DISEASES OF AQUATIC ORGANISMSDis Aquat Org

Vol 113 59ndash68 2015doi 103354dao02828

Published February 10

INTRODUCTION

Diseases have the potential to severely alter thestructure and function of marine ecosystems (Ward ampLafferty 2004) In some species such as the long-spined sea urchin in the Caribbean (Lessios 1988)and the abalone in California USA (Lafferty amp Kuris1993) infectious diseases have reduced populationdensities to such an extent that recovery is uncertain(Lafferty et al 2004) The same is true for coral dis-

eases Coral disease outbreaks have caused declinesin coral cover in many reef systems (Nugues 2002Croquer et al 2005 Bruckner amp Hill 2009) For exam-ple disease outbreaks in the Caribbean have causeda severe reduction in the abundance of the 2 previ-ously most abundant hard corals Acropora palmataand A cervicornis causing a shift in the reef commu-nity structure (Gladfelter 1982 Aronson amp Precht2001 Patterson et al 2002) Even though the Indo-Pacific appears to be less affected by coral diseases

copy Inter-Research 2015 middot wwwint-rescomCorresponding author simondavyvuwacnz

Disease dynamics of Porites bleaching with tissue loss prevalence virulence transmission

and environmental drivers

M Sudek14 G J Williams2 C Runyon3 G S Aeby3 S K Davy1

1School of Biological Sciences Victoria University of Wellington PO Box 600 Wellington 6140 New Zealand2Center for Marine Biodiversity amp Conservation Scripps Institution of Oceanography La Jolla California 92083 USA

3Hawaii Institute of Marine Biology University of Hawaii PO Box 1346 Kaneohe Hawaii 96744 USA4Present address Division of Marine and Wildlife Resources PO Box 3730 Pago Pago 96799 American Samoa

ABSTRACT The prevalence number of species affected and geographical extent of coral dis-eases have been increasing worldwide We present ecological data on the coral disease Poritesbleaching with tissue loss (PBTL) from Kaneohe Bay Oahu (Hawaii USA) affecting P compressaThis disease is prevalent throughout the year although it shows spatio-temporal variability withpeak prevalence during the warmer summer months Temporal variability in disease prevalenceshowed a strong positive relationship with elevated water temperature Spatially PBTL preva-lence peaked in clearer waters (lower turbidity) with higher water flow and higher densities ofparrotfish together explaining approximately 26 of the spatial variability in PBTL prevalenceHowever the relatively poor performance of the spatial model suggests that other unmeasuredfactors may be more important in driving spatial prevalence PBTL was not transmissible throughdirect contact or the water column in controlled aquaria experiments suggesting that this diseasemay not be caused by a pathogen is not highly infectious or perhaps requires a vector for trans-mission In general PBTL results in partial tissue mortality of affected colonies on average one-third of the tissue is lost This disease can affect the same colonies repeatedly suggesting a poten-tial for progressive damage which could cause increased tissue loss over time P compressa is themain framework-building species in Kaneohe Bay PBTL therefore has the potential to negativelyimpact the structure of the reefs at this location

KEY WORDS Porites compressa middot Tissue mortality middot Branching coral middot Framework species middot Coralreef ecology middot Etiology middot Coral disease middot Water temperature

Resale or republication not permitted without written consent of the publisher

Dis Aquat Org 113 59ndash68 2015

than the Caribbean an increasing amount of evi-dence suggests that coral diseases are common(Sutherland et al 2004 Willis et al 2004 Raymundoet al 2005) even at remote uninhabited islands(Williams et al 2008 2011b Vargas-Angel 2009)with the types of diseases and their prevalence varying across multiple spatial scales (Aeby et al2011ab) In fact the geographical extent number ofspecies affected and incidence of new diseases areincreasing globally (Harvell et al 1999 Ward amp Lafferty 2004 Sokolow 2009) Environmental stressshifts in virulence of existing pathogens introductionof novel pathogens from anthropogenic activitiesand global climate change are associated with thisincrease (Harvell et al 1999 2004 Sokolow 2009)

Coral disease prevalence can be expected to showintricate interactions with a variety of driving factors(Williams et al 2010 2014) For example an increasein temperature can lead to an increase in pathogenvirulence or cause stress to the host which canincrease its susceptibility to disease (Harvell et al2002) Coral disease outbreaks and increases in dis-ease prevalence and progression have been linked tovarious environmental factors (eg Bruno et al 20032007 Sato et al 2009 Williams et al 2010 Aeby et al2011b) It is likely that several environmental factorssimultaneously influence disease dynamics within asystem with the relative importance of each factorvarying among regions spatial scales and species(Aeby et al 2011ab) We have just begun to under-stand the complex web of interactions between envi-ronmental factors and disease prevalence exempli-fied by studies that applied a multi-factor approachto studying coral disease dynamics (eg Bruno et al2007 Haapkylauml et al 2007 McClanahan et al 2009Williams et al 2010 2014 Aeby et al 2011ab)

Additionally different coral diseases can showvarying levels of ecological impact For exampleblack band disease (BBD) and white syndromes often cause severe colony mortality (Edmunds 1991Bruckner et al 1997 Roff et al 2006 Aeby et al2010 Williams et al 2011a) whereas corals withPorites ulcerative white spot syndrome (PUWS) in thePhilippines often show complete recovery after infec-tion (Kaczmarsky 2006) In addition different coralspeciestaxa appear to vary in their susceptibility todisease infection For example on the Great BarrierReef (GBR) BBD affects 25 out of approximately 350hard coral species with branching Acropora sppbeing most affected (Page amp Willis 2006) The degreeof damage to the ecosystem therefore depends on thesuite of coral species and diseases that occur on thereef If we are to successfully manage our reef sys-

tems it is vital to understand the often intricate dis-easeminusenvironment interactions that lead to complextemporal and spatial disease dynamics as well as thenature and causes of the different diseases that affecta system

Kaneohe Bay in Oahu Hawaii (USA) containsmany fringing and patch reefs with high coral coverPorites compressa and Montipora capitata are the 2dominant framework-building corals on these reefswith P compressa accounting for up to 80 of thehard coral cover at some sites (Williams et al 2010)Several coral diseases have been reported fromKaneohe Bay of which Porites trematodiasis (Aeby2007 Williams et al 2010) Porites growth anomalies(Domart-Coulon et al 2006 Williams et al 2010Stimson 2011) and Montipora white syndrome(MWS) (Aeby et al 2010 Williams et al 2010) are themost studied In the present study we describe thedynamics of another disease in Kaneohe Bay Poritesbleaching with tissue loss (PBTL) which affects Pcompressa This disease manifests as bleaching ofthe coenenchyme with the polyps remaining browngiving the coral a lsquospeckledrsquo appearance (Fig 1B)which is distinct from the uniform color loss associ-ated with thermal bleaching PBTL causes tissue lossdue to necrosis and tissue fragmentation (Sudek et al2012b) and a significant reduction in gamete devel-opment (Sudek et al 2012a) Preliminary observa-tions of an apparent increase of PBTL prevalenceduring the summer months suggested a potential linkto temperature (M Sudek unpubl data) but overalllittle is known about the ecology of this disease Theobjectives of the present study were therefore to (1) examine the variability in disease prevalence(proportion of individuals affected) over the courseof 1 yr (temporal variability) and determine the spa-tial distribution of PBTL within Kaneohe Bay andHanauma Bay (another reef with high P compressacover located on the windward side of Oahu) (2)examine virulence (degree of harm to the host) (3)investigate disease transmissibility and (4) deter-mine the environmental correlates of variations indisease prevalence

MATERIALS AND METHODS

Prevalence and spatial distribution

Prevalence surveys were conducted at 8 perma-nent sites (AminusD and GminusJ) around Coconut IslandMarine Reserve (CIMR) Kaneohe Bay OahuHawaii USA (21deg 26rsquo N 157deg 47rsquo W Fig 2) on an ap -

60

Sudek et al Disease dynamics of Porites bleaching with tissue loss

proximately monthly basis during 2011 Five 10 times 2 mbelt transects were deployed at each site in whichevery Porites compressa colony was counted andexamined for signs of PBTL

To investigate the larger spatial extent of PBTLrapid visual surveys were conducted on 9 patch reefswithin Kaneohe Bay and on the reef in Hanuama BayMarine Reserve (Fig 2) Within Kaneohe Bay asnorkeler swam for 10 min at a speed of approxi-mately 10 m minminus1 along a haphazardly selectedpatch reef and recorded every PBTL-affected colonyobserved In Hanauma Bay 2 divers swam across thereef at approximately the same speed and recordedthe number of PBTL-affected colonies encounteredwithin 30 min This rapid survey method allows alarger spatial coverage of the reefs and provides asemi-quantitative measure of disease abundance Allrapid surveys were conducted in October 2011

Disease virulence

To determine disease virulence 42 individualPBTL-affected colonies were tagged in 2010 (Sudeket al 2012a) and an additional 36 PBTL-affectedcolonies were tagged in 2011 and followed throughtime (approximately monthly examinations) The co -lo nies that were tagged in 2010 were resurveyed andchecked for new PBTL signs andor signs of tissuerecovery Due to the 3-dimensional structure ofPorites compressa and the often poor visibility inKaneohe Bay we could not rely on photographic sur-veys with post hoc image analysis Instead the per-centage of healthy dead and affected tissue wasestimated visually in situ in addition to photo docu-mentation

Transmission

To determine whether PBTL is transmissible throughthe water column or via direct contact healthy aswell as PBTL-affected coral samples (ap proximately

61

Fig 1 (A) Healthy Porites compressa Note regular browncoloration (B) Porites bleaching with tissue loss (PBTL)-affected P compressa Note pigmented polyps and bleachedcoenenchyme (lsquospeckled appearancersquo) with onset of tissue loss

Fig 2 Kaneohe Bay Hawaii (USA) showing the 9 rapid sur-vey sites (black dots marked 7 to 46) numbered after Roy(1970) with an inset of Oahu showing the location ofKaneohe Bay and Hanauma Bay (arrow) and another insetof Coconut Island Marine Reserve (CIMR) showing the 8

permanent sites (AminusJ)

Dis Aquat Org 113 59ndash68 2015

3 cm2 each) were collected from the reef crest aroundCIMR using a bone cutter Samples were transportedto the lab in individual plastic bags to avoid any crosscontamination Manipulative ex periments were runin aquaria (8 l) under closed conditions To maintainwater quality a bubbler was placed in each aquar-ium to ensure water movement and partial waterchanges using 02 microm filtered seawater were car-ried out every 5 d Aquaria were kept outside undernatural light

To conduct the transmission experiment a coralfragment showing signs of PBTL was placed in anaquarium touching a healthy fragment with anotherfragment from the same healthy colony placed about10 cm away from the PBTL-affected fragment (n =10) As a control for possible effects of intraspecificcompetition the same setup was used with frag-ments from the same healthy colony as those used inthe transmission treatment but the diseased frag-ment was replaced by a healthy fragment from a dif-ferent colony (n = 10) The healthy fragments weremonitored daily for signs of PBTL over the course ofat least 3 wk or until the affected fragment died (max5 wk) The transmission experiment was carried outat ambient (25degC) and increased (28degC) water tem-peratures to determine whether transmission wouldoccur more readily at higher temperatures (n = 10treatmentminus1) Thermal bleaching of Hawaiian coralsoccurs with prolonged exposure to 29minus30degC (Jokiel ampColes 1990) so by choosing a temperature of 28degCthe experimental setup stayed below the range inwhich temperature-induced bleaching would usuallyoccur Additionally the lsquospeckledrsquo appearance ofPBTL is not observed during thermal bleaching orbleaching due to competition therefore only the typ-ical lsquospeckledrsquo bleaching appearance was considereda sign of PBTL

Environmental drivers

All measurements of environmental variables wereconducted at the depths of the transects Tem p e ra -ture data were collected at each site using HOBOregProdata loggers (wwwonsetcompcom) with an accu-racy of plusmn02degC The loggers recorded continuouslyevery 30 min from late February to late December2011 Turbidity chlorophyll a (chl a) and salinitywere measured at each site using an RBRreg XR-420data logger (wwwrbr-globalcom) recording everyminute over a 36 to 48 h period on 4 to 6 differentoccasions per site in 2010 and 2011 The logger wasmoved randomly between sites to maximize spatial

coverage over time Water motion was estimatedusing the clod card technique (Jokiel amp Morrissey1993) Two clod cards were placed at the beginningof each survey site and left overnight (21 to 23 h) Inaddition 2 clod cards were placed into a large bucketcontaining seawater (ca 60 l) to serve as a diffusioncontrol The exact time that the clod cards wereimmersed in water was recorded and the diffusionfactor (DF a dimensionless index of water motion)was calculated for each site (Jokiel amp Morrissey1993) Clod cards were deployed 4 times over thecourse of 6 mo in 2011 and the average DF for eachsite was used in subsequent data analyses

Corallivorous fish can be potential vectors of dis-ease (Aeby amp Santavy 2006) or a source of injurywhich can promote the spread of certain diseases(Page amp Willis 2008 Raymundo et al 2009) The den-sities of all corallivorous butterflyfish (facultative andobligate) and parrotfish were recorded over an areaof 50 times 4 m at all 8 sites The observer swam at aspeed of approximately 10 m minminus1 and recorded allbutterflyfish to species level (Chaetodon auriga Cephippium C lineolatus C lunulatus C multicinc-tus C ornatissimus C unimaculatus) Due to diffi-culties with species-level identification all parrotfish(adults and juveniles) were grouped Fish countswere carried out during 4 different months in 2011(July August September December) and all siteswere surveyed on the same day within 2 to 3 h ofeach other Total numbers of fish were used in thesubsequent data analyses

Statistical analyses

Prevalence and spatial distribution

Prevalence data by transect did not display a nor-mal distribution even after transformation We there-fore used a repeated measures permutational ana -lysis of variance based on a binomial deviance matrix(the technique does not assume normality) in PERM-ANOVA+ (Anderson et al 2008) to test the effect of2 fixed factors (site month) and their inter action withdisease prevalence

Environmental drivers

To investigate temporal variations in disease pre -valence the relationship of temperature and preva-lence was explored over a period of 10 mo using ageneral linear model (GLM) performed with SPSS

62

Sudek et al Disease dynamics of Porites bleaching with tissue loss 63

(PASW 18) Prevalence was averaged for each site(over the 5 individual transects) and the data dis-played a normal distribution Temperature data wereaveraged over the 10 d period before each survey forevery site

To examine spatial variations in disease prevalence(differences in PBTL prevalence across sites regard-less of month) 8 environmental predictor variableswere modeled against spatial variations in preva-lence across sites Measurements for predictor vari-ables were not continuous through time and weretherefore averaged for each site To achieve the sameresolution for temperature and prevalence data alltemperature data (10 d before each survey) andprevalence values (February to December) wereaveraged for each site Predictor variables were hostcover turbidity water temperature chl a watermotion salinity parrotfish density and butterflyfishdensity (Table 1) Because most butterflyfish speciesshowed low abundances on the reef all butterflyfishcounts were grouped The mean and 1 SD of all pre-dictor variables were initially examined to alsoaccount for the variability of factors at the individualsites Inter-correlation of predictor variables wastested using Pearsonrsquos correlation with predictorsexceeding a correlation value of gt075 considered forremoval and further examined using principal coor-dinates analysis (PCO) plots (Anderson et al 2008)Variables chosen for inclusion in the model weremean values for host cover temperature chl a andwater motion and the variability (SD) in turbiditysalinity butterflyfish density and parrotfish densityA permutational distance-based linear model (DIS-TLM) was used (McArdle amp Anderson 2001) to ana-lyze the data DISTLM is a multivariate multipleregression technique that quantifies the proportionof the variation in the response variable (in this case

PBTL prevalence) explained by the predictor vari-ables Environmental data were normalized and the DISTLM routine was run using the lsquobestrsquo selectionprocedure based on 9999 permutations Akaikersquosinformation criterion (Akaike 1973) with a second-order bias correction applied (AICc) (Hurvich amp Tsai1989 Burnham amp Anderson 2004) was used for modelselection The most parsimonious model with thelowest AICc and highest R2 value was selected Mod-eling analyses were based on 0-adjusted Bray-Curtissimilarity matrices (Clarke et al 2006) and carriedout using PRIMER v6 (Clarke amp Gorley 2006) andPERMANOVA+ (Anderson et al 2008)

RESULTS

Prevalence and spatial distribution

Overall average PBTL prevalence at CIMR was15 (plusmn02 SE) PBTL-affected colonies werefound at all 8 survey sites but prevalence differedsignificantly between sites (df = 7 pseudo-F = 95969p lt 0001 Fig 3A) and between months (df = 9pseudo-F = 85552 p lt 0001 Fig 3B) but no signifi-cant interaction between the two was detected (df =63 pseudo-F = 083471 p = 07473) PBTL-affectedcolonies were observed throughout the year with apeak during the summer months (Fig 3B) and thehighest average prevalence observed in June (25 plusmn03 SE) Rapid visual surveys showed that PBTLwas present in all of Kaneohe Bay but that it wasabsent from Hanauma Bay (Table 2)

Disease virulence and transmission

Most colonies affected by PBTL showed the typicallsquospeckledrsquo bleaching for a period of approximately 2to 3 mo and then the disease regressed in most cases(no more signs of bleaching) Within this time themajority of colonies (85) showed tissue loss rangingfrom 5 to 100 of the colony with a case fatality rate(total mortality) of 3 On average a colony lost athird (30) of its tissue within 2 mo Of the 42colonies tagged in 2010 55 showed no signs ofrecovery in 2011 and 24 showed partial tissue re-growth In addition 31 of these colonies becameaffected again by PBTL during 2011

No disease transmission occurred between individ-uals via the water column or direct contact in eitherthe ambient or the increased temperature treatments(n = 10 treatmentminus1)

Variable Description and units Min Max

Water temp degC 221 283

Host cover Porites compressa cover 300 756

Turbidity Formazin turbidity unit (FTU) 03 226

Chl a microg lminus1 002 21

Water motion Diffusion factor (DF) 126 795

Salinity ppt 330 357

Parrotfish Number per 200 m2 0 68density

Butterflyfish Number per 200 m2 0 13density

Table 1 Predictor variables used in model analyses with their units and minimum and maximum values

Dis Aquat Org 113 59ndash68 2015

Environmental drivers

A significant positive linear relationship was foundbetween temporal water temperature and diseaseprevalence (GLM Wald χ2 = 38128 df = 1 p lt 0001Fig 4) Based on this relationship each degree in -crease in temperature can be expected to result in anincrease of 029 to 056 in disease prevalence

Modeling of the spatial variation in PBTL preva-lence across sites identified water motion and thevariability in turbidity and parrotfish density as thestrongest predictors with 262 of the total varia -bility in PBTL prevalence across sites explained(Table 3) Water motion and parrotfish density showeda positive correlation to variations in PBTL preva-lence whereas turbidity showed a weak negativecorrelation Butterflyfish density chl a spatial tem-perature (difference between sites) salinity and hostcover were not found to be important spatial predic-tors of PBTL prevalence across sites

DISCUSSION

Prevalence and distribution

PBTL was found to be widely distributed on reefswithin Kaneohe Bay but was absent from HanaumaBay which also has a high abundance of the affectedcoral species (Porites compressa) PBTL has also notbeen reported from other reefs within the Main orNorthwestern Hawaiian Islands (Aeby et al 2011a)

64

Fig 3 Porites bleaching with tissue loss (PBTL) prevalenceat Coconut Island Marine Reserve (CIMR) (A) Averageprevalence (plusmnSE) for each permanent site around CIMR (B)Average prevalence over sites (plusmnSE) for each month in 2011with corresponding average temperature (the average over10 d prior to surveying) nd no prevalence data obtained

Area Reef No of colonies minminus1

no of search time

Kaneohe BayNorth Bay 46 16

44 2443 19

Central Bay 31 1227 3024 21

South Bay 14 249 257 08

SE Oahu Hanauma Bay 0

Table 2 Number of Porites bleaching with tissue loss(PBTL)-affected Porites compressa colonies observed (seeFig 1 for reef locations) during rapid surveys within

Kaneohe Bay and Hanauma Bay Hawaii USA

Fig 4 Relationship of average Porites bleaching with tissueloss (PBTL) prevalence (plusmnSE) for each month (February toDecember) with corresponding temperature (the average

over 10 d prior to surveying)

Sudek et al Disease dynamics of Porites bleaching with tissue loss

suggesting that PBTL may be restricted to KaneoheBay

In Hawaii average coral disease prevalence (ex -cluding Porites trematodiasis) is less than 1 (Aebyet al 2011a) which is lower than what was docu-mented for PBTL (average prevalence 15 plusmn 02SE range 0 to 37) Compared to other diseaseswithin Kaneohe Bay PBTL prevalence was higherthan that of MWS (average prevalence 023 plusmn 009SE Aeby et al 2010) but lower than Porites growthanomalies (Por GAs) (average prevalence 217 plusmn83 SE at a particular site Domart-Coulon et al2006) However both MWS and Por GAs have awider range in prevalence across sites withinKaneohe Bay (0 to 29 and 1 to 56 respectively)(Williams et al 2010)

The ecological damage from disease on a host pop-ulation depends on a combination of the spatial dis-tribution prevalence and virulence of the diseaseFor example MWS has a much lower prevalencethan Por GAs but MWS can cause extensive tissueloss and high colony mortality (Aeby et al 2010)whereas Por GAs only result in colony morbidity(reduced growth) (Stimson 2011) PBTL has a rela-tively low prevalence but it can cause extensive tis-sue loss and recovery rates (tissue re-growth) appearto be very slow It was found to be an ephemeral dis-ease with disease signs (speckled bleaching) disap-pearing in most cases within a couple of months anda small proportion of colonies showed disease regres-sion (ie repigmentation) without any signs of tissueloss However a third of the colonies showed signs ofPBTL again after complete cessation of the disease Anumber of other coral diseases have also been foundto reoccur (eg Kuta amp Richardson 1996 Sato et al2009 Aeby et al 2010) and it has been suggestedthat recurrent infections can cause cumulative tissueloss leading to colony mortality and resulting inincreased damage to the reef system over time(Borger amp Steiner 2005) Even though PBTL preva-lence is relatively low a cumulative effect of periodictissue loss could have a negative impact on Poritescompressa-dominated reefs

Transmission

No disease transmission was observed betweenhealthy and PBTL-affected fragments suggestingthat PBTL does not easily transmit via direct contactor the water column (at least over a period of approx-imately 1 to 2 mo) It may be that the environmentalconditions needed for successful transmission werenot replicated by our experimental treatment How-ever direct transmission between touching colonieswas also not observed in the field In contrast othermanipulative experiments have successfully showndisease transmission in aquaria For example MWSwas shown to be transmissible through direct contactin aquarium conditions with direct transmission alsoobserved in the field (Aeby et al 2010) In our modelwe found that host abundance was not an importantfactor in predicting PBTL prevalence The relation-ship between disease prevalence and host abun-dance is a central element in the theory of infectiousdisease ecology (Lloyd-Smith et al 2005) becausetransmission is a key process in hostminuspathogen inter-actions and increased host density can increase theprobability of horizontal transmission of an infectiousdisease (Altizer amp Augustin 1997) As such we suggest that PBTL is either not caused by a pathogenis not highly infectious or that some other variablesuch as a vector may be needed for disease trans mission

Environmental drivers

Variations in turbidity were identified as the over-all strongest predictor of spatial variation in diseaseprevalence (higher PBTL prevalence) across sitesTurbidity showed a weak negative relationship withPBTL prevalence indicating that clearer waters areassociated with higher disease prevalence We alsofound that PBTL prevalence across all sites washighest during the summer months strongly corre-lated with water temperature Increased diseaseprevalence on coral reefs often correlates with ele-

65

AICc Predictor Pseudo-F p variability explained Relationship with prevalence

27494 Turbidity SD 60364 00026 128 NegativeWater motion 37671 00281 90 PositiveParrotfish density SD 21351 01148 44 PositiveTotal 262

Table 3 Summary results of a distance-based linear model (DISTLM) analysis showing the lsquobestrsquo model with the lowestAkaikersquos information criterion with a second-order bias correction (AICc) value and highest amount of variability explained

Sudek et al Disease dynamics of Porites bleaching with tissue loss

suggesting that PBTL may be restricted to KaneoheBay

In Hawaii average coral disease prevalence (ex -cluding Porites trematodiasis) is less than 1 (Aebyet al 2011a) which is lower than what was docu-mented for PBTL (average prevalence 15 plusmn 02SE range 0 to 37) Compared to other diseaseswithin Kaneohe Bay PBTL prevalence was higherthan that of MWS (average prevalence 023 plusmn 009SE Aeby et al 2010) but lower than Porites growthanomalies (Por GAs) (average prevalence 217 plusmn83 SE at a particular site Domart-Coulon et al2006) However both MWS and Por GAs have awider range in prevalence across sites withinKaneohe Bay (0 to 29 and 1 to 56 respectively)(Williams et al 2010)

The ecological damage from disease on a host pop-ulation depends on a combination of the spatial dis-tribution prevalence and virulence of the diseaseFor example MWS has a much lower prevalencethan Por GAs but MWS can cause extensive tissueloss and high colony mortality (Aeby et al 2010)whereas Por GAs only result in colony morbidity(reduced growth) (Stimson 2011) PBTL has a rela-tively low prevalence but it can cause extensive tis-sue loss and recovery rates (tissue re-growth) appearto be very slow It was found to be an ephemeral dis-ease with disease signs (speckled bleaching) disap-pearing in most cases within a couple of months anda small proportion of colonies showed disease regres-sion (ie repigmentation) without any signs of tissueloss However a third of the colonies showed signs ofPBTL again after complete cessation of the disease Anumber of other coral diseases have also been foundto reoccur (eg Kuta amp Richardson 1996 Sato et al2009 Aeby et al 2010) and it has been suggestedthat recurrent infections can cause cumulative tissueloss leading to colony mortality and resulting inincreased damage to the reef system over time(Borger amp Steiner 2005) Even though PBTL preva-lence is relatively low a cumulative effect of periodictissue loss could have a negative impact on Poritescompressa-dominated reefs

Transmission

No disease transmission was observed betweenhealthy and PBTL-affected fragments suggestingthat PBTL does not easily transmit via direct contactor the water column (at least over a period of approx-imately 1 to 2 mo) It may be that the environmentalconditions needed for successful transmission werenot replicated by our experimental treatment How-ever direct transmission between touching colonieswas also not observed in the field In contrast othermanipulative experiments have successfully showndisease transmission in aquaria For example MWSwas shown to be transmissible through direct contactin aquarium conditions with direct transmission alsoobserved in the field (Aeby et al 2010) In our modelwe found that host abundance was not an importantfactor in predicting PBTL prevalence The relation-ship between disease prevalence and host abun-dance is a central element in the theory of infectiousdisease ecology (Lloyd-Smith et al 2005) becausetransmission is a key process in hostminuspathogen inter-actions and increased host density can increase theprobability of horizontal transmission of an infectiousdisease (Altizer amp Augustin 1997) As such we suggest that PBTL is either not caused by a pathogenis not highly infectious or that some other variablesuch as a vector may be needed for disease trans mission

Environmental drivers

Variations in turbidity were identified as the over-all strongest predictor of spatial variation in diseaseprevalence (higher PBTL prevalence) across sitesTurbidity showed a weak negative relationship withPBTL prevalence indicating that clearer waters areassociated with higher disease prevalence We alsofound that PBTL prevalence across all sites washighest during the summer months strongly corre-lated with water temperature Increased diseaseprevalence on coral reefs often correlates with ele-

65

AICc Predictor Pseudo-F p variability explained Relationship with prevalence

27494 Turbidity SD 60364 00026 128 NegativeWater motion 37671 00281 90 PositiveParrotfish density SD 21351 01148 44 PositiveTotal 262

Table 3 Summary results of a distance-based linear model (DISTLM) analysis showing the lsquobestrsquo model with the lowestAkaikersquos information criterion with a second-order bias correction (AICc) value and highest amount of variability explained

Dis Aquat Org 113 59ndash68 2015

vated seawater temperature for example as reportedfor BBD (Boyett et al 2007 Rodriguez amp Croquer2008) some white syndromes (Selig et al 2006Bruno et al 2007 Williams et al 2010 2011a) and afungal disease affecting tropical crustose corallinealgae (Williams et al 2014) Increased temperaturescan lead to an increase in pathogen virulence andorcause stress to the host making it more susceptible todisease (Harvell et al 2007) However water temper-ature is not the only abiotic factor that varies season-ally on reefs For example Sato et al (2011) foundthat high light and elevated seawater temperaturedrive the occurrence of BBD on the GBR They proposed that seasonally increased light levels maybe even more important for inducing new infectionsthan increased water temperature A link to in -creased light could explain the spotty appearance ofPBTL (bleached coenenchyme and pigmented polyps)as Symbiodinium cells may be more shielded in thepolyps because they can retract into the skeletonHowever manipulative experiments are needed toclarify the link between light temperature andPBTL

PBTL prevalence was also correlated with higherwater motion and higher parrotfish densities al -though the link appeared rather weak and is there-fore not discussed further Overall only a quarter ofthe variability in PBTL prevalence could be ex -plained by the measured factors suggesting thatother unmeasured abiotic or biotic factors could bemore important drivers of PBTL prevalence Alterna-tively our predictor variables may not have beencaptured at an appropriate temporal scale with sea-sonal variations in these factors missed this mayhave caused a reduction of the predictive power ofour model Cause of disease is dependent on theintricate interactions between the host environmentand pathogen (Work et al 2008) One can thereforeexpect coral disease spatio-temporal dynamics to behighly complex and to be correlated with multipleand possibly co-interacting environmental drivers(Williams et al 2010)

CONCLUSION

This is the first study examining the disease dy -namics of PBTL in Kaneohe Bay Hawaii PBTLcauses partial colony mortality in the host coralPorites compressa appears to be non-infectious andwas found to have the highest prevalence occurringin the warmer summer months indicating possibleseasonal dynamics Spatial variation in disease

prevalence (higher PBTL prevalence) across siteswas correlated with higher water motion lower tur-bidity and higher parrotfish densities but the modeldid not sufficiently explain the spatial variabilityThis highlights the complex nature of hostminuspatho-genminusenvironment interactions and the need forinvestigating and understanding coral disease ecol-ogy Further research into the causative agent andlinks to environmental drivers specifically at a finertemporal scale are needed to better understand thedynamics of this disease Porites compressa is amongthe main framework-building corals in Kaneohe Bayand so chronic recurrent diseases such as PBTLcould have a negative impact on the health andstructure of these reefs

Acknowledgements We thank all field assistants for theirdedicated help Timothy Jones for advice on the GLM andJamie Sziklay for assistance in constructing the site mapMS was supported by a VUW PhD Scholarship Coral col-lection was authorized under Special Activity Permit (SAP)2011-67

LITERATURE CITED

Aeby GS (2007) Spatial and temporal patterns of Poritestrematodiasis on the reefs of Kaneohe Bay OahuHawaii Bull Mar Sci 80 209minus218

Aeby GS Santavy DL (2006) Factors affecting susceptibilityof the coral Montastraea faveolata to black-band diseaseMar Ecol Prog Ser 318 103minus110

Aeby GS Ross M Williams GJ Lewis TD Work TM (2010)Disease dynamics of Montipora white syndrome withinKaneohe Bay Oahu Hawaii distribution seasonalityvirulence and transmissibility Dis Aquat Org 91 1minus8

Aeby GS Williams GJ Franklin EC Kenyon J Cox EFColes S Work TM (2011a) Patterns of coral diseaseacross the Hawaiian Archipelago relating disease toenvironment PLoS ONE 6 e20370

Aeby GS Williams GJ Franklin EC Haapkylauml J and others(2011b) Growth anomalies on the coral genera Acroporaand Porites are strongly associated with host density andhuman population size across the Indo-Pacific PLoSONE 6 e16887

Akaike H (1973) Information theory as an extension of themaximum likelihood principal Proceedings 2nd Interna-tional Symposium of Information Theory AkademicKiado Budapest p 261minus281

Altizer SM Augustin DJ (1997) Interactions between fre-quency-dependent and vertical transmission in host-par-asite systems Proc R Soc Lond B Biol Sci 264 807minus814

Anderson MJ Gorley RN Clarke KR (2008) PERMANOVA+for PRIMER guide to software and statistical methodsPRIMER-E Plymouth

Aronson RB Precht WF (2001) White-band disease and thechanging face of Caribbean coral reefs Hydrobiologia460 25minus38

Borger JL Steiner SCC (2005) The spatial and temporaldynamics of coral diseases in Dominica West Indies BullMar Sci 77 137minus154

Boyett HV Bourne DG Willis BL (2007) Elevated tempera-

66

Sudek et al Disease dynamics of Porites bleaching with tissue loss

ture and light enhance progression and spread of blackband disease on staghorn corals of the Great BarrierReef Mar Biol 151 1711minus1720

Bruckner AW Hill RL (2009) Ten years of change to coralcommunities off Mona and Desecheo Islands PuertoRico from disease and bleaching Dis Aquat Org 87 19minus31

Bruckner AW Bruckner RJ Williams EH (1997) Spread of ablack-band disease epizootic through the coral reef sys-tem in St Annrsquos Bay Jamaica Bull Mar Sci 61 919minus928

Bruno JF Petes LE Harvell CD Hettinger A (2003) Nutrientenrichment can increase the severity of coral diseasesEcol Lett 6 1056minus1061

Bruno JF Selig ER Casey KS Page CA and others (2007)Thermal stress and coral cover as drivers of coral diseaseoutbreaks PLoS Biol 5 e124

Burnham KP Anderson DR (2004) Multimodel inferencemdashunderstanding AIC and BIC in model selection SociolMethods Res 33 261minus304

Clarke KR Gorley RN (2006) PRIMER v6 user manualtuto-rial Primer-E Plymouth

Clarke KR Somerfield PJ Chapman MG (2006) On resem-blance measures for ecological studies including taxo-nomic dissimilarities and a zero-adjusted Bray-Curtiscoefficient for denuded assemblages J Exp Mar Biol Ecol330 55minus80

Croquer A Weil E Zubillaga AL Pauls SM (2005) Impact ofa white plague-II outbreak on a coral reef in the archipel-ago Los Roques National Park Venezuela Caribb J Sci41 815minus823

Domart-Coulon IJ Traylor-Knowles N Peters E Elbert Dand others (2006) Comprehensive characterization ofskeletal tissue growth anomalies of the finger coralPorites compressa Coral Reefs 25 531minus543

Edmunds PJ (1991) Extent and effect of black band diseaseon a Caribbean reef Coral Reefs 10 161minus165

Gladfelter WB (1982) White band disease in Acroporapalmatamdashimplications for the structure and growth ofshallow reefs Bull Mar Sci 32 639minus643

Haapkylauml J Seymour AS Trebilco J Smith D (2007) Coraldisease prevalence and coral health in the WakatobiMarine Park south-east Sulawesi Indonesia J Mar BiolAssoc UK 87 403minus414

Harvell CD Kim K Burkholder JM Colwell RR and others(1999) Emerging marine diseasesmdashclimate links andanthropogenic factors Science 285 1505minus1510

Harvell CD Mitchell CE Ward JR Altizer S Dobson APOstfeld RS Samuel MD (2002) Climate warming and dis-ease risks for terrestrial and marine biota Science 296 2158minus2162

Harvell D Aronson R Baron N Connell J and others (2004)The rising tide of ocean diseases unsolved problems andresearch priorities Front Ecol Environ 2 375minus382

Harvell D Jordan-Dahlgren E Merkel S Rosenberg E andothers (2007) Coral disease environmental drivers andthe balance between coral and microbial associatesOceanography 20 172minus195

Hurvich CM Tsai CL (1989) Regression and time-seriesmodel selection in small samples Biometrika 76 297minus307

Jokiel PL Coles SL (1990) Response of Hawaiian and otherIndo-Pacific reef corals to elevated temperature CoralReefs 8 155minus162

Jokiel PL Morrissey JI (1993) Water motion on coral reefs evaluation of the lsquoclod cardrsquo technique Mar Ecol ProgSer 93 175minus181

Kaczmarsky LT (2006) Coral disease dynamics in the centralPhilippines Dis Aquat Org 69 9minus21

Kuta KG Richardson LL (1996) Abundance and distributionof black band disease on coral reefs in the northernFlorida Keys Coral Reefs 15 219minus223

Lafferty KD Kuris AM (1993) Mass mortality of abaloneHaliotis cracherodii on the California Channel Islands tests of epidemiological hypotheses Mar Ecol Prog Ser96 239minus248

Lafferty KD Porter JW Ford SE (2004) Are diseases increas-ing in the ocean Annu Rev Ecol Evol Syst 35 31minus54

Lessios HA (1988) Mass mortality of Diadema antillarum inthe Caribbean What have we learned Annu Rev EcolSyst 19 371minus393

Lloyd-Smith JO Cross PC Briggs CJ Daugherty M and oth-ers (2005) Should we expect population thresholds forwildlife disease Trends Ecol Evol 20 511minus519

McArdle BH Anderson MJ (2001) Fitting multivariate mod-els to community data a comment on distance-basedredundancy analysis Ecology 82 290minus297

McClanahan TR Weil E Maina J (2009) Strong relationshipbetween coral bleaching and growth anomalies in mas-sive Porites Glob Change Biol 15 1804minus1816

Nugues MM (2002) Impact of a coral disease outbreak oncoral communities in St Lucia What and how much hasbeen lost Mar Ecol Prog Ser 229 61minus71

Page C Willis B (2006) Distribution host range and large-scale spatial variability in black band disease prevalenceon the Great Barrier Reef Australia Dis Aquat Org 69 41minus51

Page CA Willis BL (2008) Epidemiology of skeletal erodingband on the Great Barrier Reef and the role of injury inthe initiation of this widespread coral disease CoralReefs 27257ndash272

Patterson KL Porter JW Ritchie KE Polson SW and others(2002) The etiology of white pox a lethal disease of theCaribbean elkhorn coral Acropora palmata Proc NatlAcad Sci USA 99 8725minus8730

Raymundo LJ Rosell KB Reboton CT Kaczmarsky L (2005)Coral diseases on Philippine reefs genus Porites is adominant host Dis Aquat Org 64 181minus191

Raymundo LJ Halford AR Maypa AP Kerr AM (2009)Functionally diverse reef fish communities amelioratecoral disease Proc Natl Acad Sci USA 106 17067minus17070

Rodriguez S Croquer A (2008) Dynamics of black band dis-ease in a Diploria strigosa population subjected to annualupwelling on the northeastern coast of Venezuela CoralReefs 27 381minus388

Roff G Hoegh-Guldberg O Fine M (2006) Intra-colonialresponse to Acroporid lsquowhite syndromersquo lesions in tabu-lar Acropora spp (Scleractinia) Coral Reefs 25 255minus264

Roy KJ (1970) Change in bathymetric configurationKaneohe Bay Oahu 1882ndash1969 Hawaii Institute of Geo-physics University of Hawaii Honolulu HI

Sato Y Bourne DG Willis BL (2009) Dynamics of seasonaloutbreaks of black band disease in an assemblage ofMontipora species at Pelorus Island (Great Barrier ReefAustralia) Proc R Soc Lond B Biol Sci 276 2795minus2803

Sato Y Bourne DG Willis BL (2011) Effects of temperatureand light on the progression of black band disease on thereef coral Montipora hispida Coral Reefs 30 753minus761

Selig E Harvell C Bruno J Willis B Page C Casey KSweatman H (2006) Analyzing the relationship betweenocean temperature anomalies and coral disease out-breaks at broad spatial scales In Phinney J Hoegh-

67

Dis Aquat Org 113 59ndash68 2015

Guldberg O Kleypas J Skirving W Strong A (eds) Coralreefs and climate change science and managementCoastal and estuarine studies American GeophysicalUnion Washington DC p 111minus128

Sokolow S (2009) Effects of a changing climate on thedynamics of coral infectious disease a review of the evidence Dis Aquat Org 87 5minus18

Stimson J (2011) Ecological characterization of coral growthanomalies on Porites compressa in Hawailsquoi Coral Reefs30 133minus142

Sudek M Aeby GS Davy SK (2012a) Localized bleaching inHawaii causes tissue loss and a reduction in the numberof gametes in Porites compressa Coral Reefs 31 351minus355

Sudek M Work TM Aeby GS Davy SK (2012b) Histologicalobservations in the Hawaiian reef coral Porites com-pressa affected by Porites bleaching with tissue loss(PBTL) J Invertebr Pathol 111 121minus125

Sutherland KP Porter JW Torres C (2004) Disease andimmunity in Caribbean and Indo-Pacific zooxanthellatecorals Mar Ecol Prog Ser 266 273minus302

Vargas-Angel B (2009) Coral health and disease assessmentin the US Pacific remote island areas Bull Mar Sci 84 211minus227

Ward JR Lafferty KD (2004) The elusive baseline of marinedisease Are diseases in ocean ecosystems increasingPLoS Biol 2 e120

Williams GJ Davy SK Aeby GS (2008) Coral disease atPalmyra Atoll a remote reef system in the CentralPacific Coral Reefs 27 207

Williams GJ Aeby GS Cowie ROM Davy SK (2010) Predic-tive modeling of coral disease distribution within a reefsystem PLoS ONE 5 e9264

Williams GJ Knapp IS Work TM Conklin EJ (2011a) Out-break of Acropora white syndrome following a mildbleaching event at Palmyra Atoll Northern Line IslandsCentral Pacific Coral Reefs 30 621

Williams GJ Knapp IS Aeby GS Davy SK (2011b) Spatialand temporal patterns of scleractinian coral soft coraland zoanthid disease on a remote near-pristine coralreef (Palmyra Atoll central Pacific) Dis Aquat Org 94 89minus100

Williams GJ Price NN Ushijima B Aeby GS and others(2014) Ocean warming and acidification have complexinteractive effects on the dynamics of a marine fungaldisease Proc R Soc Lond B Biol Sci 281 20133069

Willis BL Page CA Dinsdale EA (2004) Coral disease on theGreat Barrier Reef In Rosenberg E Loya Y (eds) Coralhealth and disease Springer Berlin p 69minus104

Work TM Richardson LL Reynolds TL Willis BL (2008)Biomedical and veterinary science can increase ourunderstanding of coral disease J Exp Mar Biol Ecol 362 63minus70

68

Editorial responsibility Garriet SmithAiken South Carolina USA

Submitted June 3 2014 Accepted November 11 2014Proofs received from author(s) January 28 2015