chlorophyll a,skeletal density of P. damicornis, and colony brightness ofPorites were related to seasonal changes in water quality.Laboratory studies have previously demonstrated physiologicalresponses of corals to changes in water quality (Hoegh-Guldbergand Smith 1989; Muscatine et al. 1989; Marubini and Davies1996). Similarly, long-term studies in the field have foundpositive relationships between symbiont density and nutrientavailability (Stimson 1997; Fagoonee et al. 1999) althoughBrown et al. (1999) did not find any relationships betweenthese parameters in Indonesia, most likely owing to the lack

of seasonality of nutrient availability at their study site. Here,levels of water column chlorophyll a were ∼1.6-fold greater dur-ing the Wet than Dry Season, a finding that is in agreement withBrodie et al. (2007). The predictive models showed that seasonalchanges in the WQI had some influence on responses such assymbiont density, concentrations of chlorophyll a, skeletal den-sity of P. damicornis and colony brightness of Porites. However,other environmental parameters such as SST, turbidity, benthicirradiance and spatial differences among locations were alsoimportant predictors for these bioindicators. Interestingly, thedensity of macro-bioeroders in Porites did not show any responseto seasonal changes in SST, WQI or to the acute effects of amajor flood event (sensu Hutchings et al. 2005). Nevertheless,the density of macro-bioeroders was consistently greater at thecoastal location than the mid-shelf reference locations indicatinga high specificity to spatial differences in water quality. Thus,the density of macro-bioeroders may be an appropriate indicatorof the effects of chronic exposures to changes in water qual-ity on coastal coral reefs. These results demonstrated that coralbioindicators have differences in specificity to changes in waterquality on coastal coral reefs, which has important implicationsfor the selection of bioindicators for monitoring programs.

The inverse relationship observed between symbiont densityand chlorophyll a concentration of P. damicornis, and colony

714 Marine and Freshwater Research T. F. Cooper et al.

brightness of both study taxa, with irradiance and SST duringYear 1 at the coastal location (Fig. 8) was consistent with sea-sonal regulation of symbiont density reported from other coralreef systems (e.g. Stimson 1997; Brown et al. 1999; Fagooneeet al. 1999). Interestingly, a 4-year study of seasonal patternsin five species of corals from the Bahamas by Fitt et al. (2000)showed that symbiont densities were consistently greater duringthe cooler winter months than summer, leading to the conclu-sion that corals undergo some degree of bleaching every yearregardless of whether it was a so-called ‘bleaching year’. InYear 1 of the present study, mean summer SSTs reached levelsknown to cause thermal stress to corals and low-level bleachingevents, i.e. <10% bleached coral cover, occurred in many regionsof the GBR including the Townsville region over the summerof 2005–06 (Great Barrier Reef Marine Park Authority 2006).Thus, the observed pattern in symbiont density at the coastallocation in Year 1 was most likely a physiological response tothermal stress. In contrast, the mean summer SSTs inYear 2 wereslightly cooler, which may have hindered decreases in symbiontdensity.Alternatively, symbiont density may have remained highthroughoutYear 2 in response to the increased nutrient availabil-ity during the summer flood event. The relative importance ofthese two factors in contributing to the observed patterns forsymbiont density, concentration of chlorophyll a and colonybrightness is uncertain and warrants further study, incorporatingmore years without bleaching conditions.

The positive relationship between chlorophyll a content ofsymbiont cells of P. damicornis with SST at the coastal andmid-shelf locations as observed in Year 1 contrasts with previ-ous studies of seasonal dynamics of chlorophyll a content andsymbiont density in corals (e.g. Fitt et al. 2000). A possibleexplanation for this may be shuffling of different Symbiodiniumgenotypes. Symbiont shuffling is a mechanism that involveschanges in the relative abundance of co-occurring Symbio-dinium types that may differ physiologically (Little et al. 2004;Berkelmans and van Oppen 2006). Shuffling of Symbiodiniumtypes may thus provide selective physiological advantages to thecoral host exposed to changing environmental conditions, e.g.changes in SST and irradiance (Rowan et al. 1997; Berkelmansand van Oppen 2006). Recently, Apprill et al. (2007) showedthat subpopulations of the same Symbiodinium genotype (cladeC) differed in their fluorescence properties owing to variablechlorophyll a content, possibly as an adaptation to specific lightregimes. Characterisation of Symbiodinium genotypes of P. dam-icornis was not done as part of this study, although Ulstrup et al.(2006) showed P. damicornis at Magnetic Island contained onlyone genotype (clade C1). However, the possibility that sym-biont shuffling may have contributed to the patterns observedin P. damicornis, particularly if the retained Symbiodinium type(or subpopulation) had greater chlorophyll a content than thosethat were expelled, cannot be discounted until further studiesare completed.

Turbidity and benthic irradianceWeather-dependent turbidity events are important influenceson environmental conditions on coastal coral reefs (Larcombeet al. 1995; Anthony et al. 2004; Orpin et al. 2004; Alongi andMcKinnon 2005). During this study, there were 10 turbidity

events lasting ≥5 days whereby turbidity was elevated to ≥5NTU, i.e. causing a 94% reduction in benthic irradiance intensity.Anthony et al. (2004) compared patterns of benthic irradiancewith predictions of the minimum saturating irradiance (Ek)for Turbinaria mesenteria and found that variation in envi-ronmental conditions resulted in alternating periods of lightlimitation and light stress on cycles of between 2 and 8 weeksrelated predominantly to changes in turbidity. The Ek of Sym-biodinium associated with P. damicornis was ∼206 ± 8 µmolphotons m−2 s−1 (mean ± s.e., n = 24) for corals sampled fromfour coastal coral reefs of the GBR at a comparable depth tothat used here (Cooper 2008). In the present study, the simulta-neous in situ measurement of turbidity and benthic irradianceby a logger deployed at an average depth of 3.5 m allowedthe quantification of turbidity beyond which light limitationfor corals may occur. The linear relationship between Kd andturbidity showed that benthic irradiance (Ez) was ∼200 µmolphotons m−2 s−1 when turbidity reached 3 NTU. Thus, long-term average turbidity >3 NTU may lead to light limitation(i.e. Ez < Ek), and hence sublethal photo-physiological stressfor Symbiodinium hosted by P. damicornis. At 4.5 NTU, benthicirradiance was ∼6–8% of surface irradiance, which was the min-imum amount of downward irradiance required for coral reefdevelopment in the Whitsunday Islands (Cooper et al. 2007).Hence, long-term average turbidity >5 NTU is likely to rep-resent a threshold for severe stress. Physiological responses ofSymbiodinium to light limitation may in part explain the con-sistency observed in spatial differences of parameters such assymbiont density and concentrations of chlorophyll a (Falkowskiand Dubinsky 1981; Dubinsky et al. 1984). However, giventhat re-mobilisation of particulate nutrients bound to sedimentsis likely to be a feature of these weather-dependent turbidityevents (Alongi and McKinnon 2005), increased nutrient avail-ability may have also influenced the population dynamics ofSymbiodinium (Hoegh-Guldberg and Smith 1989; Muscatineet al. 1989).

In conclusion, this study has highlighted the importance ofunderstanding temporal patterns of variation as part of the pro-cess of selecting bioindicators for monitoring the effects ofchanges in water quality on coastal coral reefs. Importantly,the study revealed that the specificity to changes in water qual-ity varied among bioindicators. For example, the best predictorof symbiont density of P. damicornis, which in turn influencedcolony brightness, was mean 14-day SST followed by seasonalchanges in water quality and mean 14-day benthic irradiance.Thus, temporal variability of a range of environmental param-eters needs to be considered if physiological measures such ascolony brightness are used in water quality monitoring programson coastal coral reefs. Our findings have also further enhancedthe knowledge of environmental conditions on coastal coral reefsby quantifying potential stress thresholds of turbidity for coralsoccurring at shallow depths. The extinction of benthic irradiancewhen turbidity exceeded ∼15 NTU is notable; however, as 94%extinction occurs at a much lower value of only 5 NTU, largeincreases in turbidity to 15 NTU or more make little differenceto the light levels and thus the energy balance of corals. How-ever, the biological responses of corals will vary depending ondepth and species as well as the duration of exposure to tur-bidity. A detailed understanding of the physiological responses

Coral bioindicator temporal variability Marine and Freshwater Research 715

of a range of coastal corals using dose–response experiments isrequired before long-term turbidity >3 NTU for sublethal stressand >5 NTU for severe stress can be considered as ‘thresholdsof concern’.

AcknowledgementsThis study was supported by the Reef and Rainforest Research Centrethrough the Marine and Tropical Sciences Research Facility (MTSRF), jointcontributions from the Co-operative Research Centre (CRC) for Coral Reefsand Rainforest CRC to the Catchment to Reef Program, and the AustralianInstitute of Marine Science. The research was undertaken in accordancewith Marine Parks Permit G05/13484.1 (PVR) and G06/15571.1 (AIMS).We gratefully acknowledge the support provided in the field and labora-tory from Ray Berkelmans, Joe Goiffre, Tim Hyndes, David McKinnon,Tim Phillips, Sven Uthicke, Jake van Oosterom and Madeleine van Oppen.We also thank two anonymous referees for comments that improved themanuscript.

ReferencesAkaike, H. (1974). A new look at the statistical model identification.

IEEE Transactions on Automatic Control 19, 716–723. doi:10.1109/TAC.1974.1100705

Alongi, D. M., and McKinnon, A. D. (2005). The cycling and fate ofterrestrially-derived sediments and nutrients in the coastal zone ofthe Great Barrier Reef shelf. Marine Pollution Bulletin 51, 239–252.doi:10.1016/J.MARPOLBUL.2004.10.033

Anthony, K. R. N., and Fabricius, K. E. (2000). Shifting roles of het-erotrophy and autotrophy in coral energetics under varying turbidity.Journal of Experimental Marine Biology and Ecology 252, 221–253.doi:10.1016/S0022-0981(00)00237-9

Anthony, K. R. N., Connolly, S. R., and Willis, B. L. (2002). Comparativeanalysis of energy allocation to tissue and skeletal growth in corals.Limnology and Oceanography 47, 1417–1429.

Anthony, K. R. N., Ridd, P. V., Orpin, A. R., Larcombe, P., and Lough, J.(2004).Temporal variation of light availability in coastal benthic habitats:effects of clouds, turbidity and tides. Limnology and Oceanography 49,2201–2211.

Apprill,A. M., Bidigare, R. R., and Gates, R. D. (2007).Visibly healthy coralsexhibit variable pigment concentrations and symbiont phenotypes. CoralReefs 26, 387–397. doi:10.1007/S00338-007-0209-Y

Baith, K., Lindsay, R., Fu, G., and McClain, C. R. (2001). SeaDAS: dataanalysis system developed for ocean color satellite sensors. Eos, Trans-actions, American Geophysical Union 82, 202. doi:10.1029/01EO00109

Barnes, D. J., and Chalker, B. E. (1990). Calcification and photosynthesis inreef-building corals and algae. In ‘Coral Reefs: Ecosystems of the World25’. (Ed. Z. Dubinsky.) pp. 109–131. (Elsevier: Amsterdam.)

Barnes, D. J., and Lough, J. M. (1992). Systematic variations in the depthof skeleton occupied by coral tissue in massive colonies of Porites fromthe Great Barrier Reef. Journal of Experimental Marine Biology andEcology 159, 113–128. doi:10.1016/0022-0981(92)90261-8

Berkelmans, R., and van Oppen, M. J. H. (2006).The role of zooxanthellae inthe thermal tolerance of corals: a ‘nugget of hope’for coral reefs in an eraof climate change. Proceedings of the Royal Society of London. SeriesB. Biological Sciences 273, 2305–2312. doi:10.1098/RSPB.2006.3567

Brodie, J., De’ath, G., Devlin, M. J., Furnas, M. J., and Wright, M. (2007).Spatial and temporal trends of near-surface chlorophyll a in the GreatBarrier Reef lagoon. Marine and Freshwater Research 58, 342–353.doi:10.1071/MF06236

Brown, B. E., Dunne, R. P., Ambarsari, I., Le Tissier, M. D. A., andSatapoomin, U. (1999). Seasonal fluctuations in environmental fac-tors and variations in symbiotic algae and chlorophyll pigments in fourIndo-Pacific coral species. Marine Ecology Progress Series 191, 53–69.doi:10.3354/MEPS191053

Bureau of Meteorology (2007). ‘Flooding in the North and Central Coasts ofQueensland February 2007.’Available at http://www.bom.gov.au/hydro/flood/qld/fld_reports/reports.shtml [accessed 20 June 2007].

Burnham, K. P., and Anderson, D. R. (2002). ‘Model Selection andMultimodel Inference: a Practical Information-Theoretic Approach.’(Springer: New York.)

Cooper, T. F. (2008). Coral bioindicators of environmental conditions oncoastal coral reefs. PhD thesis, James Cook University, Townsville.

Cooper, T. F., Uthicke, S., Humphrey, C., and Fabricius, K. E. (2007).Gradients in water column nutrients, sediment parameters, irradianceand coral reef development in the Whitsunday Region, central GreatBarrier Reef. Estuarine, Coastal and Shelf Science 74, 458–470.doi:10.1016/J.ECSS.2007.05.020

Dubinsky, Z., Falkowski, P. G., Porter, J. W., and Muscatine, L. (1984).Absorption and utilisation of radiant energy by light- and shade-adaptedcolonies of the hermatypic coral Stylophora pistillata. Proceedingsof the Royal Society of London. Series B. Biological Sciences 222,203–214.

Fabricius, K. E. (2005). Effects of terrestrial runoff on the ecology of coralsand coral reefs: review and synthesis. Marine Pollution Bulletin 50, 125–146. doi:10.1016/J.MARPOLBUL.2004.11.028

Fabricius, K. E., and De’ath, G. (2004). Identifying ecological change andits causes: a case study on coral reefs. Ecological Applications 14, 1448–1465. doi:10.1890/03-5320

Fabricius, K. E., De’ath, G., McCook, L., Turak, E., and Williams, D. M.(2005). Changes in algal, coral and fish assemblages along water qualitygradients on the inshore Great Barrier Reef. Marine Pollution Bulletin51, 384–398. doi:10.1016/J.MARPOLBUL.2004.10.041

Fagoonee, I., Wilson, H. B., Hassell, M. P., and Turner, J. R. (1999). Thedynamics of zooxanthellae populations: a long-term study in the field.Science 283, 843–845. doi:10.1126/SCIENCE.283.5403.843

Falkowski, P. G., and Dubinsky, Z. (1981). Light-shade adaptation of Sty-lophora pistillata, a hermatypic coral from the Gulf of Eilat. Nature 289,172–174. doi:10.1038/289172A0

Feldman, G. C., and McClain, C. R. (2007). Ocean Color Web, MODIS Level0, NASA Goddard Space Flight Center, February 2007. (Eds N. Kuringand S. W. Bailey.) Available at http://oceancolor.gsfc.nasa.gov/ [accessed22 February 2007].

Fitt, W. K., McFarland, F. K., Warner, M. E., and Chilcoat, G. C. (2000).Seasonal patterns of tissue biomass and densities of symbiotic dinoflag-ellates in reef corals and relation to coral bleaching. Limnology andOceanography 45, 677–685.

Francis, J. M., Dunbar, G. B., Dickens, G. R., Sutherland, I. A., andDroxler, A. W. (2007). Siliciclastic sediment across the North Queens-land margin (Australia): a Holocene perspective on reciprocal versuscoeval deposition in tropical mixed siliciclastic–carbonate systems. Jour-nal of Sedimentary Research 77, 572–586. doi:10.2110/JSR.2007.057

Furnas, M. J. (2003). ‘Catchments and Corals: Terrestrial Runoff to theGreat Barrier Reef.’ (Australian Institute of Marine Science, CRC Reef:Townsville.)

Great Barrier Reef Marine Park Authority (2006). ‘Final Bleaching Sum-mary Report 2005/2006.’ Available at www.gbrmpa.gov.au/corp_site/info_services/science/climate_change/conditions_report.html [accessed20 June 2007].

Hill, R., and Ralph, P. J. (2005). Diel and seasonal changes in fluorescencerise kinetics of three scleractinian corals. Functional Plant Biology 32,549–559. doi:10.1071/FP05017

Hoegh-Guldberg, O. (1999). Climate change, coral bleaching and the futureof the world’s coral reefs. Marine and Freshwater Research 50, 839–866.doi:10.1071/MF99078

Hoegh-Guldberg, O., and Smith, G. J. (1989). Influence of the popula-tion density of zooxanthellae and supply of ammonium on the biomassand metabolic characteristics of the reef corals Seriatopora hystrix andStylophora pistillata. Marine Ecology Progress Series 57, 173–186.doi:10.3354/MEPS057173

716 Marine and Freshwater Research T. F. Cooper et al.

Hutchings, P. A., and Peyrot-Clausade, M. (2002). The distribution andabundance of boring species of polychaetes and sipunculans in coralsubstrates in French Polynesia. Journal of Experimental Marine Biologyand Ecology 269, 101–121. doi:10.1016/S0022-0981(02)00004-7

Hutchings, P., Peyrot-Clausade, M., and Osnorno, A. (2005). Influence ofland runoff on rates and agents of bioerosion of coral substrates. MarinePollution Bulletin 51, 438–447. doi:10.1016/J.MARPOLBUL.2004.10.044

Jeffrey, S. W., and Humphrey, G. F. (1975). New spectrophotometric equa-tions for determining chlorophylls a, b, c1 and c2 in higher plants, algaeand natural phytoplankton. Biochemie und Physiologie der Pflanzen 167,191–194.

Jones, R. J., and Hoegh-Guldberg, O. (2001). Diurnal changes in the pho-tochemical efficiency of the symbiotic dinoflagellates (Dinophyceae) ofcorals: photoprotection, photoinactivation and the relationship to coralbleaching. Plant, Cell & Environment 24, 89–99. doi:10.1046/J.1365-3040.2001.00648.X

Jones, R. J., Hoegh-Guldberg, O., Larkum,A.W. D., and Schreiber, U. (1998).Temperature-induced bleaching of corals begins with impairment of theCO2 fixation mechanisms in zooxanthellae. Plant, Cell & Environment21, 1219–1230.

Kirk, J. T. O. (1994). ‘Light and Photosynthesis in Aquatic Ecosystems.’(Cambridge University Press: Cambridge.)

Larcombe, P., Ridd, P.V., Prytz,A., and Wilson, B. (1995). Factors controllingsuspended sediment on inner-shelf coral reefs, Townsville, Australia.Coral Reefs 14, 163–171. doi:10.1007/BF00367235

Lesser, M. P., and Gorbunov, M.Y. (2001). Diurnal and bathymetric changesin chlorophyll fluorescence yields of reef corals measured in situ witha fast repetition rate fluorometer. Marine Ecology Progress Series 212,69–77. doi:10.3354/MEPS212069

Leuzinger, S., Anthony, K. R., and Willis, B. L. (2003). Reproductive energyinvestment in corals: scaling with module size. Oecologia 136, 524–531.doi:10.1007/S00442-003-1305-5

Little, A. F., van Oppen, M. J. H., and Willis, B. L. (2004). Flexibility in algalendosymbioses shapes growth in reef corals. Science 304, 1492–1494.doi:10.1126/SCIENCE.1095733

Lough, J. M., and Barnes, D. J. (2000). Environmental controls on growth ofthe massive coral Porites. Journal of Experimental Marine Biology andEcology 245, 225–243. doi:10.1016/S0022-0981(99)00168-9

Marubini, F., and Davies, P. S. (1996). Nitrate increases zooxanthellae pop-ulation density and reduces skeletogenesis in corals. Marine Biology(Berlin) 127, 319–328. doi:10.1007/BF00942117

Meesters, E. H., Nieuwland, G., Duineveld, G. C. A., Kok, A., and Bak,R. P. M. (2002). RNA/DNA ratios of scleractinian corals suggestacclimatisation/adaptation in relation to light gradients and turbidityregimes. Marine Ecology Progress Series 227, 233–239. doi:10.3354/MEPS227233

Mobley, C. D. (1994). ‘Light and Water: Radiative Transfer in NaturalWaters.’ (Academic Press: San Diego, CA.)

Muscatine, L. (1990). The role of symbiotic algae in carbon and energyflux in coral reefs. In ‘Ecosystems of the World 25, Coral Reefs’.(Ed. Z. Dubinsky.) pp. 75–87. (Elsevier: Amsterdam.)

Muscatine, L., Falkowski, P. G., Dubinsky, Z., Cook, P. A., andMcCloskey, L. R. (1989). The effect of external nutrient resources onthe population dynamics of zooxanthellae in a reef coral. Proceedings ofthe Royal Society of London. Series B. Biological Sciences 236, 311–324.

Orpin, A. R., Ridd, P. V., Thomas, S., Anthony, K. R. N.,Marshall, P., and Oliver, J. (2004). Natural turbidity variability and

http://www.publish.csiro.au/journals/mfr

weather forecasts in risk management of anthropogenic sediment dis-charge near sensitive environments. Marine Pollution Bulletin 49,602–612. doi:10.1016/J.MARPOLBUL.2004.03.020

Pillay, R. M., Willis, B., and Terashima, H. (2005). Trends in the densityof zooxanthellae in Acropora millepora (Ehrenberg, 1834) at the PalmIsland Group, Great Barrier Reef, Australia. Symbiosis 38, 209–226.

R Development Core Team (2006). ‘R: A Language and Environmentfor Statistical Computing.’ (R Foundation for Statistical Computing:Vienna.)

Ridd, P. V., and Larcombe, P. (1994). Biofouling control for opticalbackscatter suspended sediment sensors. Marine Geology 116, 255–258.doi:10.1016/0025-3227(94)90044-2

Risk, M. J., Heikoop, J. M., Edinger, E. N., and Erdmann, M. V. (2001). Theassessment ‘toolbox’: Community-based reef evaluation methods cou-pled with geochemical techniques to identify sources of stress. Bulletinof Marine Science 69, 443–458.

Rowan, R., Knowlton, N., Baker, A., and Jara, J. (1997). Landscape ecol-ogy of algal symbionts creates variation in episodes of coral bleaching.Nature 388, 265–269. doi:10.1038/40843

Siebeck, U. E., Marshall, N. J., Klueter, A., and Hoegh-Guldberg, O. (2006).Monitoring coral bleaching using a colour reference card. Coral Reefs25, 453–460. doi:10.1007/S00338-006-0123-8

Stimson, J. (1997). The annual cycle of density of zooxanthellae in the tis-sues of field and laboratory-held Pocillopora damicornis (Linnaeus).Journal of Experimental Marine Biology and Ecology 214, 35–48.doi:10.1016/S0022-0981(96)02753-0

Stimson, J., and Kinzie, R. A., III (1991). The temporal patterns andrate of release of zooxanthellae from the reef coral Pocillopora dam-icornis (Linnaeus) under nitrogen-enrichment and control conditions.Journal of Experimental Marine Biology and Ecology 153, 63–74.doi:10.1016/S0022-0981(05)80006-1

Ulstrup, K. E., Berkelmans, R., Ralph, P. J., and van Oppen, M. J. H. (2006).Variation in bleaching sensitivity of two coral species across a latitudinalgradient on the Great Barrier Reef: the role of zooxanthellae. MarineEcology Progress Series 314, 135–148. doi:10.3354/MEPS314135

Ulstrup, K. E., Hill, R., van Oppen, M. J. H., Larkum, A. W. D., andRalph, P. J. (2008). Seasonal variation in the photo-physiology of homo-geneous and heterogeneous Symbiodinium consortia of two scleractiniancorals. Marine Ecology Progress Series 361, 139–150. doi:10.3354/MEPS07360

Underwood, A. J. (1997). ‘Experiments in Ecology: Their Logical Designand Interpretation Using Analysis of Variance.’ (Cambridge UniversityPress: Cambridge.)

Warner, M. E., Chilcoat, G. C., McFarland, F. K., and Fitt, W. K. (2002).Seasonal fluctuations in the photosynthetic capacity of photosystem IIin symbiotic dinoflagellates in the Caribbean reef-building coral Mon-tastraea. Marine Biology 141, 31–38. doi:10.1007/S00227-002-0807-8

Winters, G., Loya, Y., and Beer, S. (2006). In situ measured seasonal vari-ations in F-v/F-m of two common Red Sea corals. Coral Reefs 25,593–598. doi:10.1007/S00338-006-0144-3

Wolanski, E., Fabricius, K. E., Cooper, T. F., and Humphrey, C. (2008).Wet season fine sediment dynamics on the inner shelf of the GreatBarrier Reef. Estuarine, Coastal and Shelf Science 77, 755–762.doi:10.1016/J.ECSS.2007.10.014

Manuscript received 26 January 2008, accepted 2 June 2008